[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [RFC PATCH 15/17] block/nvme: Use per-queue AIO context
From: |
Stefan Hajnoczi |
Subject: |
Re: [RFC PATCH 15/17] block/nvme: Use per-queue AIO context |
Date: |
Fri, 26 Jun 2020 13:42:33 +0100 |
On Thu, Jun 25, 2020 at 08:48:36PM +0200, Philippe Mathieu-Daudé wrote:
> To be able to use multiple queues on the same hardware,
> we need to have each queue able to receive IRQ notifications
> in the correct AIO context.
> The context has to be proper to each queue, not to the block
> driver. Move aio_context from BDRVNVMeState to NVMeQueuePair.
>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
> RFC because I'm not familiar with AIO context
This patch looks incomplete because there is still only 1
s->irq_notifier but the code already performs aio_set_event_notifer()
calls for each queue in nvme_close().
Either there should be 1 irq_notifier and singleton
aio_set_event_notifier() calls or irq_notifiers should really be
per-queue.
signature.asc
Description: PGP signature
- Re: [PATCH 11/17] block/nvme: Simplify nvme_create_queue_pair() arguments, (continued)
- [PATCH 12/17] block/nvme: Simplify nvme_kick trace event, Philippe Mathieu-Daudé, 2020/06/25
- [PATCH 13/17] block/nvme: Simplify completion trace events, Philippe Mathieu-Daudé, 2020/06/25
- [PATCH 14/17] block/nvme: Replace BDRV_POLL_WHILE by AIO_WAIT_WHILE, Philippe Mathieu-Daudé, 2020/06/25
- [PATCH 16/17] block/nvme: Check BDRVNVMeState::plugged out of nvme_kick(), Philippe Mathieu-Daudé, 2020/06/25
- [RFC PATCH 15/17] block/nvme: Use per-queue AIO context, Philippe Mathieu-Daudé, 2020/06/25
- [PATCH 17/17] block/nvme: Check BDRVNVMeState::plugged out of nvme_process_completion, Philippe Mathieu-Daudé, 2020/06/25
- Re: [PATCH 00/17] block/nvme: Various cleanups required to use multiple queues, no-reply, 2020/06/25