[PATCH v3] io_uring: correct check for O_TMPFILE

[PATCH v3] io_uring: correct check for O_TMPFILE

[Aleksa Sarai]

O_TMPFILE is actually __O_TMPFILE|O_DIRECTORY. This means that the old
check for whether RESOLVE_CACHED can be used would incorrectly think
that O_DIRECTORY could not be used with RESOLVE_CACHED.

Cc: stable@vger.kernel.org # v5.12+
Fixes: 3a81fd02045c ("io_uring: enable LOOKUP_CACHED path resolution for filename lookups")
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
---
Changes in v3:
- drop openat2 patch, as it's already in Christian's tree
- explain __O_TMPFILE usage in io_openat_force_async comment
- v2: https://lore.kernel.org/r/20230806-resolve_cached-o_tmpfile-v2-0-058bff24fb16@cyphar.com

Changes in v2:
- fix io_uring's io_openat_force_async as well.
- v1: <https://lore.kernel.org/r/20230806-resolve_cached-o_tmpfile-v1-1-7ba16308465e@cyphar.com>
---
 io_uring/openclose.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/io_uring/openclose.c b/io_uring/openclose.c
index 10ca57f5bd24..e3fae26e025d 100644
--- a/io_uring/openclose.c
+++ b/io_uring/openclose.c
@@ -35,9 +35,11 @@ static bool io_openat_force_async(struct io_open *open)
 {
 	/*
 	 * Don't bother trying for O_TRUNC, O_CREAT, or O_TMPFILE open,
-	 * it'll always -EAGAIN
+	 * it'll always -EAGAIN. Note that we test for __O_TMPFILE because
+	 * O_TMPFILE includes O_DIRECTORY, which isn't a flag we need to force
+	 * async for.
 	 */
-	return open->how.flags & (O_TRUNC | O_CREAT | O_TMPFILE);
+	return open->how.flags & (O_TRUNC | O_CREAT | __O_TMPFILE);
 }
 
 static int __io_openat_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)

---
base-commit: 272af00d6825f19b48b9d9cfd11b1f6bdc011e2c
change-id: 20230806-resolve_cached-o_tmpfile-978cb238bd68

Best regards,
-- 
Aleksa Sarai <cyphar@cyphar.com>

Re: [PATCH v3] io_uring: correct check for O_TMPFILE

[Jens Axboe]

On Mon, 07 Aug 2023 12:24:15 +1000, Aleksa Sarai wrote:
> O_TMPFILE is actually __O_TMPFILE|O_DIRECTORY. This means that the old
> check for whether RESOLVE_CACHED can be used would incorrectly think
> that O_DIRECTORY could not be used with RESOLVE_CACHED.
> 
> 

Applied, thanks!

[1/1] io_uring: correct check for O_TMPFILE
      (no commit info)

Best regards,
-- 
Jens Axboe