> Am 11.01.2019 um 08:14 schrieb Vadim Rozenfeld :
>
>> On Thu, 2019-01-10 at 14:57 +0100, Peter Lieven wrote:
>>> Am 18.12.18 um 15:45 schrieb Peter Lieven:
Am 18.12.18 um 14:15 schrieb Vadim Rozenfeld:
Peter, I must be missing something here, but what exactly the
problem
i
Am 10.01.2019 um 20:18 hat Eric Blake geschrieben:
> Patches speak louder than words. This is my counter-proposal to
> Leonid's thread on how best to respresent the S_*iB macros in units.h,
> where my proposal is that we don't need them at all. (hence my subject
> line, even though it is completel
11.01.2019 10:54, Vladimir Sementsov-Ogievskiy wrote:
> 10.01.2019 23:51, Eric Blake wrote:
>> On 1/10/19 7:20 AM, Vladimir Sementsov-Ogievskiy wrote:
>>> drv_co_block_status digs bs->file for additional, more accurate search
>>> for hole inside region, reported as DATA by bs since 5daa74a6ebc.
>>
Am 10.01.2019 um 20:18 hat Eric Blake geschrieben:
> Instead of defining an integer to a default string value (where we
> have to be careful how we spelled the integer because of the use of
> stringify), populate a default integer value instead.
>
> Drop a useless stringify(0); a missing default i
Am 10.01.2019 um 14:20 hat Vladimir Sementsov-Ogievskiy geschrieben:
> drv_co_block_status digs bs->file for additional, more accurate search
> for hole inside region, reported as DATA by bs since 5daa74a6ebc.
>
> This accuracy is not free: assume we have qcow2 disk. Actually, qcow2
> knows, where
11.01.2019 13:41, Kevin Wolf wrote:
> Am 10.01.2019 um 14:20 hat Vladimir Sementsov-Ogievskiy geschrieben:
>> drv_co_block_status digs bs->file for additional, more accurate search
>> for hole inside region, reported as DATA by bs since 5daa74a6ebc.
>>
>> This accuracy is not free: assume we have q
Am 11.01.2019 um 12:40 hat Vladimir Sementsov-Ogievskiy geschrieben:
> 11.01.2019 13:41, Kevin Wolf wrote:
> > Am 10.01.2019 um 14:20 hat Vladimir Sementsov-Ogievskiy geschrieben:
> >> drv_co_block_status digs bs->file for additional, more accurate search
> >> for hole inside region, reported as DA
11.01.2019 15:21, Kevin Wolf wrote:
> Am 11.01.2019 um 12:40 hat Vladimir Sementsov-Ogievskiy geschrieben:
>> 11.01.2019 13:41, Kevin Wolf wrote:
>>> Am 10.01.2019 um 14:20 hat Vladimir Sementsov-Ogievskiy geschrieben:
drv_co_block_status digs bs->file for additional, more accurate search
Am 11.01.2019 um 13:59 hat Vladimir Sementsov-Ogievskiy geschrieben:
> 11.01.2019 15:21, Kevin Wolf wrote:
> > Am 11.01.2019 um 12:40 hat Vladimir Sementsov-Ogievskiy geschrieben:
> >> 11.01.2019 13:41, Kevin Wolf wrote:
> >>> Am 10.01.2019 um 14:20 hat Vladimir Sementsov-Ogievskiy geschrieben:
> >
Am 11.01.2019 um 14:19 hat Alberto Garcia geschrieben:
> On Wed 09 Jan 2019 04:34:10 PM CET, Alberto Garcia wrote:
> > On Wed 09 Jan 2019 12:01:44 PM CET, Stefan Hajnoczi wrote:
> >> The following QMP command leads to a crash when iothreads are used:
> >>
> >> { 'execute': 'device_del', 'argument
On Wed 09 Jan 2019 04:34:10 PM CET, Alberto Garcia wrote:
> On Wed 09 Jan 2019 12:01:44 PM CET, Stefan Hajnoczi wrote:
>> The following QMP command leads to a crash when iothreads are used:
>>
>> { 'execute': 'device_del', 'arguments': {'id': 'data'} }
>
> How did you reproduce this? Do you have
For what it's worth, this series seems to fix assertion failures related
to AioContext acquire/release I've seen.
On Fri 11 Jan 2019 02:24:16 PM CET, Kevin Wolf wrote:
>> >> The following QMP command leads to a crash when iothreads are used:
>> >>
>> >> { 'execute': 'device_del', 'arguments': {'id': 'data'} }
>> >
>> > How did you reproduce this? Do you have a test case?
>>
>> Ok, I finally reproduced it, t
On 1/10/19 9:18 PM, Eric Blake wrote:
> Set the framework up for declaring integer options with an integer
> default, instead of our current insane approach of requiring the
> default value to be given as a string (which then has to be reparsed
> at every use that wants a number). git grep '[^.]de
PixelFormat is used by "ui/console.h" and by "ui/qemu-pixman.h".
Create the new header "ui/pixelformat.h" to declare this structure,
and remove the forward declaration from "qemu/typedefs.h".
Signed-off-by: Philippe Mathieu-Daudé
---
RFC because "ui/console.h" has on license, so I added a defaul
Files requiring AudioState already include "audio_int.h".
To clean "qemu/typedefs.h", move the declaration to "audio_int.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
audio/audio_int.h | 4 ++--
include/qemu/typedefs.h | 1 -
2 files changed, 2 insertion
MouseTransformInfo is only used in "ui/console.h", there is no
need to expose it via "qemu/typedefs.h".
Move the declaration to "ui/console.h" (removing the forward
declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/devices.h| 1 +
include/qemu/typedefs.h | 1 -
include/ui/c
Files requiring QemuDmaBuf already include "ui/console.h".
To clean "qemu/typedefs.h", move the declaration to "ui/console.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/qemu/typedefs.h | 1 -
include/ui/console.h| 4 ++--
2 files changed, 2 inserti
Range is only used in "hw/pci-host/q35.h" in the MCHPCIState, and
in "hw/pci/pci.h" by pci_bus_get_w64_range(). Let them include
"qemu/range.h" directly to simplify "qemu/typedefs.h".
To clean "qemu/typedefs.h", move the declaration to "qemu/range.h"
(removing the forward declaration).
Signed-of
Files requiring HCIInfo already include "sysemu/bt.h".
To clean "qemu/typedefs.h", move the declaration to "sysemu/bt.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/qemu/typedefs.h | 1 -
include/sysemu/bt.h | 4 ++--
2 files changed, 2 insertions(+
From: Philippe Mathieu-Daudé
There are only three files requiring this typedef, let them
include "hw/ssi/ssi.h" directly to simplify "qemu/typedefs.h".
To clean "qemu/typedefs.h", move the forward declaration
to "hw/ssi/ssi.h".
Signed-off-by: Philippe Mathieu-Daudé
---
hw/arm/strongarm.h
Files requiring SerialState already include "hw/char/serial.h".
To clean "qemu/typedefs.h", move the declaration to "hw/char/serial.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/char/serial.h | 4 ++--
include/qemu/typedefs.h | 1 -
2 files changed
Files requiring PCMachineClass already include "hw/i386/pc.h".
To clean "qemu/typedefs.h", move the declaration to "hw/i386/pc.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/i386/pc.h| 4 ++--
include/qemu/typedefs.h | 1 -
2 files changed, 2 ins
Files requiring DisplayState/DisplaySurface already include "ui/console.h".
To clean "qemu/typedefs.h", move the declarations to "ui/console.h"
(removing DisplaySurface forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/qemu/typedefs.h | 2 --
include/ui/console.h| 6 ++
Files requiring SMBusDevice already include "hw/i2c/smbus.h".
To clean "qemu/typedefs.h", move the forward declaration
to "hw/i2c/smbus.h".
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/i2c/smbus.h | 2 ++
include/qemu/typedefs.h | 1 -
2 files changed, 2 insertions(+), 1 deletion(-)
d
Hi,
Since this file became a hot topic [*], I took previous cleanups
from last year and continue to clean it.
Regards,
Phil.
[*] "HACKING: Clarify the paragraph about typedefs"
https://lists.gnu.org/archive/html/qemu-devel/2019-01/msg02132.html
[*] "Fix duplicated typedefs to be able to compile
From: Philippe Mathieu-Daudé
Files requiring AllwinnerAHCIState already include "hw/ide/ahci.h".
To clean "qemu/typedefs.h", move the declaration to "hw/ide/ahci.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/ide/ahci.h | 4 ++--
include/qemu/typ
From: Philippe Mathieu-Daudé
There is only one file requiring this typedef (hw/arm/pxa.h),
let it include "hw/pcmcia.h" directly to simplify "qemu/typedefs.h".
To clean "qemu/typedefs.h", move the declaration to "hw/pcmcia.h"
(removing the forward declaration).
Signed-off-by: Philippe Mathieu-D
There are only three files requiring these typedefs, let them
include "hw/nvram/fw_cfg.h" directly to simplify "qemu/typedefs.h".
To clean "qemu/typedefs.h", move the forward declarations
to "hw/nvram/fw_cfg.h".
Signed-off-by: Philippe Mathieu-Daudé
---
include/hw/acpi/vmgenid.h | 1 +
include/
PS2State is only used in "hw/input/ps2.h", there is no
need to expose it via "qemu/typedefs.h".
To clean "qemu/typedefs.h", move the forward declaration
to "hw/input/ps2.h".
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: Thomas Huth
---
include/hw/input/ps2.h | 2 ++
include/qemu/typedefs
Hi,
Since this file became a hot topic [*], I took previous cleanups
from last year and continue to clean it.
Regards,
Phil.
[*] "HACKING: Clarify the paragraph about typedefs"
https://lists.gnu.org/archive/html/qemu-devel/2019-01/msg02132.html
[*] "Fix duplicated typedefs to be able to compile
On 09/01/19 12:01, Stefan Hajnoczi wrote:
> g_free(data);
> +
> +tgm->restart_pending--;
> +aio_wait_kick();
> }
>
> static void throttle_group_restart_queue(ThrottleGroupMember *tgm, bool
> is_write)
> @@ -430,6 +433,8 @@ static void
> throttle_group_restart_queue(ThrottleGroupM
On 12.11.18 08:06, Marc Olson wrote:
> If 'once' is specified, the rule should execute just once, regardless if
> it is supposed to return an error or not. Take the example where you
> want the first IO to an LBA to succeed, but subsequent IOs to fail. You
> could either use state transitions, or c
On 1/11/19 3:04 PM, Philippe Mathieu-Daudé wrote:
> Hi,
>
> Since this file became a hot topic [*], I took previous cleanups
> from last year and continue to clean it.
>
> Regards,
>
> Phil.
>
> [*] "HACKING: Clarify the paragraph about typedefs"
> https://lists.gnu.org/archive/html/qemu-devel/
On 12.11.18 08:06, Marc Olson wrote:
> Break out the more common parts of the BlkdebugRule struct, and make
> rule_check() more explicit about operating only on error injection types
> so that additional rule types can be added in the future.
>
> Signed-off-by: Marc Olson
> ---
> block/blkdebug.
On 12.11.18 08:06, Marc Olson wrote:
> Add a new rule type for blkdebug that instead of returning an error, can
> inject latency to an IO.
>
> Signed-off-by: Marc Olson
> ---
> block/blkdebug.c | 79
> +++---
> docs/devel/blkdebug.txt| 35 ++
Am 10.01.2019 um 16:03 hat Alberto Garcia geschrieben:
> This fixes the following crash:
>
> { "execute": "blockdev-add",
> "arguments": {"driver": "null-co", "node-name": "hd0"}}
> { "execute": "object-add",
> "arguments": {"qom-type": "iothread", "id": "iothread0"}}
> { "execute": "x-blockde
On Fri 11 Jan 2019 04:02:13 PM CET, Kevin Wolf wrote:
>> @@ -2568,7 +2579,10 @@ static void scsi_block_realize(SCSIDevice *dev, Error
>> **errp)
>> }
>>
>> /* check we are using a driver managing SG_IO (version 3 and after) */
>> +ctx = blk_get_aio_context(s->qdev.conf.blk);
>> +
Am 10.01.2019 um 16:03 hat Alberto Garcia geschrieben:
> As discussed yesterday, this series acquires the AioContext in the
> _realize() functions of several devices before making use of their
> block backends. This fixes at least a couple of crashes (in virtio-blk
> and scsi).
Apart from the inco
On Fri 11 Jan 2019 04:09:17 PM CET, Kevin Wolf wrote:
> Am 10.01.2019 um 16:03 hat Alberto Garcia geschrieben:
>> As discussed yesterday, this series acquires the AioContext in the
>> _realize() functions of several devices before making use of their
>> block backends. This fixes at least a couple
Am 11.01.2019 um 16:05 hat Alberto Garcia geschrieben:
> On Fri 11 Jan 2019 04:02:13 PM CET, Kevin Wolf wrote:
> >> @@ -2568,7 +2579,10 @@ static void scsi_block_realize(SCSIDevice *dev,
> >> Error **errp)
> >> }
> >>
> >> /* check we are using a driver managing SG_IO (version 3 and af
On 1/11/19 1:50 AM, Markus Armbruster wrote:
>>> We have 54 of them. I count six uses:
>>>
>>> block/qcow2.h:#define QCOW_MAX_REFTABLE_SIZE S_8MiB
>>> block/qcow2.h:#define QCOW_MAX_L1_SIZE S_32MiB
>>> block/qcow2.h:#define DEFAULT_L2_CACHE_MAX_SIZE S_32MiB
>>> block/qcow2.h:#
On 1/11/19 1:54 AM, Vladimir Sementsov-Ogievskiy wrote:
>>
>> How much performance can we buy back without any knobs at all, if we
>> just taught posix-file.c to cache lseek() results? That is, when
>> visiting a file sequentially, if lseek(fd, 0, SEEK_HOLE) returns EOF on
>> our first block stat
On 1/11/19 10:02 AM, Eric Blake wrote:
> Even a dumb most-recent use cache will speed this up: both the second
> and third queries above can be avoided because we know that both 0x4
> and 0x3 the second query at 0x4 can be skipped (0x4 is
> between our most recent lseek at 0x2
11.01.2019 16:15, Kevin Wolf wrote:
> Am 11.01.2019 um 13:59 hat Vladimir Sementsov-Ogievskiy geschrieben:
>> 11.01.2019 15:21, Kevin Wolf wrote:
>>> Am 11.01.2019 um 12:40 hat Vladimir Sementsov-Ogievskiy geschrieben:
11.01.2019 13:41, Kevin Wolf wrote:
> Am 10.01.2019 um 14:20 hat Vladim
On 1/11/19 8:14 AM, Leonid Bloch wrote:
> On 1/10/19 9:18 PM, Eric Blake wrote:
>> Set the framework up for declaring integer options with an integer
>> default, instead of our current insane approach of requiring the
>> default value to be given as a string (which then has to be reparsed
>> at eve
11.01.2019 19:02, Eric Blake wrote:
> On 1/11/19 1:54 AM, Vladimir Sementsov-Ogievskiy wrote:
>
>>>
>>> How much performance can we buy back without any knobs at all, if we
>>> just taught posix-file.c to cache lseek() results? That is, when
>>> visiting a file sequentially, if lseek(fd, 0, SEEK_
On 01/11/19 15:08, Philippe Mathieu-Daudé wrote:
> There are only three files requiring these typedefs, let them
> include "hw/nvram/fw_cfg.h" directly to simplify "qemu/typedefs.h".
>
> To clean "qemu/typedefs.h", move the forward declarations
> to "hw/nvram/fw_cfg.h".
>
> Signed-off-by: Philipp
On 1/11/19 4:38 AM, Kevin Wolf wrote:
> Am 10.01.2019 um 20:18 hat Eric Blake geschrieben:
>> Instead of defining an integer to a default string value (where we
>> have to be careful how we spelled the integer because of the use of
>> stringify), populate a default integer value instead.
>>
>> Drop
The use of a variable named 'exp' prevents includes to import .
Rename it to avoid:
qemu-nbd.c:64:19: error: ‘exp’ redeclared as different kind of symbol
static NBDExport *exp;
^~~
In file included from /usr/include/features.h:428,
from /usr/include/
Cc'ing qemu-triv...@nongnu.org
On Fri, Jan 11, 2019 at 5:35 PM Philippe Mathieu-Daudé
wrote:
>
> The use of a variable named 'exp' prevents includes to import .
>
> Rename it to avoid:
>
> qemu-nbd.c:64:19: error: ‘exp’ redeclared as different kind of symbol
>static NBDExport *exp;
>
Whenever the allocation length of a SCSI request is shorter than the size of the
VPD page list, page_idx is used blindly to index into r->buf. Even though
the stores in the insertion sort are protected against overflows, the same is
not
true of the reads and the final store of 0xb0.
This basical
On 1/11/19 10:09 AM, Vladimir Sementsov-Ogievskiy wrote:
I suggested one: Pass large contiguous allocated ranges to the protocol
driver, but just assume that the allocation status is correct in the
format layer if they are small.
>>>
>>> So, for fully allocated image, we will call l
On 1/11/19 10:22 AM, Vladimir Sementsov-Ogievskiy wrote:
>> Even a dumb most-recent use cache will speed this up: both the second
>> and third queries above can be avoided because we know that both 0x4
>> and 0x3 the second query at 0x4 can be skipped (0x4 is
>> between our most re
On 1/11/19 10:35 AM, Philippe Mathieu-Daudé wrote:
> The use of a variable named 'exp' prevents includes to import .
>
> Rename it to avoid:
>
> qemu-nbd.c:64:19: error: ‘exp’ redeclared as different kind of symbol
>static NBDExport *exp;
> ^~~
> In file included from
On 11.01.2019 17:04, Eric Blake wrote:
> On 1/11/19 10:09 AM, Vladimir Sementsov-Ogievskiy wrote:
>
> I suggested one: Pass large contiguous allocated ranges to the protocol
> driver, but just assume that the allocation status is correct in the
> format layer if they are small.
On 11/01/19 15:08, Philippe Mathieu-Daudé wrote:
> PixelFormat is used by "ui/console.h" and by "ui/qemu-pixman.h".
ui/console.h already includes ui/qemu-pixman.h, just move it there?
Paolo
> Create the new header "ui/pixelformat.h" to declare this structure,
> and remove the forward declaration
On 12/21/18 3:35 AM, John Snow wrote:
> Presently, we abort transactions in the same order they were processed in.
> Bitmap commands, though, attempt to restore backup data structures on abort.
>
> That's not valid, they need to be aborted in reverse chronological order.
>
> Replace the QSIMPLEQ
On 1/11/19 6:23 PM, Eric Blake wrote:
> On 1/11/19 8:14 AM, Leonid Bloch wrote:
>> On 1/10/19 9:18 PM, Eric Blake wrote:
>>> Set the framework up for declaring integer options with an integer
>>> default, instead of our current insane approach of requiring the
>>> default value to be given as a str
We define 54 macros for the powers of two >= 1024. We use six, in six
macro definitions. Four of them could just as well use the common MiB
macro, so do that. The remaining two can't, because they get passed
to stringify. Replace the macro by the literal number there.
Slightly harder to read in
Back in September, Leonid Block added a whole bunch of macros (commit
540b8492618) to improve readability of qcow2.h a bit (commit
b6a95c6d100). He later used them to fix the "vdi" driver's parameter
cluster_size's default value (commit 3dd5b8f4718). He has now
proposed a further patch[1] to auto
On 1/11/19 1:14 PM, Markus Armbruster wrote:
> We define 54 macros for the powers of two >= 1024. We use six, in six
> macro definitions. Four of them could just as well use the common MiB
> macro, so do that. The remaining two can't, because they get passed
> to stringify. Replace the macro by
On 1/11/19 11:52 AM, Eric Blake wrote:
> On 12/21/18 3:35 AM, John Snow wrote:
>> Presently, we abort transactions in the same order they were processed in.
>> Bitmap commands, though, attempt to restore backup data structures on abort.
>>
>> That's not valid, they need to be aborted in reverse chr
Testing success paths is important, but it's also nice to highlight
expected failure handling, to show that we don't crash, and so that
upcoming tests that change behavior can demonstrate the resulting
effects on error paths.
Add the following errors:
Attempting to export without a running server
Since we already forbid other nbd-server commands when not
in the right state, it is unlikely that any caller was relying
on a second stop to behave as a silent no-op. Update iotest
223 to show the improved behavior.
Signed-off-by: Eric Blake
---
v3: new patch
---
blockdev-nbd.c |
Or rather, move its functionality into nbd-server-add. And as
a side effect, teach qemu-nbd how to export a persistent bitmap
without having to go through a qemu process and several QMP
commands.
Based-on: <20181221093529.23855-1-js...@redhat.com>
[jsnow: 0/11 bitmaps: remove x- prefix from QMP a
We only have one caller that wants to export a bitmap name,
which it does right after creation of the export. But there is
still a brief window of time where an NBD client could see the
export but not the dirty bitmap, which a robust client would
have to interpret as meaning the entire image should
Now that nbd-server-add can do the same functionality (well, other
than making the exported bitmap name different than the underlying
bitamp - but we argued that was not essential, since it is just as
easy to create a new non-persistent bitmap with the desired name),
we no longer need the experimen
Our initial implementation of x-nbd-server-add-bitmap put
in a restriction because of incremental backups: in that
usage, we are exporting one qcow2 file (the temporary overlay
target of a blockdev-backup sync:none job) and a dirty bitmap
owned by a second qcow2 file (the source of the
blockdev-bac
The existing NBD code had a weird split where nbd_export_new()
created an export but did not add it to the list of exported
names until a later nbd_export_set_name() came along and grabbed
a second reference on the object; later, the first call to
nbd_export_close() drops the second reference while
Having to fire up qemu, then use QMP commands for nbd-server-start
and nbd-server-add, just to expose a persistent dirty bitmap, is
rather tedious. Make it possible to expose a dirty bitmap using
just qemu-nbd (of course, for now this only works when qemu-nbd is
visiting a BDS formatted as qcow2).
With the experimental x-nbd-server-add-bitmap command, there was
a window of time where an NBD client could see the export but not
the associated dirty bitmap, which can cause a client that planned
on using the dirty bitmap to be forced to treat the entire image
as dirty as a safety fallback. Furt
On 1/11/19 1:47 PM, Eric Blake wrote:
> Or rather, move its functionality into nbd-server-add. And as
> a side effect, teach qemu-nbd how to export a persistent bitmap
> without having to go through a qemu process and several QMP
> commands.
>
> Based-on: <20181221093529.23855-1-js...@redhat.com>
On 1/4/19 5:50 PM, Eric Blake wrote:
>> hmm, as I understand, even if this compatibility exists, it's not a part
>> of standard and nothing about off_t size in POSIX..
>
> off_t allows you to run on older systems with 32-bit offsets and newer
> systems with 64-bit offsets; but these days, even on
74 matches
Mail list logo
