From: Hanna Czenczek <hreitz@redhat.com>
Date: Mon, 10 Nov 2025 15:48:41 +0000 (+0100)
Subject: gluster: Do not move coroutine into BDS context
X-Git-Tag: v10.2.0-rc1~5^2~13
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=7214ad20da625c84e55536a274acdd5af9627ff7;p=thirdparty%2Fqemu.git

gluster: Do not move coroutine into BDS context

The request coroutine may not run in the BDS AioContext.  We should wake
it in its own context, not move it.

With that, we can remove GlusterAIOCB.aio_context.

Also add a comment why aio_co_schedule() is safe to use in this way.

**Note:** Due to a lack of a gluster set-up, I have not tested this
commit.  It seemed safe enough to send anyway, just maybe not to
qemu-stable.  To be clear, I don’t know of any user-visible bugs that
would arise from the state without this patch; the request coroutine is
moved into the main BDS AioContext, so guest device completion code will
run in a different context than where the request started, which can’t
be good, but I haven’t actually confirmed any bugs (due to not being
able to test it).

Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20251110154854.151484-7-hreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---

diff --git a/block/gluster.c b/block/gluster.c
index 89abd40f31..4fb25b2c6d 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -56,7 +56,6 @@ typedef struct GlusterAIOCB {
     int64_t size;
     int ret;
     Coroutine *coroutine;
-    AioContext *aio_context;
 } GlusterAIOCB;
 
 typedef struct BDRVGlusterState {
@@ -743,7 +742,17 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret,
         acb->ret = -EIO; /* Partial read/write - fail it */
     }
 
-    aio_co_schedule(acb->aio_context, acb->coroutine);
+    /*
+     * Safe to call: The coroutine will yield exactly once awaiting this
+     * scheduling, and the context is its own context, so it will be scheduled
+     * once it does yield.
+     *
+     * (aio_co_wake() would call qemu_get_current_aio_context() to check whether
+     * we are in the same context, but we are not in a qemu thread, so we cannot
+     * do that.  Use aio_co_schedule() directly.)
+     */
+    aio_co_schedule(qemu_coroutine_get_aio_context(acb->coroutine),
+                    acb->coroutine);
 }
 
 static void qemu_gluster_parse_flags(int bdrv_flags, int *open_flags)
@@ -1006,7 +1015,6 @@ static coroutine_fn int qemu_gluster_co_pwrite_zeroes(BlockDriverState *bs,
     acb.size = bytes;
     acb.ret = 0;
     acb.coroutine = qemu_coroutine_self();
-    acb.aio_context = bdrv_get_aio_context(bs);
 
     ret = glfs_zerofill_async(s->fd, offset, bytes, gluster_finish_aiocb, &acb);
     if (ret < 0) {
@@ -1184,7 +1192,6 @@ static coroutine_fn int qemu_gluster_co_rw(BlockDriverState *bs,
     acb.size = size;
     acb.ret = 0;
     acb.coroutine = qemu_coroutine_self();
-    acb.aio_context = bdrv_get_aio_context(bs);
 
     if (write) {
         ret = glfs_pwritev_async(s->fd, qiov->iov, qiov->niov, offset, 0,
@@ -1251,7 +1258,6 @@ static coroutine_fn int qemu_gluster_co_flush_to_disk(BlockDriverState *bs)
     acb.size = 0;
     acb.ret = 0;
     acb.coroutine = qemu_coroutine_self();
-    acb.aio_context = bdrv_get_aio_context(bs);
 
     ret = glfs_fsync_async(s->fd, gluster_finish_aiocb, &acb);
     if (ret < 0) {
@@ -1299,7 +1305,6 @@ static coroutine_fn int qemu_gluster_co_pdiscard(BlockDriverState *bs,
     acb.size = 0;
     acb.ret = 0;
     acb.coroutine = qemu_coroutine_self();
-    acb.aio_context = bdrv_get_aio_context(bs);
 
     ret = glfs_discard_async(s->fd, offset, bytes, gluster_finish_aiocb, &acb);
     if (ret < 0) {