Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756792AbcCWUto (ORCPT ); Wed, 23 Mar 2016 16:49:44 -0400 Received: from g9t5008.houston.hp.com ([15.240.92.66]:58092 "EHLO g9t5008.houston.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756560AbcCWUtn (ORCPT ); Wed, 23 Mar 2016 16:49:43 -0400 From: Toshi Kani To: mingo@kernel.org, bp@suse.de, hpa@zytor.com, tglx@linutronix.de Cc: mcgrof@suse.com, jgross@suse.com, paul.gortmaker@windriver.com, konrad.wilk@oracle.com, elliott@hpe.com, x86@kernel.org, xen-devel@lists.xenproject.org, linux-kernel@vger.kernel.org Subject: [PATCH v3 0/7] Enhance PAT init to fix Xorg crashes Date: Wed, 23 Mar 2016 15:41:56 -0600 Message-Id: <1458769323-24491-1-git-send-email-toshi.kani@hpe.com> X-Mailer: git-send-email 2.5.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2637 Lines: 64 A Xorg failure on qemu32 was reported as a regression [1] caused by 'commit 9cd25aac1f44 ("x86/mm/pat: Emulate PAT when it is disabled")'. This patch-set fixes the regression. Negative effects of this regression were two failures [2] in Xorg on QEMU with QEMU CPU model "qemu32" (-cpu qemu32), which were triggered by the fact that its virtual CPU does not support MTRR. #1. copy_process() failed in the check in reserve_pfn_range() #2. error path in copy_process() then hit WARN_ON_ONCE in untrack_pfn(). These negative effects are caused by two separate bugs, but they can be addressed in separate patches. This patch-set addresses the root cause, a long-standing PAT initialization issue. Please see the changelog in patch 4/7 for the details of the issue. - Patch 1-2 make necessary enhancement to PAT for the fix without breaking Xen. - Patch 3 is cleanup. - Patch 4 fixes the regression. - Patch 5 fixes an MTRR issue related with PAT init. - Patch 6 removes PAT init code from Xen. - Patch 7 adds PAT init to documentation. [1]: https://lkml.org/lkml/2016/3/3/828 [2]: https://lkml.org/lkml/2016/3/4/775 I'd appreciate if someone can test this patch-set on Xen to verify that there is no change in "x86/PAT: Configuration [0-7] .." message in dmesg. --- v3: - Change a new func name to init_cache_modes(). (Borislav Petkov) - Add check with __pat_enabled, and use WARN_ONCE() for a bug check in pat_disable(). (Borislav Petkov) - Update changelog, comments, and doc per review. (Borislav Petkov) v2: - Divide patch-set into a single change. (Borislav Petkov) - Xen's case must be handled properly. (Luis R. Rodriguez) - Change changelog and title to describe the issue. (Ingo Molnar) - Update an error message. (Robert Elliott, Borislav Petkov) --- Toshi Kani (7): 1/7 x86/mm/pat: Add support of non-default PAT MSR setting 2/7 x86/mm/pat: Add pat_disable() interface 3/7 x86/mm/pat: Replace cpu_has_pat with boot_cpu_has 4/7 x86/mtrr: Fix Xorg crashes in Qemu sessions 5/7 x86/mtrr: Fix PAT init handling when MTRR is disabled 6/7 x86/xen,pat: Remove PAT table init code from Xen 7/7 x86/pat: Document PAT initialization --- Documentation/x86/pat.txt | 32 ++++++++++++++ arch/x86/include/asm/mtrr.h | 6 ++- arch/x86/include/asm/pat.h | 2 +- arch/x86/kernel/cpu/mtrr/generic.c | 24 +++++----- arch/x86/kernel/cpu/mtrr/main.c | 13 +++++- arch/x86/kernel/cpu/mtrr/mtrr.h | 1 + arch/x86/mm/pat.c | 90 ++++++++++++++++++++++++++++---------- arch/x86/xen/enlighten.c | 9 ---- 8 files changed, 132 insertions(+), 45 deletions(-)