netdev - PROBLEM: tg3 spitting out uninitialized memory

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-next>] [day] [month] [year] [list]

Date:	Wed, 11 Apr 2007 22:49:29 +0100
From:	Jamie webb <j@...bb.sygneca.com>
To:	netdev@...r.kernel.org
Subject: PROBLEM: tg3 spitting out uninitialized memory

Hi there

I have a Dell PE860 with built-in BCM5721, which is reported as working 
fine with the tg3 driver, however I have been getting sporadic data 
corruption, mostly evident as SSH MAC errors.

This is with vanilla 2.6.20.6 on a dual-core Xeon. CpuBurn and 
Memtest86+ report no errors.

I ran:
tungsten$ head -c50000000 /dev/zero | nc radium 2000

And on another box:
radium$ netcat -l -p 2000 | hexdump -C

Result was mostly zeros, with a small burst of garbage at about the 24MB 
mark. I re-ran the test a few times, and the garbage quickly 
disappeared. So I compiled some stuff and then ran it again. Output this 
time was more interesting (attached as file 'hexdump' to stop this 
stupid mail client scrambling it).

Well, clearly the corruption is not due to line noise or the like, but 
due to the wrong data being sent. But only very occasionally. I can 
sometimes run an SSH session for an hour or more before getting chucked off.

Any suggestions for fixing this, or betting isolating the problem?

Many thanks

/J

View attachment "hexdump" of type "text/plain" (2383 bytes)

View attachment "ver_linux" of type "text/plain" (706 bytes)

View attachment "lspci" of type "text/plain" (20326 bytes)

View attachment "config" of type "text/plain" (37541 bytes)