Re: possible deadlock in smp_call_function()

David Wragg (dpw@doc.ic.ac.uk)
27 Sep 1999 12:16:07 +0000


Manfred Spraul <manfreds@colorfullife.com> writes:
> [snip]
> David Wragg wrote:
> > It isn't there in the Pentium manual,
> The Pentium has no _PAGE_GLOBAL flag, the PPro is the first processor
> which supports this flag.

I meant that the TLB-flushing effect of wbinvd isn't there.

> > and isn't in the current IA32 manual.
> I didn't know that there is a IA32 manual, where can I find it (e.g the
> number of the PDF file, or the exact title would help)

http://developer.intel.com/design/pentiumii/manuals/243192.htm

> I found the Pentium III handbook on the Intel Developer CD, and they
> have removed the sentence that WBINVD clears the TLB cache. It seems
> this is a typo in the Pentium Pro handbook.

I think it must be true for the PPro. They would have issued a
documentation change otherwise. They probably decided that they didn't
want to be held to it in future processors. I'd be interested to know
if any of the P6 processors don't TLB flush on wbinvd.

Regards,
David Wragg

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