From: Pavel Begunkov asml.silence@gmail.com
mainline inclusion from mainline-v6.10-rc1 commit 8d09a88ef9d3cb7d21d45c39b7b7c31298d23998 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/IAYRF9 CVE: CVE-2024-50060
Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i...
--------------------------------
Conditional locking is never great, in case of __io_cqring_overflow_flush(), which is a slow path, it's not justified. Don't handle IOPOLL separately, always grab uring_lock for overflow flushing.
Signed-off-by: Pavel Begunkov asml.silence@gmail.com Link: https://lore.kernel.org/r/162947df299aa12693ac4b305dacedab32ec7976.171270826... Signed-off-by: Jens Axboe axboe@kernel.dk
Conflicts: io_uring/io_uring.c [Context conflicts because there are no commits: 408024b95927 ("io_uring: open code io_cqring_overflow_flush()") a85381d8326d ("io_uring: skip overflow CQE posting for dying ring") 52ea806ad983 ("io_uring: finish waiting before flushing overflow entries") etc.] Signed-off-by: Baokun Li libaokun1@huawei.com --- io_uring/io_uring.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 9b92de913772..94d429483567 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -1663,6 +1663,8 @@ static bool __io_cqring_overflow_flush(struct io_ring_ctx *ctx, bool force) { bool all_flushed, posted;
+ lockdep_assert_held(&ctx->uring_lock); + if (!force && __io_cqring_events(ctx) == ctx->cq_entries) return false;
@@ -1706,12 +1708,9 @@ static bool io_cqring_overflow_flush(struct io_ring_ctx *ctx) bool ret = true;
if (test_bit(0, &ctx->check_cq_overflow)) { - /* iopoll syncs against uring_lock, not completion_lock */ - if (ctx->flags & IORING_SETUP_IOPOLL) - mutex_lock(&ctx->uring_lock); + mutex_lock(&ctx->uring_lock); ret = __io_cqring_overflow_flush(ctx, false); - if (ctx->flags & IORING_SETUP_IOPOLL) - mutex_unlock(&ctx->uring_lock); + mutex_unlock(&ctx->uring_lock); }
return ret;