From: Hanna Czenczek <hreitz@redhat.com>
To: qemu-block@nongnu.org
Cc: qemu-devel@nongnu.org, "Hanna Czenczek" <hreitz@redhat.com>,
"Kevin Wolf" <kwolf@redhat.com>,
"Stefan Hajnoczi" <stefanha@redhat.com>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Richard W . M . Jones" <rjones@redhat.com>,
"Ilya Dryomov" <idryomov@gmail.com>,
"Peter Lieven" <pl@dlhnet.de>,
"Philippe Mathieu-Daudé" <philmd@linaro.org>,
"Alex Bennée" <alex.bennee@linaro.org>,
"Fam Zheng" <fam@euphon.net>,
"Ronnie Sahlberg" <ronniesahlberg@gmail.com>
Subject: [PATCH v2 19/19] win32-aio: Run CB in original context
Date: Mon, 10 Nov 2025 16:48:54 +0100 [thread overview]
Message-ID: <20251110154854.151484-20-hreitz@redhat.com> (raw)
In-Reply-To: <20251110154854.151484-1-hreitz@redhat.com>
AIO callbacks must be called in the originally calling AioContext,
regardless of the BDS’s “main” AioContext.
Note: I tried to test this (under wine), but failed. Whenever I tried
to use multiqueue or even just an I/O thread for a virtio-blk (or
virtio-scsi) device, I/O stalled, both with and without this patch.
For what it’s worth, when not using an I/O thread, I/O continued to work
with this patch.
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
---
block/win32-aio.c | 31 ++++++++++++++++++++++++-------
1 file changed, 24 insertions(+), 7 deletions(-)
diff --git a/block/win32-aio.c b/block/win32-aio.c
index 6327861e1d..f0689f3ee9 100644
--- a/block/win32-aio.c
+++ b/block/win32-aio.c
@@ -48,48 +48,62 @@ struct QEMUWin32AIOState {
typedef struct QEMUWin32AIOCB {
BlockAIOCB common;
struct QEMUWin32AIOState *ctx;
+ AioContext *req_ctx;
int nbytes;
OVERLAPPED ov;
QEMUIOVector *qiov;
void *buf;
bool is_read;
bool is_linear;
+ int ret;
} QEMUWin32AIOCB;
+static void win32_aio_completion_cb_bh(void *opaque)
+{
+ QEMUWin32AIOCB *waiocb = opaque;
+
+ waiocb->common.cb(waiocb->common.opaque, waiocb->ret);
+ aio_context_unref(waiocb->req_ctx);
+ qemu_aio_unref(waiocb);
+}
+
/*
* Completes an AIO request (calls the callback and frees the ACB).
*/
static void win32_aio_process_completion(QEMUWin32AIOState *s,
QEMUWin32AIOCB *waiocb, DWORD count)
{
- int ret;
s->count--;
if (waiocb->ov.Internal != 0) {
- ret = -EIO;
+ waiocb->ret = -EIO;
} else {
- ret = 0;
+ waiocb->ret = 0;
if (count < waiocb->nbytes) {
/* Short reads mean EOF, pad with zeros. */
if (waiocb->is_read) {
qemu_iovec_memset(waiocb->qiov, count, 0,
waiocb->qiov->size - count);
} else {
- ret = -EINVAL;
+ waiocb->ret = -EINVAL;
}
}
}
if (!waiocb->is_linear) {
- if (ret == 0 && waiocb->is_read) {
+ if (waiocb->ret == 0 && waiocb->is_read) {
QEMUIOVector *qiov = waiocb->qiov;
iov_from_buf(qiov->iov, qiov->niov, 0, waiocb->buf, qiov->size);
}
qemu_vfree(waiocb->buf);
}
- waiocb->common.cb(waiocb->common.opaque, ret);
- qemu_aio_unref(waiocb);
+ if (waiocb->req_ctx == s->aio_ctx) {
+ win32_aio_completion_cb_bh(waiocb);
+ } else {
+ aio_bh_schedule_oneshot(waiocb->req_ctx, win32_aio_completion_cb_bh,
+ waiocb);
+ }
}
static void win32_aio_completion_cb(EventNotifier *e)
@@ -120,10 +134,13 @@ BlockAIOCB *win32_aio_submit(BlockDriverState *bs,
DWORD rc;
waiocb = qemu_aio_get(&win32_aiocb_info, bs, cb, opaque);
+ waiocb->req_ctx = qemu_get_current_aio_context();
waiocb->nbytes = bytes;
waiocb->qiov = qiov;
waiocb->is_read = (type == QEMU_AIO_READ);
+ aio_context_ref(waiocb->req_ctx);
+
if (qiov->niov > 1) {
waiocb->buf = qemu_try_blockalign(bs, qiov->size);
if (waiocb->buf == NULL) {
--
2.51.1
next prev parent reply other threads:[~2025-11-10 15:53 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-10 15:48 [PATCH v2 00/19] block: Some multi-threading fixes Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 01/19] block: Note on aio_co_wake use if not yet yielding Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 02/19] rbd: Run co BH CB in the coroutine’s AioContext Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 03/19] iscsi: " Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 04/19] nfs: " Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 05/19] curl: Fix coroutine waking Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 06/19] gluster: Do not move coroutine into BDS context Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 07/19] nvme: Kick and check completions in " Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 08/19] nvme: Fix coroutine waking Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 09/19] nvme: Note in which AioContext some functions run Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 10/19] block/io: Take reqs_lock for tracked_requests Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 11/19] qcow2: Re-initialize lock in invalidate_cache Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 12/19] qcow2: Fix cache_clean_timer Hanna Czenczek
2025-11-17 14:50 ` Kevin Wolf
2025-11-18 11:01 ` Hanna Czenczek
2025-11-18 17:06 ` Kevin Wolf
2025-11-18 17:19 ` Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 13/19] qcow2: Schedule cache-clean-timer in realtime Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 14/19] ssh: Run restart_coroutine in current AioContext Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 15/19] blkreplay: Run BH in coroutine’s AioContext Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 16/19] block: Note in which AioContext AIO CBs are called Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 17/19] iscsi: Create AIO BH in original AioContext Hanna Czenczek
2025-11-10 15:48 ` [PATCH v2 18/19] null-aio: Run CB " Hanna Czenczek
2025-11-10 15:48 ` Hanna Czenczek [this message]
2025-11-17 15:11 ` [PATCH v2 00/19] block: Some multi-threading fixes Kevin Wolf
2025-11-17 19:11 ` Stefan Hajnoczi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20251110154854.151484-20-hreitz@redhat.com \
--to=hreitz@redhat.com \
--cc=alex.bennee@linaro.org \
--cc=fam@euphon.net \
--cc=idryomov@gmail.com \
--cc=kwolf@redhat.com \
--cc=pbonzini@redhat.com \
--cc=philmd@linaro.org \
--cc=pl@dlhnet.de \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=rjones@redhat.com \
--cc=ronniesahlberg@gmail.com \
--cc=stefanha@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).