Date: Mon, 05 Oct 2009 00:49:27 +0900 From: "Michio \"Karl\" Jinbo" <michio.jinbo@gmail.com> To: src-committers@freebsd.org Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org Subject: Re: svn commit: r197744 - in stable/8/sys: . amd64/amd64 amd64/include/xen cddl/contrib/opensolaris contrib/dev/acpica contrib/pf dev/xen/xenpci i386/i386 Message-ID: <20091005004923.4086.22FF24F1@gmail.com> In-Reply-To: <200910041220.n94CKxTH073639@svn.freebsd.org> References: <200910041220.n94CKxTH073639@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 4 Oct 2009 12:20:59 +0000 (UTC)
Konstantin Belousov <kib@FreeBSD.org> wrote:
> Log:
> MFC r197663:
> As a workaround, for Intel CPUs, do not use CLFLUSH in
> pmap_invalidate_cache_range() when self-snoop is apparently not reported
> in cpu features.
>
> Approved by: re (bz, kensmith)
I was tested r197663/r197744, but kernel panic again on Citrix Xen Server.
using 8.0-RC1 install cd, results are
1. INTEL SU9400+HYPER-V(Windows2008 R2) -> boot OK.
2. AMD Athlon X2 TK-55+HYPER-V(Windows2008 R2) -> boot NG.
3. AMD PhenomII 940BK+Citrix Xen Server -> boot NG.
I think INTEL CPUs are no problem, but AMD CPUs appear the problem. So I tested
the following patch, kernel boot was successed on recent 9-CURRENT and environment 3.
sorry, poor English.
--- sys/i386/i386/initcpu.c.original 2009-10-01 21:52:48.000000000 +0900
+++ sys/i386/i386/initcpu.c 2009-10-05 08:29:45.000000000 +0900
@@ -721,7 +721,7 @@
* XXXKIB: (temporary) hack to work around traps generated when
* CLFLUSHing APIC registers window.
*/
- if (cpu_vendor_id == CPU_VENDOR_INTEL && !(cpu_feature & CPUID_SS))
+ if (cpu_vendor_id == CPU_VENDOR_AMD && !(cpu_feature & CPUID_SS))
cpu_feature &= ~CPUID_CLFSH;
#if defined(PC98) && !defined(CPU_UPGRADE_HW_CACHE)
--
Michio Jinbo <michio.jinbo@gmail.com>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20091005004923.4086.22FF24F1>