05.02.2021 14:15, Claudio Fontana wrote:
Just in case it helps,
I started getting this error only when I rebased to latest master a couple days
ago, after regular rebasing and testing with full make check every 1 or 2 days.
The first crash below is on assertion which was added recently. I
On Fri, 5 Feb 2021 at 10:31, Philippe Mathieu-Daudé wrote:
> 1/ virtio-mmio must be sysbus-device,
Yes.
> 2/ we can not sysbus-map out of main memory so private container
> is incorrect, and Versal can not use "mr-ps-switch"?
No. If you have a sysbus device, and you want to map it somewhere
On Tue 02 Feb 2021 01:49:43 PM CET, Vladimir Sementsov-Ogievskiy wrote:
> The recommended use of qemu error api assumes returning status together
> with setting errp and avoid void functions with errp parameter. Let's
> improve bdrv_append and some friends to reduce error-propagation
> overhead in
Just in case it helps,
I started getting this error only when I rebased to latest master a couple days
ago, after regular rebasing and testing with full make check every 1 or 2 days.
Ciao,
Claudio
On 2/5/21 12:10 PM, Vladimir Sementsov-Ogievskiy wrote:
> I reproduced the following crash fast
On Thu, Feb 04, 2021 at 01:07:07PM -0600, Eric Blake wrote:
> Supporting '0x20M' looks odd, particularly since we have an 'E' suffix
> that is ambiguous between a hex digit and the extremely large exibyte
> suffix, as well as a 'B' suffix for bytes. In practice, people using
> hex inputs are
On Thu, Feb 04, 2021 at 01:07:08PM -0600, Eric Blake wrote:
> The value '1.1k' is inexact; 1126.4 bytes is not possible, so we
> happen to truncate it to 1126. Our use of fractional sizes is
> intended for convenience, but when a user specifies a fraction that is
> not a clean translation to
I reproduced the following crash fast enough:
0 raise () at /lib64/libc.so.6
1 abort () at /lib64/libc.so.6
2 _nl_load_domain.cold () at /lib64/libc.so.6
3 annobin_assert.c_end () at /lib64/libc.so.6
4 bdrv_reopen_multiple (bs_queue=0x55de75fa9b70, errp=0x0)
at ../block.c:3820
5
On Thu, Feb 04, 2021 at 01:07:06PM -0600, Eric Blake wrote:
> We have multiple clients of qemu_strtosz (qemu-io, the opts visitor,
> the keyval visitor), and it gets annoying that edge-case testing is
> impacted by implicit rounding to 53 bits of precision due to parsing
> with strtod(). As an
On Wed, Feb 03, 2021 at 06:18:28PM +0100, Thomas Huth wrote:
> The pc-1.x machine types have been deprecated since QEMU v5.0 already, and
> nobody complained, so they can now be removed. While we're at it, also
> remove some compatibility switches and related code that are now not
> necessary
On 04/02/21 20:39, James Bottomley wrote:
v3: add bounds checking and remove misleading comment
v2: fix build issues and update comments. Also drop the first patch
which is now merged upstream
The two patches introduce a parser for the optional OVMF description
table which is placed just
On 2/4/21 11:38 PM, Vladimir Sementsov-Ogievskiy wrote:
> 04.02.2021 20:51, Peter Maydell wrote:
>> On Thu, 4 Feb 2021 at 17:48, Philippe Mathieu-Daudé
>> wrote:
>>>
>>> Hi,
>>>
>>> Based on commit 1ed9228f63e (ericb/tags/pull-nbd-2021-02-02-v2)
>>> I got:
>>>
>>>TEST iotest-qcow2: 030
On 26/01/21 18:36, Tom Lendacky wrote:
From: Tom Lendacky
This patch series provides support for launching an SEV-ES guest.
Secure Encrypted Virtualization - Encrypted State (SEV-ES) expands on the
SEV support to protect the guest register state from the hypervisor. See
"AMD64 Architecture
On 05/02/21 10:51, Daniel P. Berrangé wrote:
+if (!pc_system_ovmf_table_find(SEV_SECRET_GUID, , NULL)) {
+error_setg(errp, "SEV: no secret area found in OVMF,"
+ " gpa must be specified.");
+return;
+}
IIUC, historically QEMU has
On 05/02/21 09:33, Chenyi Qiang wrote:
Protection Keys for Supervisor Pages (PKS) is a feature that extends the
Protection Keys architecture to support thread-specific permission
restrictions on superviosr pages, which works similar to an existing
feature named PKU (protecting user-mode pages).
On Thu, Feb 04, 2021 at 02:44:03PM -0600, Eric Blake wrote:
> On 2/4/21 2:09 PM, Peter Lieven wrote:
> > Am 02.02.21 um 16:51 schrieb Eric Blake:
> >> On 1/28/21 8:07 AM, Peter Lieven wrote:
> >>> Signed-off-by: Peter Lieven
> >> Your commit message says 'what', but not 'why'. Generally, the
On 2/3/21 6:09 PM, Alexander Graf wrote:
On 03.02.21 17:55, Philippe Mathieu-Daudé wrote:
On 1/11/21 3:50 PM, Luc Michel wrote:
Hi,
We would like to work on improving QEMU to be able to create custom
machines from the command line. The goal here is to get feedback from
the community and
Public bug reported:
As of commit 1ba089f2255, on Cavium CN8890 (ThunderX cores):
$ qemu-system-aarch64 -display none -accel kvm -M virt,gic-version=3 -accel kvm
-cpu cortex-a57 --trace \*kvm_vcpu\*
kvm_vcpu_ioctl cpu_index 0, type 0x4020aeae, arg 0x9b7f9b18
qemu-system-aarch64:
04.02.2021 22:07, Eric Blake wrote:
The value '1.1k' is inexact; 1126.4 bytes is not possible, so we
happen to truncate it to 1126. Our use of fractional sizes is
intended for convenience, but when a user specifies a fraction that is
not a clean translation to binary, truncating/rounding behind
** Tags added: ui
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1914748
Title:
Confuse error message when KVM can not start requested CPU
Status in QEMU:
New
Bug description:
As of commit
On Thu, Feb 04, 2021 at 01:07:08PM -0600, Eric Blake wrote:
> The value '1.1k' is inexact; 1126.4 bytes is not possible, so we
> happen to truncate it to 1126. Our use of fractional sizes is
> intended for convenience, but when a user specifies a fraction that is
> not a clean translation to
On Fri, Feb 05, 2021 at 01:25:18PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> 04.02.2021 22:07, Eric Blake wrote:
> >Supporting '0x20M' looks odd, particularly since we have an 'E' suffix
>
> What about also deprecating 'E' suffix? (just my problem of reviewing
> previous patch)
Ha! What if
On 2/5/21 11:03 AM, Peter Maydell wrote:
> On Fri, 5 Feb 2021 at 07:53, Philippe Mathieu-Daudé wrote:
>> Indeed, you found a design issue IMO:
>>
>> Versal creates the "mr-ps-switch" to be explicitly different from
>> the main sysbus memory. TYPE_VIRTIO_MMIO is a SYSBUS device, thus
>> can not be
On Thu, Feb 04, 2021 at 01:07:06PM -0600, Eric Blake wrote:
> We have multiple clients of qemu_strtosz (qemu-io, the opts visitor,
> the keyval visitor), and it gets annoying that edge-case testing is
> impacted by implicit rounding to 53 bits of precision due to parsing
> with strtod(). As an
Proposed fix:
https://www.mail-archive.com/qemu-devel@nongnu.org/msg779652.html
** Changed in: qemu
Status: New => Confirmed
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1914638
Title:
04.02.2021 22:07, Eric Blake wrote:
Supporting '0x20M' looks odd, particularly since we have an 'E' suffix
What about also deprecating 'E' suffix? (just my problem of reviewing previous
patch)
that is ambiguous between a hex digit and the extremely large exibyte
suffix, as well as a 'B'
Retry should only be triggered when timeout is not reached, so let's check
timeout before retry. Device should also reset retry_start_time after
successful retry.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
block/block-backend.c | 25 +++-
Mark failed requests as to be retried and implement retry_request_cb to
handle these requests.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
hw/scsi/scsi-disk.c | 16
1 file changed, 16 insertions(+)
diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c
index
Add retry_request_cb in BlockDevOps to do device specific retry action.
Backend's timer would be registered only when the backend is set 'retry'
on errors and the device supports retry action.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
block/block-backend.c | 8
05.02.2021 13:06, Vladimir Sementsov-Ogievskiy wrote:
04.02.2021 23:12, Eric Blake wrote:
On 2/4/21 1:07 PM, Eric Blake wrote:
We have multiple clients of qemu_strtosz (qemu-io, the opts visitor,
the keyval visitor), and it gets annoying that edge-case testing is
impacted by implicit rounding
Move the code that retries requests from scsi_dma_restart_bh() to its own,
non-static, function. This will allow us to call it from the
retry_request_cb() of scsi-disk in a future patch.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
hw/scsi/scsi-bus.c | 16 +++-
A VM in the cloud environment may use a virutal disk as the backend storage,
and there are usually filesystems on the virtual block device. When backend
storage is temporarily down, any I/O issued to the virtual block device
will cause an error. For example, an error occurred in ext4 filesystem
This just came up on the list the other day. It should be fixed by this patch:
https://patchew.org/QEMU/20210203161340.55210-1-aa...@os.amperecomputing.com/
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to the bug report.
Add a timer to regularly trigger retry on errors.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
block/block-backend.c | 21
1 file changed, 21 insertions(+)
diff --git a/block/block-backend.c b/block/block-backend.c
index e493f17515..3a9d55cbe3 100644
---
Add "retry_interval" and "retry_timeout" parameter for drive and device
option. These parameter are valid only when werror/rerror=retry.
eg. --drive file=image,rerror=retry,retry_interval=1000,retry_timeout=5000
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
block/block-backend.c
Add a new error action 'retry' to support retry on errors.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
blockdev.c | 2 ++
qapi/block-core.json | 9 +++--
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/blockdev.c b/blockdev.c
index b250b9b959..ece1d8ae58
Enable retry action when backend's retry timer is available. It would
trigger the timer to do device specific retry action.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
block/block-backend.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/block/block-backend.c
Insert failed requests into device's list for later retry and handle
queued requests to implement retry_request_cb.
Signed-off-by: Jiahui Cen
Signed-off-by: Ying Fang
---
hw/block/virtio-blk.c | 21 +---
1 file changed, 18 insertions(+), 3 deletions(-)
diff --git
On 2/5/21 5:52 PM, Claudio Fontana wrote:
Hi LiZhiJian,
just one thing I noticed in your call stack, your $Subject talks about AArch64,
and you end up in arm_v7m_mmu_idx_for_secstate,
which means that ARM_FEATURE_M is detected, so definitely something looks off
when it comes to the feature
On 05.02.21 10:16, Max Reitz wrote:
On 05.02.21 09:47, Peter Lieven wrote:
Am 05.02.21 um 09:18 schrieb Max Reitz:
On 04.02.21 21:09, Peter Lieven wrote:
Am 02.02.21 um 16:51 schrieb Eric Blake:
On 1/28/21 8:07 AM, Peter Lieven wrote:
Signed-off-by: Peter Lieven
Your commit message says
On Thu, 4 Feb 2021 at 19:14, Dr. David Alan Gilbert (git)
wrote:
>
> From: "Dr. David Alan Gilbert"
>
> The following changes since commit 1ba089f2255bfdb071be3ce6ac6c3069e8012179:
>
> Merge remote-tracking branch 'remotes/armbru/tags/pull-qmp-2021-02-04' into
> staging (2021-02-04 14:15:35
04.02.2021 22:07, Eric Blake wrote:
We have multiple clients of qemu_strtosz (qemu-io, the opts visitor,
the keyval visitor), and it gets annoying that edge-case testing is
impacted by implicit rounding to 53 bits of precision due to parsing
with strtod(). As an example posted by Rich Jones:
04.02.2021 23:12, Eric Blake wrote:
On 2/4/21 1:07 PM, Eric Blake wrote:
We have multiple clients of qemu_strtosz (qemu-io, the opts visitor,
the keyval visitor), and it gets annoying that edge-case testing is
impacted by implicit rounding to 53 bits of precision due to parsing
with strtod().
On Fri, 5 Feb 2021 at 07:53, Philippe Mathieu-Daudé wrote:
> Indeed, you found a design issue IMO:
>
> Versal creates the "mr-ps-switch" to be explicitly different from
> the main sysbus memory. TYPE_VIRTIO_MMIO is a SYSBUS device, thus
> can not be created without being plugged on sysbus.
> We
On Thu, Feb 04, 2021 at 04:20:18PM -0600, Eric Blake wrote:
> Our default of a backlog of 1 connection is rather puny, particularly
> for scenarios where we expect multiple listeners to connect (such as
> qemu-nbd -e X). For Unix sockets, there's no real harm in supporting
> a larger backlog, and
On 05.02.21 10:39, Thomas Huth wrote:
> There is a compiler warning with GCC 9.3 when compiling with
> the -fsanitize=thread compiler flag:
>
> In function 'strncpy',
> inlined from 's390x_write_elf64_notes' at
> ../target/s390x/arch_dump.c:219:9:
>
On Thu, Feb 04, 2021 at 04:20:18PM -0600, Eric Blake wrote:
> Our default of a backlog of 1 connection is rather puny, particularly
> for scenarios where we expect multiple listeners to connect (such as
> qemu-nbd -e X). For Unix sockets, there's no real harm in supporting
> a larger backlog, and
Hi LiZhiJian,
just one thing I noticed in your call stack, your $Subject talks about AArch64,
and you end up in arm_v7m_mmu_idx_for_secstate,
which means that ARM_FEATURE_M is detected, so definitely something looks off
when it comes to the feature bits..
Ciao,
Claudio
On 2/5/21 4:04 AM,
On Thu, Feb 04, 2021 at 11:39:39AM -0800, James Bottomley wrote:
> If the gpa isn't specified, it's value is extracted from the OVMF
> properties table located below the reset vector (and if this doesn't
> exist, an error is returned). OVMF has defined the GUID for the SEV
> secret area as
There is a compiler warning with GCC 9.3 when compiling with
the -fsanitize=thread compiler flag:
In function 'strncpy',
inlined from 's390x_write_elf64_notes' at ../target/s390x/arch_dump.c:219:9:
/usr/include/x86_64-linux-gnu/bits/string_fortified.h:106:10: error:
'__builtin_strncpy'
On Fri, Feb 5, 2021 at 10:33 AM Daniel P. Berrangé wrote:
> On Thu, Feb 04, 2021 at 10:57:20PM +0100, Philippe Mathieu-Daudé wrote:
> > On 2/4/21 6:18 PM, Daniel P. Berrangé wrote:
> > > This is an incremental step in converting vmstate loading code to report
> > > via Error objects instead of
On Thu, Feb 04, 2021 at 10:57:20PM +0100, Philippe Mathieu-Daudé wrote:
> On 2/4/21 6:18 PM, Daniel P. Berrangé wrote:
> > This is an incremental step in converting vmstate loading code to report
> > via Error objects instead of printing directly to the console/monitor.
> >
> > Signed-off-by:
Simply add the flag to an existing job, no need for yet another
job here.
Reviewed-by: Wainer dos Santos Moschetta
Signed-off-by: Thomas Huth
---
.gitlab-ci.yml | 1 +
.travis.yml| 6 --
2 files changed, 1 insertion(+), 6 deletions(-)
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index
Use clang-10, so we can also use the --enable-tsan configure
option instead of only passing the flag via --extra-cflags.
Signed-off-by: Thomas Huth
---
.gitlab-ci.yml | 9 +
.travis.yml| 51 --
2 files changed, 9 insertions(+), 51
We already have similar jobs in the gitlab-CI ("build-some-softmmu" and
"build-user-plugins"), so let's switch one of them to use --enable-debug
instead of --enable-debug-tcg, then we can simply drop these jobs from
the Travis-CI.
Signed-off-by: Thomas Huth
---
.gitlab-ci.yml | 2 +-
Add it to the existing Clang job and also add a job that covers the
linux-user code with this compiler flag. To make sure that the detected
problems are not simply ignored, let's also use "-fno-sanitize-recover=..."
now instead.
Signed-off-by: Thomas Huth
---
.gitlab-ci.yml | 14 --
From: Philippe Mathieu-Daudé
Similarly to commit 8cdb2cef3f1, move the gprof/gcov test to GitLab.
The coverage-summary.sh script is not Travis-CI specific, make it
generic.
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20201108204535.2319870-10-phi...@redhat.com>
[thuth: Add gcovr and
Since Travis changed their policies, travis-ci.org will soon become
completely useless for the QEMU project. We should now really make sure
that we move the remaining tests as good as possible to the gitlab-CI
instead.
v2:
- Run "make check" in the gprof/gcov test
- Make sure that we run at
On 05.02.21 09:47, Peter Lieven wrote:
Am 05.02.21 um 09:18 schrieb Max Reitz:
On 04.02.21 21:09, Peter Lieven wrote:
Am 02.02.21 um 16:51 schrieb Eric Blake:
On 1/28/21 8:07 AM, Peter Lieven wrote:
Signed-off-by: Peter Lieven
Your commit message says 'what', but not 'why'. Generally, the
John Snow writes:
> On 2/3/21 9:47 AM, Markus Armbruster wrote:
>> John Snow writes:
>>
>>> Presently, we use a tuple to attach a dict containing annotations
>>> (comments and compile-time conditionals) to a tree node. This is
>>> undesirable because dicts are difficult to strongly type;
On Thu, Feb 04, 2021 at 04:51:39PM +0100, Thomas Huth wrote:
> On 04/02/2021 09.36, Gerd Hoffmann wrote:
> >Hi,
> >
> > > enum USBDeviceFlags {
> > > -USB_DEV_FLAG_FULL_PATH,
> > > +USB_DEV_FLAG_FULL_PATH, /* unused since QEMU v6.0 */
> >
> > Why not just drop it? Any
On 05/02/2021 09.18, Thomas Huth wrote:
On 05/02/2021 08.08, Christian Borntraeger wrote:
On 05.02.21 07:12, Thomas Huth wrote:
On 04/02/2021 18.00, Christian Borntraeger wrote:
On 04.02.21 17:41, Thomas Huth wrote:
According to the "ELF-64 Object File Format" specification:
"The first
On Feb 3 09:27, Keith Busch wrote:
> Just had the one comment on patch 4, which is really no big deal. I need
> to integrate tooling and/or kernel support in order to properly test
> this, but just from code inspection, I think it's good.
>
> Reviewed-by: Keith Busch
>
Applied, with the small
Am 05.02.21 um 09:18 schrieb Max Reitz:
> On 04.02.21 21:09, Peter Lieven wrote:
>> Am 02.02.21 um 16:51 schrieb Eric Blake:
>>> On 1/28/21 8:07 AM, Peter Lieven wrote:
Signed-off-by: Peter Lieven
>>> Your commit message says 'what', but not 'why'. Generally, the one-line
>>> 'what' works
Eduardo Habkost writes:
> On Thu, Feb 04, 2021 at 04:37:45PM +0100, Markus Armbruster wrote:
>> Eduardo Habkost writes:
>>
>> > On Wed, Feb 03, 2021 at 03:47:36PM +0100, Markus Armbruster wrote:
>> >> John Snow writes:
>> >>
>> >> > Presently, we use a tuple to attach a dict containing
Hi,
On 2/5/21 4:16 AM, Jason Wang wrote:
>
> On 2021/2/5 上午3:12, Peter Xu wrote:
>> Previous work on dev-iotlb message broke vhost on either SMMU
>
>
> Have a quick git grep and it looks to me v3 support ATS and have command
> for device iotlb (ATC) invalidation.
Yes I will do that. Should
Expose the VMX exit/entry load pkrs control bits in
VMX_TRUE_EXIT_CTLS/VMX_TRUE_ENTRY_CTLS MSRs to guest, which supports the
PKS in nested VM.
Signed-off-by: Chenyi Qiang
---
target/i386/cpu.c | 4 ++--
target/i386/cpu.h | 2 ++
2 files changed, 4 insertions(+), 2 deletions(-)
diff --git
Protection Keys for Supervisor Pages (PKS) is a feature that extends the
Protection Keys architecture to support thread-specific permission
restrictions on superviosr pages, which works similar to an existing
feature named PKU (protecting user-mode pages).
Thanks Paolo to send out the PKS QEMU
PKS introduces MSR IA32_PKRS(0x6e1) to manage the supervisor protection
key rights. Page access and writes can be managed via the MSR update
without TLB flushes when permissions change.
Add the support to save/load IA32_PKRS MSR in guest.
Signed-off-by: Chenyi Qiang
---
target/i386/kvm/kvm.c |
On 05/02/2021 08.08, Christian Borntraeger wrote:
On 05.02.21 07:12, Thomas Huth wrote:
On 04/02/2021 18.00, Christian Borntraeger wrote:
On 04.02.21 17:41, Thomas Huth wrote:
According to the "ELF-64 Object File Format" specification:
"The first word in the entry, namesz, identifies the
On 04.02.21 21:09, Peter Lieven wrote:
Am 02.02.21 um 16:51 schrieb Eric Blake:
On 1/28/21 8:07 AM, Peter Lieven wrote:
Signed-off-by: Peter Lieven
Your commit message says 'what', but not 'why'. Generally, the one-line
'what' works well as the subject line, but you want the commit body to
05.02.2021 10:56, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy writes:
Add a useful counterpart for trace_handle_qmp_command for debugging
libvirt guests.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
monitor/qmp.c| 6 ++
monitor/trace-events | 1 +
2 files
On 210204 1452, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> The next patch will notify the GL context got flush, which will resume
> the queue processing. However, if this happens within the caller
> context, it will end up with a stack overflow flush/update loop.
>
>
03.02.2021 16:00, Peter Krempa wrote:
Bitmap's source persistence is transported over the migration stream and
the destination mirrors it. In some cases the destination might want to
persist bitmaps which are not persistent on the source (e.g. the result
of merge of bitmaps from a number of
301 - 373 of 373 matches
Mail list logo