On Wed, Jul 09, 2025 at 08:10:14PM +0200, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@xxxxxxxx>
>
> Anders and Naresh found that the addition of the FS_IOC_GETLBMD_CAP
> handling in the blockdev ioctl handler breaks all ioctls with
> _IOC_NR==2, as the new command is not added to the switch but only
> a few of the command bits are check.
>
> Refine the check to also validate the direction/type/length bits,
> but still allow all supported sizes for future extensions.
>
> Move the new command to the end of the function to avoid slowing
> down normal ioctl commands with the added branches.
>
> Fixes: 9eb22f7fedfc ("fs: add ioctl to query metadata and protection info capabilities")
> Link: https://lore.kernel.org/all/CA+G9fYvk9HHE5UJ7cdJHTcY6P5JKnp+_e+sdC5U-ZQFTP9_hqQ@xxxxxxxxxxxxxx/
> Reported-by: Naresh Kamboju <naresh.kamboju@xxxxxxxxxx>
> Cc: Anders Roxell <anders.roxell@xxxxxxxxxx>
> Cc: Naresh Kamboju <naresh.kamboju@xxxxxxxxxx>
> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> ---
Thanks!
> It seems that we have a lot of drivers with the same bug, as the
> large majority of all _IOC_NR() users in the kernel fail to also
> check the other bits of the ioctl command code. There are currently
> 55 files referencing _IOC_NR, and they all need to be manually
> checked for this problem.
> ---
The current documentation in Documentation/dev-tools/checkuapi.rst needs
updating too then.
I want this to work. So as a start we should have a common static inline
helper that encapsulates the barrage of checks.
> block/ioctl.c | 14 +++++++++-----
> 1 file changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/block/ioctl.c b/block/ioctl.c
> index 9ad403733e19..5e5a422bd09f 100644
> --- a/block/ioctl.c
> +++ b/block/ioctl.c
> @@ -567,9 +567,6 @@ static int blkdev_common_ioctl(struct block_device *bdev, blk_mode_t mode,
> {
> unsigned int max_sectors;
>
> - if (_IOC_NR(cmd) == _IOC_NR(FS_IOC_GETLBMD_CAP))
> - return blk_get_meta_cap(bdev, cmd, argp);
> -
> switch (cmd) {
> case BLKFLSBUF:
> return blkdev_flushbuf(bdev, cmd, arg);
> @@ -647,9 +644,16 @@ static int blkdev_common_ioctl(struct block_device *bdev, blk_mode_t mode,
> return blkdev_pr_preempt(bdev, mode, argp, true);
> case IOC_PR_CLEAR:
> return blkdev_pr_clear(bdev, mode, argp);
> - default:
> - return -ENOIOCTLCMD;
> }
> +
> + if (_IOC_DIR(cmd) == _IOC_DIR(FS_IOC_GETLBMD_CAP) &&
> + _IOC_TYPE(cmd) == _IOC_TYPE(FS_IOC_GETLBMD_CAP) &&
> + _IOC_NR(cmd) == _IOC_NR(FS_IOC_GETLBMD_CAP) &&
> + _IOC_SIZE(cmd) >= LBMD_SIZE_VER0 &&
> + _IOC_SIZE(cmd) <= _IOC_SIZE(FS_IOC_GETLBMD_CAP))
This part is wrong as we handle larger sizes just fine via
copy_struct_{from,to}_user().
Arnd, objections to writing it as follows?:
diff --git a/block/ioctl.c b/block/ioctl.c
index 9ad403733e19..9887ec55f8ce 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -567,9 +567,6 @@ static int blkdev_common_ioctl(struct block_device *bdev, blk_mode_t mode,
{
unsigned int max_sectors;
- if (_IOC_NR(cmd) == _IOC_NR(FS_IOC_GETLBMD_CAP))
- return blk_get_meta_cap(bdev, cmd, argp);
-
switch (cmd) {
case BLKFLSBUF:
return blkdev_flushbuf(bdev, cmd, arg);
@@ -647,9 +644,25 @@ static int blkdev_common_ioctl(struct block_device *bdev, blk_mode_t mode,
return blkdev_pr_preempt(bdev, mode, argp, true);
case IOC_PR_CLEAR:
return blkdev_pr_clear(bdev, mode, argp);
- default:
- return -ENOIOCTLCMD;
}
+
+ /* extensible ioctls */
+ switch (_IOC_NR(cmd)) {
+ case _IOC_NR(FS_IOC_GETLBMD_CAP):
+ if (_IOC_DIR(cmd) != _IOC_DIR(FS_IOC_GETLBMD_CAP))
+ break;
+ if (_IOC_TYPE(cmd) != _IOC_TYPE(FS_IOC_GETLBMD_CAP))
+ break;
+ if (_IOC_NR(cmd) != _IOC_NR(FS_IOC_GETLBMD_CAP))
+ break;
+ if (_IOC_SIZE(cmd) < LBMD_SIZE_VER0)
+ break;
+ if (_IOC_SIZE(cmd) > PAGE_SIZE)
+ break;
+ return blk_get_meta_cap(bdev, cmd, argp);
+ }
+
+ return -ENOIOCTLCMD;
}
/*
And can I ask you to please take a look at fs/pidfs.c:pidfd_ioctl() and
fs/nsfs.c:ns_ioctl()?
