2009-07-15 05:55:25

by Oleksij Rempel

Subject: [PATCH] Solve the issue with memory corruption on Intel BIOS.

In Bug 13710 was reported "corrupted low memory after resume on Intel
DG45ID board". It use AMI BIOS with Intel vendor string. We need to add this
vendor to CONFIG_X86_RESERVE_LOW_64K workaround.

Signed-off-by: Alexey Fisher <[email protected]>
---
arch/x86/kernel/setup.c | 7 +++++++
1 files changed, 7 insertions(+), 0 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index de2cab1..89d00e2 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -672,6 +672,13 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = {
DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies"),
},
},
+ {
+ .callback = dmi_low_memory_corruption,
+ .ident = "Intel BIOS",
+ .matches = {
+ DMI_MATCH(DMI_BIOS_VENDOR, "Intel Corp."),
+ },
+ },
#endif
{}
};
--
1.6.3.3

2009-07-15 09:33:51

by Alan

Subject: Re: [PATCH] Solve the issue with memory corruption on Intel BIOS.

On Wed, 15 Jul 2009 07:55:14 +0200
Alexey Fisher <[email protected]> wrote:

> In Bug 13710 was reported "corrupted low memory after resume on Intel
> DG45ID board". It use AMI BIOS with Intel vendor string. We need to add this
> vendor to CONFIG_X86_RESERVE_LOW_64K workaround.

Given that this is the first report of a problem with an Intel BIOS, and
there are millions of systems not affected your match is far too wide I
think. Match the board identifier as well at least rather than randomly
making everyone suffer.

2009-07-15 15:11:37

by Pavel Machek

Subject: Re: [PATCH] Solve the issue with memory corruption on Intel BIOS.

On Wed 2009-07-15 10:34:31, Alan Cox wrote:
> On Wed, 15 Jul 2009 07:55:14 +0200
> Alexey Fisher <[email protected]> wrote:
>
> > In Bug 13710 was reported "corrupted low memory after resume on Intel
> > DG45ID board". It use AMI BIOS with Intel vendor string. We need to add this
> > vendor to CONFIG_X86_RESERVE_LOW_64K workaround.
>
> Given that this is the first report of a problem with an Intel BIOS, and
> there are millions of systems not affected your match is far too wide I
> think. Match the board identifier as well at least rather than randomly
> making everyone suffer.

Given the intel pushing linux these days... could we get them to
debug&fix their bios?

--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2009-07-15 15:57:49

by Oleksij Rempel

Subject: Re: [PATCH] Solve the issue with memory corruption on Intel BIOS.

Alan Cox schrieb:
> On Wed, 15 Jul 2009 07:55:14 +0200
> Alexey Fisher <[email protected]> wrote:
>
>> In Bug 13710 was reported "corrupted low memory after resume on Intel
>> DG45ID board". It use AMI BIOS with Intel vendor string. We need to add this
>> vendor to CONFIG_X86_RESERVE_LOW_64K workaround.
>
> Given that this is the first report of a problem with an Intel BIOS, and
> there are millions of systems not affected your match is far too wide I
> think. Match the board identifier as well at least rather than randomly
> making everyone suffer.

It did a memory dump on "windows vista sp2" before and after suspend..
there is the same corruption on same place. But it looks like windows do
not use this part so intensive like linux do.

i send new patch to mailing list:
[PATCH] Lowmemory corruption. Quirk for Intel DG45ID board.