netdev - Re: [PATCH] tcp FRTO: in-order-only "TCP proxy" fragility workaround (fwd)

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:	Thu, 30 Oct 2008 16:16:04 -0200
From:	Dâniel Fraga <fragabr@...il.com>
To:	"Ilpo Järvinen" <ilpo.jarvinen@...sinki.fi>
Cc:	Thomas Gleixner <tglx@...utronix.de>,
	David Miller <davem@...emloft.net>,
	Netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH] tcp FRTO: in-order-only "TCP proxy" fragility
 workaround (fwd)

On Thu, 30 Oct 2008 12:43:05 +0200 (EET)
"Ilpo Järvinen" <ilpo.jarvinen@...sinki.fi> wrote:

> Nothing particularly interesting in it (sorry for the delay, there were 
> some other bugs in the meantime to fix and partially because I was out 
> of ideas). Minor increase in cpu/wa after nmap, before that it was 0-2, 
> mostly zero.

	Hi Ilpo. No problem, it's better late than never ;)

> Another thing that occurred to me today is that all these events I know of 
> are related to syslog/klogd (I don't know too well how they relate to each 
> other and what functionality is provided by each, thus I refer them here 
> as one :-)).
> 
> - sudo sendmsg'ed to /dev/log and stalled
> - every-2nd-char missing in printk relates to it
> - usb does cause printks
> - server processes likely do some logging (and can then hang on waiting
>   until that finishes and therefore eventually no worker is available to 
>   process requests)
> 
> ...Some other type of case I'm missing here?

	Yes, it is that. Anyway I'm talking directly to netfilter
people and they suggested that I use ulogd to capture the exact packet
when the stall happens using pcap. I'll try that. If they can figure it
out what happens to make the log prints every other char, maybe we have
a hint.

> Perhaps we could try to solve it though stracing syslogd...

	Yes I can trace it.

> Did you ever actually try btw with something else than NO_TICK, we 
> discussed it at some point of time but I'm not sure what was the end 
> result of what got tested and what not?

	I tried with many options, with TICK enabled and disabled, high
res timer enabled and disabled and other options in the kernel, but it
didn't make any difference.

	I'll try with netfilter people and if i get something, I
will reply ok?

-- 
--
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