qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Emanuele Giuseppe Esposito <eesposit@redhat.com>
To: qemu-block@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>, Hanna Reitz <hreitz@redhat.com>,
	John Snow <jsnow@redhat.com>, Paolo Bonzini <pbonzini@redhat.com>,
	Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>,
	Stefan Hajnoczi <stefanha@redhat.com>, Fam Zheng <fam@euphon.net>,
	Eric Blake <eblake@redhat.com>, Cleber Rosa <crosa@redhat.com>,
	qemu-devel@nongnu.org,
	Emanuele Giuseppe Esposito <eesposit@redhat.com>
Subject: [PATCH 14/20] block-gen: assert that bdrv_co_pread is always called with graph rdlock taken
Date: Wed, 16 Nov 2022 08:48:44 -0500	[thread overview]
Message-ID: <20221116134850.3051419-15-eesposit@redhat.com> (raw)
In-Reply-To: <20221116134850.3051419-1-eesposit@redhat.com>

This function, in addition to be called by a generated_co_wrapper,
is also called elsewhere else.
The strategy is to always take the lock at the function called
when the coroutine is created, to avoid recursive locking.

By protecting brdv_co_pread, we also automatically protect
the following other generated_co_wrappers:
blk_co_pread
blk_co_preadv
blk_co_preadv_part

Protecting bdrv_driver_preadv() implies that the following BlockDriver
callbacks always called with graph rdlock taken:
- bdrv_co_preadv_part
- bdrv_co_preadv
- bdrv_aio_preadv
- bdrv_co_readv

Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
---
 block/block-backend.c            | 1 +
 block/io.c                       | 1 +
 block/mirror.c                   | 6 ++++--
 include/block/block_int-common.h | 5 +++++
 include/block/block_int-io.h     | 1 +
 tests/unit/test-bdrv-drain.c     | 2 ++
 6 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/block/block-backend.c b/block/block-backend.c
index d48ec3a2ac..083ed6009e 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -1289,6 +1289,7 @@ blk_co_do_preadv_part(BlockBackend *blk, int64_t offset, int64_t bytes,
     IO_CODE();
 
     blk_wait_while_drained(blk);
+    GRAPH_RDLOCK_GUARD();
 
     /* Call blk_bs() only after waiting, the graph may have changed */
     bs = blk_bs(blk);
diff --git a/block/io.c b/block/io.c
index 92c74648fb..cfc201ef91 100644
--- a/block/io.c
+++ b/block/io.c
@@ -942,6 +942,7 @@ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
     unsigned int nb_sectors;
     QEMUIOVector local_qiov;
     int ret;
+    assert_bdrv_graph_readable();
 
     bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort);
     assert(!(flags & ~bs->supported_read_flags));
diff --git a/block/mirror.c b/block/mirror.c
index 251adc5ae0..f509cc1cb1 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -389,8 +389,10 @@ static void coroutine_fn mirror_co_read(void *opaque)
     op->is_in_flight = true;
     trace_mirror_one_iteration(s, op->offset, op->bytes);
 
-    ret = bdrv_co_preadv(s->mirror_top_bs->backing, op->offset, op->bytes,
-                         &op->qiov, 0);
+    WITH_GRAPH_RDLOCK_GUARD() {
+        ret = bdrv_co_preadv(s->mirror_top_bs->backing, op->offset, op->bytes,
+                             &op->qiov, 0);
+    }
     mirror_read_complete(op, ret);
 }
 
diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h
index e8d2e4b6c7..64c5bb64de 100644
--- a/include/block/block_int-common.h
+++ b/include/block/block_int-common.h
@@ -476,6 +476,7 @@ struct BlockDriver {
                                       Error **errp);
 
     /* aio */
+    /* Called with graph rdlock held. */
     BlockAIOCB *(*bdrv_aio_preadv)(BlockDriverState *bs,
         int64_t offset, int64_t bytes, QEMUIOVector *qiov,
         BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque);
@@ -489,6 +490,7 @@ struct BlockDriver {
         int64_t offset, int bytes,
         BlockCompletionFunc *cb, void *opaque);
 
+    /* Called with graph rdlock held. */
     int coroutine_fn (*bdrv_co_readv)(BlockDriverState *bs,
         int64_t sector_num, int nb_sectors, QEMUIOVector *qiov);
 
@@ -506,11 +508,14 @@ struct BlockDriver {
      * no larger than 'max_transfer'.
      *
      * The buffer in @qiov may point directly to guest memory.
+     *
+     * Called with graph rdlock held.
      */
     int coroutine_fn (*bdrv_co_preadv)(BlockDriverState *bs,
         int64_t offset, int64_t bytes, QEMUIOVector *qiov,
         BdrvRequestFlags flags);
 
+    /* Called with graph rdlock held. */
     int coroutine_fn (*bdrv_co_preadv_part)(BlockDriverState *bs,
         int64_t offset, int64_t bytes,
         QEMUIOVector *qiov, size_t qiov_offset,
diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h
index ae88507d6a..ac6ad3b3ff 100644
--- a/include/block/block_int-io.h
+++ b/include/block/block_int-io.h
@@ -60,6 +60,7 @@ static inline int coroutine_fn bdrv_co_pread(BdrvChild *child,
 {
     QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes);
     IO_CODE();
+    assert_bdrv_graph_readable();
 
     return bdrv_co_preadv(child, offset, bytes, &qiov, flags);
 }
diff --git a/tests/unit/test-bdrv-drain.c b/tests/unit/test-bdrv-drain.c
index 2686a8acee..90edc2f5bf 100644
--- a/tests/unit/test-bdrv-drain.c
+++ b/tests/unit/test-bdrv-drain.c
@@ -967,6 +967,8 @@ static void coroutine_fn test_co_delete_by_drain(void *opaque)
     void *buffer = g_malloc(65536);
     QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buffer, 65536);
 
+    GRAPH_RDLOCK_GUARD();
+
     /* Pretend some internal write operation from parent to child.
      * Important: We have to read from the child, not from the parent!
      * Draining works by first propagating it all up the tree to the
-- 
2.31.1



  parent reply	other threads:[~2022-11-16 13:57 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-16 13:48 [PATCH 00/20] Protect the block layer with a rwlock: part 1 Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 01/20] block: introduce a lock to protect graph operations Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 02/20] graph-lock: introduce BdrvGraphRWlock structure Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 03/20] async: register/unregister aiocontext in graph lock list Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 04/20] block.c: wrlock in bdrv_replace_child_noperm Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 05/20] block: remove unnecessary assert_bdrv_graph_writable() Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 06/20] block: assert that graph read and writes are performed correctly Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 07/20] graph-lock: implement WITH_GRAPH_RDLOCK_GUARD and GRAPH_RDLOCK_GUARD macros Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 08/20] block-coroutine-wrapper.py: take the graph rdlock in bdrv_* functions Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 09/20] block-backend: introduce new generated_co_wrapper_blk annotation Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 10/20] block-gen: assert that {bdrv/blk}_co_truncate is always called with graph rdlock taken Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 11/20] block-gen: assert that bdrv_co_{check/invalidate_cache} are " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 12/20] block-gen: assert that bdrv_co_pwrite is " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 13/20] block-gen: assert that bdrv_co_pwrite_{zeros/sync} " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` Emanuele Giuseppe Esposito [this message]
2022-11-16 13:48 ` [PATCH 15/20] block-gen: assert that {bdrv/blk}_co_flush " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 16/20] block-gen: assert that bdrv_co_{read/write}v_vmstate are " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 17/20] block-gen: assert that bdrv_co_pdiscard is " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 18/20] block-gen: assert that bdrv_co_common_block_status_above " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 19/20] block-gen: assert that bdrv_co_ioctl " Emanuele Giuseppe Esposito
2022-11-16 13:48 ` [PATCH 20/20] block-gen: assert that nbd_co_do_establish_connection " Emanuele Giuseppe Esposito
2022-11-21 15:02 ` [PATCH 00/20] Protect the block layer with a rwlock: part 1 Emanuele Giuseppe Esposito

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=20221116134850.3051419-15-eesposit@redhat.com \
    --to=eesposit@redhat.com \
    --cc=crosa@redhat.com \
    --cc=eblake@redhat.com \
    --cc=fam@euphon.net \
    --cc=hreitz@redhat.com \
    --cc=jsnow@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    --cc=vsementsov@yandex-team.ru \
    /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).