On Thu 23-10-25 12:46:45, Jiri Slaby wrote:

On 10. 09. 25, 16:36, Christian Brauner wrote:

quoted

Validate extensible ioctls stricter than we do now.

Signed-off-by: Christian Brauner <brauner@kernel.org>
---
  fs/pidfs.c         |  2 +-
  include/linux/fs.h | 14 ++++++++++++++
  2 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/fs/pidfs.c b/fs/pidfs.c
index edc35522d75c..0a5083b9cce5 100644
--- a/fs/pidfs.c
+++ b/fs/pidfs.c

@@ -440,7 +440,7 @@ static bool pidfs_ioctl_valid(unsigned int cmd)
  		 * erronously mistook the file descriptor for a pidfd.
  		 * This is not perfect but will catch most cases.
  		 */
-		return (_IOC_TYPE(cmd) == _IOC_TYPE(PIDFD_GET_INFO));
+		return extensible_ioctl_valid(cmd, PIDFD_GET_INFO, PIDFD_INFO_SIZE_VER0);

Hi,

this turned EINVAL (from pidfd_info()) into ENOTTY (from pidfd_ioctl()) for
at least LTP's:
struct pidfd_info_invalid {
	uint32_t dummy;
};

#define PIDFD_GET_INFO_SHORT _IOWR(PIDFS_IOCTL_MAGIC, 11, struct
pidfd_info_invalid)

ioctl(pidfd, PIDFD_GET_INFO_SHORT, info_invalid) == EINVAL

at:
https://github.com/linux-test-project/ltp/blob/9bb94efa39bb1b08f37e56c7437db5fa13ddcae2/testcases/kernel/syscalls/ioctl/ioctl_pidfd05.c#L46

Is this expected?

We already discussed this internally but for others the problem was
discussed here [1] and we decided the new errno value is OK and LTP test is
being fixed up.

								Honza

[1] https://lore.kernel.org/all/aPIPGeWo8gtxVxQX@yuki.lan/ (local)

-- 
Jan Kara [off-list ref]
SUSE Labs, CR