Home » Mailing list archives » linux-kernel » 2008 » September » 8

Re: [PATCH 0 of 3] Low memory corruption detection and workaround

Previous thread: [GIT PULL] UBIFS updates for 2.6.27-rc6 by Artem Bityutskiy on Sunday, September 7, 2008 - 8:15 am. (1 message)

Next thread: Re: [PATCH 4/3] x86: default corruption check to off, but put parameter default in Kconfig by Ingo Molnar on Sunday, September 7, 2008 - 8:51 am. (3 messages)
[view original message]
From: Alan Jenkins
Subject: Re: [PATCH 0 of 3] Low memory corruption detection and workaround
Date: Sunday, September 7, 2008 - 8:24 am

Seems to work ok. 

(And here's the output, for the benefit of Google / kerneloops.org).

[   84.140181] Back to C!
[   84.140181] Corrupted low memory at ffff8800000083e8 (83e8 phys) =
803c85370cfc0000
[   84.140181] Corrupted low memory at ffff8800000083f0 (83f0 phys) =
00003000
[   84.140181] ------------[ cut here ]------------
[   84.140181] WARNING: at arch/x86/kernel/setup.c:698
check_for_bios_corruption+0xd9/0xe0()
[   84.140181] Memory corruption detected in low memory
[   84.140181] Modules linked in:
[   84.140181] Pid: 7813, comm: s2ram Not tainted
2.6.27-rc5-00322-gb84d508 #160
[   84.140181]
[   84.140181] Call Trace:
[   84.140181]  [<ffffffff80238a87>] warn_slowpath+0xb7/0xe0
[   84.140181]  [<ffffffff802524f6>] ? down_trylock+0x36/0x50
[   84.140181]  [<ffffffff80239330>] ? try_acquire_console_sem+0x10/0x40
[   84.140181]  [<ffffffff80212b69>] ? read_tsc+0x9/0x20
[   84.140181]  [<ffffffff80253eca>] ? getnstimeofday+0x3a/0xc0
[   84.140181]  [<ffffffff80239c07>] ? printk+0x67/0x70
[   84.140181]  [<ffffffff8036f3e3>] ? acpi_os_release_lock+0x9/0xb
[   84.140181]  [<ffffffff8037d75c>] ? acpi_set_register+0x161/0x173
[   84.140181]  [<ffffffff8020fe69>] check_for_bios_corruption+0xd9/0xe0
[   84.140181]  [<ffffffff803c1db9>] pm_dev_dbg+0x9/0x10
[   84.140181]  [<ffffffff803c28f2>] dpm_power_up+0x32/0xf0
[   84.140181]  [<ffffffff803c2b47>] device_power_up+0x17/0x20
[   84.140181]  [<ffffffff80261e6c>] suspend_devices_and_enter+0x16c/0x190
[   84.140181]  [<ffffffff8026205c>] enter_state+0x19c/0x1b0
[   84.140181]  [<ffffffff8026211f>] state_store+0xaf/0xf0
[   84.140181]  [<ffffffff803465b7>] kobj_attr_store+0x17/0x20
[   84.140181]  [<ffffffff802ff34b>] sysfs_write_file+0xeb/0x140
[   84.140181]  [<ffffffff802a9c87>] vfs_write+0xc7/0x180
[   84.140181]  [<ffffffff802aa290>] sys_write+0x50/0x90
[   84.140181]  [<ffffffff8020c35b>] system_call_fastpath+0x16/0x1b
[   84.140181]
[   84.140181] ---[ end trace e31222efe32d34ab ]---

--


[ message continues ]
[view original message]
From: Nick Piggin
Subject: Re: [PATCH 0 of 3] Low memory corruption detection and workaround
Date: Monday, September 8, 2008 - 4:23 am

Why not high memory as well? We put page tables there too...

(rant: why can we still buy non-ECC memory?)
--


[ message continues ]
[view original message]
From: Jeremy Fitzhardinge
Subject: Re: [PATCH 0 of 3] Low memory corruption detection and workaround
Date: Monday, September 8, 2008 - 9:47 am

Well, the specific problem is that the BIOS appears to corrupt low
memory.  It might corrupt other memory anywhere, but that would be
pretty pathologically evil.  The assumption is that it gets away with it

Well, it wouldn't help in this case.  It's not random hardware bit-flipping.

    J
--


[ message continues ]
[view original message]
From: Alan Cox
Subject: Re: [PATCH 0 of 3] Low memory corruption detection and workaround
Date: Monday, September 8, 2008 - 10:01 am

On Mon, 08 Sep 2008 09:47:07 -0700

Or that its hitting stuff that doesn't kill the OS...

What has the vendor had to say and if they are not being helpful has
anyone considered delivered them a report of the corruption and trace info
by say registered post with a suggestion that they might now be knowingly
exposing customers to unneccessary risk through their inaction ...

Alan
--


[ message continues ]
[view original message]
From: Nick Piggin
Subject: Re: [PATCH 0 of 3] Low memory corruption detection and workaround
Date: Monday, September 8, 2008 - 11:44 am

OK, fine, sorry I didn't follow the thread back.
--


[ message continues ]
Previous thread: [GIT PULL] UBIFS updates for 2.6.27-rc6 by Artem Bityutskiy on Sunday, September 7, 2008 - 8:15 am. (1 message)

Next thread: Re: [PATCH 4/3] x86: default corruption check to off, but put parameter default in Kconfig by Ingo Molnar on Sunday, September 7, 2008 - 8:51 am. (3 messages)