xfs
[Top] [All Lists]

Re: #undef TCP_DEBUG

To: Eric Sandeen <sandeen@xxxxxxx>, linux-xfs@xxxxxxxxxxx
Subject: Re: #undef TCP_DEBUG
From: "Nathan Scott" <nathans@xxxxxxxxxxxxxxxxxxxxxxxx>
Date: Sun, 14 Jan 2001 15:55:55 -0400
In-reply-to: Eric Sandeen <sandeen@sgi.com> "Re: #undef TCP_DEBUG" (Jan 13, 10:27pm)
References: <20010114011416.A732@main.braxis.co.uk> <3A612A9D.E97A0778@sgi.com>
Sender: owner-linux-xfs@xxxxxxxxxxx
On Jan 13, 10:27pm, Eric Sandeen wrote:
> Subject: Re: #undef TCP_DEBUG
> These are the only printks that get turned on w/ TCP_DEBUG, all
> seemingly related to bad things happening:
> 
> printk(KERN_DEBUG "TCP: peer %u.%u.%u.%u:%u/%u shrinks window %u:%u:%u.
> Bad, what else can I say?\n",
> printk("BUG: retransmit in tcp_data_queue: seq %X\n",
> TCP_SKB_CB(skb)->seq);
> printk(KERN_DEBUG "*** tcp.c:tcp_data bug acked < copied\n")
> printk(KERN_DEBUG "reset_xmit_timer sk=%p %d when=0x%lx, caller=%p\n",

I'm guessing its this one -^ right? (was recently reported
on lkml & the verdict there seemed to be that its harmless).

> > 
> > Wouldn't it be a good idea to undef TCP_DEBUG in include/tcp.h ,
> > it's only cosmetic change (unless you modify tcp code ;-) )
> > 

include/net/tcp.h - fix is probably just to nuke line 21.
i believe this is fixed since 2.4.0-ac3... not sure when we
plan to move up from 2.4.0, but it'll likely be fixed when
we do.

cheers.

-- 
Nathan

<Prev in Thread] Current Thread [Next in Thread>