Re: [RFC] 4KB stack + irq stack for x86

David Mosberger (davidm@napali.hpl.hp.com)
Thu, 6 Jun 2002 13:27:06 -0700


>>>>> On 06 Jun 2002 21:49:15 +0200, Andi Kleen <ak@suse.de> said:

Andi> "Ulrich Weigand" <Ulrich.Weigand@de.ibm.com> writes:
>> (*Really* ugly is s390x, because we need about twice as much
>> stack on average than on s390, but page size is still only 4K --
>> most other 64-bit platforms have 8K page size ...)

Andi> <minor detail, but perhaps still interesting>

Andi> Seems to be an old myth. Actually the 4K paged 64bit platforms
Andi> are in the majority.

Andi> 64bit linux platforms:

Andi> 4K page: x86-64, ppc64, s390x, mips64, parisc64(?) 8K: alpha,
Andi> sparc64 8-64K: ia64

Just a minor nit: for ia64 it's either 32KB (for page sizes up to
16KB) or 64KB (for 64KB page size). The 32KB is conservative and
based on the assumption that there can be up to 16 nested interrupts
plus some other nested traps (such as unaligned faults). A separate
irq stack should let us reduce the per-task stack size.

--david
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/