lkml.org 
[lkml]   [2009]   [Mar]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
DateThu, 26 Mar 2009 11:40:59 +0800
FromHerbert Xu <>
SubjectRe: Revert "gro: Fix legacy path napi_complete crash",
On Wed, Mar 25, 2009 at 08:20:50PM -0700, David Miller wrote:
>
> There is still a difference compared to your fix Herbert. Jarek's
> patch flushes GRO first before the unlink.
>
> I still believe that's critical, although like you I can't pinpoint
> why.
>
> I know that GRO ought to be disabled here, but what if for some reason
> it isn't? :-)

Sure, I can accept that somehow someone has enabled GRO :) But
I'd still like to know why flushing the GRO afterwards would
lead to a hang.

> Adam Richter has successfully tested Jarek's variant, and if Ingo's
> tests show that it makes his problem go away too then I'm definitely
> going with Jarek's patch.

I don't have a problem with that.

I've asked Adam to test my patch as well. So far the only failure
case against it has been on Ingo's machine, where the process_backlog
path isn't involved. Yes it's used for loopback, but he's seeing
the hang on eth0, which with that config uses netif_receive_skb
so it should continue to work even if process_backlog completely
seizes up.

Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

\
 
 \ /
  Last update: 2009-03-26 04:45    [from the cache]
©2003-2010