Re: [syzbot] [PATCH] usbhid: fix array-index-out-of-bounds in usbhid_parse UBSAN warning

[Kees Cook]

Hi,

What's happened to getting a new version of this patch? This flaw is
still reachable in -next from what I can see?

Thanks,

-Kees

On Sat, Dec 23, 2023 at 11:59:51AM -0800, syzbot wrote:
> For archival purposes, forwarding an incoming command email to
> linux-kernel@xxxxxxxxxxxxxxx, syzkaller-bugs@xxxxxxxxxxxxxxxx.
> 
> ***
> 
> Subject: [PATCH] usbhid: fix array-index-out-of-bounds in usbhid_parse UBSAN warning
> Author: tintinm2017@xxxxxxxxx
> 
> #syz test: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> 
> Look at the bug https://syzkaller.appspot.com/bug?extid=c52569baf0c843f35495 reported by syzbot. Tested a patch through syzbot, which gives an error. 
> Requesting help from the maintainers to understand what is really going wrong in the code. 
> 
> Based on my understanding, I believe the value of the number of descriptors is calculated incorrectly before the for loop.
> 
> Signed-off-by: Attreyee Mukherjee <tintinm2017@xxxxxxxxx>
> ---
>  drivers/hid/usbhid/hid-core.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c
> index a90ed2ceae84..582ddbef448f 100644
> --- a/drivers/hid/usbhid/hid-core.c
> +++ b/drivers/hid/usbhid/hid-core.c
> @@ -1021,6 +1021,8 @@ static int usbhid_parse(struct hid_device *hid)
>  	       (hdesc->bLength - offset) / sizeof(struct hid_class_descriptor));
>  
>  	for (n = 0; n < num_descriptors; n++)
> +		if (n >= ARRAY_SIZE(hdesc->desc))
> +			break;
>  		if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
>  			rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
>  
> -- 
> 2.34.1
> 

-- 
Kees Cook