2006-10-06 18:10:28

by Aneesh Kumar K.V

Subject: safe_smp_processor_id i386 and x86_64

I was looking at the git log and the below changes confused me.

dc2bc768a009b9ad8711894c544dc6b0d8c0ce57

commit dc2bc768a009b9ad8711894c544dc6b0d8c0ce57
Author: Fernando Vazquez <[email protected]>
Date: Sat Sep 30 23:29:07 2006 -0700

[PATCH] stack overflow safe kdump: safe_smp_processor_id()

..............

This patch set does the following:

* Add safe_smp_processor_id function to i386 architecture (this function was
inspired by the x86_64 function of the same name).

151f8cc1169f9052095b2be36183ab132d75c6c2

commit 151f8cc1169f9052095b2be36183ab132d75c6c2
Author: Andi Kleen <[email protected]>
Date: Tue Sep 26 10:52:37 2006 +0200

[PATCH] Remove safe_smp_processor_id()




-aneesh

2006-10-11 16:59:44

by Eric W. Biederman

Subject: Re: safe_smp_processor_id i386 and x86_64

"Aneesh Kumar K.V" <[email protected]> writes:

> I was looking at the git log and the below changes confused me.

They are being used to solve different problems.
smp_processor_id on i386 doesn't work when a stack overflow
occurs. safe_smp_processor_id does.

On x86_64 since it is a simple register value we don't have that problem.

Eric