netdev - Re: scp stalls mysteriously

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Fri, 4 Dec 2009 17:09:45 +0200 (EET)
From:	"Ilpo Järvinen" <ilpo.jarvinen@...sinki.fi>
To:	Frederic Leroy <fredo@...rox.org>,
	Eric Dumazet <eric.dumazet@...il.com>
cc:	Damian Lukowski <damian@....rwth-aachen.de>,
	Netdev <netdev@...r.kernel.org>,
	David Miller <davem@...emloft.net>,
	Herbert Xu <herbert@...dor.apana.org.au>,
	Greg KH <gregkh@...e.de>
Subject: Re: scp stalls mysteriously

On Fri, 4 Dec 2009, Frederic Leroy wrote:

> On Fri, Dec 04, 2009 at 01:14:46PM +0200, Ilpo Järvinen wrote:
> > On Fri, 4 Dec 2009, Frederic Leroy wrote:
> > 
> > I noticed that we'll absolutely need that || !retrans_stamp thing as 
> > [...] 
> > But there may be other error handling related things here too. Can you try 
> > with the following once you can (probably can print quite much, once per 
> > arriving packet when in recovery):
> 
> I made 2 more trace .15 and .16 but ... 
> I applied your patch and forgot to add "|| !retrans_stamp" thing.
> 
> I can't test another kernel until this evening :(

I think it's fine without it too (I just meant that we'll need that || 
!retrans_stamp into mainline/stable for sure even if we find the cause for 
the another problem) and your traces now confirm what I was suspecting.
The retransmissions do not succeed because of error (-EAGAIN). Possibly 
from this check: 

        if (atomic_read(&sk->sk_wmem_alloc) >
            min(sk->sk_wmem_queued + (sk->sk_wmem_queued >> 2), sk->sk_sndbuf))
                return -EAGAIN;

...Or alternatively it comes from queue_xmit.

Eric, can you please take a look if you have a clue on this?


-- 
 i.