[Skiboot] [PATCH 3/3] hdata: Prevent NULL dereference on duplicate slot map info

Stewart Smith stewart at linux.ibm.com
Mon Mar 25 10:28:51 AEDT 2019


Vasant Hegde <hegdevasant at linux.vnet.ibm.com> writes:
> On 03/20/2019 12:03 PM, Stewart Smith wrote:
>> Vasant Hegde <hegdevasant at linux.vnet.ibm.com> writes:
>>> On 03/18/2019 11:22 AM, Stewart Smith wrote:
>>>> Signed-off-by: Stewart Smith <stewart at linux.ibm.com>
>>>
>>> Patch looks good.
>>>
>>> I see few other places inside iohub.c where we don't check dt_new* return value.
>>> May be fix them as well?
>> 
>> Huh, we probably should. afl-fuzz is pretty good at crashing things
>> here, and I have mixed feelings about how much effort to put into it...
>> 
>
> May be we can add assert() inside dt_new* functions. If we fail to allocate memory
> for device tree then we have bigger problem anyway.

Although that would be it in the PCI hotplug code path, and being unable
to process the new device shouldn't be a fatal error.

-- 
Stewart Smith
OPAL Architect, IBM.



More information about the Skiboot mailing list