This issue is conceptually identical to the one fixed in 29f077d07051 ("io_uring: always use original task when preparing req identity"), so rather than reinvent the wheel, I'm shamelessly quoting the commit message from that patch - thanks Jens:
"If the ring is setup with IORING_SETUP_IOPOLL and we have more than one task doing submissions on a ring, we can up in a situation where we assign the context from the current task rather than the request originator.
Always use req->task rather than assume it's the same as current.
No upstream patch exists for this issue, as only older kernels with the non-native workers have this problem."
Cc: Jens Axboe axboe@kernel.dk Cc: Pavel Begunkov asml.silence@gmail.com Cc: Alexander Viro viro@zeniv.linux.org.uk Cc: io-uring@vger.kernel.org Cc: linux-fsdevel@vger.kernel.org Fixes: 5c3462cfd123b ("io_uring: store io_identity in io_uring_task") Signed-off-by: Lee Jones lee@kernel.org --- fs/io_uring.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/fs/io_uring.c b/fs/io_uring.c index 023c3eb34dcca..a952288b2ab8e 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -1325,7 +1325,7 @@ static void io_req_clean_work(struct io_kiocb *req) */ static bool io_identity_cow(struct io_kiocb *req) { - struct io_uring_task *tctx = current->io_uring; + struct io_uring_task *tctx = req->task->io_uring; const struct cred *creds = NULL; struct io_identity *id;
On 7/19/22 5:52 AM, Lee Jones wrote:
This issue is conceptually identical to the one fixed in 29f077d07051 ("io_uring: always use original task when preparing req identity"), so rather than reinvent the wheel, I'm shamelessly quoting the commit message from that patch - thanks Jens:
"If the ring is setup with IORING_SETUP_IOPOLL and we have more than one task doing submissions on a ring, we can up in a situation where we assign the context from the current task rather than the request originator.
Always use req->task rather than assume it's the same as current.
No upstream patch exists for this issue, as only older kernels with the non-native workers have this problem."
Greg, can you pick this one up for 5.10-stable? Thanks!
On Wed, Jul 20, 2022 at 06:48:09AM -0600, Jens Axboe wrote:
On 7/19/22 5:52 AM, Lee Jones wrote:
This issue is conceptually identical to the one fixed in 29f077d07051 ("io_uring: always use original task when preparing req identity"), so rather than reinvent the wheel, I'm shamelessly quoting the commit message from that patch - thanks Jens:
"If the ring is setup with IORING_SETUP_IOPOLL and we have more than one task doing submissions on a ring, we can up in a situation where we assign the context from the current task rather than the request originator.
Always use req->task rather than assume it's the same as current.
No upstream patch exists for this issue, as only older kernels with the non-native workers have this problem."
Greg, can you pick this one up for 5.10-stable? Thanks!
Now queued up, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org
-
Greg KH -
Jens Axboe -
Lee Jones