Re: [PATCH] vgacon: Fix an out-of-bounds in vgacon_scrollback_update()

[Jiri Slaby <jirislaby@kernel.org>]

On 30. 07. 20, 15:24, Yang Yingliang wrote:
> 
> On 2020/7/30 19:04, Jiri Slaby wrote:
>> On 13. 07. 20, 12:57, Yang Yingliang wrote:
>>> I got a slab-out-of-bounds report when I doing fuzz test.
>>>
>>> [  334.989515]
>>> ==================================================================
>>> [  334.989577] BUG: KASAN: slab-out-of-bounds in
>>> vgacon_scroll+0x57a/0x8ed
>>> [  334.989588] Write of size 1766 at addr ffff8883de69ff3e by task
>>> test/2658
>>> [  334.989593]
>>> [  334.989608] CPU: 3 PID: 2658 Comm: test Not tainted
>>> 5.7.0-rc5-00005-g152036d1379f #789
>>> [  334.989617] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
>>> BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
>>> [  334.989624] Call Trace:
>>> [  334.989646]  dump_stack+0xe4/0x14e
>>> [  334.989676]  print_address_description.constprop.5+0x3f/0x60
>>> [  334.989699]  ? vgacon_scroll+0x57a/0x8ed
>>> [  334.989710]  __kasan_report.cold.8+0x92/0xaf
>>> [  334.989735]  ? vgacon_scroll+0x57a/0x8ed
>>> [  334.989761]  kasan_report+0x37/0x50
>>> [  334.989789]  check_memory_region+0x1c1/0x1e0
>>> [  334.989806]  memcpy+0x38/0x60
>>> [  334.989824]  vgacon_scroll+0x57a/0x8ed
>>> [  334.989876]  con_scroll+0x4ef/0x5e0
>> ...
>>> Because vgacon_scrollback_cur->tail plus memcpy size is greater than
>>> vgacon_scrollback_cur->size. Fix this by checking the memcpy size.
>>>
>>> Reported-by: Hulk Robot <hulkci@huawei.com>
>>> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
>>> ---
>>>   drivers/video/console/vgacon.c | 11 ++++++++---
>>>   1 file changed, 8 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/video/console/vgacon.c
>>> b/drivers/video/console/vgacon.c
>>> index 998b0de1812f..b51ffb9a208d 100644
>>> --- a/drivers/video/console/vgacon.c
>>> +++ b/drivers/video/console/vgacon.c
>>> @@ -243,6 +243,7 @@ static void vgacon_scrollback_startup(void)
>>>   static void vgacon_scrollback_update(struct vc_data *c, int t, int
>>> count)
>>>   {
>>>       void *p;
>>> +    int size;
>>>         if (!vgacon_scrollback_cur->data ||
>>> !vgacon_scrollback_cur->size ||
>>>           c->vc_num != fg_console)
>>> @@ -251,13 +252,17 @@ static void vgacon_scrollback_update(struct
>>> vc_data *c, int t, int count)
>>>       p = (void *) (c->vc_origin + t * c->vc_size_row);
>>>         while (count--) {
>>> +        size = vgacon_scrollback_cur->size -
>>> vgacon_scrollback_cur->tail;
>>> +        if (size > c->vc_size_row)
>>> +            size = c->vc_size_row;
>>> +
>>>           scr_memcpyw(vgacon_scrollback_cur->data +
>>>                   vgacon_scrollback_cur->tail,
>>> -                p, c->vc_size_row);
>>> +                p, size);
>> Are you sure the consumer can handle split lines? As vgacon_scrolldelta
>> (soff in particular) looks to me like it doesn't.
>>
>> Have you tested you patch? I mean with soft scrollback on the vga
>> console?
> 
> I only test the patch with the reproduce program.

Out of curiosity, what is it doing? Resize and then scroll by \n (line
feed)? Can you share it?

thanks,
-- 
js