This was pushed as commit bab3081265
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
http://libguestfs.org/virt-builder.1.html
Thanks for everyone's feedback. I pushed this series with several
enhancements as:
5fbf93a92..2f7ca818f
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows
On Tue, May 09, 2023 at 01:36:13PM -0500, Eric Blake wrote:
> On Tue, May 09, 2023 at 03:51:21PM +0100, Richard W.M. Jones wrote:
> > Debug strings contain all kinds of information including some under
> > user control. Previously we simply sent everything to stderr, but
> >
On Tue, May 09, 2023 at 01:25:18PM -0500, Eric Blake wrote:
> On Tue, May 09, 2023 at 03:51:20PM +0100, Richard W.M. Jones wrote:
> > Preserve errno even if the function fails. We need to set
> > errno = err again on every exit path out of the function.
> >
> > Chec
On Tue, May 09, 2023 at 01:18:13PM -0500, Eric Blake wrote:
> On Tue, May 09, 2023 at 03:51:18PM +0100, Richard W.M. Jones wrote:
> > eg. { '\r', '\n' } -> { '\\', 'n', '\\', 'r' }
> > ---
> > common/utils/utils.h | 1 +
> &
Debug strings contain all kinds of information including some under
user control. Previously we simply sent everything to stderr, but
this is potentially insecure, as well as not dealing well with
non-printable characters. Escape these strings when printing.
---
server/debug.c | 38
Preserve errno even if the function fails. We need to set
errno = err again on every exit path out of the function.
Check if close_memstream failed and go to the error path if so. This
is most important on Windows where close_memstream is what actually
allocates the memory. But even on Unix,
eg. { '\r', '\n' } -> { '\\', 'n', '\\', 'r' }
---
common/utils/utils.h | 1 +
common/utils/quote.c | 58
2 files changed, 59 insertions(+)
diff --git a/common/utils/utils.h b/common/utils/utils.h
index 42288a5cd..1e6d8972b 100644
---
In particular return before we have allocated 'str' with
attribute((cleanup)) so that we don't waste time calling free (NULL)
on the hot-ish path out of the function.
---
server/debug.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/server/debug.c b/server/debug.c
index
This is not secure so should not be used routinely. Also we do not
attempt to filter environment variables, so even ones containing
multiple lines or special characters are all sent to nbdkit_debug.
The reason for adding this is to allow for debugging the new
nbd_set_socket_activation_name(3)
At least under Wine they are not recognized and you just see literal
"[0m" etc appearing in the output.
Updates: commit 5a011e02375912f2ad88de1ebc6d609e29d38f71
Reviewed-by: Laszlo Ersek
---
server/debug.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/server/debug.c b/server/debug.c
Add R-b tags.
Split the last patch into 3 parts, fixing the error handling issues
raised by Laszlo in the previous review.
Rich.
___
Libguestfs mailing list
Libguestfs@redhat.com
https://listman.redhat.com/mailman/listinfo/libguestfs
On Tue, May 09, 2023 at 03:05:41PM +0200, Laszlo Ersek wrote:
[...]
> Then, in the post-patch version, the following catches my eye:
>
> fp_inner = open_memstream (_inner, _inner);
> if (fp_inner == NULL) {
> fail:
> /* Try to emit what we can. */
> errno = err;
> vfprintf
On Tue, May 09, 2023 at 02:28:40PM +0200, Laszlo Ersek wrote:
> On 5/9/23 11:51, Richard W.M. Jones wrote:
> > eg. { '\r', '\n' } -> { '\\', 'n', '\\', 'r' }
> > ---
> > common/utils/utils.h | 1 +
> > common/utils/quote.c | 58 +++
Debug strings contain all kinds of information including some under
user control. Previously we simply sent everything to stderr, but
this is potentially insecure, as well as not dealing well with
non-printable characters. Escape these strings when printing.
---
server/debug.c | 52
eg. { '\r', '\n' } -> { '\\', 'n', '\\', 'r' }
---
common/utils/utils.h | 1 +
common/utils/quote.c | 58
2 files changed, 59 insertions(+)
diff --git a/common/utils/utils.h b/common/utils/utils.h
index 42288a5cd..1e6d8972b 100644
---
This grew a bit from last time ...
The basic change to allow dumping the environment is the same, except
I enabled this for Windows (and checked that it does in fact work --
Windows has environment variables, who knew?)
The second patch fixes a problem I noticed while testing the above on
This is not secure so should not be used routinely. Also we do not
attempt to filter environment variables, so even ones containing
multiple lines or special characters are all sent to nbdkit_debug.
The reason for adding this is to allow for debugging the new
nbd_set_socket_activation_name(3)
At least under Wine they are not recognized and you just see literal
"[0m" etc appearing in the output.
Updates: commit 5a011e02375912f2ad88de1ebc6d609e29d38f71
---
server/debug.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/server/debug.c b/server/debug.c
index 1b5ddaafe..2231d40a0
Also test that the expected environment variable is set when
connecting to nbdkit.
This test requires nbdkit >= 1.35.2 which added support for -D
nbdkit.environ=1 to dump the environment, and will be skipped with
other versions.
---
.gitignore | 1 +
tests/Makefile.am
I added a test of the LISTEN_FDNAMES=unknown case as well, otherwise
only minor cosmetic changes from last time.
Rich.
___
Libguestfs mailing list
Libguestfs@redhat.com
https://listman.redhat.com/mailman/listinfo/libguestfs
On Tue, May 09, 2023 at 08:56:47AM +0200, Laszlo Ersek wrote:
> On 5/8/23 23:33, Richard W.M. Jones wrote:
> > On Mon, May 08, 2023 at 08:54:48AM +0200, Laszlo Ersek wrote:
> >> On 5/7/23 12:43, Richard W.M. Jones wrote:
> >>> This is not secure so should not
On Mon, May 08, 2023 at 08:54:48AM +0200, Laszlo Ersek wrote:
> On 5/7/23 12:43, Richard W.M. Jones wrote:
> > This is not secure so should not be used routinely. Also we do not
> > attempt to filter environment variables, so even ones containing
> > multiple lines or speci
On Mon, May 08, 2023 at 08:35:53AM -0500, Eric Blake wrote:
> On Sun, May 07, 2023 at 11:43:17AM +0100, Richard W.M. Jones wrote:
> > This is not secure so should not be used routinely. Also we do not
> > attempt to filter environment variables, so even ones containing
>
Also test that the expected environment variable is set when
connecting to nbdkit.
This test requires nbdkit >= 1.35.2 which added support for -D
nbdkit.environ=1 to dump the environment, and will be skipped with
other versions.
---
.gitignore | 1 +
tests/Makefile.am
This is not secure so should not be used routinely. Also we do not
attempt to filter environment variables, so even ones containing
multiple lines or special characters are all sent to nbdkit_debug.
The reason for adding this is to allow for debugging the new
nbd_set_socket_activation_name(3)
For the series:
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows
Thanks - upstream in commit 692802bf96
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers.
On Sat, Apr 29, 2023 at 05:24:39PM +0200, Zixun LI wrote:
> Add support for lzma, zstd and auto detection of compression methods.
>
> Signed-off-by: Zixun LI
> ---
> daemon/tar.c | 12
> 1 file changed, 12 insertions(+)
>
> diff --git a/daemon/tar.c b/daemon/tar.c
> index
On Fri, Apr 28, 2023 at 03:32:57PM +0200, Philippe Midol-Monnet wrote:
> Hi
>
> I have already fulfill a an issue, but it looks like it's not the recommended
> process.
>
>
> We are using a LUKS partition containing a GPT table and several partions.
>
> With libguestfs we can open the LUKS
On Thu, Apr 27, 2023 at 10:35:30AM -0500, Eric Blake wrote:
> Reviewed-by: Eric Blake
Thanks, I pushed the two patches as e397f0643..9b5b93a07
About using fcntl locks: it seems troublesome because we'd then need
to have an actual file to lock (even one in /dev/shm). That makes the
All looks good, thanks:
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-top is 'top' for virtual machines. Tiny program with many
powerful
Thanks, patches upstream in 7674d603e9..d0d8e67384.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
Add a contrib directory which we can use to store random tests and
other contributions that we don't necessarily want to support long
term or incorporate into nbdkit.
Into this directory place a test program which tests
nbdkit-memory-plugin with allocator=sparse.
---
configure.ac |
We've long known that nbdkit-memory-plugin with the default sparse
array allocator suffers because a global lock must be taken whenever
any read or write operation is performed. This commit aims to safely
improve performance by converting the lock into a read-write lock.
The shared (read) lock
nbdkit-memory-plugin is a RAM disk, but it is quite a slow one because
of the lock we must acquire on the sparse array. Try to improve
performance by changing this mutex for a read-write lock.
The results are somewhat mixed.
Note that in the real world, workloads which constantly update the
On Wed, Apr 26, 2023 at 04:37:21PM +0200, Laszlo Ersek wrote:
> On 4/26/23 14:59, Andrey Drobyshev wrote:
> > 'X' in the setiles' stderr doesn't necessarily mean that option 'X'
> > doesn't exist. For instance, when passing '-T' we get: "setfiles:
> > option requires an argument -- 'T'".
> >
> >
Series looks good, so also:
Reviewed-by: Richard W.M. Jones
I'll get around to this in a bit, let's just wait a few hours in case
there are further comments.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog
On Mon, Apr 24, 2023 at 08:56:45AM +0200, Ingo Weinmann wrote:
> Using virt-customize of libguestfs, I can copy a file from the host to the
> root
> partition of an image like this:
>
> virt-customize -a 2023-02-21-raspios-bullseye-arm64-lite.img --copy-in
> cmdline.txt:/boot/
>
> But
On Mon, Apr 24, 2023 at 12:47:37PM +0200, Denis V. Lunev wrote:
> On 4/24/23 10:46, Richard W.M. Jones wrote:
> >On Fri, Apr 21, 2023 at 09:01:40PM +0300, Andrey Drobyshev wrote:
> >>This patch effectively limits the number of cases when we would want to
> >>do a
On Mon, Apr 24, 2023 at 12:08:57PM +0200, Laszlo Ersek wrote:
> On 4/24/23 10:46, Richard W.M. Jones wrote:
> > On Fri, Apr 21, 2023 at 09:01:40PM +0300, Andrey Drobyshev wrote:
> >> This patch effectively limits the number of cases when we would want to
> >> do a
On Fri, Apr 21, 2023 at 09:01:40PM +0300, Andrey Drobyshev wrote:
> This patch effectively limits the number of cases when we would want to
> do a complete SELinux relabeling on Linux guest conversion.
>
> This was brought to my attention as we've recently had a support case
> when the conversion
Upstream in libguestfs-common: 70c10a079..38e6988c1
and in virt-v2v: 6db3b26d2f..e83de8abe6
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from
On Fri, Apr 21, 2023 at 09:01:41PM +0300, Andrey Drobyshev wrote:
> From: Roman Kagan
>
> Even if SELinux config file and policy tools are present, SELinux may be
> turned off by a setting in that config file, "SELINUX", having a value
> of "disabled".
>
> Detect this case and skip relabeling.
On Fri, Apr 21, 2023 at 06:30:52AM -0700, alan somers wrote:
> On Fri, Apr 21, 2023 at 2:32 AM Laszlo Ersek wrote:
> >
> > On 4/21/23 11:20, Richard W.M. Jones wrote:
> > > See: https://github.com/libguestfs/nbdkit/issues/21
> > >
> > > Tested by applying
Patches 3 & 4 look good.
For the series:
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-top is 'top' for virtual machines. Tiny pro
On Fri, Apr 21, 2023 at 10:20:35AM +0100, Richard W.M. Jones wrote:
> impl Server for RamDisk {
> +fn after_fork() -> Result<()> {
> +// A place to start background threads.
> +eprintln!("forked");
Alan, this reminds me that there is no bin
On Fri, Apr 21, 2023 at 11:32:31AM +0200, Laszlo Ersek wrote:
> On 4/21/23 11:20, Richard W.M. Jones wrote:
> > See: https://github.com/libguestfs/nbdkit/issues/21
> >
> > Tested by applying the following patch to the example plugin and
> > running it with verbose enab
See: https://github.com/libguestfs/nbdkit/issues/21
Tested by applying the following patch to the example plugin and
running it with verbose enabled:
--- a/plugins/rust/examples/ramdisk.rs
+++ b/plugins/rust/examples/ramdisk.rs
@@ -43,6 +43,12 @@ struct RamDisk {
}
impl Server for RamDisk {
+
t.orderfile
> @@ -38,6 +38,7 @@ common/*.h
> common/*/*.h
>
> # Generator files.
> +generator/*.mli
> generator/*
>
> # Source files.
>
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read m
4
> --- a/scripts/git.orderfile
> +++ b/scripts/git.orderfile
> @@ -38,7 +38,6 @@ common/*.h
> common/*/*.h
>
> # Generator files.
> -generator/generator
> generator/*
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people
On Thu, Apr 20, 2023 at 04:04:58PM +0200, Laszlo Ersek wrote:
> Change the optional "parens" parameter of "print_arg_list" from bool to a
> new union type. The new union type has three data constructors, NoParens
> (matching the previous "false" value), ParensSameLine (matching the
> previous
On Wed, Apr 19, 2023 at 03:53:57PM +0200, Laszlo Ersek wrote:
> On 4/19/23 15:37, Eric Blake wrote:
> > On Tue, Apr 18, 2023 at 07:26:29PM +0200, Laszlo Ersek wrote:
> >> Embedding a shell script in a multi-line C string literal is an exercise
> >> in pain. I can see why the original author (whom
On Tue, Apr 18, 2023 at 07:26:21PM +0200, Laszlo Ersek wrote:
> (The changes in this patch are simple, but likely more controversial than
> the rest.)
>
> The following four components don't play nice together:
>
> - Needlessly spelling out "extern" for function declarations in header
> files.
Looks fine. I agree with Eric about "newstyle fixed", that ought not
to be split. For the series:
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wor
On Tue, Apr 18, 2023 at 02:33:41PM +0200, Wouter Verhelst wrote:
> On Thu, Apr 13, 2023 at 05:02:41PM -0500, Eric Blake wrote:
> > Rather than requiring all servers and clients to have a 32-bit limit
> > on maximum NBD_CMD_READ/WRITE sizes, we can choose to standardize
> > support for a 64-bit
On Tue, Apr 18, 2023 at 11:09:48AM +0100, Richard W.M. Jones wrote:
> On Tue, Apr 18, 2023 at 12:00:22PM +0200, Laszlo Ersek wrote:
> > On 4/18/23 09:25, Richard W.M. Jones wrote:
> > > Commit 6b03ec8ad6 ("common: Move exit-with-parent code from include/
> &g
On Tue, Apr 18, 2023 at 12:00:22PM +0200, Laszlo Ersek wrote:
> On 4/18/23 09:25, Richard W.M. Jones wrote:
> > Commit 6b03ec8ad6 ("common: Move exit-with-parent code from include/
> > to utils/") was supposed to be a largely neutral refactoring of the
> > --exit-wi
Commit 6b03ec8ad6 ("common: Move exit-with-parent code from include/
to utils/") was supposed to be a largely neutral refactoring of the
--exit-with-parent flag. However I made a mistake in one hunk of that
patch:
-#ifdef HAVE_EXIT_WITH_PARENT
- if (exit_with_parent) {
-if
On Mon, Apr 17, 2023 at 07:44:31PM +0300, Nir Soffer wrote:
> When building from git we need autoconf, automake and libtool.
>
> Signed-off-by: Nir Soffer
> ---
>
> Changes sinve v1:
> - Remove `,` between package namses (Laszlo)
>
> README.md | 7 +++
> 1 file changed, 7 insertions(+)
>
For this series (which I much prefer over v1):
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
nbdkit - Flexible, fast NBD server with plugins
https
So I read through the whole series this morning and it all seems
reasonable. Therefore:
Reviewed-by: Richard W.M. Jones
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder
On Fri, Apr 14, 2023 at 10:21:31AM +0200, Laszlo Ersek wrote:
> In the strictest sense, I only need to shorten
> DESTINATION_IS_ZERO_OPTION by three characters. Should I rename it to
> DEST_IS_ZERO_OPTION, or TARGET_IS_ZERO_OPTION? (The latter is better,
> because we already have a
On Fri, Apr 14, 2023 at 09:59:56AM +0200, Laszlo Ersek wrote:
> Wrap "copy/nbd-ops.c" at 80 characters.
>
> I couldn't find a way to test that this change is unobservable.
>
> Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2172516
> Signed-off-by: Laszlo Ersek
> ---
> copy/nbd-ops.c | 5
On Fri, Apr 14, 2023 at 09:59:54AM +0200, Laszlo Ersek wrote:
> Rename _OPTION to OPT_ for two reasons:
>
> - it is more idiomatic for enum constants to have the same prefix than for
> them to have the same suffix,
>
> - we hew away three characters (length-wise) from each option name, which
>
On Fri, Apr 14, 2023 at 09:59:53AM +0200, Laszlo Ersek wrote:
> Two of the enum constants that denote command line options are
> inconsistently named with the rest: all identifiers should be
> _OPTION, but LONG_OPTIONS and SHORT_OPTIONS (which are supposed
> to list the long and short options)
On Thu, Apr 13, 2023 at 06:05:56PM +0200, Laszlo Ersek wrote:
> On 4/13/23 16:32, Richard W.M. Jones wrote:
> > See: https://listman.redhat.com/archives/libguestfs/2023-April/031243.html
> > Updates: commit c6d6113fa1f38389dbc27f000d06c425cbe912cd
> > ---
> > tes
See: https://listman.redhat.com/archives/libguestfs/2023-April/031243.html
Updates: commit c6d6113fa1f38389dbc27f000d06c425cbe912cd
---
tests/Makefile.am | 10 +--
tests/test-floppy-size.sh | 61 +++
2 files changed, 69 insertions(+), 2 deletions(-)
This series looks fine to me. 'make check' still passes? It contains
some --help validation tests.
I was preparing to do a 1.34 (new stable) release, so it's best to get
this in sooner rather than later, otherwise it will miss the release.
I will look at adding a test for
On Thu, Apr 13, 2023 at 04:03:40PM +0200, Laszlo Ersek wrote:
> Note that this patch is untested. In "make check", "tests/test-floppy.sh"
> passes. However, that test does not seem to pass the "size=..." option to
> the plugin, and the code being modified in create_virtual_floppy() is
> related to
On Fri, Mar 17, 2023 at 02:03:55PM +0100, Laszlo Ersek wrote:
> On 3/17/23 13:46, Andrey Drobyshev wrote:
> > Am I missing
> > something here? And having all that said: wouldn't it be beneficial and
> > more robust to explicitly set libguestfs backend when running the test
> > suite?
>
> Yes, I
On Thu, Mar 16, 2023 at 06:14:05PM +0200, Andrey Drobyshev wrote:
> Thanks for the detailed overview. That is actually the answer to my
> original question: I was looking for a sub-target which would check the
> docs, and failed to see that instead there's a separate test for that
> purpose. And
On Wed, Apr 05, 2023 at 04:29:19PM +, Chintan Patel wrote:
> Hi Libguestfs team,
>
> I'm trying to build guestfs on Ubuntu 22.04, but getting below error.
> #17 8.021 checking for OCaml findlib package augeas... not found
> #17 8.024 configure: error: the OCaml module 'augeas' is required
>
For the series, since this is a straight backport of the upstream
patch with the filenames changed:
Reviewed-by: Richard W.M. Jones
> https://bugzilla.redhat.com/show_bug.cgi?id=2175703
Looks like we're going to do this as a 0-day z-stream backport?
You are assignee so I'll leave it al
On Mon, Mar 20, 2023 at 12:53:00PM +0100, Laszlo Ersek wrote:
> In "kernel-5.14.0-269.el9", the "kernel-modules-core" subpackage got split
> from the "kernel-core" subpackage. Therefore, a single binary RPM
> containing *both* the "/boot/vmlinuz-5.14.0-269.el9.x86_64" file *and* the
>
13:19 < tgolembi[m]> is there a trick how to convert a VM with name that starts
with a dash? It's being confused for virt-v2v argument.
:)
Yes you can do this:
$ virt-v2v -o [etc...] -- vmname
Note all -options need to come before the "--".
For example:
$ virt-builder
On Mon, Apr 03, 2023 at 12:00:38PM +0200, Laszlo Ersek wrote:
> On 4/3/23 11:16, Richard W.M. Jones wrote:
> > SFDC case 03473932, RHEL 8 only.
> >
> > See discussion in private email thread "Customer case requiring our
> > assistance" in 2023.
> >
On Mon, Apr 03, 2023 at 11:45:33AM +0200, Laszlo Ersek wrote:
> On 4/3/23 11:16, Richard W.M. Jones wrote:
> > SFDC case 03473932, RHEL 8 only.
> >
> > See discussion in private email thread "Customer case requiring our
> > assistance" in 2023.
> >
SFDC case 03473932, RHEL 8 only.
See discussion in private email thread "Customer case requiring our
assistance" in 2023.
---
mlcustomize/SELinux_relabel.ml | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/mlcustomize/SELinux_relabel.ml b/mlcustomize/SELinux_relabel.ml
index
On Tue, Mar 14, 2023 at 04:06:18PM +0200, Andrey Drobyshev wrote:
> Speaking of "make check": could you point out, for future reference,
> which particular sub-target you're referring to here? I can see these:
> check-am, check-recursive, check-slow, check-TESTS, check-valgrind. And
> none of
For the full v3 series (common/ and v2v/):
Reviewed-by: Richard W.M. Jones
I'll give these a little while for others to comment before picking
them upstream.
Thanks,
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming
On Thu, Mar 09, 2023 at 04:38:44PM +0200, Andrey Drobyshev wrote:
> On 3/8/23 22:45, Richard W.M. Jones wrote:
> > On Tue, Mar 07, 2023 at 09:40:26PM +0200, Andrey Drobyshev wrote:
> >> In case when we are injecting virtio-scsi device driver into the guest
> >> (rather
On Thu, Mar 09, 2023 at 05:03:02PM +0200, Andrey Drobyshev wrote:
> On 3/9/23 16:50, Andrey Drobyshev wrote:
> > On 3/8/23 22:46, Richard W.M. Jones wrote:
> >>
> >> The patch series looks generally fine, but it really needs a test ...
> >>
> >>
) =
> +let fb_script = "@echo off\n\
> + \n\
> + echo Wait for VirtIO drivers to be installed\n\
> + %systemroot%\\Sysnative\\PnPutil -i -a \
> + %systemroot%\\Drivers\\Virtio\\*.inf \
> +" in
> + F
On Thu, Mar 09, 2023 at 03:47:10PM +0200, Andrey Drobyshev wrote:
> On 3/8/23 22:52, Richard W.M. Jones wrote:
> > On Wed, Mar 08, 2023 at 08:05:35PM +0200, Andrey Drobyshev wrote:
> >> During conversion we copy the necessary drivers to the directory
> >> "%sys
On Thu, Mar 09, 2023 at 11:11:50AM +, Daniel P. Berrangé wrote:
> On Thu, Mar 09, 2023 at 09:50:00AM +0000, Richard W.M. Jones wrote:
> > When a highly multi-threaded program such as nbdcopy encounters an
> > error, there is a race condition in the library which can cause a
On Thu, Mar 09, 2023 at 09:13:24AM +, Daniel P. Berrangé wrote:
> On Thu, Mar 09, 2023 at 08:44:51AM +0000, Richard W.M. Jones wrote:
> > When a highly multi-threaded program such as nbdcopy encounters an
> > error, there is a race condition in the library which can cause a
When a highly multi-threaded program such as nbdcopy encounters an
error, there is a race condition in the library which can cause an
assertion failure and thus a core dump:
(1) An error occurs on one of the threads. nbdcopy calls exit(3).
(2) In lib/errors.c, the destructor calls
This causes a different problem. We actually hit the problem in this test:
https://gitlab.com/nbdkit/libnbd/-/blob/master/tests/dlopen.c
(1) The thread calls dlclose(), unmapping the library but leaving the
pthread_key_t in place, which contains a pointer to free_errors_key in
the unmapped
This version simply removes the call to pthread_key_destroy. It fixes
the problem and allows us to leave the asserts alone so we can still
catch real errors.
Of course this leaks pthread_key_t in the case where you dlclose() the
library. glibc has a limit of 128 thread-specific keys (and the
When a highly multi-threaded program such as nbdcopy encounters an
error, there is a race condition in the library which can cause an
assertion failure and thus a core dump:
(1) An error occurs on one of the threads. nbdcopy calls exit(3).
(2) In lib/errors.c, the destructor calls
n
> nbdkit 1.32.5)
> - various fuse/* (not sure if we can address that; cleaning up FUSE is
> tricky)
>
> This reverts a small part of commit 831cbf7ee14c ("generator: Allow
> DEAD state actions to run").
>
> Thanks: Richard W.M. Jones
> Signed-off-by: Eric
On Wed, Mar 08, 2023 at 08:29:41PM +, Richard W.M. Jones wrote:
> diff --git a/generator/states.c b/generator/states.c
> index fa0f8d716e..c0cf5a7f26 100644
> --- a/generator/states.c
> +++ b/generator/states.c
> @@ -191,8 +191,6 @@ STATE_MACHINE {
>retu
On Wed, Mar 08, 2023 at 08:05:35PM +0200, Andrey Drobyshev wrote:
> During conversion we copy the necessary drivers to the directory
> "%systemroot%\Drivers\Virtio", adding it to the DevicePath registry
> value. As documented in [1], this should be enough for Windows to find
> device drivers and
The patch series looks generally fine, but it really needs a test ...
We find that having tests prevents unnoticed regressions from
happening later.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog:
On Tue, Mar 07, 2023 at 09:40:26PM +0200, Andrey Drobyshev wrote:
> In case when we are injecting virtio-scsi device driver into the guest
> (rather than the default virtio-blk), make sure we write the right PCI ID
> value into the Windows guest registry. This is essential for the guest
> to be
On Tue, Mar 07, 2023 at 09:40:07PM +0200, Andrey Drobyshev wrote:
> Signed-off-by: Andrey Drobyshev
> ---
> convert/convert_windows.ml | 1 +
> 1 files changed, 1 insertions(+)
>
> diff --git a/convert/convert_windows.ml b/convert/convert_windows.ml
> index 9d8d271d..4f672487 100644
> ---
When a highly multi-threaded program such as nbdcopy encounters an
error, there is a race condition in the library which can cause an
assertion failure and thus a core dump:
(1) An error occurs on one of the threads. nbdcopy calls exit(3).
(2) In lib/errors.c, the destructor calls
On Wed, Mar 08, 2023 at 04:00:10PM +0200, Andrey Drobyshev wrote:
> Sure, thanks for pointing out. Will do as soon as Richard leaves his
> notes on the series.
I'm giving a talk this afternoon, but I should be able to look at it
later.
Rich.
--
Richard Jones, Virtualization Group, Red Hat
On Tue, Mar 07, 2023 at 11:26:37AM +0100, Laszlo Ersek wrote:
> On 3/7/23 09:35, Richard W.M. Jones wrote:
> > On Tue, Mar 07, 2023 at 08:48:56AM +0100, Laszlo Ersek wrote:
> >> On 3/6/23 17:52, Richard W.M. Jones wrote:
> >>
> >>> diff --git
401 - 500 of 10948 matches
Mail list logo