linux-kernel - Re: [PATCH] comedi: adl_pci9111: Fix possible division by zero in pci9111_ai_do_cmd

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <04320187-02bb-440c-9180-c1198dddecb3@mev.co.uk>
Date: Wed, 18 Sep 2024 13:09:04 +0100
From: Ian Abbott <abbotti@....co.uk>
To: Aleksandr Mishin <amishin@...rgos.ru>,
 H Hartley Sweeten <hsweeten@...ionengravers.com>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
 linux-kernel@...r.kernel.org, lvc-project@...uxtesting.org
Subject: Re: [PATCH] comedi: adl_pci9111: Fix possible division by zero in
 pci9111_ai_do_cmd_test()

On 18/09/2024 12:41, Ian Abbott wrote:
> On 18/09/2024 11:43, Aleksandr Mishin wrote:
>> Division by zero is possible in pci9111_ai_do_cmd_test() in case of scan
>> begin trigger source is TRIG_TIMER and either 'chanlist_len' or
>> 'convert_arg' is zero.
>>
>> Add zero value check to prevent division by zero.
>>
>> Found by Linux Verification Center (linuxtesting.org) with SVACE.
>>
>> Fixes: f1c51faabc4d ("staging: comedi: adl_pci9111: tidy up 
>> (*do_cmdtest) Step 4")
>> Signed-off-by: Aleksandr Mishin <amishin@...rgos.ru>
>> ---
>>   drivers/comedi/drivers/adl_pci9111.c | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/comedi/drivers/adl_pci9111.c 
>> b/drivers/comedi/drivers/adl_pci9111.c
>> index 086d93f40cb9..ec1fb570b98c 100644
>> --- a/drivers/comedi/drivers/adl_pci9111.c
>> +++ b/drivers/comedi/drivers/adl_pci9111.c
>> @@ -312,6 +312,8 @@ static int pci9111_ai_do_cmd_test(struct 
>> comedi_device *dev,
>>        */
>>       if (cmd->scan_begin_src == TRIG_TIMER) {
>>           arg = cmd->chanlist_len * cmd->convert_arg;
>> +        if (!arg)
>> +            return 4;
>>           if (arg < cmd->scan_begin_arg)
>>               arg *= (cmd->scan_begin_arg / arg);
> 
> Nice catch!  `cmd->convert_arg` will be non-zero due to earlier checks, 
> but `cmd->chanlist_len` could be zero for the `COMEDI_CMDTEST` ioctl. 
> (The function is called for the `COMEDI_CMDTEST` and `COMEDI_CMD` 
> ioctls, but only `COMEDI_CMD` checks that `chanlist_len` is non-zero 
> before calling the function.)
> 
> Reviewed-by: Ian Abbott <abbotti@....co.uk>

On second thoughts, the code should not return 4 (indicating a problem 
at checking step 4) here if `chanlist_len` is 0, because it is OK to use 
a zero `chanlist_len` value when using the `COMEDI_CMDTEST` ioctl.

Suggested fix is to keep your change, but to combine it with this change:

-	if (cmd->scan_begin_src == TRIG_TIMER) {
+	if (cmd->scan_begin_src == TRIG_TIMER && cmd->chanlist_len) {

I think `cmd->convert_arg` will be non-zero unless the call to 
`comedi_8254_cascade_ns_to_timer()` somehow manages to set it to zero.

-- 
-=( Ian Abbott <abbotti@....co.uk> || MEV Ltd. is a company  )=-
-=( registered in England & Wales.  Regd. number: 02862268.  )=-
-=( Regd. addr.: S11 & 12 Building 67, Europa Business Park, )=-
-=( Bird Hall Lane, STOCKPORT, SK3 0XA, UK. || www.mev.co.uk )=-