On Fri, Jul 20, Peter Maydell wrote:
I actually wrote a patch sitting on top of this one by Stefan
which moves the flag to QEMU_CFLAGS;
both are in the 11-patch series dealing with various configure
Werror issues:
http://patchwork.ozlabs.org/patch/171689/
The actual CFLAGS issue had to be
On Thu, Mar 15, Stefano Stabellini wrote:
On Thu, 15 Mar 2012, Julien Grall wrote:
When an IOREQ_TYPE_INVALIDATE is sent to QEMU, it invalidates all entry
of the map cache even if it's locked.
QEMU is not able to know that entry was invalidated, so when an IO
access is requested a
to
qemu.git since it has the same issue. Please apply to both trees.
Signed-off-by: Olaf Hering o...@aepfle.de
---
configure |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: qemu-xen-dir-remote/configure
===
--- qemu-xen-dir
On Sat, Mar 31, Andreas Färber wrote:
This is the only usage of += outside Makefile fragments, so I wonder if
its use may have been by accident. Is it safe in a POSIX context?
Or should we better use CFLAGS=$CFLAGS -march=486?
Now that I look at the shebang, configure is a sh script and += is
On Sat, Mar 31, Peter Maydell wrote:
On 31 March 2012 16:41, Andreas Färber afaer...@suse.de wrote:
This is the only usage of += outside Makefile fragments, so I wonder if
its use may have been by accident. Is it safe in a POSIX context?
Or should we better use CFLAGS=$CFLAGS -march=486?
.
This patch is against the qemu-xen tree, but it should apply also to
qemu.git since it has the same issue. Please apply to both trees.
Signed-off-by: Olaf Hering o...@aepfle.de
---
configure |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: qemu-xen-dir-remote/configure
Using __DATE__ or __TIME__ in binary pkgs changes the checksum of
compiled binaries if they get rebuilt, even if there are no other
source changes. Replace the dynamic strings with some equally
informative static strings.
Signed-off-by: Olaf Hering o...@aepfle.de
---
Or remove both lines
On Thu, Feb 20, Stefano Stabellini wrote:
On Thu, 6 Feb 2014, Olaf Hering wrote:
@@ -253,6 +254,8 @@ static int ioreq_parse(struct ioreq *ioreq)
case BLKIF_OP_WRITE:
ioreq-prot = PROT_READ; /* from memory */
break;
+case BLKIF_OP_DISCARD:
+return 0
On Thu, Feb 20, Stefano Stabellini wrote:
On Thu, 6 Feb 2014, Olaf Hering wrote:
@@ -253,6 +254,8 @@ static int ioreq_parse(struct ioreq *ioreq)
case BLKIF_OP_WRITE:
ioreq-prot = PROT_READ; /* from memory */
break;
+case BLKIF_OP_DISCARD:
+return 0
On Sun, Feb 23, Stefano Stabellini wrote:
Given that 4.2 is the first Xen release to have BLKIF_OP_DISCARD, maybe
we could just #define BLKIF_OP_DISCARD 5 in that case?
Do we actually need to #ifndef BLKIF_OP_DISCARD?
I will do just the #define and resend with that change.
Olaf
was
intentionally created non-sparse to avoid fragmentation.
Signed-off-by: Olaf Hering o...@aepfle.de
---
v3:
define BLKIF_OP_DISCARD for xen-4.1
v2:
depends on xen_disk: fix io accounting
remove call to bdrv_acct_start
hw/block/xen_blkif.h| 12
hw/block/xen_disk.c | 33
was intentionally created
non-sparse to avoid fragmentation.
Signed-off-by: Olaf Hering o...@aepfle.de
---
The counter part for libxl is here, will go into xen-4.5:
http://lists.xenproject.org/archives/html/xen-devel/2014-01/msg02632.html
checkpatch comaplains about tabs in hw/block/xen_blkif.h
On Mon, Feb 03, Kevin Wolf wrote:
Am 30.01.2014 um 16:02 hat Olaf Hering geschrieben:
+case BLKIF_OP_DISCARD:
+{
+struct blkif_request_discard *discard_req = (void *)ioreq-req;
+bdrv_acct_start(blkdev-bs, ioreq-acct,
+discard_req
On Tue, Feb 04, Olaf Hering wrote:
On Tue, Feb 04, Kevin Wolf wrote:
Now you call bdrv_acct_done() in the callback without having a matching
bdrv_acct_start(). You need to make it conditional in the callback.
Stefano,
Is ioreq_runio_qemu_aio symetric in this regard anyway? In case
On Mon, Feb 03, Kevin Wolf wrote:
Am 30.01.2014 um 16:02 hat Olaf Hering geschrieben:
+++ b/hw/block/xen_disk.c
+case BLKIF_OP_DISCARD:
+{
+struct blkif_request_discard *discard_req = (void *)ioreq-req;
+bdrv_acct_start(blkdev-bs, ioreq-acct
On Tue, Feb 04, Kevin Wolf wrote:
Now you call bdrv_acct_done() in the callback without having a matching
bdrv_acct_start(). You need to make it conditional in the callback.
I see.
Stefano,
Is ioreq_runio_qemu_aio symetric in this regard anyway? In case of
bdrv_acct_done was called unconditional. But in case the ioreq has no
segments there is no matching bdrv_acct_start call. This could lead to
bogus accounting values.
Found by code inspection.
Signed-off-by: Olaf Hering o...@aepfle.de
---
hw/block/xen_disk.c | 13 -
1 file changed
was
intentionally created non-sparse to avoid fragmentation.
Signed-off-by: Olaf Hering o...@aepfle.de
---
v2:
depends on xen_disk: fix io accounting
remove call to bdrv_acct_start
hw/block/xen_blkif.h | 12
hw/block/xen_disk.c | 33 +
2 files changed, 45
bdrv_acct_done was called unconditional. But in case the ioreq has no
segments there is no matching bdrv_acct_start call. This could lead to
bogus accounting values.
Found by code inspection.
Signed-off-by: Olaf Hering o...@aepfle.de
---
v2:
add comment /* fall through */
hw/block/xen_disk.c
On Fri, Feb 07, Stefano Stabellini wrote:
added to my queue, I'll fix the tab manually
Thanks, sorry for the TAB.
Olaf
Right now qemu does appearently just emulate a graphics card (Cirrus for
example), but not an attached monitor/display. The result is confusion
in the guest OS.
The current DRM drivers in upstream Linux seem to default to VESA modes.
These are either provided by qemu, or Xorg uses a fixed
the symbols. The requirement to pass
also --disable-strip looks odd.
Signed-off-by: Olaf Hering o...@aepfle.de
Cc: Paolo Bonzini pbonz...@redhat.com
Cc: Peter Maydell peter.mayd...@linaro.org
Cc: Michael Tokarev m...@tls.msk.ru
Cc: Stefan Hajnoczi stefa...@redhat.com
Cc: Stefan Weil s...@weilnetz.de
On Thu, Sep 25, Peter Maydell wrote:
On 25 September 2014 09:06, Olaf Hering o...@aepfle.de wrote:
During code review for xen I noticed that --enable-debug-info would
still strip the binaries because strip_opt= defaults to yes.
If --enable-debug-info is passed to configure it has
On Thu, Sep 25, Stefan Hajnoczi wrote:
If you make this change and packagers are unaware, they could
accidentally ship qemu packages that contain the full debug symbols in
the binaries.
rpm will take care of that all by itself.
Olaf
On Fri, Nov 28, Olaf Hering wrote:
I wonder if the missing disable of the tap device is intentional, or
just an oversight, or if its just to complicated to get from a
PCIDevice * to the other end and call the -cleanup function.
qemu-traditional did just close all tap devices. With qemu
vte_terminal_set_encoding takes 3 args since 0.38.0.
This fixes commit fba958c6 ("gtk: implement set_echo")
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
ui/gtk.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/ui/gtk.c b/ui/gtk.c
index 01b8216..58d20ee
On Mon, May 23, Gerd Hoffmann wrote:
> +++ b/hw/usb/Makefile.objs
> +common-obj-$(CONFIG_XEN_BACKEND) += xen-usb.o
> +++ b/hw/usb/xen-usb.c
> +usb_bus_new(>bus, sizeof(usbif->bus), _usb_bus_ops,
> xen_sysdev);
xen_sysdev is in an i386-only file, as a result qemu fails to link.
Olaf
Compiling qemu-xen at 2ce1d30 ("xenfb.c: avoid expensive loops when prod
<= out_cons") leads to this error with -O1:
xen.git/tools/qemu-xen-dir/hw/i386/intel_iommu.c: In function
‘vtd_context_device_invalidate’:
xen.git/tools/qemu-xen-dir/hw/i386/intel_iommu.c:911:46: error: ‘mask’ may be
used
On Wed, Mar 09, Kevin Wolf wrote:
> Removing integer overflow checks without removing the potentially
> overflowing operation doesn't feel like a particularly good idea,
> though.
Why does the code use signed ints anyway for sectors and offset?!
Olaf
On Wed, Mar 09, Olaf Hering wrote:
> Why does the code use signed ints anyway for sectors and offset?!
I have to check mainline (next week), at least this fixes mkfs for me:
+++ xen-4.4.3-testing/tools/qemu-xen-dir-remote/block/raw-posix.c
@@ -792,8 +792,8 @@ static BlockDriverAI
What is the purpose of the bdrv_check_request() call in bdrv_co_discard?
It seems a frontend cant possibly know what the limit is in the
qemu-of-the-day, I found no interface to propagate
BDRV_REQUEST_MAX_SECTORS into the guest.
I think to handle nb_sectors > BDRV_REQUEST_MAX_SECTORS
On Wed, Mar 09, Olaf Hering wrote:
> On Wed, Mar 09, Kevin Wolf wrote:
>
> > Removing integer overflow checks without removing the potentially
> > overflowing operation doesn't feel like a particularly good idea,
> > though.
>
> Why does the code use signed ints
On Fri, Apr 01, Peter Maydell wrote:
> On 1 April 2016 at 13:00, Olaf Hering <o...@aepfle.de> wrote:
> > Building on powerpc-linux fails with undefined reference to __atomic_load_8
> > in
> > icount_warp_rt(). Force linking to -latomic.
>
> We should instead b
On Fri, Apr 01, Max Reitz wrote:
> In any case, do you have a test case where a guest was able to submit a
> request that led to the overflow error you described in the commit message?
mkfs -t ext4 /dev/sdb1 in a xen guest with qcow2 as backing device.
When I added discard support to libxl I
Large discard requests lead to sign expansion errors in qemu.
Since there is no API to tell a guest about the limitations qmeu
has to split a large request itself.
Signed-off-by: Olaf Hering <o...@aepfle.de>
Cc: Stefan Hajnoczi <stefa...@redhat.com>
Cc: Kevin Wolf <kw...@redhat.
Building on powerpc-linux fails with undefined reference to __atomic_load_8 in
icount_warp_rt(). Force linking to -latomic.
Fixes a0aa44b ("include/qemu/atomic.h: default to __atomic functions")
Signed-off-by: Olaf Hering <o...@aepfle.de>
Cc: Paolo Bonzini <pbonz...@redhat
Building on powerpc-linux fails with undefined reference to
__atomic_load_8 in icount_warp_rt(). Force linking to -latomic.
Fixes a0aa44b ("include/qemu/atomic.h: default to __atomic functions")
Signed-off-by: Olaf Hering <o...@aepfle.de>
Cc: Paolo Bonzini <pbonz...@redhat
Implement SUSE specific unplug protocol for emulated PCI devices
in PVonHVM guests. Its a simple 'outl(1, (ioaddr + 4));'.
This protocol was implemented and used since Xen 3.0.4.
It is used in all SUSE/SLES/openSUSE releases up to SLES11SP3 and
openSUSE 12.3.
Signed-off-by: Olaf Hering &l
-off-by: Olaf Hering <o...@aepfle.de>
---
hw/i386/xen/xen_platform.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/hw/i386/xen/xen_platform.c b/hw/i386/xen/xen_platform.c
index aa78393..d94b53c 100644
--- a/hw/i386/xen/xen_platform.c
+++ b/hw/i386/xen/xen_platform.c
@@ -115,6 +
Update unplug in Xen HVM guests to cover more cases.
Please review.
Olaf
Olaf Hering (2):
xen_platform: unplug also SCSI disks
xen_platform: SUSE xenlinux unplug for emulated PCI
hw/i386/xen/xen_platform.c | 36 +++-
1 file changed, 35 insertions(+), 1
Implement SUSE specific unplug protocol for emulated PCI devices
in PVonHVM guests. Its a simple 'outl(1, (ioaddr + 4));'.
This protocol was implemented and used since Xen 3.0.4.
It is used in all SUSE/SLES/openSUSE releases up to SLES11SP3 and
openSUSE 12.3.
Signed-off-by: Olaf Hering &l
On Wed, Aug 31, Olaf Hering wrote:
> The following HVM domU.cfg crashes during boot from emulated network
> with qemu 2.7, but it works fine with qemu stable-2.6 branch:
>
> name='hvm'
> memory=1024
> vcpus=2
> boot="n"
> disk=[ 'file:/disk0.raw,hda,w', ]
>
Update unplug in Xen HVM guests to cover more cases.
Please review.
Olaf
changes in v2:
- fix issues reported by checkpatch
Olaf Hering (2):
xen_platform: unplug also SCSI disks
xen_platform: SUSE xenlinux unplug for emulated PCI
hw/i386/xen/xen_platform.c | 35
-off-by: Olaf Hering <o...@aepfle.de>
---
hw/i386/xen/xen_platform.c | 4
1 file changed, 4 insertions(+)
diff --git a/hw/i386/xen/xen_platform.c b/hw/i386/xen/xen_platform.c
index aa78393..53be3c7 100644
--- a/hw/i386/xen/xen_platform.c
+++ b/hw/i386/xen/xen_platform.c
@@ -114,6 +
The following HVM domU.cfg crashes during boot from emulated network
with qemu 2.7, but it works fine with qemu stable-2.6 branch:
name='hvm'
memory=1024
vcpus=2
boot="n"
disk=[ 'file:/disk0.raw,hda,w', ]
vif=[ 'mac=00:08:15:41:10:80,bridge=br0', ]
keymap="de"
serial="pty"
builder="hvm"
vnc=1
On Wed, Sep 14, Stefano Stabellini wrote:
> Written like this, the code will unplug any Xen SCSI disks together with
> Xen IDE disks when the guest writes "1" to ioport `0x10`. I am sorry to
> be pedantic, but the recent changes introduced to
> docs/misc/hvm-emulated-unplug.markdown do not cover
On Wed, Sep 14, Stefano Stabellini wrote:
> The doc says:
>
> "If VMDP was configured to control just NIC devices it would write the
> value 0x1 to offset 0x8. If VMDP was configured to control just storage
> devices it would write the value 0x2 to offset 0x8."
>
> So 0x1 to 0x8 to unplug NICs,
On Mon, Oct 03, Cédric Le Goater wrote:
> +++ b/include/hw/ppc/pnv_xscom.h
> +typedef struct PnvChip PnvChip;
This causes a compile error with gcc 4.5.1:
...
[ 552s] In file included from
/usr/src/packages/BUILD/qemu-20161104T092624.9226682/hw/ppc/pnv.c:29:0:
[ 552s]
On Thu, Oct 13, Stefano Stabellini wrote:
> On Fri, 2 Sep 2016, Olaf Hering wrote:
> > Implement SUSE specific unplug protocol for emulated PCI devices
> > in PVonHVM guests. Its a simple 'outl(1, (ioaddr + 4));'.
> > This protocol was implemented and used since Xen 3.0.4.
Update unplug in Xen HVM guests to cover more cases.
Please review.
changes in v3:
- adapt to API changes from 49137bf ("block-backend: remove blk_flush_all")
changes in v2:
- fix issues reported by checkpatch
Olaf Hering (2):
xen_platform: unplug also SCSI disks
xen_plat
-off-by: Olaf Hering <o...@aepfle.de>
---
hw/i386/xen/xen_platform.c | 4
1 file changed, 4 insertions(+)
diff --git a/hw/i386/xen/xen_platform.c b/hw/i386/xen/xen_platform.c
index f85635c..91d8a7a 100644
--- a/hw/i386/xen/xen_platform.c
+++ b/hw/i386/xen/xen_platform.c
@@ -114,6 +
versions are handled as well.
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
hw/i386/xen/xen_platform.c | 27 ++-
1 file changed, 26 insertions(+), 1 deletion(-)
diff --git a/hw/i386/xen/xen_platform.c b/hw/i386/xen/xen_platform.c
index 91d8a7a..2e1e543 100644
--- a/hw/i3
Am 23. November 2016 21:44:50 MEZ, schrieb Olaf Hering <o...@aepfle.de>:
>Is this a can for 2.x?
candidate
Olaf
On Wed, Nov 23, Olaf Hering wrote:
> > > +if (!blk_split_discard(ioreq, req->sector_number,
> > > req->nr_sectors)) {
> > > +goto err;
> > How is error handling supposed to work here?
In the guest the cmd is stuck, instead of getting
:
mkfs.ext4 -F /dev/xvda
Discarding device blocks: failed - Input/output error
Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS.
Add input range checking to avoid overflow.
Fixes f313520 ("xen_disk: add discard support")
Signed-off-by: Olaf Hering <o...@aepf
:
mkfs.ext4 -F /dev/xvda
Discarding device blocks: failed - Input/output error
Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS.
Add input range checking to avoid overflow.
Fixes f313520 ("xen_disk: add discard support")
Signed-off-by: Olaf Hering <o...@aepf
Ping.
On Fri, Nov 18, Olaf Hering wrote:
> On Fri, Nov 18, Olaf Hering wrote:
>
> > @@ -708,12 +743,10 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
> > +if (!blk_split_discard(ioreq, req->sector_number,
> > req->nr_sectors)) {
> > +
Am 23. November 2016 13:27:13 MEZ, schrieb Kevin Wolf :
>Am 23.11.2016 um 12:40 hat Eric Blake geschrieben:
>> Qualifies as a bug fix, so requesting 2.8 inclusion.
>> Reviewed-by: Eric Blake
Is this a can for 2.x?
Olaf
On Fri, Nov 18, Eric Blake wrote:
> if (sec_start > (INT64_MAX >> BDRV_SECTOR_BITS) - sec_count)
I have looked at this for a while now and cant spot how this would cover
all cases. Are you saying there should be just a single overflow check,
yours? My change has two: one to check for wrap around
On Tue, Nov 22, Eric Blake wrote:
> if (sec_start + sec_count < sec_count ||
> sec_start > (INT64_MAX >> BDRV_SECTOR_BITS) - sec_count) {
> return false;
> }
My point was: how does this handle sec_start=0,sec_count=UINT64_MAX or
sec_start=INT64_MAX,sec_count=INT64_MAX? For me this gets
On Fri, May 06, Max Reitz wrote:
> On 01.04.2016 19:49, Olaf Hering wrote:
> > On Fri, Apr 01, Max Reitz wrote:
> >
> >> In any case, do you have a test case where a guest was able to submit a
> >> request that led to the overflow error you described in the comm
Am 18. November 2016 14:43:18 MEZ, schrieb Eric Blake <ebl...@redhat.com>:
>On 11/18/2016 04:24 AM, Olaf Hering wrote:
>> The guest sends discard requests as u64 sector/count pairs, but the
>> block layer operates internally with s64/s32 pairs. The conversion
>> lead
On Fri, Nov 18, Eric Blake wrote:
> On 11/18/2016 04:24 AM, Olaf Hering wrote:
> > +/* Overflowing byte limit? */
> > +if ((sec_start + sec_count) > ((INT64_MAX + INT_MAX) >>
> > BDRV_SECTOR_BITS)) {
> This is undefined. INT64_MAX + anything non-negati
On Thu, Nov 17, Olaf Hering wrote:
> Looking at current master, BLKIF_OP_DISCARD is indeed broken. The values
> passed from the guest are u64 and get stashed into signed values. I will
> add a loop to repeatedly call blk_aio_pdiscard with small chunks of
> BDRV_REQUEST_MAX_SECT
:
mkfs.ext4 -F /dev/xvda
Discarding device blocks: failed - Input/output error
Fix this by splitting the request into chunks of BDRV_REQUEST_MAX_SECTORS.
Add input range checking to avoid overflow.
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
hw/block/xen_disk.
On Fri, Nov 18, Olaf Hering wrote:
> @@ -708,12 +743,10 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
> +if (!blk_split_discard(ioreq, req->sector_number, req->nr_sectors)) {
> +goto err;
How is error handling supposed to work here?
Initially I forgot
On Wed, Nov 02, Thomas Huth wrote:
> Olaf Hering reported a build failure due to an undefined reference
> to 'qemu_log_vprintf'. Explicitely including qemu/log.h seems to
> fix the issue.
>
> Signed-off-by: Thomas Huth <th...@redhat.com>
Tested-by: Olaf Hering &l
Not sure if there are automated build tests to catch such mistakes,
these errors appeared after qemu-20161028T142950.fd209e4:
[ 337s] ../hw/xen/xen_pvdev.o: In function `xen_pv_printf':
[ 337s]
/home/abuild/rpmbuild/BUILD/qemu-20161101T165305.4eb28ab/hw/xen/xen_pvdev.c:220:
undefined
On Wed, Nov 02, Thomas Huth wrote:
> Does it work if you add an explicit #include "qemu/log.h" at the
> beginning of that file (after osdep.h)?
I will try that.
> system. Which version of GTK are you using?
2.18.9 as included in SLE11.
Olaf
signature.asc
Description: PGP signature
On Wed, Nov 02, Olaf Hering wrote:
> On Wed, Nov 02, Thomas Huth wrote:
> > Does it work if you add an explicit #include "qemu/log.h" at the
> > beginning of that file (after osdep.h)?
> I will try that.
Yes, that helps. Thanks.
Olaf
signature.asc
Description: PGP signature
On Fri, Jul 28, Eric Blake wrote:
> This version is prone to multiplication overflow (well, maybe not, but
> you have to audit for that). Wouldn't it be better to use:
What could go wrong?
qemu will die either way, I think.
Olaf
signature.asc
Description: PGP signature
g_malloc0_n is available since glib-2.24. To allow build with older glib
versions use the generic g_malloc0, which is already used in many other
places in the code.
Fixes commit 3284fad728 ("xen-disk: add support for multi-page shared rings")
Signed-off-by: Olaf Hering <o...@aepfl
g_malloc0_n is available since glib-2.24. To allow build with older glib
versions use the generic g_new0, which is already used in many other
places in the code.
Fixes commit 3284fad728 ("xen-disk: add support for multi-page shared rings")
Signed-off-by: Olaf Hering <o...@aepfl
On Tue, Jun 27, Stefano Stabellini wrote:
> From: Paul Durrant
> The blkif protocol has had provision for negotiation of multi-page shared
> rings for some time now and many guest OS have support in their frontend
> drivers.
> +++ b/hw/block/xen_disk.c
> +domids =
On Wed, Sep 20, Stefano Stabellini wrote:
> From: Olaf Hering <o...@aepfle.de>
> g_malloc0_n is available since glib-2.24. To allow build with older glib
> versions use the generic g_new0, which is already used in many other
> places in the code.
> Fixes commit 3284fad728 (&
Am Fri, 4 May 2018 17:00:23 +0100
schrieb Daniel P. Berrangé :
> - I suggested following libvirt's lead in writing a policy for how
> we pick supported OS targets to inform maintainers when min versions
> can be increased.
Since Xen depends on qemu, it also means
, with g_renew.
Fixes commit 418026ca43 ("util: Introduce vfio helpers")
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
This must go into stable-2.12.
util/vfio-helpers.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/util/vfio-helpers.c b/util/vfio
Am Mon, 14 May 2018 23:04:31 -0700 (PDT)
schrieb no-re...@patchew.org:
> vfio-helpers.c:525:17: error: assignment from incompatible pointer type
Does anyone happen to know what s390x is trying to do here?
Olaf
pgpHBYpHmh42k.pgp
Description: Digitale Signatur von OpenPGP
, with g_renew.
Fixes commit 418026ca43 ("util: Introduce vfio helpers")
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
This must go into stable-2.12.
util/vfio-helpers.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/util/vfio-helpers.c b/util/vfio
Am Wed, 9 May 2018 00:15:42 +0200
schrieb Paolo Bonzini :
> Actually, --program-prefix= and --exec-prefix= have a meaning, they
> cannot just be ignored. For now I've removed this line; what is the
> exact incantation used by SUSE Linux?
Like in every other rpm.rpm I guess:
Am Thu, 7 Jun 2018 09:25:24 +0100
schrieb Daniel P. Berrangé :
> https://qemu.weilnetz.de/doc/qemu-doc.html#Supported-build-platforms
SLE12-SP2: 2.48.2
Olaf
pgpzqS0yDN306.pgp
Description: Digitale Signatur von OpenPGP
Am Thu, 26 Apr 2018 13:52:04 +0200
schrieb Paolo Bonzini :
> min-glib Dockerfile is using glib 2.22
Why did it not prevent the bad commit from entering 2.12?
Olaf
pgpLXaYUkUuzv.pgp
Description: Digitale Signatur von OpenPGP
With commit 927128222b ("ui: add keycodemapdb repository as a GIT
submodule") sed is called with -E, which requires a recent sed.
The version of sed shipped with SLE11 is 4.1.5, which does not
understand the option '-E'. As a result xen-4.11 will fail to build.
Is there a way to achive the same
Am Thu, 9 Aug 2018 16:37:05 +0200
schrieb Paolo Bonzini :
> If no RAM is allocated (i.e. xen_map_cache will return NULL), however,
> the memory should not be registered as RAM with the memory API. So I
> think the bug is in Xen code.
Someone familiar with that code has to figure that out. A
Am Thu, 9 Aug 2018 16:38:16 +0200
schrieb Olaf Hering :
> Someone familiar with that code has to figure that out. A ballooned page will
> trigger that bug.
Indeed, xen-4.4 + qemu-3.0 crashes with ballooned pages. That can easily happen
if the domU does readdir via NFS.
Olaf
Thread 1
Am Thu, 9 Aug 2018 16:52:22 +0200
schrieb Paolo Bonzini :
> I think the simplest fix is for the map
> cache to set aside a zero page and return it whenever it is asked for a
> ballooned page.
Can qemu actually know if it ran into a ballooned page? I think no.
Olaf
pgpmdqCLn4Xnn.pgp
The codepaths behind qemu_ram_ptr_length can return NULL.
Avoid crashing the device-model in such case, just move on.
Signed-off-by: Olaf Hering
---
This happens if calling xendevicemodel_create_ioreq_server() is disabled,
and eventually if that function returns an error.
---
exec.c | 3 ++-
1
Am Thu, 9 Aug 2018 16:14:03 +0200
schrieb Olaf Hering :
> The codepaths behind qemu_ram_ptr_length can return NULL.
While that might be a bug by itself, the question is why in that case no
memset(buf, 0xff, l) is done?
Olaf
pgpPcpqmT0hnz.pgp
Description: Digitale Signatur von OpenPGP
Am Fri, 10 Aug 2018 14:29:28 +0200
schrieb Paolo Bonzini :
> On 10/08/2018 12:32, Olaf Hering wrote:
> > Am Fri, 10 Aug 2018 12:25:09 +0200
> > schrieb Paolo Bonzini :
> > So you mean that function must not return NULL?
> > Or should the caller check for the result?
Am Fri, 10 Aug 2018 12:25:09 +0200
schrieb Paolo Bonzini :
> There are many more callers of qemu_ram_ptr_length, and none of them check
> the result.
So you mean that function must not return NULL?
Or should the caller check for the result?
Olaf
pgp6zq9J9aL36.pgp
Description: Digitale
Am Wed, 18 Apr 2018 10:04:03 -0500
schrieb Eric Blake :
> Rather than hacking configure, why not fix util/vfio-helpers.c to avoid
> g_realloc_n()? We've done it before; see commit 071d405
I can not really test it, this variant may work:
--- a/util/vfio-helpers.c
+++
Am Wed, 18 Apr 2018 10:04:03 -0500
schrieb Eric Blake :
> avoid g_realloc_n()?
I guess there is some global header for this, to prevent further accidents?
extern void *g_realloc_n_can_not_be_used(void);
#define g_realloc_n(a,b,c) g_realloc_n_can_not_be_used()
Olaf
Am Wed, 18 Apr 2018 14:45:38 +0200
schrieb Cornelia Huck :
> Are we ready to give up support for whatever distro is still on 2.22?
This is SLE11, in case it matters for this discussion.
Olaf
pgp5sY4ljZvqS.pgp
Description: Digitale Signatur von OpenPGP
Since usage of g_realloc_n was introduced, glib-2.22 can not be used
anymore on Linux. Leave non-Linux unchanged because one developer system
still uses it.
Fixes commit 418026ca43 ("util: Introduce vfio helpers")
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
configure |
Am Wed, 18 Apr 2018 13:47:53 +0100
schrieb Daniel P. Berrangé :
> IIRC, one of Peters desired build targets required 2.22, but I guess we
> missed this bug because that platform would not enable vfio functionality ?
Does the check in configure need a wrap "if Linux", just as
Since usage of g_realloc_n was introduced, glib-2.22 can not be used anymore.
Fixes commit 418026ca43 ("util: Introduce vfio helpers")
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
configure | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/configure b
Am Wed, 18 Apr 2018 05:53:58 -0700 (PDT)
schrieb no-re...@patchew.org:
> This series failed docker-build@min-glib build test.
I guess that test must be fixed either way because it does not test what it is
supposed to test...
Olaf
pgpgzq5OJ4fiN.pgp
Description: Digitale Signatur von OpenPGP
mmand "gcc" hardcoded. That binary may not be suitable to build ipxe.
Reuse the same compiler command that is used to build qemu by specifying
CC and HOST_CC as make commandline option to override the built-in
defaults of ipxe.
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
roms/Makefile
Am Wed, 18 Apr 2018 10:13:10 +0200
schrieb Olaf Hering <o...@aepfle.de>:
> configure allows to run a dedicated compiler binary, which is stored in
> CC and HOST_CC variables, and overrides the default 'cc' command.
> This command however is not used when building ipxe from source.
Extend the list of recognized, but ignored options from rpms %configure
macro. This fixes build on hosts running SUSE Linux.
Signed-off-by: Olaf Hering <o...@aepfle.de>
---
configure | 3 +++
1 file changed, 3 insertions(+)
diff --git a/configure b/configure
index 0a19b033bc..6e9b994f21
1 - 100 of 199 matches
Mail list logo