[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1361384183.8579.1.camel@jlt4.sipsolutions.net>
Date: Wed, 20 Feb 2013 19:16:23 +0100
From: Johannes Berg <johannes@...solutions.net>
To: Eric Dumazet <eric.dumazet@...il.com>
Cc: Marc MERLIN <marc@...lins.org>, David Miller <davem@...emloft.net>,
Larry.Finger@...inger.net, bhutchings@...arflare.com,
linux-wireless@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: 3.7.8/amd64 full interrupt hangs due to iwlwifi under big nfs
copies out
On Wed, 2013-02-20 at 09:24 -0800, Eric Dumazet wrote:
> On Wed, 2013-02-20 at 18:01 +0100, Johannes Berg wrote:
>
> > FWIW, I think just using order-0 pages and turning 8k A-MSDUs off by
> > default makes more sense, A-MSDU is rarely used to begin with ...
> >
>
> My suggested patch makes sure that if someone needs 8k A-MSDU, iwlwifi
> still works correctly.
True, but if we don't advertise 8k A-MSDU then they won't be used, and
it'll still work correctly, just not be able to receive such large
A-MSDUs. Linux can't even transmit _any_ A-MSDUs, and many (most?) APs
don't either.
> > Also, if we copy larger frames here, we should also take into account
> > the (variable) 802.11 header length to avoid copying into a position
> > where the IP header ends up being unaligned.
>
> But the current code already has this problem (if its a problem at all,
> as on x86 NET_IP_ALIGN is 0)
Yes and no. The current code doesn't pull in much data, not even the IP
header, so worst case it copies 14 bytes (ethernet header) in mac80211
to obtain alignment. By pulling in everything here, it later has to copy
everything again later to obtain alignment.
johannes
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists