On Thu, Mar 11, 2021 at 4:18 PM Bin Meng wrote:
>
> On Sat, Mar 6, 2021 at 2:01 PM Bin Meng wrote:
> >
> > From: Bin Meng
> >
> > Per SST25VF016B datasheet [1], SST flash requires a dummy byte after
> > the address bytes. Note only SPI mode is supported by SST flashes.
> >
> > [1]
From: David Edmondson
Rather than having a device specific debug implementation in
pflash_cfi01.c and pflash_cfi02.c, use the standard tracing facility.
Signed-off-by: David Edmondson
Reviewed-by: Philippe Mathieu-Daudé
Message-Id: <20210216142721.1985543-2-david.edmond...@oracle.com>
[PMD:
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: David Edmondson
Reviewed-by: Bin Meng
Message-Id: <20210310170528.1184868-9-phi...@redhat.com>
---
hw/block/pflash_cfi02.c | 8
1 file changed, 8 insertions(+)
diff --git a/hw/block/pflash_cfi02.c b/hw/block/pflash_cfi02.c
index
There is multiple places resetting the internal state machine.
Factor the code out in a new pflash_reset_state_machine() method.
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: David Edmondson
Reviewed-by: Bin Meng
Message-Id: <20210310170528.1184868-8-phi...@redhat.com>
---
The same pattern is used when setting the flash in READ_ARRAY mode:
- Set the state machine command to READ_ARRAY
- Reset the write_cycle counter
- Reset the memory region in ROMD
Refactor the current code by extracting this pattern.
It is used three times:
- When the timer expires and not in
Use the 'mode_read_array' event when we set the device in such
mode, and use the 'reset' event in DeviceReset handler.
Reviewed-by: Bin Meng
Reviewed-by: David Edmondson
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20210310170528.1184868-10-phi...@redhat.com>
---
hw/block/pflash_cfi01.c
From: David Edmondson
PFlashCFI01.ro is a bool, declare it as such.
Signed-off-by: David Edmondson
Reviewed-by: Philippe Mathieu-Daudé
Message-Id: <20210216142721.1985543-3-david.edmond...@oracle.com>
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: Bin Meng
---
hw/block/pflash_cfi01.c |
There is only one call to pflash_register_memory() with
rom_mode == false. As we want to modify pflash_register_memory()
in the next patch, open-code this trivial function in place for
the 'rom_mode == false' case.
Reviewed-by: Bin Meng
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: David
Fill the CFI table in out of DeviceRealize() in a new function:
pflash_cfi02_fill_cfi_table().
Reviewed-by: Bin Meng
Reviewed-by: David Edmondson
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20210310170528.1184868-4-phi...@redhat.com>
---
hw/block/pflash_cfi02.c | 193
There is only one call to pflash_setup_mappings(). Convert 'rom_mode'
to boolean and set it to true directly within pflash_setup_mappings().
Reviewed-by: Bin Meng
Signed-off-by: Philippe Mathieu-Daudé
Reviewed-by: David Edmondson
Message-Id: <20210310170528.1184868-5-phi...@redhat.com>
---
Fill the CFI table in out of DeviceRealize() in a new function:
pflash_cfi01_fill_cfi_table().
Reviewed-by: Bin Meng
Reviewed-by: David Edmondson
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20210310170528.1184868-3-phi...@redhat.com>
---
hw/block/pflash_cfi01.c | 140
The following changes since commit 2615a5e433aeb812c300d3a48e1a88e1303e2339:
Merge remote-tracking branch
'remotes/stefanha-gitlab/tags/block-pull-request' into staging (2021-03-15
19:23:00 +)
are available in the Git repository at:
https://github.com/philmd/qemu.git
We are going to move this code, fix its style first.
Reviewed-by: Bin Meng
Reviewed-by: David Edmondson
Signed-off-by: Philippe Mathieu-Daudé
Message-Id: <20210310170528.1184868-2-phi...@redhat.com>
---
hw/block/pflash_cfi01.c | 36
hw/block/pflash_cfi02.c
On 3/10/21 6:05 PM, Philippe Mathieu-Daudé wrote:
> I remembered this almost 2 years old series while reviewing
> David Edmondson's patches... (which are included at the end).
>
> Basically we move things around to make the code easier to maintain.
>
> David Edmondson (2):
>
On Mar 10 10:53, Klaus Jensen wrote:
> From: Klaus Jensen
>
> This is v5 of a series that adds support for metadata and end-to-end
> data protection.
>
> First, on the subject of metadata, in v1, support was restricted to
> extended logical blocks, which was pretty trivial to implement, but
>
On Mon, 15 Mar 2021 at 09:51, Stefan Hajnoczi wrote:
>
> The following changes since commit 6157b0e19721aadb4c7fdcfe57b2924af6144b14:
>
> Merge remote-tracking branch 'remotes/vivier2/tags/linux-user-for-6.0-pull-=
> request' into staging (2021-03-14 17:47:49 +)
>
> are available in the Git
15.03.2021 09:06, Roman Kagan wrote:
As the reconnect logic no longer interferes with drained sections, it
appears unnecessary to explicitly manipulate the in_flight counter.
Fixes: 5ad81b4946 ("nbd: Restrict connection_co reentrance")
And here you actually allow qemu_aio_coroutine_enter()
15.03.2021 09:06, Roman Kagan wrote:
The reconnection logic doesn't need to stop while in a drained section.
Moreover it has to be active during the drained section, as the requests
that were caught in-flight with the connection to the server broken can
only usefully get drained if the
On Mon, Mar 15, 2021 at 07:41:32PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> 15.03.2021 09:06, Roman Kagan wrote:
> > Document (via a comment and an assert) that
> > nbd_client_detach_aio_context and nbd_client_attach_aio_context_bh run
> > in the desired aio_context
> >
> > Signed-off-by:
15.03.2021 09:06, Roman Kagan wrote:
The reconnection logic doesn't need to stop while in a drained section.
Moreover it has to be active during the drained section, as the requests
that were caught in-flight with the connection to the server broken can
only usefully get drained if the
On Mar 14 19:31, Dmitry Fomichev wrote:
> LGTM,
> Reviewed-by: Dmitry Fomichev
>
Thanks Dmitry!
If you could give [10/13], [11/13] and [12/13] a look-see as well that
would be awesome :)
-
k
signature.asc
Description: PGP signature
On 3/15/21 12:45 PM, Daniel P. Berrangé wrote:
> This has been hardcoded to "false" since 2.10.0, since secrets required
> to unlock block devices are now always provided upfront instead of using
up front
> interactive prompts.
>
> Reviewed-by: Thomas Huth
> Signed-off-by: Daniel P. Berrangé
On 3/15/21 12:45 PM, Daniel P. Berrangé wrote:
> The code comment suggests removing QAPIEvent_(str|lookup) symbols too,
> however, these are both auto-generated as standard for any enum in
> QAPI. As such it they'll exist whether we use them or not.
>
> Reviewed-by: Thomas Huth
> Signed-off-by:
iOS hosts do not have these defined so we fallback to the
default behaviour.
Co-authored-by: Warner Losh
Reviewed-by: Peter Maydell
Signed-off-by: Joelle van Dyne
---
block/file-posix.c | 21 +
1 file changed, 9 insertions(+), 12 deletions(-)
diff --git
On Darwin (iOS), there are no system level APIs for directly accessing
host block devices. We detect this at configure time.
Signed-off-by: Joelle van Dyne
---
meson.build | 6 +-
qapi/block-core.json | 10 +++---
block/file-posix.c | 33 ++---
3
Some BSD platforms do not have this header.
Reviewed-by: Peter Maydell
Reviewed-by: Philippe Mathieu-Daudé
Signed-off-by: Joelle van Dyne
---
meson.build | 1 +
block.c | 2 +-
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/meson.build b/meson.build
index
The same information is available via the 'recording' and 'busy' fields.
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Daniel P. Berrangé
---
block/dirty-bitmap.c | 38
docs/system/deprecated.rst | 7 -
docs/system/removed-features.rst | 7 +
This has been hardcoded to "false" since 2.10.0, since secrets required
to unlock block devices are now always provided upfront instead of using
interactive prompts.
Reviewed-by: Thomas Huth
Signed-off-by: Daniel P. Berrangé
---
block/qapi.c | 1 -
The 'ide-hd' and 'ide-cd' devices provide suitable alternatives.
Reviewed-by: Thomas Huth
Signed-off-by: Daniel P. Berrangé
---
docs/qdev-device-use.txt | 2 +-
docs/system/deprecated.rst | 6 -
docs/system/removed-features.rst | 8 +++
hw/i386/pc.c
This only makes sense conceptually when used with listener chardevs.
Reviewed-by: Marc-André Lureau
Signed-off-by: Daniel P. Berrangé
---
chardev/char-socket.c| 12
docs/system/deprecated.rst | 6 --
docs/system/removed-features.rst | 6 ++
3 files
The newer 'query-cpus-fast' command avoids side effects on the guest
execution. Note that some of the field names are different in the
'query-cpus-fast' command.
Reviewed-by: Wainer dos Santos Moschetta
Tested-by: Wainer dos Santos Moschetta
Signed-off-by: Daniel P. Berrangé
---
The 'host_device' and 'host_cdrom' drivers must be used instead.
Reviewed-by: Eric Blake
Signed-off-by: Daniel P. Berrangé
---
block/file-posix.c | 17 ++---
docs/system/deprecated.rst | 7 ---
docs/system/removed-features.rst | 7 +++
The 'scsi-hd' and 'scsi-cd' devices provide suitable alternatives.
Reviewed-by: Thomas Huth
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst | 9 -
docs/system/removed-features.rst | 6
hw/i386/pc.c | 1 -
hw/scsi/scsi-disk.c |
The same data is available in the 'BlockDeviceInfo' struct.
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Daniel P. Berrangé
---
block/qapi.c| 5 -
docs/system/deprecated.rst | 13 -
This is only semantically useful for QMP.
Reviewed-by: Dr. David Alan Gilbert
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst | 7 ---
docs/system/removed-features.rst | 6 ++
monitor/monitor.c| 4 ++--
qemu-options.hx | 5 +++--
The VNC ACL concept has been replaced by the pluggable "authz" framework
which does not use monitor commands.
Reviewed-by: Thomas Huth
Reviewed-by: Dr. David Alan Gilbert
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst | 16 ---
docs/system/removed-features.rst | 13
Reviewed-by: Thomas Huth
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst | 6 -
docs/system/removed-features.rst | 6 +
hw/core/machine-qmp-cmds.c | 41
qapi/machine.json| 22 -
4 files
The generic 'migrate_set_parameters' command handle all types of param.
Only the QMP commands were documented in the deprecations page, but the
rationale for deprecating applies equally to HMP, and the replacements
exist. Furthermore the HMP commands are just shims to the QMP commands,
so
The code comment suggests removing QAPIEvent_(str|lookup) symbols too,
however, these are both auto-generated as standard for any enum in
QAPI. As such it they'll exist whether we use them or not.
Reviewed-by: Thomas Huth
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst |
This is an update to
https://lists.gnu.org/archive/html/qemu-devel/2021-02/msg07558.html
The following features have been deprecated for well over the 2
release cycle we promise
``-drive file=3Djson:{...{'driver':'file'}}`` (since 3.0)
``-vnc acl`` (since 4.0.0)
``-mon
15.03.2021 09:06, Roman Kagan wrote:
Document (via a comment and an assert) that
nbd_client_detach_aio_context and nbd_client_attach_aio_context_bh run
in the desired aio_context
Signed-off-by: Roman Kagan
---
block/nbd.c | 12
1 file changed, 12 insertions(+)
diff --git
Am 15.03.2021 um 16:26 hat Markus Armbruster geschrieben:
> Kevin Wolf writes:
>
> > Am 13.03.2021 um 14:40 hat Markus Armbruster geschrieben:
> >> Markus Armbruster writes:
> >>
> >> > Paolo Bonzini writes:
> >> >
> >> >> On 11/03/21 15:08, Markus Armbruster wrote:
> >> I would rather
15.03.2021 09:06, Roman Kagan wrote:
Use nbd_client_connecting_wait uniformly all over the block/nbd.c.
While at this, drop the redundant check for nbd_client_connecting_wait
in reconnect_delay_timer_init, as all its callsites do this check too.
Signed-off-by: Roman Kagan
Reviewed-by:
15.03.2021 09:06, Roman Kagan wrote:
When the NBD connection is being torn down, the connection thread gets
canceled and "detached", meaning it is about to get freed.
If this happens while the connection coroutine yielded waiting for the
connection thread to complete, when it resumes it may
On Mon, Mar 08, 2021 at 10:43:54AM -0600, Connor Kuehl wrote:
> On 3/1/21 11:28 AM, Connor Kuehl wrote:
> > The contents of this patch were initially developed and posted by Han
> > Han[1], however, it appears the original patch was not applied. Since
> > then, the relevant documentation has been
Kevin Wolf writes:
> Am 13.03.2021 um 14:40 hat Markus Armbruster geschrieben:
>> Markus Armbruster writes:
>>
>> > Paolo Bonzini writes:
>> >
>> >> On 11/03/21 15:08, Markus Armbruster wrote:
>> I would rather keep the OptsVisitor here. Do the same check for JSON
>> syntax that
09.03.2021 20:08, Alberto Garcia wrote:
Based-on: <20201127144522.29991-1-vsement...@virtuozzo.com>
Hello,
here's the third version of the patches that allow replacing bs->file
using (x-)blockdev-reopen. You can read more details here:
Am 15.03.2021 um 15:15 hat Markus Armbruster geschrieben:
> Kevin Wolf writes:
>
> > Am 13.03.2021 um 13:30 hat Markus Armbruster geschrieben:
> >> Paolo Bonzini writes:
> >>
> >> > On 13/03/21 08:40, Markus Armbruster wrote:
> >> >>> +if (!user_creatable_add_from_str(optarg,
15.03.2021 12:58, Max Reitz wrote:
On 12.03.21 19:43, Vladimir Sementsov-Ogievskiy wrote:
12.03.2021 21:15, Max Reitz wrote:
On 05.03.21 18:35, Vladimir Sementsov-Ogievskiy wrote:
Compressed writes are unaligned to 512, which works very slow in
O_DIRECT mode. Let's use the cache.
Kevin Wolf writes:
> Am 13.03.2021 um 13:30 hat Markus Armbruster geschrieben:
>> Paolo Bonzini writes:
>>
>> > On 13/03/21 08:40, Markus Armbruster wrote:
>> >>> +if (!user_creatable_add_from_str(optarg, _err)) {
>> >>> +if (local_err) {
>> >>> +
13.03.2021 08:51, Mahmoud Mandour wrote:
Thank you for the fast review and I'm sorry for the silly and obvious style
errors. Unfortunately I did not notice the section on using the checkpatch
script in the Contributing page on the wiki before committing. But I assure
you that such errors will
15.03.2021 08:36, Roman Kagan wrote:
On Fri, Mar 12, 2021 at 03:35:25PM +0300, Vladimir Sementsov-Ogievskiy wrote:
10.03.2021 12:32, Roman Kagan wrote:
NBD connect coroutine takes an extra in_flight reference as if it's a
request handler. This prevents drain from completion until the
The following changes since commit 6157b0e19721aadb4c7fdcfe57b2924af6144b14:
Merge remote-tracking branch
'remotes/vivier2/tags/linux-user-for-6.0-pull-request' into staging (2021-03-14
17:47:49 +)
are available in the Git repository at:
git://repo.or.cz/qemu/kevin.git
On 24/02/2021 14.11, Daniel P. Berrangé wrote:
The 'scsi-hd' and 'scsi-cd' devices provide suitable alternatives.
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst | 9 -
docs/system/removed-features.rst | 6
hw/i386/pc.c | 1 -
On Mon, 15 Mar 2021 at 12:09, Kevin Wolf wrote:
> Hm, okay. I guess this means we can't test that a deprecated option
> still works. I'll remove the test and resend.
You'd need a mechanism for "suppress the warning because
we expect it". (Often we do this in a kind of ad-hoc way by
having
On Mon, 15 Mar 2021 at 11:34, Paolo Bonzini wrote:
>
> On 07/03/21 23:26, Philippe Mathieu-Daudé wrote:
> > TYPE_PFLASH_CFI01 is a TYPE_SYS_BUS_DEVICE which registers its romd
> > MemoryRegion with sysbus_init_mmio(), so we can use the generic
> > sysbus_mmio_get_region() to get the region, no
Am 12.03.2021 um 19:56 hat Peter Maydell geschrieben:
> On Thu, 11 Mar 2021 at 14:48, Kevin Wolf wrote:
> >
> > The following changes since commit d689ecad073e0289afa8ca863e45879d719e5c21:
> >
> > Merge remote-tracking branch 'remotes/nvme/tags/nvme-next-pull-request'
> > into staging
On 15/03/21 12:38, Dr. David Alan Gilbert wrote:
* Kevin Wolf (kw...@redhat.com) wrote:
Am 15.03.2021 um 10:39 hat Markus Armbruster geschrieben:
Paolo Bonzini writes:
On 13/03/21 14:28, Markus Armbruster wrote:
Kevin Wolf writes:
This switches the HMP command object_add from a
On 2/24/21 7:11 AM, Daniel P. Berrangé wrote:
> The 'host_device' and 'host_cdrom' drivers must be used instead.
>
> Signed-off-by: Daniel P. Berrangé
> ---
> block/file-posix.c | 17 ++---
> docs/system/deprecated.rst | 7 ---
>
Ping for anyone, especially block maintainers, willing to review this
before soft freeze.
On Wed, Feb 24, 2021 at 01:11:42PM +, Daniel P. Berrangé wrote:
> The 'host_device' and 'host_cdrom' drivers must be used instead.
>
> Signed-off-by: Daniel P. Berrangé
> ---
> block/file-posix.c
On 24/02/2021 14.11, Daniel P. Berrangé wrote:
Signed-off-by: Daniel P. Berrangé
---
docs/system/deprecated.rst | 6 -
docs/system/removed-features.rst | 6 +
hw/core/machine-qmp-cmds.c | 41
qapi/machine.json| 22
Ping for anyone willing to review this so I can get this in before freeze.
On Wed, Feb 24, 2021 at 01:11:35PM +, Daniel P. Berrangé wrote:
> Signed-off-by: Daniel P. Berrangé
> ---
> docs/system/deprecated.rst | 6 -
> docs/system/removed-features.rst | 6 +
>
Am 13.03.2021 um 09:41 hat Markus Armbruster geschrieben:
> Observation, not objection:
>
> 1. QMP core parses JSON text into QObject, passes to generated
>marshaller.
>
> 2. Marshaller converts QObject to ObjectOptions with the QObject input
>visitor, passes to qmp_object_add().
>
> 3.
* Kevin Wolf (kw...@redhat.com) wrote:
> Am 15.03.2021 um 10:39 hat Markus Armbruster geschrieben:
> > Paolo Bonzini writes:
> >
> > > On 13/03/21 14:28, Markus Armbruster wrote:
> > >> Kevin Wolf writes:
> > >>
> > >>> This switches the HMP command object_add from a QemuOpts-based parser to
>
Am 13.03.2021 um 13:30 hat Markus Armbruster geschrieben:
> Paolo Bonzini writes:
>
> > On 13/03/21 08:40, Markus Armbruster wrote:
> >>> +if (!user_creatable_add_from_str(optarg, _err)) {
> >>> +if (local_err) {
> >>> +
Am 13.03.2021 um 14:40 hat Markus Armbruster geschrieben:
> Markus Armbruster writes:
>
> > Paolo Bonzini writes:
> >
> >> On 11/03/21 15:08, Markus Armbruster wrote:
> I would rather keep the OptsVisitor here. Do the same check for JSON
> syntax that you have in
On 07/03/21 23:26, Philippe Mathieu-Daudé wrote:
TYPE_PFLASH_CFI01 is a TYPE_SYS_BUS_DEVICE which registers its romd
MemoryRegion with sysbus_init_mmio(), so we can use the generic
sysbus_mmio_get_region() to get the region, no need for a specific
pflash_cfi01_get_memory() helper.
First replace
On Mar 15 12:19, Philippe Mathieu-Daudé wrote:
> On 3/15/21 12:03 PM, Klaus Jensen wrote:
> > From: Klaus Jensen
> >
> > Coverity complains about a possible memory corruption in the
> > nvme_ns_attach and _detach functions. While we should not (famous last
> > words) be able to reach this
On 3/15/21 12:03 PM, Klaus Jensen wrote:
> From: Klaus Jensen
>
> Coverity complains about a possible memory corruption in the
> nvme_ns_attach and _detach functions. While we should not (famous last
> words) be able to reach this function without nsid having previously
> been validated, this is
Am 15.03.2021 um 10:39 hat Markus Armbruster geschrieben:
> Paolo Bonzini writes:
>
> > On 13/03/21 14:28, Markus Armbruster wrote:
> >> Kevin Wolf writes:
> >>
> >>> This switches the HMP command object_add from a QemuOpts-based parser to
> >>> user_creatable_add_from_str() which uses a
From: Klaus Jensen
page_size is a uint32_t, and zasl is a uint8_t, so the expression
`page_size << zasl` is done using 32-bit arithmetic and might overflow.
Since we then compare this against a 64 bit data_size value, Coverity
complains that we might overflow unintentionally. An MDTS/ZASL value
From: Klaus Jensen
Fix three issues reported by coverity (CID 1450756, CID 1450757 and CID
1450758).
Klaus Jensen (2):
hw/block/nvme: fix potential overflow
hw/block/nvme: assert namespaces array indices
hw/block/nvme-subsys.h | 2 ++
hw/block/nvme.h| 10 --
From: Klaus Jensen
Coverity complains about a possible memory corruption in the
nvme_ns_attach and _detach functions. While we should not (famous last
words) be able to reach this function without nsid having previously
been validated, this is still an open door for future misuse.
Make Coverity
On 12.03.21 19:43, Vladimir Sementsov-Ogievskiy wrote:
12.03.2021 21:15, Max Reitz wrote:
On 05.03.21 18:35, Vladimir Sementsov-Ogievskiy wrote:
Compressed writes are unaligned to 512, which works very slow in
O_DIRECT mode. Let's use the cache.
Signed-off-by: Vladimir Sementsov-Ogievskiy
From: Akihiko Odaki
Report the configured granularity for discard operation to the
guest. If this is not set use the block size.
Since until now we have ignored the configured discard granularity
and always reported the block size, let's add
'report-discard-granularity' property and disable it
The following changes since commit 6157b0e19721aadb4c7fdcfe57b2924af6144b14:
Merge remote-tracking branch 'remotes/vivier2/tags/linux-user-for-6.0-pull-=
request' into staging (2021-03-14 17:47:49 +)
are available in the Git repository at:
https://gitlab.com/stefanha/qemu.git
Paolo Bonzini writes:
> On 13/03/21 14:28, Markus Armbruster wrote:
>> Kevin Wolf writes:
>>
>>> This switches the HMP command object_add from a QemuOpts-based parser to
>>> user_creatable_add_from_str() which uses a keyval parser and enforces
>>> the QAPI schema.
>>>
>>> Apart from being a
Use nbd_client_connecting_wait uniformly all over the block/nbd.c.
While at this, drop the redundant check for nbd_client_connecting_wait
in reconnect_delay_timer_init, as all its callsites do this check too.
Signed-off-by: Roman Kagan
---
block/nbd.c | 34 +++---
1
Document (via a comment and an assert) that
nbd_client_detach_aio_context and nbd_client_attach_aio_context_bh run
in the desired aio_context.
Signed-off-by: Roman Kagan
---
block/nbd.c | 12
1 file changed, 12 insertions(+)
diff --git a/block/nbd.c b/block/nbd.c
index
The reconnection logic doesn't need to stop while in a drained section.
Moreover it has to be active during the drained section, as the requests
that were caught in-flight with the connection to the server broken can
only usefully get drained if the connection is restored. Otherwise such
requests
As the reconnect logic no longer interferes with drained sections, it
appears unnecessary to explicitly manipulate the in_flight counter.
Fixes: 5ad81b4946 ("nbd: Restrict connection_co reentrance")
Signed-off-by: Roman Kagan
---
block/nbd.c | 6 --
nbd/client.c | 2 --
2 files changed, 8
The reconnection logic doesn't need to stop while in a drained section.
Moreover it has to be active during the drained section, as the requests
that were caught in-flight with the connection to the server broken can
only usefully get drained if the connection is restored. Otherwise such
requests
Cosmetic: adjust the comment and the return value in
nbd_co_establish_connection where it's entered while the connection
thread is still running.
Signed-off-by: Roman Kagan
---
block/nbd.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/block/nbd.c b/block/nbd.c
index
When the NBD connection is being torn down, the connection thread gets
canceled and "detached", meaning it is about to get freed.
If this happens while the connection coroutine yielded waiting for the
connection thread to complete, when it resumes it may access the
invalidated connection thread
Make varios pieces of reconnection logic correctly survive the
transition of the BDRVNBDState from one aio_context to another. In
particular,
- cancel the reconnect_delay_timer and rearm it in the new context;
- cancel the sleep of the connection_co between reconnect attempt so
that it
85 matches
Mail list logo
