svn commit: r277072 - head/sys/netinet6

[Adrian Chadd]

On 12 January 2015 at 07:09, Alexander V. Chernikov
<melifaro at freebsd.org> wrote:
> On 12.01.2015 17:52, Gleb Smirnoff wrote:
>> Author: glebius
>> Date: Mon Jan 12 14:52:43 2015
>> New Revision: 277072
>> URL: https://svnweb.freebsd.org/changeset/base/277072
>>
>> Log:
>>   Do not go one layer down to check ifqueue length. First, not all drivers
>>   use ifqueue at all. Second, there is no point in this lockless check.
>>   Either positive or negative result of the check could be incorrect after
>>   a tick.
> Finally, no useless ifq_maxlen check!

Yeah, I killed it in the ipv4 world. Well, commented it out. It's been
bogus ever since kernels became re-entrant and SMP. :(



-adrian

> Thanks!
>>
>>   Sponsored by:       Nginx, Inc.
>>
>> Modified:
>>   head/sys/netinet6/ip6_output.c
>>
>> Modified: head/sys/netinet6/ip6_output.c
>> ==============================================================================
>> --- head/sys/netinet6/ip6_output.c    Mon Jan 12 13:53:40 2015        (r277071)
>> +++ head/sys/netinet6/ip6_output.c    Mon Jan 12 14:52:43 2015        (r277072)
>> @@ -905,8 +905,6 @@ passout:
>>               u_int32_t id = htonl(ip6_randomid());
>>               u_char nextproto;
>>
>> -             int qslots = ifp->if_snd.ifq_maxlen - ifp->if_snd.ifq_len;
>> -
>>               /*
>>                * Too large for the destination or interface;
>>                * fragment if possible.
>> @@ -924,18 +922,6 @@ passout:
>>               }
>>
>>               /*
>> -              * Verify that we have any chance at all of being able to queue
>> -              *      the packet or packet fragments
>> -              */
>> -             if (qslots <= 0 || ((u_int)qslots * (mtu - hlen)
>> -                 < tlen  /* - hlen */)) {
>> -                     error = ENOBUFS;
>> -                     IP6STAT_INC(ip6s_odropped);
>> -                     goto bad;
>> -             }
>> -
>> -
>> -             /*
>>                * If the interface will not calculate checksums on
>>                * fragmented packets, then do it here.
>>                * XXX-BZ handle the hw offloading case.  Need flags.
>>
>>
>
>