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 <[email protected]> --- block/gluster.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) 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) { -- 2.51.1
