Bug#610085: [2.6.26 -> 2.6.32 regression] [xeon x3430] symptoms of memory corruption when more than one CPU core is activated
- To: Jonathan Nieder <firstname.lastname@example.org>
- Cc: email@example.com
- Subject: Bug#610085: [2.6.26 -> 2.6.32 regression] [xeon x3430] symptoms of memory corruption when more than one CPU core is activated
- From: Adrian Kieß <firstname.lastname@example.org>
- Date: Thu, 26 Jan 2012 04:57:28 +0100
- Message-id: <CADAQ8f1=-RkkWiUQy5Ost=ni7rYsTePthGVYT02u6RZPycL2Sg@mail.gmail.com>
- Reply-to: Adrian Kieß <email@example.com>, firstname.lastname@example.org
- In-reply-to: <20120126001325.GA4479@burratino>
- References: <BANLkTi=N1LSnXcLz1eeAd556WnVJe4Asow@mail.gmail.com> <20120126001325.GA4479@burratino>
the problem went away with kernel 3.0 and 3.1.
Im now happily running Debian/testing on my HP ProLiant G6 with Intel Xeon processor.
Previous kernels are crashing though.
On Thu, Jan 26, 2012 at 1:13 AM, Jonathan Nieder <email@example.com>
Adrian Kieß wrote:
> i just wanted to ask if the bug is resolved already. Booting into kernel
> 2.6.38 also brings the system to halt when more then one
> CPU core is activated. The system boots fine when only one CPU core (of
> four) is activated.
Sorry for the slow reply. The symptoms point to memory corruption,
so the question becomes _what_ is corrupting the memory.
The first step to answering that is to try as recent a kernel as
possible, so we can get help from upstream. Could you try a 3.x.y
kernel from sid or experimental and let us know how it goes?
Please also attach full "dmesg" output from booting such a kernel.
Sorry for the trouble,
Adrian Immanuel Kieß (adrian at immanuelK.net)