Re: [Qemu-devel] [PATCH v9 4/4] xilinx_zynq: machine model initial version

[Peter Crosthwaite]

nvm,

youve got it covered. Thanks Anthony.

On Sat, Mar 10, 2012 at 8:15 PM, Peter Crosthwaite
<address@hidden> wrote:
> On it
>
> On Fri, Mar 9, 2012 at 2:42 AM, Anthony Liguori <address@hidden> wrote:
>> On 03/04/2012 10:39 PM, Peter A. G. Crosthwaite wrote:
>>>
>>> Xilinx zynq-7000 machine model. Also includes device model for the
>>> zynq-specific
>>> system level control register (SLCR) module.
>>>
>>> Signed-off-by: Peter A. G. Crosthwaite<address@hidden>
>>> Acked-by: Edgar E. Iglesias<address@hidden>
>>
>>
>> This breaks the build for me:
>>
>> cc1: warnings being treated as errors
>> /home/anthony/git/qemu/hw/zynq_slcr.c: In function ‘zynq_slcr_read’:
>> /home/anthony/git/qemu/hw/zynq_slcr.c:314:25: error: array subscript is
>> above array bounds
>> /home/anthony/git/qemu/hw/zynq_slcr.c: In function ‘zynq_slcr_write’:
>> /home/anthony/git/qemu/hw/zynq_slcr.c:457:22: error: array subscript is
>> above array bounds
>> make[1]: *** [zynq_slcr.o] Error 1
>>
>>
>>
>>> +        case 0xA60 ... 0xA8C:
>>> +            s->afi[0][(offset - 0x700) / 4] = val;
>>> +            break;
>>
>>
>> And GCC is definitely right here.  This will pretty clearly overflow the
>> array if offset is any of the allowable values per the switch statement.
>>
>> Regards,
>>
>> Anthony Liguori