Re: Corruption Stats: Suggested Blacklist from the data

Linus Torvalds (torvalds@transmeta.com)
30 Jul 1998 06:17:45 GMT


In article <19980730122531.A13155@caffeine.ix.net.nz>,
Chris Wedgwood <chris@cybernet.co.nz> wrote:
>On Thu, Jul 30, 1998 at 12:52:22AM +0200, MOLNAR Ingo wrote:
>
>> so the only thing left is to poke the PCI bus directly and measure that
>> speed. Unfortunately i have found no way to create some reliable
>> nonintrusive transaction on the PCI bus that can be measured. maybe
>> reading the config space takes some constant number of PCI cycles?
>
>reading/writing to a frame-buffer - using rdtsc ?

Note that even if you have a frame-buffer, there is nothing to guarantee
that it does single-cycle PCI accesses.

As far as I can tell, the only way to get the PCI bus speed is to try to
access all the devices you have on the thing, time them all (discounting
the ones that are behind a bridge, like a PCI-ISA thing), and try to
look for the lowest common denominator. And personally I wouldn't trust
even that - with write posting potentially on the CPU and having to take
into account the synchronize of the CPU clock to the bus lock etc I
don't see anything that could be considered workable.

Accessing the north or south bridge sounds like the most likely
candidate to get single-cycle accesses, but even with them I'd wonder.

Linus

-
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.altern.org/andrebalsa/doc/lkml-faq.html