** Description changed:
QEMU does not support sprite/overlay in DMA-buf mode for Intel gvt
vGPUs. Some use cases relies on sprite plane support, e.g. hw
accelerated video playback in Windows 10 guest.
To support this feature, functions in both kernel and QEMU should be
implemented:
On 5/6/19 4:57 PM, Laszlo Ersek wrote:
> On 05/05/19 22:06, Philippe Mathieu-Daudé wrote:
>> The reset() code is used in various places, refactor it.
>>
>> Signed-off-by: Philippe Mathieu-Daudé
>> ---
>> hw/block/pflash_cfi02.c | 25 +++--
>> 1 file changed, 15 insertions(+),
On 05/05/19 22:06, Philippe Mathieu-Daudé wrote:
> The pflash device is a child of TYPE_DEVICE, so it can implement
> the DeviceReset handler. Actually it has to implement it, else
> on machine reset it might stay in an incoherent state, as it has
> been reported in the buglink listed below.
>
>
This patch introduces variable length suffixes for being used for inode mapping
instead of the fixed 16 bit size prefixes of patch 1. With this patch the inode
numbers on guest will typically be much smaller (e.g. around >2^1 .. >2^7
instead of >2^48 with patch 1).
I preserved both solutions in
> Btw, I'm assuming where you say...
My bad. Yes, you are correct. The problem is seen when decompiling the blob to
source.
> 1) The powernv machine in qemu is attempting to use a phandle for node
> that doesn't have one.
True.
> 2) dtc is crashing with an assertion - that shouldn't happen,
On 5/6/19 4:04 PM, Laszlo Ersek wrote:
> On 05/04/19 15:35, Philippe Mathieu-Daudé wrote:
>> In commit 1cab464136b4 we incorrectly described the
>> EDK2_BASETOOLS_OPTFLAGS can pass CPPFLAGS and CFLAGS
>> options to the EDK2 build tools, but it only expands
>> the CFLAGS (not to the CPPFLAGS).
>>
In commit 1cab464136b4 we incorrectly described the
EDK2_BASETOOLS_OPTFLAGS can pass CPPFLAGS and CFLAGS
options to the EDK2 build tools, but it only expands
the CFLAGS (not to the CPPFLAGS).
Update the description to be more accurate.
Reported-by: Laszlo Ersek
Reviewed-by: Laszlo Ersek
Signed-off-by: Cédric Le Goater
---
hw/arm/aspeed_soc.c | 50 ++---
1 file changed, 20 insertions(+), 30 deletions(-)
diff --git a/hw/arm/aspeed_soc.c b/hw/arm/aspeed_soc.c
index bd83618ceba9..4b705afd096a 100644
--- a/hw/arm/aspeed_soc.c
+++
Encryption will be done in threads, to take benefit of it, we should
move it out of the lock first.
Signed-off-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Alberto Garcia
Reviewed-by: Max Reitz
---
block/qcow2.c | 35 +--
1 file changed, 21 insertions(+), 14
v6:
- Rebase on master, so 02 changed to reflect changes in
qcow2_compress(), keep r-b.
- Rewrite 06 to even simpler locking [Max], drop r-b
Drop following
"qcow2: qcow2_co_preadv: skip using hd_qiov when possible", I'll do
it in my other coming soon series, and it is actually
Background: decryption will be done in threads, to take benefit of it,
we should move it out of the lock first.
But let's go further: it turns out, that only
qcow2_get_cluster_offset() needs locking, so reduce locking to it.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
block/qcow2.c | 12
Do encryption/decryption in threads, like it is already done for
compression. This improves asynchronous encrypted io.
Signed-off-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Alberto Garcia
Reviewed-by: Max Reitz
---
block/qcow2.h | 8 ++
block/qcow2-cluster.c | 7 ++---
On 05/05/19 22:06, Philippe Mathieu-Daudé wrote:
> The pflash device is a child of TYPE_DEVICE, so it can implement
> the DeviceReset handler. Actually it has to implement it, else
> on machine reset it might stay in an incoherent state, as it has
> been reported in the buglink listed below.
>
>
ping2
Hi!
My approach of adding perf test disliked by Max in other thread, so
01 may be just skipped.
What about 02?
08.04.2019 19:26, Vladimir Sementsov-Ogievskiy wrote:
> Hi!
>
> It's a continuation for
> "[PATCH] qcow2: avoid lseek on block_status if possible"
>
This patch adds an optional qemu device parameter "vii" (very important
inode[s]), which accepts a colon separated list of pathes relative below
the 9p export root, which shall be given the smallest inode suffix (and
hence largest possible inode namespace on guest) according to the given
order.
The goto is unnecessary in the stream_run() since the common exit
code was removed in the commit eb23654dbe43b549ea2a9ebff9d8e:
"jobs: utilize job_exit shim".
Signed-off-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Andrey Shinkevich
Reviewed-by: Alberto Garcia
---
block/stream.c | 13
01.05.2019 21:13, Alberto Garcia wrote:
> There's only one bdrv_write() call left in the qcow2 code, and it can
> be trivially replaced with the byte-based bdrv_pwrite().
>
> Signed-off-by: Alberto Garcia
Reviewed-by: Vladimir Sementsov-Ogievskiy
--
Best regards,
Vladimir
Hi Axel,
can you link us to your test code, such that we can try to reproduce it.
Cheers,
Bastian
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1826568
Title:
RISC-V Disassembler/translator
233 generally filters the port, but in two cases does not. If some
other concurrently running application has already taken port 10809,
this will result in an output mismatch. Fix this by applying the
filter in these two cases, too.
Signed-off-by: Max Reitz
---
tests/qemu-iotests/233 | 6
On 05/05/19 04:06, Peter Xu wrote:
>> I wonder if it's possible to untangle this and make the irqchip
>> option stop affecting guest ABI on 4.1+ machine-types? This way
>> QEMU could choose smarter defaults in the future without the
>> compatibility code hassle.
> Hi, Eduardo,
>
> Do you mean to
On 5/6/19 11:05 AM, Max Reitz wrote:
> 233 generally filters the port, but in two cases does not. If some
> other concurrently running application has already taken port 10809,
> this will result in an output mismatch. Fix this by applying the
> filter in these two cases, too.
>
>
On 03/05/19 15:00, Michael S. Tsirkin wrote:
> On Mon, Apr 29, 2019 at 11:55:56AM -0300, Eduardo Habkost wrote:
>> irqchip=split and irqchip=kernel aren't guest ABI compatible, are
>> they?
>
> Can you remind me why they aren't?
They are compatible if the userspace IOAPIC is configured with
There is an analogous change for ARM here:
https://patchwork.kernel.org/patch/10649857
---
target/riscv/csr.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/target/riscv/csr.c b/target/riscv/csr.c
index 6083c782a1..1ec1222da1 100644
--- a/target/riscv/csr.c
+++
On 01.05.19 15:41, Thomas Huth wrote:
> The output of qemu-io changed recently - most tests have been fixed in
> commit 36b9986b08787019ef42 ("tests/qemu-iotests: Fix output of qemu-io
> related tests") already, but the qcow1 and the vmdk test was still missing.
>
> Fixes: 99e98d7c9fc1a1639fad
This series introduces a bottom intermediate node that eliminates the
dependency on the base that may change while stream job is running.
It happens when stream/commit parallel jobs are running on the same
backing chain. The base node of the stream job may be a top node of
the parallel commit job
From: Andrey Shinkevich
This patch is used in the 'block/stream: introduce a bottom node'
that is following. Instead of the base node, the caller may pass
the node that has the base as its backing image to the function
bdrv_is_allocated_above() with a new parameter include_base = true
and get
From: Andrey Shinkevich
The bottom node is the intermediate block device that has the base as its
backing image. It is used instead of the base node while a block stream
job is running to avoid dependency on the base that may change due to the
parallel jobs. The change may take place due to a
Argh, meant to include a signed off by line:
Signed-off-by: Jonathan Behrens
On Mon, May 6, 2019 at 11:31 AM Jonathan Behrens wrote:
> There is an analogous change for ARM here:
> https://patchwork.kernel.org/patch/10649857
> ---
> target/riscv/csr.c | 4 +++-
> 1 file changed, 3
According to the spec, "All bits besides SSIP, USIP, and UEIP in the sip
register are read-only." Further, if an interrupt is not delegated to mode
x,
then "the corresponding bits in xip [...] should appear to be hardwired to
zero. This patch implements both of those requirements.
Signed-off-by:
01.05.2019 21:13, Alberto Garcia wrote:
> The last user of this field disappeared when we replace the
> sector-based bdrv_write() with the byte-based bdrv_pwrite().
>
> Signed-off-by: Alberto Garcia
Reviewed-by: Vladimir Sementsov-Ogievskiy
--
Best regards,
Vladimir
Never mind, 0ec7e6779fc830e5b4e6a448d75317fafcf69477 fixed this.
This can be closed.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1827871
Title:
Race condition when rebooting with the TCG
** Changed in: qemu
Status: New => Fix Released
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1827871
Title:
Race condition when rebooting with the TCG backend
Status in QEMU:
Fix
Fixed in commit range 8482ff2eb3bb..24496b8d27d9.
** Changed in: qemu
Status: In Progress => Fix Committed
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1821884
Title:
Extend
On Mon, May 06, 2019 at 10:09:29AM -0500, Eric Blake wrote:
> On 5/5/19 5:55 AM, Yuval Shaia wrote:
> > Any GID change in guest must be propogate to host. This is already done
>
> s/propogate to/propagated to the/
>
> > by firing QMP event to managment system such as libvirt which in turn
>
>
FYI: we've hit this as will with syzkaller testing; this is still
reproducible as-is with latest qemu (commit a6ae238), and the latest
Linux kernel (5.1-rc7).
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
01.05.2019 21:13, Alberto Garcia wrote:
> There's only a couple of bdrv_read() and bdrv_write() calls left in
> the vdi code, and they can be trivially replaced with the byte-based
> bdrv_pread() and bdrv_pwrite().
>
> Signed-off-by: Alberto Garcia
> ---
> block/vdi.c | 15 +--
>
Any GID change in guest must be propagated to the host. This is already
done by firing QMP event to management system such as libvirt which in
turn will update the host with the relevant change.
When qemu is executed on non-qmp framework (ex from command-line) we
need to update the host instead.
01.05.2019 21:13, Alberto Garcia wrote:
> There's only a couple of bdrv_read() and bdrv_write() calls left in
> the vvfat code, and they can be trivially replaced with the byte-based
> bdrv_pread() and bdrv_pwrite().
>
> Signed-off-by: Alberto Garcia
> ---
> block/vvfat.c | 12 +++-
>
On 5/6/19 10:32 AM, Max Reitz wrote:
> On 01.05.19 15:41, Thomas Huth wrote:
>> The output of qemu-io changed recently - most tests have been fixed in
>> commit 36b9986b08787019ef42 ("tests/qemu-iotests: Fix output of qemu-io
>> related tests") already, but the qcow1 and the vmdk test was still
01.05.2019 21:13, Alberto Garcia wrote:
> No one is using these functions anymore, all callers have switched to
> the byte-based bdrv_pread() and bdrv_pwrite()
>
> Signed-off-by: Alberto Garcia
Reviewed-by: Vladimir Sementsov-Ogievskiy
However, I doubt that we want to save error codes
Please let me know if further work or another patch submission is
required.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1814352
Title:
SIOCGIFNAME takes a struct ifreq not an integer
Status in
On 4/29/19 5:57 AM, Peter Maydell wrote:
> gcc: error: /usr/lib/x86_64-linux-gnu/libunistring.so: No such file or
> directory
>
> which is almost certainly because we're now including
> a bunch of extra libraries in the link for the static
> linux-user binaries which don't work statically
On 5/6/19 11:47 AM, Vladimir Sementsov-Ogievskiy wrote:
> 01.05.2019 21:13, Alberto Garcia wrote:
>> There's only a couple of bdrv_read() and bdrv_write() calls left in
>> the vvfat code, and they can be trivially replaced with the byte-based
>> bdrv_pread() and bdrv_pwrite().
>>
>> Signed-off-by:
05.05.2019 18:54, Thomas Huth wrote:
> On 03/05/2019 18.15, Alex Bennée wrote:
>>
>> Thomas Huth writes:
>>
>>> On 03/05/2019 16.39, Alex Bennée wrote:
This attempts to clean-up the output to better match the output of the
rest of the QEMU check system. This includes:
-
If a ccw has CCW_FLAG_SKIP set, and the command is of type
read, read backwards, or sense, no data should be written
to the guest for that command.
Signed-off-by: Cornelia Huck
---
Only extremely lightly tested (i.e., can boot a guest.)
---
hw/s390x/css.c | 22 ++
Recently, a few bugs were reported that resulted from an inconsistent
state regarding AioContexts. Block nodes can end up in different
contexts than their users expect - the AioContext of a node can even
change under the feet of a device with no way for the device to forbid
this. We recently added
Eventually, we want to make sure that all parents and all children of a
node are in the same AioContext as the node itself. This means that
changing the AioContext may fail because one of the other involved
parties (e.g. a guest device that was configured with an iothread)
cannot allow switching
Signed-off-by: Kevin Wolf
---
tests/test-block-iothread.c | 71 +
1 file changed, 71 insertions(+)
diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c
index 938831c9bd..59f692892e 100644
--- a/tests/test-block-iothread.c
+++
Signed-off-by: Kevin Wolf
---
tests/test-block-iothread.c | 131
1 file changed, 131 insertions(+)
diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c
index 036ed9a3b3..938831c9bd 100644
--- a/tests/test-block-iothread.c
+++
Instead of having two recursions, in bdrv_attach_aio_context() and in
bdrv_detach_aio_context(), just having one recursion is enough. Said
functions are only about a single node now.
It is important that the recursion doesn't happen between detaching and
attaching a context to the current node
06.05.2019 20:06, Eric Blake wrote:
> On 5/6/19 11:47 AM, Vladimir Sementsov-Ogievskiy wrote:
>> 01.05.2019 21:13, Alberto Garcia wrote:
>>> There's only a couple of bdrv_read() and bdrv_write() calls left in
>>> the vvfat code, and they can be trivially replaced with the byte-based
>>>
All block nodes and users in any connected component of the block graph
must be in the same AioContext, so changing the AioContext of one node
must not only change all of its children, but all of its parents (and
in turn their children etc.) as well.
Signed-off-by: Kevin Wolf
---
Block jobs require that all of the nodes the job is using are in the
same AioContext. Therefore all BdrvChild objects of the job propagate
.(can_)set_aio_context to all other job nodes, so that the switch is
checked and performed consistently even if both nodes are in different
subtrees.
Some users (like block jobs) can tolerate an AioContext change for their
BlockBackend. Add a function that tells the BlockBackend that it can
allow changes.
Signed-off-by: Kevin Wolf
---
include/sysemu/block-backend.h | 1 +
block/block-backend.c | 10 ++
2 files changed, 11
Since commit b97511c7bc8, there is no reason for block drivers any more
to call these functions (see the function comment in block_int.h). They
are now just internal helper functions for bdrv_set_aio_context()
and can be made static.
Signed-off-by: Kevin Wolf
---
include/block/block_int.h | 21
On 5/3/19 3:54 PM, Eric Blake wrote:
>>> Ugh. Hole granularities are file-system specific, so we need to figure
>>> out some way to fuzz the input. It might also be possible to pick nicer
>>> size numbers - perhaps if the test image is sized at 64k+1 instead of
>>> 43009 (84*512, but NOT evenly
The notifiers made sure that the job is quiesced and that the
job->aio_context field is updated. The first part is unnecessary today
since bdrv_set_aio_context_ignore() drains the block node, and this
means drainig the block job, too. The second part can be done in the
.set_aio_ctx callback of the
bdrv_try_set_aio_context() currently fails if a BlockBackend is attached
to a node because it doesn't implement the BdrvChildRole callbacks for
AioContext management.
We can allow changing the AioContext of monitor-owned BlockBackends as
long as no device is attached to them.
When setting the
At least ubuntu 18.04 does not package static gnutls libraries.
Signed-off-by: Richard Henderson
---
configure | 19 +--
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/configure b/configure
index 5b183c2e39..daadfe9ad4 100755
--- a/configure
+++ b/configure
@@
On 5/5/19 10:54 AM, Thomas Huth wrote:
>>> That's why I thought that having a TAP mode for the check script could
>>> be a good idea, too. Then we could pipe the output through the
>>> tap-driver.pl script, too, so we get uniform output for all tests...?
>>
>> That would probably be a cleaner
On 5/3/19 3:27 AM, Antonio Ospite wrote:
> From: Antonio Ospite
>
> The configure script breaks when the qemu source directory is in a path
> containing white spaces, in particular the list of targets is not
> correctly generated when calling "./configure --help" because of how the
>
Use #ifdef _WIN32 instead of #ifndef _WIN32.
This will make other tests easier to sequence.
Reviewed-by: Philippe Mathieu-Daudé
Reviewed-by: Daniel P. Berrangé
Signed-off-by: Richard Henderson
---
crypto/random-platform.c | 35 +--
1 file changed, 17
On 5/3/19 3:27 AM, Antonio Ospite wrote:
> From: Antonio Ospite
>
> Since commit 79d77bcd36 (configure: Remove --source-path option,
> 2019-04-29) source_path cannot be overridden anymore, move it out of the
> "default parameters" block since the word "default" may suggest that the
> value can
For some particular configurations of ext4, sizing an image to 84
sectors + 1 byte causes test failures when the size of the hole is
rounded to a 4k alignment. Let's instead size things to 128 sectors +
1 byte, as the 64k boundary is more likely to work with various hole
granularities.
We will shortly need this in the user-only binaries, so drop the split
into system and tools binaries. This also means that crypto-aes-obj-y
can be merged back into crypto-obj-y.
Cc: Daniel P. Berrangé
Signed-off-by: Richard Henderson
---
Makefile | 12 +---
Makefile.objs
Changes since v3:
* Do not autoenable gnutls if linking is broken.
Fixes --static on ubuntu 18.04.
Changes since v2:
* Changes from review.
- getrandom is not exclusive of /dev/urandom fallback.
- vnc fails gracefully on crypto failure.
- a great renaming.
* Drop the
There were 3 copies of this code, one of which used the wrong
data size for the failure indicator.
Reviewed-by: Philippe Mathieu-Daudé
Reviewed-by: Gerd Hoffmann
Reviewed-by: Daniel P. Berrangé
Signed-off-by: Richard Henderson
---
ui/vnc.c | 37 +++--
1 file
When the -seed option is given, call qemu_guest_random_seed_main,
putting the subsystem into deterministic mode. Pass derived seeds
to each cpu created during clone; which is a no-op unless the
subsystem is in deterministic mode.
Cc: Laurent Vivier
Reviewed-by: Philippe Mathieu-Daudé
We can always get EINTR for read; /dev/urandom is no exception.
Rearrange the order of tests for likelihood; allow degenerate buflen==0
case to perform a no-op zero-length read. This means that the normal
success path is a straight line with a single test for success.
Reviewed-by: Daniel P.
Prefer it to direct use of /dev/urandom.
Reviewed-by: Daniel P. Berrangé
Signed-off-by: Richard Henderson
---
v3: If getrandom is not present, fall back on /dev/(u)random.
---
crypto/random-platform.c | 37 -
configure| 18 +-
201 - 270 of 270 matches
Mail list logo