On Fri, Sep 02, 2011 at 11:37:25AM +0300, Avi Kivity wrote:
> On 09/01/2011 07:05 PM, Anthony Liguori wrote:
> >
> > The challenge is what you do about something like ne2k where the core
> > chipset can either be a PCI device or an ISA device. You would have
> > to implement a wrapper around pci
On 09/02/2011 12:46 AM, Adnan Khaleel wrote:
> Is there anyway we can prevent Qemu breaking up 64,128 and 256bit XMM or YMM
> instructions into smaller chunks and have them issue as a single transaction
> of the original width?
Not yet. The new MemoryRegion API will allow this, but we need to c
On 09/01/2011 09:25 PM, Anthony Liguori wrote:
> I think this is the wrong approach given the introduction of the memory API.
>
> I think we should have a generic memory access function that takes a
> MemoryRegion as it's first argument.
>
> The PCI bus should then expose one memory region for e
The Buildbot has detected a new failure on builder block_x86_64_debian_6_0
while building qemu.
Full details are available at:
http://buildbot.b1-systems.de/qemu/builders/block_x86_64_debian_6_0/builds/19
Buildbot URL: http://buildbot.b1-systems.de/qemu/
Buildslave for this Build: yuzuki
Build
On 2 September 2011 22:48, Anthony Liguori wrote:
> diff --git a/translate-all.c b/translate-all.c
> index 041c108..ecb035a 100644
> --- a/translate-all.c
> +++ b/translate-all.c
> @@ -67,7 +67,9 @@ int cpu_gen_code(CPUState *env, TranslationBlock *tb, int
> *gen_code_size_ptr)
> #endif
> tc
On 2 September 2011 22:48, Anthony Liguori wrote:
> diff --git a/target-i386/helper.c b/target-i386/helper.c
> index 5df40d4..793e467 100644
> --- a/target-i386/helper.c
> +++ b/target-i386/helper.c
> @@ -32,6 +32,24 @@
>
> //#define DEBUG_MMU
>
> +void cpu_get_fp80(uint64_t *pmant, uint16_t *pe
Signed-off-by: Anthony Liguori
---
Makefile.target |8 +---
cpu-exec.c | 19 ---
cpus.c | 11 ++-
exec.c | 29 ++---
softmmu_template.h | 14 ++
target-i386/help
Signed-off-by: Anthony Liguori
---
Makefile.target |3 ++-
cpu-all.h |6 ++
exec.c |2 +-
qemu-common.h |2 +-
translate-all.c | 12 +---
5 files changed, 19 insertions(+), 6 deletions(-)
diff --git a/Makefile.target b/Makefile.target
index 0787758.
Hi,
There have been a few attempts in the past to allow TCG to be disabled
at build time. Recently, Alex made the suggestion that we could do it by using
the same trick that we used to introduce kvm support. That involves introducing
a tcg_enabled() macro that will be (0) if TCG is disabled in t
Signed-off-by: Anthony Liguori
---
arch_init.c |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/arch_init.c b/arch_init.c
index 567ab32..5b37f8d 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -711,7 +711,11 @@ int audio_available(void)
int tcg_available(void)
{
+#if
Signed-off-by: Anthony Liguori
---
Makefile.target |6 +++---
exec.c | 44 +---
target-i386/op_helper.c |2 +-
translate-all.c | 12 +++-
4 files changed, 36 insertions(+), 28 deletions(-)
diff --git a/M
This lets you build without TCG (KVM only). When this flag is passed to
configure, it will automatically filter out the target list to only those that
support KVM.
I guess the same could be done for Xen but I'll let the Xen folks figure that
one out :-)
Signed-off-by: Anthony Liguori
---
confi
On Fri, 2 Sep 2011 18:06:13 -0300
Luiz Capitulino wrote:
> On Fri, 2 Sep 2011 12:34:56 -0500
> Anthony Liguori wrote:
>
> > A new QMP only command to change the blockdev associated with a block
> > device.
> > The semantics of change right now are just plain scary. This command
> > introduc
On Fri, 2 Sep 2011 12:34:56 -0500
Anthony Liguori wrote:
> A new QMP only command to change the blockdev associated with a block device.
> The semantics of change right now are just plain scary. This command
> introduces
> sane semantics. For more details, see the documentation in the schema
Allow overriding the location of Samba's smbd.
Pretty much every OS I look at has some means of
changing this path (patching) so lets just make
it easier for OS developers creating packages
and/or end users to override the location.
Signed-off-by: Brad Smith
---
configure |9
On Fri, 2 Sep 2011 12:34:55 -0500
Anthony Liguori wrote:
> New QMP only command to change the VNC server's listening address.
>
> Signed-off-by: Anthony Liguori
> ---
> v1 -> v2
> - Enhanced docs (Luiz)
> ---
> qapi-schema.json | 15 +++
> qmp-commands.hx |8
> qm
On Fri, 2 Sep 2011 12:34:47 -0500
Anthony Liguori wrote:
> Use the new middle mode within the existing QMP server.
>
> Signed-off-by: Anthony Liguori
> ---
> v1 -> v2
> - fixed some of the generated header dependencies
> ---
> Makefile | 12
> Makefile.objs|2 +
Forgot to mention, I'm using linux 3.0.4, but the problem also is
present on all 3.0.* kernels:
% pacman -Ss linux | 'grep' core
...
core/linux 3.0.4-1 (base) [installed]
And the kvm module used is the intel one:
% lsmod | 'grep' kvm
kvm_intel 53373 3
kvm 328912
On Fri, Sep 02, 2011 at 05:19:54PM +0200, Gerd Hoffmann wrote:
> reds_stream_free() may call the channel_event callback which is not
> supposed to be callsed from worker thread context. This patch moves
> the reds_stream_free call for the display channel from the worker to
> the dispatcher to fix
This provides a bridge between Error (new error mechanism) and QError (old error
mechanism). Errors can be propagated whereas QError cannot.
The minor evilness avoids layering violations. Since QError should go away RSN,
it seems like a reasonable hack.
Signed-off-by: Anthony Liguori
---
v1 ->
Mark formats that cannot be probed with 100% reliability (just raw). This is
necessary to implement change-blockdev.
Signed-off-by: Anthony Liguori
---
block/raw.c |2 ++
block_int.h |2 ++
2 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/block/raw.c b/block/raw.c
index 55
Currently when disabling login in VNC, the password is cleared out and the
authentication protocol is forced to AUTH_VNC. If you're using a stronger
authentication protocol, this has the effect of downgrading your security
protocol.
Fix this by only changing the authentication protocol if the cur
Use the new middle mode within the existing QMP server.
Signed-off-by: Anthony Liguori
---
v1 -> v2
- fixed some of the generated header dependencies
---
Makefile | 12
Makefile.objs|2 ++
Makefile.target |6 +++---
monitor.c| 11 ---
qapi-s
This is a new QMP only command that only changes the VNC password.
Signed-off-by: Anthony Liguori
---
v1 -> v2
- fix schema documentation
---
qapi-schema.json | 13 +
qmp-commands.hx |8
qmp.c| 11 ++-
3 files changed, 31 insertions(+), 1 deleti
N.B. This temporarily changes the error semantics of eject. Errors are not
stable in QMP so this is not technically a breakage. The last patch restores
the previous error semantics.
Signed-off-by: Anthony Liguori
---
blockdev.c | 22 +++---
blockdev.h |1 -
hm
A simple example conversion 'info name'. This also adds the new files for
QMP and HMP.
Signed-off-by: Anthony Liguori
---
Makefile.objs|1 +
hmp.c| 26 ++
hmp.h| 22 ++
monitor.c| 27 +-
HMP is now implemented in terms of QMP. The monitor has a bunch of logic to
deal with HMP right now like readline support. Export it from the monitor so
we can consume it in hmp.c.
In short time, hmp.c will take over all of the readline bits.
Signed-off-by: Anthony Liguori
---
monitor.c | 1
Public bug reported:
I'm using the qemu-kvm package from arch (not the qemu package), and on
prior versions of qemu-kvm I was able to use -usbdevide tablet without
problems. The absolute mouse position is great... With current version
of qemu-kvm, when I use -usbdevice tablet, I got no mouse at
Am 02.09.2011 um 18:38 schrieb Peter Maydell:
2011/8/28 Peter Maydell :
On 28 August 2011 14:02, Alexander Graf wrote:
There are two ways we can fix the naming clash here:
(1) change int16 -> int_fast16_t. This was argued against because
it means that the type might have a different width dep
To get the ball rolling merging QAPI, this patch introduces a "middle mode" to
the code generator. In middle mode, the code generator generates marshalling
functions that are compatible with the current QMP server. We absolutely need
to replace the current QMP server in order to support proper as
Changing an encrypted block device requires a flow of:
1) change device
2) receive error from change
3) block_passwd
DeviceEncrypted receives the device name but with a block backend with multiple
backing files, any one of the backing files could be encrypted. From a UI
perspective, you have
block_passwd is unfortunately named so while converting block_passwd to QAPI,
introduce a more properly named alias.
Signed-off-by: Anthony Liguori
---
blockdev.c | 29 +++--
hmp-commands.hx |2 +-
hmp.c| 12
hmp.h|1
New QMP only command to change the VNC server's listening address.
Signed-off-by: Anthony Liguori
---
v1 -> v2
- Enhanced docs (Luiz)
---
qapi-schema.json | 15 +++
qmp-commands.hx |8
qmp.c|7 +++
3 files changed, 30 insertions(+), 0 deletions(-)
On 09/02/2011 12:23 PM, Luiz Capitulino wrote:
On Fri, 26 Aug 2011 16:48:13 +0200
Jan Kiszka wrote:
This reserves JSON objects that contain the key '__class__' for QMP-specific
complex objects. First user will be the buffer class.
BTW, we need to teach QAPI how handle these types.
QAPI alre
On 09/02/2011 10:57 AM, Michael S. Tsirkin wrote:
> On Thu, Sep 01, 2011 at 03:26:01PM -0500, Scott Wood wrote:
>> On 09/01/2011 03:00 PM, Michael S. Tsirkin wrote:
>>> That's a very rich interface, and easy to get wrong.
>>> AFAIK the only reason vfio uses read/write for PCI was to avoid inventing
On 2011-09-02 19:23, Luiz Capitulino wrote:
> On Fri, 26 Aug 2011 16:48:13 +0200
> Jan Kiszka wrote:
>
>> This reserves JSON objects that contain the key '__class__' for QMP-specific
>> complex objects. First user will be the buffer class.
>>
>> CC: Luiz Capitulino
>> Signed-off-by: Jan Kiszka
Convert the change command to QAPI and document all of its gloriousness.
Signed-off-by: Anthony Liguori
---
v1 -> v2
- Fix docs (Luiz)
fix hmp change command
---
blockdev.c | 27 --
blockdev.h |6 +++-
hmp-commands.hx |3 +-
hmp.c| 59 ++
A new QMP only command to change the blockdev associated with a block device.
The semantics of change right now are just plain scary. This command introduces
sane semantics. For more details, see the documentation in the schema file.
Signed-off-by: Anthony Liguori
---
v1 -> v2
- Rename command
This is my attempt to jump-start the QAPI switch over. All of the hard work
is already done in my qapi branch, we just need to start merging stuff.
To simplify the merge process, I've introduced a new mode to the code generator
that lets us do conversions without using the new QMP server.
Once t
Reported-by: Luiz Capitulino
Signed-off-by: Anthony Liguori
---
error.c |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/error.c b/error.c
index b802752..68c0039 100644
--- a/error.c
+++ b/error.c
@@ -97,6 +97,10 @@ bool error_is_type(Error *err, const char *fmt)
On Fri, 26 Aug 2011 16:48:10 +0200
Jan Kiszka wrote:
> More than one year ago I posted some patches to add a monitor command
> callend device_show. The purpose of that command is to dump the state of
> some qdev device based on its vmstate.
>
> To improve the usability of that interface, the pre
On Fri, 26 Aug 2011 16:48:13 +0200
Jan Kiszka wrote:
> This reserves JSON objects that contain the key '__class__' for QMP-specific
> complex objects. First user will be the buffer class.
>
> CC: Luiz Capitulino
> Signed-off-by: Jan Kiszka
> ---
> QMP/qmp-spec.txt | 16 +---
> 1
On Fri, 26 Aug 2011 20:02:37 +0200
Jan Kiszka wrote:
> On 2011-08-26 17:47, Jan Kiszka wrote:
> > On 2011-08-26 17:23, Jan Kiszka wrote:
> >> On 2011-08-26 17:21, Peter Maydell wrote:
> >>> On 26 August 2011 15:48, Jan Kiszka wrote:
> Will be used by QBuffer.
> >>>
> >>> Is it possible to u
On Fri, 02 Sep 2011 11:45:28 -0500
Anthony Liguori wrote:
> On 09/02/2011 11:31 AM, Michael Roth wrote:
>
> >>> This ended up being a bit harder than I initially thought so I'd prefer
> >>> to delay this to another series so we can start converting commands to
> >>> QAPI ASAP.
> >>
> >> Fine wit
On Fri, Sep 2, 2011 at 5:04 PM, Julien Heyman wrote:
> Thanks Dave.
> I use system emulation, and my main concern is "just" to know that the
> actual board will run faster than the emulation. So based on your example,
> and even though my target board (mini2440) is nowhere as fast as a Panda
> boa
Kevin Wolf writes:
> Am 02.09.2011 17:30, schrieb Markus Armbruster:
>> Kevin Wolf writes:
>>
>>> Am 03.08.2011 15:08, schrieb Markus Armbruster:
Device models should be able to set it without an unclean include of
block_int.h.
Signed-off-by: Markus Armbruster
---
On 09/02/2011 11:31 AM, Michael Roth wrote:
This ended up being a bit harder than I initially thought so I'd prefer
to delay this to another series so we can start converting commands to
QAPI ASAP.
Fine with me.
Same. Curious what the complications are though, I wouldn't mind looking
at it.
Kevin Wolf writes:
> Am 02.09.2011 17:35, schrieb Markus Armbruster:
>> Kevin Wolf writes:
>>
>>> Am 03.08.2011 15:08, schrieb Markus Armbruster:
Breaks migration of qdevs "scsi-cd" and legacy "scsi-disk" to older
versions. We normally use subsections to avoid that. Not possible
>>>
On Fri, Sep 02, 2011 at 04:37:24PM +0100, Stefan Hajnoczi wrote:
> On Fri, Sep 2, 2011 at 4:01 PM, Daniel P. Berrange wrote:
> > From: "Daniel P. Berrange"
> >
> > To aid in knowing whether a 'block_resize' was succesful, display
> > the sector count in the 'info block' output
> >
> > Signed-off-
2011/8/28 Peter Maydell :
> On 28 August 2011 14:02, Alexander Graf wrote:
> There are two ways we can fix the naming clash here:
>
> (1) change int16 -> int_fast16_t. This was argued against because
> it means that the type might have a different width depending on
> the host system, which means
From: "Daniel P. Berrange"
To aid in knowing whether a 'block_resize' was succesful, display
the logical disk size in bytes, in the 'info block' output
In v2:
- Replace sectors with bytes
Signed-off-by: Daniel P. Berrange
---
block.c |6 --
qmp-commands.hx |1 +
2 files
Am 02.09.2011 18:05, schrieb Anthony Liguori:
> On 08/25/2011 08:52 AM, Kevin Wolf wrote:
>> Am 25.08.2011 15:40, schrieb Anthony Liguori:
>>> On 08/25/2011 07:19 AM, Kevin Wolf wrote:
Am 24.08.2011 20:43, schrieb Anthony Liguori:
> Signed-off-by: Anthony Liguori
> ---
>blockde
On 09/02/2011 11:09 AM, Luiz Capitulino wrote:
On Fri, 02 Sep 2011 11:00:50 -0500
Anthony Liguori wrote:
On 08/25/2011 11:24 AM, Michael Roth wrote:
On 08/24/2011 01:42 PM, Anthony Liguori wrote:
Use the new middle mode within the existing QMP server.
Signed-off-by: Anthony Liguori
---
Make
Am 02.09.2011 17:30, schrieb Markus Armbruster:
> Kevin Wolf writes:
>
>> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>>> Device models should be able to set it without an unclean include of
>>> block_int.h.
>>>
>>> Signed-off-by: Markus Armbruster
>>> ---
>>> block.c |6 --
Am 02.09.2011 17:29, schrieb Markus Armbruster:
> Kevin Wolf writes:
>
>> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>>> Need to ask the device, so this requires new BlockDevOps member
>>> is_medium_ejected().
>>>
>>> Signed-off-by: Markus Armbruster
>>
>> I find the name confusing. "medium
Am 02.09.2011 17:35, schrieb Markus Armbruster:
> Kevin Wolf writes:
>
>> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>>> Breaks migration of qdevs "scsi-cd" and legacy "scsi-disk" to older
>>> versions. We normally use subsections to avoid that. Not possible
>>> here, because we don't have
Remove the 'tarbin' target -- it isn't used as part of the official
QEMU release process, and it's out of date (various new bios files
were never added to its list of files). It's better not to provide
it at all than to have a broken makefile target we never use or test.
(Creating a tarball by jus
On 08/25/2011 08:32 AM, Luiz Capitulino wrote:
On Wed, 24 Aug 2011 13:43:06 -0500
Anthony Liguori wrote:
New QMP only command to change the VNC server's listening address.
Signed-off-by: Anthony Liguori
---
qapi-schema.json | 14 ++
qmp-commands.hx |8
qmp.c
On 2 September 2011 17:04, Julien Heyman wrote:
> Thanks Dave.
> I use system emulation, and my main concern is "just" to know that the
> actual board will run faster than the emulation. So based on your example,
> and even though my target board (mini2440) is nowhere as fast as a Panda
> board, t
On Fri, 02 Sep 2011 11:00:50 -0500
Anthony Liguori wrote:
> On 08/25/2011 11:24 AM, Michael Roth wrote:
> > On 08/24/2011 01:42 PM, Anthony Liguori wrote:
> >> Use the new middle mode within the existing QMP server.
> >>
> >> Signed-off-by: Anthony Liguori
> >> ---
> >> Makefile | 11 +++
On 08/25/2011 08:33 AM, Luiz Capitulino wrote:
On Wed, 24 Aug 2011 13:43:05 -0500
Anthony Liguori wrote:
This is a new QMP only command that only changes the VNC password.
Isn't this useful in HMP too?
I think the right design goal for HMP is to minimize the number of
commands since a hum
On 08/25/2011 08:52 AM, Kevin Wolf wrote:
Am 25.08.2011 15:40, schrieb Anthony Liguori:
On 08/25/2011 07:19 AM, Kevin Wolf wrote:
Am 24.08.2011 20:43, schrieb Anthony Liguori:
Signed-off-by: Anthony Liguori
---
blockdev.c | 22 +++---
blockdev.h |1 -
h
Thanks Dave.
I use system emulation, and my main concern is "just" to know that the
actual board will run faster than the emulation. So based on your example,
and even though my target board (mini2440) is nowhere as fast as a Panda
board, this should be the case by a comfortable margin. Now, as I a
On Fri, Sep 02, 2011 at 08:11:29AM +0200, Paolo Bonzini wrote:
> On 09/02/2011 02:11 AM, David Gibson wrote:
> >Why not limit the change to ppc then?
> >>>
> >>> Because the bug is masked by the x86 memory model, but it is still
> >>> there even there conceptually. It is not really true th
On 08/25/2011 11:24 AM, Michael Roth wrote:
On 08/24/2011 01:42 PM, Anthony Liguori wrote:
Use the new middle mode within the existing QMP server.
Signed-off-by: Anthony Liguori
---
Makefile | 11 +++
Makefile.objs | 2 ++
Makefile.target | 6 +++---
monitor.c | 11 ---
qapi-schema.
On 08/24/2011 03:15 PM, Luiz Capitulino wrote:
On Wed, 24 Aug 2011 13:42:56 -0500
Anthony Liguori wrote:
This provides a bridge between Error (new error mechanism) and QError (old error
mechanism). Errors can be propagated whereas QError cannot.
The minor evilness avoids layering violations.
On Thu, Sep 01, 2011 at 03:26:01PM -0500, Scott Wood wrote:
> On 09/01/2011 03:00 PM, Michael S. Tsirkin wrote:
> > That's a very rich interface, and easy to get wrong.
> > AFAIK the only reason vfio uses read/write for PCI was to avoid inventing
> > a custom interface. But if you do, it looks like
Am 02.09.2011 17:20, schrieb Markus Armbruster:
> Kevin Wolf writes:
>
>> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>>> Use a subsection, so that migration to older version still works,
>>> provided the tray is closed and unlocked.
>>>
>>> Signed-off-by: Markus Armbruster
>>
>> Ah, here is
On Thu, Sep 01, 2011 at 04:31:09PM -0400, Paolo Bonzini wrote:
> > > > Why not limit the change to ppc then?
> > >
> > > Because the bug is masked by the x86 memory model, but it is still
> > > there even there conceptually. It is not really true that x86 does
> > > not need memory barriers, though
Hi,
As of commit 12d4536f7d911b6d87a766ad7300482ea663cea2, the I/O thread is
now enabled by default.
I think we've done about as much testing and preparation as we can. If
you run into any odd regressions, you should be able to easily revert
this commit still. We can hold off making signif
On 08/31/2011 02:25 AM, Frediano Ziglio wrote:
Signed-off-by: Frediano Ziglio
Applied. Thanks.
Regards,
Anthony Liguori
---
trace-events | 10 +-
vl.c |6 +++---
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/trace-events b/trace-events
index dc300
On 09/02/2011 09:42 AM, Paolo Bonzini wrote:
On 09/02/2011 04:31 PM, Anthony Liguori wrote:
For a platform to be supported, it needs to be actively maintained and
fixed. If there aren't enough folks testing/fixing Darwin/ppc64, then
it's not a platform we can reasonable support :-/
I agree unf
On 08/29/2011 09:55 AM, Anthony Liguori wrote:
This has been discussed before in the past. The special casing really makes no
sense anymore. This seems like a good change to make for 1.0.
Signed-off-by: Anthony Liguori
Applied.
Regards,
Anthony Liguori
---
Makefile|5 ++---
On 08/30/2011 02:24 PM, Blue Swirl wrote:
On Mon, Aug 29, 2011 at 2:55 PM, Anthony Liguori wrote:
This has been discussed before in the past. The special casing really makes no
sense anymore. This seems like a good change to make for 1.0.
Signed-off-by: Anthony Liguori
---
Makefile
On 08/28/2011 03:01 AM, Brad wrote:
Fix install(1) usage to be compatible with OpenBSD's install(1).
When creating a directory via the -d flag the -p flag cannot be
used at the same time. Also in the context of installing QEMU it
doesn't make sense to use the -p flag anyway so use the [default]
On 09/02/2011 10:31 AM, Paolo Bonzini wrote:
On 09/02/2011 05:18 PM, Gerd Hoffmann wrote:
Can you just use a bottom half to defer this work to the I/O thread?
Bottom half scheduling has to be signal safe which means it will also be
thread safe.
Not that straight forward as I would have to pa
On Fri, Sep 2, 2011 at 4:01 PM, Daniel P. Berrange wrote:
> From: "Daniel P. Berrange"
>
> To aid in knowing whether a 'block_resize' was succesful, display
> the sector count in the 'info block' output
>
> Signed-off-by: Daniel P. Berrange
> ---
> block.c | 6 --
> qmp-commands.
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> Breaks migration of qdevs "scsi-cd" and legacy "scsi-disk" to older
>> versions. We normally use subsections to avoid that. Not possible
>> here, because we don't have a section to begin with. Too bad.
>>
>> Signed-off-by
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> To let device models distinguish between eject and load.
>>
>> Signed-off-by: Markus Armbruster
>> ---
>> block.c| 12 ++--
>> block.h|3 ++-
>> hw/fdc.c |2 +-
>> hw/ide/core.c |
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> BlockDriverState member buffer_alignment is initially 512. The device
>> model may set them, with bdrv_set_buffer_alignment(). If the device
>> model gets detached (hot unplug), the device's alignment is left
>> behind. On
On 09/02/2011 05:18 PM, Gerd Hoffmann wrote:
Can you just use a bottom half to defer this work to the I/O thread?
Bottom half scheduling has to be signal safe which means it will also be
thread safe.
Not that straight forward as I would have to pass arguments to the
bottom half.
Can you add
Anthony Liguori writes:
> On 09/02/2011 04:39 AM, Stefan Hajnoczi wrote:
>> On Thu, Sep 1, 2011 at 8:08 PM, Anthony Liguori wrote:
>>> On 09/01/2011 03:06 AM, Stefan Hajnoczi wrote:
The following changes since commit
f0fb8b7180fdcf536ea635a0720e1496110ecb3b:
Merge branc
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> Device models should be able to set it without an unclean include of
>> block_int.h.
>>
>> Signed-off-by: Markus Armbruster
>> ---
>> block.c |6 --
>> block.h |1 +
>> hw/ide/core.c |2 +-
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> Need to ask the device, so this requires new BlockDevOps member
>> is_medium_ejected().
>>
>> Signed-off-by: Markus Armbruster
>
> I find the name confusing. "medium is ejected" seems to mean "tray is
> open". It isn't obvi
On 09/02/2011 05:22 PM, Markus Armbruster wrote:
>> On another note, have our subsections been fixed meanwhile? If not,
>> increasing the version number instead might be the right choice. Paolo?
>
> No, they haven't, but there's hope they will before 1.0.
Use a subsection here anyway? Or sh
Paolo Bonzini writes:
> On 09/02/2011 02:20 PM, Kevin Wolf wrote:
>> Ah, here is the migration part. The state wasn't migrated before, so not
>> doing it in patch 10 doesn't break anything. I would still prefer to add
>> migration in the same patch that adds the fields.
>>
>> You could just move
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> Use a subsection, so that migration to older version still works,
>> provided the tray is closed and unlocked.
>>
>> Signed-off-by: Markus Armbruster
>
> Ah, here is the migration part. The state wasn't migrated before, so
On 09/02/2011 10:18 AM, Gerd Hoffmann wrote:
Hi,
A patch like the attached (warning: untested) should do as quick&dirty
fix for stable. But IMO we really should fix spice instead.
I agree. I'm not sure I like the idea of still calling QEMU code without
holding the mutex (even the QObject code
reds_stream_free() may call the channel_event callback which is not
supposed to be callsed from worker thread context. This patch moves
the reds_stream_free call for the display channel from the worker to
the dispatcher to fix this issue.
[ Note: not tested yet, against 0.8 branch, sending out fo
Kevin Wolf writes:
> Am 03.08.2011 15:08, schrieb Markus Armbruster:
>> When ide-cd is backed by a physical drive, we want the physical tray
>> match the virtual one. To that end, we call bdrv_eject() on guest's
>> load/eject, and bdrv_lock_medium() on guest's prevent/allow removal.
>> But we do
Hi,
A patch like the attached (warning: untested) should do as quick&dirty
fix for stable. But IMO we really should fix spice instead.
I agree. I'm not sure I like the idea of still calling QEMU code without
holding the mutex (even the QObject code).
I though just creating the objects isn'
From: "Daniel P. Berrange"
To aid in knowing whether a 'block_resize' was succesful, display
the sector count in the 'info block' output
Signed-off-by: Daniel P. Berrange
---
block.c |6 --
qmp-commands.hx |1 +
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a
On Thu, Sep 01, 2011 at 05:55:43PM +0200, Christoph Hellwig wrote:
> Please try the attached (untested) patch:
Yes, this patch worked succesfully with the following test case:
$ dd if=/dev/zero of=loop.img bs=1M seek=100 count=0
$ losetup -f loop.img
$ ./x86_64-softmmu/qemu-system-x86_64 -
On 09/02/2011 09:54 AM, Anthony Liguori wrote:
On 09/02/2011 04:39 AM, Stefan Hajnoczi wrote:
On Thu, Sep 1, 2011 at 8:08 PM, Anthony Liguori
wrote:
On 09/01/2011 03:06 AM, Stefan Hajnoczi wrote:
The following changes since commit
f0fb8b7180fdcf536ea635a0720e1496110ecb3b:
Merge branch 'omap-
Kevin Wolf writes:
> Am 03.08.2011 15:07, schrieb Markus Armbruster:
>> Signed-off-by: Markus Armbruster
>> ---
>> block.c |8
>> block.h |2 +-
>> block/raw-posix.c |8
>> block/raw.c |6 +++---
>> block_int.h |2 +-
>> hw/i
On 09/02/2011 04:39 AM, Stefan Hajnoczi wrote:
On Thu, Sep 1, 2011 at 8:08 PM, Anthony Liguori wrote:
On 09/01/2011 03:06 AM, Stefan Hajnoczi wrote:
The following changes since commit
f0fb8b7180fdcf536ea635a0720e1496110ecb3b:
Merge branch 'omap-for-upstream' of
git://git.linaro.org/people
Kevin Wolf writes:
> Am 03.08.2011 15:07, schrieb Markus Armbruster:
>> We already track it in BlockDriverState. Just like tray open/close
>> state, we should track it in the device models instead, because it's
>> device state.
>>
>> Signed-off-by: Markus Armbruster
>> ---
>> hw/ide/atapi.c
On 09/02/2011 03:42 AM, Stefan Hajnoczi wrote:
2011/9/1 Michael Roth:
On 09/01/2011 01:54 PM, Anthony Liguori wrote:
On 08/25/2011 03:18 AM, Stefan Hajnoczi wrote:
Avoid duplicate object files during the link. There are legitimate
cases where a link command-line would include duplicate objec
Kevin Wolf writes:
> Am 03.08.2011 15:07, schrieb Markus Armbruster:
>> Signed-off-by: Markus Armbruster
>> ---
>> hw/scsi-disk.c | 17 +
>> 1 files changed, 13 insertions(+), 4 deletions(-)
>>
>> diff --git a/hw/scsi-disk.c b/hw/scsi-disk.c
>> index fa198f9..d549cb7 100644
>
Kevin Wolf writes:
> Am 03.08.2011 15:07, schrieb Markus Armbruster:
>> We already track it in BlockDriverState since commit 4be9762a. As
>> discussed in that commit's message, we should track it in the device
>> device models instead, because it's device state.
>>
>> Signed-off-by: Markus Armb
1 - 100 of 183 matches
Mail list logo
