Re: [PATCH][linux 2.6.18] remove pointless error handling in scsiback

["Jürgen Groß" <jgross@suse.com>]

On 07/02/2014 02:33 PM, Jan Beulich wrote:
>>>> On 02.07.14 at 14:27, <JBeulich@suse.com> wrote:
>>>>> On 02.07.14 at 14:13, <JGross@suse.com> wrote:
>>> On 07/02/2014 01:57 PM, Jan Beulich wrote:
>>>> @@ -229,8 +230,8 @@ static void __report_luns(pending_req_t
>>>>    	unsigned int req_bufflen = 0;
>>>>    	unsigned int actual_len = 0;
>>>>    	unsigned int retry_cnt = 0;
>>>
>>> When you are cleaning up local variables, IMHO you could remove above 3
>>> unneeded initializations, too.
>>
>> Certainly not retry_cnt, but perhaps you meant (apart from the other
>> two visible here) alloc_luns (immediately prior to the first context line).

Sorry, yes.

>> Yes, I guess I'll do that.
>
> And indeed very useful to have this done, as it resulted in me spotting
> that alloc_len was of type unsigned char, i.e. setting us up for memory
> corruption as soon as the total buffer size needed would exceed 255
> bytes. Fixed at once (and I take it that this doesn't invalidate your
> Reviewed-by).

In this case covered by "In any case" :-)

Juergen