On Sun, 14 Oct 2007 23:23:16 -0700 Andrew Morton <[email protected]> wrote: > > Also, it's bad that the operator has to provide some special boot command-line > option to make the machine work properly. Please consider this to be a > bug. Has it always needed apm=power-off? It has always been that way because the behaviour of APM on SMP is not defined. We discovered early that it will work (at least for powering off) for most SMP machines as long as the APM calls are done on CPU 0. However on an SMP machine that does not cope, very bad things happen - thus the operator has to explictly enable the power off behaviour on an SMP box. (On UP power off is enabled by default.) -- Cheers, Stephen Rothwell [email protected] http://www.canb.auug.org.au/~sfr/
Attachment:
pgpzBRwF4yCY3.pgp
Description: PGP signature
- Follow-Ups:
- Re: apm system does not power off anymore
- From: Andrew Morton <[email protected]>
- Re: apm system does not power off anymore
- References:
- apm system does not power off anymore
- From: Olaf Hering <[email protected]>
- Re: apm system does not power off anymore
- From: Andrew Morton <[email protected]>
- apm system does not power off anymore
- Prev by Date: Re: [RFC] [Patch] calgary iommu: Use the first kernel's tce tables in kdump
- Next by Date: Re: [PATCH] hptiop: avoid buffer overflow when returning sense data
- Previous by thread: Re: apm system does not power off anymore
- Next by thread: Re: apm system does not power off anymore
- Index(es):
 |