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

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

[Khazhismel Kumykov]

[ Upstream commit ccf16413e520164eb718cf8b22a30438da80ff23 ]

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>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Link: https://lore.kernel.org/r/20220414224056.2875681-1-khazhy@google.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
[compat_ioctl is it's own file in 5.4-stable and earlier]
---

The original commit should apply to the newer stables, this should apply
to all the older stables.

 block/compat_ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/compat_ioctl.c b/block/compat_ioctl.c
index 7f053468b50d..d490ac220ba8 100644
--- a/block/compat_ioctl.c
+++ b/block/compat_ioctl.c
@@ -393,7 +393,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 		return 0;
 	case BLKGETSIZE:
 		size = i_size_read(bdev->bd_inode);
-		if ((size >> 9) > ~0UL)
+		if ((size >> 9) > ~(compat_ulong_t)0)
 			return -EFBIG;
 		return compat_put_ulong(arg, size >> 9);
 
-- 
2.36.0.rc0.470.gd361397f0d-goog

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

[Greg KH]

On Tue, Apr 19, 2022 at 12:12:39PM -0700, Khazhismel Kumykov wrote:
> [ Upstream commit ccf16413e520164eb718cf8b22a30438da80ff23 ]
> 
> 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>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> Link: https://lore.kernel.org/r/20220414224056.2875681-1-khazhy@google.com
> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> [compat_ioctl is it's own file in 5.4-stable and earlier]
> ---
> 
> The original commit should apply to the newer stables

It does not, it only applied to 5.17.y.

Please provide working backports for all of the others.

> this should apply
> to all the older stables.

I'll wait for the 5.10.y and 5.15.y backport first before applying this
one.

thanks,

greg k-h

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

[Khazhy Kumykov]

[-- Attachment #1: Type: text/plain, Size: 1174 bytes --]

On Thu, Apr 21, 2022 at 3:02 AM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Tue, Apr 19, 2022 at 12:12:39PM -0700, Khazhismel Kumykov wrote:
> > [ Upstream commit ccf16413e520164eb718cf8b22a30438da80ff23 ]
> >
> > 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>
> > Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> > Link: https://lore.kernel.org/r/20220414224056.2875681-1-khazhy@google.com
> > Signed-off-by: Jens Axboe <axboe@kernel.dk>
> > [compat_ioctl is it's own file in 5.4-stable and earlier]
> > ---
> >
> > The original commit should apply to the newer stables
>
> It does not, it only applied to 5.17.y.
>
> Please provide working backports for all of the others.
>
> > this should apply
> > to all the older stables.
>
> I'll wait for the 5.10.y and 5.15.y backport first before applying this
> one.
I double checked and the above patch applied to 4.9-5.4 for me
>
> thanks,
>
> greg k-h

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 3999 bytes --]

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

[Greg KH]

On Mon, Apr 25, 2022 at 12:29:41PM -0700, Khazhy Kumykov wrote:
> On Thu, Apr 21, 2022 at 3:02 AM Greg KH <gregkh@linuxfoundation.org> wrote:
> >
> > On Tue, Apr 19, 2022 at 12:12:39PM -0700, Khazhismel Kumykov wrote:
> > > [ Upstream commit ccf16413e520164eb718cf8b22a30438da80ff23 ]
> > >
> > > 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>
> > > Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> > > Link: https://lore.kernel.org/r/20220414224056.2875681-1-khazhy@google.com
> > > Signed-off-by: Jens Axboe <axboe@kernel.dk>
> > > [compat_ioctl is it's own file in 5.4-stable and earlier]
> > > ---
> > >
> > > The original commit should apply to the newer stables
> >
> > It does not, it only applied to 5.17.y.
> >
> > Please provide working backports for all of the others.
> >
> > > this should apply
> > > to all the older stables.
> >
> > I'll wait for the 5.10.y and 5.15.y backport first before applying this
> > one.
> I double checked and the above patch applied to 4.9-5.4 for me

All now queued up, thanks.

greg k-h

[PATCH v5.10] block/compat_ioctl: fix range check in BLKGETSIZE

[Khazhismel Kumykov]

[ Upstream commit ccf16413e520164eb718cf8b22a30438da80ff23 ]

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>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Link: https://lore.kernel.org/r/20220414224056.2875681-1-khazhy@google.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 block/ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/ioctl.c b/block/ioctl.c
index ed240e170e14..e7eed7dadb5c 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -679,7 +679,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 			       (bdev->bd_bdi->ra_pages * PAGE_SIZE) / 512);
 	case BLKGETSIZE:
 		size = i_size_read(bdev->bd_inode);
-		if ((size >> 9) > ~0UL)
+		if ((size >> 9) > ~(compat_ulong_t)0)
 			return -EFBIG;
 		return compat_put_ulong(argp, size >> 9);
 
-- 
2.36.0.rc2.479.g8af0fa9b8e-goog

[PATCH v5.15] block/compat_ioctl: fix range check in BLKGETSIZE

[Khazhismel Kumykov]

[ Upstream commit ccf16413e520164eb718cf8b22a30438da80ff23 ]

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>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Link: https://lore.kernel.org/r/20220414224056.2875681-1-khazhy@google.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 block/ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/ioctl.c b/block/ioctl.c
index a31be7fa31a5..cd506a902963 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -645,7 +645,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 			(bdev->bd_disk->bdi->ra_pages * PAGE_SIZE) / 512);
 	case BLKGETSIZE:
 		size = i_size_read(bdev->bd_inode);
-		if ((size >> 9) > ~0UL)
+		if ((size >> 9) > ~(compat_ulong_t)0)
 			return -EFBIG;
 		return compat_put_ulong(argp, size >> 9);
 
-- 
2.36.0.rc2.479.g8af0fa9b8e-goog