Implement the code which updates the GIC state when an interrupt
input into the GIC is asserted.
Signed-off-by: Peter Maydell
Reviewed-by: Shannon Zhao
Tested-by: Shannon Zhao
Message-id:
On 17/06/2016 15:11, Efimov Vasily wrote:
> The isa_bus_irqs function initializes ISA bus IRQ array pointer with specified
> value.
>
> Previously the ICH9 LPC bridge model did not have its own IRQs but
> only IRQ pointer cache. And same GSI were used for ISA bus and other sources
> behind the
Some architectures require the stack to be executable; notably
this includes MIPS, because the kernel's floating point emulator
may try to put trampoline code on the stack to handle some cases.
(See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=815409
for an example of this causing QEMU to
From: Pavel Fedin
Add state information to GICv3 object structure and implement
arm_gicv3_common_reset().
This commit includes accessor functions for the fields which are
stored as bitmaps in uint32_t arrays.
Signed-off-by: Pavel Fedin
Reviewed-by:
On 17 June 2016 at 14:06, Amit Shah <amit.s...@redhat.com> wrote:
> The following changes since commit 4acc8fdfd315f7ee474bea28fcbcc4dca9717d13:
>
> Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.7-20160617'
> into staging (2016-06-17 12:36:27 +0100)
>
>
Moving trace_init_backends() into trace_opt_parse() is not possible. This
should be called after daemonize() in vl.c.
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric Blake
CC: Paolo Bonzini
CC: Stefan Hajnoczi
CC:
Implement the gicv3_cpuif_update() function which deals with correctly
asserting IRQ and FIQ based on the current running priority of the CPU,
the priority of the highest priority pending interrupt and the CPU's
current exception level and security state.
Signed-off-by: Peter Maydell
On 17/06/16 13:57, Artyom Tarasenko wrote:
> On Fri, Jun 17, 2016 at 2:44 PM, Mark Cave-Ayland
> wrote:
>> On 17/06/16 12:36, Artyom Tarasenko wrote:
>>
>>> Hi Mark,
>>>
>>> On Fri, Jun 17, 2016 at 1:27 PM, Mark Cave-Ayland
>>>
On Thu, Jun 16, 2016 at 10:15:11AM +0300, Denis V. Lunev wrote:
> This is necessary to enable creation of common qemu-img options which will
> be specified before command.
>
> The patch also enables '-V' alias to '--version' (exactly like in other
> block utilities) and documents this change.
>
The patch also creates trace_opt_parse() helper in trace/control.c to reuse
this code in next patches for qemu-nbd and qemu-io.
The patch also makes trace_init_events() static, as this call is not used
outside the module anymore.
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric
From: Shlomo Pongratz
This patch includes the device class itself, some ID register
value functions which will be needed by both distributor
and redistributor, and some skeleton functions for handling
interrupts coming in and going out, which will be filled in
in a
On Thu, Jun 16, 2016 at 08:53:12PM -0600, Eric Blake wrote:
> On 06/16/2016 04:10 AM, Stefan Hajnoczi wrote:
>
> >
> > io_sectors currently only accounts for bytes written, not bytes read.
> >
> > Therefore, I think we need:
> >
> > /* Don't charge for efficient zero writes */
> > if
On Thu, Jun 16, 2016 at 10:15:12AM +0300, Denis V. Lunev wrote:
> The command will work this way:
> qemu-img --trace qcow2* create -f qcow2 1.img 64G
>
> Signed-off-by: Denis V. Lunev
> Suggested by: Daniel P. Berrange
> Reviewed-by: Eric Blake
This is necessary to enable creation of common qemu-img options which will
be specified before command.
The patch also enables '-V' alias to '--version' (exactly like in other
block utilities) and documents this change.
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric Blake
A half-shuffle operation takes a word with zeros in the high half:
ABCD EFGH IJKL MNOP
and spreads the bits out so they are in every other bit of the word:
0A0B 0C0D 0E0F 0G0H 0I0J 0K0L 0M0N 0O0P
A half-unshuffle performs the reverse operation.
Provide functions in bitops.h
On Thu, Jun 16, 2016 at 10:15:09AM +0300, Denis V. Lunev wrote:
> Moving trace_init_backends() into trace_opt_parse() is not possible. This
> should be called after daemonize() in vl.c.
>
> Signed-off-by: Denis V. Lunev
> Reviewed-by: Eric Blake
> CC: Paolo
Changes from v6:
- changed order of patches 1 & 2
Changes from v5:
- added missed hunk into patch #7
Changes from v4:
- synced help descriprion for --trace with man for qemu.1/qemu-img.1/qemu-nbd.8
- moved @findex from qemu-option-trace.texi
Changes from v3:
- fixed difference in help/man for
Move the GICv3 parent_irq and parent_fiq pointers into the
GICv3CPUState structure rather than giving them their own array.
This will make it easy to assert the IRQ and FIQ lines for a
particular CPU interface without having to know or calculate
the CPU index for the GICv3CPUState we are working
On 17/06/16 14:35, Paolo Bonzini wrote:
> On 17/06/2016 15:13, Mark Cave-Ayland wrote:
>> Unforunately this causes regressions on a few of my SPARC32 images:
>> NextStep, Solaris 1.1.2, NetBSD and Debian Etch all hang whilst
>> enumerating the SCSI bus with this patch applied.
>
> Ok, I'll
On Thu, Jun 16, 2016 at 10:15:10AM +0300, Denis V. Lunev wrote:
> Please note, trace_init_backends() must be called in the final process,
> i.e. after daemonization. This is necessary to keep tracing thread in
> the proper process.
>
> Signed-off-by: Denis V. Lunev
>
On Thu, Jun 16, 2016 at 09:39:46AM +0100, Daniel P. Berrange wrote:
> This is a followup to:
>
> v1: https://lists.gnu.org/archive/html/qemu-devel/2016-06/msg02537.html
>
> To quote Peter:
>
> pm215: "trace-events: causing merge conflicts since 2010"
>
> There's really no good reason why
/migration-for-2.7-5' into staging (2016-06-17
14:09:46 +0100)
are available in the git repository at:
git://git.linaro.org/people/pmaydell/qemu-arm.git
tags/pull-target-arm-20160617
for you to fetch changes up to f06765a94a31bdd8b65fc83fd91a6c3f8e8a1195:
ACPI: ARM: Present GIC version in MADT
This text will be included to qemu-nbd/qemu-img mans in the next patches.
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric Blake
CC: Paolo Bonzini
CC: Stefan Hajnoczi
CC: Kevin Wolf
---
Makefile
On Fri, Jun 17, 2016 at 03:01:55PM +0200, Paolo Bonzini wrote:
>
>
> On 17/06/2016 14:46, Eduardo Habkost wrote:
> >> >
> >> > The bits are reserved anyway, so we can do whatever we want with them.
> >> > In fact I think it's weird for the architecture to make them
> >> > must-be-zero, it might
* Paolo Bonzini (pbonz...@redhat.com) wrote:
> Note that the failure mode is pretty brutal since KVM reports an
> internal error right after restarting on the destination, and who knows
> what used to happen before the assertion was introduced. All MSRs after
> MTRRs would be ignored by KVM!
I
On 06/17/2016 04:18 PM, Eduardo Habkost wrote:
On Fri, Jun 17, 2016 at 09:15:06AM +0100, Dr. David Alan Gilbert wrote:
* Eduardo Habkost (ehabk...@redhat.com) wrote:
On Thu, Jun 16, 2016 at 06:12:12PM +0100, Dr. David Alan Gilbert (git) wrote:
From: "Dr. David Alan Gilbert"
On Thu, Jun 16, 2016 at 10:15:06AM +0300, Denis V. Lunev wrote:
> Signed-off-by: Denis V. Lunev
> Reviewed-by: Eric Blake
> CC: Paolo Bonzini
> CC: Stefan Hajnoczi
> CC: Kevin Wolf
> ---
>
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric Blake
CC: Paolo Bonzini
CC: Stefan Hajnoczi
CC: Kevin Wolf
---
qemu-options.hx | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
On Thu, Jun 16, 2016 at 10:15:08AM +0300, Denis V. Lunev wrote:
> +/**
> + * Definition of QEMU options describing trace subsystem configuration
> + */
> +extern QemuOptsList qemu_trace_opts;
> +
> +/**
> + * trace_opt_parse:
> + * @optarg: A string argument of --trace command line argument
> + *
e know.
The following changes since commit 4acc8fdfd315f7ee474bea28fcbcc4dca9717d13:
Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.7-20160617' into
staging (2016-06-17 12:36:27 +0100)
are available in the git repository at:
git://github.com/ehabkost/qemu.git tags/machin
On 17/06/2016 10:14, Chao Peng wrote:
> Basically:
> - it removes old ISA devices and support only PCI devices;
I think you need to keep at least the RTC, otherwise where does Linux
get the time of day from?
> - it removes 8259, instead use MSI as much as possible. IOAPIC and PCI
> PIN are
On 17/06/2016 15:11, Efimov Vasily wrote:
> ICH9 LPC bridge is used to route PCI IRQs to GSI. The root PCI bus reference
> is
> required to setup the routing. According to specification, the bridge is
> connected to root bus. Hence, there is no reason to setup the routing
> externally.
>
>
On 17/06/2016 15:11, Efimov Vasily wrote:
> Currently, Q35 instance is configured using direct access to structure fields.
> The patch uses property interface to set the fields.
>
> Signed-off-by: Efimov Vasily
> ---
> hw/i386/pc_q35.c | 18 --
> 1 file
The command will work this way:
qemu-img --trace qcow2* create -f qcow2 1.img 64G
Signed-off-by: Denis V. Lunev
Suggested by: Daniel P. Berrange
Reviewed-by: Eric Blake
CC: Paolo Bonzini
CC: Stefan Hajnoczi
On 17/06/2016 15:13, Mark Cave-Ayland wrote:
> Unforunately this causes regressions on a few of my SPARC32 images:
> NextStep, Solaris 1.1.2, NetBSD and Debian Etch all hang whilst
> enumerating the SCSI bus with this patch applied.
Ok, I'll either fix migration (adding a subsection) or bump
On 17/06/2016 15:10, Efimov Vasily wrote:
> Currently vmport device is identified by the string literal. Using a
> preprocessor alias instead is preferable.
>
> Signed-off-by: Efimov Vasily
> ---
> hw/misc/vmport.c | 1 -
> include/hw/i386/pc.h | 3 ++-
> 2 files changed,
Please note, trace_init_backends() must be called in the final process,
i.e. after daemonization. This is necessary to keep tracing thread in
the proper process.
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric Blake
CC: Paolo Bonzini
CC:
On 17/06/2016 15:11, Efimov Vasily wrote:
> ICH9 SMB bridge can be created using qdev API despite existence of helper
> function. The type name is needed for such creation. Using a preprocessor
> alias instead the string type name itself is preferable.
>
> The patch makes the alias accessible
On Thu, Jun 16, 2016 at 11:58:28AM +0200, Cornelia Huck wrote:
> On Thu, 16 Jun 2016 09:40:23 +0100
> "Daniel P. Berrange" wrote:
>
> > Move all trace-events for files in the target-s390x/ directory to
> > their own file.
> >
> > Signed-off-by: Daniel P. Berrange
On Fri, Jun 17, 2016 at 09:15:06AM +0100, Dr. David Alan Gilbert wrote:
> * Eduardo Habkost (ehabk...@redhat.com) wrote:
> > On Thu, Jun 16, 2016 at 06:12:12PM +0100, Dr. David Alan Gilbert (git)
> > wrote:
> > > From: "Dr. David Alan Gilbert"
> > >
> > > Currently QEMU
On 17/06/2016 15:11, Efimov Vasily wrote:
> During creation of Q35 instance several parameters are set using direct
> access.
> It violates Qemu device model. Correctly, the parameters should be handled as
> object properties.
>
> The patch adds four link type properties for fields:
>
On 06/17/2016 04:36 PM, Stefan Hajnoczi wrote:
On Thu, Jun 16, 2016 at 10:15:07AM +0300, Denis V. Lunev wrote:
This text will be included to qemu-nbd/qemu-img mans in the next patches.
Signed-off-by: Denis V. Lunev
Reviewed-by: Eric Blake
CC: Paolo Bonzini
On 17/06/2016 14:42, Peter Maydell wrote:
> The trace format string in nbd_send_request uses PRIu16 for
> request->type, but request->type is a uint32_t. This provokes
> compiler warnings on the OSX clang. Use PRIu32 instead.
>
> Signed-off-by: Peter Maydell
> ---
> I
On 17/06/2016 15:11, Efimov Vasily wrote:
> The i8042 device has outgouing IRQ line A20. Currently the IRQ is referenced
> by a pointer which normally is set during machine initialization. The pointer
> is never changed at runtime. So common GPIO model can be applied to A20 IRQ
> line. Note that
Due to rounding down performed by ptimer_get_count, it returns counter - 1 for
the active timer. That's incorrect because counter should decrement only after
period been expired, not before. I.e. if running timer has been loaded with
value X, then timer counter should stay with X until period
The patch moves "hw/ide/achi.h", "hw/ide/pci.h" and "hw/ide/internal.h" headers
to corresponding folders inside "include" folder alike other Qemu headers.
Signed-off-by: Efimov Vasily
---
This patch only moves headers (with the exception of include path correction
in 'ahci.c').
On 17/06/2016 15:18, Eduardo Habkost wrote:
> On Fri, Jun 17, 2016 at 09:15:06AM +0100, Dr. David Alan Gilbert wrote:
>> * Eduardo Habkost (ehabk...@redhat.com) wrote:
>>> On Thu, Jun 16, 2016 at 06:12:12PM +0100, Dr. David Alan Gilbert (git)
>>> wrote:
From: "Dr. David Alan Gilbert"
ICH9 LPC bridge is used to route PCI IRQs to GSI. The root PCI bus reference is
required to setup the routing. According to specification, the bridge is
connected to root bus. Hence, there is no reason to setup the routing
externally.
The patch moves the setup code to 'realize' method. Also
On 17/06/2016 15:10, Efimov Vasily wrote:
> The speaker device needs pointer to ISA PIT device to operate. But according
> to
> qdev-properties.h, properties of pointer type should be avoided. It seems a
> link type property is a good substitution.
>
> Signed-off-by: Efimov Vasily
The MC146818 RTC device has output IRQ line.
Currently the corresponding field is only accessible through direct access.
Such access violates Qemu model.
The patch makes the field accessible through GPIO. It also updates the setting
of the IRQ during initialization in case of IRQ intercepting.
On Thu, Jun 16, 2016 at 10:15:07AM +0300, Denis V. Lunev wrote:
> This text will be included to qemu-nbd/qemu-img mans in the next patches.
>
> Signed-off-by: Denis V. Lunev
> Reviewed-by: Eric Blake
> CC: Paolo Bonzini
> CC: Stefan
On 17/06/2016 15:11, Efimov Vasily wrote:
> The port92 device has outgouing IRQ line A20. Currently the IRQ is referenced
> by a pointer which normally is set during machine initialization. The
> pointer is never changed at runtime. Hence, common GPIO model can be applied
> to A20 IRQ line. Note
On 17/06/2016 15:11, Efimov Vasily wrote:
> The ICH9 LPC bridge has 24 output IRQs connected to I/O APIC and 16 output
> IRQs
> connected through GSI to both PIC and I/O APIC. Currently the IRQs are
> referenced by pointers. The pointers are initialized at startup by direct
> access to the
Currently ptimer prints error message and stops the running timer that has
delta (counter) = 0, this is an incorrect behaviour for some of the ptimer
users. There are different variants of how particular timer could handle that
case besides stopping the timer, like immediate or deferred IRQ
On 17/06/2016 15:10, Efimov Vasily wrote:
> The patch moves "hw/ide/achi.h", "hw/ide/pci.h" and "hw/ide/internal.h"
> headers
> to corresponding folders inside "include" folder alike other Qemu headers.
>
> Signed-off-by: Efimov Vasily
> ---
> This patch only moves headers
On 17/06/2016 15:11, Efimov Vasily wrote:
> qdev API can be used to create CFI pflash devices despite existance of helper
> functions. The type name is needed in course of such creation. Using the
> preprocessor alias instead of the string literal itself is preferable.
>
> The patch makes the
Current ARM MPTimer implementation uses QEMUTimer for the actual timer,
this implementation isn't complete and mostly tries to duplicate of what
generic ptimer is already doing fine.
Conversion to ptimer brings the following benefits and fixes:
- Simple timer pausing implementation
On 17/06/2016 15:10, Efimov Vasily wrote:
> The speaker device needs pointer to ISA PIT device to operate. But according
> to
> qdev-properties.h, properties of pointer type should be avoided. It seems a
> link type property is a good substitution.
>
> Signed-off-by: Efimov Vasily
Hello,
Current QEMU ARM MPTimer device model provides only a certain subset of the
emulation behavior, so this patch series is supposed to add missing parts by
converting the MPTimer to use generic ptimer helper. It fixes some important
ptimer bugs and provides new features that are required for
On 17.06.2016 08:36, David Gibson wrote:
> From: Thomas Huth
>
> Since the mac99 and g3beige PowerPC machines recently broke without
> being noticed, it would be good to have a tester for "make check"
> that detects such issues immediately. A simple way to test the firmware
>
During creation of Q35 instance several parameters are set using direct access.
It violates Qemu device model. Correctly, the parameters should be handled as
object properties.
The patch adds four link type properties for fields:
mch.ram_memory
mch.pci_address_space
mch.system_memory
On 17/06/16 10:55, Paolo Bonzini wrote:
> The implementation of SATN/STOP is completely busted. The idea
> would be that the next DMA read is for a SCSI message and after
> that the adapter would transition to the command phase.
>
> The recent fix to SATN/STOP broke migration, which is one more
The port92 device has outgouing IRQ line A20. Currently the IRQ is referenced
by a pointer which normally is set during machine initialization. The
pointer is never changed at runtime. Hence, common GPIO model can be applied
to A20 IRQ line. Note that checking for IRQ to be connected as in
Currently vmport device is identified by the string literal. Using a
preprocessor alias instead is preferable.
Signed-off-by: Efimov Vasily
---
hw/misc/vmport.c | 1 -
include/hw/i386/pc.h | 3 ++-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git
The i8042 device has outgouing IRQ line A20. Currently the IRQ is referenced
by a pointer which normally is set during machine initialization. The pointer
is never changed at runtime. So common GPIO model can be applied to A20 IRQ
line. Note that checking for IRQ to be connected as in previous
Currently, Q35 instance is configured using direct access to structure fields.
The patch uses property interface to set the fields.
Signed-off-by: Efimov Vasily
---
hw/i386/pc_q35.c | 18 --
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git
The recent commit that added the prom-env-test accidentially
overwrote the check-qtest-ppc-y, check-qtest-ppc64-y and
check-qtest-sparc-y variables instead of extending them.
Fixes: fcbf4a3c0c576eec1321f9cff4fa0dd8e0b1a82f
Signed-off-by: Thomas Huth
---
tests/Makefile.include
qdev API can be used to create CFI pflash devices despite existance of helper
functions. The type name is needed in course of such creation. Using the
preprocessor alias instead of the string literal itself is preferable.
The patch makes the aliases accessible through the header.
Signed-off-by:
On 06/17/2016 02:32 PM, Paolo Bonzini wrote:
>
>
> On 17/06/2016 14:15, Cédric Le Goater wrote:
>> From: Paolo Bonzini
>>
>> This allows specifying the property via -drive if=none and creating
>> the flash device with -device.
>>
>> Signed-off-by: Paolo Bonzini
The isa_bus_irqs function initializes ISA bus IRQ array pointer with specified
value.
Previously the ICH9 LPC bridge model did not have its own IRQs but
only IRQ pointer cache. And same GSI were used for ISA bus and other sources
behind the bridge (PCI, SCI). Hence, the pc_q35_init was only
The ICH9 LPC bridge has 24 output IRQs connected to I/O APIC and 16 output IRQs
connected through GSI to both PIC and I/O APIC. Currently the IRQs are
referenced by pointers. The pointers are initialized at startup by direct
access to the structure fields. This violates Qemu device model.
The
ICH9 SMB bridge can be created using qdev API despite existence of helper
function. The type name is needed for such creation. Using a preprocessor
alias instead the string type name itself is preferable.
The patch makes the alias accessible through the header.
Signed-off-by: Efimov Vasily
The speaker device needs pointer to ISA PIT device to operate. But according to
qdev-properties.h, properties of pointer type should be avoided. It seems a
link type property is a good substitution.
Signed-off-by: Efimov Vasily
---
hw/audio/pcspk.c | 7 +--
From: Liang Li
Use 'QemuMutex comp_done_lock' and 'QemuCond comp_done_cond' instead
of 'QemuMutex *comp_done_lock' and 'QemuCond comp_done_cond'. To keep
consistent with 'QemuMutex decomp_done_lock' and
'QemuCond comp_done_cond'.
Signed-off-by: Liang Li
Am 17.06.2016 um 14:32 hat Kevin Wolf geschrieben:
> Am 03.06.2016 um 10:49 hat Fam Zheng geschrieben:
> > This takes care of both the CLOEXEC flag and fd-path mapping for image
> > locking.
> >
> > Signed-off-by: Fam Zheng
> > ---
> > include/qemu/osdep.h | 3 +++
> >
The patch series makes several devices closer to Qemu object model.
I am developing a tool that automatize creation of device and machine models.
Recently, I've take part in development of several models. And I noticed that
a significant part of code is same. The examples are:
- Each device
From: Liang Li
quit_comp_thread and quit_decomp_thread are accessed by several
thread, it's better to protect them with locks. We use a per
thread flag to replace the global one, and the new flag is protected
by a lock.
Signed-off-by: Liang Li
From: Liang Li
The current code for multi-thread compression is not clear,
especially in the aspect of using lock. Refine the code
to make it clear.
Signed-off-by: Liang Li
Message-Id: <1462433579-13691-8-git-send-email-liang.z...@intel.com>
Am 03.06.2016 um 10:49 hat Fam Zheng geschrieben:
> virtlockd in libvirt locks the first byte, we lock byte 1 to avoid
> the intervene.
>
> Both file and host device protocols are covered.
>
> The complication is with reopen. We have three different locking states,
> namely "unlocked", "shared
From: Liang Li
At the end of live migration and before vm_start() on the destination
side, we should make sure all the decompression tasks are finished, if
this can not be guaranteed, the VM may get the incorrect memory data,
or the updated memory may be overwritten by the
From: Liang Li
Use qemu_put_compression_data to do the compression directly
instead of using do_compress_ram_page, avoid some data copy.
very small improvement, at the same time, add code to check
if the compression is successful.
Signed-off-by: Liang Li
From: Liang Li
Current qemu_put_compression_data can only work with no writable
QEMUFile, and can't work with the writable QEMUFile. But it does
not provide any measure to prevent users from using it with a
writable QEMUFile.
We should fix this flaw to make it works with
From: "Dr. David Alan Gilbert"
A couple of improvements to tracing that have come out of helping
people with migration problems:
* vmstate_n_elems trace the count/name - for when you have problems
getting array counts right
* vmstate_subsection_load_bad - add the
If a field changed from something to unused, the checker wasn't flagging
if the field size mismatched. This was noticed in:
http://thread.gmane.org/gmane.comp.emulators.qemu/419802
where the 4->1 size change along with field name change to 'unused'
wasn't being flagged. Fix this.
From: Liang Li
page_buffer is set twice repeatedly, remove the previous set.
Signed-off-by: Liang Li
Reviewed-by: Dr. David Alan Gilbert
Reviewed-by: Juan Quintela
Message-Id:
From: "Denis V. Lunev"
The following sequence of operations fails:
virsh start vm
virsh snapshot-create vm
virshh save vm --file file
with the following error
error: Failed to save domain vm to file
error: internal error: unable to execute QEMU command
From: Peter Maydell
The *_to_cpup() and cpu_to_*w() functions just compose a pointer
dereference with a byteswap. Instead use ld*_p() and st*_p(),
which handle potential pointer misalignment and avoid the need
to cast the pointer.
Signed-off-by: Peter Maydell
From: Liang Li
The current code for multi-thread decompression is not clear,
especially in the aspect of using lock. Refine the code
to make it clear.
Signed-off-by: Liang Li
Message-Id: <1462433579-13691-9-git-send-email-liang.z...@intel.com>
From: Liang Li
Recently, a bug related to multiple thread compression feature for
live migration is reported. The destination side will be blocked
during live migration if there are heavy workload in host and
memory intensive workload in guest, this is most likely to happen
The following changes since commit 4acc8fdfd315f7ee474bea28fcbcc4dca9717d13:
Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.7-20160617' into
staging (2016-06-17 12:36:27 +0100)
are available in the git repository at:
https://git.kernel.org/pub/scm/virt/qemu/amit/migration.git
On 17/06/2016 14:46, Eduardo Habkost wrote:
>> >
>> > The bits are reserved anyway, so we can do whatever we want with them.
>> > In fact I think it's weird for the architecture to make them
>> > must-be-zero, it might even make more sense to make them must-be-one...
>> > It's a mask after all,
On Fri, Jun 17, 2016 at 2:44 PM, Mark Cave-Ayland
wrote:
> On 17/06/16 12:36, Artyom Tarasenko wrote:
>
>> Hi Mark,
>>
>> On Fri, Jun 17, 2016 at 1:27 PM, Mark Cave-Ayland
>> wrote:
>>> On 17/06/16 07:07, David Gibson wrote:
>>>
On 17 June 2016 at 13:41, Mark Cave-Ayland
wrote:
> Thank you looking into this. I know that some workarounds have been
> applied for similar bugs in Peter's setup, but in general should I
> assume that out-of-the-box Debian oldstable is now no longer supported
> to
On 17/06/16 12:36, Artyom Tarasenko wrote:
> Hi Mark,
>
> On Fri, Jun 17, 2016 at 1:27 PM, Mark Cave-Ayland
> wrote:
>> On 17/06/16 07:07, David Gibson wrote:
>>
>>> On Wed, Jun 15, 2016 at 01:10:18PM +1000, David Gibson wrote:
On Tue, Jun 14, 2016 at
On 06/15/2016 07:40 PM, Krzeminski, Marcin (Nokia - PL/Wroclaw) wrote:
>
>
> W dniu 15.06.2016 o 16:25, Cédric Le Goater pisze:
>> On 06/15/2016 03:41 PM, marcin.krzemin...@nokia.com wrote:
>>> From: Marcin Krzeminski
>>>
>>> Quad and Equad modes for Spansion and
On (Fri) 17 Jun 2016 [11:55:48], Paolo Bonzini wrote:
> The implementation of SATN/STOP is completely busted. The idea
> would be that the next DMA read is for a SCSI message and after
> that the adapter would transition to the command phase.
>
> The recent fix to SATN/STOP broke migration,
On Fri, Jun 17, 2016 at 09:47:27AM +0200, Paolo Bonzini wrote:
> On 16/06/2016 22:14, Eduardo Habkost wrote:
> > This is necessary only when phys_bits is higher on the
> > destination, right?
> >
> > Should we really default this to true? I would like to enable
> > this hack only when really
On 15/06/16 13:35, Laszlo Ersek wrote:
> On 06/15/16 12:08, Mark Cave-Ayland wrote:
>> On 15/06/16 10:48, Andrew Jones wrote:
>>
>>> On Wed, Jun 15, 2016 at 08:14:18AM +0100, Mark Cave-Ayland wrote:
Hi Peter,
Just to let you know that after pulling git master
Thanks, much better! Just one thing:
On 16/06/2016 12:20, Ashijeet Acharya wrote:
> +case SOCKET_ADDRESS_KIND_INET:
> +buf = g_strdup(addr->u.inet.data->host);
Please include the port too here. Also, if the host contains a colon
(which you can check with strchr), please print it as
The trace format string in nbd_send_request uses PRIu16 for
request->type, but request->type is a uint32_t. This provokes
compiler warnings on the OSX clang. Use PRIu32 instead.
Signed-off-by: Peter Maydell
---
I didn't catch this when I ran the merge build test for the
201 - 300 of 473 matches
Mail list logo