Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750759AbWBUXXG (ORCPT ); Tue, 21 Feb 2006 18:23:06 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750948AbWBUXXG (ORCPT ); Tue, 21 Feb 2006 18:23:06 -0500 Received: from liaag2af.mx.compuserve.com ([149.174.40.157]:51924 "EHLO liaag2af.mx.compuserve.com") by vger.kernel.org with ESMTP id S1750759AbWBUXXF (ORCPT ); Tue, 21 Feb 2006 18:23:05 -0500 Date: Tue, 21 Feb 2006 18:19:58 -0500 From: Chuck Ebbert <76306.1226@compuserve.com> Subject: Re: 2.6.16-rc4-mm1 To: Andrew Morton Cc: linux-kernel@vger.kernel.org Message-ID: <200602211822_MC3-1-B8F6-689@compuserve.com> MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3136 Lines: 95 In-Reply-To: <20060221134139.11b8668b.akpm@osdl.org> On Tue, 21 Feb 2006 at 13:41:39 -0800, Andrew Morton wrote: > None of these are must-have fixes, are they? I had them queued for 2.6.17. Well i386-allow-disabling-x86_feature_sep-at-boot.patch is for me. I had to backport it to -rc4 for testing differences between sysenter and int80 syscall paths. I could live without it (I once forgot to apply it and wondered why it wasn't working), but anyway here's the backport: From: Chuck Ebbert <76306.1226@compuserve.com> Allow the x86 "sep" feature to be disabled at bootup. This forces use of the int80 vsyscall. Mainly for testing or benchmarking the int80 vsyscall code. Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com> Signed-off-by: Andrew Morton --- Documentation/kernel-parameters.txt | 6 +++++- arch/i386/kernel/cpu/common.c | 15 ++++++++++++++- 2 files changed, 19 insertions(+), 2 deletions(-) --- 2.6.16-rc3-nb.orig/Documentation/kernel-parameters.txt +++ 2.6.16-rc3-nb/Documentation/kernel-parameters.txt @@ -1002,7 +1002,9 @@ running once the system is up. noexec=on: enable non-executable mappings (default) noexec=off: disable nn-executable mappings - nofxsr [BUGS=IA-32] + nofxsr [BUGS=IA-32] Disables x86 floating point extended + register save and restore. The kernel will only save + legacy floating-point registers on task switch. nohlt [BUGS=ARM] @@ -1045,6 +1047,8 @@ running once the system is up. nosbagart [IA-64] + nosep [BUGS=IA-32] Disables x86 SYSENTER/SYSEXIT support. + nosmp [SMP] Tells an SMP kernel to act as a UP kernel. nosync [HW,M68K] Disables sync negotiation for all devices. --- 2.6.16-rc3-nb.orig/arch/i386/kernel/cpu/common.c +++ 2.6.16-rc3-nb/arch/i386/kernel/cpu/common.c @@ -22,8 +22,9 @@ DEFINE_PER_CPU(unsigned char, cpu_16bit_ EXPORT_PER_CPU_SYMBOL(cpu_16bit_stack); static int cachesize_override __devinitdata = -1; -static int disable_x86_fxsr __devinitdata = 0; +static int disable_x86_fxsr __devinitdata; static int disable_x86_serial_nr __devinitdata = 1; +static int disable_x86_sep __devinitdata; struct cpu_dev * cpu_devs[X86_VENDOR_NUM] = {}; @@ -183,6 +184,14 @@ static int __init x86_fxsr_setup(char * __setup("nofxsr", x86_fxsr_setup); +static int __init x86_sep_setup(char * s) +{ + disable_x86_sep = 1; + return 1; +} +__setup("nosep", x86_sep_setup); + + /* Standard macro to see if a specific flag is changeable */ static inline int flag_is_changeable_p(u32 flag) { @@ -401,6 +410,10 @@ void __devinit identify_cpu(struct cpuin clear_bit(X86_FEATURE_XMM, c->x86_capability); } + /* SEP disabled? */ + if (disable_x86_sep) + clear_bit(X86_FEATURE_SEP, c->x86_capability); + if (disable_pse) clear_bit(X86_FEATURE_PSE, c->x86_capability); -- Chuck "Equations are the Devil's sentences." --Stephen Colbert - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/