[Web10g-user] patch error for 3.3 kernel

[rapier]

Srikanth,

I'm guessing that this is in the do_tcp_sendpages function. We have that 
set up as follows:

wait_for_memory:
      tcp_push(sk, flags & ~MSG_MORE, mss_now, TCP_NAGLE_PUSH);

      if (copied)
         TCP_ESTATS_UPDATE(tp, tcp_estats_update_writeq(sk));

      if ((err = sk_stream_wait_memory(sk, &timeo)) != 0)
         goto do_error;

      mss_now = tcp_send_mss(sk, &size_goal, flags);
  }

I believe in your fix you'll end up over counting. It would build and 
seem to work correctly but your stats will be off.

We only really provide patches for the minor revision (3.3, 3.4, etc). 
Mostly because it's a *lot* of work to keep up with all of the sub 
revisions. So when people are having a patching problem it often helps 
to take a look at what we've done in the following revision to see how 
we've dealt with all of the changes that have happened. Of course, we 
are also here to help out with these questions.

Chris




On 9/6/13 9:50 AM, Srikanth Sundaresan wrote:
> Hi all,
>
> I tried the web10g-estats-0.1-3.3 patch for the 3.3.8 kernel. All the patches applied seamlessly, except one.
>
> Here's the part that caused the problem (this is the "fixed" version that applies):
> https://github.com/projectbismark/bismark-openwrt-overlay/blob/c9af799be0f0f85ed9dca54f518192f4abe17042/target/linux/generic/patches-3.3/999-estats-nl-0.1-3.3.patch
>
> The diff from the original patch is very small - the original is here:
> https://gist.github.com/ssundaresan/6463951
>
> This patch compiles, and I tested it with the userland, and it works.
>
> I still would like to confirm that this change is safe, though.
>
> Thanks,
> - Srikanth
>
>
> _______________________________________________
> Web10g-user mailing list
> Web10g-user at web10g.org
> https://lists.psc.edu/mailman/listinfo/web10g-user
>
> To UNSUBSCRIBE visit https://lists.psc.edu/mailman/unsubscribe/web10g-user
>