*** This bug is a duplicate of bug 1450881 ***
https://bugs.launchpad.net/bugs/1450881
** This bug has been marked a duplicate of bug 1450881
qemu-system-sparc MUTEX_HELD assert and libC lock errors
--
You received this bug notification because you are a member of qemu-
devel-ml, which is
Fix is included in the 2.6.0-rc2 release.
** 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/1450881
Title:
qemu-system-sparc MUTEX_HELD assert an
On Thu, Apr 07, 2016 at 04:23:12PM +0100, Mark Cave-Ayland wrote:
> Using a new DEFINE_PPC_MACHINE macro, make sure that all PPC machines now
> derive from
> the new PPCMachineClass.
> Signed-off-by: Mark Cave-Ayland
Hmm.. I'm a little bit dubious about this, since the timebase sync
will be pre
On Thu, Apr 07, 2016 at 04:23:11PM +0100, Mark Cave-Ayland wrote:
> Introduce PPCMachineClass in anticipation of making it the superclass for
> all PPC machines.
>
> Signed-off-by: Mark Cave-Ayland
> ---
> hw/ppc/ppc.c| 16
> include/hw/ppc/ppc.h| 16
On Thu, Apr 07, 2016 at 04:23:10PM +0100, Mark Cave-Ayland wrote:
> This RFC series follows on from discussions in the "Migrating decrementer"
> thread relating to handling migration of the timebase and decrementer
> registers for both TCG and KVM. The aim is to provide a consistent virtual
> clock
On Thu, Apr 07, 2016 at 04:23:13PM +0100, Mark Cave-Ayland wrote:
> Recalculate the tb_offset between the guest and host, applying it to all CPUs
> when (re)starting the virtual machine. This has the effect of providing a
> near-seamless virtual timebase for KVM guests that support
> KVM_REG_PPC_TB
From: Wen Congyang
Some programs that add a dependency on it will use
the block layer directly.
Signed-off-by: Wen Congyang
Signed-off-by: zhanghailiang
Signed-off-by: Gonglei
Signed-off-by: Changlong Xie
Reviewed-by: Stefan Hajnoczi
Reviewed-by: Jeff Cody
---
block/Makefile.objs | 2 +-
From: Wen Congyang
Auto complete mirror job in background to prevent from
blocking synchronously
Signed-off-by: Wen Congyang
Signed-off-by: Changlong Xie
---
block/mirror.c| 13 +
blockdev.c| 2 +-
include/block/block_int.h | 3 ++-
qemu-img.c
From: Wen Congyang
Signed-off-by: Wen Congyang
Signed-off-by: zhanghailiang
Signed-off-by: Gonglei
Signed-off-by: Changlong Xie
---
block/backup.c| 17 +
include/block/block_int.h | 2 ++
2 files changed, 19 insertions(+)
diff --git a/block/backup.c b/block/back
From: Wen Congyang
Signed-off-by: Wen Congyang
Signed-off-by: Changlong Xie
---
block.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/block.c b/block.c
index 4bdc6b3..1e5a4fd 100644
--- a/block.c
+++ b/block.c
@@ -1255,6 +1255,23 @@ void bdrv_set_backing_hd(BlockDriver
Block replication is a very important feature which is used for
continuous checkpoints(for example: COLO).
You can get the detailed information about block replication from here:
http://wiki.qemu.org/Features/BlockReplication
Usage:
Please refer to docs/block-replication.txt
This patch series is
From: Wen Congyang
Signed-off-by: Wen Congyang
Signed-off-by: zhanghailiang
Signed-off-by: Gonglei
Signed-off-by: Changlong Xie
---
docs/block-replication.txt | 239 +
1 file changed, 239 insertions(+)
create mode 100644 docs/block-replication.txt
From: Wen Congyang
Signed-off-by: Wen Congyang
Signed-off-by: zhanghailiang
Signed-off-by: Gonglei
Signed-off-by: Changlong Xie
---
block/Makefile.objs | 1 +
block/replication.c | 625 +++
tests/.gitignore | 1 +
tests/Makefi
From: Wen Congyang
Signed-off-by: Wen Congyang
Signed-off-by: zhanghailiang
Signed-off-by: Gonglei
Signed-off-by: Changlong Xie
Reviewed-by: Eric Blake
---
qapi/block-core.json | 20 ++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/qapi/block-core.json b/q
Signed-off-by: Wen Congyang
Signed-off-by: zhanghailiang
Signed-off-by: Gonglei
Signed-off-by: Changlong Xie
---
Makefile.objs| 1 +
qapi/block-core.json | 13
replication.c| 105 ++
replication.h| 176
Am 15.04.2016 um 01:14 hat Max Reitz geschrieben:
> On 13.04.2016 13:06, Kevin Wolf wrote:
> > This allows testing blk_aio_write_zeroes().
> >
> > Signed-off-by: Kevin Wolf
> > ---
> > qemu-io-cmds.c | 64
> > +++---
> > 1 file changed, 48 ins
Commit 57d6a428 neglected to pass the given flags to blk_aio_prwv(),
which broke discard by WRITE SAME for scsi-disk (the UNMAP bit would be
ignored).
Reported-by: Max Reitz
Signed-off-by: Kevin Wolf
---
block/block-backend.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a
Eric Blake writes:
> Implement the new type_null() callback for the qmp input and
> output visitors. While we don't yet have a use for this in QAPI
> input (the generator will need some tweaks first), one usage
> is already envisioned: when changing blockdev parameters, it
> would be nice to have
Am 15.04.2016 um 01:30 hat Max Reitz geschrieben:
> On 13.04.2016 13:06, Kevin Wolf wrote:
> > Commit 57d6a428 broke blk_aio_write_zeroes() because in some write
> > functions in the call path don't have an explicit length argument but
> > reuse qiov->size instead. Which is great, except that write
Hi Peter,
This contains some register corruption fixes for qemu-system-sparc64. Please
pull.
ATB,
Mark.
The following changes since commit bc8995cafa0c36b9e4be682e9a60d59484b33500:
Update version for v2.6.0-rc2 release (2016-04-14 17:30:28 +0100)
are available in the git repository at:
Agreed. This looks much better, and is simpler and faster. Though, it looks to
me that "fst" should be called "status" instead (just not to break consistency
throughout SoftFloat library and in other places of assigning name "status" to
the last argument of any function if its type is float_stat
env->xer only holds the lower bits of the XER register nowadays, the
SO, OV and CA bits are stored in separate variables (see the function
cpu_write_xer() for details). Since the migration code currently only
reads the "xer" variable, the upper bits are lost during migration.
Fix it by using cpu_re
Eric Blake writes:
> Add a new qmp_output_visitor_reset(), which must be called before
> reusing an exising QmpOutputVisitor on a new root object. Tighten
> assertions to require that qmp_output_get_qobject() can only be
> called after pairing a visit_end_* for every visit_start_* (rather
> than
On Thu, Apr 14, 2016 at 07:46:17PM +0200, Stefan Weil wrote:
> It is broken since commit c619644067f98098dcdbc951e2dda79e97560afa.
>
> Reported-by: Michael Fritscher
> Signed-off-by: Stefan Weil
> ---
>
> Networking with QEMU for Windows is currently not usable,
> see bug report https://bugs.la
On Thu, Apr 14, 2016 at 08:12:00PM +0100, Peter Maydell wrote:
> On 14 April 2016 at 18:46, Stefan Weil wrote:
> > It is broken since commit c619644067f98098dcdbc951e2dda79e97560afa.
> >
> > Reported-by: Michael Fritscher
> > Signed-off-by: Stefan Weil
> > ---
> >
> > Networking with QEMU for Wi
Hello Samuel,
> @Ipv6: I'll test it tomorrow.
Interesting results:
On my native windows builds with msys2 and mingw64, a time wget
http://www.heise.de gives me (tested under native Win7)
real 0m24.081s
user 0m2.313s
sys 0m9.620s
or
real 0m52.250s
user 0m3.647s
sys 0m21.117s
The results are qui
On 12/04/2016 10:20, Markus Armbruster wrote:
> QOM replaced the static DeviceInfo. The static qdev property arrays
> remain, but they're now stored into their DeviceClass's member props
> dynamically, by the TypeInfo's class_init() method.
>
> The qdev properties so stored get mapped to QOM pr
On 15/04/2016 08:43, Gerd Hoffmann wrote:
> This reverts commit 7070e085d490c396f9237c8f10bf8b6e69cd0066.
>
> Commit message claims locking is not needed, but that appears
> to not be true, seabios ehci driver runs into timekeeping problems
> with this, see
> https://bugzilla.redhat.com/sh
On 15/04/16 10:03, Thomas Huth wrote:
> env->xer only holds the lower bits of the XER register nowadays, the
> SO, OV and CA bits are stored in separate variables (see the function
> cpu_write_xer() for details). Since the migration code currently only
> reads the "xer" variable, the upper bits ar
On 14/04/16 16:14, Thomas Huth wrote:
> These two patches fix the bad range checks in the LSWI and LSWX
> instructions.
>
> To see the change in behavior for the lswx instruction, you can use the
> "emulator" test from the kvm-unit-tests suite - code can be found here:
>
> https://git.kernel.org
On 15 April 2016 at 09:40, Mark Cave-Ayland
wrote:
> Hi Peter,
>
> This contains some register corruption fixes for qemu-system-sparc64. Please
> pull.
>
>
> ATB,
>
> Mark.
>
>
> The following changes since commit bc8995cafa0c36b9e4be682e9a60d59484b33500:
>
> Update version for v2.6.0-rc2 relea
Libvirt currently rejects using host /dev/urandom as an input source for a
virtio-rng device. The only accepted sources are /dev/random and /dev/hwrng.
This is the result of discussions on qemu-devel around when the feature was
first added (2013). Examples:
http://lists.gnu.org/archive/html/qemu-d
On 15/04/16 09:48, Aleksandar Markovic wrote:
> Agreed. This looks much better, and is simpler and faster. Though, it looks
> to me that "fst" should be called "status" instead (just not to break
> consistency throughout SoftFloat library and in other places of assigning
> name "status" to the l
On 14/04/2016 17:30, Peter Maydell wrote:
> On 4 April 2016 at 11:40, Paolo Bonzini wrote:
>>
>>
>> On 01/04/2016 19:58, Dr. David Alan Gilbert wrote:
>>> In the continuing journeys of trying to migrate a q35 guest with ovmf,
>>> I've just hit this assert:
>>>
>>> qemu-system-x86_64: /root/git/q
Eric Blake writes:
> As mentioned in previous patches, we want to call visit_end_struct()
> functions unconditionally, so that visitors can release resources
> tied up since the matching visit_start_struct() without also having
> to worry about error priority if more than one error occurs.
>
> Ev
On 04/15/2016 06:06 PM, Paolo Bonzini wrote:
Here's my try rewording the comment (sticking to GTK-Doc conventions
even though I dislike them):
/**
* @qdev_property_add_static:
* @dev: Device to add the property to
* @prop: The qdev property definition
* @err: location to store erro
CoQueue do not need to remove any element but the head of the list;
processing is always strictly FIFO. Therefore, the simpler singly-linked
QSIMPLEQ can be used instead.
Signed-off-by: Paolo Bonzini
---
include/qemu/coroutine.h | 2 +-
include/qemu/coroutine_int.h | 4 ++--
util/qemu-cor
Signed-off-by: Paolo Bonzini
---
include/qemu/coroutine.h | 5 -
util/qemu-coroutine-lock.c | 14 --
2 files changed, 19 deletions(-)
diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index 63ae7fe..bb23be0 100644
--- a/include/qemu/coroutine.h
+++ b/include/qem
Signed-off-by: Paolo Bonzini
---
include/block/aio.h| 5 +
include/qemu/coroutine.h | 18 ++
iothread.c | 16
stubs/iothread.c | 11 +++
tests/iothread.c | 16
trace-events |
This uses the lock-free mutex described in the paper '"Blocking without
Locking", or LFTHREADS: A lock-free thread library' by Gidenstam and
Papatriantafilou. The same technique is used in OSv, and in fact
the code is essentially a conversion to C of OSv's code.
Signed-off-by: Paolo Bonzini
---
[this time including the mailing list]
This is yet another tiny bit of the multiqueue work, this time affecting
the synchronization infrastructure for coroutines. Currently, coroutines
synchronize between the main I/O thread and the dataplane iothread through
the AioContext lock. However, for mu
Signed-off-by: Paolo Bonzini
---
block/throttle-groups.c | 69 +++--
1 file changed, 49 insertions(+), 20 deletions(-)
diff --git a/block/throttle-groups.c b/block/throttle-groups.c
index 1938e90..53e910e 100644
--- a/block/throttle-groups.c
+++ b/bloc
Running a very small critical section on pthread_mutex_t and CoMutex
shows that pthread_mutex_t is much faster because it doesn't actually
go to sleep. What happens is that the critical section is shorter
than the latency of entering the kernel and thus FUTEX_WAIT always
fails. With CoMutex there
This provides the infrastructure to start a coroutine on a remote
AioContext. It will be used by CoMutex and CoQueue, so that
coroutines don't jump from one context to another when they
go to sleep on a mutex or waitqueue.
aio_co_schedule is based on a lock-free multiple-producer,
single-consumer
This will avoid forward references in the next patch. It is also
more logical because CoQueue is not anymore the basic primitive.
Signed-off-by: Paolo Bonzini
---
include/qemu/coroutine.h | 79
1 file changed, 39 insertions(+), 40 deletions(-)
d
Add two implementations of the same benchmark as the previous patch,
but using pthreads. One uses a normal QemuMutex, the other is Linux
only and implements a fair mutex based on MCS locks and futexes.
This shows that the slower performance of the 5-thread case is due to
the fairness of CoMutex, r
Hi, Leon,
There is one more subtle point here. The question is: Where to put declarations
of two new functions? This is a new case in the organization of mips-specific
source code files. File helpers.h does not seem to be a good place, cpu.h -
maybe, but still looks clumsy to me.
IMHO, my solu
All that CoQueue needs in order to become thread-safe is help
from an external mutex. Add this to the API.
Signed-off-by: Paolo Bonzini
---
block/backup.c | 2 +-
block/io.c | 2 +-
block/qcow2-cluster.c | 4 +---
block/sheepdog.c | 2 +-
block/thr
This adds a CoMutex around the existing CoQueue. Because the write-side
can just take CoMutex, the old "writer" field is not necessary anymore.
Instead of removing it altogether, count the number of pending writers
during a read-side critical section and forbid further readers from
entering.
Sign
From: "Dr. David Alan Gilbert"
Pflash migration (e.g. q35 + EFI variable storage) fails
with the assert:
bdrv_co_do_pwritev: Assertion `!(bs->open_flags & 0x0800)' failed.
This avoids the problem by delaying the pflash update until after
the device loads complete.
Tested by:
Migrating Q35/EF
Eric Blake writes:
> The qmp-input visitor was playing rather fast and loose: when
I guess (some of) its *users* are playing fast and loose, and the
visitor itself lets them. The patch's title suggests "some of its
users" == qapi-commands.py.
> visiting a QDict, you could grab members of the r
On Fri, Apr 15, 2016 at 06:41:34AM -0400, Cole Robinson wrote:
> Libvirt currently rejects using host /dev/urandom as an input source for a
> virtio-rng device. The only accepted sources are /dev/random and /dev/hwrng.
> This is the result of discussions on qemu-devel around when the feature was
>
Eric Blake writes:
> The qmp-input visitor was playing rather fast and loose: when
> visiting a QDict, you could grab members of the root dictionary
> without first pushing into the dict. But we are about to tighten
> the input visitor, at which point user_creatable_add_type() MUST
> follow the
On Fri, Apr 15, 2016 at 12:46:46PM +0100, Richard W.M. Jones wrote:
> On Fri, Apr 15, 2016 at 06:41:34AM -0400, Cole Robinson wrote:
> > Libvirt currently rejects using host /dev/urandom as an input source for a
> > virtio-rng device. The only accepted sources are /dev/random and /dev/hwrng.
> > Th
Hi Jevon,
More findings in fsdev/qemu-fsdev.c and hw/9pfs/Makefile.objs, see below.
Cheers.
--
Greg
On Tue, 15 Mar 2016 00:02:48 +0800
Jevon Qiao wrote:
> Ceph as a promising unified distributed storage system is widely used in the
> world of OpenStack. OpenStack users deploying Ceph for blo
On 15.04.2016 12:17, Mark Cave-Ayland wrote:
> On 15/04/16 10:03, Thomas Huth wrote:
>
>> env->xer only holds the lower bits of the XER register nowadays, the
>> SO, OV and CA bits are stored in separate variables (see the function
>> cpu_write_xer() for details). Since the migration code currentl
On 04/14/2016 06:19 PM, Gerd Hoffmann wrote:
> On Do, 2016-04-14 at 17:29 +0200, Alexander Graf wrote:
>> On 04/14/2016 05:17 PM, Gerd Hoffmann wrote:
>>> On Do, 2016-04-14 at 16:25 +0200, Alexander Graf wrote:
We can currently buffer up to 16 events in our event queue for input
devices.
On 04/15/16 13:41, Dr. David Alan Gilbert (git) wrote:
> From: "Dr. David Alan Gilbert"
>
> Pflash migration (e.g. q35 + EFI variable storage) fails
> with the assert:
>
> bdrv_co_do_pwritev: Assertion `!(bs->open_flags & 0x0800)' failed.
>
> This avoids the problem by delaying the pflash updat
On Fri, Apr 15, 2016 at 11:27:56AM +0800, Fam Zheng wrote:
> Signed-off-by: Fam Zheng
> ---
> block/gluster.c | 30 ++
> 1 file changed, 30 insertions(+)
The Gluster changes look good to me.
Reviewed-by: Niels de Vos
>
> diff --git a/block/gluster.c b/block/gluste
Eric Blake writes:
> Don't embed the root of the visit into the stack of current
> containers being visited. That way, we no longer get confused
I got confused pretty much every time I looked at this code...
> on whether the first visit of a dictionary is to the dictionary
> itself or to one o
One more nit in configure. See below.
Cheers.
--
Greg
On Tue, 15 Mar 2016 00:02:48 +0800
Jevon Qiao wrote:
> Ceph as a promising unified distributed storage system is widely used in the
> world of OpenStack. OpenStack users deploying Ceph for block (Cinder) and
> object (S3/Swift) are unsurpri
Commit 43dc2a64 replaced assert() with abort(), but didnt remove statements
that followed these calls. So current code still has return values set after
a call to abort(). Such statements will never execute and need to be cleaned
up.
Signed-off-by: Prerna Saxena
---
block/vvfat.c | 17 +++---
Today, some calls to abort() do not have a preceding error string that
might hint to the end user why QEMU died. Debugging such scenarios is
painful today.
This patchset attempts to clean up some dead code in vvfat.c;
it also aims to improve qemu error-reporting by placing error messages
that prec
Qemu code has abort() calls in various places which raises a SIGABRT;
This patch adds error messages before (most)calls to abort(), so that
it is easier to determine why QEMU died.
Signed-off-by: Prerna Saxena
---
block.c| 1 +
block/block-backend.c | 4
block/curl.c
On Thu, Apr 14, 2016 at 12:22:30PM -0400, Andrea Arcangeli wrote:
> Adding linux-mm too,
>
> On Thu, Apr 14, 2016 at 01:34:41PM +0100, Dr. David Alan Gilbert wrote:
> > * Andrea Arcangeli (aarca...@redhat.com) wrote:
> >
> > > The next suspect is the massive THP refcounting change that went
> > >
* Kirill A. Shutemov (kir...@shutemov.name) wrote:
> On Thu, Apr 14, 2016 at 12:22:30PM -0400, Andrea Arcangeli wrote:
> > Adding linux-mm too,
> >
> > On Thu, Apr 14, 2016 at 01:34:41PM +0100, Dr. David Alan Gilbert wrote:
> > > * Andrea Arcangeli (aarca...@redhat.com) wrote:
> > >
> > > > The n
On 15/04/2016 08:56, Markus Armbruster wrote:
> Paolo Bonzini writes:
>
>> On 14/04/2016 17:44, Igor Mammedov wrote:
>>> On Thu, 14 Apr 2016 15:24:10 +0200
>>> Paolo Bonzini wrote:
>>>
On 14/04/2016 13:34, Igor Mammedov wrote:
> On Wed, 13 Apr 2016 18:57:40 +0200
> marcandre.lur..
On 04/15/2016 02:27 AM, Kevin Wolf wrote:
> Commit 57d6a428 neglected to pass the given flags to blk_aio_prwv(),
> which broke discard by WRITE SAME for scsi-disk (the UNMAP bit would be
> ignored).
>
> Reported-by: Max Reitz
> Signed-off-by: Kevin Wolf
> ---
> block/block-backend.c | 3 ++-
>
Eric Blake writes:
> We have two uses of list visits in the entire code base: one in
> spapr_drc (which completely avoids visit_next_list(), feeding in
> integers from a different source than uint8List), and one in
> qapi-visit.py
Period here, convert the parenthesis to a sentence.
>
Hi,
This series finally completes the re-build of Fred's multi_tcg_v8 tree
by enabling MTTCG for armv7 guests on x86 hosts. This applies on top
of the previous series:
- [RFC v2 00/11] Base enabling patches for MTTCG
You can find the final tree at:
https://github.com/stsquad/qemu/tree/mttcg
From: Sergey Fedorov
Ensure direct jump patching in i386 is atomic by:
* naturally aligning a location of direct jump address;
* using atomic_read()/atomic_set() for code patching.
Signed-off-by: Sergey Fedorov
Signed-off-by: Sergey Fedorov
---
include/exec/exec-all.h | 2 +-
tcg/i386/tc
Useful for counting down.
Signed-off-by: Alex Bennée
---
include/qemu/atomic.h | 4
1 file changed, 4 insertions(+)
diff --git a/include/qemu/atomic.h b/include/qemu/atomic.h
index 8f1d8d9..5dba7db 100644
--- a/include/qemu/atomic.h
+++ b/include/qemu/atomic.h
@@ -131,6 +131,8 @@
#define
This is a temporary squashed patch of the following patches from
Emilio's qht series. This includes:
- include/processor.h: define cpu_relax()
- qemu-thread: add simple test-and-set spinlock
- qemu-thread: call cpu_relax() while spinning
- qemu-thread: optimize spin_lock for uncontended lo
From: KONRAD Frederic
This mechanism replaces the existing load/store exclusive mechanism which seems
to be broken for MTTCG. It follows the intention of the existing
mechanism and stores the target address and data values during a load
operation and checks that they remain unchanged before a sto
From: Paolo Bonzini
Signed-off-by: Paolo Bonzini
---
include/qemu-common.h | 24
include/qemu/timer.h | 1 +
include/qom/cpu.h | 9 +
include/sysemu/cpus.h | 13 +
stubs/cpu-get-icount.c | 1 +
translate-common.c | 1 +
vl.c
This allows for slightly neater code when checking for success of a
cmpxchg operation.
Signed-off-by: Alex Bennée
---
include/qemu/atomic.h | 11 +++
1 file changed, 11 insertions(+)
diff --git a/include/qemu/atomic.h b/include/qemu/atomic.h
index 5dba7db..94e7110 100644
--- a/include/q
Hi,
The finally completes the re-build of Fred's multi_tcg_v8 tree by
enabling MTTCG for armv7 guests on x86 hosts. This applies on top of
the previous series:
- [RFC v2 00/11] Base enabling patches for MTTCG
You can find the final tree at:
https://github.com/stsquad/qemu/tree/mttcg/enable-
On Fri, 2016-15-04 at 02:06:13 UTC, Unknown sender due to SPF wrote:
> The real LE feature entry in the ibm_pa_feature struct has the
> wrong number of elements. Instead of checking for byte 5, bit 0,
> we check for byte 0, bit 0, and we also incorrectly update cpu user
> feature bit 5.
>
> Fixes:
From: KONRAD Frederic
tb_flush is not thread safe we definitely need to exit VCPUs to do that.
This introduces tb_flush_safe which just creates an async safe work which will
do a tb_flush later. This is called when we run out of space.
Signed-off-by: KONRAD Frederic
[AJB: merge the various tb_f
CPUState is a fairly common pointer to pass to these helpers. This means
if you need other arguments for the async_run_on_cpu case you end up
having to do a g_malloc to stuff additional data into the routine. For
the current users this isn't a massive deal but for MTTCG this gets
cumbersome when th
From: KONRAD Frederic
We already had async_run_on_cpu but for some tasks we need to ensure all
vCPUs have stopped running before updating shared structures. We call
this safe work as it is safe to make the modifications.
Work is scheduled with async_safe_run_on_cpu() which is passed the
CPUState
From: KONRAD Frederic
Some architectures allow to flush the tlb of other VCPUs. This is not a problem
when we have only one thread for all VCPUs but it definitely needs to be an
asynchronous work when we are in true multithreaded work.
This patch doesn't do anything to protect other cputlb funct
This makes multi-threading the default for 32 bit ARM on x86. It has
been tested with Debian Jessie as well as my extended KVM unit tests
which stress the SMC and TB invalidation code. Those tests can be found
at:
https://github.com/stsquad/kvm-unit-tests/tree/mttcg/current-tests-v5
Signed-off-
From: KONRAD Frederic
Instead of flushing each individual vCPU use the tlb_flush_page_all
functions which is async enabled for MTTCG.
Signed-off-by: KONRAD Frederic
Signed-off-by: Alex Bennée
---
include/exec/exec-all.h | 3 +++
target-arm/helper.c | 12 ++--
2 files changed, 5 i
Am 15.04.2016 um 15:59 hat Eric Blake geschrieben:
> On 04/15/2016 02:27 AM, Kevin Wolf wrote:
> > Commit 57d6a428 neglected to pass the given flags to blk_aio_prwv(),
> > which broke discard by WRITE SAME for scsi-disk (the UNMAP bit would be
> > ignored).
> >
> > Reported-by: Max Reitz
> > Sign
Eric Blake writes:
> Returning a partial object on error is an invitation for a careless
> caller to leak memory. As no one outside the testsuite was actually
> relying on these semantics, it is cleaner to just document and
> guarantee that ALL pointer-based visit_type_FOO() functions always
> l
Commit 57d6a428 neglected to pass the given flags to blk_aio_prwv(),
which broke discard by WRITE SAME for scsi-disk (the UNMAP bit would be
ignored).
Commit fc1453cd introduced the same bug for blk_write_zeroes(). This is
used for 'qemu-img convert' without has_zero_init (e.g. on a block
device)
Eric Blake writes:
> On 04/13/2016 06:48 AM, Markus Armbruster wrote:
>> Eric Blake writes:
>>
>>> Rather than having two separate visitor callbacks with items
>>> already broken out, pass the actual QAPISchemaObjectType object
>>> to the visitor. This lets the visitor access things like
>>> t
Am 15.04.2016 um 05:11 hat Fam Zheng geschrieben:
> qcow2 is not necessarily the top layer node. Since bdrv_inactivate()
> doesn't recurse, we should ensure all block nodes are inactivated.
>
> Signed-off-by: Fam Zheng
Hm, I don't think that's quite right. bdrv_inactivate_all() should be
the opp
On 04/15/2016 08:52 AM, Kevin Wolf wrote:
> Commit 57d6a428 neglected to pass the given flags to blk_aio_prwv(),
> which broke discard by WRITE SAME for scsi-disk (the UNMAP bit would be
> ignored).
>
> Commit fc1453cd introduced the same bug for blk_write_zeroes(). This is
> used for 'qemu-img co
On Fri, Apr 15, 2016 at 02:42:33PM +0100, Dr. David Alan Gilbert wrote:
> * Kirill A. Shutemov (kir...@shutemov.name) wrote:
> > On Thu, Apr 14, 2016 at 12:22:30PM -0400, Andrea Arcangeli wrote:
> > > Adding linux-mm too,
> > >
> > > On Thu, Apr 14, 2016 at 01:34:41PM +0100, Dr. David Alan Gilbert
Eric Blake writes:
> On 04/13/2016 07:49 AM, Markus Armbruster wrote:
>> Eric Blake writes:
>>
>>> We have three classes of QAPI visitors: input, output, and dealloc.
>>> Currently, all implementations of these visitors have one thing in
>>> common based on their visitor type: the implementatio
Eric Blake writes:
> On 04/13/2016 09:53 AM, Markus Armbruster wrote:
>> Eric Blake writes:
>>
>>> Management of the top of stack was a bit verbose; creating a
>>> temporary variable and adding some comments makes the existing
>>> code more legible before the next few patches improve things.
>>
Am 15.04.2016 um 14:19 hat Laszlo Ersek geschrieben:
> On 04/15/16 13:41, Dr. David Alan Gilbert (git) wrote:
> > From: "Dr. David Alan Gilbert"
> >
> > Pflash migration (e.g. q35 + EFI variable storage) fails
> > with the assert:
> >
> > bdrv_co_do_pwritev: Assertion `!(bs->open_flags & 0x0800)
Eric Blake writes:
> On 04/13/2016 10:06 AM, Markus Armbruster wrote:
>> Eric Blake writes:
>>
>>> Commit e8316d7 mistakenly passed consume=true
>>
>> in qmp_input_optional(), right?
>
> yes
>
>>
>>> when checking if
>>> an optional member was pre
2016-04-15 11:31+0800, Peter Xu:
> This patch allows Intel IR work with splitted irqchip. Two more fields
> are added to IOAPICCommonState to support the translation process (For
> future AMD IR support, we will need to provide another AMD-specific
> callback for int_remap()). In split irqchip mode
Am 14.04.2016 um 13:08 hat Sascha Silbe geschrieben:
> Dear developers,
>
> Sascha Silbe writes:
>
> > The current rate limit implementation for block jobs is ineffective
> > below a certain minimum rate. It will permit writes at least once per
> > time slice. The resulting minimum write speed (
[Context restored]
"Michael S. Tsirkin" writes:
> On Thu, Apr 14, 2016 at 01:29:20PM +0200, Markus Armbruster wrote:
>> The next use case to consider is a user picking a new name for a new
>> interface between host and guest. I find the idea that such a user
>> won't notice warnings farfetched.
Eric Blake writes:
> This is now 2.7 material (it is too invasive for hard freeze).
Yes.
> Based on master, with no prerequisite patches.
I had quite a few comments, but I guess that's more because of me than
because of your patches :) Seriously, the patches feel close to ready.
Most of my co
On 04/15/2016 04:41 AM, Cole Robinson wrote:
> Libvirt currently rejects using host /dev/urandom as an input source for a
> virtio-rng device. The only accepted sources are /dev/random and /dev/hwrng.
> This is the result of discussions on qemu-devel around when the feature was
> first added (2013)
1 - 100 of 158 matches
Mail list logo
