Re: [PATCH V2 0/4] Add IPv6 support for iWARP

[Steve Wise <swise@opengridcomputing.com>]

On 6/19/2013 11:08 PM, David Miller wrote:
> From: Steve Wise <swise@opengridcomputing.com>
> Date: Wed, 19 Jun 2013 21:19:13 -0500
>
>> On 6/19/2013 8:01 PM, David Miller wrote:
>>> From: Vipul Pandya <vipul@chelsio.com>
>>> Date: Wed, 12 Jun 2013 17:11:38 +0530
>>>
>>>> We have included all the maintainers of respective drivers. Kindly
>>>> review the change and let us know in case of any review comments.
>>> I have not seen anyone review v2 of this patch series.
>>>
>> Reviewed-by: Steve Wise <swise@opengridcomputing.com>
> You wrote the first patch, and I bet you didn't even read the code in
> the cxgb4 driver.  So your review is sort of pointless... UNLESS you
> spotted the obvious bugs in these changes, that would have been
> interesting.
>
> Because NOBODY, and I mean NOBODY, even looked at the build of the
> cxgb4 changes.
>
> Tell me what this does:
>
>   	struct tid_info *t = dev->rdev.lldi.tids;
>   	int status = GET_AOPEN_STATUS(ntohl(rpl->atid_status));
> +	struct sockaddr_in *la = (struct sockaddr_in *)&ep->com.local_addr;
> +	struct sockaddr_in *ra = (struct sockaddr_in *)&ep->com.remote_addr;
> +	struct sockaddr_in6 *la6 = (struct sockaddr_in6 *)&ep->com.local_addr;
> +	struct sockaddr_in6 *ra6 = (struct sockaddr_in6 *)&ep->com.remote_addr;
> +
> +
>   
>   	ep = lookup_atid(t, atid);
>
> Dereferencing 'ep' before initializing it.
>
> The compiler complains loudly about this, therefore nobody even looked at
> the build logs from these changes before submitting them to me.
>
> That translates to "don't care", and if the people submitting this
> code don't care why should I?
>
> Sorry, not impressed.  I'm seriously going to take my time reviewing
> any future submissions of these changes, because it's obvious that
> even the people writing and submitting this code DO NOT CARE.
>

We do care.  We screwed up.