13 messages in org.freebsd.freebsd-smpRe: Dog Sloooow SMP| From | Sent On | Attachments |
|---|---|---|
| John Capo | Nov 3, 1998 3:35 pm | |
| John Capo | Nov 4, 1998 11:20 am | |
| Mike Smith | Nov 4, 1998 11:43 pm | |
| Mike Smith | Nov 7, 1998 6:26 pm | |
| Bruce Evans | Nov 7, 1998 6:43 pm | |
| Peter Wemm | Nov 7, 1998 7:25 pm | |
| Mike Smith | Nov 7, 1998 7:30 pm | |
| Bruce Evans | Nov 9, 1998 12:48 am | |
| Narvi | Nov 9, 1998 1:36 am | |
| Bruce Evans | Nov 9, 1998 2:22 am | |
| Peter Wemm | Nov 9, 1998 3:21 am | |
| Bruce Evans | Nov 10, 1998 1:08 am | |
| Mike Smith | Nov 10, 1998 1:24 am |
| Subject: | Re: Dog Sloooow SMP | |
|---|---|---|
| From: | Bruce Evans (bd...@zeta.org.au) | |
| Date: | Nov 9, 1998 12:48:11 am | |
| List: | org.freebsd.freebsd-smp | |
It's only OK for MII's because of various `#if 0's and `#ifdef SMP's that prevent non-OK code from running on MII's.
I think it should be CPU specific, not cpu class specific. The model-specific-registers are very specific to the Intel family. I'd be a lot happier if it was 'if (cpu == CPU_686 || cpu == CPU_PII) ...' Of course, feature tests would be better. 'if (cpu_features & CF_PPRO_MSR)...' The problem is that there is a 'cpu_feature' already for the CPUID. We need more general flags than what Intel choose to tell us.
FreeBSD should use its own bitmap of capabilities and not test the Intel flags except once to translate them. 32 general flags might even be enough.
Bruce
To Unsubscribe: send mail to majo...@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message