* Re: [Qemu-devel] [PATCH for-2.9] blockjob: avoid recursive AioContext locking
2017-03-21 17:48 [Qemu-devel] [PATCH for-2.9] blockjob: avoid recursive AioContext locking Paolo Bonzini
@ 2017-03-21 19:15 ` Eric Blake
2017-03-22 12:05 ` [Qemu-devel] [Qemu-block] " Jeff Cody
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Eric Blake @ 2017-03-21 19:15 UTC (permalink / raw)
To: Paolo Bonzini, qemu-devel; +Cc: jsnow, famz, stefanha, qemu-block
[-- Attachment #1: Type: text/plain, Size: 1128 bytes --]
On 03/21/2017 12:48 PM, Paolo Bonzini wrote:
> Streaming or any other block job hangs when performed on a block device
> that has a non-default iothread. This happens because the AioContext
> is acquired twice by block_job_defer_to_main_loop_bh and then released
> only once by BDRV_POLL_WHILE. (Insert rants on recursive mutexes, which
>
> unfortunately are a temporary but necessary evil for iothreads at the
Why the blank line?
> moment).
>
> Luckily, the reason for the double acquisition is simple; the function
> acquires the AioContext for both the job iothread and the BDS iothread,
> in case the BDS iothread was changed while the job was running. It
> is therefore enough to skip the second acquisition when the two
> AioContexts are one and the same.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> blockjob.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
Makes sense from the description.
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Qemu-devel] [Qemu-block] [PATCH for-2.9] blockjob: avoid recursive AioContext locking
2017-03-21 17:48 [Qemu-devel] [PATCH for-2.9] blockjob: avoid recursive AioContext locking Paolo Bonzini
2017-03-21 19:15 ` Eric Blake
@ 2017-03-22 12:05 ` Jeff Cody
2017-03-22 12:15 ` Jeff Cody
2017-03-22 15:32 ` [Qemu-devel] " John Snow
3 siblings, 0 replies; 5+ messages in thread
From: Jeff Cody @ 2017-03-22 12:05 UTC (permalink / raw)
To: Paolo Bonzini; +Cc: qemu-devel, famz, stefanha, qemu-block
On Tue, Mar 21, 2017 at 06:48:10PM +0100, Paolo Bonzini wrote:
> Streaming or any other block job hangs when performed on a block device
> that has a non-default iothread. This happens because the AioContext
> is acquired twice by block_job_defer_to_main_loop_bh and then released
> only once by BDRV_POLL_WHILE. (Insert rants on recursive mutexes, which
>
> unfortunately are a temporary but necessary evil for iothreads at the
> moment).
>
> Luckily, the reason for the double acquisition is simple; the function
> acquires the AioContext for both the job iothread and the BDS iothread,
> in case the BDS iothread was changed while the job was running. It
> is therefore enough to skip the second acquisition when the two
> AioContexts are one and the same.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> blockjob.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/blockjob.c b/blockjob.c
> index 69126af..2159df7 100644
> --- a/blockjob.c
> +++ b/blockjob.c
> @@ -755,12 +755,16 @@ static void block_job_defer_to_main_loop_bh(void *opaque)
>
> /* Fetch BDS AioContext again, in case it has changed */
> aio_context = blk_get_aio_context(data->job->blk);
> - aio_context_acquire(aio_context);
> + if (aio_context != data->aio_context) {
> + aio_context_acquire(aio_context);
> + }
>
> data->job->deferred_to_main_loop = false;
> data->fn(data->job, data->opaque);
>
> - aio_context_release(aio_context);
> + if (aio_context != data->aio_context) {
> + aio_context_release(aio_context);
> + }
>
> aio_context_release(data->aio_context);
>
> --
> 1.8.3.1
>
>
Reviewed-by: Jeff Cody <jcody@redhat.com>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Qemu-devel] [Qemu-block] [PATCH for-2.9] blockjob: avoid recursive AioContext locking
2017-03-21 17:48 [Qemu-devel] [PATCH for-2.9] blockjob: avoid recursive AioContext locking Paolo Bonzini
2017-03-21 19:15 ` Eric Blake
2017-03-22 12:05 ` [Qemu-devel] [Qemu-block] " Jeff Cody
@ 2017-03-22 12:15 ` Jeff Cody
2017-03-22 15:32 ` [Qemu-devel] " John Snow
3 siblings, 0 replies; 5+ messages in thread
From: Jeff Cody @ 2017-03-22 12:15 UTC (permalink / raw)
To: Paolo Bonzini; +Cc: qemu-devel, famz, stefanha, qemu-block
On Tue, Mar 21, 2017 at 06:48:10PM +0100, Paolo Bonzini wrote:
> Streaming or any other block job hangs when performed on a block device
> that has a non-default iothread. This happens because the AioContext
> is acquired twice by block_job_defer_to_main_loop_bh and then released
> only once by BDRV_POLL_WHILE. (Insert rants on recursive mutexes, which
>
> unfortunately are a temporary but necessary evil for iothreads at the
> moment).
>
> Luckily, the reason for the double acquisition is simple; the function
> acquires the AioContext for both the job iothread and the BDS iothread,
> in case the BDS iothread was changed while the job was running. It
> is therefore enough to skip the second acquisition when the two
> AioContexts are one and the same.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> blockjob.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/blockjob.c b/blockjob.c
> index 69126af..2159df7 100644
> --- a/blockjob.c
> +++ b/blockjob.c
> @@ -755,12 +755,16 @@ static void block_job_defer_to_main_loop_bh(void *opaque)
>
> /* Fetch BDS AioContext again, in case it has changed */
> aio_context = blk_get_aio_context(data->job->blk);
> - aio_context_acquire(aio_context);
> + if (aio_context != data->aio_context) {
> + aio_context_acquire(aio_context);
> + }
>
> data->job->deferred_to_main_loop = false;
> data->fn(data->job, data->opaque);
>
> - aio_context_release(aio_context);
> + if (aio_context != data->aio_context) {
> + aio_context_release(aio_context);
> + }
>
> aio_context_release(data->aio_context);
>
> --
> 1.8.3.1
>
>
Deleted the blank line in the commit message, and:
Thanks,
Applied to my block branch:
git://github.com/codyprime/qemu-kvm-jtc.git block
-Jeff
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Qemu-devel] [PATCH for-2.9] blockjob: avoid recursive AioContext locking
2017-03-21 17:48 [Qemu-devel] [PATCH for-2.9] blockjob: avoid recursive AioContext locking Paolo Bonzini
` (2 preceding siblings ...)
2017-03-22 12:15 ` Jeff Cody
@ 2017-03-22 15:32 ` John Snow
3 siblings, 0 replies; 5+ messages in thread
From: John Snow @ 2017-03-22 15:32 UTC (permalink / raw)
To: Paolo Bonzini, qemu-devel; +Cc: famz, stefanha, qemu-block
On 03/21/2017 01:48 PM, Paolo Bonzini wrote:
> Streaming or any other block job hangs when performed on a block device
> that has a non-default iothread. This happens because the AioContext
> is acquired twice by block_job_defer_to_main_loop_bh and then released
> only once by BDRV_POLL_WHILE. (Insert rants on recursive mutexes, which
>
> unfortunately are a temporary but necessary evil for iothreads at the
> moment).
>
> Luckily, the reason for the double acquisition is simple; the function
> acquires the AioContext for both the job iothread and the BDS iothread,
> in case the BDS iothread was changed while the job was running. It
> is therefore enough to skip the second acquisition when the two
> AioContexts are one and the same.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> blockjob.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/blockjob.c b/blockjob.c
> index 69126af..2159df7 100644
> --- a/blockjob.c
> +++ b/blockjob.c
> @@ -755,12 +755,16 @@ static void block_job_defer_to_main_loop_bh(void *opaque)
>
> /* Fetch BDS AioContext again, in case it has changed */
> aio_context = blk_get_aio_context(data->job->blk);
> - aio_context_acquire(aio_context);
> + if (aio_context != data->aio_context) {
> + aio_context_acquire(aio_context);
> + }
>
> data->job->deferred_to_main_loop = false;
> data->fn(data->job, data->opaque);
>
> - aio_context_release(aio_context);
> + if (aio_context != data->aio_context) {
> + aio_context_release(aio_context);
> + }
>
> aio_context_release(data->aio_context);
>
>
Reviewed-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 5+ messages in thread