netdev - Re: r8169 device jumps from 12k/sec interrupts with traffic to over 100k/sec interrupts without traffic

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] [day] [month] [year] [list]

Date:	Tue, 15 Jul 2008 07:41:00 -0400
From:	"Andrew Paprocki" <andrew@...iboo.com>
To:	"Francois Romieu" <romieu@...zoreil.com>
Cc:	netdev@...r.kernel.org
Subject: Re: r8169 device jumps from 12k/sec interrupts with traffic to over 100k/sec interrupts without traffic

On Tue, Jul 15, 2008 at 2:56 AM, Francois Romieu <romieu@...zoreil.com> wrote:
> Andrew Paprocki <andrew@...iboo.com> :
>> I'm currently running 2.6.26-rc8 (NAPI enabled) and my r8169 device is
>> going a little crazy after a short time under load.
> [...]
>> What can I do to further debug this problem?
>
> Can you send your complete dmesg (especially the XID line) and a complete
> lspci -vvxx  of your system to identify the network device ?
>
> The content of the /proc/interrupts file would be welcome too.

Attached.

I modified r8169.c to print out the %04x status inside of every
interrupt. Obviously these would fly by the screen, but I could at
least make out what was going on when the behavior changed. While the
device was functioning normally, all the interrupts were either 0x0484
or 0x0001. It makes sense that 0x0004 (TxOK) would exist, but I have
no idea why 0x0080 (TxDescUnavail) is set or 0x0400, which isn't
defined in the driver. After the device stopped functioning normally
and the system hung, the entire screen was constantly filled with
0x0010 (RxOverflow).

> You may try http://userweb.kernel.org/~romieu/r8169/2.6.26-rc9/20080710/
> but it's more or less a shot in the dark for now.

Does the above information indicate any of these might help the situation?

Thanks,
-Andrew

Download attachment "dmesg" of type "application/octet-stream" (14176 bytes)

Download attachment "lspci" of type "application/octet-stream" (9745 bytes)

Download attachment "interrupts" of type "application/octet-stream" (571 bytes)