On 09/16/2014 10:34 PM, Kevin Wolf wrote:
> Am 16.09.2014 um 14:10 hat Paolo Bonzini geschrieben:
>> Il 16/09/2014 14:02, Alexey Kardashevskiy ha scritto:
>>> I am having problems when migrate a guest via libvirt like this:
>>>
>>> virsh migrate --live --persistent --undefinesource --copy-storage-a
Jeff Cody writes:
> In vhdx_create_metadata(), we allocate 40 bytes to entry_buffer for
> the various metadata table entries. However, we write out 64kB from
> that buffer into the new file. Only write out the correct 40 bytes.
>
> Signed-off-by: Jeff Cody
> ---
> block/vhdx.c | 16 --
> What if instead of having a "mmu_index" for the mmu arrays, we have a pointer
> to the "mmu context". This does imply an extra memory load on the fast path,
> but probably not an extra instruction.
>
> With this, we can suddenly afford to have a relatively large number of mmu
> contexts, with
"Michael S. Tsirkin" writes:
> On Tue, Sep 16, 2014 at 10:01:19PM +0300, Michael S. Tsirkin wrote:
>> On Tue, Sep 16, 2014 at 08:31:08PM +0200, Paolo Bonzini wrote:
>> > Il 16/09/2014 18:56, Michael S. Tsirkin ha scritto:
>> > > On Tue, Sep 16, 2014 at 06:27:51PM +0200, Paolo Bonzini wrote:
>> >
Eric Blake writes:
> On 09/16/2014 12:31 PM, Paolo Bonzini wrote:
>
>>> Change legacy_name to point to a detailed human-readable
>>> description of the type?
>>> E.g. "Ethernet 6-byte MAC Address, format: AA:BB:CC:DD:EE:FF"?
>>
>> If libvirt can cope with
>>
>> e1000.mac=str (Ethernet 6-byte MA
Hi,
The VHOST_NET_SET_BACKEND could tell the user when the backend is created or
destroyed.is usefull for the user but this command is lost in the protocol
Hi,
There is two memory regions when receive VHOST_SET_MEM_TABLE message:
region[0]
gpa = 0x0
size = 655360
ua = 0x2ac0
offset = 0
region[1]
gpa = 0xC
size = 2146697216
ua = 0x2acc
offset = 786432
region[0] always
> From: Paolo Bonzini [mailto:pbonz...@redhat.com]
> Sent: Tuesday, September 16, 2014 10:37 PM
> Subject: Re: [Qemu-devel] [PATCH 0/3] Fix confused output for alias properties
>
> Il 16/09/2014 16:32, Markus Armbruster ha scritto:
> > Paolo Bonzini writes:
> >
> >> Il 16/09/2014 11:16, Markus Ar
On Tue, 09/16 15:24, Stefan Hajnoczi wrote:
> diff --git a/qapi/block-core.json b/qapi/block-core.json
> index a685d02..56302c7 100644
> --- a/qapi/block-core.json
> +++ b/qapi/block-core.json
> @@ -197,7 +197,7 @@
> # 'file', 'file', 'ftp', 'ftps', 'host_cdrom', 'host_device',
> # 'h
Introduce help function acpi_set_local_sts() to simplify acpi_cpu_plug_cb
and acpi_cpu_hotplug_init, so that we can keep bit setting in one place.
Signed-off-by: Gu Zheng
---
hw/acpi/cpu_hotplug.c | 22 +++---
1 files changed, 15 insertions(+), 7 deletions(-)
diff --git a/hw/a
Update rtc_cmos in pc_cpu_plug directly instead of the notifier, with
this change, there will no user of CPU hot-plug notifier any more, so
remove it.
Signed-off-by: Gu Zheng
---
hw/i386/pc.c| 25 ++---
include/sysemu/sysemu.h |3 ---
qom/cpu.c
Rename:
AcpiCpuHotplug_init --> acpi_cpu_hotplug_init
AcpiCpuHotplug_ops --> acpi_cpu_hotplug_ops
for better readability, just cleanup.
Signed-off-by: Gu Zheng
---
hw/acpi/cpu_hotplug.c |4 ++--
hw/acpi/ich9.c|4 ++--
hw/acpi/piix4.c |4 ++--
inc
Hi
在 08/01/2014 10:52 PM, Dr. David Alan Gilbert 写道:
* Yang Hongyang (yan...@cn.fujitsu.com) wrote:
We need a buffer to store migration data.
On save side:
all saved data was write into colo buffer first, so that we can know
the total size of the migration data. this can also separate the d
Add cpu hotplug handler to PC_MACHINE, which will perform the acpi
cpu hotplug callback via hotplug_handler API.
v3:
-deal with start up cpus in a more neat way as Igor suggested.
v2:
-just rebase.
Signed-off-by: Gu Zheng
---
hw/i386/pc.c | 26 +-
1 files changed, 25
Convert notifier based hotplug handle to hotplug_handler API,
and remove the unused AcpiCpuHotplug_add().
v2:
-remove the unused AcpiCpuHotplug_add().
Signed-off-by: Gu Zheng
---
hw/acpi/cpu_hotplug.c | 14 ++
hw/acpi/piix4.c | 14 ++
include/hw
Convert notifier based hotplug handle to hotplug_handler API.
Signed-off-by: Gu Zheng
---
hw/acpi/ich9.c | 14 +++---
include/hw/acpi/ich9.h |1 -
2 files changed, 3 insertions(+), 12 deletions(-)
diff --git a/hw/acpi/ich9.c b/hw/acpi/ich9.c
index 7b14bbb..c53d4ab 100644
-
---
v2:
-add errp argument to catch error.
-return error instead of aborting if cpu id is invalid.
-make acpi_cpu_plug_cb as a wrapper around AcpiCpuHotplug_add.
---
Signed-off-by: Gu Zheng
---
hw/acpi/cpu_hotplug.c | 17 +
include/hw/acpi/cpu_hotplug.h |3 +++
2
Previously we use cpu_added_notifiers to register cpu hotplug notifier callback
which is not able to pass/handle errors, so we switch it to unified hotplug
handler API which allows to pass errors and would allow to cancel device_add
in case of error.
Thanks very much for Igor's review and suggestio
On Tue, Sep 16, 2014 at 12:19:58PM +0200, Kevin Wolf wrote:
> Am 16.09.2014 um 12:10 hat Hu Tao geschrieben:
> > ping...
>
> Sorry, forgot to send the mail when I merged it. This is in master now.
Thank you very much!
Regards,
Hu
>
> Kevin
>
> > On Fri, Sep 12, 2014 at 05:22:45PM +0800, Hu Ta
> From: Markus Armbruster [mailto:arm...@redhat.com]
> Sent: Tuesday, September 16, 2014 11:12 PM
> To: Gonglei (Arei)
> Cc: qemu-devel@nongnu.org; kw...@redhat.com; Huangweidong (C);
> lcapitul...@redhat.com; stefa...@redhat.com
> Subject: Re: [Qemu-devel] [PATCH] hmp: fix memory leak at
> hmp_inf
> From: Markus Armbruster [mailto:arm...@redhat.com]
> Sent: Wednesday, September 17, 2014 12:16 AM
> To: Gerd Hoffmann
> Cc: Gonglei (Arei); Huangweidong (C); qemu-devel@nongnu.org
> Subject: Re: [Qemu-devel] [PATCH v2 0/2] usb: Don't use qerror_report
>
> Gerd Hoffmann writes:
>
> > Hi,
> >
> From: Markus Armbruster [mailto:arm...@redhat.com]
> Sent: Tuesday, September 16, 2014 10:26 PM
> To: Paolo Bonzini
> Cc: Gonglei (Arei); qemu-devel@nongnu.org; Huangweidong (C);
> kra...@redhat.com
> Subject: Re: [Qemu-devel] [PATCH v2 0/2] usb: Don't use qerror_report
>
> Paolo Bonzini writes
Hi,
Ping...
Is this acceptable?
Thanks,
zhanghailiang
On 2014/9/12 15:06, zhanghailiang wrote:
Add some trace events for easier debugging
Signed-off-by: zhanghailiang
---
hw/virtio/virtio-balloon.c | 6 ++
trace-events | 4
2 files changed, 10 insertions(+)
diff --
On 9 September 2014 01:11, Peter Maydell wrote:
> On 9 September 2014 07:47, Paolo Bonzini wrote:
>> Il 02/09/2014 12:24, Peter Maydell ha scritto:
>>> This patchset converts the two remaining legacy chardevs
>>> ('socket' and 'udp') to use the new-style parse/kind
>>> mechanisms, and removes all
On 09/16/2014 11:41 AM, Richard Henderson wrote:
>> In practice, only 3 to 7 are---hence my original attempt at using some
>> kind of FIFO caching:
>>
>>user mode, translation enabled
>>kernel mode, paging disabled
>>kernel mode, paging enabled
>>supervisor mode, paging disabled
>>
On 09/16/2014 11:49 AM, Peter Maydell wrote:
> On 16 September 2014 10:20, Tom Musta wrote:
>>
>> 1389 /* Compensate for very large offsets. */
>> 1390 if (add_off >= 0x8000) {
>> 1391 /* Most target env are smaller than 32k; none are larger
>> than 64k.
>> 1392
On Tue, 2014-09-16 at 14:51 -0600, Alex Williamson wrote:
> On Tue, 2014-09-16 at 00:01 +0200, Eric Auger wrote:
> > On 09/12/2014 01:05 AM, Christoffer Dall wrote:
> > > On Thu, Sep 11, 2014 at 04:51:14PM -0600, Alex Williamson wrote:
> > >> On Thu, 2014-09-11 at 15:23 -0700, Christoffer Dall wrot
On 09/16/2014 10:51 PM, Alex Williamson wrote:
> On Tue, 2014-09-16 at 00:01 +0200, Eric Auger wrote:
>> On 09/12/2014 01:05 AM, Christoffer Dall wrote:
>>> On Thu, Sep 11, 2014 at 04:51:14PM -0600, Alex Williamson wrote:
On Thu, 2014-09-11 at 15:23 -0700, Christoffer Dall wrote:
> On Thu,
On Tue, 2014-09-16 at 00:01 +0200, Eric Auger wrote:
> On 09/12/2014 01:05 AM, Christoffer Dall wrote:
> > On Thu, Sep 11, 2014 at 04:51:14PM -0600, Alex Williamson wrote:
> >> On Thu, 2014-09-11 at 15:23 -0700, Christoffer Dall wrote:
> >>> On Thu, Sep 11, 2014 at 04:14:09PM -0600, Alex Williamson
On 09/16/2014 12:31 PM, Paolo Bonzini wrote:
>> Change legacy_name to point to a detailed human-readable
>> description of the type?
>> E.g. "Ethernet 6-byte MAC Address, format: AA:BB:CC:DD:EE:FF"?
>
> If libvirt can cope with
>
> e1000.mac=str (Ethernet 6-byte MAC Address, format: AA:BB:CC:DD:
The Tuesday 16 Sep 2014 à 13:40:24 (-0600), Eric Blake wrote :
> On 09/16/2014 12:04 PM, Chrysostomos Nanakos wrote:
> > If event_notifier_init fails QEMU exits without printing
> > any error information to the user. This commit adds an error
> > message on failure:
> >
> > # qemu [...]
>
> Show
On 09/16/2014 12:04 PM, Chrysostomos Nanakos wrote:
> If event_notifier_init fails QEMU exits without printing
> any error information to the user. This commit adds an error
> message on failure:
>
> # qemu [...]
Showing the actual command line you used would be helpful.
> qemu: Failed to init
In vhdx_create_metadata(), we allocate 40 bytes to entry_buffer for
the various metadata table entries. However, we write out 64kB from
that buffer into the new file. Only write out the correct 40 bytes.
Signed-off-by: Jeff Cody
---
block/vhdx.c | 16
1 file changed, 8 inserti
On 09/16/2014 11:18 AM, Alex Bennée wrote:
> Do this mean when all this unwinding has occured cc->do_interrupt will
> no longer be needed at all entries will go in via
> $ARCH_cpu_exec_interrupt?
No.
Don't forget the call to do_interrupt in the block immediately following the
sigsetjmp. And even
On Tue, Sep 16, 2014 at 10:01:19PM +0300, Michael S. Tsirkin wrote:
> On Tue, Sep 16, 2014 at 08:31:08PM +0200, Paolo Bonzini wrote:
> > Il 16/09/2014 18:56, Michael S. Tsirkin ha scritto:
> > > On Tue, Sep 16, 2014 at 06:27:51PM +0200, Paolo Bonzini wrote:
> > >> Il 16/09/2014 18:26, Michael S. Ts
On Tue, Sep 16, 2014 at 08:31:08PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 18:56, Michael S. Tsirkin ha scritto:
> > On Tue, Sep 16, 2014 at 06:27:51PM +0200, Paolo Bonzini wrote:
> >> Il 16/09/2014 18:26, Michael S. Tsirkin ha scritto:
> >>> Right so types should be explicit.
> >>> If an arbit
On 16 September 2014 10:20, Tom Musta wrote:
>
> 1389 /* Compensate for very large offsets. */
> 1390 if (add_off >= 0x8000) {
> 1391 /* Most target env are smaller than 32k; none are larger than
> 64k.
> 1392 Simplify the logic here merely to offset by 0x7
On 09/16/2014 11:27 AM, Paolo Bonzini wrote:
> Il 16/09/2014 20:02, Richard Henderson ha scritto:
>> While we could probably fix this for ppc (using addis), it's not nearly so
>> easily fixable for arm -- without impacting performance anyway.
>>
>> Does 96k worth of TLBs really help that much? Are
Richard Henderson writes:
> Since do_interrupt_m68k_hardirq is no longer used outside
> op_helper.c, make it static.
>
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
> ---
> cpu-exec.c | 13 -
> target-m68k/cpu-qom.h | 1 +
> target-m68k/cpu.c |
Richard Henderson writes:
> Cc: Alexander Graf
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
> ---
> cpu-exec.c | 6 --
> target-s390x/cpu-qom.h | 1 +
> target-s390x/cpu.c | 1 +
> target-s390x/helper.c | 13 +
> 4 files changed, 15 inserti
27; into staging
> (2014-09-15 19:44:34 +0100)
>
> are available in the git repository at:
>
>
> git://git.kraxel.org/qemu tags/pull-sdl-20140916-1
>
> for you to fetch changes up to 0d61f7dcc6c04e7e825417c74e5036
Il 16/09/2014 18:56, Michael S. Tsirkin ha scritto:
> On Tue, Sep 16, 2014 at 06:27:51PM +0200, Paolo Bonzini wrote:
>> Il 16/09/2014 18:26, Michael S. Tsirkin ha scritto:
>>> Right so types should be explicit.
>>> If an arbitrary string isn't allowed, this should be documented.
>>> It's not great
Richard Henderson writes:
> Cc: Max Filippov
> Signed-off-by: Richard Henderson
> @@ -669,12 +669,6 @@ int cpu_exec(CPUArchState *env)
> cc->do_interrupt(cpu);
> next_tb = 0;
> }
> -#elif defined(TARGET_XTENSA)
> -
Il 16/09/2014 20:02, Richard Henderson ha scritto:
> While we could probably fix this for ppc (using addis), it's not nearly so
> easily fixable for arm -- without impacting performance anyway.
>
> Does 96k worth of TLBs really help that much? Are all 12 of them actually
> used? Can we use a mor
Signed-off-by: Markus Armbruster
---
include/hw/isa/pc87312.h | 3 ---
1 file changed, 3 deletions(-)
diff --git a/include/hw/isa/pc87312.h b/include/hw/isa/pc87312.h
index befc8bd..bf74470 100644
--- a/include/hw/isa/pc87312.h
+++ b/include/hw/isa/pc87312.h
@@ -47,13 +47,10 @@ typedef struct PC
Move device model attachment / detachment and the BlockDevOps device
model callbacks and their wrappers from BlockDriverState to
BlockBackend.
Signed-off-by: Markus Armbruster
---
block.c| 126 --
block/block-backend.c | 151 ++
I'll use BlockDriverAIOCB with block backends shortly, and the name is
going to fit badly there. It's a block layer thing anyway, not just a
block driver thing.
Signed-off-by: Markus Armbruster
---
block-migration.c | 2 +-
block.c | 151 ++---
Much more command code needs conversion. I'm converting these now
because they's using bdrv_dev_* functions, which I'm about to lift
into BlockBackend.
Signed-off-by: Markus Armbruster
---
blockdev.c | 20
1 file changed, 12 insertions(+), 8 deletions(-)
diff --git a/block
Much more command code needs conversion. I start with this one
because it's using bdrv_dev_* functions, which I'm about to lift into
BlockBackend.
While there, give bdrv_query_info() internal linkage.
Signed-off-by: Markus Armbruster
---
block/qapi.c | 15 ---
include/block
blockdev_init() always creates a DriveInfo, but only drive_new() fills
it in. qmp_blockdev_add() leaves it blank. This results in a drive
with type = IF_IDE, bus = 0, unit = 0. Screwed up in commit ee13ed1c.
Board initialization code looking for IDE drive (0,0) can pick up one
of these bogus dr
Add a BlockBackend member to TrimAIOCB, so ide_issue_trim_cb() can use
blk_aio_discard() instead of bdrv_aio_discard().
Signed-off-by: Markus Armbruster
---
hw/ide/core.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/hw/ide/core.c b/hw/ide/core.c
index f02ce4a..23964a
Commit 12c5674 turned it into a pointer to member blk.conf.
Signed-off-by: Markus Armbruster
---
hw/block/virtio-blk.c | 28 ++--
include/hw/virtio/virtio-blk.h | 1 -
2 files changed, 14 insertions(+), 15 deletions(-)
diff --git a/hw/block/virtio-blk.c b/hw/bl
Doesn't make a difference just yet, but it's the right thing to do.
Signed-off-by: Markus Armbruster
---
block/block-backend.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/block/block-backend.c b/block/block-backend.c
index d49c988..5646628 100644
--- a/block/block-back
On BlockBackend destruction, unref its BlockDriverState. Replaces the
callers' unrefs.
Signed-off-by: Markus Armbruster
---
block/block-backend.c | 6 ++
blockdev.c| 8 ++--
hw/block/xen_disk.c | 6 +++---
qemu-img.c| 35 +--
This is consistent with how VirtIOFOOConf variables are named
elsewhere, and makes blk available for BlockBackend variables.
Signed-off-by: Markus Armbruster
---
hw/block/dataplane/virtio-blk.c | 33 +-
hw/block/dataplane/virtio-blk.h | 2 +-
hw/block/virtio-blk.c
Signed-off-by: Markus Armbruster
---
block/block-backend.c | 1 -
blockdev.c| 3 +--
include/sysemu/blockdev.h | 1 -
3 files changed, 1 insertion(+), 4 deletions(-)
diff --git a/block/block-backend.c b/block/block-backend.c
index 5f796b4..141a31b 100644
--- a/block/block-bac
A block device consists of a frontend device model and a backend.
A block backend has a tree of block drivers doing the actual work.
The tree is managed by the block layer.
We currently use a single abstraction BlockDriverState both for tree
nodes and the backend as a whole. Drawbacks:
* Its AP
Make the BlockBackend own the DriveInfo. Change blockdev_init() to
return the BlockBackend instead of the DriveInfo.
Signed-off-by: Markus Armbruster
---
block.c | 2 --
block/block-backend.c | 38
blockdev.c| 73 ++
device_name[] can become non-empty only in bdrv_new_root() and
bdrv_move_feature_fields(). The latter is used only to undo damage
done by bdrv_swap(). The former is called only by blk_new_with_bs().
Therefore, when a BlockDriverState's device_name[] is non-empty, then
it's been created with a Blo
Commit 2d246f0 introduced DriveInfo member enable_auto_del to
distinguish DriveInfo created via drive_new() from DriveInfo created
via qmp_blockdev_add(). The latter no longer exist. Drop
enable_auto_del.
Signed-off-by: Markus Armbruster
---
blockdev.c| 11 +++
include/
BlockBackend's name space is separate only to keep the initial patches
simple. Time to merge the two.
Signed-off-by: Markus Armbruster
Reviewed-by: Benoît Canet
---
block.c | 11 +++
block/block-backend.c | 13 ++---
2 files changed, 9 insertions(+), 15 deletions(
My last attempt got bogged down because I tried to do a reasonably
complete job, and the complexity proved more than I could handle with
the limited amount of uninterrupted time available. This time, I'm
cutting BlockBackend off with an axe, leaving most of the work for
later.
Done in this series
Signed-off-by: Markus Armbruster
Reviewed-by: Benoît Canet
---
block-migration.c | 30 +++---
block.c | 9 -
blockdev.c| 31 +--
include/block/block.h | 2 --
monitor.c | 32 +-
The patch is big, but all it really does is replacing
dinfo->bdrv
by
blk_bs(blk_by_legacy_dinfo(dinfo))
The replacement is repetitive, but the conversion of device models to
BlockBackend is imminent, and will shorten it to just
blk_legacy_dinfo(dinfo).
Line wrapping muddies the waters
Creating an anonymous BDS can't fail. Make that obvious.
Signed-off-by: Markus Armbruster
Reviewed-by: Max Reitz
Reviewed-by: Benoît Canet
---
block.c | 28 +++-
block/iscsi.c | 2 +-
block/vvfat.c | 2 +-
blockdev.c| 2 +-
Signed-off-by: Markus Armbruster
Reviewed-by: Benoît Canet
---
block/block-backend.c | 15 +++
blockdev.c| 13 -
include/sysemu/blockdev.h | 1 -
stubs/Makefile.objs | 1 -
stubs/blockdev.c | 12
5 files changed, 15 insert
The pointer from BlockBackend to BlockDriverState is a strong
reference, managed with bdrv_ref() / bdrv_unref(), the back-pointer is
a weak one.
Convenience function blk_new_with_bs() creates a BlockBackend with its
BlockDriverState. Callers have to unref both. The commit after next
will relieve
On Tue, Sep 16, 2014 at 06:22:15PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 18:07, Marcelo Tosatti ha scritto:
> >> > The cpu_synchronize_all_states() call in kvmclock_vm_state_change() is
> >> > needed to make env->tsc up to date with the value on the source, right?
> > Its there to make sure t
Richard Henderson writes:
> Continuing the removal of ifdefs from cpu_exec.
>
> Cc: Andreas Färber
> Signed-off-by: Richard Henderson
Reviewed-by: Alex Bennée
> ---
> cpu-exec.c| 14 +-
> include/qom/cpu.h | 2 ++
> qom/cpu.c | 6 ++
> 3 files changed, 17 i
On 16 September 2014 10:46, Richard Henderson wrote:
> Instead of using structures, which imply some amount of overhead
> on certain ABIs, use pointer types.
>
> This actually reduces the size of the binaries vs a NON-debug
> build on ppc64 and x86_64, due to a reduction in the number of
> sign-ex
On 16 September 2014 10:46, Richard Henderson wrote:
> The combination of always_inline + artificial allows tiny inline
> functions to be written that do not interfere with debugging.
> In particular, gdb will not step into an artificial function.
>
> The always_inline attribute was introduced in
If event_notifier_init fails QEMU exits without printing
any error information to the user. This commit adds an error
message on failure:
# qemu [...]
qemu: Failed to initialize event notifier: Too many open files in system
Signed-off-by: Chrysostomos Nanakos
---
async.c |
v2->v3
--
* Remove errno usage and print the detailed message based on errno when
event_notifier_init() fails.
* Propagate error and return from iothread_complete() if aio_context_new()
fails.
* Return if !iothread->ctx from iothread_instance_finalize(), used by QOM
when object_unref(obj)
On 09/16/2014 10:20 AM, Tom Musta wrote:
> On 9/15/2014 10:03 AM, Paolo Bonzini wrote:
>> PowerPC TCG flushes the TLB on every IR/DR change, which basically
>> means on every user<->kernel context switch. Encode IR/DR in the
>> MMU index.
>>
>> This brings the number of TLB flushes down from ~9000
On Tue, Sep 16, 2014 at 06:50:38PM +0300, Chrysostomos Nanakos wrote:
> On Tue, Sep 16, 2014 at 05:45:16PM +0200, Paolo Bonzini wrote:
> > Il 16/09/2014 17:43, Chrysostomos Nanakos ha scritto:
> > >> > error_propagate(errp, local_error);
> > >> > return;
> > > Just to note that after propag
On Tue, Sep 16, 2014 at 10:23:03AM -0700, Peter Maydell wrote:
> On 16 September 2014 09:57, Michael S. Tsirkin wrote:
> > Could you pls tell me whether my patch works, in your testing?
> > My box which had such an old glib seems to be dead.
>
> Can you give me a Subject: line or a URL in patchwo
We've had a number of bugs slip through recently that would have
been caught by --enable-tcg-debug.
r~
Richard Henderson (2):
qemu/compiler: Define QEMU_ARTIFICIAL
tcg: Always enable TCGv type checking
include/qemu/compiler.h | 6
tcg/tcg.h | 89 +++
Instead of using structures, which imply some amount of overhead
on certain ABIs, use pointer types.
This actually reduces the size of the binaries vs a NON-debug
build on ppc64 and x86_64, due to a reduction in the number of
sign-extension insns.
Signed-off-by: Richard Henderson
---
tcg/tcg.h
The combination of always_inline + artificial allows tiny inline
functions to be written that do not interfere with debugging.
In particular, gdb will not step into an artificial function.
The always_inline attribute was introduced in gcc 4.2,
and the artificial attribute was introduced in gcc 4.3
Paolo Bonzini writes:
> Il 16/09/2014 16:08, Markus Armbruster ha scritto:
>> +if (bs->device_list.tqe_prev) {
>> QTAILQ_REMOVE(&bdrv_states, bs, device_list);
>> +bs->device_list.tqe_prev = NULL;
>>> >
>>> > I think a comments explaining the trick you are doin
emotes/qmp-unstable/queue/qmp' into staging
> (2014-09-15 19:44:34 +0100)
>
> are available in the git repository at:
>
>
> git://anongit.freedesktop.org/spice/qemu tags/pull-spice-20140916-2
>
> for you to fetch changes up to 07d49a53b6394941ed833486a3acb5c480d87db2:
>
On 16 September 2014 09:57, Michael S. Tsirkin wrote:
> Could you pls tell me whether my patch works, in your testing?
> My box which had such an old glib seems to be dead.
Can you give me a Subject: line or a URL in patchwork,
please? (I can't conveniently search by message-id.)
thanks
-- PMM
On 9/15/2014 10:03 AM, Paolo Bonzini wrote:
> PowerPC TCG flushes the TLB on every IR/DR change, which basically
> means on every user<->kernel context switch. Encode IR/DR in the
> MMU index.
>
> This brings the number of TLB flushes down from ~90 to ~5
> for starting up the Debian insta
On Tue, Sep 16, 2014 at 06:23:29PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 18:20, Peter Maydell ha scritto:
> >> > - bumping the minimum required version from 2.12 to 2.16. I suggest
> >> > bumping to the currently required version for Windows, which is 2.20
> >> > (released March 2009).
> > T
On Tue, Sep 16, 2014 at 05:42:16PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 17:14, Marcelo Tosatti ha scritto:
> > + /*
> > +* Make sure that CPU state is synchronized from KVM
> > +* once every VM state change callback has finished.
>
> Which other callback could affect the in-kernel
On Tue, Sep 16, 2014 at 01:48:24PM -0300, Marcelo Tosatti wrote:
> On Tue, Sep 16, 2014 at 06:22:15PM +0200, Paolo Bonzini wrote:
> > Il 16/09/2014 18:07, Marcelo Tosatti ha scritto:
> > >> > The cpu_synchronize_all_states() call in kvmclock_vm_state_change() is
> > >> > needed to make env->tsc up
On Tue, Sep 16, 2014 at 09:45:03AM -0700, Peter Maydell wrote:
> On 16 September 2014 09:33, Paolo Bonzini wrote:
> > Il 16/09/2014 18:28, Peter Maydell ha scritto:
> >>> > Though you would just drop support for "make check" on RHEL5. Even
> >>> > "make -k check" would roughly work.
> >> Can't we
On Tue, Sep 16, 2014 at 06:27:51PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 18:26, Michael S. Tsirkin ha scritto:
> > Right so types should be explicit.
> > If an arbitrary string isn't allowed, this should be documented.
> > It's not great as is: what's the format for macaddr? AA:BB:CC:DD:EE:FF
On Tue, Sep 16, 2014 at 06:22:15PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 18:07, Marcelo Tosatti ha scritto:
> >> > The cpu_synchronize_all_states() call in kvmclock_vm_state_change() is
> >> > needed to make env->tsc up to date with the value on the source, right?
> > Its there to make sure t
On 16 September 2014 09:33, Paolo Bonzini wrote:
> Il 16/09/2014 18:28, Peter Maydell ha scritto:
>>> > Though you would just drop support for "make check" on RHEL5. Even
>>> > "make -k check" would roughly work.
>> Can't we just put in the makefile and configure magic to skip
>> the test if the
On Tue, Sep 16, 2014 at 09:28:03AM -0700, Peter Maydell wrote:
> On 16 September 2014 09:23, Paolo Bonzini wrote:
> > Il 16/09/2014 18:20, Peter Maydell ha scritto:
> >>> > - bumping the minimum required version from 2.12 to 2.16. I suggest
> >>> > bumping to the currently required version for Wi
Il 16/09/2014 18:28, Peter Maydell ha scritto:
>> > Though you would just drop support for "make check" on RHEL5. Even
>> > "make -k check" would roughly work.
> Can't we just put in the makefile and configure magic to skip
> the test if the glib version is too old for it?
Yes, that's what Michae
On 16 September 2014 09:23, Paolo Bonzini wrote:
> Il 16/09/2014 18:20, Peter Maydell ha scritto:
>>> > - bumping the minimum required version from 2.12 to 2.16. I suggest
>>> > bumping to the currently required version for Windows, which is 2.20
>>> > (released March 2009).
>> The commit message
On Tue, Sep 16, 2014 at 12:45:43PM +0200, Paolo Bonzini wrote:
> Il 16/09/2014 12:37, Michael S. Tsirkin ha scritto:
> > str really should be for free-form strings.
> > It makes as much sense to call it as string
>
> Yes, that's why I said drive->str is a degradation. The question is,
> how impor
Il 16/09/2014 18:26, Michael S. Tsirkin ha scritto:
> Right so types should be explicit.
> If an arbitrary string isn't allowed, this should be documented.
> It's not great as is: what's the format for macaddr? AA:BB:CC:DD:EE:FF?
> aa:bb:cc:dd:ee:ff? aabbccddeeff? 0xaabbccddeeff?
> But just saying
zhanghailiang writes:
> Functions shouldn't return an error code and an Error object at the same time.
> Turn all these functions that returning Error object to void.
> We also judge if a function success or fail by reference to the *errp.
>
> Signed-off-by: zhanghailiang
> ---
> dump.c | 244
Il 16/09/2014 18:20, Peter Maydell ha scritto:
>> > - bumping the minimum required version from 2.12 to 2.16. I suggest
>> > bumping to the currently required version for Windows, which is 2.20
>> > (released March 2009).
> The commit message for a52d28afb suggests that this bump would be
> droppi
Il 16/09/2014 18:07, Marcelo Tosatti ha scritto:
>> > The cpu_synchronize_all_states() call in kvmclock_vm_state_change() is
>> > needed to make env->tsc up to date with the value on the source, right?
> Its there to make sure the pair
>
> env->tsc, s->clock = data.clock
>
> are relative to point
On 16 September 2014 08:43, Paolo Bonzini wrote:
> Glib recently introduced a robust way to run tests in a subprocess,
> which is used in test-qdev-global-props. However, we would like
> to have the same tests run with older versions of glib, and the
> older fork-based mechanisms works well enoug
Gerd Hoffmann writes:
> Hi,
>
>> > Gonglei (2):
>> > redirect.c: Don't use qerror_report()
>> > dev-network: Don't use qerror_report_err()
>
>> Hi, Gerd
>> Would you like to apply this patch series at present? Thanks!
>
> Picked it up now. Originally dropped due to review comments from pa
1 - 100 of 244 matches
Mail list logo
