[PATCH] block/compat_ioctl: fix range check in BLKGETSIZE

Khazhismel Kumykov posted 1 patch 4 years, 2 months ago
There is a newer version of this series
block/ioctl.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
[PATCH] block/compat_ioctl: fix range check in BLKGETSIZE
Posted by Khazhismel Kumykov 4 years, 2 months ago
kernel ulong and compat_ulong_t may not be same width. Use type directly
to eliminate mismatches.

This would result in truncation rather than EFBIG for 32bit mode for
large disks.

Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
---
 block/ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Noticed this one was sitting in my "not landed yet" pile, third time's
the charm? :)

diff --git a/block/ioctl.c b/block/ioctl.c
index 4a86340133e4..959e93a90b29 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -629,7 +629,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 		return compat_put_long(argp,
 			(bdev->bd_disk->bdi->ra_pages * PAGE_SIZE) / 512);
 	case BLKGETSIZE:
-		if (bdev_nr_sectors(bdev) > ~0UL)
+		if (bdev_nr_sectors(bdev) > ~((compat_ulong_t)0UL))
 			return -EFBIG;
 		return compat_put_ulong(argp, bdev_nr_sectors(bdev));
 
-- 
2.35.1.1178.g4f1659d476-goog
Re: [PATCH] block/compat_ioctl: fix range check in BLKGETSIZE
Posted by Bart Van Assche 4 years, 2 months ago
On 4/8/22 16:47, Khazhismel Kumykov wrote:
> kernel ulong and compat_ulong_t may not be same width. Use type directly
> to eliminate mismatches.
> 
> This would result in truncation rather than EFBIG for 32bit mode for
> large disks.
> 
> Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
> ---
>   block/ioctl.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Noticed this one was sitting in my "not landed yet" pile, third time's
> the charm? :)
> 
> diff --git a/block/ioctl.c b/block/ioctl.c
> index 4a86340133e4..959e93a90b29 100644
> --- a/block/ioctl.c
> +++ b/block/ioctl.c
> @@ -629,7 +629,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
>   		return compat_put_long(argp,
>   			(bdev->bd_disk->bdi->ra_pages * PAGE_SIZE) / 512);
>   	case BLKGETSIZE:
> -		if (bdev_nr_sectors(bdev) > ~0UL)
> +		if (bdev_nr_sectors(bdev) > ~((compat_ulong_t)0UL))
>   			return -EFBIG;
>   		return compat_put_ulong(argp, bdev_nr_sectors(bdev));

A nit: the "UL" and two parentheses can be left out. Anyway:

Reviewed-by: Bart Van Assche <bvanassche@acm.org>
[PATCH v2] block/compat_ioctl: fix range check in BLKGETSIZE
Posted by Khazhismel Kumykov 4 years, 2 months ago
kernel ulong and compat_ulong_t may not be same width. Use type directly
to eliminate mismatches.

This would result in truncation rather than EFBIG for 32bit mode for
large disks.

Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
---
 block/ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

v2: addressed bart's comment

diff --git a/block/ioctl.c b/block/ioctl.c
index 4a86340133e4..f8703db99c73 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -629,7 +629,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 		return compat_put_long(argp,
 			(bdev->bd_disk->bdi->ra_pages * PAGE_SIZE) / 512);
 	case BLKGETSIZE:
-		if (bdev_nr_sectors(bdev) > ~0UL)
+		if (bdev_nr_sectors(bdev) > ~(compat_ulong_t)0)
 			return -EFBIG;
 		return compat_put_ulong(argp, bdev_nr_sectors(bdev));
 
-- 
2.36.0.rc0.470.gd361397f0d-goog
Re: [PATCH v2] block/compat_ioctl: fix range check in BLKGETSIZE
Posted by Jens Axboe 4 years, 2 months ago
On Thu, 14 Apr 2022 15:40:56 -0700, Khazhismel Kumykov wrote:
> kernel ulong and compat_ulong_t may not be same width. Use type directly
> to eliminate mismatches.
> 
> This would result in truncation rather than EFBIG for 32bit mode for
> large disks.
> 
> 
> [...]

Applied, thanks!

[1/1] block/compat_ioctl: fix range check in BLKGETSIZE
      commit: ccf16413e520164eb718cf8b22a30438da80ff23

Best regards,
-- 
Jens Axboe
Re: [PATCH v2] block/compat_ioctl: fix range check in BLKGETSIZE
Posted by Chaitanya Kulkarni 4 years, 2 months ago
On 4/14/22 15:40, Khazhismel Kumykov wrote:
> kernel ulong and compat_ulong_t may not be same width. Use type directly
> to eliminate mismatches.
> 
> This would result in truncation rather than EFBIG for 32bit mode for
> large disks.
> 
> Reviewed-by: Bart Van Assche <bvanassche@acm.org>
> Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
> ---

Looks good.

Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>

-ck