System clock died on e420r
Poul-Henning Kamp
phk at phk.freebsd.dk
Tue Feb 15 00:39:44 PST 2005
In message <20050214230635.GA86332 at xor.obsecurity.org>, Kris Kennaway writes:
>On Mon, Feb 14, 2005 at 11:14:51PM +0100, Poul-Henning Kamp wrote:
>>=20
>> If hardclock dies, timecounters (like much else) suffers.
>
>Surely, so why is it happening and how can it be fixed? :)
>
>jhb suspected there might be unserialized access to the time counter
>on sparc64.
As far as I know that timecounter is a single register so there is no
need for serialization. It's only when the hardware requires you
to access multiple times/registers that you need to put locks around it.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
More information about the freebsd-sparc64
mailing list