Re: [libvirt PATCH v2 0/1] meson: Improve nbdkit configurability
On 10/5/23 4:22 AM, Andrea Bolognani wrote: Changes from [v1]: * disable nbdkit on anything older than Fedora 40 in the RPM. [v1] https://listman.redhat.com/archives/libvir-list/2023-October/242498.html Andrea Bolognani (1): meson: Improve nbdkit configurability libvirt.spec.in| 28 +--- meson.build| 29 + meson_options.txt | 2 +- src/qemu/qemu_nbdkit.c | 6 +++--- 4 files changed, 50 insertions(+), 15 deletions(-) Reviewed-by: Jonathon Jongsma
Re: [libvirt PATCH] qemu: DomainGetGuestVcpusParams: reduce scope of tmp
On 9/29/23 8:48 AM, Ján Tomko wrote:
Wrap the macro body in a new block and move the declaration of 'tmp'
into it, to avoid the need to mix g_autofree with manual freeing.
Signed-off-by: Ján Tomko
---
src/qemu/qemu_driver.c | 13 +++--
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 056b5cec98..f3c9730cd8 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -18396,7 +18396,6 @@ qemuDomainGetGuestVcpusParams(virTypedParameterPtr
*params,
g_autoptr(virBitmap) vcpus = virBitmapNew(QEMU_GUEST_VCPU_MAX_ID);
g_autoptr(virBitmap) online = virBitmapNew(QEMU_GUEST_VCPU_MAX_ID);
g_autoptr(virBitmap) offlinable = virBitmapNew(QEMU_GUEST_VCPU_MAX_ID);
-g_autofree char *tmp = NULL;
size_t i;
int ret = -1;
@@ -18416,11 +18415,13 @@ qemuDomainGetGuestVcpusParams(virTypedParameterPtr *params,
}
#define ADD_BITMAP(name) \
-if (!(tmp = virBitmapFormat(name))) \
-goto cleanup; \
-if (virTypedParamsAddString(, , , #name, tmp) < 0) \
-goto cleanup; \
-VIR_FREE(tmp)
+do { \
+g_autofree char *tmp = NULL; \
+if (!(tmp = virBitmapFormat(name))) \
+goto cleanup; \
+if (virTypedParamsAddString(, , , #name, tmp) < 0) \
+goto cleanup; \
+} while (0)
ADD_BITMAP(vcpus);
ADD_BITMAP(online);
Reviewed-by: Jonathon Jongsma
Re: [libvirt PATCH 0/2] Add vdpablock and nbdkit to NEWS
On 9/19/23 3:47 PM, Jonathon Jongsma wrote: Jonathon Jongsma (2): news: document support for vdpa block devices news: document nbdkit support for network disks NEWS.rst | 18 ++ 1 file changed, 18 insertions(+) ping
[libvirt PATCH v2] util: fix success return for virProcessKillPainfullyDelay()
virProcessKillPainfullyDelay() currently almost always returns 1 or -1, even though the documentation indicates that it should return 0 if the process was terminated gracefully. But the computation of the return code is faulty and the only case that it currently returns 0 is when it is called with the pid of a process that does not exist. Since no callers ever even distinguish between the 0 and 1 response codes, simply get rid of the distinction and return 0 for both cases. Signed-off-by: Jonathon Jongsma --- Change in v2: - just drop the distinction between 0 and 1 and always return 0. Suggested by Ján Tomko src/util/virprocess.c | 7 +++ 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/src/util/virprocess.c b/src/util/virprocess.c index 6ce5ef99a9..b6fb17db83 100644 --- a/src/util/virprocess.c +++ b/src/util/virprocess.c @@ -363,9 +363,8 @@ pid_t virProcessGroupGet(pid_t pid) /* * Try to kill the process and verify it has exited * - * Returns 0 if it was killed gracefully, 1 if it - * was killed forcibly, -1 if it is still alive, - * or another error occurred. + * Returns 0 if it was killed, -1 if it is still alive or another error + * occurred. * * Callers can provide an extra delay in seconds to * wait longer than the default. @@ -426,7 +425,7 @@ virProcessKillPainfullyDelay(pid_t pid, bool force, unsigned int extradelay, boo (long long)pid, signame); return -1; } -return signum == SIGTERM ? 0 : 1; +return 0; } g_usleep(200 * 1000); -- 2.41.0
[libvirt PATCH 2/2] util: fix success return for virProcessKillPainfullyDelay()
virProcessKillPainfullyDelay() currently almost always returns 1 or -1,
even though the documentation indicates that it should return 0 if the
process was terminated gracefully. The only case that it currently
returns 0 is when it is called with the pid of a process that does not
exist.
This function initially sends a SIGTERM signal to the process. If
the signal was sent successfully (i.e. virProcessKill() returns 0) we
immediately sleep for 200ms before iterating the loop. On the second
iteration of the loop, signum will be set to 0 and we call
virProcessKill() again to check whether the process still exists. If the
process does not exist (virProcessKill() returns -1 and errno is ESRCH),
we consider the process killed successfully. But in order to determine
what value to return from the function, we compare the value of signum
to SIGTERM. But signum has already been set to 0; it will only ever be
SIGTERM in the very first loop iteration. So it always indicates a
forcible killing even when it is killed gracefully.
Signed-off-by: Jonathon Jongsma
---
src/util/virprocess.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/util/virprocess.c b/src/util/virprocess.c
index 6ce5ef99a9..42b3b76eaa 100644
--- a/src/util/virprocess.c
+++ b/src/util/virprocess.c
@@ -377,6 +377,7 @@ virProcessKillPainfullyDelay(pid_t pid, bool force,
unsigned int extradelay, boo
/* This is in 1/5th seconds since polling is on a 0.2s interval */
unsigned int polldelay = (force ? 200 : 75) + (extradelay*5);
const char *signame = "TERM";
+int forced = 0;
VIR_DEBUG("vpid=%lld force=%d extradelay=%u group=%d",
(long long)pid, force, extradelay, group);
@@ -399,6 +400,7 @@ virProcessKillPainfullyDelay(pid_t pid, bool force,
unsigned int extradelay, boo
if (i == 0) {
signum = SIGTERM; /* kindly suggest it should exit */
} else if (i == 50 && force) {
+forced = 1;
VIR_DEBUG("Timed out waiting after SIGTERM to process %lld, "
"sending SIGKILL", (long long)pid);
/* No SIGKILL kill on Win32 ! Use SIGABRT instead which our
@@ -426,7 +428,7 @@ virProcessKillPainfullyDelay(pid_t pid, bool force,
unsigned int extradelay, boo
(long long)pid, signame);
return -1;
}
-return signum == SIGTERM ? 0 : 1;
+return forced;
}
g_usleep(200 * 1000);
--
2.41.0
[libvirt PATCH 1/2] util: Fix error return for virProcessKillPainfullyDelay()
Commit 93af79fb removed a cleanup label in favor of returning error
values directly in certain cases. But the final return value was changed
from -1 to 0. If we get to the end of the function, that means that
we've waited for the process to exit but it still exists. So we should
return -1. The error message was still being set correctly, but we were
returning a success status (0).
Signed-off-by: Jonathon Jongsma
---
src/util/virprocess.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/util/virprocess.c b/src/util/virprocess.c
index 31b833e5c8..6ce5ef99a9 100644
--- a/src/util/virprocess.c
+++ b/src/util/virprocess.c
@@ -436,7 +436,7 @@ virProcessKillPainfullyDelay(pid_t pid, bool force,
unsigned int extradelay, boo
_("Failed to terminate process %1$lld with SIG%2$s"),
(long long)pid, signame);
-return 0;
+return -1;
}
--
2.41.0
Re: [libvirt PATCH v8 10/37] qemu: add functions to start and stop nbdkit
On 9/21/23 1:10 PM, Peter Krempa wrote:
On Thu, Sep 21, 2023 at 12:50:52 -0500, Jonathon Jongsma wrote:
On 9/20/23 7:24 AM, Pavel Hrdina wrote:
On Thu, Aug 31, 2023 at 04:39:50PM -0500, Jonathon Jongsma wrote:
Add some helper functions to build a virCommand object and run the
nbdkit process for a given virStorageSource.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 250 +
src/qemu/qemu_nbdkit.h | 10 ++
2 files changed, 260 insertions(+)
[...]
+VIR_DEBUG("Stopping nbdkit process %i", proc->pid);
+virProcessKill(proc->pid, SIGTERM);
Coverity complains here that the return value of virProcessKill() is not
checked which leads me to a question if we should use
virProcessKillPainfully() instead.
With the code that is pushed the function qemuNbdkitProcessStop() is
called only within the qemu_nbdkit.c for these cases:
- in qemuNbdkitProcessRestart() before starting the process again
where we do not check if the original process was killed correctly
or not,
- in qemuNbdkitStopStorageSource() where we check return value of
qemuNbdkitProcessStop() but it will always be 0,
- in qemuNbdkitProcessStart() as error path where we don't check any
return value.
To me it seems that the return value qemuNbdkitProcessStop can be changed
to void as we always return 0 and use virProcessKillPainfully() or
properly pass return value of virProcessKill() and check it for every
use of qemuNbdkitProcessStop().
Pavel
Good question. In one of my earlier series I had actually used
virProcessKillPainfully(), but changed it based on a suggestion from Peter
that it would be bad to kill it painfully if nbdkit was ever used in
read-write mode. But apparently I forgot to handle a shutdown failure. An
alternative would be to simply refuse to use nbdkit if the user requests
read-write mode.
We can use the same algorithm as with the qemu process where we first
issue SIGTERM, thus if the process is responsive it can execute the
shutdown actions. Otherwise it'll get SIGKILL right after.
That sounds almost identical to what virProcessKillPainfully() does.
Jonathon
Re: [libvirt PATCH v8 10/37] qemu: add functions to start and stop nbdkit
On 9/20/23 7:24 AM, Pavel Hrdina wrote:
On Thu, Aug 31, 2023 at 04:39:50PM -0500, Jonathon Jongsma wrote:
Add some helper functions to build a virCommand object and run the
nbdkit process for a given virStorageSource.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 250 +
src/qemu/qemu_nbdkit.h | 10 ++
2 files changed, 260 insertions(+)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 9a2a89224d..6bf962d0f1 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -24,6 +24,8 @@
#include "virerror.h"
#include "virlog.h"
#include "virpidfile.h"
+#include "virsecureerase.h"
+#include "virtime.h"
#include "virutil.h"
#include "qemu_block.h"
#include "qemu_conf.h"
@@ -666,6 +668,168 @@ qemuNbdkitInitStorageSource(qemuNbdkitCaps *caps,
}
+static int
+qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
+ virCommand *cmd)
+{
+g_autoptr(virURI) uri = qemuBlockStorageSourceGetURI(proc->source);
+g_autofree char *uristring = virURIFormat(uri);
+
+/* nbdkit plugin name */
+virCommandAddArg(cmd, "curl");
+if (proc->source->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP) {
+/* allow http to be upgraded to https via e.g. redirect */
+virCommandAddArgPair(cmd, "protocols", "http,https");
+} else {
+virCommandAddArgPair(cmd, "protocols",
+
virStorageNetProtocolTypeToString(proc->source->protocol));
+}
+virCommandAddArgPair(cmd, "url", uristring);
+
+if (proc->source->auth) {
+g_autoptr(virConnect) conn = virGetConnectSecret();
+g_autofree uint8_t *secret = NULL;
+size_t secretlen = 0;
+g_autofree char *password = NULL;
+int secrettype;
+virStorageAuthDef *authdef = proc->source->auth;
+
+virCommandAddArgPair(cmd, "user",
+ proc->source->auth->username);
+
+if ((secrettype =
virSecretUsageTypeFromString(proc->source->auth->secrettype)) < 0) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("invalid secret type %1$s"),
+ proc->source->auth->secrettype);
+return -1;
+}
+
+if (virSecretGetSecretString(conn,
+ >seclookupdef,
+ secrettype,
+ ,
+ ) < 0) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("failed to get auth secret for storage"));
+return -1;
+}
+
+/* ensure that the secret is a NULL-terminated string */
+password = g_strndup((char*)secret, secretlen);
+virSecureErase(secret, secretlen);
+
+/* for now, just report an error rather than passing the password in
+ * cleartext on the commandline */
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("Password not yet supported for nbdkit sources"));
+
+virSecureEraseString(password);
+
+return -1;
+}
+
+if (proc->source->ncookies > 0) {
+/* for now, just report an error rather than passing cookies in
+ * cleartext on the commandline */
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("Cookies not yet supported for nbdkit sources"));
+return -1;
+}
+
+if (proc->source->sslverify == VIR_TRISTATE_BOOL_NO) {
+virCommandAddArgPair(cmd, "sslverify", "false");
+}
+
+if (proc->source->timeout > 0) {
+g_autofree char *timeout = g_strdup_printf("%llu",
proc->source->timeout);
+virCommandAddArgPair(cmd, "timeout", timeout);
+}
+
+return 0;
+}
+
+
+static int
+qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
+ virCommand *cmd)
+{
+const char *user = NULL;
+virStorageNetHostDef *host = >source->hosts[0];
+g_autofree char *portstr = g_strdup_printf("%u", host->port);
+
+/* nbdkit plugin name */
+virCommandAddArg(cmd, "ssh");
+
+virCommandAddArgPair(cmd, "host", host->name);
+virCommandAddArgPair(cmd, "port", portstr);
+virCommandAddArgPair(cmd, "path", proc->source->path);
+
+if (proc->source->auth)
+user = proc->source->auth->username;
+else if (proc->source->ssh_user)
+user = proc->source-&
[libvirt PATCH 2/2] news: document nbdkit support for network disks
Signed-off-by: Jonathon Jongsma --- I put this under 'Improvements' since it's not really intended to be a user-visible feature, but it could be moved to 'New Features' instead... NEWS.rst | 9 + 1 file changed, 9 insertions(+) diff --git a/NEWS.rst b/NEWS.rst index b57ba5c7d1..768e6fc326 100644 --- a/NEWS.rst +++ b/NEWS.rst @@ -34,6 +34,15 @@ v9.8.0 (unreleased) * **Improvements** + * qemu: add nbdkit backend for network disks + +Up until now, libvirt supported network disks (http, ftp, ssh) by passing +the URL to qemu and having the appropriate qemu block drivers handle the +disk I/O. However, by handling the network I/O outside of the qemu process, +we get several advantages, such as reduced attack surface and improved +stability of qemu. Therefore, when available, libvirt will use nbdkit as a +backend for these network disks and export an NBD disk to qemu. + * **Bug fixes** -- 2.41.0
Re: [libvirt PATCH v8 00/37] Use nbdkit for http/ftp/ssh network drives in libvirt
On 9/19/23 2:44 AM, Peter Krempa wrote: On Thu, Aug 31, 2023 at 16:39:40 -0500, Jonathon Jongsma wrote: [...] Jonathon Jongsma (37): schema: allow 'ssh' as a protocol for network disks qemu: Add functions for determining nbdkit availability qemu: expand nbdkit capabilities util: Allow virFileCache data to be any GObject qemu: implement basic virFileCache for nbdkit caps qemu: implement persistent file cache for nbdkit caps qemu: use file cache for nbdkit caps qemu: Add qemuNbdkitProcess qemu: query nbdkit module dir from binary qemu: add functions to start and stop nbdkit Generalize qemuDomainLogContextNew() qemu: Extract qemuDomainLogContext into a new file qemu: move qemuProcessReadLog() to qemuLogContext qemu: log error output from nbdkit tests: add ability to test various nbdkit capabilities qemu: split qemuDomainSecretStorageSourcePrepare qemu: include nbdkit state in private xml util: secure erase virCommand send buffers qemu: pass sensitive data to nbdkit via pipe qemu: use nbdkit to serve network disks if available util: make virCommandSetSendBuffer testable tests: add tests for nbdkit invocation qemu: add test for authenticating a https network disk qemu: Add Taint for nbdkit restart failure qemu: Monitor nbdkit process for exit qemu: improve error handling when restarting nbdkit qemu: try to connect to nbdkit early to detect errors schema: add password configuration for ssh disk qemu: implement password auth for ssh disks with nbdkit schema: add configuration for host verification of ssh disks qemu: implement knownHosts for ssh disks with nbdkit schema: add keyfile configuration for ssh disks qemu: implement keyfile auth for ssh disks with nbdkit schema: add ssh-agent configuration for ssh disks qemu: implement ssh-agent auth for ssh disks with nbdkit Please push these patches. I plan working on qemu storage daemon integration which has the potential to conflict with this. rpm: update spec file for for nbdkit support This one can be omitted if you're still waiting for the selinux policy. ci: add libnbd to build This one probably should work even now. Sorry for the delay. It should be pushed now. Jonathon
[libvirt PATCH 1/2] news: document support for vdpa block devices
Signed-off-by: Jonathon Jongsma --- NEWS.rst | 9 + 1 file changed, 9 insertions(+) diff --git a/NEWS.rst b/NEWS.rst index 926620b03f..b57ba5c7d1 100644 --- a/NEWS.rst +++ b/NEWS.rst @@ -23,6 +23,15 @@ v9.8.0 (unreleased) * **New features** + * qemu: Add support for vDPA block devices + +With a new enough version of qemu, libvirt will allow you to assign vDPA block +devices to a domain. This is configured with:: + + + +... + * **Improvements** * **Bug fixes** -- 2.41.0
[libvirt PATCH 0/2] Add vdpablock and nbdkit to NEWS
Jonathon Jongsma (2): news: document support for vdpa block devices news: document nbdkit support for network disks NEWS.rst | 18 ++ 1 file changed, 18 insertions(+) -- 2.41.0
Re: [libvirt PATCH v2 0/5] Add support for vDPA block devices
On 9/12/23 7:00 AM, Peter Krempa wrote: On Mon, Sep 11, 2023 at 16:53:42 -0500, Jonathon Jongsma wrote: see https://bugzilla.redhat.com/show_bug.cgi?id=1900770. Changes in v2: - Don't use virStorageSource->path for vdpa device path to avoid clashing with existing path functionality - Move vdpa device opening to the qemuProcessPrepareHostStorageSource() function rather than the qemuDomainPrepareStorageSource() function. This also required some additional support in the tests for setting up the objects properly for testing. - rebased to latest master branch Reviewed-by: Peter Krempa I pushed this series, but for some reason only the ubuntu images are failing CI with no useful output: https://gitlab.com/libvirt/libvirt/-/pipelines/1001459836 I suspect it may be related to address sanitizer stuff, does anybody have tips on getting more information about this failure? Jonathon
[libvirt PATCH v2 2/5] qemu: add virtio-blk-vhost-vdpa capability
Check whether the qemu binary supports the vdpa block driver. We can't
rely simply on the existence of the virtio-blk-vhost-vdpa block driver
since the first releases of qemu didn't support fd-passing for this
driver. So we have to check for the 'fdset' feature on the driver
object. This feature will be present in the qemu 8.1.0 release and was
merged to qemu in commit 98b126f5.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_capabilities.c | 2 ++
src/qemu/qemu_capabilities.h | 1 +
tests/qemucapabilitiesdata/caps_8.1.0_x86_64.xml | 1 +
3 files changed, 4 insertions(+)
diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c
index 87412dd4ec..3a1bfbf74d 100644
--- a/src/qemu/qemu_capabilities.c
+++ b/src/qemu/qemu_capabilities.c
@@ -697,6 +697,7 @@ VIR_ENUM_IMPL(virQEMUCaps,
/* 450 */
"run-with.async-teardown", /* QEMU_CAPS_RUN_WITH_ASYNC_TEARDOWN
*/
+ "virtio-blk-vhost-vdpa", /*
QEMU_CAPS_DEVICE_VIRTIO_BLK_VHOST_VDPA */
);
@@ -1531,6 +1532,7 @@ static struct virQEMUCapsStringFlags
virQEMUCapsQMPSchemaQueries[] = {
{ "blockdev-add/arg-type/+rbd/encrypt/format/^luks-any",
QEMU_CAPS_RBD_ENCRYPTION_LUKS_ANY },
{ "blockdev-add/arg-type/+nbd/tls-hostname",
QEMU_CAPS_BLOCKDEV_NBD_TLS_HOSTNAME },
{ "blockdev-add/arg-type/+qcow2/discard-no-unref",
QEMU_CAPS_QCOW2_DISCARD_NO_UNREF },
+{ "blockdev-add/arg-type/+virtio-blk-vhost-vdpa/$fdset",
QEMU_CAPS_DEVICE_VIRTIO_BLK_VHOST_VDPA},
{ "blockdev-snapshot/$allow-write-only-overlay",
QEMU_CAPS_BLOCKDEV_SNAPSHOT_ALLOW_WRITE_ONLY },
{ "chardev-add/arg-type/backend/+socket/data/reconnect",
QEMU_CAPS_CHARDEV_RECONNECT },
{ "device_add/$json-cli-hotplug", QEMU_CAPS_DEVICE_JSON },
diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h
index e51d3fffdc..3c4f7f625b 100644
--- a/src/qemu/qemu_capabilities.h
+++ b/src/qemu/qemu_capabilities.h
@@ -676,6 +676,7 @@ typedef enum { /* virQEMUCapsFlags grouping marker for
syntax-check */
/* 450 */
QEMU_CAPS_RUN_WITH_ASYNC_TEARDOWN, /* asynchronous teardown -run-with
async-teardown=on|off */
+QEMU_CAPS_DEVICE_VIRTIO_BLK_VHOST_VDPA, /* virtio-blk-vhost-vdpa block
driver */
QEMU_CAPS_LAST /* this must always be the last item */
} virQEMUCapsFlags;
diff --git a/tests/qemucapabilitiesdata/caps_8.1.0_x86_64.xml
b/tests/qemucapabilitiesdata/caps_8.1.0_x86_64.xml
index 6f8c5a57b7..d266dd0f31 100644
--- a/tests/qemucapabilitiesdata/caps_8.1.0_x86_64.xml
+++ b/tests/qemucapabilitiesdata/caps_8.1.0_x86_64.xml
@@ -197,6 +197,7 @@
+
8001000
43100245
v8.1.0
--
2.41.0
[libvirt PATCH v2 3/5] qemu: make vdpa connect function more generic
qemuInterfaceVDPAConnect() was a helper function for connecting to the
vdpa device file. But in order to support other vdpa devices besides
network interfaces (e.g. vdpa block devices) make this function a bit
more generic.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_command.c | 23 ++-
src/qemu/qemu_command.h | 1 +
src/qemu/qemu_interface.c | 23 ---
src/qemu/qemu_interface.h | 2 --
tests/qemuhotplugmock.c | 4 ++--
tests/qemuxml2argvmock.c | 2 +-
6 files changed, 26 insertions(+), 29 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index 778958700b..e84374b4cf 100644
--- a/src/qemu/qemu_command.c
+++ b/src/qemu/qemu_command.c
@@ -8533,7 +8533,7 @@ qemuBuildInterfaceConnect(virDomainObj *vm,
break;
case VIR_DOMAIN_NET_TYPE_VDPA:
-if ((vdpafd = qemuInterfaceVDPAConnect(net)) < 0)
+if ((vdpafd = qemuVDPAConnect(net->data.vdpa.devicepath)) < 0)
return -1;
netpriv->vdpafd = qemuFDPassNew(net->info.alias, priv);
@@ -10993,3 +10993,24 @@
qemuBuildStorageSourceChainAttachPrepareBlockdevTop(virStorageSource *top,
return g_steal_pointer();
}
+
+
+/* qemuVDPAConnect:
+ * @devicepath: the path to the vdpa device
+ *
+ * returns: file descriptor of the vdpa device
+ */
+int
+qemuVDPAConnect(const char *devicepath)
+{
+int fd;
+
+if ((fd = open(devicepath, O_RDWR)) < 0) {
+virReportSystemError(errno,
+ _("Unable to open '%1$s' for vdpa device"),
+ devicepath);
+return -1;
+}
+
+return fd;
+}
diff --git a/src/qemu/qemu_command.h b/src/qemu/qemu_command.h
index 55efa45601..341ec43f9a 100644
--- a/src/qemu/qemu_command.h
+++ b/src/qemu/qemu_command.h
@@ -248,3 +248,4 @@ qemuBuildTPMOpenBackendFDs(const char *tpmdev,
const char * qemuAudioDriverTypeToString(virDomainAudioType type);
virDomainAudioType qemuAudioDriverTypeFromString(const char *str);
+int qemuVDPAConnect(const char *devicepath) G_NO_INLINE;
diff --git a/src/qemu/qemu_interface.c b/src/qemu/qemu_interface.c
index e875de48ee..8856bb95a8 100644
--- a/src/qemu/qemu_interface.c
+++ b/src/qemu/qemu_interface.c
@@ -648,29 +648,6 @@ qemuInterfaceBridgeConnect(virDomainDef *def,
}
-/* qemuInterfaceVDPAConnect:
- * @net: pointer to the VM's interface description
- *
- * returns: file descriptor of the vdpa device
- *
- * Called *only* called if actualType is VIR_DOMAIN_NET_TYPE_VDPA
- */
-int
-qemuInterfaceVDPAConnect(virDomainNetDef *net)
-{
-int fd;
-
-if ((fd = open(net->data.vdpa.devicepath, O_RDWR)) < 0) {
-virReportSystemError(errno,
- _("Unable to open '%1$s' for vdpa device"),
- net->data.vdpa.devicepath);
-return -1;
-}
-
-return fd;
-}
-
-
/*
* Returns: -1 on error, 0 on success. Populates net->privateData->slirp if
* the slirp helper is needed.
diff --git a/src/qemu/qemu_interface.h b/src/qemu/qemu_interface.h
index d866beb184..6eed3e6bd7 100644
--- a/src/qemu/qemu_interface.h
+++ b/src/qemu/qemu_interface.h
@@ -55,5 +55,3 @@ int qemuInterfaceOpenVhostNet(virDomainObj *def,
int qemuInterfacePrepareSlirp(virQEMUDriver *driver,
virDomainNetDef *net);
-
-int qemuInterfaceVDPAConnect(virDomainNetDef *net) G_NO_INLINE;
diff --git a/tests/qemuhotplugmock.c b/tests/qemuhotplugmock.c
index 89d287945a..dd7e2c67e0 100644
--- a/tests/qemuhotplugmock.c
+++ b/tests/qemuhotplugmock.c
@@ -18,8 +18,8 @@
#include
+#include "qemu/qemu_command.h"
#include "qemu/qemu_hotplug.h"
-#include "qemu/qemu_interface.h"
#include "qemu/qemu_process.h"
#include "testutilsqemu.h"
#include "conf/domain_conf.h"
@@ -94,7 +94,7 @@ qemuProcessKillManagedPRDaemon(virDomainObj *vm G_GNUC_UNUSED)
}
int
-qemuInterfaceVDPAConnect(virDomainNetDef *net G_GNUC_UNUSED)
+qemuVDPAConnect(const char *devicepath G_GNUC_UNUSED)
{
/* need a valid fd or sendmsg won't work. Just open /dev/null */
return open("/dev/null", O_RDONLY);
diff --git a/tests/qemuxml2argvmock.c b/tests/qemuxml2argvmock.c
index 400dd5c020..52c44b2ed0 100644
--- a/tests/qemuxml2argvmock.c
+++ b/tests/qemuxml2argvmock.c
@@ -255,7 +255,7 @@ virNetDevBandwidthSetRootQDisc(const char *ifname
G_GNUC_UNUSED,
int
-qemuInterfaceVDPAConnect(virDomainNetDef *net G_GNUC_UNUSED)
+qemuVDPAConnect(const char *devicepath G_GNUC_UNUSED)
{
if (fcntl(1732, F_GETFD) != -1)
abort();
--
2.41.0
[libvirt PATCH v2 4/5] qemu: consider vdpa block devices for memlock limits
vDPA block devices will also need the same consideration for memlock
limits as other vdpa devices, so consider these devices when calculating
memlock limits.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c | 12 ++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index c7d64e1b5c..52ea8f649d 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -9652,7 +9652,7 @@ qemuDomainGetNumNVMeDisks(const virDomainDef *def)
static int
-qemuDomainGetNumVDPANetDevices(const virDomainDef *def)
+qemuDomainGetNumVDPADevices(const virDomainDef *def)
{
size_t i;
int n = 0;
@@ -9662,6 +9662,14 @@ qemuDomainGetNumVDPANetDevices(const virDomainDef *def)
n++;
}
+for (i = 0; i < def->ndisks; i++) {
+virStorageSource *src;
+for (src = def->disks[i]->src; src; src = src->backingStore) {
+if (src->type == VIR_STORAGE_TYPE_VHOST_VDPA)
+n++;
+}
+}
+
return n;
}
@@ -9704,7 +9712,7 @@ qemuDomainGetMemLockLimitBytes(virDomainDef *def)
nvfio = qemuDomainGetNumVFIOHostdevs(def);
nnvme = qemuDomainGetNumNVMeDisks(def);
-nvdpa = qemuDomainGetNumVDPANetDevices(def);
+nvdpa = qemuDomainGetNumVDPADevices(def);
/* For device passthrough using VFIO the guest memory and MMIO memory
* regions need to be locked persistent in order to allow DMA.
*
--
2.41.0
[libvirt PATCH v2 1/5] conf: add ability to configure a vdpa block disk device
vDPA block devices can be configured as follows:
Signed-off-by: Jonathon Jongsma
---
docs/formatdomain.rst | 19 +--
src/ch/ch_monitor.c | 1 +
src/conf/domain_conf.c| 8
src/conf/schemas/domaincommon.rng | 13 +
src/conf/storage_source_conf.c| 7 ++-
src/conf/storage_source_conf.h| 2 ++
src/libxl/xen_xl.c| 1 +
src/qemu/qemu_block.c | 6 ++
src/qemu/qemu_command.c | 1 +
src/qemu/qemu_migration.c | 2 ++
src/qemu/qemu_snapshot.c | 4
src/qemu/qemu_validate.c | 1 +
src/storage_file/storage_source.c | 1 +
13 files changed, 63 insertions(+), 3 deletions(-)
diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst
index bc469e5f9f..a65edc6703 100644
--- a/docs/formatdomain.rst
+++ b/docs/formatdomain.rst
@@ -2678,6 +2678,11 @@ paravirtualized driver is specified via the ``disk``
element.
+
+
+
+
+
...
@@ -2688,8 +2693,9 @@ paravirtualized driver is specified via the ``disk``
element.
``type``
Valid values are "file", "block", "dir" ( :since:`since 0.7.5` ),
"network" ( :since:`since 0.8.7` ), or "volume" ( :since:`since 1.0.5` ),
- or "nvme" ( :since:`since 6.0.0` ), or "vhostuser" ( :since:`since
7.1.0` )
- and refer to the underlying source for the disk. :since:`Since 0.0.3`
+ or "nvme" ( :since:`since 6.0.0` ), or "vhostuser" ( :since:`since
7.1.0` ),
+ or "vhostvdpa" ( :since:`since 9.8.0 (QEMU 8.1.0)`) and refer to the
+ underlying source for the disk. :since:`Since 0.0.3`
``device``
Indicates how the disk is to be exposed to the guest OS. Possible values
for this attribute are "floppy", "disk", "cdrom", and "lun", defaulting
to
@@ -2879,6 +2885,15 @@ paravirtualized driver is specified via the ``disk``
element.
XML for this disk type. Additionally features such as
blockjobs,
incremental backups and snapshots are not supported for this disk type.
+ ``vhostvdpa``
+ Enables the hypervisor to connect to a vDPA block device. Requires shared
+ memory configured for the VM, for more details see ``access`` mode for
+ ``memoryBacking`` element (See `Memory Backing`_).
+
+ The ``source`` element has a mandatory attribute ``dev`` that specifies
+ the fully-qualified path to the vhost-vdpa character device (e.g.
+ ``/dev/vhost-vdpa-0``).
+
With "file", "block", and "volume", one or more optional sub-elements
``seclabel`` (See `Security label`_) can be used to override the domain
security labeling policy for just that source file.
diff --git a/src/ch/ch_monitor.c b/src/ch/ch_monitor.c
index 200ad6c77b..1691a4efb6 100644
--- a/src/ch/ch_monitor.c
+++ b/src/ch/ch_monitor.c
@@ -197,6 +197,7 @@ virCHMonitorBuildDiskJson(virJSONValue *disks,
virDomainDiskDef *diskdef)
case VIR_STORAGE_TYPE_VOLUME:
case VIR_STORAGE_TYPE_NVME:
case VIR_STORAGE_TYPE_VHOST_USER:
+case VIR_STORAGE_TYPE_VHOST_VDPA:
case VIR_STORAGE_TYPE_LAST:
default:
virReportEnumRangeError(virStorageType, diskdef->src->type);
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 2c8727de54..1f14ef6f23 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7522,6 +7522,10 @@ virDomainStorageSourceParse(xmlNodePtr node,
if (virDomainDiskSourceVHostUserParse(node, src, xmlopt, ctxt) < 0)
return -1;
break;
+case VIR_STORAGE_TYPE_VHOST_VDPA:
+if (!(src->vdpadev = virXMLPropStringRequired(node, "dev")))
+return -1;
+break;
case VIR_STORAGE_TYPE_NONE:
case VIR_STORAGE_TYPE_LAST:
virReportError(VIR_ERR_INTERNAL_ERROR,
@@ -22386,6 +22390,10 @@ virDomainDiskSourceFormat(virBuffer *buf,
virDomainDiskSourceVhostuserFormat(, ,
src->vhostuser);
break;
+case VIR_STORAGE_TYPE_VHOST_VDPA:
+virBufferEscapeString(, " dev='%s'", src->vdpadev);
+break;
+
case VIR_STORAGE_TYPE_NONE:
case VIR_STORAGE_TYPE_LAST:
virReportError(VIR_ERR_INTERNAL_ERROR,
diff --git a/src/conf/schemas/domaincommon.rng
b/src/conf/schemas/domaincommon.rng
index 2f9ba31c0a..1fe9ccb70e 100644
--- a/src/conf/schemas/domaincommon.rng
+++ b/src/conf/schemas/domaincommon.rng
@@ -1811,6 +1811,7 @@
+
@@ -2381,6 +2382,18 @@
+
+
+ vhostvdpa
+
+
+
+
+
+
+
+
+
(ioemu:)?(fd|hd|sd|vd|xvd|ubd)[a-zA-Z0-9_]+
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index dcac3a8ff6.
[libvirt PATCH v2 5/5] qemu: Implement support for vDPA block devices
Requires recent qemu with support for the virtio-blk-vhost-vdpa device
and the ability to pass a /dev/fdset/N path for the vdpa path (8.1.0)
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1900770
Signed-off-by: Jonathon Jongsma
---
src/qemu/qemu_block.c | 20 --
src/qemu/qemu_process.c | 34 +
src/qemu/qemu_validate.c | 32 ++--
.../disk-vhostvdpa.x86_64-latest.args | 37 +++
tests/qemuxml2argvdata/disk-vhostvdpa.xml | 21 +++
tests/qemuxml2argvtest.c | 34 +
tests/testutilsqemu.c | 11 ++
tests/testutilsqemu.h | 2 +
8 files changed, 185 insertions(+), 6 deletions(-)
create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.xml
diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
index 0b4c2dbcf4..d31bbde0f4 100644
--- a/src/qemu/qemu_block.c
+++ b/src/qemu/qemu_block.c
@@ -778,6 +778,20 @@ qemuBlockStorageSourceGetNVMeProps(virStorageSource *src)
}
+static virJSONValue *
+qemuBlockStorageSourceGetVhostVdpaProps(virStorageSource *src)
+{
+virJSONValue *ret = NULL;
+qemuDomainStorageSourcePrivate *srcpriv =
QEMU_DOMAIN_STORAGE_SOURCE_PRIVATE(src);
+
+ignore_value(virJSONValueObjectAdd(,
+ "s:driver", "virtio-blk-vhost-vdpa",
+ "s:path",
qemuFDPassGetPath(srcpriv->fdpass),
+ NULL));
+return ret;
+}
+
+
static int
qemuBlockStorageSourceGetBlockdevGetCacheProps(virStorageSource *src,
virJSONValue *props)
@@ -874,9 +888,9 @@ qemuBlockStorageSourceGetBackendProps(virStorageSource *src,
break;
case VIR_STORAGE_TYPE_VHOST_VDPA:
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("vhostvdpa disk type not yet supported"));
-return NULL;
+if (!(fileprops = qemuBlockStorageSourceGetVhostVdpaProps(src)))
+return NULL;
+break;
case VIR_STORAGE_TYPE_VHOST_USER:
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index 7a1cdb0302..42837c4a8a 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -6820,6 +6820,28 @@ qemuProcessPrepareLaunchSecurityGuestInput(virDomainObj
*vm)
}
+static int
+qemuProcessPrepareHostStorageSourceVDPA(virStorageSource *src,
+qemuDomainObjPrivate *priv)
+{
+qemuDomainStorageSourcePrivate *srcpriv = NULL;
+virStorageType actualType = virStorageSourceGetActualType(src);
+int vdpafd = -1;
+
+if (actualType != VIR_STORAGE_TYPE_VHOST_VDPA)
+return 0;
+
+if ((vdpafd = qemuVDPAConnect(src->vdpadev)) < 0)
+return -1;
+
+srcpriv = qemuDomainStorageSourcePrivateFetch(src);
+
+srcpriv->fdpass = qemuFDPassNew(src->nodestorage, priv);
+qemuFDPassAddFD(srcpriv->fdpass, , "-vdpa");
+return 0;
+}
+
+
static int
qemuProcessPrepareHostStorage(virQEMUDriver *driver,
virDomainObj *vm,
@@ -6856,6 +6878,18 @@ qemuProcessPrepareHostStorage(virQEMUDriver *driver,
return -1;
}
+/* connect to any necessary vdpa block devices */
+for (i = vm->def->ndisks; i > 0; i--) {
+size_t idx = i - 1;
+virDomainDiskDef *disk = vm->def->disks[idx];
+virStorageSource *src;
+
+for (src = disk->src; virStorageSourceIsBacking(src); src =
src->backingStore) {
+if (qemuProcessPrepareHostStorageSourceVDPA(src, vm->privateData)
< 0)
+return -1;
+}
+}
+
return 0;
}
diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c
index 165ab3a66a..5bae56b00f 100644
--- a/src/qemu/qemu_validate.c
+++ b/src/qemu/qemu_validate.c
@@ -3175,13 +3175,39 @@ qemuValidateDomainDeviceDefDisk(const virDomainDiskDef
*disk,
}
if (disk->src->type == VIR_STORAGE_TYPE_VHOST_USER) {
+const char *vhosttype = virStorageTypeToString(disk->src->type);
+
if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VHOST_USER_BLK)) {
-virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
- _("vhostuser disk is not supported with this QEMU
binary"));
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("%1$s disk is not supported with this QEMU
binary"),
+ vhosttype);
+return -1;
+}
+
+if (qemuValidateDomainDefVhostUserRequireSharedMemory(def, vhosttype)
< 0)
+
[libvirt PATCH v2 0/5] Add support for vDPA block devices
see https://bugzilla.redhat.com/show_bug.cgi?id=1900770. Changes in v2: - Don't use virStorageSource->path for vdpa device path to avoid clashing with existing path functionality - Move vdpa device opening to the qemuProcessPrepareHostStorageSource() function rather than the qemuDomainPrepareStorageSource() function. This also required some additional support in the tests for setting up the objects properly for testing. - rebased to latest master branch Jonathon Jongsma (5): conf: add ability to configure a vdpa block disk device qemu: add virtio-blk-vhost-vdpa capability qemu: make vdpa connect function more generic qemu: consider vdpa block devices for memlock limits qemu: Implement support for vDPA block devices docs/formatdomain.rst | 19 +- src/ch/ch_monitor.c | 1 + src/conf/domain_conf.c| 8 src/conf/schemas/domaincommon.rng | 13 +++ src/conf/storage_source_conf.c| 7 +++- src/conf/storage_source_conf.h| 2 + src/libxl/xen_xl.c| 1 + src/qemu/qemu_block.c | 20 ++ src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 24 +++- src/qemu/qemu_command.h | 1 + src/qemu/qemu_domain.c| 12 +- src/qemu/qemu_interface.c | 23 src/qemu/qemu_interface.h | 2 - src/qemu/qemu_migration.c | 2 + src/qemu/qemu_process.c | 34 + src/qemu/qemu_snapshot.c | 4 ++ src/qemu/qemu_validate.c | 33 +++-- src/storage_file/storage_source.c | 1 + .../caps_8.1.0_x86_64.xml | 1 + tests/qemuhotplugmock.c | 4 +- .../disk-vhostvdpa.x86_64-latest.args | 37 +++ tests/qemuxml2argvdata/disk-vhostvdpa.xml | 21 +++ tests/qemuxml2argvmock.c | 2 +- tests/qemuxml2argvtest.c | 34 + tests/testutilsqemu.c | 11 ++ tests/testutilsqemu.h | 2 + 28 files changed, 285 insertions(+), 37 deletions(-) create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.x86_64-latest.args create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.xml -- 2.41.0
Re: [libvirt PATCH 5/5] qemu: Implement support for vDPA block devices
On 7/24/23 8:05 AM, Peter Krempa wrote:
As I've promised a long time ago I gave your patches some testing in
regards of cooperation with blockjobs and snapshots.
Since the new version of the patches was not yet posted on the list I'm
replying here including my observations from testing patches from your
gitlab branch:
On Tue, Jun 06, 2023 at 16:15:30 -0500, Jonathon Jongsma wrote:
Requires recent qemu with support for the virtio-blk-vhost-vdpa device
and the ability to pass a /dev/fdset/N path for the vdpa path (8.1.0)
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1900770
Since this is a feature addition the 'Fixes' keyword doesn't make sense.
Use e.g. 'Resolves' instead.
Additionally you're missing the DCO certification here.
---
src/qemu/qemu_block.c | 20 --
src/qemu/qemu_domain.c | 25
src/qemu/qemu_validate.c | 44 +++---
tests/qemuxml2argvdata/disk-vhostvdpa.args | 35 +
tests/qemuxml2argvdata/disk-vhostvdpa.xml | 21 +++
tests/qemuxml2argvtest.c | 2 +
6 files changed, 139 insertions(+), 8 deletions(-)
create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.args
create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.xml
[...]
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 2f6b32e394..119e52a7d7 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -11157,6 +11157,28 @@ qemuDomainPrepareStorageSourceFDs(virStorageSource
*src,
}
+static int
+qemuDomainPrepareStorageSourceVDPA(virStorageSource *src,
+ qemuDomainObjPrivate *priv)
+{
+qemuDomainStorageSourcePrivate *srcpriv = NULL;
+virStorageType actualType = virStorageSourceGetActualType(src);
+int vdpafd = -1;
+
+if (actualType != VIR_STORAGE_TYPE_VHOST_VDPA)
+return 0;
+
+if ((vdpafd = qemuVDPAConnect(src->path)) < 0)
+return -1;
This function call directly touches the host filesystem, which is not
supposed to be in the *DomainPrepareStorageSource* functions but we
rather have a completely separate machinery in
qemuProcessPrepareHostStorage.
Unfortunately that one doesn't yet need to handle individual backing
chain members though.
This ensures that the code doesn't get accidentally called from tests
even without mocking the code as the tests reimplement the functions
differently for testing purposes.
Somehow I missed this comment earlier. Unfortunately, it doesn't seem
straightforward to move this code. We can't simply move all of the logic
to qemuProcessPrepareHostStorage() because that function doesn't get
called during the tests. I thought I could move only the opening of the
fd to the PrepareHostStorage() function and then keep the qemuFDPass
construction in this function, but that doesn't work: the
PrepareHostStorage() function actually gets called *after* this
function. So the fd would not even be open yet at the time this function
gets called.
So... it seems that the options are either:
- leave everything in qemuDomainPrepareStorageSourceVDPA() (as is)
- move the fd opening to PrepareHostStorage() and then move the rest to
a different common function that is called after that, such as
qemuBuildDiskSourceCommandLine()
- a third option (suggestions?)
It's worth noting that the vdpa *network* device essentially does
everything (opening the file, creating the qemuFDPass object, etc) in
the qemuBuildInterfaceCommandLine() function. This was done to match
other network devices that connect to open file descriptors (see
qemuBuildInterfaceConnect()). But based on your comments above, it
sounds like this may not be a the ideal situation even though the
function is mocked to not actually open any file descriptors from the
host filesystem under test.
Jonathon
+
+srcpriv = qemuDomainStorageSourcePrivateFetch(src);
+
+srcpriv->fdpass = qemuFDPassNew(src->nodestorage, priv);
+qemuFDPassAddFD(srcpriv->fdpass, , "-vdpa");
+return 0;
+}
[...]
diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c
index 9dce908cfe..67b0944162 100644
--- a/src/qemu/qemu_validate.c
+++ b/src/qemu/qemu_validate.c
@@ -3220,6 +3220,28 @@ qemuValidateDomainDeviceDefDiskTransient(const
virDomainDiskDef *disk,
}
+static int
+qemuValidateDomainDeviceDefDiskVhost(const virDomainDef *def,
+ virStorageType storagetype,
+ virQEMUCapsFlags flag,
+ virQEMUCaps *qemuCaps)
+{
+const char *vhosttype = virStorageTypeToString(storagetype);
+
+if (!virQEMUCapsGet(qemuCaps, flag)) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("%1$s disk is not supported with this QEMU binary"),
+ vhosttype);
+return -1;
I'd prefer if both thing
Re: [libvirt PATCH v8 25/37] qemu: Monitor nbdkit process for exit
On 9/5/23 2:54 AM, Peter Krempa wrote:
On Thu, Aug 31, 2023 at 16:40:05 -0500, Jonathon Jongsma wrote:
Adds the ability to monitor the nbdkit process so that we can take
action in case the child exits unexpectedly.
When the nbdkit process exits, we pause the vm, restart nbdkit, and then
resume the vm. This allows the vm to continue working in the event of a
nbdkit failure.
Eventually we may want to generalize this functionality since we may
need something similar for e.g. qemu-storage-daemon, etc.
The process is monitored with the pidfd_open() syscall if it exists
(since linux 5.3). Otherwise it resorts to checking whether the process
is alive once a second. The one-second time period was chosen somewhat
arbitrarily.
Signed-off-by: Jonathon Jongsma
---
meson.build | 11 +++
src/qemu/qemu_domain.c | 1 +
src/qemu/qemu_domain.h | 1 +
src/qemu/qemu_driver.c | 16
src/qemu/qemu_nbdkit.c | 169 ---
src/qemu/qemu_nbdkit.h | 8 +-
src/qemu/qemu_process.c | 15 +++-
src/qemu/qemu_process.h | 3 +
tests/meson.build| 6 +-
tests/qemuxml2argvtest.c | 6 +-
10 files changed, 219 insertions(+), 17 deletions(-)
[...]
diff --git a/tests/meson.build b/tests/meson.build
index f05774263c..b235c5f4dd 100644
--- a/tests/meson.build
+++ b/tests/meson.build
@@ -456,8 +456,12 @@ if conf.has('WITH_QEMU')
{ 'name': 'qemuvhostusertest', 'link_with': [ test_qemu_driver_lib ],
'link_whole': [ test_file_wrapper_lib ] },
{ 'name': 'qemuxml2argvtest', 'link_with': [ test_qemu_driver_lib,
test_utils_qemu_monitor_lib ], 'link_whole': [ test_utils_qemu_lib,
test_file_wrapper_lib ] },
{ 'name': 'qemuxml2xmltest', 'link_with': [ test_qemu_driver_lib ],
'link_whole': [ test_utils_qemu_lib, test_file_wrapper_lib ] },
-{ 'name': 'qemunbdkittest', 'link_with': [ test_qemu_driver_lib ],
'link_whole': [ test_utils_qemu_lib, test_file_wrapper_lib ] },
]
+ if conf.has('WITH_NBDKIT')
+tests += [
+ { 'name': 'qemunbdkittest', 'link_with': [ test_qemu_driver_lib ],
'link_whole': [ test_utils_qemu_lib, test_file_wrapper_lib ] },
+]
+ endif
endif
if conf.has('WITH_REMOTE')
diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c
index 0304f66f1d..216fd3a841 100644
--- a/tests/qemuxml2argvtest.c
+++ b/tests/qemuxml2argvtest.c
@@ -839,8 +839,12 @@ mymain(void)
# define DO_TEST_CAPS_ARCH_VER(name, arch, ver) \
DO_TEST_CAPS_ARCH_VER_FULL(name, arch, ver, ARG_END)
-# define DO_TEST_CAPS_LATEST_NBDKIT(name, ...) \
+# if WITH_NBDKIT
+# define DO_TEST_CAPS_LATEST_NBDKIT(name, ...) \
DO_TEST_CAPS_ARCH_LATEST_FULL(name, "x86_64", ARG_NBDKIT_CAPS,
__VA_ARGS__, QEMU_NBDKIT_CAPS_LAST, ARG_END)
+# else
+# define DO_TEST_CAPS_LATEST_NBDKIT(name, ...)
+# endif /* WITH_NBDKIT */
# define DO_TEST_CAPS_LATEST(name) \
DO_TEST_CAPS_ARCH_LATEST(name, "x86_64")
These two hunks seem misplaced and don't semantically align with the
rest of the patch.
Reviewed-by: Peter Krempa
Yes, they don't appear very related on first glance. But this is the
first commit that added a dependency on the pidfd_open syscall (and thus
added the "WITH_NBDKIT" preprocessor symbol). Before this patch, nbdkit
tests would run fine and generate an nbdkit commandline on any platform
as long as the nbdkit capabilities were set. But after this patch, any
platform that did not provide this syscall would fall back to using the
qemu block driver and generate a different commandline, causing tests to
fail.
Jonathon
Re: [libvirt PATCH v8 26/37] qemu: improve error handling when restarting nbdkit
On 9/5/23 2:55 AM, Peter Krempa wrote: On Thu, Aug 31, 2023 at 16:40:06 -0500, Jonathon Jongsma wrote: Change the return value for qemuNbdkitProcessRestart() and qemuNbdkitStorageSourceManageProcess() to return an error status. The main effect of this change is that when libvirt starts up and reconnects to an already-running domain with an nbdkit-backed disk, it will return an error if it fails to restart any nbdkit processes that are not found to be running or if it fails to monitor any running nbdkit processes. These failures will result in the domain being killed. I don't quite understand why this is a separate patch, but ... Signed-off-by: Jonathon Jongsma --- src/qemu/qemu_driver.c | 3 ++- src/qemu/qemu_nbdkit.c | 34 -- src/qemu/qemu_nbdkit.h | 4 ++-- src/qemu/qemu_process.c | 6 -- 4 files changed, 24 insertions(+), 23 deletions(-) Reviewed-by: Peter Krempa Yes, I probably should have explained why it was separate. I initially made it separate in case you felt that the behavior change was not acceptable and wanted me to drop it. I will squash it into the previous patch. Jonathon
Re: [libvirt PATCH 5/5] qemu: Implement support for vDPA block devices
On 8/16/23 4:19 PM, Jonathon Jongsma wrote:
On 8/8/23 6:00 AM, Stefano Garzarella wrote:
On Mon, Aug 07, 2023 at 03:41:21PM +0200, Peter Krempa wrote:
On Thu, Aug 03, 2023 at 09:48:01 +0200, Stefano Garzarella wrote:
On Wed, Aug 2, 2023 at 10:33 PM Jonathon Jongsma
wrote:
> On 7/24/23 8:05 AM, Peter Krempa wrote:
[...]
> >
> > I've also noticed that using 'qcow2' format for the device
doesn't work:
> >
> > error: internal error: process exited while connecting to
monitor: 2023-07-24T12:54:15.818631Z qemu-system-x86_64: -blockdev
{"node-name":"libvirt-1-format","read-only":false,"cache":{"direct":true,"no-flush":false},"driver":"qcow2","file":"libvirt-1-storage"}: Could not read qcow2 header: Invalid argument
> >
> > If that is supposed to work, then qemu devs will probably need
to know
> > about that, if that is not supposed to work, libvirt needs to add a
> > check, because the error doesn't tell much. It's also possible I've
> > messed up when formatting the image though, as didn't really try to
> > figure out what's happening.
> >
>
>
> That's a good question, and I don't actually know the answer. Were
you
> using an actual vdpa block device for your tests or were you using
the
> vdpa block simulator kernel module? How did you set it up? Adding
> Stefano to cc for his thoughts.
Yep, I would also like to understand how you initialized the device
with a qcow2 format.
Naively and originally I've simply used it as 'raw' at first and
formatted it from the guest OS. Then I've shut-down the VM and started
it back reconfiguring the image format as qcow2. This normally works
with real-file backed storage, and since the vdpa simulator seems to
persist the contents I supposed this would work.
Cool, I'll try that.
Can you try to reboot the VM, use it as `raw`, and read the qcow2 in the
vm from the guest OS?
Note: there could be some bugs in the simulator!
Theoretically, the best use case for vDPA block is that the backend
handles formats, for QEMU it should just be a virtio device, but being
a blockdev, we should be able to use formats anyway, so it should
work.
Yeah, ideally there will be no format driver in qemu used for these
devices (this is not yet the case, I'll need to fix libvirt to stop
using the 'raw' driver if not needed).
Here I'm more interested whether it is supposed to work, in which case
we want to allow using qcow2 as a format in libvirt, or it's not
supposed to work and we should forbid it before the user gets a
suboptimal error message such as now.
This is a good question. We certainly haven't tested it, because it's an
uncommon scenario, but as I said before, maybe it should work. I need to
check it better.
For now, waiting for real hardware, the only way to test vDPA block
support in QEMU is to use the simulator in the kernel or VDUSE.
With the kernel simulator we only have a 128 MB ramdisk available,
with VDUSE you can use QSD with any file:
$ modprobe -a vhost_vdpa vduse
$ qemu-storage-daemon \
--blockdev
file,filename=/path/to/image.qcow2,cache.direct=on,aio=native,node-name=file
\
--blockdev qcow2,file=file,node-name=qcow2 \
--export
vduse-blk,id=vduse0,name=vduse0,num-queues=1,node-name=qcow2,writable=on
$ vdpa dev add name vduse0 mgmtdev vduse
Then you have a /dev/vhost-vdpa-X device that you can use with the
`virtio-blk-vhost-vdpa` blockdev (note: vduse requires QEMU with a
memory-backed with `share=on`), but using raw since the qcow2 is
handled by QSD.
Of course, we should be able to use raw file with QSD and qcow2 on
qemu (although it's not the optimal configuration), but I don't know
how to initialize a `virtio-blk-vhost-vdpa` blockdev with a qcow2
image :-(
With the above qemu storage daemon you should be able to do that by
simply dropping the qcow2 format driver and simply exposing a qcow2
formatted image. It similarly works with NBD:
I've formatted 2 qcow2 images:
# qemu-img create -f qcow2 /root/image1.qcow2 100M
# qemu-img create -f qcow2 /root/image2.qcow2 100M
And then exported them both via vduse and nbd without interpreting
qcow2, thus making the QSD into just a dumb storage device:
# qemu-storage-daemon \
--blockdev
file,filename=/root/image1.qcow2,cache.direct=on,aio=native,node-name=file1 \
--export
vduse-blk,id=vduse0,name=vduse0,num-queues=1,node-name=file1,writable=on \
--blockdev
file,filename=/root/image2.qcow2,cache.direct=on,aio=native,node-name=file2 \
--nbd-server addr.type=unix,addr.path=/tmp/nbd.sock \
--export nbd,id=nbd0,node-name=file2,writable=on,name=exportname
Cool! Thanks for sharing!
Now when I start a VM using the NBD export in qcow2 format:
function='0x0'/>
The VM starts fine, but when using:
function='0x
[libvirt PATCH v8 33/37] qemu: implement keyfile auth for ssh disks with nbdkit
For ssh disks that are served by nbdkit, we can support logging in with
an ssh key file. Pass the path to the configured key file and the
username to the nbdkit process.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/domain_conf.c| 30 +
src/conf/storage_source_conf.c| 2 ++
src/conf/storage_source_conf.h| 5 ++-
src/qemu/qemu_nbdkit.c| 15 +++--
.../disk-network-ssh-key.args.disk0 | 9 +
.../disk-network-ssh.args.disk2 | 9 +
tests/qemunbdkittest.c| 1 +
.../qemuxml2argvdata/disk-network-ssh-key.xml | 33 +++
8 files changed, 93 insertions(+), 11 deletions(-)
create mode 100644 tests/qemunbdkitdata/disk-network-ssh-key.args.disk0
create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk2
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh-key.xml
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 842b6404b5..929e115bce 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7268,10 +7268,18 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
return -1;
}
}
-if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
-(tmpnode = virXPathNode("./knownHosts", ctxt))) {
-if (!(src->ssh_known_hosts_file = virXMLPropStringRequired(tmpnode,
"path")))
-return -1;
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH) {
+if ((tmpnode = virXPathNode("./knownHosts", ctxt))) {
+if (!(src->ssh_known_hosts_file =
virXMLPropStringRequired(tmpnode, "path")))
+return -1;
+}
+if ((tmpnode = virXPathNode("./identity", ctxt))) {
+if (!(src->ssh_user = virXMLPropStringRequired(tmpnode,
"username")))
+return -1;
+
+if (!(src->ssh_keyfile = virXMLPropStringRequired(tmpnode,
"keyfile")))
+return -1;
+}
}
return 0;
@@ -22280,8 +22288,18 @@ virDomainDiskSourceFormatNetwork(virBuffer *attrBuf,
if (src->timeout)
virBufferAsprintf(childBuf, "\n",
src->timeout);
-if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
src->ssh_known_hosts_file)
-virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH) {
+if (src->ssh_known_hosts_file)
+virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
+if (src->ssh_keyfile) {
+virBufferAddLit(childBuf, "ssh_user);
+virBufferEscapeString(childBuf, " keyfile='%s'", src->ssh_keyfile);
+
+virBufferAddLit(childBuf, "/>\n");
+}
+}
}
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index 906bc36a9b..ce9c1f66c2 100644
--- a/src/conf/storage_source_conf.c
+++ b/src/conf/storage_source_conf.c
@@ -896,6 +896,7 @@ virStorageSourceCopy(const virStorageSource *src,
def->ssh_host_key_check_disabled = src->ssh_host_key_check_disabled;
def->ssh_user = g_strdup(src->ssh_user);
def->ssh_known_hosts_file = g_strdup(src->ssh_known_hosts_file);
+def->ssh_keyfile = g_strdup(src->ssh_keyfile);
def->nfs_user = g_strdup(src->nfs_user);
def->nfs_group = g_strdup(src->nfs_group);
@@ -1172,6 +1173,7 @@ virStorageSourceClear(virStorageSource *def)
VIR_FREE(def->ssh_user);
VIR_FREE(def->ssh_known_hosts_file);
+VIR_FREE(def->ssh_keyfile);
VIR_FREE(def->nfs_user);
VIR_FREE(def->nfs_group);
diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
index 8a9c7d07e2..8c805664af 100644
--- a/src/conf/storage_source_conf.h
+++ b/src/conf/storage_source_conf.h
@@ -406,12 +406,11 @@ struct _virStorageSource {
bool hostcdrom; /* backing device is a cdrom */
-/* passthrough variables for the ssh driver which we don't handle properly
*/
-/* these must not be used apart from formatting the output JSON in the
qemu driver */
+/* ssh variables */
char *ssh_user;
bool ssh_host_key_check_disabled;
-/* additional ssh variables */
char *ssh_known_hosts_file;
+char *ssh_keyfile;
/* nfs_user and nfs_group store the strings passed in by the user for NFS
params.
* nfs_uid and nfs_gid represent the converted/looked up ID numbers which
are used
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index afac71e21a..0a6c7962b0 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -1049,8 +1049,12 @@ qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
if (proc->source->auth) {
if (q
[libvirt PATCH v8 15/37] tests: add ability to test various nbdkit capabilities
Add new DO_TEST_CAPS_LATEST_NBDKIT macro to test xml2argv for various
nbdkit capability scenarios.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 20 +---
tests/qemuxml2argvtest.c | 11 +++
tests/testutilsqemu.c| 26 ++
tests/testutilsqemu.h| 4
4 files changed, 58 insertions(+), 3 deletions(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 2d70e72c42..81861bae4a 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -290,10 +290,16 @@ virNbkditCapsCheckModdir(const char *moddir,
static bool
virNbdkitCapsIsValid(void *data,
- void *privData G_GNUC_UNUSED)
+ void *privData)
{
qemuNbdkitCaps *nbdkitCaps = data;
struct stat st;
+/* when run under test, we will use privData as a signal to indicate that
+ * we shouldn't touch the filesystem */
+bool skipValidation = (privData != NULL);
+
+if (skipValidation)
+return true;
if (!nbdkitCaps->path)
return true;
@@ -334,9 +340,17 @@ virNbdkitCapsIsValid(void *data,
static void*
virNbdkitCapsNewData(const char *binary,
- void *privData G_GNUC_UNUSED)
+ void *privData)
{
-qemuNbdkitCaps *caps = qemuNbdkitCapsNew(binary);
+/* when run under test, we will use privData as a signal to indicate that
+ * we shouldn't touch the filesystem */
+bool skipNewData = (privData != NULL);
+qemuNbdkitCaps *caps = NULL;
+
+if (skipNewData)
+return NULL;
+
+caps = qemuNbdkitCapsNew(binary);
qemuNbdkitCapsQuery(caps);
return caps;
diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c
index 1b76b32812..d64c21ae17 100644
--- a/tests/qemuxml2argvtest.c
+++ b/tests/qemuxml2argvtest.c
@@ -611,6 +611,14 @@ testCompareXMLToArgv(const void *data)
if (qemuTestCapsCacheInsert(driver.qemuCapsCache, info->qemuCaps) < 0)
goto cleanup;
+if (info->nbdkitCaps) {
+if (virFileCacheInsertData(driver.nbdkitCapsCache, TEST_NBDKIT_PATH,
+ g_object_ref(info->nbdkitCaps)) < 0) {
+g_object_unref(info->nbdkitCaps);
+goto cleanup;
+}
+}
+
if (info->migrateFrom &&
!(migrateURI = qemuMigrationDstGetURI(info->migrateFrom,
info->migrateFd)))
@@ -831,6 +839,9 @@ mymain(void)
# define DO_TEST_CAPS_ARCH_VER(name, arch, ver) \
DO_TEST_CAPS_ARCH_VER_FULL(name, arch, ver, ARG_END)
+# define DO_TEST_CAPS_LATEST_NBDKIT(name, ...) \
+DO_TEST_CAPS_ARCH_LATEST_FULL(name, "x86_64", ARG_NBDKIT_CAPS,
__VA_ARGS__, QEMU_NBDKIT_CAPS_LAST, ARG_END)
+
# define DO_TEST_CAPS_LATEST(name) \
DO_TEST_CAPS_ARCH_LATEST(name, "x86_64")
diff --git a/tests/testutilsqemu.c b/tests/testutilsqemu.c
index fdbad16abe..9a607ab5a3 100644
--- a/tests/testutilsqemu.c
+++ b/tests/testutilsqemu.c
@@ -50,6 +50,10 @@ virFindFileInPath(const char *file)
return g_strdup_printf("/usr/bin/%s", file);
}
+if (g_str_equal(file, "nbdkit")) {
+return g_strdup(TEST_NBDKIT_PATH);
+}
+
/* Nothing in tests should be relying on real files
* in host OS, so we return NULL to try to force
* an error in such a case
@@ -288,6 +292,7 @@ void qemuTestDriverFree(virQEMUDriver *driver)
virObjectUnref(driver->caps);
virObjectUnref(driver->config);
virObjectUnref(driver->securityManager);
+g_clear_object(>nbdkitCapsCache);
virCPUDefFree(cpuDefault);
virCPUDefFree(cpuHaswell);
@@ -487,6 +492,12 @@ int qemuTestDriverInit(virQEMUDriver *driver)
if (!driver->qemuCapsCache)
goto error;
+driver->nbdkitCapsCache = qemuNbdkitCapsCacheNew("/dev/null");
+/* the nbdkitCapsCache just interprets the presence of a non-null private
+ * data pointer as a signal to skip cache validation. This prevents the
+ * cache from trying to validate the plugindir mtime, etc during test */
+virFileCacheSetPriv(driver->nbdkitCapsCache, GUINT_TO_POINTER(1));
+
driver->xmlopt = virQEMUDriverCreateXMLConf(driver, "none");
if (!driver->xmlopt)
goto error;
@@ -780,6 +791,14 @@ testQemuInfoSetArgs(struct testQemuInfo *info,
ignore_value(virBitmapSetBit(info->args.fakeCapsDel, flag));
break;
+case ARG_NBDKIT_CAPS:
+if (!(info->args.fakeNbdkitCaps))
+info->args.fakeNbdkitCaps =
virBitmapNew(QEMU_NBDKIT_CAPS_LAST);
+
+while ((flag = va_arg(argptr, int)) < QEMU_NBDKIT_CAPS_LAST)
+ignore_value(virBitmapSetBit(info->args.fakeNbdkitCaps, flag));
+break;
+
case ARG_GIC:
info->args.gic = va_arg(argptr, int);
[libvirt PATCH v8 21/37] util: make virCommandSetSendBuffer testable
Add a private function to peek at the list of send buffers in virCommand
so that it is testable
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/libvirt_private.syms | 1 +
src/util/vircommand.c | 17 +
src/util/vircommand.h | 8
src/util/vircommandpriv.h | 4
4 files changed, 22 insertions(+), 8 deletions(-)
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
index 1e3e407097..e4da5388e7 100644
--- a/src/libvirt_private.syms
+++ b/src/libvirt_private.syms
@@ -2117,6 +2117,7 @@ virCommandNewArgs;
virCommandNewVAList;
virCommandNonblockingFDs;
virCommandPassFD;
+virCommandPeekSendBuffers;
virCommandRawStatus;
virCommandRequireHandshake;
virCommandRun;
diff --git a/src/util/vircommand.c b/src/util/vircommand.c
index 899d413dd2..7d7ce4297e 100644
--- a/src/util/vircommand.c
+++ b/src/util/vircommand.c
@@ -78,14 +78,6 @@ struct _virCommandFD {
unsigned int flags;
};
-typedef struct _virCommandSendBuffer virCommandSendBuffer;
-struct _virCommandSendBuffer {
-int fd;
-unsigned char *buffer;
-size_t buflen;
-off_t offset;
-};
-
struct _virCommand {
int has_error; /* 0 on success, -1 on error */
@@ -3515,3 +3507,12 @@ virCommandSetRunAmong(virCommand *cmd,
cmd->schedCore = pid;
}
+
+void
+virCommandPeekSendBuffers(virCommand *cmd,
+ virCommandSendBuffer **buffers,
+ int *nbuffers)
+{
+*buffers = cmd->sendBuffers;
+*nbuffers = cmd->numSendBuffers;
+}
diff --git a/src/util/vircommand.h b/src/util/vircommand.h
index d51449ac90..9bcdce35b9 100644
--- a/src/util/vircommand.h
+++ b/src/util/vircommand.h
@@ -24,6 +24,14 @@
#include "internal.h"
#include "virbuffer.h"
+typedef struct _virCommandSendBuffer virCommandSendBuffer;
+struct _virCommandSendBuffer {
+int fd;
+unsigned char *buffer;
+size_t buflen;
+off_t offset;
+};
+
typedef struct _virCommand virCommand;
/* This will execute in the context of the first child
diff --git a/src/util/vircommandpriv.h b/src/util/vircommandpriv.h
index ff17fa5ded..d579810bb5 100644
--- a/src/util/vircommandpriv.h
+++ b/src/util/vircommandpriv.h
@@ -47,3 +47,7 @@ void virCommandSetDryRun(virCommandDryRunToken *tok,
bool bufCommandStripPath,
virCommandDryRunCallback cb,
void *opaque);
+
+void virCommandPeekSendBuffers(virCommand *cmd,
+ virCommandSendBuffer **buffers,
+ int *nbuffers);
--
2.41.0
[libvirt PATCH v8 12/37] qemu: Extract qemuDomainLogContext into a new file
This will allow us to use it for nbdkit logging in upcoming commits.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
po/POTFILES| 1 +
src/qemu/meson.build | 1 +
src/qemu/qemu_domain.c | 247 ++
src/qemu/qemu_domain.h | 27 +---
src/qemu/qemu_logcontext.c | 264 +
src/qemu/qemu_logcontext.h | 38 ++
src/qemu/qemu_process.c| 44 +++
7 files changed, 346 insertions(+), 276 deletions(-)
create mode 100644 src/qemu/qemu_logcontext.c
create mode 100644 src/qemu/qemu_logcontext.h
diff --git a/po/POTFILES b/po/POTFILES
index 6167f98ac5..3a51aea5cb 100644
--- a/po/POTFILES
+++ b/po/POTFILES
@@ -174,6 +174,7 @@ src/qemu/qemu_hostdev.c
src/qemu/qemu_hotplug.c
src/qemu/qemu_interface.c
src/qemu/qemu_interop_config.c
+src/qemu/qemu_logcontext.c
src/qemu/qemu_migration.c
src/qemu/qemu_migration_cookie.c
src/qemu/qemu_migration_params.c
diff --git a/src/qemu/meson.build b/src/qemu/meson.build
index 9be6996195..6d7a1bfbb0 100644
--- a/src/qemu/meson.build
+++ b/src/qemu/meson.build
@@ -21,6 +21,7 @@ qemu_driver_sources = [
'qemu_hotplug.c',
'qemu_interface.c',
'qemu_interop_config.c',
+ 'qemu_logcontext.c',
'qemu_migration.c',
'qemu_migration_cookie.c',
'qemu_migration_params.c',
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index d79f9879df..23608f95bd 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -455,21 +455,8 @@ qemuDomainObjFromDomain(virDomainPtr domain)
}
-struct _qemuDomainLogContext {
-GObject parent;
-
-int writefd;
-int readfd; /* Only used if manager == NULL */
-off_t pos;
-ino_t inode; /* Only used if manager != NULL */
-char *path;
-virLogManager *manager;
-};
-
-G_DEFINE_TYPE(qemuDomainLogContext, qemu_domain_log_context, G_TYPE_OBJECT);
static virClass *qemuDomainSaveCookieClass;
-static void qemuDomainLogContextFinalize(GObject *obj);
static void qemuDomainSaveCookieDispose(void *obj);
@@ -482,32 +469,8 @@ qemuDomainOnceInit(void)
return 0;
}
-static void qemu_domain_log_context_init(qemuDomainLogContext *logctxt
G_GNUC_UNUSED)
-{
-}
-
-static void qemu_domain_log_context_class_init(qemuDomainLogContextClass
*klass)
-{
-GObjectClass *obj = G_OBJECT_CLASS(klass);
-
-obj->finalize = qemuDomainLogContextFinalize;
-}
-
VIR_ONCE_GLOBAL_INIT(qemuDomain);
-static void
-qemuDomainLogContextFinalize(GObject *object)
-{
-qemuDomainLogContext *ctxt = QEMU_DOMAIN_LOG_CONTEXT(object);
-VIR_DEBUG("ctxt=%p", ctxt);
-
-virLogManagerFree(ctxt->manager);
-VIR_FREE(ctxt->path);
-VIR_FORCE_CLOSE(ctxt->writefd);
-VIR_FORCE_CLOSE(ctxt->readfd);
-G_OBJECT_CLASS(qemu_domain_log_context_parent_class)->finalize(object);
-}
-
/* qemuDomainGetMasterKeyFilePath:
* @libDir: Directory path to domain lib files
*
@@ -6882,7 +6845,7 @@ static void G_GNUC_PRINTF(5, 6)
qemuDomainObjTaintMsg(virQEMUDriver *driver,
virDomainObj *obj,
virDomainTaintFlags taint,
- qemuDomainLogContext *logCtxt,
+ qemuLogContext *logCtxt,
const char *fmt, ...)
{
virErrorPtr orig_err = NULL;
@@ -6935,12 +6898,12 @@ qemuDomainObjTaintMsg(virQEMUDriver *driver,
goto cleanup;
if (logCtxt) {
-rc = qemuDomainLogContextWrite(logCtxt,
- "%s: Domain id=%d is tainted:
%s%s%s%s\n",
- timestamp,
- obj->def->id,
- virDomainTaintTypeToString(taint),
- extraprefix, extramsg, extrasuffix);
+rc = qemuLogContextWrite(logCtxt,
+ "%s: Domain id=%d is tainted: %s%s%s%s\n",
+ timestamp,
+ obj->def->id,
+ virDomainTaintTypeToString(taint),
+ extraprefix, extramsg, extrasuffix);
} else {
rc = qemuDomainLogAppendMessage(driver, obj,
"%s: Domain id=%d is tainted:
%s%s%s%s\n",
@@ -6961,7 +6924,7 @@ qemuDomainObjTaintMsg(virQEMUDriver *driver,
void qemuDomainObjTaint(virQEMUDriver *driver,
virDomainObj *obj,
virDomainTaintFlags taint,
-qemuDomainLogContext *logCtxt)
+qemuLogContext *logCtxt)
{
qemuDomainObjTaintMsg(driver, obj, taint, logCtxt, NULL);
qemuDomainSaveStatus(obj);
@@ -6970,7 +6933,7 @@ void qemuDomainObjTaint(virQEMUDriver *driver,
static void
qemuDomainObjCheckMachineTaint(virQEMUDriver *driver,
virDomainO
[libvirt PATCH v8 35/37] qemu: implement ssh-agent auth for ssh disks with nbdkit
It's not possible to use password-protected ssh keys directly with
libvirt because libvirt doesn't have any way to prompt a user for the
password. To accomodate password-protected key files, an administrator
can add these keys to an ssh agent and then configure the domain with
the path to the ssh-agent socket.
Note that this requires an administrator or management app to
configure the ssh-agent with an appropriate socket path and add the
necessary keys to it. In addition, it does not currently work with
selinux enabled. The ssh-agent socket would need a label that libvirt
would be allowed to access rather than unconfined_t.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/domain_conf.c | 14 --
src/conf/storage_source_conf.c | 2 ++
src/conf/storage_source_conf.h | 1 +
src/qemu/qemu_nbdkit.c | 10 ++
.../disk-network-ssh-key.args.disk0 | 6 +++---
.../disk-network-ssh-key.args.disk1 | 9 +
tests/qemuxml2argvdata/disk-network-ssh-key.xml | 17 ++---
7 files changed, 51 insertions(+), 8 deletions(-)
create mode 100644 tests/qemunbdkitdata/disk-network-ssh-key.args.disk1
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 929e115bce..398f40d2be 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7277,8 +7277,17 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
if (!(src->ssh_user = virXMLPropStringRequired(tmpnode,
"username")))
return -1;
-if (!(src->ssh_keyfile = virXMLPropStringRequired(tmpnode,
"keyfile")))
+/* optional path to an ssh key file */
+src->ssh_keyfile = virXMLPropString(tmpnode, "keyfile");
+
+/* optional ssh-agent socket location */
+src->ssh_agent = virXMLPropString(tmpnode, "agentsock");
+if (!src->ssh_keyfile && !src->ssh_agent) {
+virReportError(VIR_ERR_XML_ERROR,
+ _("element '%1$s' requires either 'keyfile' or
'agentsock' attribute"),
+ tmpnode->name);
return -1;
+}
}
}
@@ -22291,11 +22300,12 @@ virDomainDiskSourceFormatNetwork(virBuffer *attrBuf,
if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH) {
if (src->ssh_known_hosts_file)
virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
-if (src->ssh_keyfile) {
+if (src->ssh_keyfile || src->ssh_agent) {
virBufferAddLit(childBuf, "ssh_user);
virBufferEscapeString(childBuf, " keyfile='%s'", src->ssh_keyfile);
+virBufferEscapeString(childBuf, " agentsock='%s'", src->ssh_agent);
virBufferAddLit(childBuf, "/>\n");
}
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index ce9c1f66c2..cafa031dfe 100644
--- a/src/conf/storage_source_conf.c
+++ b/src/conf/storage_source_conf.c
@@ -897,6 +897,7 @@ virStorageSourceCopy(const virStorageSource *src,
def->ssh_user = g_strdup(src->ssh_user);
def->ssh_known_hosts_file = g_strdup(src->ssh_known_hosts_file);
def->ssh_keyfile = g_strdup(src->ssh_keyfile);
+def->ssh_agent = g_strdup(src->ssh_agent);
def->nfs_user = g_strdup(src->nfs_user);
def->nfs_group = g_strdup(src->nfs_group);
@@ -1174,6 +1175,7 @@ virStorageSourceClear(virStorageSource *def)
VIR_FREE(def->ssh_user);
VIR_FREE(def->ssh_known_hosts_file);
VIR_FREE(def->ssh_keyfile);
+VIR_FREE(def->ssh_agent);
VIR_FREE(def->nfs_user);
VIR_FREE(def->nfs_group);
diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
index 8c805664af..061faa66cb 100644
--- a/src/conf/storage_source_conf.h
+++ b/src/conf/storage_source_conf.h
@@ -411,6 +411,7 @@ struct _virStorageSource {
bool ssh_host_key_check_disabled;
char *ssh_known_hosts_file;
char *ssh_keyfile;
+char *ssh_agent;
/* nfs_user and nfs_group store the strings passed in by the user for NFS
params.
* nfs_uid and nfs_gid represent the converted/looked up ID numbers which
are used
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 0a6c7962b0..66b09cd240 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -1057,6 +1057,9 @@ qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
virCommandAddArgPair(cmd, "user", proc->source->ssh_user);
}
+if (proc->source->ssh_agent)
+virCommandAddEnvPair(cmd, "SSH_AUTH_SOCK", proc->source->ssh_agent);
+
if (proc->source->ssh_host_key_check_disabled)
virCommand
[libvirt PATCH v8 34/37] schema: add ssh-agent configuration for ssh disks
Add the ability to specify a path to a ssh-agent socket in order to use the ssh-agent to authenticate to remote ssh disks. Example configuration: ... ... Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 13 - src/conf/schemas/domaincommon.rng | 11 --- 2 files changed, 16 insertions(+), 8 deletions(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index baa2fdce7d..714fee4fbf 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3007,11 +3007,14 @@ paravirtualized driver is specified via the ``disk`` element. are intended to be default, then the entire element may be omitted. When using an ``ssh`` protocol, this element is used to enable - authentication via ssh keys. In this configuration, the element has two - attributes. The ``username`` attribute specifies the name of the user on - the remote server and the ``keyfile`` attribute specifies the path to the - keyfile. Note that this only works for ssh keys that are not - password-protected. + authentication via ssh keys. In this configuration, the element has three + possible attributes. The ``username`` attribute is required and specifies + the name of the user on the remote server. ssh keys can be specified in + one of two ways. The first way is by adding them to an ssh-agent and + providing the path to the ssh-agent socket in the ``agentsock`` + attribute. This method works for ssh keys with or without password + protection. Alternatively, for ssh keys without a password, the ssh key + can be specified directly by setting the ``keyfile`` attribute. ``reconnect`` For disk type ``vhostuser`` configures reconnect timeout if the connection is lost. This is set with the two mandatory attributes ``enabled`` and diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index 47c5ee2a31..d8dd1b8c69 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2186,9 +2186,14 @@ - - - + + + + + + + + -- 2.41.0
[libvirt PATCH v8 29/37] qemu: implement password auth for ssh disks with nbdkit
For ssh disks that are served by nbdkit, lookup the password from the
configured secret and securely pass it to the nbdkit process using fd
passing.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c| 84 ++-
.../disk-network-ssh-password.args.disk0 | 8 ++
...k-network-ssh-password.args.disk0.pipe.778 | 1 +
.../disk-network-ssh.args.disk1 | 8 ++
.../disk-network-ssh.args.disk1.pipe.778 | 1 +
tests/qemunbdkittest.c| 1 +
...sk-network-ssh-password.x86_64-latest.args | 35
.../disk-network-ssh-password.xml | 34
tests/qemuxml2argvtest.c | 1 +
9 files changed, 134 insertions(+), 39 deletions(-)
create mode 100644 tests/qemunbdkitdata/disk-network-ssh-password.args.disk0
create mode 100644
tests/qemunbdkitdata/disk-network-ssh-password.args.disk0.pipe.778
create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk1
create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk1.pipe.778
create mode 100644
tests/qemuxml2argvdata/disk-network-ssh-password.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh-password.xml
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index db86a18321..4cd91e282b 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -950,6 +950,43 @@ qemuNbdkitCommandPassDataByPipe(virCommand *cmd,
}
+static int
+qemuNbdkitProcessBuildCommandAuth(virStorageAuthDef *authdef,
+ virCommand *cmd)
+{
+g_autoptr(virConnect) conn = NULL;
+g_autofree uint8_t *secret = NULL;
+size_t secretlen = 0;
+int secrettype;
+
+if (!authdef)
+return 0;
+
+if ((secrettype = virSecretUsageTypeFromString(authdef->secrettype)) < 0) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("invalid secret type %1$s"),
+ authdef->secrettype);
+return -1;
+}
+
+conn = virGetConnectSecret();
+if (virSecretGetSecretString(conn,
+ >seclookupdef,
+ secrettype,
+ ,
+ ) < 0)
+return -1;
+
+virCommandAddArgPair(cmd, "user", authdef->username);
+
+if (qemuNbdkitCommandPassDataByPipe(cmd, "password",
+, secretlen) < 0)
+return -1;
+
+return 0;
+}
+
+
static int
qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
virCommand *cmd)
@@ -968,37 +1005,8 @@ qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
}
virCommandAddArgPair(cmd, "url", uristring);
-if (proc->source->auth) {
-g_autoptr(virConnect) conn = virGetConnectSecret();
-g_autofree uint8_t *secret = NULL;
-size_t secretlen = 0;
-int secrettype;
-virStorageAuthDef *authdef = proc->source->auth;
-
-virCommandAddArgPair(cmd, "user",
- proc->source->auth->username);
-
-if ((secrettype =
virSecretUsageTypeFromString(proc->source->auth->secrettype)) < 0) {
-virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
- _("invalid secret type %1$s"),
- proc->source->auth->secrettype);
-return -1;
-}
-
-if (virSecretGetSecretString(conn,
- >seclookupdef,
- secrettype,
- ,
- ) < 0) {
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("failed to get auth secret for storage"));
-return -1;
-}
-
-if (qemuNbdkitCommandPassDataByPipe(cmd, "password",
-, secretlen) < 0)
-return -1;
-}
+if (proc->source->auth &&
qemuNbdkitProcessBuildCommandAuth(proc->source->auth, cmd) < 0)
+return -1;
/* Create a pipe to send the cookies to the nbdkit process. */
if (proc->source->ncookies) {
@@ -1027,7 +1035,6 @@ static int
qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
virCommand *cmd)
{
-const char *user = NULL;
virStorageNetHostDef *host = >source->hosts[0];
g_autofree char *portstr = g_strdup_printf("%u", host->port);
@@ -1038,13 +1045,12 @@ qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess
*proc,
virCommandAddArgPair(cmd, "port", portstr);
virCommandAddArgPair(cmd, "path", proc->source->path);
-
[libvirt PATCH v8 37/37] ci: add libnbd to build
Signed-off-by: Jonathon Jongsma
Reviewed-by: Erik Skultety
---
ci/buildenv/almalinux-8.sh| 1 +
ci/buildenv/centos-stream-8.sh| 1 +
ci/buildenv/centos-stream-9.sh| 1 +
ci/buildenv/debian-12-cross-aarch64.sh| 1 +
ci/buildenv/debian-12-cross-armv6l.sh | 1 +
ci/buildenv/debian-12-cross-armv7l.sh | 1 +
ci/buildenv/debian-12-cross-i686.sh | 1 +
ci/buildenv/debian-12-cross-mips64el.sh | 1 +
ci/buildenv/debian-12-cross-mipsel.sh | 1 +
ci/buildenv/debian-12-cross-ppc64le.sh| 1 +
ci/buildenv/debian-12-cross-s390x.sh | 1 +
ci/buildenv/debian-12.sh | 1 +
ci/buildenv/debian-sid-cross-aarch64.sh | 1 +
ci/buildenv/debian-sid-cross-armv6l.sh| 1 +
ci/buildenv/debian-sid-cross-armv7l.sh| 1 +
ci/buildenv/debian-sid-cross-i686.sh | 1 +
ci/buildenv/debian-sid-cross-mips64el.sh | 1 +
ci/buildenv/debian-sid-cross-mipsel.sh| 1 +
ci/buildenv/debian-sid-cross-ppc64le.sh | 1 +
ci/buildenv/debian-sid-cross-s390x.sh | 1 +
ci/buildenv/debian-sid.sh | 1 +
ci/buildenv/fedora-37.sh | 1 +
ci/buildenv/fedora-38-cross-mingw32.sh| 1 +
ci/buildenv/fedora-38-cross-mingw64.sh| 1 +
ci/buildenv/fedora-38.sh | 1 +
ci/buildenv/fedora-rawhide-cross-mingw32.sh | 1 +
ci/buildenv/fedora-rawhide-cross-mingw64.sh | 1 +
ci/buildenv/fedora-rawhide.sh | 1 +
ci/buildenv/opensuse-leap-15.sh | 1 +
ci/buildenv/opensuse-tumbleweed.sh| 1 +
ci/buildenv/ubuntu-2204.sh| 1 +
ci/containers/almalinux-8.Dockerfile | 1 +
ci/containers/centos-stream-8.Dockerfile | 1 +
ci/containers/centos-stream-9.Dockerfile | 1 +
ci/containers/debian-12-cross-aarch64.Dockerfile | 1 +
ci/containers/debian-12-cross-armv6l.Dockerfile | 1 +
ci/containers/debian-12-cross-armv7l.Dockerfile | 1 +
ci/containers/debian-12-cross-i686.Dockerfile | 1 +
ci/containers/debian-12-cross-mips64el.Dockerfile | 1 +
ci/containers/debian-12-cross-mipsel.Dockerfile | 1 +
ci/containers/debian-12-cross-ppc64le.Dockerfile | 1 +
ci/containers/debian-12-cross-s390x.Dockerfile| 1 +
ci/containers/debian-12.Dockerfile| 1 +
ci/containers/debian-sid-cross-aarch64.Dockerfile | 1 +
ci/containers/debian-sid-cross-armv6l.Dockerfile | 1 +
ci/containers/debian-sid-cross-armv7l.Dockerfile | 1 +
ci/containers/debian-sid-cross-i686.Dockerfile| 1 +
ci/containers/debian-sid-cross-mips64el.Dockerfile| 1 +
ci/containers/debian-sid-cross-mipsel.Dockerfile | 1 +
ci/containers/debian-sid-cross-ppc64le.Dockerfile | 1 +
ci/containers/debian-sid-cross-s390x.Dockerfile | 1 +
ci/containers/debian-sid.Dockerfile | 1 +
ci/containers/fedora-37.Dockerfile| 1 +
ci/containers/fedora-38-cross-mingw32.Dockerfile | 1 +
ci/containers/fedora-38-cross-mingw64.Dockerfile | 1 +
ci/containers/fedora-38.Dockerfile| 1 +
ci/containers/fedora-rawhide-cross-mingw32.Dockerfile | 1 +
ci/containers/fedora-rawhide-cross-mingw64.Dockerfile | 1 +
ci/containers/fedora-rawhide.Dockerfile | 1 +
ci/containers/opensuse-leap-15.Dockerfile | 1 +
ci/containers/opensuse-tumbleweed.Dockerfile | 1 +
ci/containers/ubuntu-2204.Dockerfile | 1 +
ci/lcitool/projects/libvirt.yml | 1 +
63 files changed, 63 insertions(+)
diff --git a/ci/buildenv/almalinux-8.sh b/ci/buildenv/almalinux-8.sh
index 086b4d946b..e3c0909557 100644
--- a/ci/buildenv/almalinux-8.sh
+++ b/ci/buildenv/almalinux-8.sh
@@ -45,6 +45,7 @@ function install_buildenv() {
libcap-ng-devel \
libcurl-devel \
libiscsi-devel \
+libnbd-devel \
libnl3-devel \
libpcap-devel \
libpciaccess-devel \
diff --git a/ci/buildenv/centos-stream-8.sh b/ci/buildenv/centos-stream-8.sh
index 6b3de502df..f6de8ee7f7 100644
--- a/ci/buildenv/centos-stream-8.sh
+++ b/ci/buildenv/centos-stream-8.sh
@@ -46,6 +46,7 @@ function install_buildenv() {
libcap-ng-devel \
libcurl-devel \
libiscsi-devel \
+libnbd-devel \
libnl3-devel \
libpcap-devel \
libpciaccess-devel \
diff --git a/ci/buildenv/centos-stream-9.sh b/ci/buildenv/centos-stream-9.sh
index 454e1f6322..d6657425a2 100644
--- a/ci/buildenv/centos-stream-9.sh
+++ b/ci/buildenv/centos-stream-9.sh
@@ -43,6 +43,7 @@ function install_buildenv
[libvirt PATCH v8 10/37] qemu: add functions to start and stop nbdkit
Add some helper functions to build a virCommand object and run the
nbdkit process for a given virStorageSource.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 250 +
src/qemu/qemu_nbdkit.h | 10 ++
2 files changed, 260 insertions(+)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 9a2a89224d..6bf962d0f1 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -24,6 +24,8 @@
#include "virerror.h"
#include "virlog.h"
#include "virpidfile.h"
+#include "virsecureerase.h"
+#include "virtime.h"
#include "virutil.h"
#include "qemu_block.h"
#include "qemu_conf.h"
@@ -666,6 +668,168 @@ qemuNbdkitInitStorageSource(qemuNbdkitCaps *caps,
}
+static int
+qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
+ virCommand *cmd)
+{
+g_autoptr(virURI) uri = qemuBlockStorageSourceGetURI(proc->source);
+g_autofree char *uristring = virURIFormat(uri);
+
+/* nbdkit plugin name */
+virCommandAddArg(cmd, "curl");
+if (proc->source->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP) {
+/* allow http to be upgraded to https via e.g. redirect */
+virCommandAddArgPair(cmd, "protocols", "http,https");
+} else {
+virCommandAddArgPair(cmd, "protocols",
+
virStorageNetProtocolTypeToString(proc->source->protocol));
+}
+virCommandAddArgPair(cmd, "url", uristring);
+
+if (proc->source->auth) {
+g_autoptr(virConnect) conn = virGetConnectSecret();
+g_autofree uint8_t *secret = NULL;
+size_t secretlen = 0;
+g_autofree char *password = NULL;
+int secrettype;
+virStorageAuthDef *authdef = proc->source->auth;
+
+virCommandAddArgPair(cmd, "user",
+ proc->source->auth->username);
+
+if ((secrettype =
virSecretUsageTypeFromString(proc->source->auth->secrettype)) < 0) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("invalid secret type %1$s"),
+ proc->source->auth->secrettype);
+return -1;
+}
+
+if (virSecretGetSecretString(conn,
+ >seclookupdef,
+ secrettype,
+ ,
+ ) < 0) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("failed to get auth secret for storage"));
+return -1;
+}
+
+/* ensure that the secret is a NULL-terminated string */
+password = g_strndup((char*)secret, secretlen);
+virSecureErase(secret, secretlen);
+
+/* for now, just report an error rather than passing the password in
+ * cleartext on the commandline */
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("Password not yet supported for nbdkit sources"));
+
+virSecureEraseString(password);
+
+return -1;
+}
+
+if (proc->source->ncookies > 0) {
+/* for now, just report an error rather than passing cookies in
+ * cleartext on the commandline */
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("Cookies not yet supported for nbdkit sources"));
+return -1;
+}
+
+if (proc->source->sslverify == VIR_TRISTATE_BOOL_NO) {
+virCommandAddArgPair(cmd, "sslverify", "false");
+}
+
+if (proc->source->timeout > 0) {
+g_autofree char *timeout = g_strdup_printf("%llu",
proc->source->timeout);
+virCommandAddArgPair(cmd, "timeout", timeout);
+}
+
+return 0;
+}
+
+
+static int
+qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
+ virCommand *cmd)
+{
+const char *user = NULL;
+virStorageNetHostDef *host = >source->hosts[0];
+g_autofree char *portstr = g_strdup_printf("%u", host->port);
+
+/* nbdkit plugin name */
+virCommandAddArg(cmd, "ssh");
+
+virCommandAddArgPair(cmd, "host", host->name);
+virCommandAddArgPair(cmd, "port", portstr);
+virCommandAddArgPair(cmd, "path", proc->source->path);
+
+if (proc->source->auth)
+user = proc->source->auth->username;
+else if (proc->source->ssh_user)
+user = proc->source->ssh_user;
+
+if (user)
+virCommandAddArgPair(cmd, "user", user);
+
+if (proc->source->ssh_host_key_check_di
[libvirt PATCH v8 14/37] qemu: log error output from nbdkit
log stderr and stdout from nbdkit into its own log so that
nbdkit-related issues can be debugged more easily.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 16 +++-
1 file changed, 15 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 6bf962d0f1..2d70e72c42 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -852,12 +852,23 @@ qemuNbdkitProcessStart(qemuNbdkitProcess *proc,
virTimeBackOffVar timebackoff;
g_autoptr(virURI) uri = NULL;
g_autofree char *uristring = NULL;
+g_autofree char *basename = g_strdup_printf("%s-nbdkit-%i", vm->def->name,
proc->source->id);
+int logfd = -1;
+g_autoptr(qemuLogContext) logContext = NULL;
if (!(cmd = qemuNbdkitProcessBuildCommand(proc)))
return -1;
+if (!(logContext = qemuLogContextNew(driver, vm, basename))) {
+virLastErrorPrefixMessage("%s", _("can't connect to virtlogd"));
+return -1;
+}
+
+logfd = qemuLogContextGetWriteFD(logContext);
+
VIR_DEBUG("starting nbdkit process for %s", proc->source->nodestorage);
-virCommandSetErrorBuffer(cmd, );
+virCommandSetErrorFD(cmd, );
+virCommandSetOutputFD(cmd, );
virCommandSetPidFile(cmd, proc->pidfile);
if (qemuExtDeviceLogCommand(driver, vm, cmd, "nbdkit") < 0)
@@ -899,6 +910,9 @@ qemuNbdkitProcessStart(qemuNbdkitProcess *proc,
if ((uri = qemuBlockStorageSourceGetURI(proc->source)))
uristring = virURIFormat(uri);
+if (qemuLogContextReadFiltered(logContext, , 1024) < 0)
+VIR_WARN("Unable to read from nbdkit log");
+
virReportError(VIR_ERR_OPERATION_FAILED,
_("Failed to connect to nbdkit for '%1$s': %2$s"),
NULLSTR(uristring), NULLSTR(errbuf));
--
2.41.0
[libvirt PATCH v8 28/37] schema: add password configuration for ssh disk
Right now, ssh network disks are not usable. There is some basic support in libvirt that is meant to support disk chains that have backing disks located at ssh urls, but there is no real way for a user to configure a ssh-based disk. This commit allows users to configure an ssh disk with password authentication. Implementation will follow. Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 27 ++- src/conf/schemas/domaincommon.rng | 23 ++- 2 files changed, 36 insertions(+), 14 deletions(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 68f54ab3ed..39d4230ec0 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -2778,7 +2778,7 @@ paravirtualized driver is specified via the ``disk`` element. ``network`` The ``protocol`` attribute specifies the protocol to access to the requested image. Possible values are "nbd", "iscsi", "rbd", "sheepdog", - "gluster", "vxhs", "nfs", "http", "https", "ftp", ftps", or "tftp". + "gluster", "vxhs", "nfs", "http", "https", "ftp", ftps", "tftp", or "ssh". For any ``protocol`` other than ``nbd`` an additional attribute ``name`` is mandatory to specify which volume/image will be used. @@ -2930,18 +2930,19 @@ paravirtualized driver is specified via the ``disk`` element. ``auth`` :since:`Since libvirt 3.9.0` , the ``auth`` element is supported for a disk ``type`` "network" that is using a ``source`` element with the - ``protocol`` attributes "rbd" or "iscsi". If present, the ``auth`` element - provides the authentication credentials needed to access the source. It - includes a mandatory attribute ``username``, which identifies the username - to use during authentication, as well as a sub-element ``secret`` with - mandatory attribute ``type``, to tie back to a `libvirt secret - object `__ that holds the actual password or other - credentials (the domain XML intentionally does not expose the password, - only the reference to the object that does manage the password). Known - secret types are "ceph" for Ceph RBD network sources and "iscsi" for CHAP - authentication of iSCSI targets. Both will require either a ``uuid`` - attribute with the UUID of the secret object or a ``usage`` attribute - matching the key that was specified in the secret object. + ``protocol`` attributes "rbd", "iscsi", or "ssh". If present, the + ``auth`` element provides the authentication credentials needed to access + the source. It includes a mandatory attribute ``username``, which + identifies the username to use during authentication, as well as a + sub-element ``secret`` with mandatory attribute ``type``, to tie back to + a `libvirt secret object `__ that holds the actual + password or other credentials (the domain XML intentionally does not + expose the password, only the reference to the object that does manage + the password). Known secret types are "ceph" for Ceph RBD network sources + and "iscsi" for CHAP authentication of iSCSI targets. Both will require + either a ``uuid`` attribute with the UUID of the secret object or a + ``usage`` attribute matching the key that was specified in the secret + object. ``encryption`` :since:`Since libvirt 3.9.0` , the ``encryption`` can be a sub-element of the ``source`` element for encrypted storage sources. If present, diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index 4a475f5c36..cd838a475c 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2172,6 +2172,27 @@ + + + + + +ssh + + + + + + + + + + + + + + + @@ -2179,7 +2200,6 @@ sheepdog tftp -ssh @@ -2289,6 +2309,7 @@ + -- 2.41.0
[libvirt PATCH v8 25/37] qemu: Monitor nbdkit process for exit
Adds the ability to monitor the nbdkit process so that we can take
action in case the child exits unexpectedly.
When the nbdkit process exits, we pause the vm, restart nbdkit, and then
resume the vm. This allows the vm to continue working in the event of a
nbdkit failure.
Eventually we may want to generalize this functionality since we may
need something similar for e.g. qemu-storage-daemon, etc.
The process is monitored with the pidfd_open() syscall if it exists
(since linux 5.3). Otherwise it resorts to checking whether the process
is alive once a second. The one-second time period was chosen somewhat
arbitrarily.
Signed-off-by: Jonathon Jongsma
---
meson.build | 11 +++
src/qemu/qemu_domain.c | 1 +
src/qemu/qemu_domain.h | 1 +
src/qemu/qemu_driver.c | 16
src/qemu/qemu_nbdkit.c | 169 ---
src/qemu/qemu_nbdkit.h | 8 +-
src/qemu/qemu_process.c | 15 +++-
src/qemu/qemu_process.h | 3 +
tests/meson.build| 6 +-
tests/qemuxml2argvtest.c | 6 +-
10 files changed, 219 insertions(+), 17 deletions(-)
diff --git a/meson.build b/meson.build
index 965ada483b..18ec312ee6 100644
--- a/meson.build
+++ b/meson.build
@@ -682,6 +682,13 @@ symbols = [
[ 'sched.h', 'cpu_set_t' ],
]
+if host_machine.system() == 'linux'
+ symbols += [
+# process management
+[ 'sys/syscall.h', 'SYS_pidfd_open' ],
+ ]
+endif
+
foreach symbol : symbols
if cc.has_header_symbol(symbol[0], symbol[1], args: '-D_GNU_SOURCE', prefix:
symbol.get(2, ''))
conf.set('WITH_DECL_@0@'.format(symbol[1].to_upper()), 1)
@@ -2002,6 +2009,9 @@ endif
conf.set_quoted('TLS_PRIORITY', get_option('tls_priority'))
+if conf.has('WITH_DECL_SYS_PIDFD_OPEN')
+ conf.set('WITH_NBDKIT', 1)
+endif
# Various definitions
@@ -2259,6 +2269,7 @@ misc_summary = {
'firewalld-zone': conf.has('WITH_FIREWALLD_ZONE'),
'nss': conf.has('WITH_NSS'),
'numad': conf.has('WITH_NUMAD'),
+ 'nbdkit': conf.has('WITH_NBDKIT'),
'Init script': init_script,
'Char device locks': chrdev_lock_files,
'Loader/NVRAM': loader_res,
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 46fe5a1cf4..2b8ece8f19 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -11511,6 +11511,7 @@ qemuProcessEventFree(struct qemuProcessEvent *event)
case QEMU_PROCESS_EVENT_PR_DISCONNECT:
case QEMU_PROCESS_EVENT_UNATTENDED_MIGRATION:
case QEMU_PROCESS_EVENT_RESET:
+case QEMU_PROCESS_EVENT_NBDKIT_EXITED:
case QEMU_PROCESS_EVENT_LAST:
break;
}
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h
index ddd20e67b4..f018b45eb6 100644
--- a/src/qemu/qemu_domain.h
+++ b/src/qemu/qemu_domain.h
@@ -465,6 +465,7 @@ typedef enum {
QEMU_PROCESS_EVENT_MEMORY_DEVICE_SIZE_CHANGE,
QEMU_PROCESS_EVENT_UNATTENDED_MIGRATION,
QEMU_PROCESS_EVENT_RESET,
+QEMU_PROCESS_EVENT_NBDKIT_EXITED,
QEMU_PROCESS_EVENT_LAST
} qemuProcessEventType;
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index ad8428948b..c4236b872f 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -4040,6 +4040,19 @@ processResetEvent(virQEMUDriver *driver,
}
+static void
+processNbdkitExitedEvent(virDomainObj *vm,
+ qemuNbdkitProcess *nbdkit)
+{
+if (virDomainObjBeginJob(vm, VIR_JOB_MODIFY) < 0)
+return;
+
+qemuNbdkitProcessRestart(nbdkit, vm);
+
+virDomainObjEndJob(vm);
+}
+
+
static void qemuProcessEventHandler(void *data, void *opaque)
{
struct qemuProcessEvent *processEvent = data;
@@ -4097,6 +4110,9 @@ static void qemuProcessEventHandler(void *data, void
*opaque)
case QEMU_PROCESS_EVENT_RESET:
processResetEvent(driver, vm);
break;
+case QEMU_PROCESS_EVENT_NBDKIT_EXITED:
+processNbdkitExitedEvent(vm, processEvent->data);
+break;
case QEMU_PROCESS_EVENT_LAST:
break;
}
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index df638e99c0..b9ced1e5cf 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -19,6 +19,7 @@
#include
#include
+#include
#include "vircommand.h"
#include "virerror.h"
@@ -33,6 +34,7 @@
#include "qemu_nbdkit.h"
#define LIBVIRT_QEMU_NBDKITPRIV_H_ALLOW
#include "qemu_nbdkitpriv.h"
+#include "qemu_process.h"
#include "qemu_security.h"
#include
@@ -41,6 +43,12 @@
VIR_LOG_INIT("qemu.nbdkit");
+#if WITH_NBDKIT
+# define WITHOUT_NBDKIT_UNUSED
+#else
+# define WITHOUT_NBDKIT_UNUSED G_GNUC_UNUSED
+#endif
+
VIR_ENUM_IMPL(qemuNbdkitCaps,
QEMU_NBDKIT_CAPS_LAST,
/* 0 */
@@ -611,6 +619,116 @@ qemuNbdkitCapsCacheNew(const char *cachedir)
}
+void
+qemuNbdkitProcessRestart(qemuNbdkitProcess *proc,
+ virDomainObj *vm)
+{
+qemuDomainObjPrivate *vmpriv = vm->privateData;
+virQEMUDriver *driver = vmpriv->driver;
[libvirt PATCH v8 27/37] qemu: try to connect to nbdkit early to detect errors
When using nbdkit to serve a network disk source, the nbdkit process
will start and wait for an nbd connection before actually attempting to
connect to the (remote) disk location. Because of this, nbdkit will not
report an error until after qemu is launched and tries to read from the
disk. This results in a fairly user-unfriendly error saying that qemu
was unable to start because "Requested export not available".
Ideally we'd like to be able to tell the user *why* the export is not
available, but this sort of information is only available to nbdkit, not
qemu. It could be because the url was incorrect, or because of an
authentication failure, or one of many other possibilities.
To make this friendlier for users and easier to detect
misconfigurations, try to connect to nbdkit immediately after starting
nbdkit and before we try to start qemu. This requires adding a
dependency on libnbd. If an error occurs when connecting to nbdkit, read
back from the nbdkit error log and provide that information in the error
report from qemuNbdkitProcessStart().
User-visible change demonstrated below:
Previous error:
$ virsh start nbdkit-test
2023-01-18 19:47:45.778+: 30895: error :
virNetClientProgramDispatchError:172 : internal
error: process exited while connecting to monitor:
2023-01-18T19:47:45.704658Z
qemu-system-x86_64: -blockdev {"driver":"nbd","server":{"type":"unix",
"path":"/var/lib/libvirt/qemu/domain-1-nbdkit-test/nbdkit-libvirt-1-storage.socket"},
"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}:
Requested export not
available
error: Failed to start domain 'nbdkit-test'
error: internal error: process exited while connecting to monitor:
2023-01-18T19:47:45.704658Z
qemu-system-x86_64: -blockdev {"driver":"nbd","server":{"type":"unix",
"path":"/var/lib/libvirt/qemu/domain-1-nbdkit-test/nbdkit-libvirt-1-storage.socket"},
"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}:
Requested export not
available
After this change:
$ virsh start nbdkit-test
2023-01-18 19:44:36.242+: 30895: error :
virNetClientProgramDispatchError:172 : internal
error: Failed to connect to nbdkit for
'http://localhost:/nonexistent.iso': nbdkit: curl[1]:
error: problem doing HEAD request to fetch size of URL
[http://localhost:/nonexistent.iso]:
HTTP response code said error: The requested URL returned error: 404
error: Failed to start domain 'nbdkit-test'
error: internal error: Failed to connect to nbdkit for
'http://localhost:/nonexistent.iso]:
error: problem doing HEAD request to fetch size of URL
[http://localhost:/nonexistent.iso]:
HTTP response code said error: The requested URL returned error: 404
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
meson.build| 7 +++
meson_options.txt | 1 +
src/qemu/meson.build | 1 +
src/qemu/qemu_nbdkit.c | 23 +++
4 files changed, 32 insertions(+)
diff --git a/meson.build b/meson.build
index 18ec312ee6..132269addc 100644
--- a/meson.build
+++ b/meson.build
@@ -1002,6 +1002,12 @@ endif
libiscsi_version = '1.18.0'
libiscsi_dep = dependency('libiscsi', version: '>=' + libiscsi_version,
required: get_option('libiscsi'))
+libnbd_version = '1.0'
+libnbd_dep = dependency('libnbd', version: '>=' + libnbd_version, required:
get_option('libnbd'))
+if libnbd_dep.found()
+ conf.set('WITH_LIBNBD', 1)
+endif
+
libnl_version = '3.0'
if not get_option('libnl').disabled() and host_machine.system() == 'linux'
libnl_dep = dependency('libnl-3.0', version: '>=' + libnl_version, required:
get_option('libnl'))
@@ -2219,6 +2225,7 @@ libs_summary = {
'glusterfs': glusterfs_dep.found(),
'libiscsi': libiscsi_dep.found(),
'libkvm': libkvm_dep.found(),
+ 'libnbd': libnbd_dep.found(),
'libnl': libnl_dep.found(),
'libparted': libparted_dep.found(),
'libpcap': libpcap_dep.found(),
diff --git a/meson_options.txt b/meson_options.txt
index 9174c4021c..ba6e49afc5 100644
--- a/meson_options.txt
+++ b/meson_options.txt
@@ -25,6 +25,7 @@ option('curl', type: 'feature', value: 'auto', description:
'curl support')
option('fuse', type: 'feature', value: 'auto', description: 'fuse support')
option('glusterfs', type: 'feature', value: 'auto', description: 'glusterfs
support')
option('libiscsi', type: 'feature', value: 'auto', description: 'libiscsi
support')
+option('libnbd', type: 'feature', value: 'auto', description: 'libnbd support')
option('libnl', type: 'feature', value: 'auto', description: 'libnl support')
option('libpcap', type: 'feature', value: 'auto', description: 'libpcap
support')
option('
[libvirt PATCH v8 06/37] qemu: implement persistent file cache for nbdkit caps
Implement the loadFile and saveFile virFileCacheHandlers callbacks so
that nbdkit capabilities are cached perstistently across daemon
restarts. The format and implementation is modeled on the qemu
capabilities, but simplified slightly.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
po/POTFILES| 1 +
src/qemu/qemu_nbdkit.c | 226 -
2 files changed, 225 insertions(+), 2 deletions(-)
diff --git a/po/POTFILES b/po/POTFILES
index 5d6ec195b4..6167f98ac5 100644
--- a/po/POTFILES
+++ b/po/POTFILES
@@ -181,6 +181,7 @@ src/qemu/qemu_monitor.c
src/qemu/qemu_monitor_json.c
src/qemu/qemu_monitor_text.c
src/qemu/qemu_namespace.c
+src/qemu/qemu_nbdkit.c
src/qemu/qemu_passt.c
src/qemu/qemu_process.c
src/qemu/qemu_qapi.c
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 9828f562f7..97612b637f 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -312,11 +312,233 @@ virNbdkitCapsNewData(const char *binary,
}
+static int
+qemuNbdkitCapsValidateBinary(qemuNbdkitCaps *nbdkitCaps,
+ xmlXPathContextPtr ctxt)
+{
+g_autofree char *str = NULL;
+
+if (!(str = virXPathString("string(./path)", ctxt))) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("missing path in nbdkit capabilities cache"));
+return -1;
+}
+
+if (STRNEQ(str, nbdkitCaps->path)) {
+virReportError(VIR_ERR_INTERNAL_ERROR,
+ _("Expected caps for '%1$s' but saw '%2$s'"),
+ nbdkitCaps->path, str);
+return -1;
+}
+
+return 0;
+}
+
+
+static int
+qemuNbdkitCapsParseFlags(qemuNbdkitCaps *nbdkitCaps,
+ xmlXPathContextPtr ctxt)
+{
+g_autofree xmlNodePtr *nodes = NULL;
+size_t i;
+int n;
+
+if ((n = virXPathNodeSet("./flag", ctxt, )) < 0) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("failed to parse qemu capabilities flags"));
+return -1;
+}
+
+VIR_DEBUG("Got flags %d", n);
+for (i = 0; i < n; i++) {
+unsigned int flag;
+
+if (virXMLPropEnum(nodes[i], "name", qemuNbdkitCapsTypeFromString,
+ VIR_XML_PROP_REQUIRED, ) < 0)
+return -1;
+
+qemuNbdkitCapsSet(nbdkitCaps, flag);
+}
+
+return 0;
+}
+
+
+/*
+ * Parsing a doc that looks like
+ *
+ *
+ * /some/path
+ * 234235253
+ * 234235253
+ * 234235253
+ * 234235253
+ * 1002016
+ *
+ *
+ * ...
+ *
+ *
+ * Returns 0 on success, 1 if outdated, -1 on error
+ */
+static int
+qemuNbdkitCapsLoadCache(qemuNbdkitCaps *nbdkitCaps,
+const char *filename)
+{
+g_autoptr(xmlDoc) doc = NULL;
+g_autoptr(xmlXPathContext) ctxt = NULL;
+long long int l;
+
+if (!(doc = virXMLParse(filename, NULL, NULL, "nbdkitCaps", , NULL,
false)))
+return -1;
+
+if (virXPathLongLong("string(./selfctime)", ctxt, ) < 0) {
+VIR_DEBUG("missing selfctime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->libvirtCtime = (time_t)l;
+
+nbdkitCaps->libvirtVersion = 0;
+virXPathUInt("string(./selfvers)", ctxt, >libvirtVersion);
+
+if (nbdkitCaps->libvirtCtime != virGetSelfLastChanged() ||
+nbdkitCaps->libvirtVersion != LIBVIR_VERSION_NUMBER) {
+VIR_DEBUG("Outdated capabilities in %s: libvirt changed (%lld vs %lld,
%lu vs %lu), stopping load",
+ nbdkitCaps->path,
+ (long long)nbdkitCaps->libvirtCtime,
+ (long long)virGetSelfLastChanged(),
+ (unsigned long)nbdkitCaps->libvirtVersion,
+ (unsigned long)LIBVIR_VERSION_NUMBER);
+return 1;
+}
+
+if (qemuNbdkitCapsValidateBinary(nbdkitCaps, ctxt) < 0)
+return -1;
+
+if (virXPathLongLong("string(./nbdkitctime)", ctxt, ) < 0) {
+VIR_DEBUG("missing nbdkitctime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->ctime = (time_t)l;
+
+if (virXPathLongLong("string(./plugindirmtime)", ctxt, ) < 0) {
+VIR_DEBUG("missing plugindirmtime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->pluginDirMtime = (time_t)l;
+
+if (virXPathLongLong("string(./filterdirmtime)", ctxt, ) < 0) {
+VIR_DEBUG("missing filterdirmtime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->filterDirMtime = (time_t)l;
+
+if (qemuNbdkitCapsParseFlags(nbdkitCaps, ctxt) < 0)
+return -1;
+
+if ((nbdkitCaps->version = virXPathString("string(./version)", ctxt)) ==
NULL) {
+VIR_DEBUG("missing
[libvirt PATCH v8 17/37] qemu: include nbdkit state in private xml
Add xml to the private data for a disk source to represent the nbdkit
process so that the state can be re-created if the libvirt daemon is
restarted. Format:
/path/to/nbdkit.pid
/path/to/nbdkit.socket
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c| 52 +
src/qemu/qemu_nbdkit.c| 71 +++
src/qemu/qemu_nbdkit.h| 8 +++
src/qemu/qemu_process.c | 6 ++
tests/qemustatusxml2xmldata/modern-in.xml | 4 ++
5 files changed, 141 insertions(+)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 951f3127d9..8429ce1028 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -1958,6 +1958,33 @@
qemuStorageSourcePrivateDataAssignSecinfo(qemuDomainSecretInfo **secinfo,
}
+static int
+qemuStorageSourcePrivateDataParseNbdkit(xmlNodePtr node,
+xmlXPathContextPtr ctxt,
+virStorageSource *src)
+{
+g_autofree char *pidfile = NULL;
+g_autofree char *socketfile = NULL;
+VIR_XPATH_NODE_AUTORESTORE(ctxt);
+
+ctxt->node = node;
+
+if (!(pidfile = virXPathString("string(./pidfile)", ctxt))) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing nbdkit
pidfile"));
+return -1;
+}
+
+if (!(socketfile = virXPathString("string(./socketfile)", ctxt))) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing nbdkit
socketfile"));
+return -1;
+}
+
+qemuNbdkitReconnectStorageSource(src, pidfile, socketfile);
+
+return 0;
+}
+
+
static int
qemuStorageSourcePrivateDataParse(xmlXPathContextPtr ctxt,
virStorageSource *src)
@@ -1971,6 +1998,7 @@ qemuStorageSourcePrivateDataParse(xmlXPathContextPtr ctxt,
bool fdsetPresent = false;
unsigned int fdSetID;
int enccount;
+xmlNodePtr nbdkitnode = NULL;
src->nodestorage =
virXPathString("string(./nodenames/nodename[@type='storage']/@name)", ctxt);
src->nodeformat =
virXPathString("string(./nodenames/nodename[@type='format']/@name)", ctxt);
@@ -2036,6 +2064,10 @@ qemuStorageSourcePrivateDataParse(xmlXPathContextPtr
ctxt,
virTristateBoolTypeFromString(thresholdEventWithIndex) ==
VIR_TRISTATE_BOOL_YES)
src->thresholdEventWithIndex = true;
+if ((nbdkitnode = virXPathNode("nbdkit", ctxt))) {
+if (qemuStorageSourcePrivateDataParseNbdkit(nbdkitnode, ctxt, src) < 0)
+return -1;
+}
return 0;
}
@@ -2053,6 +2085,23 @@ qemuStorageSourcePrivateDataFormatSecinfo(virBuffer *buf,
}
+static void
+qemuStorageSourcePrivateDataFormatNbdkit(qemuNbdkitProcess *nbdkit,
+ virBuffer *buf)
+{
+g_auto(virBuffer) childBuf = VIR_BUFFER_INIT_CHILD(buf);
+
+if (!nbdkit)
+return;
+
+virBufferEscapeString(, "%s\n",
+ nbdkit->pidfile);
+virBufferEscapeString(, "%s\n",
+ nbdkit->socketfile);
+virXMLFormatElement(buf, "nbdkit", NULL, );
+}
+
+
static int
qemuStorageSourcePrivateDataFormat(virStorageSource *src,
virBuffer *buf)
@@ -2102,6 +2151,9 @@ qemuStorageSourcePrivateDataFormat(virStorageSource *src,
if (src->thresholdEventWithIndex)
virBufferAddLit(buf, "\n");
+if (srcPriv)
+qemuStorageSourcePrivateDataFormatNbdkit(srcPriv->nbdkitProcess, buf);
+
return 0;
}
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 81861bae4a..e3923ab4f2 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -627,6 +627,77 @@ qemuNbdkitProcessNew(virStorageSource *source,
return nbdkit;
}
+/**
+ * qemuNbdkitReconnectStorageSource:
+ * @source: a storage source
+ * @pidfile: a pidfile for an nbdkit process
+ * @socketfile: the socket file associated with the nbdkit process
+ *
+ * This function constructs a new qemuNbdkitProcess object with the given
values for @pidfile and
+ * @socketfile and stores it in @source. This is intended to be called when
the libvirt daemon is
+ * restarted and tries to reconnect to all currently-running domains. Since
this function is called
+ * from the code that parses the current daemon state, it should not perform
any filesystem
+ * operations, or anything else that might fail. Additional initialization
will be done later by
+ * calling qemuNbdkitStorageSourceManageProcess().
+ */
+void
+qemuNbdkitReconnectStorageSource(virStorageSource *source,
+ const char *pidfile,
+ const char *socketfile)
+{
+qemuDomainStorageSourcePrivate *srcpriv =
qemuDomainStorageSourcePrivateFetch(s
[libvirt PATCH v8 36/37] rpm: update spec file for for nbdkit support
Require libnbd-devel when building the qemu driver, recommend nbdkit
packages.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
libvirt.spec.in | 8
1 file changed, 8 insertions(+)
diff --git a/libvirt.spec.in b/libvirt.spec.in
index b471afebb1..744fa5c88d 100644
--- a/libvirt.spec.in
+++ b/libvirt.spec.in
@@ -312,6 +312,7 @@ BuildRequires: util-linux
BuildRequires: libacl-devel
# From QEMU RPMs, used by virstoragetest
BuildRequires: /usr/bin/qemu-img
+BuildRequires: libnbd-devel
%endif
# For LVM drivers
BuildRequires: lvm2
@@ -768,6 +769,9 @@ Requires: numad
Recommends: passt
Recommends: passt-selinux
%endif
+Recommends: nbdkit
+Recommends: nbdkit-curl-plugin
+Recommends: nbdkit-ssh-plugin
%description daemon-driver-qemu
The qemu driver plugin for the libvirtd daemon, providing
@@ -1074,8 +1078,10 @@ exit 1
%if %{with_qemu}
%define arg_qemu -Ddriver_qemu=enabled
+%define arg_libnbd -Dlibnbd=enabled
%else
%define arg_qemu -Ddriver_qemu=disabled
+%define arg_libnbd -Dlibnbd=disabled
%endif
%if %{with_openvz}
@@ -1264,6 +1270,7 @@ export SOURCE_DATE_EPOCH=$(stat --printf='%Y'
%{_specdir}/libvirt.spec)
-Dyajl=enabled \
%{?arg_sanlock} \
-Dlibpcap=enabled \
+ %{?arg_libnbd} \
-Dlibnl=enabled \
-Daudit=enabled \
-Ddtrace=enabled \
@@ -1327,6 +1334,7 @@ export SOURCE_DATE_EPOCH=$(stat --printf='%Y'
%{_specdir}/libvirt.spec)
-Dglusterfs=disabled \
-Dhost_validate=disabled \
-Dlibiscsi=disabled \
+ -Dlibnbd=disabled \
-Dlibnl=disabled \
-Dlibpcap=disabled \
-Dlibssh2=disabled \
--
2.41.0
[libvirt PATCH v8 13/37] qemu: move qemuProcessReadLog() to qemuLogContext
This code can be used by the nbdkit implementation for reading back
filtered log data for error reporting. Move it to qemuLogContext so that
it can be shared. Renamed to qemuLogContextReadFiltered().
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_logcontext.c | 65
src/qemu/qemu_logcontext.h | 3 ++
src/qemu/qemu_process.c| 67 +-
3 files changed, 69 insertions(+), 66 deletions(-)
diff --git a/src/qemu/qemu_logcontext.c b/src/qemu/qemu_logcontext.c
index 0121ae5173..6e20f58bfa 100644
--- a/src/qemu/qemu_logcontext.c
+++ b/src/qemu/qemu_logcontext.c
@@ -21,6 +21,7 @@
#include "qemu_logcontext.h"
#include "viralloc.h"
#include "virlog.h"
+#include "virstring.h"
#include "virutil.h"
#include
@@ -236,6 +237,70 @@ qemuLogContextRead(qemuLogContext *ctxt,
}
+/**
+ * qemuLogContextFilter: Read and filter log for relevant messages
+ * @ctxt: the domain log context
+ * @msg: pointer to buffer to store the read messages in
+ * @max: maximum length of the message returned in @msg after filtering
+ *
+ * Reads log output from @ctxt and filters it. Skips messages not produced by
+ * the target executable or irrelevant messages. If @max is not zero, @buf will
+ * contain at most @max characters from the end of the log and @buf will start
+ * after a new line if possible.
+ */
+int
+qemuLogContextReadFiltered(qemuLogContext *ctxt,
+ char **msg,
+ size_t max)
+{
+char *buf;
+char *eol;
+char *filter_next;
+size_t skip;
+ssize_t got;
+
+if ((got = qemuLogContextRead(ctxt, )) < 0)
+return -1;
+
+/* Filter out debug messages from intermediate libvirt process */
+filter_next = buf;
+while ((eol = strchr(filter_next, '\n'))) {
+*eol = '\0';
+if (virLogProbablyLogMessage(filter_next) ||
+strstr(filter_next, "char device redirected to")) {
+skip = (eol + 1) - filter_next;
+memmove(filter_next, eol + 1, buf + got - eol);
+got -= skip;
+} else {
+filter_next = eol + 1;
+*eol = '\n';
+}
+}
+
+if (got > 0 &&
+buf[got - 1] == '\n') {
+buf[got - 1] = '\0';
+got--;
+}
+
+if (max > 0 && got > max) {
+skip = got - max;
+
+if (buf[skip - 1] != '\n' &&
+(eol = strchr(buf + skip, '\n')) &&
+!virStringIsEmpty(eol + 1))
+skip = eol + 1 - buf;
+
+memmove(buf, buf + skip, got - skip + 1);
+got -= skip;
+}
+
+buf = g_renew(char, buf, got + 1);
+*msg = buf;
+return 0;
+}
+
+
int
qemuLogContextGetWriteFD(qemuLogContext *ctxt)
{
diff --git a/src/qemu/qemu_logcontext.h b/src/qemu/qemu_logcontext.h
index 6ad60b7b4a..738e908bc3 100644
--- a/src/qemu/qemu_logcontext.h
+++ b/src/qemu/qemu_logcontext.h
@@ -32,6 +32,9 @@ int qemuLogContextWrite(qemuLogContext *ctxt,
const char *fmt, ...) G_GNUC_PRINTF(2, 3);
ssize_t qemuLogContextRead(qemuLogContext *ctxt,
char **msg);
+int qemuLogContextReadFiltered(qemuLogContext *ctxt,
+ char **msg,
+ size_t max);
int qemuLogContextGetWriteFD(qemuLogContext *ctxt);
void qemuLogContextMarkPosition(qemuLogContext *ctxt);
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index e03d0d8b4d..a77d2ba7de 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -1908,71 +1908,6 @@ qemuConnectMonitor(virQEMUDriver *driver,
}
-/**
- * qemuProcessReadLog: Read log file of a qemu VM
- * @logCtxt: the domain log context
- * @msg: pointer to buffer to store the read messages in
- * @max: maximum length of the message returned in @msg
- *
- * Reads log of a qemu VM. Skips messages not produced by qemu or irrelevant
- * messages. If @max is not zero, @msg will contain at most @max characters
- * from the end of the log and @msg will start after a new line if possible.
- *
- * Returns 0 on success or -1 on error
- */
-static int
-qemuProcessReadLog(qemuLogContext *logCtxt,
- char **msg,
- size_t max)
-{
-char *buf;
-ssize_t got;
-char *eol;
-char *filter_next;
-size_t skip;
-
-if ((got = qemuLogContextRead(logCtxt, )) < 0)
-return -1;
-
-/* Filter out debug messages from intermediate libvirt process */
-filter_next = buf;
-while ((eol = strchr(filter_next, '\n'))) {
-*eol = '\0';
-if (virLogProbablyLogMessage(filter_next) ||
-strstr(filter_next, "char device redirected to")) {
-skip = (eol + 1) - filter_next;
-memmove(filter_next, eol + 1, buf + got - eol);
-got -= skip;
-} else {
-
[libvirt PATCH v8 09/37] qemu: query nbdkit module dir from binary
Rather than hard-coding the nbdkit module directory, query the nbdkit
binary for the location to these directories. nbdkit provides a
--dump-config optiont that outputs this information and can be easily
parsed. We can also get the version from this output rather than
executing `nbdkit --version` separately.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 77 --
1 file changed, 60 insertions(+), 17 deletions(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 12c721f7f1..9a2a89224d 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -20,7 +20,6 @@
#include
#include
-#include "configmake.h"
#include "vircommand.h"
#include "virerror.h"
#include "virlog.h"
@@ -39,10 +38,6 @@
VIR_LOG_INIT("qemu.nbdkit");
-#define NBDKIT_MODDIR LIBDIR "/nbdkit"
-#define NBDKIT_PLUGINDIR NBDKIT_MODDIR "/plugins"
-#define NBDKIT_FILTERDIR NBDKIT_MODDIR "/filters"
-
VIR_ENUM_IMPL(qemuNbdkitCaps,
QEMU_NBDKIT_CAPS_LAST,
/* 0 */
@@ -56,6 +51,9 @@ struct _qemuNbdkitCaps {
char *path;
char *version;
+char *filterDir;
+char *pluginDir;
+
time_t ctime;
time_t libvirtCtime;
time_t pluginDirMtime;
@@ -129,18 +127,47 @@ qemuNbdkitCapsQueryFilters(qemuNbdkitCaps *nbdkit)
static int
-qemuNbdkitCapsQueryVersion(qemuNbdkitCaps *nbdkit)
+qemuNbdkitCapsQueryBuildConfig(qemuNbdkitCaps *nbdkit)
{
+size_t i;
+g_autofree char *output = NULL;
+g_auto(GStrv) lines = NULL;
+const char *line;
g_autoptr(virCommand) cmd = virCommandNewArgList(nbdkit->path,
- "--version",
+ "--dump-config",
NULL);
-virCommandSetOutputBuffer(cmd, >version);
+virCommandSetOutputBuffer(cmd, );
if (virCommandRun(cmd, NULL) != 0)
return -1;
-VIR_DEBUG("Got nbdkit version %s", nbdkit->version);
+lines = g_strsplit(output, "\n", 0);
+if (!lines)
+return -1;
+
+for (i = 0; (line = lines[i]); i++) {
+const char *key;
+const char *val;
+char *p;
+
+p = strchr(line, '=');
+if (!p)
+continue;
+
+*p = '\0';
+key = line;
+val = p + 1;
+
+VIR_DEBUG("Got nbdkit config value %s=%s", key, val);
+
+if (STREQ(key, "version"))
+nbdkit->version = g_strdup(val);
+else if (STREQ(key, "filterdir"))
+nbdkit->filterDir = g_strdup(val);
+else if (STREQ(key, "plugindir"))
+nbdkit->pluginDir = g_strdup(val);
+}
return 0;
}
@@ -152,6 +179,8 @@ qemuNbdkitCapsFinalize(GObject *object)
g_clear_pointer(>path, g_free);
g_clear_pointer(>version, g_free);
+g_clear_pointer(>filterDir, g_free);
+g_clear_pointer(>pluginDir, g_free);
g_clear_pointer(>flags, virBitmapFree);
G_OBJECT_CLASS(qemu_nbdkit_caps_parent_class)->finalize(object);
@@ -214,15 +243,15 @@ qemuNbdkitCapsQuery(qemuNbdkitCaps *caps)
return;
}
+qemuNbdkitCapsQueryBuildConfig(caps);
+qemuNbdkitCapsQueryPlugins(caps);
+qemuNbdkitCapsQueryFilters(caps);
+
caps->ctime = st.st_ctime;
-caps->filterDirMtime = qemuNbdkitGetDirMtime(NBDKIT_FILTERDIR);
-caps->pluginDirMtime = qemuNbdkitGetDirMtime(NBDKIT_PLUGINDIR);
+caps->filterDirMtime = qemuNbdkitGetDirMtime(caps->filterDir);
+caps->pluginDirMtime = qemuNbdkitGetDirMtime(caps->pluginDir);
caps->libvirtCtime = virGetSelfLastChanged();
caps->libvirtVersion = LIBVIR_VERSION_NUMBER;
-
-qemuNbdkitCapsQueryPlugins(caps);
-qemuNbdkitCapsQueryFilters(caps);
-qemuNbdkitCapsQueryVersion(caps);
}
@@ -267,9 +296,9 @@ virNbdkitCapsIsValid(void *data,
if (!nbdkitCaps->path)
return true;
-if (!virNbkditCapsCheckModdir(NBDKIT_PLUGINDIR,
nbdkitCaps->pluginDirMtime))
+if (!virNbkditCapsCheckModdir(nbdkitCaps->pluginDir,
nbdkitCaps->pluginDirMtime))
return false;
-if (!virNbkditCapsCheckModdir(NBDKIT_FILTERDIR,
nbdkitCaps->filterDirMtime))
+if (!virNbkditCapsCheckModdir(nbdkitCaps->filterDir,
nbdkitCaps->filterDirMtime))
return false;
if (nbdkitCaps->libvirtCtime != virGetSelfLastChanged() ||
@@ -421,12 +450,22 @@ qemuNbdkitCapsLoadCache(qemuNbdkitCaps *nbdkitCaps,
}
nbdkitCaps->ctime = (time_t)l;
+if ((nbdkitCaps->pluginDir = virXPathString("string(./plugindir)", ctxt))
== NULL) {
+VIR_DEBUG("missing plugindir in nbdkit capabilities cache");
+return -1;
+}
+
if (virXPathLongLong(&quo
[libvirt PATCH v8 31/37] qemu: implement knownHosts for ssh disks with nbdkit
For ssh disks that are served by nbdkit, use the configured value for
knownHosts and pass it to the nbdkit process.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/domain_conf.c| 8 ++
src/conf/storage_source_conf.c| 2 ++
src/conf/storage_source_conf.h| 2 ++
src/qemu/qemu_extdevice.c | 4 +--
src/qemu/qemu_hotplug.c | 4 +--
src/qemu/qemu_nbdkit.c| 25 +++
src/qemu/qemu_nbdkit.h| 6 +++--
.../disk-network-ssh-password.args.disk0 | 3 ++-
.../disk-network-ssh.args.disk0 | 3 ++-
.../disk-network-ssh-password.xml | 1 +
tests/qemuxml2argvdata/disk-network-ssh.xml | 1 +
11 files changed, 46 insertions(+), 13 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 8feaf5d055..842b6404b5 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7268,6 +7268,11 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
return -1;
}
}
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
+(tmpnode = virXPathNode("./knownHosts", ctxt))) {
+if (!(src->ssh_known_hosts_file = virXMLPropStringRequired(tmpnode,
"path")))
+return -1;
+}
return 0;
}
@@ -22274,6 +22279,9 @@ virDomainDiskSourceFormatNetwork(virBuffer *attrBuf,
if (src->timeout)
virBufferAsprintf(childBuf, "\n",
src->timeout);
+
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
src->ssh_known_hosts_file)
+virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
}
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index dcac3a8ff6..906bc36a9b 100644
--- a/src/conf/storage_source_conf.c
+++ b/src/conf/storage_source_conf.c
@@ -895,6 +895,7 @@ virStorageSourceCopy(const virStorageSource *src,
/* ssh config passthrough for libguestfs */
def->ssh_host_key_check_disabled = src->ssh_host_key_check_disabled;
def->ssh_user = g_strdup(src->ssh_user);
+def->ssh_known_hosts_file = g_strdup(src->ssh_known_hosts_file);
def->nfs_user = g_strdup(src->nfs_user);
def->nfs_group = g_strdup(src->nfs_group);
@@ -1170,6 +1171,7 @@ virStorageSourceClear(virStorageSource *def)
VIR_FREE(def->tlsHostname);
VIR_FREE(def->ssh_user);
+VIR_FREE(def->ssh_known_hosts_file);
VIR_FREE(def->nfs_user);
VIR_FREE(def->nfs_group);
diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
index f13e7c756a..8a9c7d07e2 100644
--- a/src/conf/storage_source_conf.h
+++ b/src/conf/storage_source_conf.h
@@ -410,6 +410,8 @@ struct _virStorageSource {
/* these must not be used apart from formatting the output JSON in the
qemu driver */
char *ssh_user;
bool ssh_host_key_check_disabled;
+/* additional ssh variables */
+char *ssh_known_hosts_file;
/* nfs_user and nfs_group store the strings passed in by the user for NFS
params.
* nfs_uid and nfs_gid represent the converted/looked up ID numbers which
are used
diff --git a/src/qemu/qemu_extdevice.c b/src/qemu/qemu_extdevice.c
index 42ecdf13d5..3cf3867056 100644
--- a/src/qemu/qemu_extdevice.c
+++ b/src/qemu/qemu_extdevice.c
@@ -297,11 +297,11 @@ qemuExtDevicesStop(virQEMUDriver *driver,
for (i = 0; i < def->ndisks; i++) {
virDomainDiskDef *disk = def->disks[i];
-qemuNbdkitStopStorageSource(disk->src);
+qemuNbdkitStopStorageSource(disk->src, vm);
}
if (def->os.loader && def->os.loader->nvram)
-qemuNbdkitStopStorageSource(def->os.loader->nvram);
+qemuNbdkitStopStorageSource(def->os.loader->nvram, vm);
}
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index dc06486922..bbc5177206 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -1036,7 +1036,7 @@ qemuDomainAttachDeviceDiskLiveInternal(virQEMUDriver
*driver,
if (virStorageSourceChainHasManagedPR(disk->src))
ignore_value(qemuHotplugRemoveManagedPR(vm, VIR_ASYNC_JOB_NONE));
-qemuNbdkitStopStorageSource(disk->src);
+qemuNbdkitStopStorageSource(disk->src, vm);
}
qemuDomainSecretDiskDestroy(disk);
qemuDomainCleanupStorageSourceFD(disk->src);
@@ -4496,7 +4496,7 @@ qemuDomainRemoveDiskDevice(virQEMUDriver *driver,
qemuHotplugRemoveManagedPR(vm, VIR_ASYNC_JOB_NONE) < 0)
goto cleanup;
-qemuNbdkitStopStorageSource(disk->src);
+qemuNbdkitStopStorageSource(disk->src, vm);
if (disk->transient) {
VIR_DEBUG("Removing transient overlay '%s' of disk '%s'",
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu
[libvirt PATCH v8 24/37] qemu: Add Taint for nbdkit restart failure
Since the restart handler will trigger at an arbitrary time (when the
nbdkit process crashes, for instance), it's difficult to provide
feedback to the user if the restart is unsuccessful. Rather than just
relying on a warning in the log, taint the domain so that there will be
a slightly more user-visible notification.
Signed-off-by: Jonathon Jongsma
---
src/conf/domain_conf.c | 2 ++
src/conf/domain_conf.h | 1 +
2 files changed, 3 insertions(+)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index bb4f1fdb94..8feaf5d055 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -87,6 +87,7 @@ VIR_ENUM_IMPL(virDomainTaint,
"custom-hypervisor-feature",
"deprecated-config",
"custom-device",
+ "nbdkit-restart",
);
VIR_ENUM_IMPL(virDomainTaintMessage,
@@ -105,6 +106,7 @@ VIR_ENUM_IMPL(virDomainTaintMessage,
N_("hypervisor feature autodetection override"),
N_("use of deprecated configuration settings"),
N_("custom device configuration"),
+ N_("nbdkit restart failed"),
);
VIR_ENUM_IMPL(virDomainVirt,
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index ca195a52d2..c0729905a8 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -3194,6 +3194,7 @@ typedef enum {
VIR_DOMAIN_TAINT_CUSTOM_HYPERVISOR_FEATURE, /* custom hypervisor feature
control */
VIR_DOMAIN_TAINT_DEPRECATED_CONFIG, /* Configuration that is marked
deprecated */
VIR_DOMAIN_TAINT_CUSTOM_DEVICE, /* hypervisor device config customized */
+VIR_DOMAIN_TAINT_NBDKIT_RESTART,/* nbdkit could not be restarted */
VIR_DOMAIN_TAINT_LAST
} virDomainTaintFlags;
--
2.41.0
[libvirt PATCH v8 32/37] schema: add keyfile configuration for ssh disks
Authenticating via key file to an ssh server is often preferable to logging in via password. In order to support this functionality add a new xml element for ssh disks that allows the user to specify a keyfile and username. Example configuration: ... ... Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 7 +++ src/conf/schemas/domaincommon.rng | 19 ++- 2 files changed, 25 insertions(+), 1 deletion(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 496a8ebfbe..baa2fdce7d 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3005,6 +3005,13 @@ paravirtualized driver is specified via the ``disk`` element. of these attributes is omitted, then that field is assumed to be the default value for the current system. If both ``user`` and ``group`` are intended to be default, then the entire element may be omitted. + + When using an ``ssh`` protocol, this element is used to enable + authentication via ssh keys. In this configuration, the element has two + attributes. The ``username`` attribute specifies the name of the user on + the remote server and the ``keyfile`` attribute specifies the path to the + keyfile. Note that this only works for ssh keys that are not + password-protected. ``reconnect`` For disk type ``vhostuser`` configures reconnect timeout if the connection is lost. This is set with the two mandatory attributes ``enabled`` and diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index ca43586323..47c5ee2a31 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2180,6 +2180,19 @@ + + + + + + + + + + + + + @@ -2199,11 +2212,15 @@ - + + + + + -- 2.41.0
[libvirt PATCH v8 18/37] util: secure erase virCommand send buffers
All users of virCommandSetSendBuffer() are using it to send sensitive
data to a child process. So, since these buffers contain sensitive
information, clear it with virSecureErase().
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/util/vircommand.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/util/vircommand.c b/src/util/vircommand.c
index 5f094c625a..899d413dd2 100644
--- a/src/util/vircommand.c
+++ b/src/util/vircommand.c
@@ -54,6 +54,7 @@
#include "virpidfile.h"
#include "virprocess.h"
#include "virbuffer.h"
+#include "virsecureerase.h"
#include "virthread.h"
#include "virstring.h"
@@ -1697,6 +1698,7 @@ virCommandFreeSendBuffers(virCommand *cmd)
for (i = 0; i < virCommandGetNumSendBuffers(cmd); i++) {
VIR_FORCE_CLOSE(cmd->sendBuffers[i].fd);
+virSecureErase(cmd->sendBuffers[i].buffer, cmd->sendBuffers[i].buflen);
VIR_FREE(cmd->sendBuffers[i].buffer);
}
VIR_FREE(cmd->sendBuffers);
--
2.41.0
[libvirt PATCH v8 30/37] schema: add configuration for host verification of ssh disks
In order to make ssh disks usable, we need to be able to validate a remote host. To do this, add a xml element for ssh disks to allow the user to specify a location for a file that contains known host keys. Implementation to follow. Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 8 src/conf/schemas/domaincommon.rng | 11 +++ 2 files changed, 19 insertions(+) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 39d4230ec0..496a8ebfbe 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3021,6 +3021,14 @@ paravirtualized driver is specified via the ``disk`` element. paused and will be rerun after a successful reconnect. After that time, any delayed requests and all future requests before a successful reconnect will immediately fail. If not set the default QEMU value is 0. + ``knownHosts`` + For storage accessed via the ``ssh`` protocol, this element configures a + path to a file that will be used to verify the remote host. This file + must contain the expected host key for the remote host or the connection + will fail. The location of the file is specified via the ``path`` + attribute. + :since:`Since 9.8.0` + For a "file" or "volume" disk type which represents a cdrom or floppy (the ``device`` attribute), it is possible to define policy what to do with the diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index cd838a475c..ca43586323 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2172,6 +2172,14 @@ + + + + + + + + @@ -2187,6 +2195,9 @@ + + + -- 2.41.0
[libvirt PATCH v8 05/37] qemu: implement basic virFileCache for nbdkit caps
Preparatory step for caching nbdkit capabilities. This patch implements
the newData and isValid virFileCacheHandlers callback functions.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 89 +-
src/qemu/qemu_nbdkit.h | 4 ++
2 files changed, 92 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 58828dd89a..9828f562f7 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -201,7 +201,7 @@ qemuNbdkitGetDirMtime(const char *moddir)
}
-G_GNUC_UNUSED static void
+static void
qemuNbdkitCapsQuery(qemuNbdkitCaps *caps)
{
struct stat st;
@@ -240,3 +240,90 @@ qemuNbdkitCapsSet(qemuNbdkitCaps *nbdkitCaps,
{
ignore_value(virBitmapSetBit(nbdkitCaps->flags, flag));
}
+
+
+static bool
+virNbkditCapsCheckModdir(const char *moddir,
+ time_t expectedMtime)
+{
+time_t mtime = qemuNbdkitGetDirMtime(moddir);
+
+if (mtime != expectedMtime) {
+VIR_DEBUG("Outdated capabilities for nbdkit: module directory '%s'
changed (%lld vs %lld)",
+ moddir, (long long)mtime, (long long)expectedMtime);
+return false;
+}
+return true;
+}
+
+
+static bool
+virNbdkitCapsIsValid(void *data,
+ void *privData G_GNUC_UNUSED)
+{
+qemuNbdkitCaps *nbdkitCaps = data;
+struct stat st;
+
+if (!nbdkitCaps->path)
+return true;
+
+if (!virNbkditCapsCheckModdir(NBDKIT_PLUGINDIR,
nbdkitCaps->pluginDirMtime))
+return false;
+if (!virNbkditCapsCheckModdir(NBDKIT_FILTERDIR,
nbdkitCaps->filterDirMtime))
+return false;
+
+if (nbdkitCaps->libvirtCtime != virGetSelfLastChanged() ||
+nbdkitCaps->libvirtVersion != LIBVIR_VERSION_NUMBER) {
+VIR_DEBUG("Outdated capabilities for '%s': libvirt changed (%lld vs
%lld, %lu vs %lu)",
+ nbdkitCaps->path,
+ (long long)nbdkitCaps->libvirtCtime,
+ (long long)virGetSelfLastChanged(),
+ (unsigned long)nbdkitCaps->libvirtVersion,
+ (unsigned long)LIBVIR_VERSION_NUMBER);
+return false;
+}
+
+if (stat(nbdkitCaps->path, ) < 0) {
+VIR_DEBUG("Failed to stat nbdkit binary '%s': %s",
+ nbdkitCaps->path,
+ g_strerror(errno));
+return false;
+}
+
+if (st.st_ctime != nbdkitCaps->ctime) {
+VIR_DEBUG("Outdated capabilities for '%s': nbdkit binary changed (%lld
vs %lld)",
+ nbdkitCaps->path,
+ (long long)st.st_ctime, (long long)nbdkitCaps->ctime);
+return false;
+}
+
+return true;
+}
+
+
+static void*
+virNbdkitCapsNewData(const char *binary,
+ void *privData G_GNUC_UNUSED)
+{
+qemuNbdkitCaps *caps = qemuNbdkitCapsNew(binary);
+qemuNbdkitCapsQuery(caps);
+
+return caps;
+}
+
+
+virFileCacheHandlers nbdkitCapsCacheHandlers = {
+.isValid = virNbdkitCapsIsValid,
+.newData = virNbdkitCapsNewData,
+.loadFile = NULL,
+.saveFile = NULL,
+.privFree = NULL,
+};
+
+
+virFileCache*
+qemuNbdkitCapsCacheNew(const char *cachedir)
+{
+g_autofree char *dir = g_build_filename(cachedir, "nbdkitcapabilities",
NULL);
+return virFileCacheNew(dir, "xml", );
+}
diff --git a/src/qemu/qemu_nbdkit.h b/src/qemu/qemu_nbdkit.h
index e191e1fdb4..4aba7c8455 100644
--- a/src/qemu/qemu_nbdkit.h
+++ b/src/qemu/qemu_nbdkit.h
@@ -21,6 +21,7 @@
#include "internal.h"
#include "virenum.h"
+#include "virfilecache.h"
typedef struct _qemuNbdkitCaps qemuNbdkitCaps;
@@ -38,6 +39,9 @@ VIR_ENUM_DECL(qemuNbdkitCaps);
qemuNbdkitCaps *
qemuNbdkitCapsNew(const char *path);
+virFileCache *
+qemuNbdkitCapsCacheNew(const char *cachedir);
+
bool
qemuNbdkitCapsGet(qemuNbdkitCaps *nbdkitCaps,
qemuNbdkitCapsFlags flag);
--
2.41.0
[libvirt PATCH v8 20/37] qemu: use nbdkit to serve network disks if available
For virStorageSource objects that contain an nbdkitProcess, start that
nbdkit process to serve that network drive and then pass the nbdkit
socket to qemu rather than sending the network url to qemu directly.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_block.c | 162 +++---
src/qemu/qemu_domain.c| 13 +-
src/qemu/qemu_extdevice.c | 62 +++
src/qemu/qemu_hotplug.c | 7 +
src/qemu/qemu_nbdkit.c| 42 +
src/qemu/qemu_nbdkit.h| 13 ++
...sk-cdrom-network-nbdkit.x86_64-latest.args | 42 +
.../disk-cdrom-network-nbdkit.xml | 1 +
...isk-network-http-nbdkit.x86_64-latest.args | 44 +
.../disk-network-http-nbdkit.xml | 1 +
...rce-curl-nbdkit-backing.x86_64-latest.args | 37
...isk-network-source-curl-nbdkit-backing.xml | 45 +
...work-source-curl-nbdkit.x86_64-latest.args | 49 ++
.../disk-network-source-curl-nbdkit.xml | 1 +
...isk-network-source-curl.x86_64-latest.args | 52 ++
.../disk-network-source-curl.xml | 71
...disk-network-ssh-nbdkit.x86_64-latest.args | 35
.../disk-network-ssh-nbdkit.xml | 1 +
tests/qemuxml2argvtest.c | 6 +
19 files changed, 618 insertions(+), 66 deletions(-)
create mode 100644
tests/qemuxml2argvdata/disk-cdrom-network-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-cdrom-network-nbdkit.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-http-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-network-http-nbdkit.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl-nbdkit-backing.x86_64-latest.args
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl-nbdkit-backing.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-network-source-curl-nbdkit.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-network-source-curl.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-ssh-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-network-ssh-nbdkit.xml
diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
index dcdf883926..1a2dc8ffb4 100644
--- a/src/qemu/qemu_block.c
+++ b/src/qemu/qemu_block.c
@@ -438,6 +438,32 @@ qemuBlockStorageSourceGetCURLProps(virStorageSource *src,
}
+static virJSONValue *
+qemuBlockStorageSourceGetNbdkitProps(virStorageSource *src)
+{
+qemuDomainStorageSourcePrivate *srcPriv =
QEMU_DOMAIN_STORAGE_SOURCE_PRIVATE(src);
+virJSONValue *ret = NULL;
+g_autoptr(virJSONValue) serverprops = NULL;
+virStorageNetHostDef host = { .transport = VIR_STORAGE_NET_HOST_TRANS_UNIX
};
+
+/* srcPriv->nbdkitProcess will already be initialized if we can use nbdkit
+ * to proxy this storage source */
+if (!(srcPriv && srcPriv->nbdkitProcess))
+return NULL;
+
+host.socket = srcPriv->nbdkitProcess->socketfile;
+serverprops = qemuBlockStorageSourceBuildJSONSocketAddress();
+
+if (!serverprops)
+return NULL;
+
+if (virJSONValueObjectAdd(, "a:server", , NULL) < 0)
+return NULL;
+
+return ret;
+}
+
+
static virJSONValue *
qemuBlockStorageSourceGetISCSIProps(virStorageSource *src,
bool onlytarget)
@@ -890,69 +916,75 @@ qemuBlockStorageSourceGetBackendProps(virStorageSource
*src,
return NULL;
case VIR_STORAGE_TYPE_NETWORK:
-switch ((virStorageNetProtocol) src->protocol) {
-case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-driver = "gluster";
-if (!(fileprops = qemuBlockStorageSourceGetGlusterProps(src,
onlytarget)))
-return NULL;
-break;
-
-case VIR_STORAGE_NET_PROTOCOL_VXHS:
-driver = "vxhs";
-if (!(fileprops = qemuBlockStorageSourceGetVxHSProps(src,
onlytarget)))
-return NULL;
-break;
-
-case VIR_STORAGE_NET_PROTOCOL_HTTP:
-case VIR_STORAGE_NET_PROTOCOL_HTTPS:
-case VIR_STORAGE_NET_PROTOCOL_FTP:
-case VIR_STORAGE_NET_PROTOCOL_FTPS:
-case VIR_STORAGE_NET_PROTOCOL_TFTP:
-driver = virStorageNetProtocolTypeToString(src->protocol);
-if (!(fileprops = qemuBlockStorageSourceGetCURLProps(src,
onlytarget)))
-return NULL;
-break;
-
-case VIR_STORAGE_NET_PROTOCOL_ISCSI:
-driver = "iscsi";
-if (!(fileprops = qemuBlockStorageSourceGetISCSIProps(src,
onlytarget)))
-return NULL;
-break;
-
-
[libvirt PATCH v8 26/37] qemu: improve error handling when restarting nbdkit
Change the return value for qemuNbdkitProcessRestart() and
qemuNbdkitStorageSourceManageProcess() to return an error status. The
main effect of this change is that when libvirt starts up and reconnects
to an already-running domain with an nbdkit-backed disk, it will return
an error if it fails to restart any nbdkit processes that are not found
to be running or if it fails to monitor any running nbdkit processes.
These failures will result in the domain being killed.
Signed-off-by: Jonathon Jongsma
---
src/qemu/qemu_driver.c | 3 ++-
src/qemu/qemu_nbdkit.c | 34 --
src/qemu/qemu_nbdkit.h | 4 ++--
src/qemu/qemu_process.c | 6 --
4 files changed, 24 insertions(+), 23 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index c4236b872f..a469b1d04a 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -4047,7 +4047,8 @@ processNbdkitExitedEvent(virDomainObj *vm,
if (virDomainObjBeginJob(vm, VIR_JOB_MODIFY) < 0)
return;
-qemuNbdkitProcessRestart(nbdkit, vm);
+if (qemuNbdkitProcessRestart(nbdkit, vm) < 0)
+virDomainObjTaint(vm, VIR_DOMAIN_TAINT_NBDKIT_RESTART);
virDomainObjEndJob(vm);
}
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index b9ced1e5cf..2ad34d6484 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -619,7 +619,7 @@ qemuNbdkitCapsCacheNew(const char *cachedir)
}
-void
+int
qemuNbdkitProcessRestart(qemuNbdkitProcess *proc,
virDomainObj *vm)
{
@@ -629,10 +629,7 @@ qemuNbdkitProcessRestart(qemuNbdkitProcess *proc,
/* clean up resources associated with process */
qemuNbdkitProcessStop(proc);
-if (qemuNbdkitProcessStart(proc, vm, driver) < 0) {
-VIR_DEBUG("Unable to restart nbkdit process");
-virDomainObjTaint(vm, VIR_DOMAIN_TAINT_NBDKIT_RESTART);
-}
+return qemuNbdkitProcessStart(proc, vm, driver);
}
@@ -775,7 +772,7 @@ qemuNbdkitReconnectStorageSource(virStorageSource *source,
}
-static void
+static int
qemuNbdkitStorageSourceManageProcessOne(virStorageSource *source,
virDomainObj *vm)
{
@@ -784,33 +781,31 @@ qemuNbdkitStorageSourceManageProcessOne(virStorageSource
*source,
qemuNbdkitProcess *proc;
if (!srcpriv)
-return;
+return 0;
proc = srcpriv->nbdkitProcess;
if (!proc)
-return;
+return 0;
if (!proc->caps)
proc->caps = qemuGetNbdkitCaps(vmpriv->driver);
if (proc->pid <= 0) {
if (virPidFileReadPath(proc->pidfile, >pid) < 0) {
-VIR_WARN("Unable to read pidfile '%s'", proc->pidfile);
-return;
+virReportError(VIR_ERR_INTERNAL_ERROR,
+ _("Unable to read pidfile '%1$s'"),
+ proc->pidfile);
+return -1;
}
}
if (virProcessKill(proc->pid, 0) < 0) {
VIR_DEBUG("nbdkit process %i is not alive", proc->pid);
-qemuNbdkitProcessRestart(proc, vm);
-return;
+return qemuNbdkitProcessRestart(proc, vm);
}
-if (qemuNbdkitProcessStartMonitor(proc, vm) < 0) {
-VIR_DEBUG("unable monitor nbdkit process");
-virDomainObjTaint(vm, VIR_DOMAIN_TAINT_NBDKIT_RESTART);
-}
+return qemuNbdkitProcessStartMonitor(proc, vm);
}
/**
@@ -822,13 +817,16 @@ qemuNbdkitStorageSourceManageProcessOne(virStorageSource
*source,
* @source. It is intended to be called after libvirt restarts and has loaded
its current state from
* disk and is attempting to re-connect to active domains.
*/
-void
+int
qemuNbdkitStorageSourceManageProcess(virStorageSource *source,
virDomainObj *vm)
{
virStorageSource *backing;
for (backing = source; backing != NULL; backing = backing->backingStore)
-qemuNbdkitStorageSourceManageProcessOne(backing, vm);
+if (qemuNbdkitStorageSourceManageProcessOne(backing, vm) < 0)
+return -1;
+
+return 0;
}
diff --git a/src/qemu/qemu_nbdkit.h b/src/qemu/qemu_nbdkit.h
index f33b049d38..a818ff65e1 100644
--- a/src/qemu/qemu_nbdkit.h
+++ b/src/qemu/qemu_nbdkit.h
@@ -68,7 +68,7 @@ qemuNbdkitStartStorageSource(virQEMUDriver *driver,
void
qemuNbdkitStopStorageSource(virStorageSource *src);
-void
+int
qemuNbdkitStorageSourceManageProcess(virStorageSource *src,
virDomainObj *vm);
@@ -100,7 +100,7 @@ qemuNbdkitProcessStart(qemuNbdkitProcess *proc,
virDomainObj *vm,
virQEMUDriver *driver);
-void
+int
qemuNbdkitProcessRestart(qemuNbdkitProcess *proc,
virDomainObj *vm);
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index a1c6cfd91e..ddf95a064f 100644
--- a/s
[libvirt PATCH v8 22/37] tests: add tests for nbdkit invocation
We were testing the arguments that were being passed to qemu when a disk was being served by nbdkit, but the arguments used to start nbdkit itself were not testable. This adds a test to ensure that we're invoking nbdkit correctly for various disk source definitions. Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- build-aux/syntax-check.mk | 2 +- src/qemu/qemu_nbdkit.c| 4 +- src/qemu/qemu_nbdkitpriv.h| 31 ++ tests/meson.build | 1 + .../disk-cdrom-network.args.disk0 | 6 + .../disk-cdrom-network.args.disk1 | 8 + .../disk-cdrom-network.args.disk1.pipe.778| 1 + .../disk-cdrom-network.args.disk2 | 8 + .../disk-cdrom-network.args.disk2.pipe.780| 1 + .../disk-network-http.args.disk0 | 6 + .../disk-network-http.args.disk1 | 5 + .../disk-network-http.args.disk2 | 6 + .../disk-network-http.args.disk2.pipe.778 | 1 + .../disk-network-http.args.disk3 | 7 + .../disk-network-http.args.disk3.pipe.780 | 1 + ...work-source-curl-nbdkit-backing.args.disk0 | 7 + ...ce-curl-nbdkit-backing.args.disk0.pipe.778 | 1 + .../disk-network-source-curl.args.disk0 | 7 + ...sk-network-source-curl.args.disk0.pipe.778 | 1 + .../disk-network-source-curl.args.disk1 | 7 + ...sk-network-source-curl.args.disk1.pipe.780 | 1 + .../disk-network-source-curl.args.disk2 | 7 + ...sk-network-source-curl.args.disk2.pipe.782 | 1 + .../disk-network-source-curl.args.disk3 | 6 + .../disk-network-source-curl.args.disk4 | 6 + .../disk-network-ssh.args.disk0 | 6 + tests/qemunbdkittest.c| 308 ++ 27 files changed, 444 insertions(+), 2 deletions(-) create mode 100644 src/qemu/qemu_nbdkitpriv.h create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk1 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk1.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk2 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk2.pipe.780 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk1 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk2 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk2.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk3 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk3.pipe.780 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl-nbdkit-backing.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl-nbdkit-backing.args.disk0.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk0.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk1 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk2 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.782 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk3 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk4 create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk0 create mode 100644 tests/qemunbdkittest.c diff --git a/build-aux/syntax-check.mk b/build-aux/syntax-check.mk index 64c1e2773e..ec04402133 100644 --- a/build-aux/syntax-check.mk +++ b/build-aux/syntax-check.mk @@ -1370,7 +1370,7 @@ exclude_file_name_regexp--sc_prohibit_close = \ (\.p[yl]$$|\.spec\.in$$|^docs/|^(src/util/vir(file|event)\.c|src/libvirt-stream\.c|tests/(vir.+mock\.c|commandhelper\.c|qemusecuritymock\.c)|tools/nss/libvirt_nss_(leases|macs)\.c)|tools/virt-qemu-qmp-proxy$$) exclude_file_name_regexp--sc_prohibit_empty_lines_at_EOF = \ - (^tests/(nodedevmdevctl|viracpi|virhostcpu|virpcitest|virstoragetest)data/|docs/js/.*\.js|docs/fonts/.*\.woff|\.diff|tests/virconfdata/no-newline\.conf$$) + (^tests/(nodedevmdevctl|viracpi|virhostcpu|virpcitest|virstoragetest|qemunbdkit)data/|docs/js/.*\.js|docs/fonts/.*\.woff|\.diff|tests/virconfdata/no-newline\.conf$$) exclude_file_name_regexp--sc_prohibit_fork_wrappers = \ (^(src/(util/(vircommand|virdaemon)|lxc/lxc_controller)|tests/testutils)\.c$$) diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c index 299d8824f2..df638e99c0 100644 --- a/src/qemu/qemu_nbdkit.c +++ b/src/qemu/qemu_nbdkit.c @@ -31,6 +31,8 @@ #include "qemu_domain.h" #include "qemu_extdevice.h" #include "qemu_nbdkit.h" +#define LIBVIRT_QEM
[libvirt PATCH v8 04/37] util: Allow virFileCache data to be any GObject
Since the libvirt documentation suggests to prefer GObject over
virObject, and since virObject is a GObject, change virFileCache to
allow GObjects as data.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/util/virfilecache.c | 14 --
src/util/virfilecache.h | 2 +-
2 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/util/virfilecache.c b/src/util/virfilecache.c
index c730de066e..6f698016a1 100644
--- a/src/util/virfilecache.c
+++ b/src/util/virfilecache.c
@@ -170,7 +170,7 @@ virFileCacheLoad(virFileCache *cache,
*data = g_steal_pointer();
cleanup:
-virObjectUnref(loadData);
+g_clear_pointer(, g_object_unref);
return ret;
}
@@ -207,7 +207,7 @@ virFileCacheNewData(virFileCache *cache,
return NULL;
if (virFileCacheSave(cache, name, data) < 0) {
-g_clear_pointer(, virObjectUnref);
+g_clear_object();
}
}
@@ -239,7 +239,7 @@ virFileCacheNew(const char *dir,
if (!(cache = virObjectNew(virFileCacheClass)))
return NULL;
-cache->table = virHashNew(virObjectUnref);
+cache->table = virHashNew(g_object_unref);
cache->dir = g_strdup(dir);
@@ -270,7 +270,7 @@ virFileCacheValidate(virFileCache *cache,
if (*data) {
VIR_DEBUG("Caching data '%p' for '%s'", *data, name);
if (virHashAddEntry(cache->table, name, *data) < 0) {
-g_clear_pointer(data, virObjectUnref);
+g_clear_pointer(data, g_object_unref);
}
}
}
@@ -300,7 +300,8 @@ virFileCacheLookup(virFileCache *cache,
data = virHashLookup(cache->table, name);
virFileCacheValidate(cache, name, );
-virObjectRef(data);
+if (data)
+g_object_ref(data);
virObjectUnlock(cache);
return data;
@@ -331,7 +332,8 @@ virFileCacheLookupByFunc(virFileCache *cache,
data = virHashSearch(cache->table, iter, iterData, );
virFileCacheValidate(cache, name, );
-virObjectRef(data);
+if (data)
+g_object_ref(data);
virObjectUnlock(cache);
return data;
diff --git a/src/util/virfilecache.h b/src/util/virfilecache.h
index c3bc0f529c..944741c0a7 100644
--- a/src/util/virfilecache.h
+++ b/src/util/virfilecache.h
@@ -48,7 +48,7 @@ typedef bool
* @priv: private data created together with cache
*
* Creates a new data based on the @name. The returned data must be
- * an instance of virObject.
+ * an instance of GObject.
*
* Returns data object or NULL on error.
*/
--
2.41.0
[libvirt PATCH v8 23/37] qemu: add test for authenticating a https network disk
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
tests/qemunbdkitdata/disk-network-source-curl.args.disk1 | 4 +++-
.../disk-network-source-curl.args.disk1.pipe.780 | 2 +-
.../disk-network-source-curl.args.disk1.pipe.782 | 1 +
tests/qemunbdkitdata/disk-network-source-curl.args.disk2 | 2 +-
.../disk-network-source-curl.args.disk2.pipe.784 | 1 +
.../disk-network-source-curl.x86_64-latest.args | 3 ++-
tests/qemuxml2argvdata/disk-network-source-curl.xml | 3 +++
7 files changed, 12 insertions(+), 4 deletions(-)
create mode 100644
tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
create mode 100644
tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
index fde6a4f533..d1288dd242 100644
--- a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
@@ -3,5 +3,7 @@ nbdkit \
--foreground curl \
protocols=https \
'url=https://https.example.org:8443/path/to/disk5.iso?foo=bar' \
-cookie=-779 \
+user=myname \
+password=-779 \
+cookie=-781 \
sslverify=false
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
index 20af4ae383..ccdd4033fc 100644
--- a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
@@ -1 +1 @@
-cookie1=cookievalue1; cookie2=cookievalue2
\ No newline at end of file
+iscsi-mycluster_myname-secret
\ No newline at end of file
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
new file mode 100644
index 00..20af4ae383
--- /dev/null
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
@@ -0,0 +1 @@
+cookie1=cookievalue1; cookie2=cookievalue2
\ No newline at end of file
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
index 88c9fa35a1..f1d0e1929e 100644
--- a/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
@@ -4,4 +4,4 @@ nbdkit \
--readonly curl \
protocols=http,https \
url=http://http.example.org:8080/path/to/disk2.iso \
-cookie=-781
+cookie=-783
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
new file mode 100644
index 00..5c035e84c5
--- /dev/null
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
@@ -0,0 +1 @@
+cookie1=cookievalue1; cookie2=cookievalue2; cookie3=cookievalue3
\ No newline at end of file
diff --git a/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
b/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
index cb0e5a92ea..f6ab5532cc 100644
--- a/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
+++ b/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
@@ -32,9 +32,10 @@
XDG_CONFIG_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.config \
-blockdev
'{"driver":"https","url":"https://https.example.org:8443/path/to/disk1.iso","cookie-secret":"libvirt-5-storage-httpcookie-secret0","node-name":"libvirt-5-storage","auto-read-only":true,"discard":"unmap"}'
\
-blockdev
'{"node-name":"libvirt-5-format","read-only":true,"driver":"raw","file":"libvirt-5-storage"}'
\
-device
'{"driver":"virtio-blk-pci","bus":"pci.0","addr":"0x3","drive":"libvirt-5-format","id":"virtio-disk0","bootindex":1}'
\
+-object
'{"qom-type":"secret","id":"libvirt-4-storage-auth-secret0","data":"9eao5F8qtkGt+seB1HYivWIxbtwUu6MQtg1zpj/oDtUsPr1q8wBYM91uEHCn6j/1","keyid":"masterKey0","iv":"AAECAwQFBgcICQoLDA0ODw==","format":"base64"}'
\
-object
'{"qom-type":"secret","id":"libvirt-4-storage-httpcookie-secret0","data":"BUU0KmnWfonHdjzhYhwVQZ5iTI1KweTJ22q8XWUVoBCVu1z70reDuczPBIabZtC3","keyid":"masterKey0","iv":"AAECAwQFBgcICQoLDA0ODw==","format":"base64"}'
\
-object
'{"qom-type":"secret","id":"libvirt-4-format-encryption-secret0","data":"9eao5F8qtkGt+seB1HYivWIxbtwUu6MQtg1zpj/oD
[libvirt PATCH v8 16/37] qemu: split qemuDomainSecretStorageSourcePrepare
This prepares encryption secrets and authentication secrets. When we add
nbdkit-backed network storage sources, we will not need to send
authentication secrets to qemu, since they will be sent to nbdkit
instead. So split this into two different functions.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c | 103 +
1 file changed, 62 insertions(+), 41 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 23608f95bd..951f3127d9 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -1398,38 +1398,70 @@
qemuDomainSecretStorageSourcePrepareCookies(qemuDomainObjPrivate *priv,
/**
- * qemuDomainSecretStorageSourcePrepare:
+ * qemuDomainSecretStorageSourcePrepareEncryption:
* @priv: domain private object
* @src: storage source struct to setup
- * @authalias: prefix of the alias for secret holding authentication data
- * @encalias: prefix of the alias for secret holding encryption password
+ * @alias: prefix of the alias for secret holding encryption password
*
- * Prepares data necessary for encryption and authentication of @src. The two
- * alias prefixes are provided since in the backing chain authentication
belongs
- * to the storage protocol data whereas encryption is relevant to the format
- * driver in qemu. The two will have different node names.
+ * Prepares data necessary for encryption of @src.
*
* Returns 0 on success; -1 on error while reporting an libvirt error.
*/
static int
-qemuDomainSecretStorageSourcePrepare(qemuDomainObjPrivate *priv,
- virStorageSource *src,
- const char *aliasprotocol,
- const char *aliasformat)
+qemuDomainSecretStorageSourcePrepareEncryption(qemuDomainObjPrivate *priv,
+ virStorageSource *src,
+ const char *alias)
{
qemuDomainStorageSourcePrivate *srcPriv;
-bool hasEnc = src->encryption && src->encryption->nsecrets > 0;
+size_t nsecrets = 0;
+size_t i;
-if (virStorageSourceIsEmpty(src))
+if (!(src->encryption && src->encryption->nsecrets > 0))
return 0;
-if (!src->auth && !hasEnc && src->ncookies == 0)
+if (virStorageSourceIsEmpty(src))
return 0;
-if (!(src->privateData = qemuDomainStorageSourcePrivateNew()))
-return -1;
+nsecrets = src->encryption->nsecrets;
+
+srcPriv = qemuDomainStorageSourcePrivateFetch(src);
+
+srcPriv->enccount = nsecrets;
+srcPriv->encinfo = g_new0(qemuDomainSecretInfo *, nsecrets);
+for (i = 0; i < nsecrets; ++i) {
+if (!(srcPriv->encinfo[i] = qemuDomainSecretInfoSetupFromSecret(priv,
alias,
+
"encryption", i,
+
VIR_SECRET_USAGE_TYPE_VOLUME,
+NULL,
+
>encryption->secrets[i]->seclookupdef)))
+return -1;
+}
+
+return 0;
+}
+
+
+/**
+ * qemuDomainSecretStorageSourcePrepareAuth:
+ * @priv: domain private object
+ * @src: storage source struct to setup
+ * @alias: prefix of the alias for secret holding authentication data
+ *
+ * Prepares data necessary for authentication of @src.
+ *
+ * Returns 0 on success; -1 on error while reporting an libvirt error.
+ */
+static int
+qemuDomainSecretStorageSourcePrepareAuth(qemuDomainObjPrivate *priv,
+ virStorageSource *src,
+ const char *alias)
+{
+qemuDomainStorageSourcePrivate *srcPriv;
+
+if (virStorageSourceIsEmpty(src))
+return 0;
-srcPriv = QEMU_DOMAIN_STORAGE_SOURCE_PRIVATE(src);
+srcPriv = qemuDomainStorageSourcePrivateFetch(src);
if (src->auth) {
virSecretUsageType usageType = VIR_SECRET_USAGE_TYPE_ISCSI;
@@ -1437,7 +1469,7 @@ qemuDomainSecretStorageSourcePrepare(qemuDomainObjPrivate
*priv,
if (src->protocol == VIR_STORAGE_NET_PROTOCOL_RBD)
usageType = VIR_SECRET_USAGE_TYPE_CEPH;
-if (!(srcPriv->secinfo = qemuDomainSecretInfoSetupFromSecret(priv,
aliasprotocol,
+if (!(srcPriv->secinfo = qemuDomainSecretInfoSetupFromSecret(priv,
alias,
"auth", 0,
usageType,
src->auth->username,
@@ -1445,26 +1477,10 @@
qemuDomainSecretStorageSourcePrepare(qemuDomainObjPrivate
[libvirt PATCH v8 07/37] qemu: use file cache for nbdkit caps
Add the virFileCache implementation for nbdkit capabilities to the qemu
driver. This allows us to determine whether nbdkit is installed and
which plugins are supported. it also has persistent caching and the
capabilities are re-queried whenever something changes.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_conf.h | 3 +++
src/qemu/qemu_driver.c | 3 +++
2 files changed, 6 insertions(+)
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index 4f610d86a1..a44985fb8b 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -318,6 +318,9 @@ struct _virQEMUDriver {
/* Immutable pointer, self-locking APIs */
virHashAtomic *migrationErrors;
+
+/* Immutable pointer, self-locking APIs */
+virFileCache *nbdkitCapsCache;
};
virQEMUDriverConfig *virQEMUDriverConfigNew(bool privileged,
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 5db42f0753..ad8428948b 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -845,6 +845,8 @@ qemuStateInitialize(bool privileged,
defsecmodel)))
goto error;
+qemu_driver->nbdkitCapsCache = qemuNbdkitCapsCacheNew(cfg->cacheDir);
+
/* If hugetlbfs is present, then we need to create a sub-directory within
* it, since we can't assume the root mount point has permissions that
* will let our spawned QEMU instances use it. */
@@ -1078,6 +1080,7 @@ qemuStateCleanup(void)
ebtablesContextFree(qemu_driver->ebtables);
VIR_FREE(qemu_driver->qemuImgBinary);
virObjectUnref(qemu_driver->domains);
+virObjectUnref(qemu_driver->nbdkitCapsCache);
if (qemu_driver->lockFD != -1)
virPidFileRelease(qemu_driver->config->stateDir, "driver",
qemu_driver->lockFD);
--
2.41.0
[libvirt PATCH v8 03/37] qemu: expand nbdkit capabilities
In order to add caching of the nbdkit capabilities, we will need to
compare against file modification times, etc. So look up this
information when creating the nbdkit caps.
Add a nbdkit_moddir build option to allow the builder to specify the
location to look for nbdkit plugins and filters.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 42 ++
1 file changed, 42 insertions(+)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index e4e8fd568e..58828dd89a 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -20,6 +20,7 @@
#include
#include
+#include "configmake.h"
#include "vircommand.h"
#include "virerror.h"
#include "virlog.h"
@@ -38,6 +39,10 @@
VIR_LOG_INIT("qemu.nbdkit");
+#define NBDKIT_MODDIR LIBDIR "/nbdkit"
+#define NBDKIT_PLUGINDIR NBDKIT_MODDIR "/plugins"
+#define NBDKIT_FILTERDIR NBDKIT_MODDIR "/filters"
+
VIR_ENUM_IMPL(qemuNbdkitCaps,
QEMU_NBDKIT_CAPS_LAST,
/* 0 */
@@ -51,6 +56,11 @@ struct _qemuNbdkitCaps {
char *path;
char *version;
+time_t ctime;
+time_t libvirtCtime;
+time_t pluginDirMtime;
+time_t filterDirMtime;
+unsigned int libvirtVersion;
virBitmap *flags;
};
@@ -175,9 +185,41 @@ qemuNbdkitCapsNew(const char *path)
}
+static time_t
+qemuNbdkitGetDirMtime(const char *moddir)
+{
+struct stat st;
+
+if (stat(moddir, ) < 0) {
+VIR_DEBUG("Failed to stat nbdkit module directory '%s': %s",
+ moddir,
+ g_strerror(errno));
+return 0;
+}
+
+return st.st_mtime;
+}
+
+
G_GNUC_UNUSED static void
qemuNbdkitCapsQuery(qemuNbdkitCaps *caps)
{
+struct stat st;
+
+if (stat(caps->path, ) < 0) {
+VIR_DEBUG("Failed to stat nbdkit binary '%s': %s",
+ caps->path,
+ g_strerror(errno));
+caps->ctime = 0;
+return;
+}
+
+caps->ctime = st.st_ctime;
+caps->filterDirMtime = qemuNbdkitGetDirMtime(NBDKIT_FILTERDIR);
+caps->pluginDirMtime = qemuNbdkitGetDirMtime(NBDKIT_PLUGINDIR);
+caps->libvirtCtime = virGetSelfLastChanged();
+caps->libvirtVersion = LIBVIR_VERSION_NUMBER;
+
qemuNbdkitCapsQueryPlugins(caps);
qemuNbdkitCapsQueryFilters(caps);
qemuNbdkitCapsQueryVersion(caps);
--
2.41.0
[libvirt PATCH v8 19/37] qemu: pass sensitive data to nbdkit via pipe
Rather than passing passwords and cookies (which could contain
passwords) to nbdkit via commandline arguments, use the alternate format
that nbdkit supports where we can specify a file descriptor which nbdkit
will read to get the password or cookies.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 54 ++
1 file changed, 34 insertions(+), 20 deletions(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index e3923ab4f2..22a67b0748 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -24,7 +24,6 @@
#include "virerror.h"
#include "virlog.h"
#include "virpidfile.h"
-#include "virsecureerase.h"
#include "virtime.h"
#include "virutil.h"
#include "qemu_block.h"
@@ -753,6 +752,29 @@ qemuNbdkitInitStorageSource(qemuNbdkitCaps *caps,
}
+static int
+qemuNbdkitCommandPassDataByPipe(virCommand *cmd,
+const char *argName,
+unsigned char **buf,
+size_t buflen)
+{
+g_autofree char *fdfmt = NULL;
+int fd = virCommandSetSendBuffer(cmd, buf, buflen);
+
+if (fd < 0)
+return -1;
+
+/* some nbdkit arguments accept a variation where nbdkit will read the data
+ * from a file descriptor, e.g. password=-FD */
+fdfmt = g_strdup_printf("-%i", fd);
+virCommandAddArgPair(cmd, argName, fdfmt);
+
+virCommandDoAsyncIO(cmd);
+
+return 0;
+}
+
+
static int
qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
virCommand *cmd)
@@ -775,7 +797,6 @@ qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
g_autoptr(virConnect) conn = virGetConnectSecret();
g_autofree uint8_t *secret = NULL;
size_t secretlen = 0;
-g_autofree char *password = NULL;
int secrettype;
virStorageAuthDef *authdef = proc->source->auth;
@@ -799,26 +820,19 @@ qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
return -1;
}
-/* ensure that the secret is a NULL-terminated string */
-password = g_strndup((char*)secret, secretlen);
-virSecureErase(secret, secretlen);
-
-/* for now, just report an error rather than passing the password in
- * cleartext on the commandline */
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("Password not yet supported for nbdkit sources"));
-
-virSecureEraseString(password);
-
-return -1;
+if (qemuNbdkitCommandPassDataByPipe(cmd, "password",
+, secretlen) < 0)
+return -1;
}
-if (proc->source->ncookies > 0) {
-/* for now, just report an error rather than passing cookies in
- * cleartext on the commandline */
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("Cookies not yet supported for nbdkit sources"));
-return -1;
+/* Create a pipe to send the cookies to the nbdkit process. */
+if (proc->source->ncookies) {
+g_autofree char *cookies =
qemuBlockStorageSourceGetCookieString(proc->source);
+
+if (qemuNbdkitCommandPassDataByPipe(cmd, "cookie",
+(unsigned char**),
+strlen(cookies)) < 0)
+return -1;
}
if (proc->source->sslverify == VIR_TRISTATE_BOOL_NO) {
--
2.41.0
[libvirt PATCH v8 11/37] Generalize qemuDomainLogContextNew()
Allow to specify a basename for the log file so that
qemuDomainLogContextNew() can be used to create log contexts for
secondary loggers.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c | 5 +++--
src/qemu/qemu_domain.h | 3 ++-
src/qemu/qemu_process.c | 2 +-
3 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 5a2eb4868a..d79f9879df 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -7119,7 +7119,8 @@ void qemuDomainObjCheckNetTaint(virQEMUDriver *driver,
qemuDomainLogContext *qemuDomainLogContextNew(virQEMUDriver *driver,
- virDomainObj *vm)
+ virDomainObj *vm,
+ const char *basename)
{
g_autoptr(virQEMUDriverConfig) cfg = virQEMUDriverGetConfig(driver);
qemuDomainLogContext *ctxt =
QEMU_DOMAIN_LOG_CONTEXT(g_object_new(QEMU_TYPE_DOMAIN_LOG_CONTEXT, NULL));
@@ -7128,7 +7129,7 @@ qemuDomainLogContext
*qemuDomainLogContextNew(virQEMUDriver *driver,
ctxt->writefd = -1;
ctxt->readfd = -1;
-ctxt->path = g_strdup_printf("%s/%s.log", cfg->logDir, vm->def->name);
+ctxt->path = g_strdup_printf("%s/%s.log", cfg->logDir, basename);
if (cfg->stdioLogD) {
ctxt->manager = virLogManagerNew(driver->privileged);
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h
index 89edc75fcf..a262555c8c 100644
--- a/src/qemu/qemu_domain.h
+++ b/src/qemu/qemu_domain.h
@@ -657,7 +657,8 @@ void qemuDomainObjCheckNetTaint(virQEMUDriver *driver,
qemuDomainLogContext *logCtxt);
qemuDomainLogContext *qemuDomainLogContextNew(virQEMUDriver *driver,
- virDomainObj *vm);
+ virDomainObj *vm,
+ const char *basename);
int qemuDomainLogContextWrite(qemuDomainLogContext *ctxt,
const char *fmt, ...) G_GNUC_PRINTF(2, 3);
ssize_t qemuDomainLogContextRead(qemuDomainLogContext *ctxt,
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index a6ed69cfe2..e0385d11be 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -7616,7 +7616,7 @@ qemuProcessLaunch(virConnectPtr conn,
hookData.cfg = cfg;
VIR_DEBUG("Creating domain log file");
-if (!(logCtxt = qemuDomainLogContextNew(driver, vm))) {
+if (!(logCtxt = qemuDomainLogContextNew(driver, vm, vm->def->name))) {
virLastErrorPrefixMessage("%s", _("can't connect to virtlogd"));
goto cleanup;
}
--
2.41.0
[libvirt PATCH v8 08/37] qemu: Add qemuNbdkitProcess
An object for storing information about a nbdkit process that is serving
a specific virStorageSource. At the moment, this information is just
stored in the private data of virStorageSource and not used at all.
Future commits will use this data to actually start a nbdkit process.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_conf.c | 22
src/qemu/qemu_conf.h | 2 ++
src/qemu/qemu_domain.c | 31
src/qemu/qemu_domain.h | 4 +++
src/qemu/qemu_nbdkit.c | 82 ++
src/qemu/qemu_nbdkit.h | 26 ++
6 files changed, 167 insertions(+)
diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c
index 3f811d064f..a0360e8d1b 100644
--- a/src/qemu/qemu_conf.c
+++ b/src/qemu/qemu_conf.c
@@ -1656,3 +1656,25 @@ qemuHugepageMakeBasedir(virQEMUDriver *driver,
return 0;
}
+
+
+/*
+ * qemuGetNbdkitCaps:
+ * @driver: the qemu driver
+ *
+ * Gets the capabilities for Nbdkit for the specified driver. These can be used
+ * to determine whether a particular disk source can be served by nbdkit or
+ * not.
+ *
+ * Returns: a reference to qemuNbdkitCaps or NULL
+ */
+qemuNbdkitCaps*
+qemuGetNbdkitCaps(virQEMUDriver *driver)
+{
+char *nbdkitBinary = virFindFileInPath("nbdkit");
+
+if (!nbdkitBinary)
+return NULL;
+
+return virFileCacheLookup(driver->nbdkitCapsCache, nbdkitBinary);
+}
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index a44985fb8b..1a3ba3a0fb 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -377,3 +377,5 @@ int qemuGetMemoryBackingPath(virQEMUDriver *driver,
int qemuHugepageMakeBasedir(virQEMUDriver *driver,
virHugeTLBFS *hugepage);
+
+qemuNbdkitCaps* qemuGetNbdkitCaps(virQEMUDriver *driver);
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index bfeddc7746..5a2eb4868a 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -882,6 +882,7 @@ qemuDomainStorageSourcePrivateDispose(void *obj)
g_clear_pointer(>httpcookie, qemuDomainSecretInfoFree);
g_clear_pointer(>tlsKeySecret, qemuDomainSecretInfoFree);
g_clear_pointer(>fdpass, qemuFDPassFree);
+g_clear_pointer(>nbdkitProcess, qemuNbdkitProcessFree);
}
@@ -10471,6 +10472,34 @@ qemuDomainPrepareStorageSourceNFS(virStorageSource
*src)
}
+/* qemuPrepareStorageSourceNbdkit:
+ * @src: source for a disk
+ *
+ * If src is an network source that is managed by nbdkit, prepare data so that
+ * nbdkit can be launched before the domain is started
+ *
+ * Returns true if nbdkit will be used for this source,
+ */
+static bool
+qemuDomainPrepareStorageSourceNbdkit(virStorageSource *src,
+ virQEMUDriverConfig *cfg,
+ const char *alias,
+ qemuDomainObjPrivate *priv)
+{
+g_autoptr(qemuNbdkitCaps) nbdkit = NULL;
+
+if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_NETWORK)
+return false;
+
+nbdkit = qemuGetNbdkitCaps(priv->driver);
+if (!nbdkit)
+return false;
+
+return qemuNbdkitInitStorageSource(nbdkit, src, priv->libDir,
+ alias, cfg->user, cfg->group);
+}
+
+
/* qemuProcessPrepareStorageSourceTLS:
* @source: source for a disk
* @cfg: driver configuration
@@ -11300,6 +11329,8 @@
qemuDomainPrepareStorageSourceBlockdevNodename(virDomainDiskDef *disk,
if (qemuDomainPrepareStorageSourceFDs(src, priv) < 0)
return -1;
+qemuDomainPrepareStorageSourceNbdkit(src, cfg, src->nodestorage, priv);
+
return 0;
}
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h
index 786f239495..89edc75fcf 100644
--- a/src/qemu/qemu_domain.h
+++ b/src/qemu/qemu_domain.h
@@ -33,6 +33,7 @@
#include "qemu_conf.h"
#include "qemu_capabilities.h"
#include "qemu_migration_params.h"
+#include "qemu_nbdkit.h"
#include "qemu_slirp.h"
#include "qemu_fd.h"
#include "virchrdev.h"
@@ -308,6 +309,9 @@ struct _qemuDomainStorageSourcePrivate {
/* file descriptors if user asks for FDs to be passed */
qemuFDPass *fdpass;
+
+/* an nbdkit process for serving network storage sources */
+qemuNbdkitProcess *nbdkitProcess;
};
virObject *qemuDomainStorageSourcePrivateNew(void);
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 97612b637f..12c721f7f1 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -549,3 +549,85 @@ qemuNbdkitCapsCacheNew(const char *cachedir)
g_autofree char *dir = g_build_filename(cachedir, "nbdkitcapabilities",
NULL);
return virFileCacheNew(dir, "xml", );
}
+
+
+static qemuNbdkitProcess *
+qemuNbdkitProcessNew(virStorageSource *source,
+ const char *pidfile,
+ const char *socketfile)
+{
+qemu
[libvirt PATCH v8 02/37] qemu: Add functions for determining nbdkit availability
In future commits, we will optionally use nbdkit to serve some remote
disk sources. This patch queries to see whether nbdkit is installed on
the host and queries it for capabilities. The data will be used in later
commits.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/meson.build | 1 +
src/qemu/qemu_conf.h | 1 +
src/qemu/qemu_nbdkit.c | 200 +
src/qemu/qemu_nbdkit.h | 50 +++
4 files changed, 252 insertions(+)
create mode 100644 src/qemu/qemu_nbdkit.c
create mode 100644 src/qemu/qemu_nbdkit.h
diff --git a/src/qemu/meson.build b/src/qemu/meson.build
index c8806bbc36..9be6996195 100644
--- a/src/qemu/meson.build
+++ b/src/qemu/meson.build
@@ -28,6 +28,7 @@ qemu_driver_sources = [
'qemu_monitor_json.c',
'qemu_monitor_text.c',
'qemu_namespace.c',
+ 'qemu_nbdkit.c',
'qemu_passt.c',
'qemu_process.c',
'qemu_qapi.c',
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index 11c740d28f..4f610d86a1 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -36,6 +36,7 @@
#include "virthreadpool.h"
#include "locking/lock_manager.h"
#include "qemu_capabilities.h"
+#include "qemu_nbdkit.h"
#include "virclosecallbacks.h"
#include "virhostdev.h"
#include "virfile.h"
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
new file mode 100644
index 00..e4e8fd568e
--- /dev/null
+++ b/src/qemu/qemu_nbdkit.c
@@ -0,0 +1,200 @@
+/*
+ * qemu_nbdkit.c: helpers for using nbdkit
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library. If not, see
+ * <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include
+#include
+
+#include "vircommand.h"
+#include "virerror.h"
+#include "virlog.h"
+#include "virpidfile.h"
+#include "virutil.h"
+#include "qemu_block.h"
+#include "qemu_conf.h"
+#include "qemu_domain.h"
+#include "qemu_extdevice.h"
+#include "qemu_nbdkit.h"
+#include "qemu_security.h"
+
+#include
+
+#define VIR_FROM_THIS VIR_FROM_QEMU
+
+VIR_LOG_INIT("qemu.nbdkit");
+
+VIR_ENUM_IMPL(qemuNbdkitCaps,
+QEMU_NBDKIT_CAPS_LAST,
+/* 0 */
+"plugin-curl", /* QEMU_NBDKIT_CAPS_PLUGIN_CURL */
+"plugin-ssh", /* QEMU_NBDKIT_CAPS_PLUGIN_SSH */
+"filter-readahead", /* QEMU_NBDKIT_CAPS_FILTER_READAHEAD */
+);
+
+struct _qemuNbdkitCaps {
+GObject parent;
+
+char *path;
+char *version;
+
+virBitmap *flags;
+};
+G_DEFINE_TYPE(qemuNbdkitCaps, qemu_nbdkit_caps, G_TYPE_OBJECT);
+
+
+static void
+qemuNbdkitCheckCommandCap(qemuNbdkitCaps *nbdkit,
+ virCommand *cmd,
+ qemuNbdkitCapsFlags cap)
+{
+if (virCommandRun(cmd, NULL) != 0)
+return;
+
+VIR_DEBUG("Setting nbdkit capability %i", cap);
+ignore_value(virBitmapSetBit(nbdkit->flags, cap));
+}
+
+
+static void
+qemuNbdkitQueryFilter(qemuNbdkitCaps *nbdkit,
+ const char *filter,
+ qemuNbdkitCapsFlags cap)
+{
+g_autoptr(virCommand) cmd = virCommandNewArgList(nbdkit->path,
+ "--version",
+ NULL);
+
+virCommandAddArgPair(cmd, "--filter", filter);
+
+/* use null plugin to check for filter */
+virCommandAddArg(cmd, "null");
+
+qemuNbdkitCheckCommandCap(nbdkit, cmd, cap);
+}
+
+
+static void
+qemuNbdkitQueryPlugin(qemuNbdkitCaps *nbdkit,
+ const char *plugin,
+ qemuNbdkitCapsFlags cap)
+{
+g_autoptr(virCommand) cmd = virCommandNewArgList(nbdkit->path,
+ plugin,
+ "--version",
+ NULL);
+
+qemuNbdkitCheckCommandCap(nbdkit, cmd, cap);
+}
+
+
+static void
+qemuNbdkitCapsQueryPlugins(qemuNbdkitCaps *nbdkit)
+{
+qemuNbdkitQueryPlugin(nbdkit, "curl", QEMU_NBDKIT_CAPS_PLUGIN_CURL);
+qemuNbdkitQueryPlugin(nbdkit, "ssh", QEMU_NBDKIT_CAPS_PLUGIN_SSH);
+}
+
+
+static void
+qemuNbdkitCapsQueryFilters(qemuNbdkitCaps *nbdkit)
+{
+qemuNbd
[libvirt PATCH v8 00/37] Use nbdkit for http/ftp/ssh network drives in libvirt
This is the eighth version of this patch series. See https://bugzilla.redhat.com/show_bug.cgi?id=2016527 for more information. Note that testing this requires selinux policy changes which are not fully done, but there is a new policy in development that has allowed me to run with selinux in enforcing mode for the common cases. See https://bugzilla.redhat.com/show_bug.cgi?id=2182505 for more information. The following scenarios should work now with selinux enabled using the selinux policy from that bug: - http/https disks - ssh disks with password authentication - ssh disks with passwordless keyfile The one major thing that doesn't work and is difficult to get working with selinux enabled is the ssh-agent. This is because there doesn't seem to be any selinux policy for ssh-agent, so by default the ssh-agent socket is labeled unconfined_t. We cannot allow access from the libvirt/qemu to unconfined_t because that would open up access to just about anything on the host. So additional work will likely be necessary for ssh-agent/libvirt interaction in the future. Fortunately ssh-agent is something that never was really supported with the old qemu block driver either, so I think we could potentially merge this patchset either without the ssh-agent patches or with a note that ssh-agent won't work with selinux enabled. Changes in v8: - Hopefully addressed all of Peter's issues, in addition to: - updated documentation to say 9.8.0, since 9.7.0 is currently in freeze - used WITH_NBDKIT instead of WITH_DECL_SYS_PIDFD_OPEN to make the code a bit more concise and understandable - enabled ci by adding libnbd to the dependencies, which uncovered a couple additional minor issues with those platforms that don't support the pidfd_open syscall - don't run nbdkit tests when WITH_NBDKIT is not defined - avoid warnings with unused function arguments - note that the ubuntu containers are currently failing due to a LeakSanitizer error, but I haven't reproduced it locally and can't figure out how to get better information from the leak sanitizer. Pointers appreciated: https://gitlab.com/jjongsma/libvirt/-/jobs/4991631193 - One change of note is a new patch "qemu: improve error handling when restarting nbdkit". In order to provide better error reporting to the user and avoid VIR_WARN as suggested by Peter, some functions now return an error and this error is propagated up to qemuProcessReconnect(). This could potentially result in running domains being killed upon a libvirt restart, but only if they were in a state where they were was not a running nbdkit backend or libvirt couldn't monitor the process nbdkit. Jonathon Jongsma (37): schema: allow 'ssh' as a protocol for network disks qemu: Add functions for determining nbdkit availability qemu: expand nbdkit capabilities util: Allow virFileCache data to be any GObject qemu: implement basic virFileCache for nbdkit caps qemu: implement persistent file cache for nbdkit caps qemu: use file cache for nbdkit caps qemu: Add qemuNbdkitProcess qemu: query nbdkit module dir from binary qemu: add functions to start and stop nbdkit Generalize qemuDomainLogContextNew() qemu: Extract qemuDomainLogContext into a new file qemu: move qemuProcessReadLog() to qemuLogContext qemu: log error output from nbdkit tests: add ability to test various nbdkit capabilities qemu: split qemuDomainSecretStorageSourcePrepare qemu: include nbdkit state in private xml util: secure erase virCommand send buffers qemu: pass sensitive data to nbdkit via pipe qemu: use nbdkit to serve network disks if available util: make virCommandSetSendBuffer testable tests: add tests for nbdkit invocation qemu: add test for authenticating a https network disk qemu: Add Taint for nbdkit restart failure qemu: Monitor nbdkit process for exit qemu: improve error handling when restarting nbdkit qemu: try to connect to nbdkit early to detect errors schema: add password configuration for ssh disk qemu: implement password auth for ssh disks with nbdkit schema: add configuration for host verification of ssh disks qemu: implement knownHosts for ssh disks with nbdkit schema: add keyfile configuration for ssh disks qemu: implement keyfile auth for ssh disks with nbdkit schema: add ssh-agent configuration for ssh disks qemu: implement ssh-agent auth for ssh disks with nbdkit rpm: update spec file for for nbdkit support ci: add libnbd to build build-aux/syntax-check.mk |2 +- ci/buildenv/almalinux-8.sh|1 + ci/buildenv/centos-stream-8.sh|1 + ci/buildenv/centos-stream-9.sh|1 + ci/buildenv/debian-12-cross-aarch64.sh|1 + ci/buildenv/debian-12-cross-armv6l.sh |1 + ci/buildenv/debian-12-cross-armv7l.sh |1 + ci/buildenv/debian-12-cross-i686.sh |1 + ci/buildenv/debia
[libvirt PATCH v8 01/37] schema: allow 'ssh' as a protocol for network disks
There was support in the code for parsing protocol='ssh' on network disk
sources, but it was not present in the xml schema. Add this to the
schema.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/schemas/domaincommon.rng | 1 +
tests/qemublocktest.c | 2 +-
...w2-invalid.json => network-ssh-qcow2.json} | 0
...cow2-invalid.xml => network-ssh-qcow2.xml} | 0
.../disk-network-ssh.x86_64-latest.args | 35 +++
tests/qemuxml2argvdata/disk-network-ssh.xml | 31
tests/qemuxml2argvtest.c | 1 +
7 files changed, 69 insertions(+), 1 deletion(-)
rename tests/qemublocktestdata/imagecreate/{network-ssh-qcow2-invalid.json =>
network-ssh-qcow2.json} (100%)
rename tests/qemublocktestdata/imagecreate/{network-ssh-qcow2-invalid.xml =>
network-ssh-qcow2.xml} (100%)
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh.xml
diff --git a/src/conf/schemas/domaincommon.rng
b/src/conf/schemas/domaincommon.rng
index de3bd1c35c..4a475f5c36 100644
--- a/src/conf/schemas/domaincommon.rng
+++ b/src/conf/schemas/domaincommon.rng
@@ -2179,6 +2179,7 @@
sheepdog
tftp
+ssh
diff --git a/tests/qemublocktest.c b/tests/qemublocktest.c
index 9a968477d7..8bad69e7ac 100644
--- a/tests/qemublocktest.c
+++ b/tests/qemublocktest.c
@@ -1213,7 +1213,7 @@ mymain(void)
TEST_IMAGE_CREATE("network-gluster-qcow2", NULL);
TEST_IMAGE_CREATE("network-rbd-qcow2", NULL);
-TEST_IMAGE_CREATE("network-ssh-qcow2-invalid", NULL);
+TEST_IMAGE_CREATE("network-ssh-qcow2", NULL);
#define TEST_BITMAP_DETECT(testname) \
do { \
diff --git a/tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.json
b/tests/qemublocktestdata/imagecreate/network-ssh-qcow2.json
similarity index 100%
rename from tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.json
rename to tests/qemublocktestdata/imagecreate/network-ssh-qcow2.json
diff --git a/tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.xml
b/tests/qemublocktestdata/imagecreate/network-ssh-qcow2.xml
similarity index 100%
rename from tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.xml
rename to tests/qemublocktestdata/imagecreate/network-ssh-qcow2.xml
diff --git a/tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
b/tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
new file mode 100644
index 00..b7fd30032b
--- /dev/null
+++ b/tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
@@ -0,0 +1,35 @@
+LC_ALL=C \
+PATH=/bin \
+HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1 \
+USER=test \
+LOGNAME=test \
+XDG_DATA_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.local/share \
+XDG_CACHE_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.cache \
+XDG_CONFIG_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.config \
+/usr/bin/qemu-system-x86_64 \
+-name guest=QEMUGuest1,debug-threads=on \
+-S \
+-object
'{"qom-type":"secret","id":"masterKey0","format":"raw","file":"/var/lib/libvirt/qemu/domain--1-QEMUGuest1/master-key.aes"}'
\
+-machine pc,usb=off,dump-guest-core=off,memory-backend=pc.ram,acpi=off \
+-accel kvm \
+-cpu qemu64 \
+-m size=219136k \
+-object '{"qom-type":"memory-backend-ram","id":"pc.ram","size":224395264}' \
+-overcommit mem-lock=off \
+-smp 1,sockets=1,cores=1,threads=1 \
+-uuid c7a5fdbd-edaf-9455-926a-d65c16db1809 \
+-display none \
+-no-user-config \
+-nodefaults \
+-chardev socket,id=charmonitor,fd=1729,server=on,wait=off \
+-mon chardev=charmonitor,id=monitor,mode=control \
+-rtc base=utc \
+-no-shutdown \
+-boot strict=on \
+-device
'{"driver":"piix3-usb-uhci","id":"usb","bus":"pci.0","addr":"0x1.0x2"}' \
+-blockdev
'{"driver":"ssh","path":"test.img","server":{"host":"example.org","port":""},"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}'
\
+-blockdev
'{"node-name":"libvirt-1-format","read-only":false,"driver":"raw","file":"libvirt-1-storage"}'
\
+-device
'{"driver":"virtio-blk-pci","bus":"pci.0","addr":"0x2","drive":"libvirt-1-format","id":"virtio-disk0","bootindex":1}'
\
+-audiodev '{"id":"audio1","driver":"none"}' \
+-sandbox
on,obsolete=deny,elevateprivileges=deny,sp
[libvirt PATCH v7 22/35] tests: add tests for nbdkit invocation
We were testing the arguments that were being passed to qemu when a disk was being served by nbdkit, but the arguments used to start nbdkit itself were not testable. This adds a test to ensure that we're invoking nbdkit correctly for various disk source definitions. Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- build-aux/syntax-check.mk | 2 +- src/qemu/qemu_nbdkit.c| 4 +- src/qemu/qemu_nbdkitpriv.h| 31 ++ tests/meson.build | 1 + .../disk-cdrom-network.args.disk0 | 6 + .../disk-cdrom-network.args.disk1 | 8 + .../disk-cdrom-network.args.disk1.pipe.778| 1 + .../disk-cdrom-network.args.disk2 | 8 + .../disk-cdrom-network.args.disk2.pipe.780| 1 + .../disk-network-http.args.disk0 | 6 + .../disk-network-http.args.disk1 | 5 + .../disk-network-http.args.disk2 | 6 + .../disk-network-http.args.disk2.pipe.778 | 1 + .../disk-network-http.args.disk3 | 7 + .../disk-network-http.args.disk3.pipe.780 | 1 + ...work-source-curl-nbdkit-backing.args.disk0 | 7 + ...ce-curl-nbdkit-backing.args.disk0.pipe.778 | 1 + .../disk-network-source-curl.args.disk0 | 7 + ...sk-network-source-curl.args.disk0.pipe.778 | 1 + .../disk-network-source-curl.args.disk1 | 7 + ...sk-network-source-curl.args.disk1.pipe.780 | 1 + .../disk-network-source-curl.args.disk2 | 7 + ...sk-network-source-curl.args.disk2.pipe.782 | 1 + .../disk-network-source-curl.args.disk3 | 6 + .../disk-network-source-curl.args.disk4 | 6 + .../disk-network-ssh.args.disk0 | 6 + tests/qemunbdkittest.c| 308 ++ 27 files changed, 444 insertions(+), 2 deletions(-) create mode 100644 src/qemu/qemu_nbdkitpriv.h create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk1 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk1.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk2 create mode 100644 tests/qemunbdkitdata/disk-cdrom-network.args.disk2.pipe.780 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk1 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk2 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk2.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk3 create mode 100644 tests/qemunbdkitdata/disk-network-http.args.disk3.pipe.780 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl-nbdkit-backing.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl-nbdkit-backing.args.disk0.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk0 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk0.pipe.778 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk1 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk2 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.782 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk3 create mode 100644 tests/qemunbdkitdata/disk-network-source-curl.args.disk4 create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk0 create mode 100644 tests/qemunbdkittest.c diff --git a/build-aux/syntax-check.mk b/build-aux/syntax-check.mk index 64c1e2773e..ec04402133 100644 --- a/build-aux/syntax-check.mk +++ b/build-aux/syntax-check.mk @@ -1370,7 +1370,7 @@ exclude_file_name_regexp--sc_prohibit_close = \ (\.p[yl]$$|\.spec\.in$$|^docs/|^(src/util/vir(file|event)\.c|src/libvirt-stream\.c|tests/(vir.+mock\.c|commandhelper\.c|qemusecuritymock\.c)|tools/nss/libvirt_nss_(leases|macs)\.c)|tools/virt-qemu-qmp-proxy$$) exclude_file_name_regexp--sc_prohibit_empty_lines_at_EOF = \ - (^tests/(nodedevmdevctl|viracpi|virhostcpu|virpcitest|virstoragetest)data/|docs/js/.*\.js|docs/fonts/.*\.woff|\.diff|tests/virconfdata/no-newline\.conf$$) + (^tests/(nodedevmdevctl|viracpi|virhostcpu|virpcitest|virstoragetest|qemunbdkit)data/|docs/js/.*\.js|docs/fonts/.*\.woff|\.diff|tests/virconfdata/no-newline\.conf$$) exclude_file_name_regexp--sc_prohibit_fork_wrappers = \ (^(src/(util/(vircommand|virdaemon)|lxc/lxc_controller)|tests/testutils)\.c$$) diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c index 299d8824f2..df638e99c0 100644 --- a/src/qemu/qemu_nbdkit.c +++ b/src/qemu/qemu_nbdkit.c @@ -31,6 +31,8 @@ #include "qemu_domain.h" #include "qemu_extdevice.h" #include "qemu_nbdkit.h" +#define LIBVIRT_QEM
[libvirt PATCH v7 04/35] util: Allow virFileCache data to be any GObject
Since the libvirt documentation suggests to prefer GObject over
virObject, and since virObject is a GObject, change virFileCache to
allow GObjects as data.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/util/virfilecache.c | 14 --
src/util/virfilecache.h | 2 +-
2 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/util/virfilecache.c b/src/util/virfilecache.c
index c730de066e..6f698016a1 100644
--- a/src/util/virfilecache.c
+++ b/src/util/virfilecache.c
@@ -170,7 +170,7 @@ virFileCacheLoad(virFileCache *cache,
*data = g_steal_pointer();
cleanup:
-virObjectUnref(loadData);
+g_clear_pointer(, g_object_unref);
return ret;
}
@@ -207,7 +207,7 @@ virFileCacheNewData(virFileCache *cache,
return NULL;
if (virFileCacheSave(cache, name, data) < 0) {
-g_clear_pointer(, virObjectUnref);
+g_clear_object();
}
}
@@ -239,7 +239,7 @@ virFileCacheNew(const char *dir,
if (!(cache = virObjectNew(virFileCacheClass)))
return NULL;
-cache->table = virHashNew(virObjectUnref);
+cache->table = virHashNew(g_object_unref);
cache->dir = g_strdup(dir);
@@ -270,7 +270,7 @@ virFileCacheValidate(virFileCache *cache,
if (*data) {
VIR_DEBUG("Caching data '%p' for '%s'", *data, name);
if (virHashAddEntry(cache->table, name, *data) < 0) {
-g_clear_pointer(data, virObjectUnref);
+g_clear_pointer(data, g_object_unref);
}
}
}
@@ -300,7 +300,8 @@ virFileCacheLookup(virFileCache *cache,
data = virHashLookup(cache->table, name);
virFileCacheValidate(cache, name, );
-virObjectRef(data);
+if (data)
+g_object_ref(data);
virObjectUnlock(cache);
return data;
@@ -331,7 +332,8 @@ virFileCacheLookupByFunc(virFileCache *cache,
data = virHashSearch(cache->table, iter, iterData, );
virFileCacheValidate(cache, name, );
-virObjectRef(data);
+if (data)
+g_object_ref(data);
virObjectUnlock(cache);
return data;
diff --git a/src/util/virfilecache.h b/src/util/virfilecache.h
index c3bc0f529c..944741c0a7 100644
--- a/src/util/virfilecache.h
+++ b/src/util/virfilecache.h
@@ -48,7 +48,7 @@ typedef bool
* @priv: private data created together with cache
*
* Creates a new data based on the @name. The returned data must be
- * an instance of virObject.
+ * an instance of GObject.
*
* Returns data object or NULL on error.
*/
--
2.41.0
[libvirt PATCH v7 27/35] schema: add password configuration for ssh disk
Right now, ssh network disks are not usable. There is some basic support in libvirt that is meant to support disk chains that have backing disks located at ssh urls, but there is no real way for a user to configure a ssh-based disk. This commit allows users to configure an ssh disk with password authentication. Implementation will follow. Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 27 ++- src/conf/schemas/domaincommon.rng | 23 ++- 2 files changed, 36 insertions(+), 14 deletions(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 68f54ab3ed..39d4230ec0 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -2778,7 +2778,7 @@ paravirtualized driver is specified via the ``disk`` element. ``network`` The ``protocol`` attribute specifies the protocol to access to the requested image. Possible values are "nbd", "iscsi", "rbd", "sheepdog", - "gluster", "vxhs", "nfs", "http", "https", "ftp", ftps", or "tftp". + "gluster", "vxhs", "nfs", "http", "https", "ftp", ftps", "tftp", or "ssh". For any ``protocol`` other than ``nbd`` an additional attribute ``name`` is mandatory to specify which volume/image will be used. @@ -2930,18 +2930,19 @@ paravirtualized driver is specified via the ``disk`` element. ``auth`` :since:`Since libvirt 3.9.0` , the ``auth`` element is supported for a disk ``type`` "network" that is using a ``source`` element with the - ``protocol`` attributes "rbd" or "iscsi". If present, the ``auth`` element - provides the authentication credentials needed to access the source. It - includes a mandatory attribute ``username``, which identifies the username - to use during authentication, as well as a sub-element ``secret`` with - mandatory attribute ``type``, to tie back to a `libvirt secret - object `__ that holds the actual password or other - credentials (the domain XML intentionally does not expose the password, - only the reference to the object that does manage the password). Known - secret types are "ceph" for Ceph RBD network sources and "iscsi" for CHAP - authentication of iSCSI targets. Both will require either a ``uuid`` - attribute with the UUID of the secret object or a ``usage`` attribute - matching the key that was specified in the secret object. + ``protocol`` attributes "rbd", "iscsi", or "ssh". If present, the + ``auth`` element provides the authentication credentials needed to access + the source. It includes a mandatory attribute ``username``, which + identifies the username to use during authentication, as well as a + sub-element ``secret`` with mandatory attribute ``type``, to tie back to + a `libvirt secret object `__ that holds the actual + password or other credentials (the domain XML intentionally does not + expose the password, only the reference to the object that does manage + the password). Known secret types are "ceph" for Ceph RBD network sources + and "iscsi" for CHAP authentication of iSCSI targets. Both will require + either a ``uuid`` attribute with the UUID of the secret object or a + ``usage`` attribute matching the key that was specified in the secret + object. ``encryption`` :since:`Since libvirt 3.9.0` , the ``encryption`` can be a sub-element of the ``source`` element for encrypted storage sources. If present, diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index 4a475f5c36..cd838a475c 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2172,6 +2172,27 @@ + + + + + +ssh + + + + + + + + + + + + + + + @@ -2179,7 +2200,6 @@ sheepdog tftp -ssh @@ -2289,6 +2309,7 @@ + -- 2.41.0
[libvirt PATCH v7 32/35] qemu: implement keyfile auth for ssh disks with nbdkit
For ssh disks that are served by nbdkit, we can support logging in with
an ssh key file. Pass the path to the configured key file and the
username to the nbdkit process.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/domain_conf.c| 30 +
src/conf/storage_source_conf.c| 2 ++
src/conf/storage_source_conf.h| 5 ++-
src/qemu/qemu_nbdkit.c| 15 +++--
.../disk-network-ssh-key.args.disk0 | 9 +
.../disk-network-ssh.args.disk2 | 9 +
tests/qemunbdkittest.c| 1 +
.../qemuxml2argvdata/disk-network-ssh-key.xml | 33 +++
8 files changed, 93 insertions(+), 11 deletions(-)
create mode 100644 tests/qemunbdkitdata/disk-network-ssh-key.args.disk0
create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk2
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh-key.xml
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 842b6404b5..929e115bce 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7268,10 +7268,18 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
return -1;
}
}
-if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
-(tmpnode = virXPathNode("./knownHosts", ctxt))) {
-if (!(src->ssh_known_hosts_file = virXMLPropStringRequired(tmpnode,
"path")))
-return -1;
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH) {
+if ((tmpnode = virXPathNode("./knownHosts", ctxt))) {
+if (!(src->ssh_known_hosts_file =
virXMLPropStringRequired(tmpnode, "path")))
+return -1;
+}
+if ((tmpnode = virXPathNode("./identity", ctxt))) {
+if (!(src->ssh_user = virXMLPropStringRequired(tmpnode,
"username")))
+return -1;
+
+if (!(src->ssh_keyfile = virXMLPropStringRequired(tmpnode,
"keyfile")))
+return -1;
+}
}
return 0;
@@ -22280,8 +22288,18 @@ virDomainDiskSourceFormatNetwork(virBuffer *attrBuf,
if (src->timeout)
virBufferAsprintf(childBuf, "\n",
src->timeout);
-if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
src->ssh_known_hosts_file)
-virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH) {
+if (src->ssh_known_hosts_file)
+virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
+if (src->ssh_keyfile) {
+virBufferAddLit(childBuf, "ssh_user);
+virBufferEscapeString(childBuf, " keyfile='%s'", src->ssh_keyfile);
+
+virBufferAddLit(childBuf, "/>\n");
+}
+}
}
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index 906bc36a9b..ce9c1f66c2 100644
--- a/src/conf/storage_source_conf.c
+++ b/src/conf/storage_source_conf.c
@@ -896,6 +896,7 @@ virStorageSourceCopy(const virStorageSource *src,
def->ssh_host_key_check_disabled = src->ssh_host_key_check_disabled;
def->ssh_user = g_strdup(src->ssh_user);
def->ssh_known_hosts_file = g_strdup(src->ssh_known_hosts_file);
+def->ssh_keyfile = g_strdup(src->ssh_keyfile);
def->nfs_user = g_strdup(src->nfs_user);
def->nfs_group = g_strdup(src->nfs_group);
@@ -1172,6 +1173,7 @@ virStorageSourceClear(virStorageSource *def)
VIR_FREE(def->ssh_user);
VIR_FREE(def->ssh_known_hosts_file);
+VIR_FREE(def->ssh_keyfile);
VIR_FREE(def->nfs_user);
VIR_FREE(def->nfs_group);
diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
index 8a9c7d07e2..8c805664af 100644
--- a/src/conf/storage_source_conf.h
+++ b/src/conf/storage_source_conf.h
@@ -406,12 +406,11 @@ struct _virStorageSource {
bool hostcdrom; /* backing device is a cdrom */
-/* passthrough variables for the ssh driver which we don't handle properly
*/
-/* these must not be used apart from formatting the output JSON in the
qemu driver */
+/* ssh variables */
char *ssh_user;
bool ssh_host_key_check_disabled;
-/* additional ssh variables */
char *ssh_known_hosts_file;
+char *ssh_keyfile;
/* nfs_user and nfs_group store the strings passed in by the user for NFS
params.
* nfs_uid and nfs_gid represent the converted/looked up ID numbers which
are used
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index d0eca1ab89..0393850ddc 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -1035,8 +1035,12 @@ qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
if (proc->source->auth) {
if (q
[libvirt PATCH v7 17/35] qemu: include nbdkit state in private xml
Add xml to the private data for a disk source to represent the nbdkit
process so that the state can be re-created if the libvirt daemon is
restarted. Format:
/path/to/nbdkit.pid
/path/to/nbdkit.socket
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c| 52 +
src/qemu/qemu_nbdkit.c| 71 +++
src/qemu/qemu_nbdkit.h| 8 +++
src/qemu/qemu_process.c | 6 ++
tests/qemustatusxml2xmldata/modern-in.xml | 4 ++
5 files changed, 141 insertions(+)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 951f3127d9..8429ce1028 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -1958,6 +1958,33 @@
qemuStorageSourcePrivateDataAssignSecinfo(qemuDomainSecretInfo **secinfo,
}
+static int
+qemuStorageSourcePrivateDataParseNbdkit(xmlNodePtr node,
+xmlXPathContextPtr ctxt,
+virStorageSource *src)
+{
+g_autofree char *pidfile = NULL;
+g_autofree char *socketfile = NULL;
+VIR_XPATH_NODE_AUTORESTORE(ctxt);
+
+ctxt->node = node;
+
+if (!(pidfile = virXPathString("string(./pidfile)", ctxt))) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing nbdkit
pidfile"));
+return -1;
+}
+
+if (!(socketfile = virXPathString("string(./socketfile)", ctxt))) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing nbdkit
socketfile"));
+return -1;
+}
+
+qemuNbdkitReconnectStorageSource(src, pidfile, socketfile);
+
+return 0;
+}
+
+
static int
qemuStorageSourcePrivateDataParse(xmlXPathContextPtr ctxt,
virStorageSource *src)
@@ -1971,6 +1998,7 @@ qemuStorageSourcePrivateDataParse(xmlXPathContextPtr ctxt,
bool fdsetPresent = false;
unsigned int fdSetID;
int enccount;
+xmlNodePtr nbdkitnode = NULL;
src->nodestorage =
virXPathString("string(./nodenames/nodename[@type='storage']/@name)", ctxt);
src->nodeformat =
virXPathString("string(./nodenames/nodename[@type='format']/@name)", ctxt);
@@ -2036,6 +2064,10 @@ qemuStorageSourcePrivateDataParse(xmlXPathContextPtr
ctxt,
virTristateBoolTypeFromString(thresholdEventWithIndex) ==
VIR_TRISTATE_BOOL_YES)
src->thresholdEventWithIndex = true;
+if ((nbdkitnode = virXPathNode("nbdkit", ctxt))) {
+if (qemuStorageSourcePrivateDataParseNbdkit(nbdkitnode, ctxt, src) < 0)
+return -1;
+}
return 0;
}
@@ -2053,6 +2085,23 @@ qemuStorageSourcePrivateDataFormatSecinfo(virBuffer *buf,
}
+static void
+qemuStorageSourcePrivateDataFormatNbdkit(qemuNbdkitProcess *nbdkit,
+ virBuffer *buf)
+{
+g_auto(virBuffer) childBuf = VIR_BUFFER_INIT_CHILD(buf);
+
+if (!nbdkit)
+return;
+
+virBufferEscapeString(, "%s\n",
+ nbdkit->pidfile);
+virBufferEscapeString(, "%s\n",
+ nbdkit->socketfile);
+virXMLFormatElement(buf, "nbdkit", NULL, );
+}
+
+
static int
qemuStorageSourcePrivateDataFormat(virStorageSource *src,
virBuffer *buf)
@@ -2102,6 +2151,9 @@ qemuStorageSourcePrivateDataFormat(virStorageSource *src,
if (src->thresholdEventWithIndex)
virBufferAddLit(buf, "\n");
+if (srcPriv)
+qemuStorageSourcePrivateDataFormatNbdkit(srcPriv->nbdkitProcess, buf);
+
return 0;
}
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 81861bae4a..e3923ab4f2 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -627,6 +627,77 @@ qemuNbdkitProcessNew(virStorageSource *source,
return nbdkit;
}
+/**
+ * qemuNbdkitReconnectStorageSource:
+ * @source: a storage source
+ * @pidfile: a pidfile for an nbdkit process
+ * @socketfile: the socket file associated with the nbdkit process
+ *
+ * This function constructs a new qemuNbdkitProcess object with the given
values for @pidfile and
+ * @socketfile and stores it in @source. This is intended to be called when
the libvirt daemon is
+ * restarted and tries to reconnect to all currently-running domains. Since
this function is called
+ * from the code that parses the current daemon state, it should not perform
any filesystem
+ * operations, or anything else that might fail. Additional initialization
will be done later by
+ * calling qemuNbdkitStorageSourceManageProcess().
+ */
+void
+qemuNbdkitReconnectStorageSource(virStorageSource *source,
+ const char *pidfile,
+ const char *socketfile)
+{
+qemuDomainStorageSourcePrivate *srcpriv =
qemuDomainStorageSourcePrivateFetch(s
[libvirt PATCH v7 05/35] qemu: implement basic virFileCache for nbdkit caps
Preparatory step for caching nbdkit capabilities. This patch implements
the newData and isValid virFileCacheHandlers callback functions.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 89 +-
src/qemu/qemu_nbdkit.h | 4 ++
2 files changed, 92 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 58828dd89a..9828f562f7 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -201,7 +201,7 @@ qemuNbdkitGetDirMtime(const char *moddir)
}
-G_GNUC_UNUSED static void
+static void
qemuNbdkitCapsQuery(qemuNbdkitCaps *caps)
{
struct stat st;
@@ -240,3 +240,90 @@ qemuNbdkitCapsSet(qemuNbdkitCaps *nbdkitCaps,
{
ignore_value(virBitmapSetBit(nbdkitCaps->flags, flag));
}
+
+
+static bool
+virNbkditCapsCheckModdir(const char *moddir,
+ time_t expectedMtime)
+{
+time_t mtime = qemuNbdkitGetDirMtime(moddir);
+
+if (mtime != expectedMtime) {
+VIR_DEBUG("Outdated capabilities for nbdkit: module directory '%s'
changed (%lld vs %lld)",
+ moddir, (long long)mtime, (long long)expectedMtime);
+return false;
+}
+return true;
+}
+
+
+static bool
+virNbdkitCapsIsValid(void *data,
+ void *privData G_GNUC_UNUSED)
+{
+qemuNbdkitCaps *nbdkitCaps = data;
+struct stat st;
+
+if (!nbdkitCaps->path)
+return true;
+
+if (!virNbkditCapsCheckModdir(NBDKIT_PLUGINDIR,
nbdkitCaps->pluginDirMtime))
+return false;
+if (!virNbkditCapsCheckModdir(NBDKIT_FILTERDIR,
nbdkitCaps->filterDirMtime))
+return false;
+
+if (nbdkitCaps->libvirtCtime != virGetSelfLastChanged() ||
+nbdkitCaps->libvirtVersion != LIBVIR_VERSION_NUMBER) {
+VIR_DEBUG("Outdated capabilities for '%s': libvirt changed (%lld vs
%lld, %lu vs %lu)",
+ nbdkitCaps->path,
+ (long long)nbdkitCaps->libvirtCtime,
+ (long long)virGetSelfLastChanged(),
+ (unsigned long)nbdkitCaps->libvirtVersion,
+ (unsigned long)LIBVIR_VERSION_NUMBER);
+return false;
+}
+
+if (stat(nbdkitCaps->path, ) < 0) {
+VIR_DEBUG("Failed to stat nbdkit binary '%s': %s",
+ nbdkitCaps->path,
+ g_strerror(errno));
+return false;
+}
+
+if (st.st_ctime != nbdkitCaps->ctime) {
+VIR_DEBUG("Outdated capabilities for '%s': nbdkit binary changed (%lld
vs %lld)",
+ nbdkitCaps->path,
+ (long long)st.st_ctime, (long long)nbdkitCaps->ctime);
+return false;
+}
+
+return true;
+}
+
+
+static void*
+virNbdkitCapsNewData(const char *binary,
+ void *privData G_GNUC_UNUSED)
+{
+qemuNbdkitCaps *caps = qemuNbdkitCapsNew(binary);
+qemuNbdkitCapsQuery(caps);
+
+return caps;
+}
+
+
+virFileCacheHandlers nbdkitCapsCacheHandlers = {
+.isValid = virNbdkitCapsIsValid,
+.newData = virNbdkitCapsNewData,
+.loadFile = NULL,
+.saveFile = NULL,
+.privFree = NULL,
+};
+
+
+virFileCache*
+qemuNbdkitCapsCacheNew(const char *cachedir)
+{
+g_autofree char *dir = g_build_filename(cachedir, "nbdkitcapabilities",
NULL);
+return virFileCacheNew(dir, "xml", );
+}
diff --git a/src/qemu/qemu_nbdkit.h b/src/qemu/qemu_nbdkit.h
index e191e1fdb4..4aba7c8455 100644
--- a/src/qemu/qemu_nbdkit.h
+++ b/src/qemu/qemu_nbdkit.h
@@ -21,6 +21,7 @@
#include "internal.h"
#include "virenum.h"
+#include "virfilecache.h"
typedef struct _qemuNbdkitCaps qemuNbdkitCaps;
@@ -38,6 +39,9 @@ VIR_ENUM_DECL(qemuNbdkitCaps);
qemuNbdkitCaps *
qemuNbdkitCapsNew(const char *path);
+virFileCache *
+qemuNbdkitCapsCacheNew(const char *cachedir);
+
bool
qemuNbdkitCapsGet(qemuNbdkitCaps *nbdkitCaps,
qemuNbdkitCapsFlags flag);
--
2.41.0
[libvirt PATCH v7 14/35] qemu: log error output from nbdkit
log stderr and stdout from nbdkit into its own log so that
nbdkit-related issues can be debugged more easily.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 16 +++-
1 file changed, 15 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 6bf962d0f1..2d70e72c42 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -852,12 +852,23 @@ qemuNbdkitProcessStart(qemuNbdkitProcess *proc,
virTimeBackOffVar timebackoff;
g_autoptr(virURI) uri = NULL;
g_autofree char *uristring = NULL;
+g_autofree char *basename = g_strdup_printf("%s-nbdkit-%i", vm->def->name,
proc->source->id);
+int logfd = -1;
+g_autoptr(qemuLogContext) logContext = NULL;
if (!(cmd = qemuNbdkitProcessBuildCommand(proc)))
return -1;
+if (!(logContext = qemuLogContextNew(driver, vm, basename))) {
+virLastErrorPrefixMessage("%s", _("can't connect to virtlogd"));
+return -1;
+}
+
+logfd = qemuLogContextGetWriteFD(logContext);
+
VIR_DEBUG("starting nbdkit process for %s", proc->source->nodestorage);
-virCommandSetErrorBuffer(cmd, );
+virCommandSetErrorFD(cmd, );
+virCommandSetOutputFD(cmd, );
virCommandSetPidFile(cmd, proc->pidfile);
if (qemuExtDeviceLogCommand(driver, vm, cmd, "nbdkit") < 0)
@@ -899,6 +910,9 @@ qemuNbdkitProcessStart(qemuNbdkitProcess *proc,
if ((uri = qemuBlockStorageSourceGetURI(proc->source)))
uristring = virURIFormat(uri);
+if (qemuLogContextReadFiltered(logContext, , 1024) < 0)
+VIR_WARN("Unable to read from nbdkit log");
+
virReportError(VIR_ERR_OPERATION_FAILED,
_("Failed to connect to nbdkit for '%1$s': %2$s"),
NULLSTR(uristring), NULLSTR(errbuf));
--
2.41.0
[libvirt PATCH v7 33/35] schema: add ssh-agent configuration for ssh disks
Add the ability to specify a path to a ssh-agent socket in order to use the ssh-agent to authenticate to remote ssh disks. Example configuration: ... ... Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 13 - src/conf/schemas/domaincommon.rng | 11 --- 2 files changed, 16 insertions(+), 8 deletions(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 2ebd7bb1fc..720433617d 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3007,11 +3007,14 @@ paravirtualized driver is specified via the ``disk`` element. are intended to be default, then the entire element may be omitted. When using an ``ssh`` protocol, this element is used to enable - authentication via ssh keys. In this configuration, the element has two - attributes. The ``username`` attribute specifies the name of the user on - the remote server and the ``keyfile`` attribute specifies the path to the - keyfile. Note that this only works for ssh keys that are not - password-protected. + authentication via ssh keys. In this configuration, the element has three + possible attributes. The ``username`` attribute is required and specifies + the name of the user on the remote server. ssh keys can be specified in + one of two ways. The first way is by adding them to an ssh-agent and + providing the path to the ssh-agent socket in the ``agentsock`` + attribute. This method works for ssh keys with or without password + protection. Alternatively, for ssh keys without a password, the ssh key + can be specified directly by setting the ``keyfile`` attribute. ``reconnect`` For disk type ``vhostuser`` configures reconnect timeout if the connection is lost. This is set with the two mandatory attributes ``enabled`` and diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index 47c5ee2a31..d8dd1b8c69 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2186,9 +2186,14 @@ - - - + + + + + + + + -- 2.41.0
[libvirt PATCH v7 23/35] qemu: add test for authenticating a https network disk
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
tests/qemunbdkitdata/disk-network-source-curl.args.disk1 | 4 +++-
.../disk-network-source-curl.args.disk1.pipe.780 | 2 +-
.../disk-network-source-curl.args.disk1.pipe.782 | 1 +
tests/qemunbdkitdata/disk-network-source-curl.args.disk2 | 2 +-
.../disk-network-source-curl.args.disk2.pipe.784 | 1 +
.../disk-network-source-curl.x86_64-latest.args | 3 ++-
tests/qemuxml2argvdata/disk-network-source-curl.xml | 3 +++
7 files changed, 12 insertions(+), 4 deletions(-)
create mode 100644
tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
create mode 100644
tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
index fde6a4f533..d1288dd242 100644
--- a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1
@@ -3,5 +3,7 @@ nbdkit \
--foreground curl \
protocols=https \
'url=https://https.example.org:8443/path/to/disk5.iso?foo=bar' \
-cookie=-779 \
+user=myname \
+password=-779 \
+cookie=-781 \
sslverify=false
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
index 20af4ae383..ccdd4033fc 100644
--- a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.780
@@ -1 +1 @@
-cookie1=cookievalue1; cookie2=cookievalue2
\ No newline at end of file
+iscsi-mycluster_myname-secret
\ No newline at end of file
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
new file mode 100644
index 00..20af4ae383
--- /dev/null
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk1.pipe.782
@@ -0,0 +1 @@
+cookie1=cookievalue1; cookie2=cookievalue2
\ No newline at end of file
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
index 88c9fa35a1..f1d0e1929e 100644
--- a/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2
@@ -4,4 +4,4 @@ nbdkit \
--readonly curl \
protocols=http,https \
url=http://http.example.org:8080/path/to/disk2.iso \
-cookie=-781
+cookie=-783
diff --git a/tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
new file mode 100644
index 00..5c035e84c5
--- /dev/null
+++ b/tests/qemunbdkitdata/disk-network-source-curl.args.disk2.pipe.784
@@ -0,0 +1 @@
+cookie1=cookievalue1; cookie2=cookievalue2; cookie3=cookievalue3
\ No newline at end of file
diff --git a/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
b/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
index cb0e5a92ea..f6ab5532cc 100644
--- a/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
+++ b/tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
@@ -32,9 +32,10 @@
XDG_CONFIG_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.config \
-blockdev
'{"driver":"https","url":"https://https.example.org:8443/path/to/disk1.iso","cookie-secret":"libvirt-5-storage-httpcookie-secret0","node-name":"libvirt-5-storage","auto-read-only":true,"discard":"unmap"}'
\
-blockdev
'{"node-name":"libvirt-5-format","read-only":true,"driver":"raw","file":"libvirt-5-storage"}'
\
-device
'{"driver":"virtio-blk-pci","bus":"pci.0","addr":"0x3","drive":"libvirt-5-format","id":"virtio-disk0","bootindex":1}'
\
+-object
'{"qom-type":"secret","id":"libvirt-4-storage-auth-secret0","data":"9eao5F8qtkGt+seB1HYivWIxbtwUu6MQtg1zpj/oDtUsPr1q8wBYM91uEHCn6j/1","keyid":"masterKey0","iv":"AAECAwQFBgcICQoLDA0ODw==","format":"base64"}'
\
-object
'{"qom-type":"secret","id":"libvirt-4-storage-httpcookie-secret0","data":"BUU0KmnWfonHdjzhYhwVQZ5iTI1KweTJ22q8XWUVoBCVu1z70reDuczPBIabZtC3","keyid":"masterKey0","iv":"AAECAwQFBgcICQoLDA0ODw==","format":"base64"}'
\
-object
'{"qom-type":"secret","id":"libvirt-4-format-encryption-secret0","data":"9eao5F8qtkGt+seB1HYivWIxbtwUu6MQtg1zpj/oD
[libvirt PATCH v7 18/35] util: secure erase virCommand send buffers
All users of virCommandSetSendBuffer() are using it to send sensitive
data to a child process. So, since these buffers contain sensitive
information, clear it with virSecureErase().
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/util/vircommand.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/util/vircommand.c b/src/util/vircommand.c
index 5f094c625a..899d413dd2 100644
--- a/src/util/vircommand.c
+++ b/src/util/vircommand.c
@@ -54,6 +54,7 @@
#include "virpidfile.h"
#include "virprocess.h"
#include "virbuffer.h"
+#include "virsecureerase.h"
#include "virthread.h"
#include "virstring.h"
@@ -1697,6 +1698,7 @@ virCommandFreeSendBuffers(virCommand *cmd)
for (i = 0; i < virCommandGetNumSendBuffers(cmd); i++) {
VIR_FORCE_CLOSE(cmd->sendBuffers[i].fd);
+virSecureErase(cmd->sendBuffers[i].buffer, cmd->sendBuffers[i].buflen);
VIR_FREE(cmd->sendBuffers[i].buffer);
}
VIR_FREE(cmd->sendBuffers);
--
2.41.0
[libvirt PATCH v7 31/35] schema: add keyfile configuration for ssh disks
Authenticating via key file to an ssh server is often preferable to logging in via password. In order to support this functionality add a new xml element for ssh disks that allows the user to specify a keyfile and username. Example configuration: ... ... Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 7 +++ src/conf/schemas/domaincommon.rng | 19 ++- 2 files changed, 25 insertions(+), 1 deletion(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 9230f540f4..2ebd7bb1fc 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3005,6 +3005,13 @@ paravirtualized driver is specified via the ``disk`` element. of these attributes is omitted, then that field is assumed to be the default value for the current system. If both ``user`` and ``group`` are intended to be default, then the entire element may be omitted. + + When using an ``ssh`` protocol, this element is used to enable + authentication via ssh keys. In this configuration, the element has two + attributes. The ``username`` attribute specifies the name of the user on + the remote server and the ``keyfile`` attribute specifies the path to the + keyfile. Note that this only works for ssh keys that are not + password-protected. ``reconnect`` For disk type ``vhostuser`` configures reconnect timeout if the connection is lost. This is set with the two mandatory attributes ``enabled`` and diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index ca43586323..47c5ee2a31 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2180,6 +2180,19 @@ + + + + + + + + + + + + + @@ -2199,11 +2212,15 @@ - + + + + + -- 2.41.0
[libvirt PATCH v7 19/35] qemu: pass sensitive data to nbdkit via pipe
Rather than passing passwords and cookies (which could contain
passwords) to nbdkit via commandline arguments, use the alternate format
that nbdkit supports where we can specify a file descriptor which nbdkit
will read to get the password or cookies.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 54 ++
1 file changed, 34 insertions(+), 20 deletions(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index e3923ab4f2..22a67b0748 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -24,7 +24,6 @@
#include "virerror.h"
#include "virlog.h"
#include "virpidfile.h"
-#include "virsecureerase.h"
#include "virtime.h"
#include "virutil.h"
#include "qemu_block.h"
@@ -753,6 +752,29 @@ qemuNbdkitInitStorageSource(qemuNbdkitCaps *caps,
}
+static int
+qemuNbdkitCommandPassDataByPipe(virCommand *cmd,
+const char *argName,
+unsigned char **buf,
+size_t buflen)
+{
+g_autofree char *fdfmt = NULL;
+int fd = virCommandSetSendBuffer(cmd, buf, buflen);
+
+if (fd < 0)
+return -1;
+
+/* some nbdkit arguments accept a variation where nbdkit will read the data
+ * from a file descriptor, e.g. password=-FD */
+fdfmt = g_strdup_printf("-%i", fd);
+virCommandAddArgPair(cmd, argName, fdfmt);
+
+virCommandDoAsyncIO(cmd);
+
+return 0;
+}
+
+
static int
qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
virCommand *cmd)
@@ -775,7 +797,6 @@ qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
g_autoptr(virConnect) conn = virGetConnectSecret();
g_autofree uint8_t *secret = NULL;
size_t secretlen = 0;
-g_autofree char *password = NULL;
int secrettype;
virStorageAuthDef *authdef = proc->source->auth;
@@ -799,26 +820,19 @@ qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
return -1;
}
-/* ensure that the secret is a NULL-terminated string */
-password = g_strndup((char*)secret, secretlen);
-virSecureErase(secret, secretlen);
-
-/* for now, just report an error rather than passing the password in
- * cleartext on the commandline */
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("Password not yet supported for nbdkit sources"));
-
-virSecureEraseString(password);
-
-return -1;
+if (qemuNbdkitCommandPassDataByPipe(cmd, "password",
+, secretlen) < 0)
+return -1;
}
-if (proc->source->ncookies > 0) {
-/* for now, just report an error rather than passing cookies in
- * cleartext on the commandline */
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("Cookies not yet supported for nbdkit sources"));
-return -1;
+/* Create a pipe to send the cookies to the nbdkit process. */
+if (proc->source->ncookies) {
+g_autofree char *cookies =
qemuBlockStorageSourceGetCookieString(proc->source);
+
+if (qemuNbdkitCommandPassDataByPipe(cmd, "cookie",
+(unsigned char**),
+strlen(cookies)) < 0)
+return -1;
}
if (proc->source->sslverify == VIR_TRISTATE_BOOL_NO) {
--
2.41.0
[libvirt PATCH v7 06/35] qemu: implement persistent file cache for nbdkit caps
Implement the loadFile and saveFile virFileCacheHandlers callbacks so
that nbdkit capabilities are cached perstistently across daemon
restarts. The format and implementation is modeled on the qemu
capabilities, but simplified slightly.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
po/POTFILES| 1 +
src/qemu/qemu_nbdkit.c | 226 -
2 files changed, 225 insertions(+), 2 deletions(-)
diff --git a/po/POTFILES b/po/POTFILES
index 5d6ec195b4..6167f98ac5 100644
--- a/po/POTFILES
+++ b/po/POTFILES
@@ -181,6 +181,7 @@ src/qemu/qemu_monitor.c
src/qemu/qemu_monitor_json.c
src/qemu/qemu_monitor_text.c
src/qemu/qemu_namespace.c
+src/qemu/qemu_nbdkit.c
src/qemu/qemu_passt.c
src/qemu/qemu_process.c
src/qemu/qemu_qapi.c
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 9828f562f7..97612b637f 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -312,11 +312,233 @@ virNbdkitCapsNewData(const char *binary,
}
+static int
+qemuNbdkitCapsValidateBinary(qemuNbdkitCaps *nbdkitCaps,
+ xmlXPathContextPtr ctxt)
+{
+g_autofree char *str = NULL;
+
+if (!(str = virXPathString("string(./path)", ctxt))) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("missing path in nbdkit capabilities cache"));
+return -1;
+}
+
+if (STRNEQ(str, nbdkitCaps->path)) {
+virReportError(VIR_ERR_INTERNAL_ERROR,
+ _("Expected caps for '%1$s' but saw '%2$s'"),
+ nbdkitCaps->path, str);
+return -1;
+}
+
+return 0;
+}
+
+
+static int
+qemuNbdkitCapsParseFlags(qemuNbdkitCaps *nbdkitCaps,
+ xmlXPathContextPtr ctxt)
+{
+g_autofree xmlNodePtr *nodes = NULL;
+size_t i;
+int n;
+
+if ((n = virXPathNodeSet("./flag", ctxt, )) < 0) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("failed to parse qemu capabilities flags"));
+return -1;
+}
+
+VIR_DEBUG("Got flags %d", n);
+for (i = 0; i < n; i++) {
+unsigned int flag;
+
+if (virXMLPropEnum(nodes[i], "name", qemuNbdkitCapsTypeFromString,
+ VIR_XML_PROP_REQUIRED, ) < 0)
+return -1;
+
+qemuNbdkitCapsSet(nbdkitCaps, flag);
+}
+
+return 0;
+}
+
+
+/*
+ * Parsing a doc that looks like
+ *
+ *
+ * /some/path
+ * 234235253
+ * 234235253
+ * 234235253
+ * 234235253
+ * 1002016
+ *
+ *
+ * ...
+ *
+ *
+ * Returns 0 on success, 1 if outdated, -1 on error
+ */
+static int
+qemuNbdkitCapsLoadCache(qemuNbdkitCaps *nbdkitCaps,
+const char *filename)
+{
+g_autoptr(xmlDoc) doc = NULL;
+g_autoptr(xmlXPathContext) ctxt = NULL;
+long long int l;
+
+if (!(doc = virXMLParse(filename, NULL, NULL, "nbdkitCaps", , NULL,
false)))
+return -1;
+
+if (virXPathLongLong("string(./selfctime)", ctxt, ) < 0) {
+VIR_DEBUG("missing selfctime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->libvirtCtime = (time_t)l;
+
+nbdkitCaps->libvirtVersion = 0;
+virXPathUInt("string(./selfvers)", ctxt, >libvirtVersion);
+
+if (nbdkitCaps->libvirtCtime != virGetSelfLastChanged() ||
+nbdkitCaps->libvirtVersion != LIBVIR_VERSION_NUMBER) {
+VIR_DEBUG("Outdated capabilities in %s: libvirt changed (%lld vs %lld,
%lu vs %lu), stopping load",
+ nbdkitCaps->path,
+ (long long)nbdkitCaps->libvirtCtime,
+ (long long)virGetSelfLastChanged(),
+ (unsigned long)nbdkitCaps->libvirtVersion,
+ (unsigned long)LIBVIR_VERSION_NUMBER);
+return 1;
+}
+
+if (qemuNbdkitCapsValidateBinary(nbdkitCaps, ctxt) < 0)
+return -1;
+
+if (virXPathLongLong("string(./nbdkitctime)", ctxt, ) < 0) {
+VIR_DEBUG("missing nbdkitctime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->ctime = (time_t)l;
+
+if (virXPathLongLong("string(./plugindirmtime)", ctxt, ) < 0) {
+VIR_DEBUG("missing plugindirmtime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->pluginDirMtime = (time_t)l;
+
+if (virXPathLongLong("string(./filterdirmtime)", ctxt, ) < 0) {
+VIR_DEBUG("missing filterdirmtime in nbdkit capabilities XML");
+return -1;
+}
+nbdkitCaps->filterDirMtime = (time_t)l;
+
+if (qemuNbdkitCapsParseFlags(nbdkitCaps, ctxt) < 0)
+return -1;
+
+if ((nbdkitCaps->version = virXPathString("string(./version)", ctxt)) ==
NULL) {
+VIR_DEBUG("missing
[libvirt PATCH v7 21/35] util: make virCommandSetSendBuffer testable
Add a private function to peek at the list of send buffers in virCommand
so that it is testable
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/libvirt_private.syms | 1 +
src/util/vircommand.c | 17 +
src/util/vircommand.h | 8
src/util/vircommandpriv.h | 4
4 files changed, 22 insertions(+), 8 deletions(-)
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
index 1e3e407097..e4da5388e7 100644
--- a/src/libvirt_private.syms
+++ b/src/libvirt_private.syms
@@ -2117,6 +2117,7 @@ virCommandNewArgs;
virCommandNewVAList;
virCommandNonblockingFDs;
virCommandPassFD;
+virCommandPeekSendBuffers;
virCommandRawStatus;
virCommandRequireHandshake;
virCommandRun;
diff --git a/src/util/vircommand.c b/src/util/vircommand.c
index 899d413dd2..7d7ce4297e 100644
--- a/src/util/vircommand.c
+++ b/src/util/vircommand.c
@@ -78,14 +78,6 @@ struct _virCommandFD {
unsigned int flags;
};
-typedef struct _virCommandSendBuffer virCommandSendBuffer;
-struct _virCommandSendBuffer {
-int fd;
-unsigned char *buffer;
-size_t buflen;
-off_t offset;
-};
-
struct _virCommand {
int has_error; /* 0 on success, -1 on error */
@@ -3515,3 +3507,12 @@ virCommandSetRunAmong(virCommand *cmd,
cmd->schedCore = pid;
}
+
+void
+virCommandPeekSendBuffers(virCommand *cmd,
+ virCommandSendBuffer **buffers,
+ int *nbuffers)
+{
+*buffers = cmd->sendBuffers;
+*nbuffers = cmd->numSendBuffers;
+}
diff --git a/src/util/vircommand.h b/src/util/vircommand.h
index d51449ac90..9bcdce35b9 100644
--- a/src/util/vircommand.h
+++ b/src/util/vircommand.h
@@ -24,6 +24,14 @@
#include "internal.h"
#include "virbuffer.h"
+typedef struct _virCommandSendBuffer virCommandSendBuffer;
+struct _virCommandSendBuffer {
+int fd;
+unsigned char *buffer;
+size_t buflen;
+off_t offset;
+};
+
typedef struct _virCommand virCommand;
/* This will execute in the context of the first child
diff --git a/src/util/vircommandpriv.h b/src/util/vircommandpriv.h
index ff17fa5ded..d579810bb5 100644
--- a/src/util/vircommandpriv.h
+++ b/src/util/vircommandpriv.h
@@ -47,3 +47,7 @@ void virCommandSetDryRun(virCommandDryRunToken *tok,
bool bufCommandStripPath,
virCommandDryRunCallback cb,
void *opaque);
+
+void virCommandPeekSendBuffers(virCommand *cmd,
+ virCommandSendBuffer **buffers,
+ int *nbuffers);
--
2.41.0
[libvirt PATCH v7 24/35] qemu: Monitor nbdkit process for exit
Adds the ability to monitor the nbdkit process so that we can take
action in case the child exits unexpectedly.
When the nbdkit process exits, we pause the vm, restart nbdkit, and then
resume the vm. This allows the vm to continue working in the event of a
nbdkit failure.
Eventually we may want to generalize this functionality since we may
need something similar for e.g. qemu-storage-daemon, etc.
The process is monitored with the pidfd_open() syscall if it exists
(since linux 5.3). Otherwise it resorts to checking whether the process
is alive once a second. The one-second time period was chosen somewhat
arbitrarily.
Signed-off-by: Jonathon Jongsma
---
meson.build | 7 ++
src/qemu/qemu_domain.c | 1 +
src/qemu/qemu_domain.h | 1 +
src/qemu/qemu_driver.c | 18 ++
src/qemu/qemu_nbdkit.c | 137 ++--
src/qemu/qemu_nbdkit.h | 8 ++-
src/qemu/qemu_process.c | 13 +++-
src/qemu/qemu_process.h | 3 +
8 files changed, 180 insertions(+), 8 deletions(-)
diff --git a/meson.build b/meson.build
index 965ada483b..8da3b59070 100644
--- a/meson.build
+++ b/meson.build
@@ -682,6 +682,13 @@ symbols = [
[ 'sched.h', 'cpu_set_t' ],
]
+if host_machine.system() == 'linux'
+ symbols += [
+# process management
+[ 'sys/syscall.h', 'SYS_pidfd_open' ],
+ ]
+endif
+
foreach symbol : symbols
if cc.has_header_symbol(symbol[0], symbol[1], args: '-D_GNU_SOURCE', prefix:
symbol.get(2, ''))
conf.set('WITH_DECL_@0@'.format(symbol[1].to_upper()), 1)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 46fe5a1cf4..2b8ece8f19 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -11511,6 +11511,7 @@ qemuProcessEventFree(struct qemuProcessEvent *event)
case QEMU_PROCESS_EVENT_PR_DISCONNECT:
case QEMU_PROCESS_EVENT_UNATTENDED_MIGRATION:
case QEMU_PROCESS_EVENT_RESET:
+case QEMU_PROCESS_EVENT_NBDKIT_EXITED:
case QEMU_PROCESS_EVENT_LAST:
break;
}
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h
index ddd20e67b4..f018b45eb6 100644
--- a/src/qemu/qemu_domain.h
+++ b/src/qemu/qemu_domain.h
@@ -465,6 +465,7 @@ typedef enum {
QEMU_PROCESS_EVENT_MEMORY_DEVICE_SIZE_CHANGE,
QEMU_PROCESS_EVENT_UNATTENDED_MIGRATION,
QEMU_PROCESS_EVENT_RESET,
+QEMU_PROCESS_EVENT_NBDKIT_EXITED,
QEMU_PROCESS_EVENT_LAST
} qemuProcessEventType;
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index ad8428948b..46e089fe0f 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -4040,6 +4040,21 @@ processResetEvent(virQEMUDriver *driver,
}
+
+
+static void
+processNbdkitExitedEvent(virDomainObj *vm,
+ qemuNbdkitProcess *nbdkit)
+{
+if (virDomainObjBeginJob(vm, VIR_JOB_MODIFY) < 0)
+return;
+
+qemuNbdkitProcessRestart(nbdkit, vm);
+
+virDomainObjEndJob(vm);
+}
+
+
static void qemuProcessEventHandler(void *data, void *opaque)
{
struct qemuProcessEvent *processEvent = data;
@@ -4097,6 +4112,9 @@ static void qemuProcessEventHandler(void *data, void
*opaque)
case QEMU_PROCESS_EVENT_RESET:
processResetEvent(driver, vm);
break;
+case QEMU_PROCESS_EVENT_NBDKIT_EXITED:
+processNbdkitExitedEvent(vm, processEvent->data);
+break;
case QEMU_PROCESS_EVENT_LAST:
break;
}
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index df638e99c0..c3fa349922 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -19,6 +19,7 @@
#include
#include
+#include
#include "vircommand.h"
#include "virerror.h"
@@ -33,6 +34,7 @@
#include "qemu_nbdkit.h"
#define LIBVIRT_QEMU_NBDKITPRIV_H_ALLOW
#include "qemu_nbdkitpriv.h"
+#include "qemu_process.h"
#include "qemu_security.h"
#include
@@ -611,6 +613,106 @@ qemuNbdkitCapsCacheNew(const char *cachedir)
}
+void
+qemuNbdkitProcessRestart(qemuNbdkitProcess *proc,
+ virDomainObj *vm)
+{
+qemuDomainObjPrivate *vmpriv = vm->privateData;
+virQEMUDriver *driver = vmpriv->driver;
+
+/* clean up resources associated with process */
+qemuNbdkitProcessStop(proc);
+
+if (qemuNbdkitProcessStart(proc, vm, driver) < 0)
+VIR_WARN("Unable to restart nbkdit process");
+}
+
+
+typedef struct {
+qemuNbdkitProcess *proc;
+virDomainObj *vm;
+} qemuNbdkitProcessEventData;
+
+
+static qemuNbdkitProcessEventData*
+qemuNbdkitProcessEventDataNew(qemuNbdkitProcess *proc,
+ virDomainObj *vm)
+{
+qemuNbdkitProcessEventData *d = g_new(qemuNbdkitProcessEventData, 1);
+d->proc = proc;
+d->vm = virObjectRef(vm);
+return d;
+}
+
+
+static void
+qemuNbdkitProcessEventDataFree(qemuNbdkitProcessEventData *d)
+{
+virObjectUnref(d->vm);
+g_free(d);
+}
+
+
+#if WITH_DECL_SYS_PIDFD_OPEN
+static void
+qemuNbdkitProcessPi
[libvirt PATCH v7 35/35] rpm: update spec file for for nbdkit support
Require libnbd-devel when building the qemu driver, recommend nbdkit
packages.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
libvirt.spec.in | 8
1 file changed, 8 insertions(+)
diff --git a/libvirt.spec.in b/libvirt.spec.in
index 7157cfe3b4..94cae34496 100644
--- a/libvirt.spec.in
+++ b/libvirt.spec.in
@@ -312,6 +312,7 @@ BuildRequires: util-linux
BuildRequires: libacl-devel
# From QEMU RPMs, used by virstoragetest
BuildRequires: /usr/bin/qemu-img
+BuildRequires: libnbd-devel
%endif
# For LVM drivers
BuildRequires: lvm2
@@ -768,6 +769,9 @@ Requires: numad
Recommends: passt
Recommends: passt-selinux
%endif
+Recommends: nbdkit
+Recommends: nbdkit-curl-plugin
+Recommends: nbdkit-ssh-plugin
%description daemon-driver-qemu
The qemu driver plugin for the libvirtd daemon, providing
@@ -1069,8 +1073,10 @@ exit 1
%if %{with_qemu}
%define arg_qemu -Ddriver_qemu=enabled
+%define arg_libnbd -Dlibndb=enabled
%else
%define arg_qemu -Ddriver_qemu=disabled
+%define arg_libnbd -Dlibndb=disabled
%endif
%if %{with_openvz}
@@ -1259,6 +1265,7 @@ export SOURCE_DATE_EPOCH=$(stat --printf='%Y'
%{_specdir}/libvirt.spec)
-Dyajl=enabled \
%{?arg_sanlock} \
-Dlibpcap=enabled \
+ %{?arg_libnbd} \
-Dlibnl=enabled \
-Daudit=enabled \
-Ddtrace=enabled \
@@ -1322,6 +1329,7 @@ export SOURCE_DATE_EPOCH=$(stat --printf='%Y'
%{_specdir}/libvirt.spec)
-Dglusterfs=disabled \
-Dhost_validate=disabled \
-Dlibiscsi=disabled \
+ -Dlibnbd=disabled \
-Dlibnl=disabled \
-Dlibpcap=disabled \
-Dlibssh2=disabled \
--
2.41.0
[libvirt PATCH v7 20/35] qemu: use nbdkit to serve network disks if available
For virStorageSource objects that contain an nbdkitProcess, start that
nbdkit process to serve that network drive and then pass the nbdkit
socket to qemu rather than sending the network url to qemu directly.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_block.c | 162 +++---
src/qemu/qemu_domain.c| 13 +-
src/qemu/qemu_extdevice.c | 62 +++
src/qemu/qemu_hotplug.c | 7 +
src/qemu/qemu_nbdkit.c| 42 +
src/qemu/qemu_nbdkit.h| 13 ++
...sk-cdrom-network-nbdkit.x86_64-latest.args | 42 +
.../disk-cdrom-network-nbdkit.xml | 1 +
...isk-network-http-nbdkit.x86_64-latest.args | 44 +
.../disk-network-http-nbdkit.xml | 1 +
...rce-curl-nbdkit-backing.x86_64-latest.args | 37
...isk-network-source-curl-nbdkit-backing.xml | 45 +
...work-source-curl-nbdkit.x86_64-latest.args | 49 ++
.../disk-network-source-curl-nbdkit.xml | 1 +
...isk-network-source-curl.x86_64-latest.args | 52 ++
.../disk-network-source-curl.xml | 71
...disk-network-ssh-nbdkit.x86_64-latest.args | 35
.../disk-network-ssh-nbdkit.xml | 1 +
tests/qemuxml2argvtest.c | 6 +
19 files changed, 618 insertions(+), 66 deletions(-)
create mode 100644
tests/qemuxml2argvdata/disk-cdrom-network-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-cdrom-network-nbdkit.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-http-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-network-http-nbdkit.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl-nbdkit-backing.x86_64-latest.args
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl-nbdkit-backing.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-network-source-curl-nbdkit.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-source-curl.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-network-source-curl.xml
create mode 100644
tests/qemuxml2argvdata/disk-network-ssh-nbdkit.x86_64-latest.args
create mode 12 tests/qemuxml2argvdata/disk-network-ssh-nbdkit.xml
diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
index dcdf883926..1a2dc8ffb4 100644
--- a/src/qemu/qemu_block.c
+++ b/src/qemu/qemu_block.c
@@ -438,6 +438,32 @@ qemuBlockStorageSourceGetCURLProps(virStorageSource *src,
}
+static virJSONValue *
+qemuBlockStorageSourceGetNbdkitProps(virStorageSource *src)
+{
+qemuDomainStorageSourcePrivate *srcPriv =
QEMU_DOMAIN_STORAGE_SOURCE_PRIVATE(src);
+virJSONValue *ret = NULL;
+g_autoptr(virJSONValue) serverprops = NULL;
+virStorageNetHostDef host = { .transport = VIR_STORAGE_NET_HOST_TRANS_UNIX
};
+
+/* srcPriv->nbdkitProcess will already be initialized if we can use nbdkit
+ * to proxy this storage source */
+if (!(srcPriv && srcPriv->nbdkitProcess))
+return NULL;
+
+host.socket = srcPriv->nbdkitProcess->socketfile;
+serverprops = qemuBlockStorageSourceBuildJSONSocketAddress();
+
+if (!serverprops)
+return NULL;
+
+if (virJSONValueObjectAdd(, "a:server", , NULL) < 0)
+return NULL;
+
+return ret;
+}
+
+
static virJSONValue *
qemuBlockStorageSourceGetISCSIProps(virStorageSource *src,
bool onlytarget)
@@ -890,69 +916,75 @@ qemuBlockStorageSourceGetBackendProps(virStorageSource
*src,
return NULL;
case VIR_STORAGE_TYPE_NETWORK:
-switch ((virStorageNetProtocol) src->protocol) {
-case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-driver = "gluster";
-if (!(fileprops = qemuBlockStorageSourceGetGlusterProps(src,
onlytarget)))
-return NULL;
-break;
-
-case VIR_STORAGE_NET_PROTOCOL_VXHS:
-driver = "vxhs";
-if (!(fileprops = qemuBlockStorageSourceGetVxHSProps(src,
onlytarget)))
-return NULL;
-break;
-
-case VIR_STORAGE_NET_PROTOCOL_HTTP:
-case VIR_STORAGE_NET_PROTOCOL_HTTPS:
-case VIR_STORAGE_NET_PROTOCOL_FTP:
-case VIR_STORAGE_NET_PROTOCOL_FTPS:
-case VIR_STORAGE_NET_PROTOCOL_TFTP:
-driver = virStorageNetProtocolTypeToString(src->protocol);
-if (!(fileprops = qemuBlockStorageSourceGetCURLProps(src,
onlytarget)))
-return NULL;
-break;
-
-case VIR_STORAGE_NET_PROTOCOL_ISCSI:
-driver = "iscsi";
-if (!(fileprops = qemuBlockStorageSourceGetISCSIProps(src,
onlytarget)))
-return NULL;
-break;
-
-
[libvirt PATCH v7 28/35] qemu: implement password auth for ssh disks with nbdkit
For ssh disks that are served by nbdkit, lookup the password from the
configured secret and securely pass it to the nbdkit process using fd
passing.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c| 84 ++-
.../disk-network-ssh-password.args.disk0 | 8 ++
...k-network-ssh-password.args.disk0.pipe.778 | 1 +
.../disk-network-ssh.args.disk1 | 8 ++
.../disk-network-ssh.args.disk1.pipe.778 | 1 +
tests/qemunbdkittest.c| 1 +
...sk-network-ssh-password.x86_64-latest.args | 35
.../disk-network-ssh-password.xml | 34
tests/qemuxml2argvtest.c | 1 +
9 files changed, 134 insertions(+), 39 deletions(-)
create mode 100644 tests/qemunbdkitdata/disk-network-ssh-password.args.disk0
create mode 100644
tests/qemunbdkitdata/disk-network-ssh-password.args.disk0.pipe.778
create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk1
create mode 100644 tests/qemunbdkitdata/disk-network-ssh.args.disk1.pipe.778
create mode 100644
tests/qemuxml2argvdata/disk-network-ssh-password.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh-password.xml
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 3e82c90850..cedeaa880b 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -936,6 +936,43 @@ qemuNbdkitCommandPassDataByPipe(virCommand *cmd,
}
+static int
+qemuNbdkitProcessBuildCommandAuth(virStorageAuthDef *authdef,
+ virCommand *cmd)
+{
+g_autoptr(virConnect) conn = NULL;
+g_autofree uint8_t *secret = NULL;
+size_t secretlen = 0;
+int secrettype;
+
+if (!authdef)
+return 0;
+
+if ((secrettype = virSecretUsageTypeFromString(authdef->secrettype)) < 0) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("invalid secret type %1$s"),
+ authdef->secrettype);
+return -1;
+}
+
+conn = virGetConnectSecret();
+if (virSecretGetSecretString(conn,
+ >seclookupdef,
+ secrettype,
+ ,
+ ) < 0)
+return -1;
+
+virCommandAddArgPair(cmd, "user", authdef->username);
+
+if (qemuNbdkitCommandPassDataByPipe(cmd, "password",
+, secretlen) < 0)
+return -1;
+
+return 0;
+}
+
+
static int
qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
virCommand *cmd)
@@ -954,37 +991,8 @@ qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
}
virCommandAddArgPair(cmd, "url", uristring);
-if (proc->source->auth) {
-g_autoptr(virConnect) conn = virGetConnectSecret();
-g_autofree uint8_t *secret = NULL;
-size_t secretlen = 0;
-int secrettype;
-virStorageAuthDef *authdef = proc->source->auth;
-
-virCommandAddArgPair(cmd, "user",
- proc->source->auth->username);
-
-if ((secrettype =
virSecretUsageTypeFromString(proc->source->auth->secrettype)) < 0) {
-virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
- _("invalid secret type %1$s"),
- proc->source->auth->secrettype);
-return -1;
-}
-
-if (virSecretGetSecretString(conn,
- >seclookupdef,
- secrettype,
- ,
- ) < 0) {
-virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
- _("failed to get auth secret for storage"));
-return -1;
-}
-
-if (qemuNbdkitCommandPassDataByPipe(cmd, "password",
-, secretlen) < 0)
-return -1;
-}
+if (proc->source->auth &&
qemuNbdkitProcessBuildCommandAuth(proc->source->auth, cmd) < 0)
+return -1;
/* Create a pipe to send the cookies to the nbdkit process. */
if (proc->source->ncookies) {
@@ -1013,7 +1021,6 @@ static int
qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
virCommand *cmd)
{
-const char *user = NULL;
virStorageNetHostDef *host = >source->hosts[0];
g_autofree char *portstr = g_strdup_printf("%u", host->port);
@@ -1024,13 +1031,12 @@ qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess
*proc,
virCommandAddArgPair(cmd, "port", portstr);
virCommandAddArgPair(cmd, "path", proc->source->path);
-if (pr
[libvirt PATCH v7 34/35] qemu: implement ssh-agent auth for ssh disks with nbdkit
It's not possible to use password-protected ssh keys directly with
libvirt because libvirt doesn't have any way to prompt a user for the
password. To accomodate password-protected key files, an administrator
can add these keys to an ssh agent and then configure the domain with
the path to the ssh-agent socket.
Note that this requires an administrator or management app to
configure the ssh-agent with an appropriate socket path and add the
necessary keys to it. In addition, it does not currently work with
selinux enabled. The ssh-agent socket would need a label that libvirt
would be allowed to access rather than unconfined_t.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/domain_conf.c | 14 --
src/conf/storage_source_conf.c | 2 ++
src/conf/storage_source_conf.h | 1 +
src/qemu/qemu_nbdkit.c | 10 ++
.../disk-network-ssh-key.args.disk0 | 6 +++---
.../disk-network-ssh-key.args.disk1 | 9 +
tests/qemuxml2argvdata/disk-network-ssh-key.xml | 17 ++---
7 files changed, 51 insertions(+), 8 deletions(-)
create mode 100644 tests/qemunbdkitdata/disk-network-ssh-key.args.disk1
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 929e115bce..398f40d2be 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7277,8 +7277,17 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
if (!(src->ssh_user = virXMLPropStringRequired(tmpnode,
"username")))
return -1;
-if (!(src->ssh_keyfile = virXMLPropStringRequired(tmpnode,
"keyfile")))
+/* optional path to an ssh key file */
+src->ssh_keyfile = virXMLPropString(tmpnode, "keyfile");
+
+/* optional ssh-agent socket location */
+src->ssh_agent = virXMLPropString(tmpnode, "agentsock");
+if (!src->ssh_keyfile && !src->ssh_agent) {
+virReportError(VIR_ERR_XML_ERROR,
+ _("element '%1$s' requires either 'keyfile' or
'agentsock' attribute"),
+ tmpnode->name);
return -1;
+}
}
}
@@ -22291,11 +22300,12 @@ virDomainDiskSourceFormatNetwork(virBuffer *attrBuf,
if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH) {
if (src->ssh_known_hosts_file)
virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
-if (src->ssh_keyfile) {
+if (src->ssh_keyfile || src->ssh_agent) {
virBufferAddLit(childBuf, "ssh_user);
virBufferEscapeString(childBuf, " keyfile='%s'", src->ssh_keyfile);
+virBufferEscapeString(childBuf, " agentsock='%s'", src->ssh_agent);
virBufferAddLit(childBuf, "/>\n");
}
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index ce9c1f66c2..cafa031dfe 100644
--- a/src/conf/storage_source_conf.c
+++ b/src/conf/storage_source_conf.c
@@ -897,6 +897,7 @@ virStorageSourceCopy(const virStorageSource *src,
def->ssh_user = g_strdup(src->ssh_user);
def->ssh_known_hosts_file = g_strdup(src->ssh_known_hosts_file);
def->ssh_keyfile = g_strdup(src->ssh_keyfile);
+def->ssh_agent = g_strdup(src->ssh_agent);
def->nfs_user = g_strdup(src->nfs_user);
def->nfs_group = g_strdup(src->nfs_group);
@@ -1174,6 +1175,7 @@ virStorageSourceClear(virStorageSource *def)
VIR_FREE(def->ssh_user);
VIR_FREE(def->ssh_known_hosts_file);
VIR_FREE(def->ssh_keyfile);
+VIR_FREE(def->ssh_agent);
VIR_FREE(def->nfs_user);
VIR_FREE(def->nfs_group);
diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
index 8c805664af..061faa66cb 100644
--- a/src/conf/storage_source_conf.h
+++ b/src/conf/storage_source_conf.h
@@ -411,6 +411,7 @@ struct _virStorageSource {
bool ssh_host_key_check_disabled;
char *ssh_known_hosts_file;
char *ssh_keyfile;
+char *ssh_agent;
/* nfs_user and nfs_group store the strings passed in by the user for NFS
params.
* nfs_uid and nfs_gid represent the converted/looked up ID numbers which
are used
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 0393850ddc..a417146426 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -1043,6 +1043,9 @@ qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
virCommandAddArgPair(cmd, "user", proc->source->ssh_user);
}
+if (proc->source->ssh_agent)
+virCommandAddEnvPair(cmd, "SSH_AUTH_SOCK", proc->source->ssh_agent);
+
if (proc->source->ssh_host_key_check_disabled)
virCommand
[libvirt PATCH v7 25/35] qemu: Taint domain if nbdkit restart fails
Since the restart handler will trigger at an arbitrary time (when the
nbdkit process crashes, for instance), it's difficult to provide
feedback to the user if the restart is unsuccessful. Rather than just
relying on a warning in the log, taint the domain so that there will be
a slightly more user-visible notification.
Signed-off-by: Jonathon Jongsma
---
src/conf/domain_conf.c | 2 ++
src/conf/domain_conf.h | 1 +
src/qemu/qemu_nbdkit.c | 4 +++-
3 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index bb4f1fdb94..8feaf5d055 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -87,6 +87,7 @@ VIR_ENUM_IMPL(virDomainTaint,
"custom-hypervisor-feature",
"deprecated-config",
"custom-device",
+ "nbdkit-restart",
);
VIR_ENUM_IMPL(virDomainTaintMessage,
@@ -105,6 +106,7 @@ VIR_ENUM_IMPL(virDomainTaintMessage,
N_("hypervisor feature autodetection override"),
N_("use of deprecated configuration settings"),
N_("custom device configuration"),
+ N_("nbdkit restart failed"),
);
VIR_ENUM_IMPL(virDomainVirt,
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index ca195a52d2..c0729905a8 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -3194,6 +3194,7 @@ typedef enum {
VIR_DOMAIN_TAINT_CUSTOM_HYPERVISOR_FEATURE, /* custom hypervisor feature
control */
VIR_DOMAIN_TAINT_DEPRECATED_CONFIG, /* Configuration that is marked
deprecated */
VIR_DOMAIN_TAINT_CUSTOM_DEVICE, /* hypervisor device config customized */
+VIR_DOMAIN_TAINT_NBDKIT_RESTART,/* nbdkit could not be restarted */
VIR_DOMAIN_TAINT_LAST
} virDomainTaintFlags;
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index c3fa349922..ff5f1c0d05 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -623,8 +623,10 @@ qemuNbdkitProcessRestart(qemuNbdkitProcess *proc,
/* clean up resources associated with process */
qemuNbdkitProcessStop(proc);
-if (qemuNbdkitProcessStart(proc, vm, driver) < 0)
+if (qemuNbdkitProcessStart(proc, vm, driver) < 0) {
VIR_WARN("Unable to restart nbkdit process");
+virDomainObjTaint(vm, VIR_DOMAIN_TAINT_NBDKIT_RESTART);
+}
}
--
2.41.0
[libvirt PATCH v7 13/35] qemu: move qemuProcessReadLog() to qemuLogContext
This code can be used by the nbdkit implementation for reading back
filtered log data for error reporting. Move it to qemuLogContext so that
it can be shared. Renamed to qemuLogContextReadFiltered().
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_logcontext.c | 65
src/qemu/qemu_logcontext.h | 3 ++
src/qemu/qemu_process.c| 67 +-
3 files changed, 69 insertions(+), 66 deletions(-)
diff --git a/src/qemu/qemu_logcontext.c b/src/qemu/qemu_logcontext.c
index 0121ae5173..6e20f58bfa 100644
--- a/src/qemu/qemu_logcontext.c
+++ b/src/qemu/qemu_logcontext.c
@@ -21,6 +21,7 @@
#include "qemu_logcontext.h"
#include "viralloc.h"
#include "virlog.h"
+#include "virstring.h"
#include "virutil.h"
#include
@@ -236,6 +237,70 @@ qemuLogContextRead(qemuLogContext *ctxt,
}
+/**
+ * qemuLogContextFilter: Read and filter log for relevant messages
+ * @ctxt: the domain log context
+ * @msg: pointer to buffer to store the read messages in
+ * @max: maximum length of the message returned in @msg after filtering
+ *
+ * Reads log output from @ctxt and filters it. Skips messages not produced by
+ * the target executable or irrelevant messages. If @max is not zero, @buf will
+ * contain at most @max characters from the end of the log and @buf will start
+ * after a new line if possible.
+ */
+int
+qemuLogContextReadFiltered(qemuLogContext *ctxt,
+ char **msg,
+ size_t max)
+{
+char *buf;
+char *eol;
+char *filter_next;
+size_t skip;
+ssize_t got;
+
+if ((got = qemuLogContextRead(ctxt, )) < 0)
+return -1;
+
+/* Filter out debug messages from intermediate libvirt process */
+filter_next = buf;
+while ((eol = strchr(filter_next, '\n'))) {
+*eol = '\0';
+if (virLogProbablyLogMessage(filter_next) ||
+strstr(filter_next, "char device redirected to")) {
+skip = (eol + 1) - filter_next;
+memmove(filter_next, eol + 1, buf + got - eol);
+got -= skip;
+} else {
+filter_next = eol + 1;
+*eol = '\n';
+}
+}
+
+if (got > 0 &&
+buf[got - 1] == '\n') {
+buf[got - 1] = '\0';
+got--;
+}
+
+if (max > 0 && got > max) {
+skip = got - max;
+
+if (buf[skip - 1] != '\n' &&
+(eol = strchr(buf + skip, '\n')) &&
+!virStringIsEmpty(eol + 1))
+skip = eol + 1 - buf;
+
+memmove(buf, buf + skip, got - skip + 1);
+got -= skip;
+}
+
+buf = g_renew(char, buf, got + 1);
+*msg = buf;
+return 0;
+}
+
+
int
qemuLogContextGetWriteFD(qemuLogContext *ctxt)
{
diff --git a/src/qemu/qemu_logcontext.h b/src/qemu/qemu_logcontext.h
index 6ad60b7b4a..738e908bc3 100644
--- a/src/qemu/qemu_logcontext.h
+++ b/src/qemu/qemu_logcontext.h
@@ -32,6 +32,9 @@ int qemuLogContextWrite(qemuLogContext *ctxt,
const char *fmt, ...) G_GNUC_PRINTF(2, 3);
ssize_t qemuLogContextRead(qemuLogContext *ctxt,
char **msg);
+int qemuLogContextReadFiltered(qemuLogContext *ctxt,
+ char **msg,
+ size_t max);
int qemuLogContextGetWriteFD(qemuLogContext *ctxt);
void qemuLogContextMarkPosition(qemuLogContext *ctxt);
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index e03d0d8b4d..a77d2ba7de 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -1908,71 +1908,6 @@ qemuConnectMonitor(virQEMUDriver *driver,
}
-/**
- * qemuProcessReadLog: Read log file of a qemu VM
- * @logCtxt: the domain log context
- * @msg: pointer to buffer to store the read messages in
- * @max: maximum length of the message returned in @msg
- *
- * Reads log of a qemu VM. Skips messages not produced by qemu or irrelevant
- * messages. If @max is not zero, @msg will contain at most @max characters
- * from the end of the log and @msg will start after a new line if possible.
- *
- * Returns 0 on success or -1 on error
- */
-static int
-qemuProcessReadLog(qemuLogContext *logCtxt,
- char **msg,
- size_t max)
-{
-char *buf;
-ssize_t got;
-char *eol;
-char *filter_next;
-size_t skip;
-
-if ((got = qemuLogContextRead(logCtxt, )) < 0)
-return -1;
-
-/* Filter out debug messages from intermediate libvirt process */
-filter_next = buf;
-while ((eol = strchr(filter_next, '\n'))) {
-*eol = '\0';
-if (virLogProbablyLogMessage(filter_next) ||
-strstr(filter_next, "char device redirected to")) {
-skip = (eol + 1) - filter_next;
-memmove(filter_next, eol + 1, buf + got - eol);
-got -= skip;
-} else {
-
[libvirt PATCH v7 16/35] qemu: split qemuDomainSecretStorageSourcePrepare
This prepares encryption secrets and authentication secrets. When we add
nbdkit-backed network storage sources, we will not need to send
authentication secrets to qemu, since they will be sent to nbdkit
instead. So split this into two different functions.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c | 103 +
1 file changed, 62 insertions(+), 41 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 23608f95bd..951f3127d9 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -1398,38 +1398,70 @@
qemuDomainSecretStorageSourcePrepareCookies(qemuDomainObjPrivate *priv,
/**
- * qemuDomainSecretStorageSourcePrepare:
+ * qemuDomainSecretStorageSourcePrepareEncryption:
* @priv: domain private object
* @src: storage source struct to setup
- * @authalias: prefix of the alias for secret holding authentication data
- * @encalias: prefix of the alias for secret holding encryption password
+ * @alias: prefix of the alias for secret holding encryption password
*
- * Prepares data necessary for encryption and authentication of @src. The two
- * alias prefixes are provided since in the backing chain authentication
belongs
- * to the storage protocol data whereas encryption is relevant to the format
- * driver in qemu. The two will have different node names.
+ * Prepares data necessary for encryption of @src.
*
* Returns 0 on success; -1 on error while reporting an libvirt error.
*/
static int
-qemuDomainSecretStorageSourcePrepare(qemuDomainObjPrivate *priv,
- virStorageSource *src,
- const char *aliasprotocol,
- const char *aliasformat)
+qemuDomainSecretStorageSourcePrepareEncryption(qemuDomainObjPrivate *priv,
+ virStorageSource *src,
+ const char *alias)
{
qemuDomainStorageSourcePrivate *srcPriv;
-bool hasEnc = src->encryption && src->encryption->nsecrets > 0;
+size_t nsecrets = 0;
+size_t i;
-if (virStorageSourceIsEmpty(src))
+if (!(src->encryption && src->encryption->nsecrets > 0))
return 0;
-if (!src->auth && !hasEnc && src->ncookies == 0)
+if (virStorageSourceIsEmpty(src))
return 0;
-if (!(src->privateData = qemuDomainStorageSourcePrivateNew()))
-return -1;
+nsecrets = src->encryption->nsecrets;
+
+srcPriv = qemuDomainStorageSourcePrivateFetch(src);
+
+srcPriv->enccount = nsecrets;
+srcPriv->encinfo = g_new0(qemuDomainSecretInfo *, nsecrets);
+for (i = 0; i < nsecrets; ++i) {
+if (!(srcPriv->encinfo[i] = qemuDomainSecretInfoSetupFromSecret(priv,
alias,
+
"encryption", i,
+
VIR_SECRET_USAGE_TYPE_VOLUME,
+NULL,
+
>encryption->secrets[i]->seclookupdef)))
+return -1;
+}
+
+return 0;
+}
+
+
+/**
+ * qemuDomainSecretStorageSourcePrepareAuth:
+ * @priv: domain private object
+ * @src: storage source struct to setup
+ * @alias: prefix of the alias for secret holding authentication data
+ *
+ * Prepares data necessary for authentication of @src.
+ *
+ * Returns 0 on success; -1 on error while reporting an libvirt error.
+ */
+static int
+qemuDomainSecretStorageSourcePrepareAuth(qemuDomainObjPrivate *priv,
+ virStorageSource *src,
+ const char *alias)
+{
+qemuDomainStorageSourcePrivate *srcPriv;
+
+if (virStorageSourceIsEmpty(src))
+return 0;
-srcPriv = QEMU_DOMAIN_STORAGE_SOURCE_PRIVATE(src);
+srcPriv = qemuDomainStorageSourcePrivateFetch(src);
if (src->auth) {
virSecretUsageType usageType = VIR_SECRET_USAGE_TYPE_ISCSI;
@@ -1437,7 +1469,7 @@ qemuDomainSecretStorageSourcePrepare(qemuDomainObjPrivate
*priv,
if (src->protocol == VIR_STORAGE_NET_PROTOCOL_RBD)
usageType = VIR_SECRET_USAGE_TYPE_CEPH;
-if (!(srcPriv->secinfo = qemuDomainSecretInfoSetupFromSecret(priv,
aliasprotocol,
+if (!(srcPriv->secinfo = qemuDomainSecretInfoSetupFromSecret(priv,
alias,
"auth", 0,
usageType,
src->auth->username,
@@ -1445,26 +1477,10 @@
qemuDomainSecretStorageSourcePrepare(qemuDomainObjPrivate
[libvirt PATCH v7 30/35] qemu: implement knownHosts for ssh disks with nbdkit
For ssh disks that are served by nbdkit, use the configured value for
knownHosts and pass it to the nbdkit process.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/domain_conf.c| 8 ++
src/conf/storage_source_conf.c| 2 ++
src/conf/storage_source_conf.h| 2 ++
src/qemu/qemu_extdevice.c | 4 +--
src/qemu/qemu_hotplug.c | 4 +--
src/qemu/qemu_nbdkit.c| 25 +++
src/qemu/qemu_nbdkit.h| 6 +++--
.../disk-network-ssh-password.args.disk0 | 3 ++-
.../disk-network-ssh.args.disk0 | 3 ++-
.../disk-network-ssh-password.xml | 1 +
tests/qemuxml2argvdata/disk-network-ssh.xml | 1 +
11 files changed, 46 insertions(+), 13 deletions(-)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 8feaf5d055..842b6404b5 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -7268,6 +7268,11 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
return -1;
}
}
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
+(tmpnode = virXPathNode("./knownHosts", ctxt))) {
+if (!(src->ssh_known_hosts_file = virXMLPropStringRequired(tmpnode,
"path")))
+return -1;
+}
return 0;
}
@@ -22274,6 +22279,9 @@ virDomainDiskSourceFormatNetwork(virBuffer *attrBuf,
if (src->timeout)
virBufferAsprintf(childBuf, "\n",
src->timeout);
+
+if (src->protocol == VIR_STORAGE_NET_PROTOCOL_SSH &&
src->ssh_known_hosts_file)
+virBufferEscapeString(childBuf, "\n",
src->ssh_known_hosts_file);
}
diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
index dcac3a8ff6..906bc36a9b 100644
--- a/src/conf/storage_source_conf.c
+++ b/src/conf/storage_source_conf.c
@@ -895,6 +895,7 @@ virStorageSourceCopy(const virStorageSource *src,
/* ssh config passthrough for libguestfs */
def->ssh_host_key_check_disabled = src->ssh_host_key_check_disabled;
def->ssh_user = g_strdup(src->ssh_user);
+def->ssh_known_hosts_file = g_strdup(src->ssh_known_hosts_file);
def->nfs_user = g_strdup(src->nfs_user);
def->nfs_group = g_strdup(src->nfs_group);
@@ -1170,6 +1171,7 @@ virStorageSourceClear(virStorageSource *def)
VIR_FREE(def->tlsHostname);
VIR_FREE(def->ssh_user);
+VIR_FREE(def->ssh_known_hosts_file);
VIR_FREE(def->nfs_user);
VIR_FREE(def->nfs_group);
diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
index f13e7c756a..8a9c7d07e2 100644
--- a/src/conf/storage_source_conf.h
+++ b/src/conf/storage_source_conf.h
@@ -410,6 +410,8 @@ struct _virStorageSource {
/* these must not be used apart from formatting the output JSON in the
qemu driver */
char *ssh_user;
bool ssh_host_key_check_disabled;
+/* additional ssh variables */
+char *ssh_known_hosts_file;
/* nfs_user and nfs_group store the strings passed in by the user for NFS
params.
* nfs_uid and nfs_gid represent the converted/looked up ID numbers which
are used
diff --git a/src/qemu/qemu_extdevice.c b/src/qemu/qemu_extdevice.c
index 42ecdf13d5..3cf3867056 100644
--- a/src/qemu/qemu_extdevice.c
+++ b/src/qemu/qemu_extdevice.c
@@ -297,11 +297,11 @@ qemuExtDevicesStop(virQEMUDriver *driver,
for (i = 0; i < def->ndisks; i++) {
virDomainDiskDef *disk = def->disks[i];
-qemuNbdkitStopStorageSource(disk->src);
+qemuNbdkitStopStorageSource(disk->src, vm);
}
if (def->os.loader && def->os.loader->nvram)
-qemuNbdkitStopStorageSource(def->os.loader->nvram);
+qemuNbdkitStopStorageSource(def->os.loader->nvram, vm);
}
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index dc06486922..bbc5177206 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -1036,7 +1036,7 @@ qemuDomainAttachDeviceDiskLiveInternal(virQEMUDriver
*driver,
if (virStorageSourceChainHasManagedPR(disk->src))
ignore_value(qemuHotplugRemoveManagedPR(vm, VIR_ASYNC_JOB_NONE));
-qemuNbdkitStopStorageSource(disk->src);
+qemuNbdkitStopStorageSource(disk->src, vm);
}
qemuDomainSecretDiskDestroy(disk);
qemuDomainCleanupStorageSourceFD(disk->src);
@@ -4496,7 +4496,7 @@ qemuDomainRemoveDiskDevice(virQEMUDriver *driver,
qemuHotplugRemoveManagedPR(vm, VIR_ASYNC_JOB_NONE) < 0)
goto cleanup;
-qemuNbdkitStopStorageSource(disk->src);
+qemuNbdkitStopStorageSource(disk->src, vm);
if (disk->transient) {
VIR_DEBUG("Removing transient overlay '%s' of disk '%s'",
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu
[libvirt PATCH v7 26/35] qemu: try to connect to nbdkit early to detect errors
When using nbdkit to serve a network disk source, the nbdkit process
will start and wait for an nbd connection before actually attempting to
connect to the (remote) disk location. Because of this, nbdkit will not
report an error until after qemu is launched and tries to read from the
disk. This results in a fairly user-unfriendly error saying that qemu
was unable to start because "Requested export not available".
Ideally we'd like to be able to tell the user *why* the export is not
available, but this sort of information is only available to nbdkit, not
qemu. It could be because the url was incorrect, or because of an
authentication failure, or one of many other possibilities.
To make this friendlier for users and easier to detect
misconfigurations, try to connect to nbdkit immediately after starting
nbdkit and before we try to start qemu. This requires adding a
dependency on libnbd. If an error occurs when connecting to nbdkit, read
back from the nbdkit error log and provide that information in the error
report from qemuNbdkitProcessStart().
User-visible change demonstrated below:
Previous error:
$ virsh start nbdkit-test
2023-01-18 19:47:45.778+: 30895: error :
virNetClientProgramDispatchError:172 : internal
error: process exited while connecting to monitor:
2023-01-18T19:47:45.704658Z
qemu-system-x86_64: -blockdev {"driver":"nbd","server":{"type":"unix",
"path":"/var/lib/libvirt/qemu/domain-1-nbdkit-test/nbdkit-libvirt-1-storage.socket"},
"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}:
Requested export not
available
error: Failed to start domain 'nbdkit-test'
error: internal error: process exited while connecting to monitor:
2023-01-18T19:47:45.704658Z
qemu-system-x86_64: -blockdev {"driver":"nbd","server":{"type":"unix",
"path":"/var/lib/libvirt/qemu/domain-1-nbdkit-test/nbdkit-libvirt-1-storage.socket"},
"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}:
Requested export not
available
After this change:
$ virsh start nbdkit-test
2023-01-18 19:44:36.242+: 30895: error :
virNetClientProgramDispatchError:172 : internal
error: Failed to connect to nbdkit for
'http://localhost:/nonexistent.iso': nbdkit: curl[1]:
error: problem doing HEAD request to fetch size of URL
[http://localhost:/nonexistent.iso]:
HTTP response code said error: The requested URL returned error: 404
error: Failed to start domain 'nbdkit-test'
error: internal error: Failed to connect to nbdkit for
'http://localhost:/nonexistent.iso]:
error: problem doing HEAD request to fetch size of URL
[http://localhost:/nonexistent.iso]:
HTTP response code said error: The requested URL returned error: 404
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
meson.build| 7 +++
meson_options.txt | 1 +
src/qemu/meson.build | 1 +
src/qemu/qemu_nbdkit.c | 23 +++
4 files changed, 32 insertions(+)
diff --git a/meson.build b/meson.build
index 8da3b59070..a5ab94c0ba 100644
--- a/meson.build
+++ b/meson.build
@@ -1002,6 +1002,12 @@ endif
libiscsi_version = '1.18.0'
libiscsi_dep = dependency('libiscsi', version: '>=' + libiscsi_version,
required: get_option('libiscsi'))
+libnbd_version = '1.0'
+libnbd_dep = dependency('libnbd', version: '>=' + libnbd_version, required:
get_option('libnbd'))
+if libnbd_dep.found()
+ conf.set('WITH_LIBNBD', 1)
+endif
+
libnl_version = '3.0'
if not get_option('libnl').disabled() and host_machine.system() == 'linux'
libnl_dep = dependency('libnl-3.0', version: '>=' + libnl_version, required:
get_option('libnl'))
@@ -2216,6 +,7 @@ libs_summary = {
'glusterfs': glusterfs_dep.found(),
'libiscsi': libiscsi_dep.found(),
'libkvm': libkvm_dep.found(),
+ 'libnbd': libnbd_dep.found(),
'libnl': libnl_dep.found(),
'libparted': libparted_dep.found(),
'libpcap': libpcap_dep.found(),
diff --git a/meson_options.txt b/meson_options.txt
index 9174c4021c..ba6e49afc5 100644
--- a/meson_options.txt
+++ b/meson_options.txt
@@ -25,6 +25,7 @@ option('curl', type: 'feature', value: 'auto', description:
'curl support')
option('fuse', type: 'feature', value: 'auto', description: 'fuse support')
option('glusterfs', type: 'feature', value: 'auto', description: 'glusterfs
support')
option('libiscsi', type: 'feature', value: 'auto', description: 'libiscsi
support')
+option('libnbd', type: 'feature', value: 'auto', description: 'libnbd support')
option('libnl', type: 'feature', value: 'auto', description: 'libnl support')
option('libpcap', type: 'feature', value: 'auto', description: 'libpcap
support')
option('
[libvirt PATCH v7 07/35] qemu: use file cache for nbdkit caps
Add the virFileCache implementation for nbdkit capabilities to the qemu
driver. This allows us to determine whether nbdkit is installed and
which plugins are supported. it also has persistent caching and the
capabilities are re-queried whenever something changes.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_conf.h | 3 +++
src/qemu/qemu_driver.c | 3 +++
2 files changed, 6 insertions(+)
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index 4f610d86a1..a44985fb8b 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -318,6 +318,9 @@ struct _virQEMUDriver {
/* Immutable pointer, self-locking APIs */
virHashAtomic *migrationErrors;
+
+/* Immutable pointer, self-locking APIs */
+virFileCache *nbdkitCapsCache;
};
virQEMUDriverConfig *virQEMUDriverConfigNew(bool privileged,
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 5db42f0753..ad8428948b 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -845,6 +845,8 @@ qemuStateInitialize(bool privileged,
defsecmodel)))
goto error;
+qemu_driver->nbdkitCapsCache = qemuNbdkitCapsCacheNew(cfg->cacheDir);
+
/* If hugetlbfs is present, then we need to create a sub-directory within
* it, since we can't assume the root mount point has permissions that
* will let our spawned QEMU instances use it. */
@@ -1078,6 +1080,7 @@ qemuStateCleanup(void)
ebtablesContextFree(qemu_driver->ebtables);
VIR_FREE(qemu_driver->qemuImgBinary);
virObjectUnref(qemu_driver->domains);
+virObjectUnref(qemu_driver->nbdkitCapsCache);
if (qemu_driver->lockFD != -1)
virPidFileRelease(qemu_driver->config->stateDir, "driver",
qemu_driver->lockFD);
--
2.41.0
[libvirt PATCH v7 29/35] schema: add configuration for host verification of ssh disks
In order to make ssh disks usable, we need to be able to validate a remote host. To do this, add a xml element for ssh disks to allow the user to specify a location for a file that contains known host keys. Implementation to follow. Signed-off-by: Jonathon Jongsma Reviewed-by: Peter Krempa --- docs/formatdomain.rst | 8 src/conf/schemas/domaincommon.rng | 11 +++ 2 files changed, 19 insertions(+) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 39d4230ec0..9230f540f4 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3021,6 +3021,14 @@ paravirtualized driver is specified via the ``disk`` element. paused and will be rerun after a successful reconnect. After that time, any delayed requests and all future requests before a successful reconnect will immediately fail. If not set the default QEMU value is 0. + ``knownHosts`` + For storage accessed via the ``ssh`` protocol, this element configures a + path to a file that will be used to verify the remote host. This file + must contain the expected host key for the remote host or the connection + will fail. The location of the file is specified via the ``path`` + attribute. + :since:`Since 9.6.0` + For a "file" or "volume" disk type which represents a cdrom or floppy (the ``device`` attribute), it is possible to define policy what to do with the diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index cd838a475c..ca43586323 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -2172,6 +2172,14 @@ + + + + + + + + @@ -2187,6 +2195,9 @@ + + + -- 2.41.0
[libvirt PATCH v7 10/35] qemu: add functions to start and stop nbdkit
Add some helper functions to build a virCommand object and run the
nbdkit process for a given virStorageSource.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 250 +
src/qemu/qemu_nbdkit.h | 10 ++
2 files changed, 260 insertions(+)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 9a2a89224d..6bf962d0f1 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -24,6 +24,8 @@
#include "virerror.h"
#include "virlog.h"
#include "virpidfile.h"
+#include "virsecureerase.h"
+#include "virtime.h"
#include "virutil.h"
#include "qemu_block.h"
#include "qemu_conf.h"
@@ -666,6 +668,168 @@ qemuNbdkitInitStorageSource(qemuNbdkitCaps *caps,
}
+static int
+qemuNbdkitProcessBuildCommandCurl(qemuNbdkitProcess *proc,
+ virCommand *cmd)
+{
+g_autoptr(virURI) uri = qemuBlockStorageSourceGetURI(proc->source);
+g_autofree char *uristring = virURIFormat(uri);
+
+/* nbdkit plugin name */
+virCommandAddArg(cmd, "curl");
+if (proc->source->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP) {
+/* allow http to be upgraded to https via e.g. redirect */
+virCommandAddArgPair(cmd, "protocols", "http,https");
+} else {
+virCommandAddArgPair(cmd, "protocols",
+
virStorageNetProtocolTypeToString(proc->source->protocol));
+}
+virCommandAddArgPair(cmd, "url", uristring);
+
+if (proc->source->auth) {
+g_autoptr(virConnect) conn = virGetConnectSecret();
+g_autofree uint8_t *secret = NULL;
+size_t secretlen = 0;
+g_autofree char *password = NULL;
+int secrettype;
+virStorageAuthDef *authdef = proc->source->auth;
+
+virCommandAddArgPair(cmd, "user",
+ proc->source->auth->username);
+
+if ((secrettype =
virSecretUsageTypeFromString(proc->source->auth->secrettype)) < 0) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("invalid secret type %1$s"),
+ proc->source->auth->secrettype);
+return -1;
+}
+
+if (virSecretGetSecretString(conn,
+ >seclookupdef,
+ secrettype,
+ ,
+ ) < 0) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("failed to get auth secret for storage"));
+return -1;
+}
+
+/* ensure that the secret is a NULL-terminated string */
+password = g_strndup((char*)secret, secretlen);
+virSecureErase(secret, secretlen);
+
+/* for now, just report an error rather than passing the password in
+ * cleartext on the commandline */
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("Password not yet supported for nbdkit sources"));
+
+virSecureEraseString(password);
+
+return -1;
+}
+
+if (proc->source->ncookies > 0) {
+/* for now, just report an error rather than passing cookies in
+ * cleartext on the commandline */
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+ _("Cookies not yet supported for nbdkit sources"));
+return -1;
+}
+
+if (proc->source->sslverify == VIR_TRISTATE_BOOL_NO) {
+virCommandAddArgPair(cmd, "sslverify", "false");
+}
+
+if (proc->source->timeout > 0) {
+g_autofree char *timeout = g_strdup_printf("%llu",
proc->source->timeout);
+virCommandAddArgPair(cmd, "timeout", timeout);
+}
+
+return 0;
+}
+
+
+static int
+qemuNbdkitProcessBuildCommandSSH(qemuNbdkitProcess *proc,
+ virCommand *cmd)
+{
+const char *user = NULL;
+virStorageNetHostDef *host = >source->hosts[0];
+g_autofree char *portstr = g_strdup_printf("%u", host->port);
+
+/* nbdkit plugin name */
+virCommandAddArg(cmd, "ssh");
+
+virCommandAddArgPair(cmd, "host", host->name);
+virCommandAddArgPair(cmd, "port", portstr);
+virCommandAddArgPair(cmd, "path", proc->source->path);
+
+if (proc->source->auth)
+user = proc->source->auth->username;
+else if (proc->source->ssh_user)
+user = proc->source->ssh_user;
+
+if (user)
+virCommandAddArgPair(cmd, "user", user);
+
+if (proc->source->ssh_host_key_check_di
[libvirt PATCH v7 15/35] tests: add ability to test various nbdkit capabilities
Add new DO_TEST_CAPS_LATEST_NBDKIT macro to test xml2argv for various
nbdkit capability scenarios.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 20 +---
tests/qemuxml2argvtest.c | 11 +++
tests/testutilsqemu.c| 26 ++
tests/testutilsqemu.h| 4
4 files changed, 58 insertions(+), 3 deletions(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 2d70e72c42..81861bae4a 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -290,10 +290,16 @@ virNbkditCapsCheckModdir(const char *moddir,
static bool
virNbdkitCapsIsValid(void *data,
- void *privData G_GNUC_UNUSED)
+ void *privData)
{
qemuNbdkitCaps *nbdkitCaps = data;
struct stat st;
+/* when run under test, we will use privData as a signal to indicate that
+ * we shouldn't touch the filesystem */
+bool skipValidation = (privData != NULL);
+
+if (skipValidation)
+return true;
if (!nbdkitCaps->path)
return true;
@@ -334,9 +340,17 @@ virNbdkitCapsIsValid(void *data,
static void*
virNbdkitCapsNewData(const char *binary,
- void *privData G_GNUC_UNUSED)
+ void *privData)
{
-qemuNbdkitCaps *caps = qemuNbdkitCapsNew(binary);
+/* when run under test, we will use privData as a signal to indicate that
+ * we shouldn't touch the filesystem */
+bool skipNewData = (privData != NULL);
+qemuNbdkitCaps *caps = NULL;
+
+if (skipNewData)
+return NULL;
+
+caps = qemuNbdkitCapsNew(binary);
qemuNbdkitCapsQuery(caps);
return caps;
diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c
index 1b76b32812..d64c21ae17 100644
--- a/tests/qemuxml2argvtest.c
+++ b/tests/qemuxml2argvtest.c
@@ -611,6 +611,14 @@ testCompareXMLToArgv(const void *data)
if (qemuTestCapsCacheInsert(driver.qemuCapsCache, info->qemuCaps) < 0)
goto cleanup;
+if (info->nbdkitCaps) {
+if (virFileCacheInsertData(driver.nbdkitCapsCache, TEST_NBDKIT_PATH,
+ g_object_ref(info->nbdkitCaps)) < 0) {
+g_object_unref(info->nbdkitCaps);
+goto cleanup;
+}
+}
+
if (info->migrateFrom &&
!(migrateURI = qemuMigrationDstGetURI(info->migrateFrom,
info->migrateFd)))
@@ -831,6 +839,9 @@ mymain(void)
# define DO_TEST_CAPS_ARCH_VER(name, arch, ver) \
DO_TEST_CAPS_ARCH_VER_FULL(name, arch, ver, ARG_END)
+# define DO_TEST_CAPS_LATEST_NBDKIT(name, ...) \
+DO_TEST_CAPS_ARCH_LATEST_FULL(name, "x86_64", ARG_NBDKIT_CAPS,
__VA_ARGS__, QEMU_NBDKIT_CAPS_LAST, ARG_END)
+
# define DO_TEST_CAPS_LATEST(name) \
DO_TEST_CAPS_ARCH_LATEST(name, "x86_64")
diff --git a/tests/testutilsqemu.c b/tests/testutilsqemu.c
index fdbad16abe..9a607ab5a3 100644
--- a/tests/testutilsqemu.c
+++ b/tests/testutilsqemu.c
@@ -50,6 +50,10 @@ virFindFileInPath(const char *file)
return g_strdup_printf("/usr/bin/%s", file);
}
+if (g_str_equal(file, "nbdkit")) {
+return g_strdup(TEST_NBDKIT_PATH);
+}
+
/* Nothing in tests should be relying on real files
* in host OS, so we return NULL to try to force
* an error in such a case
@@ -288,6 +292,7 @@ void qemuTestDriverFree(virQEMUDriver *driver)
virObjectUnref(driver->caps);
virObjectUnref(driver->config);
virObjectUnref(driver->securityManager);
+g_clear_object(>nbdkitCapsCache);
virCPUDefFree(cpuDefault);
virCPUDefFree(cpuHaswell);
@@ -487,6 +492,12 @@ int qemuTestDriverInit(virQEMUDriver *driver)
if (!driver->qemuCapsCache)
goto error;
+driver->nbdkitCapsCache = qemuNbdkitCapsCacheNew("/dev/null");
+/* the nbdkitCapsCache just interprets the presence of a non-null private
+ * data pointer as a signal to skip cache validation. This prevents the
+ * cache from trying to validate the plugindir mtime, etc during test */
+virFileCacheSetPriv(driver->nbdkitCapsCache, GUINT_TO_POINTER(1));
+
driver->xmlopt = virQEMUDriverCreateXMLConf(driver, "none");
if (!driver->xmlopt)
goto error;
@@ -780,6 +791,14 @@ testQemuInfoSetArgs(struct testQemuInfo *info,
ignore_value(virBitmapSetBit(info->args.fakeCapsDel, flag));
break;
+case ARG_NBDKIT_CAPS:
+if (!(info->args.fakeNbdkitCaps))
+info->args.fakeNbdkitCaps =
virBitmapNew(QEMU_NBDKIT_CAPS_LAST);
+
+while ((flag = va_arg(argptr, int)) < QEMU_NBDKIT_CAPS_LAST)
+ignore_value(virBitmapSetBit(info->args.fakeNbdkitCaps, flag));
+break;
+
case ARG_GIC:
info->args.gic = va_arg(argptr, int);
[libvirt PATCH v7 11/35] Generalize qemuDomainLogContextNew()
Allow to specify a basename for the log file so that
qemuDomainLogContextNew() can be used to create log contexts for
secondary loggers.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_domain.c | 5 +++--
src/qemu/qemu_domain.h | 3 ++-
src/qemu/qemu_process.c | 2 +-
3 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 5a2eb4868a..d79f9879df 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -7119,7 +7119,8 @@ void qemuDomainObjCheckNetTaint(virQEMUDriver *driver,
qemuDomainLogContext *qemuDomainLogContextNew(virQEMUDriver *driver,
- virDomainObj *vm)
+ virDomainObj *vm,
+ const char *basename)
{
g_autoptr(virQEMUDriverConfig) cfg = virQEMUDriverGetConfig(driver);
qemuDomainLogContext *ctxt =
QEMU_DOMAIN_LOG_CONTEXT(g_object_new(QEMU_TYPE_DOMAIN_LOG_CONTEXT, NULL));
@@ -7128,7 +7129,7 @@ qemuDomainLogContext
*qemuDomainLogContextNew(virQEMUDriver *driver,
ctxt->writefd = -1;
ctxt->readfd = -1;
-ctxt->path = g_strdup_printf("%s/%s.log", cfg->logDir, vm->def->name);
+ctxt->path = g_strdup_printf("%s/%s.log", cfg->logDir, basename);
if (cfg->stdioLogD) {
ctxt->manager = virLogManagerNew(driver->privileged);
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h
index 89edc75fcf..a262555c8c 100644
--- a/src/qemu/qemu_domain.h
+++ b/src/qemu/qemu_domain.h
@@ -657,7 +657,8 @@ void qemuDomainObjCheckNetTaint(virQEMUDriver *driver,
qemuDomainLogContext *logCtxt);
qemuDomainLogContext *qemuDomainLogContextNew(virQEMUDriver *driver,
- virDomainObj *vm);
+ virDomainObj *vm,
+ const char *basename);
int qemuDomainLogContextWrite(qemuDomainLogContext *ctxt,
const char *fmt, ...) G_GNUC_PRINTF(2, 3);
ssize_t qemuDomainLogContextRead(qemuDomainLogContext *ctxt,
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index a6ed69cfe2..e0385d11be 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -7616,7 +7616,7 @@ qemuProcessLaunch(virConnectPtr conn,
hookData.cfg = cfg;
VIR_DEBUG("Creating domain log file");
-if (!(logCtxt = qemuDomainLogContextNew(driver, vm))) {
+if (!(logCtxt = qemuDomainLogContextNew(driver, vm, vm->def->name))) {
virLastErrorPrefixMessage("%s", _("can't connect to virtlogd"));
goto cleanup;
}
--
2.41.0
[libvirt PATCH v7 12/35] qemu: Extract qemuDomainLogContext into a new file
This will allow us to use it for nbdkit logging in upcoming commits.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
po/POTFILES| 1 +
src/qemu/meson.build | 1 +
src/qemu/qemu_domain.c | 247 ++
src/qemu/qemu_domain.h | 27 +---
src/qemu/qemu_logcontext.c | 264 +
src/qemu/qemu_logcontext.h | 38 ++
src/qemu/qemu_process.c| 44 +++
7 files changed, 346 insertions(+), 276 deletions(-)
create mode 100644 src/qemu/qemu_logcontext.c
create mode 100644 src/qemu/qemu_logcontext.h
diff --git a/po/POTFILES b/po/POTFILES
index 6167f98ac5..3a51aea5cb 100644
--- a/po/POTFILES
+++ b/po/POTFILES
@@ -174,6 +174,7 @@ src/qemu/qemu_hostdev.c
src/qemu/qemu_hotplug.c
src/qemu/qemu_interface.c
src/qemu/qemu_interop_config.c
+src/qemu/qemu_logcontext.c
src/qemu/qemu_migration.c
src/qemu/qemu_migration_cookie.c
src/qemu/qemu_migration_params.c
diff --git a/src/qemu/meson.build b/src/qemu/meson.build
index 9be6996195..6d7a1bfbb0 100644
--- a/src/qemu/meson.build
+++ b/src/qemu/meson.build
@@ -21,6 +21,7 @@ qemu_driver_sources = [
'qemu_hotplug.c',
'qemu_interface.c',
'qemu_interop_config.c',
+ 'qemu_logcontext.c',
'qemu_migration.c',
'qemu_migration_cookie.c',
'qemu_migration_params.c',
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index d79f9879df..23608f95bd 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -455,21 +455,8 @@ qemuDomainObjFromDomain(virDomainPtr domain)
}
-struct _qemuDomainLogContext {
-GObject parent;
-
-int writefd;
-int readfd; /* Only used if manager == NULL */
-off_t pos;
-ino_t inode; /* Only used if manager != NULL */
-char *path;
-virLogManager *manager;
-};
-
-G_DEFINE_TYPE(qemuDomainLogContext, qemu_domain_log_context, G_TYPE_OBJECT);
static virClass *qemuDomainSaveCookieClass;
-static void qemuDomainLogContextFinalize(GObject *obj);
static void qemuDomainSaveCookieDispose(void *obj);
@@ -482,32 +469,8 @@ qemuDomainOnceInit(void)
return 0;
}
-static void qemu_domain_log_context_init(qemuDomainLogContext *logctxt
G_GNUC_UNUSED)
-{
-}
-
-static void qemu_domain_log_context_class_init(qemuDomainLogContextClass
*klass)
-{
-GObjectClass *obj = G_OBJECT_CLASS(klass);
-
-obj->finalize = qemuDomainLogContextFinalize;
-}
-
VIR_ONCE_GLOBAL_INIT(qemuDomain);
-static void
-qemuDomainLogContextFinalize(GObject *object)
-{
-qemuDomainLogContext *ctxt = QEMU_DOMAIN_LOG_CONTEXT(object);
-VIR_DEBUG("ctxt=%p", ctxt);
-
-virLogManagerFree(ctxt->manager);
-VIR_FREE(ctxt->path);
-VIR_FORCE_CLOSE(ctxt->writefd);
-VIR_FORCE_CLOSE(ctxt->readfd);
-G_OBJECT_CLASS(qemu_domain_log_context_parent_class)->finalize(object);
-}
-
/* qemuDomainGetMasterKeyFilePath:
* @libDir: Directory path to domain lib files
*
@@ -6882,7 +6845,7 @@ static void G_GNUC_PRINTF(5, 6)
qemuDomainObjTaintMsg(virQEMUDriver *driver,
virDomainObj *obj,
virDomainTaintFlags taint,
- qemuDomainLogContext *logCtxt,
+ qemuLogContext *logCtxt,
const char *fmt, ...)
{
virErrorPtr orig_err = NULL;
@@ -6935,12 +6898,12 @@ qemuDomainObjTaintMsg(virQEMUDriver *driver,
goto cleanup;
if (logCtxt) {
-rc = qemuDomainLogContextWrite(logCtxt,
- "%s: Domain id=%d is tainted:
%s%s%s%s\n",
- timestamp,
- obj->def->id,
- virDomainTaintTypeToString(taint),
- extraprefix, extramsg, extrasuffix);
+rc = qemuLogContextWrite(logCtxt,
+ "%s: Domain id=%d is tainted: %s%s%s%s\n",
+ timestamp,
+ obj->def->id,
+ virDomainTaintTypeToString(taint),
+ extraprefix, extramsg, extrasuffix);
} else {
rc = qemuDomainLogAppendMessage(driver, obj,
"%s: Domain id=%d is tainted:
%s%s%s%s\n",
@@ -6961,7 +6924,7 @@ qemuDomainObjTaintMsg(virQEMUDriver *driver,
void qemuDomainObjTaint(virQEMUDriver *driver,
virDomainObj *obj,
virDomainTaintFlags taint,
-qemuDomainLogContext *logCtxt)
+qemuLogContext *logCtxt)
{
qemuDomainObjTaintMsg(driver, obj, taint, logCtxt, NULL);
qemuDomainSaveStatus(obj);
@@ -6970,7 +6933,7 @@ void qemuDomainObjTaint(virQEMUDriver *driver,
static void
qemuDomainObjCheckMachineTaint(virQEMUDriver *driver,
virDomainO
[libvirt PATCH v7 08/35] qemu: Add qemuNbdkitProcess
An object for storing information about a nbdkit process that is serving
a specific virStorageSource. At the moment, this information is just
stored in the private data of virStorageSource and not used at all.
Future commits will use this data to actually start a nbdkit process.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_conf.c | 22
src/qemu/qemu_conf.h | 2 ++
src/qemu/qemu_domain.c | 31
src/qemu/qemu_domain.h | 4 +++
src/qemu/qemu_nbdkit.c | 82 ++
src/qemu/qemu_nbdkit.h | 26 ++
6 files changed, 167 insertions(+)
diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c
index 3f811d064f..a0360e8d1b 100644
--- a/src/qemu/qemu_conf.c
+++ b/src/qemu/qemu_conf.c
@@ -1656,3 +1656,25 @@ qemuHugepageMakeBasedir(virQEMUDriver *driver,
return 0;
}
+
+
+/*
+ * qemuGetNbdkitCaps:
+ * @driver: the qemu driver
+ *
+ * Gets the capabilities for Nbdkit for the specified driver. These can be used
+ * to determine whether a particular disk source can be served by nbdkit or
+ * not.
+ *
+ * Returns: a reference to qemuNbdkitCaps or NULL
+ */
+qemuNbdkitCaps*
+qemuGetNbdkitCaps(virQEMUDriver *driver)
+{
+char *nbdkitBinary = virFindFileInPath("nbdkit");
+
+if (!nbdkitBinary)
+return NULL;
+
+return virFileCacheLookup(driver->nbdkitCapsCache, nbdkitBinary);
+}
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index a44985fb8b..1a3ba3a0fb 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -377,3 +377,5 @@ int qemuGetMemoryBackingPath(virQEMUDriver *driver,
int qemuHugepageMakeBasedir(virQEMUDriver *driver,
virHugeTLBFS *hugepage);
+
+qemuNbdkitCaps* qemuGetNbdkitCaps(virQEMUDriver *driver);
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index bfeddc7746..5a2eb4868a 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -882,6 +882,7 @@ qemuDomainStorageSourcePrivateDispose(void *obj)
g_clear_pointer(>httpcookie, qemuDomainSecretInfoFree);
g_clear_pointer(>tlsKeySecret, qemuDomainSecretInfoFree);
g_clear_pointer(>fdpass, qemuFDPassFree);
+g_clear_pointer(>nbdkitProcess, qemuNbdkitProcessFree);
}
@@ -10471,6 +10472,34 @@ qemuDomainPrepareStorageSourceNFS(virStorageSource
*src)
}
+/* qemuPrepareStorageSourceNbdkit:
+ * @src: source for a disk
+ *
+ * If src is an network source that is managed by nbdkit, prepare data so that
+ * nbdkit can be launched before the domain is started
+ *
+ * Returns true if nbdkit will be used for this source,
+ */
+static bool
+qemuDomainPrepareStorageSourceNbdkit(virStorageSource *src,
+ virQEMUDriverConfig *cfg,
+ const char *alias,
+ qemuDomainObjPrivate *priv)
+{
+g_autoptr(qemuNbdkitCaps) nbdkit = NULL;
+
+if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_NETWORK)
+return false;
+
+nbdkit = qemuGetNbdkitCaps(priv->driver);
+if (!nbdkit)
+return false;
+
+return qemuNbdkitInitStorageSource(nbdkit, src, priv->libDir,
+ alias, cfg->user, cfg->group);
+}
+
+
/* qemuProcessPrepareStorageSourceTLS:
* @source: source for a disk
* @cfg: driver configuration
@@ -11300,6 +11329,8 @@
qemuDomainPrepareStorageSourceBlockdevNodename(virDomainDiskDef *disk,
if (qemuDomainPrepareStorageSourceFDs(src, priv) < 0)
return -1;
+qemuDomainPrepareStorageSourceNbdkit(src, cfg, src->nodestorage, priv);
+
return 0;
}
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h
index 786f239495..89edc75fcf 100644
--- a/src/qemu/qemu_domain.h
+++ b/src/qemu/qemu_domain.h
@@ -33,6 +33,7 @@
#include "qemu_conf.h"
#include "qemu_capabilities.h"
#include "qemu_migration_params.h"
+#include "qemu_nbdkit.h"
#include "qemu_slirp.h"
#include "qemu_fd.h"
#include "virchrdev.h"
@@ -308,6 +309,9 @@ struct _qemuDomainStorageSourcePrivate {
/* file descriptors if user asks for FDs to be passed */
qemuFDPass *fdpass;
+
+/* an nbdkit process for serving network storage sources */
+qemuNbdkitProcess *nbdkitProcess;
};
virObject *qemuDomainStorageSourcePrivateNew(void);
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 97612b637f..12c721f7f1 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -549,3 +549,85 @@ qemuNbdkitCapsCacheNew(const char *cachedir)
g_autofree char *dir = g_build_filename(cachedir, "nbdkitcapabilities",
NULL);
return virFileCacheNew(dir, "xml", );
}
+
+
+static qemuNbdkitProcess *
+qemuNbdkitProcessNew(virStorageSource *source,
+ const char *pidfile,
+ const char *socketfile)
+{
+qemu
[libvirt PATCH v7 09/35] qemu: query nbdkit module dir from binary
Rather than hard-coding the nbdkit module directory, query the nbdkit
binary for the location to these directories. nbdkit provides a
--dump-config optiont that outputs this information and can be easily
parsed. We can also get the version from this output rather than
executing `nbdkit --version` separately.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 77 --
1 file changed, 60 insertions(+), 17 deletions(-)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index 12c721f7f1..9a2a89224d 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -20,7 +20,6 @@
#include
#include
-#include "configmake.h"
#include "vircommand.h"
#include "virerror.h"
#include "virlog.h"
@@ -39,10 +38,6 @@
VIR_LOG_INIT("qemu.nbdkit");
-#define NBDKIT_MODDIR LIBDIR "/nbdkit"
-#define NBDKIT_PLUGINDIR NBDKIT_MODDIR "/plugins"
-#define NBDKIT_FILTERDIR NBDKIT_MODDIR "/filters"
-
VIR_ENUM_IMPL(qemuNbdkitCaps,
QEMU_NBDKIT_CAPS_LAST,
/* 0 */
@@ -56,6 +51,9 @@ struct _qemuNbdkitCaps {
char *path;
char *version;
+char *filterDir;
+char *pluginDir;
+
time_t ctime;
time_t libvirtCtime;
time_t pluginDirMtime;
@@ -129,18 +127,47 @@ qemuNbdkitCapsQueryFilters(qemuNbdkitCaps *nbdkit)
static int
-qemuNbdkitCapsQueryVersion(qemuNbdkitCaps *nbdkit)
+qemuNbdkitCapsQueryBuildConfig(qemuNbdkitCaps *nbdkit)
{
+size_t i;
+g_autofree char *output = NULL;
+g_auto(GStrv) lines = NULL;
+const char *line;
g_autoptr(virCommand) cmd = virCommandNewArgList(nbdkit->path,
- "--version",
+ "--dump-config",
NULL);
-virCommandSetOutputBuffer(cmd, >version);
+virCommandSetOutputBuffer(cmd, );
if (virCommandRun(cmd, NULL) != 0)
return -1;
-VIR_DEBUG("Got nbdkit version %s", nbdkit->version);
+lines = g_strsplit(output, "\n", 0);
+if (!lines)
+return -1;
+
+for (i = 0; (line = lines[i]); i++) {
+const char *key;
+const char *val;
+char *p;
+
+p = strchr(line, '=');
+if (!p)
+continue;
+
+*p = '\0';
+key = line;
+val = p + 1;
+
+VIR_DEBUG("Got nbdkit config value %s=%s", key, val);
+
+if (STREQ(key, "version"))
+nbdkit->version = g_strdup(val);
+else if (STREQ(key, "filterdir"))
+nbdkit->filterDir = g_strdup(val);
+else if (STREQ(key, "plugindir"))
+nbdkit->pluginDir = g_strdup(val);
+}
return 0;
}
@@ -152,6 +179,8 @@ qemuNbdkitCapsFinalize(GObject *object)
g_clear_pointer(>path, g_free);
g_clear_pointer(>version, g_free);
+g_clear_pointer(>filterDir, g_free);
+g_clear_pointer(>pluginDir, g_free);
g_clear_pointer(>flags, virBitmapFree);
G_OBJECT_CLASS(qemu_nbdkit_caps_parent_class)->finalize(object);
@@ -214,15 +243,15 @@ qemuNbdkitCapsQuery(qemuNbdkitCaps *caps)
return;
}
+qemuNbdkitCapsQueryBuildConfig(caps);
+qemuNbdkitCapsQueryPlugins(caps);
+qemuNbdkitCapsQueryFilters(caps);
+
caps->ctime = st.st_ctime;
-caps->filterDirMtime = qemuNbdkitGetDirMtime(NBDKIT_FILTERDIR);
-caps->pluginDirMtime = qemuNbdkitGetDirMtime(NBDKIT_PLUGINDIR);
+caps->filterDirMtime = qemuNbdkitGetDirMtime(caps->filterDir);
+caps->pluginDirMtime = qemuNbdkitGetDirMtime(caps->pluginDir);
caps->libvirtCtime = virGetSelfLastChanged();
caps->libvirtVersion = LIBVIR_VERSION_NUMBER;
-
-qemuNbdkitCapsQueryPlugins(caps);
-qemuNbdkitCapsQueryFilters(caps);
-qemuNbdkitCapsQueryVersion(caps);
}
@@ -267,9 +296,9 @@ virNbdkitCapsIsValid(void *data,
if (!nbdkitCaps->path)
return true;
-if (!virNbkditCapsCheckModdir(NBDKIT_PLUGINDIR,
nbdkitCaps->pluginDirMtime))
+if (!virNbkditCapsCheckModdir(nbdkitCaps->pluginDir,
nbdkitCaps->pluginDirMtime))
return false;
-if (!virNbkditCapsCheckModdir(NBDKIT_FILTERDIR,
nbdkitCaps->filterDirMtime))
+if (!virNbkditCapsCheckModdir(nbdkitCaps->filterDir,
nbdkitCaps->filterDirMtime))
return false;
if (nbdkitCaps->libvirtCtime != virGetSelfLastChanged() ||
@@ -421,12 +450,22 @@ qemuNbdkitCapsLoadCache(qemuNbdkitCaps *nbdkitCaps,
}
nbdkitCaps->ctime = (time_t)l;
+if ((nbdkitCaps->pluginDir = virXPathString("string(./plugindir)", ctxt))
== NULL) {
+VIR_DEBUG("missing plugindir in nbdkit capabilities cache");
+return -1;
+}
+
if (virXPathLongLong(&quo
[libvirt PATCH v7 03/35] qemu: expand nbdkit capabilities
In order to add caching of the nbdkit capabilities, we will need to
compare against file modification times, etc. So look up this
information when creating the nbdkit caps.
Add a nbdkit_moddir build option to allow the builder to specify the
location to look for nbdkit plugins and filters.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/qemu_nbdkit.c | 42 ++
1 file changed, 42 insertions(+)
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
index e4e8fd568e..58828dd89a 100644
--- a/src/qemu/qemu_nbdkit.c
+++ b/src/qemu/qemu_nbdkit.c
@@ -20,6 +20,7 @@
#include
#include
+#include "configmake.h"
#include "vircommand.h"
#include "virerror.h"
#include "virlog.h"
@@ -38,6 +39,10 @@
VIR_LOG_INIT("qemu.nbdkit");
+#define NBDKIT_MODDIR LIBDIR "/nbdkit"
+#define NBDKIT_PLUGINDIR NBDKIT_MODDIR "/plugins"
+#define NBDKIT_FILTERDIR NBDKIT_MODDIR "/filters"
+
VIR_ENUM_IMPL(qemuNbdkitCaps,
QEMU_NBDKIT_CAPS_LAST,
/* 0 */
@@ -51,6 +56,11 @@ struct _qemuNbdkitCaps {
char *path;
char *version;
+time_t ctime;
+time_t libvirtCtime;
+time_t pluginDirMtime;
+time_t filterDirMtime;
+unsigned int libvirtVersion;
virBitmap *flags;
};
@@ -175,9 +185,41 @@ qemuNbdkitCapsNew(const char *path)
}
+static time_t
+qemuNbdkitGetDirMtime(const char *moddir)
+{
+struct stat st;
+
+if (stat(moddir, ) < 0) {
+VIR_DEBUG("Failed to stat nbdkit module directory '%s': %s",
+ moddir,
+ g_strerror(errno));
+return 0;
+}
+
+return st.st_mtime;
+}
+
+
G_GNUC_UNUSED static void
qemuNbdkitCapsQuery(qemuNbdkitCaps *caps)
{
+struct stat st;
+
+if (stat(caps->path, ) < 0) {
+VIR_DEBUG("Failed to stat nbdkit binary '%s': %s",
+ caps->path,
+ g_strerror(errno));
+caps->ctime = 0;
+return;
+}
+
+caps->ctime = st.st_ctime;
+caps->filterDirMtime = qemuNbdkitGetDirMtime(NBDKIT_FILTERDIR);
+caps->pluginDirMtime = qemuNbdkitGetDirMtime(NBDKIT_PLUGINDIR);
+caps->libvirtCtime = virGetSelfLastChanged();
+caps->libvirtVersion = LIBVIR_VERSION_NUMBER;
+
qemuNbdkitCapsQueryPlugins(caps);
qemuNbdkitCapsQueryFilters(caps);
qemuNbdkitCapsQueryVersion(caps);
--
2.41.0
[libvirt PATCH v7 01/35] schema: allow 'ssh' as a protocol for network disks
There was support in the code for parsing protocol='ssh' on network disk
sources, but it was not present in the xml schema. Add this to the
schema.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/conf/schemas/domaincommon.rng | 1 +
tests/qemublocktest.c | 2 +-
...w2-invalid.json => network-ssh-qcow2.json} | 0
...cow2-invalid.xml => network-ssh-qcow2.xml} | 0
.../disk-network-ssh.x86_64-latest.args | 35 +++
tests/qemuxml2argvdata/disk-network-ssh.xml | 31
tests/qemuxml2argvtest.c | 1 +
7 files changed, 69 insertions(+), 1 deletion(-)
rename tests/qemublocktestdata/imagecreate/{network-ssh-qcow2-invalid.json =>
network-ssh-qcow2.json} (100%)
rename tests/qemublocktestdata/imagecreate/{network-ssh-qcow2-invalid.xml =>
network-ssh-qcow2.xml} (100%)
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
create mode 100644 tests/qemuxml2argvdata/disk-network-ssh.xml
diff --git a/src/conf/schemas/domaincommon.rng
b/src/conf/schemas/domaincommon.rng
index de3bd1c35c..4a475f5c36 100644
--- a/src/conf/schemas/domaincommon.rng
+++ b/src/conf/schemas/domaincommon.rng
@@ -2179,6 +2179,7 @@
sheepdog
tftp
+ssh
diff --git a/tests/qemublocktest.c b/tests/qemublocktest.c
index 9a968477d7..8bad69e7ac 100644
--- a/tests/qemublocktest.c
+++ b/tests/qemublocktest.c
@@ -1213,7 +1213,7 @@ mymain(void)
TEST_IMAGE_CREATE("network-gluster-qcow2", NULL);
TEST_IMAGE_CREATE("network-rbd-qcow2", NULL);
-TEST_IMAGE_CREATE("network-ssh-qcow2-invalid", NULL);
+TEST_IMAGE_CREATE("network-ssh-qcow2", NULL);
#define TEST_BITMAP_DETECT(testname) \
do { \
diff --git a/tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.json
b/tests/qemublocktestdata/imagecreate/network-ssh-qcow2.json
similarity index 100%
rename from tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.json
rename to tests/qemublocktestdata/imagecreate/network-ssh-qcow2.json
diff --git a/tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.xml
b/tests/qemublocktestdata/imagecreate/network-ssh-qcow2.xml
similarity index 100%
rename from tests/qemublocktestdata/imagecreate/network-ssh-qcow2-invalid.xml
rename to tests/qemublocktestdata/imagecreate/network-ssh-qcow2.xml
diff --git a/tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
b/tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
new file mode 100644
index 00..b7fd30032b
--- /dev/null
+++ b/tests/qemuxml2argvdata/disk-network-ssh.x86_64-latest.args
@@ -0,0 +1,35 @@
+LC_ALL=C \
+PATH=/bin \
+HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1 \
+USER=test \
+LOGNAME=test \
+XDG_DATA_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.local/share \
+XDG_CACHE_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.cache \
+XDG_CONFIG_HOME=/var/lib/libvirt/qemu/domain--1-QEMUGuest1/.config \
+/usr/bin/qemu-system-x86_64 \
+-name guest=QEMUGuest1,debug-threads=on \
+-S \
+-object
'{"qom-type":"secret","id":"masterKey0","format":"raw","file":"/var/lib/libvirt/qemu/domain--1-QEMUGuest1/master-key.aes"}'
\
+-machine pc,usb=off,dump-guest-core=off,memory-backend=pc.ram,acpi=off \
+-accel kvm \
+-cpu qemu64 \
+-m size=219136k \
+-object '{"qom-type":"memory-backend-ram","id":"pc.ram","size":224395264}' \
+-overcommit mem-lock=off \
+-smp 1,sockets=1,cores=1,threads=1 \
+-uuid c7a5fdbd-edaf-9455-926a-d65c16db1809 \
+-display none \
+-no-user-config \
+-nodefaults \
+-chardev socket,id=charmonitor,fd=1729,server=on,wait=off \
+-mon chardev=charmonitor,id=monitor,mode=control \
+-rtc base=utc \
+-no-shutdown \
+-boot strict=on \
+-device
'{"driver":"piix3-usb-uhci","id":"usb","bus":"pci.0","addr":"0x1.0x2"}' \
+-blockdev
'{"driver":"ssh","path":"test.img","server":{"host":"example.org","port":""},"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}'
\
+-blockdev
'{"node-name":"libvirt-1-format","read-only":false,"driver":"raw","file":"libvirt-1-storage"}'
\
+-device
'{"driver":"virtio-blk-pci","bus":"pci.0","addr":"0x2","drive":"libvirt-1-format","id":"virtio-disk0","bootindex":1}'
\
+-audiodev '{"id":"audio1","driver":"none"}' \
+-sandbox
on,obsolete=deny,elevateprivileges=deny,sp
[libvirt PATCH v7 02/35] qemu: Add functions for determining nbdkit availability
In future commits, we will optionally use nbdkit to serve some remote
disk sources. This patch queries to see whether nbdkit is installed on
the host and queries it for capabilities. The data will be used in later
commits.
Signed-off-by: Jonathon Jongsma
Reviewed-by: Peter Krempa
---
src/qemu/meson.build | 1 +
src/qemu/qemu_conf.h | 1 +
src/qemu/qemu_nbdkit.c | 200 +
src/qemu/qemu_nbdkit.h | 50 +++
4 files changed, 252 insertions(+)
create mode 100644 src/qemu/qemu_nbdkit.c
create mode 100644 src/qemu/qemu_nbdkit.h
diff --git a/src/qemu/meson.build b/src/qemu/meson.build
index c8806bbc36..9be6996195 100644
--- a/src/qemu/meson.build
+++ b/src/qemu/meson.build
@@ -28,6 +28,7 @@ qemu_driver_sources = [
'qemu_monitor_json.c',
'qemu_monitor_text.c',
'qemu_namespace.c',
+ 'qemu_nbdkit.c',
'qemu_passt.c',
'qemu_process.c',
'qemu_qapi.c',
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index 11c740d28f..4f610d86a1 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -36,6 +36,7 @@
#include "virthreadpool.h"
#include "locking/lock_manager.h"
#include "qemu_capabilities.h"
+#include "qemu_nbdkit.h"
#include "virclosecallbacks.h"
#include "virhostdev.h"
#include "virfile.h"
diff --git a/src/qemu/qemu_nbdkit.c b/src/qemu/qemu_nbdkit.c
new file mode 100644
index 00..e4e8fd568e
--- /dev/null
+++ b/src/qemu/qemu_nbdkit.c
@@ -0,0 +1,200 @@
+/*
+ * qemu_nbdkit.c: helpers for using nbdkit
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library. If not, see
+ * <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include
+#include
+
+#include "vircommand.h"
+#include "virerror.h"
+#include "virlog.h"
+#include "virpidfile.h"
+#include "virutil.h"
+#include "qemu_block.h"
+#include "qemu_conf.h"
+#include "qemu_domain.h"
+#include "qemu_extdevice.h"
+#include "qemu_nbdkit.h"
+#include "qemu_security.h"
+
+#include
+
+#define VIR_FROM_THIS VIR_FROM_QEMU
+
+VIR_LOG_INIT("qemu.nbdkit");
+
+VIR_ENUM_IMPL(qemuNbdkitCaps,
+QEMU_NBDKIT_CAPS_LAST,
+/* 0 */
+"plugin-curl", /* QEMU_NBDKIT_CAPS_PLUGIN_CURL */
+"plugin-ssh", /* QEMU_NBDKIT_CAPS_PLUGIN_SSH */
+"filter-readahead", /* QEMU_NBDKIT_CAPS_FILTER_READAHEAD */
+);
+
+struct _qemuNbdkitCaps {
+GObject parent;
+
+char *path;
+char *version;
+
+virBitmap *flags;
+};
+G_DEFINE_TYPE(qemuNbdkitCaps, qemu_nbdkit_caps, G_TYPE_OBJECT);
+
+
+static void
+qemuNbdkitCheckCommandCap(qemuNbdkitCaps *nbdkit,
+ virCommand *cmd,
+ qemuNbdkitCapsFlags cap)
+{
+if (virCommandRun(cmd, NULL) != 0)
+return;
+
+VIR_DEBUG("Setting nbdkit capability %i", cap);
+ignore_value(virBitmapSetBit(nbdkit->flags, cap));
+}
+
+
+static void
+qemuNbdkitQueryFilter(qemuNbdkitCaps *nbdkit,
+ const char *filter,
+ qemuNbdkitCapsFlags cap)
+{
+g_autoptr(virCommand) cmd = virCommandNewArgList(nbdkit->path,
+ "--version",
+ NULL);
+
+virCommandAddArgPair(cmd, "--filter", filter);
+
+/* use null plugin to check for filter */
+virCommandAddArg(cmd, "null");
+
+qemuNbdkitCheckCommandCap(nbdkit, cmd, cap);
+}
+
+
+static void
+qemuNbdkitQueryPlugin(qemuNbdkitCaps *nbdkit,
+ const char *plugin,
+ qemuNbdkitCapsFlags cap)
+{
+g_autoptr(virCommand) cmd = virCommandNewArgList(nbdkit->path,
+ plugin,
+ "--version",
+ NULL);
+
+qemuNbdkitCheckCommandCap(nbdkit, cmd, cap);
+}
+
+
+static void
+qemuNbdkitCapsQueryPlugins(qemuNbdkitCaps *nbdkit)
+{
+qemuNbdkitQueryPlugin(nbdkit, "curl", QEMU_NBDKIT_CAPS_PLUGIN_CURL);
+qemuNbdkitQueryPlugin(nbdkit, "ssh", QEMU_NBDKIT_CAPS_PLUGIN_SSH);
+}
+
+
+static void
+qemuNbdkitCapsQueryFilters(qemuNbdkitCaps *nbdkit)
+{
+qemuNbd
[libvirt PATCH v7 00/35] Use nbdkit for http/ftp/ssh network drives in libvirt
This is the seventh version of this patch series. See https://bugzilla.redhat.com/show_bug.cgi?id=2016527 for more information. Note that testing this requires selinux policy changes which are not fully done, but there is a new policy in development that has allowed me to run with selinux in enforcing mode for the common cases. See https://bugzilla.redhat.com/show_bug.cgi?id=2182505 for more information. The following scenarios should work now with selinux enabled using the selinux policy from that bug: - http/https disks - ssh disks with password authentication - ssh disks with passwordless keyfile The one major thing that doesn't work and is difficult to get working with selinux enabled is the ssh-agent. This is because there doesn't seem to be any selinux policy for ssh-agent, so by default the ssh-agent socket is labeled unconfined_t. We cannot allow access from the libvirt/qemu to unconfined_t because that would open up access to just about anything on the host. So additional work will likely be necessary for ssh-agent/libvirt interaction in the future. Fortunately ssh-agent is something that never was really supported with the old qemu block driver either, so I think we could potentially merge this patchset either without the ssh-agent patches or with a note that ssh-agent won't work with selinux enabled. Note also that gitlab CI will not work for this series without changes to the ci definitions due to the addition of libnbd dependency. Changes in v7: - rebased to latest master - moved restarting of nbdkit process to per-domain event thread - a few other smaller changes suggested by Peter in v6 Jonathon Jongsma (35): schema: allow 'ssh' as a protocol for network disks qemu: Add functions for determining nbdkit availability qemu: expand nbdkit capabilities util: Allow virFileCache data to be any GObject qemu: implement basic virFileCache for nbdkit caps qemu: implement persistent file cache for nbdkit caps qemu: use file cache for nbdkit caps qemu: Add qemuNbdkitProcess qemu: query nbdkit module dir from binary qemu: add functions to start and stop nbdkit Generalize qemuDomainLogContextNew() qemu: Extract qemuDomainLogContext into a new file qemu: move qemuProcessReadLog() to qemuLogContext qemu: log error output from nbdkit tests: add ability to test various nbdkit capabilities qemu: split qemuDomainSecretStorageSourcePrepare qemu: include nbdkit state in private xml util: secure erase virCommand send buffers qemu: pass sensitive data to nbdkit via pipe qemu: use nbdkit to serve network disks if available util: make virCommandSetSendBuffer testable tests: add tests for nbdkit invocation qemu: add test for authenticating a https network disk qemu: Monitor nbdkit process for exit qemu: Taint domain if nbdkit restart fails qemu: try to connect to nbdkit early to detect errors schema: add password configuration for ssh disk qemu: implement password auth for ssh disks with nbdkit schema: add configuration for host verification of ssh disks qemu: implement knownHosts for ssh disks with nbdkit schema: add keyfile configuration for ssh disks qemu: implement keyfile auth for ssh disks with nbdkit schema: add ssh-agent configuration for ssh disks qemu: implement ssh-agent auth for ssh disks with nbdkit rpm: update spec file for for nbdkit support build-aux/syntax-check.mk |2 +- docs/formatdomain.rst | 45 +- libvirt.spec.in |8 + meson.build | 14 + meson_options.txt |1 + po/POTFILES |2 + src/conf/domain_conf.c| 38 + src/conf/domain_conf.h|1 + src/conf/schemas/domaincommon.rng | 55 + src/conf/storage_source_conf.c|6 + src/conf/storage_source_conf.h|6 +- src/libvirt_private.syms |1 + src/qemu/meson.build |3 + src/qemu/qemu_block.c | 162 ++- src/qemu/qemu_conf.c | 22 + src/qemu/qemu_conf.h |6 + src/qemu/qemu_domain.c| 436 +++--- src/qemu/qemu_domain.h| 31 +- src/qemu/qemu_driver.c| 21 + src/qemu/qemu_extdevice.c | 62 + src/qemu/qemu_hotplug.c |7 + src/qemu/qemu_logcontext.c| 329 + src/qemu/qemu_logcontext.h| 41 + src/qemu/qemu_nbdkit.c| 1277 + src/qemu/qemu_nbdkit.h| 119 ++ src/qemu/qemu_nbdkitpriv.h| 31 + src/qemu/qemu_process.c | 122 +- src/qemu/qemu_process.h |3 + src/util
Re: [libvirt PATCH v6 35/36] run: add ability to set selinux context
On 8/23/23 2:29 AM, Erik Skultety wrote: On Tue, Aug 22, 2023 at 03:24:03PM +0200, Peter Krempa wrote: On Thu, Jul 20, 2023 at 17:20:02 -0500, Jonathon Jongsma wrote: When running libvirt from the build directory with the 'run' script, it will run as unconfined_t. This can result in unexpected behavior when selinux is enforcing due to the fact that the selinux policies are written assuming that libvirt is running with the system_u:system_r:virtd_t context. This patch adds a new --selinux option to the run script. When this option is specified, it will launch the specified binary using the 'runcon' utility to set its selinux context to the one mentioned above. Since this may require root privileges, setting the selinux context is not the default behavior and must be enabled with the command line switch. Signed-off-by: Jonathon Jongsma --- run.in | 100 + 1 file changed, 80 insertions(+), 20 deletions(-) Please send this one separately again. The idea of doing this is cool, but I don't really fancy to review what's happening in 'run.in' at this point. No need actually - this patch has already been posted as a v2 here [1] and ACKed by Martin Kletzander, but hasn't been pushed yet. [1] https://listman.redhat.com/archives/libvir-list/2023-June/240358.html Regards, Erik Oops, I pushed this one now.
Re: [libvirt PATCH 5/5] qemu: Implement support for vDPA block devices
On 8/8/23 6:00 AM, Stefano Garzarella wrote:
On Mon, Aug 07, 2023 at 03:41:21PM +0200, Peter Krempa wrote:
On Thu, Aug 03, 2023 at 09:48:01 +0200, Stefano Garzarella wrote:
On Wed, Aug 2, 2023 at 10:33 PM Jonathon Jongsma
wrote:
> On 7/24/23 8:05 AM, Peter Krempa wrote:
[...]
> >
> > I've also noticed that using 'qcow2' format for the device
doesn't work:
> >
> > error: internal error: process exited while connecting to
monitor: 2023-07-24T12:54:15.818631Z qemu-system-x86_64: -blockdev
{"node-name":"libvirt-1-format","read-only":false,"cache":{"direct":true,"no-flush":false},"driver":"qcow2","file":"libvirt-1-storage"}: Could not read qcow2 header: Invalid argument
> >
> > If that is supposed to work, then qemu devs will probably need to
know
> > about that, if that is not supposed to work, libvirt needs to add a
> > check, because the error doesn't tell much. It's also possible I've
> > messed up when formatting the image though, as didn't really try to
> > figure out what's happening.
> >
>
>
> That's a good question, and I don't actually know the answer. Were you
> using an actual vdpa block device for your tests or were you using the
> vdpa block simulator kernel module? How did you set it up? Adding
> Stefano to cc for his thoughts.
Yep, I would also like to understand how you initialized the device
with a qcow2 format.
Naively and originally I've simply used it as 'raw' at first and
formatted it from the guest OS. Then I've shut-down the VM and started
it back reconfiguring the image format as qcow2. This normally works
with real-file backed storage, and since the vdpa simulator seems to
persist the contents I supposed this would work.
Cool, I'll try that.
Can you try to reboot the VM, use it as `raw`, and read the qcow2 in the
vm from the guest OS?
Note: there could be some bugs in the simulator!
Theoretically, the best use case for vDPA block is that the backend
handles formats, for QEMU it should just be a virtio device, but being
a blockdev, we should be able to use formats anyway, so it should
work.
Yeah, ideally there will be no format driver in qemu used for these
devices (this is not yet the case, I'll need to fix libvirt to stop
using the 'raw' driver if not needed).
Here I'm more interested whether it is supposed to work, in which case
we want to allow using qcow2 as a format in libvirt, or it's not
supposed to work and we should forbid it before the user gets a
suboptimal error message such as now.
This is a good question. We certainly haven't tested it, because it's an
uncommon scenario, but as I said before, maybe it should work. I need to
check it better.
For now, waiting for real hardware, the only way to test vDPA block
support in QEMU is to use the simulator in the kernel or VDUSE.
With the kernel simulator we only have a 128 MB ramdisk available,
with VDUSE you can use QSD with any file:
$ modprobe -a vhost_vdpa vduse
$ qemu-storage-daemon \
--blockdev
file,filename=/path/to/image.qcow2,cache.direct=on,aio=native,node-name=file
\
--blockdev qcow2,file=file,node-name=qcow2 \
--export
vduse-blk,id=vduse0,name=vduse0,num-queues=1,node-name=qcow2,writable=on
$ vdpa dev add name vduse0 mgmtdev vduse
Then you have a /dev/vhost-vdpa-X device that you can use with the
`virtio-blk-vhost-vdpa` blockdev (note: vduse requires QEMU with a
memory-backed with `share=on`), but using raw since the qcow2 is
handled by QSD.
Of course, we should be able to use raw file with QSD and qcow2 on
qemu (although it's not the optimal configuration), but I don't know
how to initialize a `virtio-blk-vhost-vdpa` blockdev with a qcow2
image :-(
With the above qemu storage daemon you should be able to do that by
simply dropping the qcow2 format driver and simply exposing a qcow2
formatted image. It similarly works with NBD:
I've formatted 2 qcow2 images:
# qemu-img create -f qcow2 /root/image1.qcow2 100M
# qemu-img create -f qcow2 /root/image2.qcow2 100M
And then exported them both via vduse and nbd without interpreting
qcow2, thus making the QSD into just a dumb storage device:
# qemu-storage-daemon \
--blockdev
file,filename=/root/image1.qcow2,cache.direct=on,aio=native,node-name=file1 \
--export
vduse-blk,id=vduse0,name=vduse0,num-queues=1,node-name=file1,writable=on \
--blockdev
file,filename=/root/image2.qcow2,cache.direct=on,aio=native,node-name=file2 \
--nbd-server addr.type=unix,addr.path=/tmp/nbd.sock \
--export nbd,id=nbd0,node-name=file2,writable=on,name=exportname
Cool! Thanks for sharing!
Now when I start a VM using the NBD export in qcow2 format:
function='0x0'/>
The VM starts fine, but when using:
function='0x0'/>
I get:
error: internal error: QEMU u
Re: [libvirt PATCH 5/5] qemu: Implement support for vDPA block devices
On 7/24/23 8:05 AM, Peter Krempa wrote:
As I've promised a long time ago I gave your patches some testing in
regards of cooperation with blockjobs and snapshots.
Since the new version of the patches was not yet posted on the list I'm
replying here including my observations from testing patches from your
gitlab branch:
On Tue, Jun 06, 2023 at 16:15:30 -0500, Jonathon Jongsma wrote:
Requires recent qemu with support for the virtio-blk-vhost-vdpa device
and the ability to pass a /dev/fdset/N path for the vdpa path (8.1.0)
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1900770
Since this is a feature addition the 'Fixes' keyword doesn't make sense.
Use e.g. 'Resolves' instead.
Additionally you're missing the DCO certification here.
---
src/qemu/qemu_block.c | 20 --
src/qemu/qemu_domain.c | 25
src/qemu/qemu_validate.c | 44 +++---
tests/qemuxml2argvdata/disk-vhostvdpa.args | 35 +
tests/qemuxml2argvdata/disk-vhostvdpa.xml | 21 +++
tests/qemuxml2argvtest.c | 2 +
6 files changed, 139 insertions(+), 8 deletions(-)
create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.args
create mode 100644 tests/qemuxml2argvdata/disk-vhostvdpa.xml
[...]
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 2f6b32e394..119e52a7d7 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -11157,6 +11157,28 @@ qemuDomainPrepareStorageSourceFDs(virStorageSource
*src,
}
+static int
+qemuDomainPrepareStorageSourceVDPA(virStorageSource *src,
+ qemuDomainObjPrivate *priv)
+{
+qemuDomainStorageSourcePrivate *srcpriv = NULL;
+virStorageType actualType = virStorageSourceGetActualType(src);
+int vdpafd = -1;
+
+if (actualType != VIR_STORAGE_TYPE_VHOST_VDPA)
+return 0;
+
+if ((vdpafd = qemuVDPAConnect(src->path)) < 0)
+return -1;
This function call directly touches the host filesystem, which is not
supposed to be in the *DomainPrepareStorageSource* functions but we
rather have a completely separate machinery in
qemuProcessPrepareHostStorage.
Unfortunately that one doesn't yet need to handle individual backing
chain members though.
This ensures that the code doesn't get accidentally called from tests
even without mocking the code as the tests reimplement the functions
differently for testing purposes.
+
+srcpriv = qemuDomainStorageSourcePrivateFetch(src);
+
+srcpriv->fdpass = qemuFDPassNew(src->nodestorage, priv);
+qemuFDPassAddFD(srcpriv->fdpass, , "-vdpa");
+return 0;
+}
[...]
diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c
index 9dce908cfe..67b0944162 100644
--- a/src/qemu/qemu_validate.c
+++ b/src/qemu/qemu_validate.c
@@ -3220,6 +3220,28 @@ qemuValidateDomainDeviceDefDiskTransient(const
virDomainDiskDef *disk,
}
+static int
+qemuValidateDomainDeviceDefDiskVhost(const virDomainDef *def,
+ virStorageType storagetype,
+ virQEMUCapsFlags flag,
+ virQEMUCaps *qemuCaps)
+{
+const char *vhosttype = virStorageTypeToString(storagetype);
+
+if (!virQEMUCapsGet(qemuCaps, flag)) {
+virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
+ _("%1$s disk is not supported with this QEMU binary"),
+ vhosttype);
+return -1;
I'd prefer if both things this function does are duplicated inline below
rather than passing it via arguments here. It makes it harder to follow.
+}
+
+if (qemuValidateDomainDefVhostUserRequireSharedMemory(def, vhosttype) < 0)
+return -1;
+
+return 0;
+}
In my testing of the code from the new branch I've observed that
blockjobs and snapshot creation work well thanks to libblkio, so we
don't have to add any additional checks or limitations.
I'll still need to go ahead and finish the series removing the 'raw'
driver when it's not necessary so that the fast-path, once implemented
will be possible. Waiting for that is not necessary for this series as
it works properly even with the 'raw' driver in place.
With your new version of the patches I've noticed the following
problems:
- After converting to store the vdpa device path in src->vdpadev:
- rejecting of the empty disk source doesn't work for vdpa. If you use
in stead of the proper path, the XML will be defined but
broken.
- virStorageSourceCopy doesn't copy the new member (as instructed in
the comment above the struct), thus block job code which uses this
extensively to work on the inactive definition creates broken
configurations.
I've also noticed that using 'qcow2' format for the device doesn't work:
error: internal error: process exited while connecting to monitor: 2023-07-24T12:54:15.8186
