[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [Qemu-block] [PATCH v5 05/12] block: Introduce "drained
From: |
Jeff Cody |
Subject: |
Re: [Qemu-devel] [Qemu-block] [PATCH v5 05/12] block: Introduce "drained begin/end" API |
Date: |
Wed, 21 Oct 2015 12:11:18 -0400 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Wed, Oct 21, 2015 at 10:06:42AM +0800, Fam Zheng wrote:
> The semantics is that after bdrv_drained_begin(bs), bs will not get new
> external
> requests until the matching bdrv_drained_end(bs).
>
> Signed-off-by: Fam Zheng <address@hidden>
> ---
> block/io.c | 17 +++++++++++++++++
> include/block/block.h | 21 +++++++++++++++++++++
> include/block/block_int.h | 2 ++
> 3 files changed, 40 insertions(+)
>
> diff --git a/block/io.c b/block/io.c
> index 2fd7a1d..5ac6256 100644
> --- a/block/io.c
> +++ b/block/io.c
> @@ -2624,3 +2624,20 @@ void bdrv_flush_io_queue(BlockDriverState *bs)
> }
> bdrv_start_throttled_reqs(bs);
> }
> +
> +void bdrv_drained_begin(BlockDriverState *bs)
> +{
> + if (!bs->quiesce_counter++) {
> + aio_disable_external(bdrv_get_aio_context(bs));
> + }
> + bdrv_drain(bs);
> +}
> +
> +void bdrv_drained_end(BlockDriverState *bs)
> +{
> + assert(bs->quiesce_counter > 0);
> + if (--bs->quiesce_counter > 0) {
> + return;
> + }
> + aio_enable_external(bdrv_get_aio_context(bs));
> +}
Why do we need a quiesce counter, given that
aio_{disable, enable}_external() increments / decrements a counter?
> diff --git a/include/block/block.h b/include/block/block.h
> index 28d903c..6d38b62 100644
> --- a/include/block/block.h
> +++ b/include/block/block.h
> @@ -610,4 +610,25 @@ void bdrv_io_plug(BlockDriverState *bs);
> void bdrv_io_unplug(BlockDriverState *bs);
> void bdrv_flush_io_queue(BlockDriverState *bs);
>
> +BlockAcctStats *bdrv_get_stats(BlockDriverState *bs);
> +
Is the above line from a bad rebase?
> +/**
> + * bdrv_drained_begin:
> + *
> + * Begin a quiesced section for exclusive access to the BDS, by disabling
> + * external request sources including NBD server and device model. Note that
> + * this doesn't block timers or coroutines from submitting more requests,
> which
> + * means block_job_pause is still necessary.
> + *
> + * This function can be recursive.
> + */
> +void bdrv_drained_begin(BlockDriverState *bs);
> +
> +/**
> + * bdrv_drained_end:
> + *
> + * End a quiescent section started by bdrv_drained_begin().
> + */
> +void bdrv_drained_end(BlockDriverState *bs);
> +
> #endif
> diff --git a/include/block/block_int.h b/include/block/block_int.h
> index e472a03..e317b14 100644
> --- a/include/block/block_int.h
> +++ b/include/block/block_int.h
> @@ -448,6 +448,8 @@ struct BlockDriverState {
> /* threshold limit for writes, in bytes. "High water mark". */
> uint64_t write_threshold_offset;
> NotifierWithReturn write_threshold_notifier;
> +
> + int quiesce_counter;
> };
>
> struct BlockBackendRootState {
> --
> 2.4.3
>
>
- [Qemu-devel] [PATCH v5 00/12] block: Protect nested event loop with bdrv_drained_begin and bdrv_drained_end, Fam Zheng, 2015/10/20
- [Qemu-devel] [PATCH v5 05/12] block: Introduce "drained begin/end" API, Fam Zheng, 2015/10/20
- Re: [Qemu-devel] [Qemu-block] [PATCH v5 05/12] block: Introduce "drained begin/end" API,
Jeff Cody <=
- [Qemu-devel] [PATCH v5 06/12] block: Add "drained begin/end" for transactional external snapshot, Fam Zheng, 2015/10/20
- [Qemu-devel] [PATCH v5 07/12] block: Add "drained begin/end" for transactional backup, Fam Zheng, 2015/10/20
- [Qemu-devel] [PATCH v5 08/12] block: Add "drained begin/end" for transactional blockdev-backup, Fam Zheng, 2015/10/20
- [Qemu-devel] [PATCH v5 09/12] block: Add "drained begin/end" for internal snapshot, Fam Zheng, 2015/10/20