From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756110AbbHYUM3 (ORCPT ); Tue, 25 Aug 2015 16:12:29 -0400 Received: from mga01.intel.com ([192.55.52.88]:11064 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756049AbbHYUMH (ORCPT ); Tue, 25 Aug 2015 16:12:07 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.17,747,1437462000"; d="scan'208";a="775534040" Subject: [PATCH 11/11] x86, fpu: check CPU-provided sizes against struct declarations To: dave@sr71.net Cc: dave.hansen@linux.intel.com, mingo@redhat.com, x86@kernel.org, bp@alien8.de, fenghua.yu@intel.com, tim.c.chen@linux.intel.com, linux-kernel@vger.kernel.org From: Dave Hansen Date: Tue, 25 Aug 2015 13:12:06 -0700 References: <20150825201201.CF766C1B@viggo.jf.intel.com> In-Reply-To: <20150825201201.CF766C1B@viggo.jf.intel.com> Message-Id: <20150825201206.93136028@viggo.jf.intel.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Dave Hansen We now have C structures defined for each of the XSAVE state components that we support. This patch adds checks during our verification pass to ensure that the CPU-provided data enumerated in CPUID leaves matches our C structures. If not, we warn and dump all the XSAVE CPUID leaves. Note: this *actually* found an inconsistency with the MPX 'bndcsr' state. The hardware pads it out differently from our C structures. This patch caught it and warned. Signed-off-by: Dave Hansen Cc: Ingo Molnar Cc: x86@kernel.org Cc: Borislav Petkov Cc: Fenghua Yu Cc: Tim Chen Cc: linux-kernel@vger.kernel.org --- b/appendme | 6 +++ b/arch/x86/include/asm/fpu/types.h | 1 b/arch/x86/kernel/fpu/xstate.c | 62 +++++++++++++++++++++++++++++++++++-- 3 files changed, 67 insertions(+), 2 deletions(-) diff -puN /dev/null appendme --- /dev/null 2015-07-13 14:24:11.435656502 -0700 +++ b/appendme 2015-08-25 12:50:01.857661314 -0700 @@ -0,0 +1,6 @@ +Cc: Ingo Molnar +Cc: x86@kernel.org +Cc: Borislav Petkov +Cc: Fenghua Yu +Cc: Tim Chen +Cc: linux-kernel@vger.kernel.org diff -puN arch/x86/include/asm/fpu/types.h~x86-fpu-check-against-struct-declarations arch/x86/include/asm/fpu/types.h --- a/arch/x86/include/asm/fpu/types.h~x86-fpu-check-against-struct-declarations 2015-08-25 12:50:01.853661133 -0700 +++ b/arch/x86/include/asm/fpu/types.h 2015-08-25 12:50:01.858661359 -0700 @@ -108,6 +108,7 @@ enum xfeature_nr { XFEATURE_NR_OPMASK, XFEATURE_NR_ZMM_Hi256, XFEATURE_NR_Hi16_ZMM, + XFEATURE_NR_PT_UNIMPLEMENTED_SO_FAR, XFEATURES_NR_MAX, }; diff -puN arch/x86/kernel/fpu/xstate.c~x86-fpu-check-against-struct-declarations arch/x86/kernel/fpu/xstate.c --- a/arch/x86/kernel/fpu/xstate.c~x86-fpu-check-against-struct-declarations 2015-08-25 12:50:01.854661178 -0700 +++ b/arch/x86/kernel/fpu/xstate.c 2015-08-25 12:50:01.858661359 -0700 @@ -182,7 +182,8 @@ static int xfeature_nr_enabled(enum xfea static void __init setup_xstate_features(void) { u32 eax, ebx, ecx, edx, i; - unsigned int last_good_offset = -1; + /* start at the beginnning of the "extended state" */ + unsigned int last_good_offset = offsetof(struct xregs_state, __reserved); for (i = FIRST_EXTENDED_XFEATURE_NR; i < XFEATURES_NR_MAX; i++) { if (!xfeature_nr_enabled(i)) @@ -196,7 +197,8 @@ static void __init setup_xstate_features * highest-numbered xstate feature has the * highest offset in the buffer. Ensure it does. */ - WARN_ON(last_good_offset > xstate_offsets[i]); + WARN_ONCE(last_good_offset > xstate_offsets[i], + "x86/fpu: misordered xstate at %d\n", last_good_offset); last_good_offset = xstate_offsets[i]; printk(KERN_INFO "x86/fpu: xstate_offset[%d]: %04x, xstate_sizes[%d]: %04x\n", i, ebx, i, eax); @@ -407,6 +409,52 @@ static void __xstate_dump_leaves(void) } \ } while (0) + +#define XCHECK_SZ(sz, nr, nr_macro, __struct) do { \ + if ((nr == nr_macro) && \ + WARN_ONCE(sz != sizeof(__struct), \ + "%s: struct is %ld bytes, cpu state %d bytes\n", \ + __stringify(nr_macro), sizeof(__struct), sz)) { \ + __xstate_dump_leaves(); \ + } \ +} while (0) + + +/* + * We have a C struct for each 'xstate'. We need to ensure + * that our software representation matches what the CPU + * tells us about the state's size. + */ +static void check_xstate_against_struct(int nr) +{ + /* + * Ask the CPU for the size of the state. + */ + int sz = xfeature_size(nr); + /* + * Match each CPU state with the corresponding software + * structure. + */ + XCHECK_SZ(sz, nr, XFEATURE_NR_YMM, struct ymmh_struct); + XCHECK_SZ(sz, nr, XFEATURE_NR_BNDREGS, struct mpx_bndreg_state); + XCHECK_SZ(sz, nr, XFEATURE_NR_BNDCSR, struct mpx_bndcsr_state); + XCHECK_SZ(sz, nr, XFEATURE_NR_OPMASK, struct avx_512_opmask_state); + XCHECK_SZ(sz, nr, XFEATURE_NR_ZMM_Hi256, struct avx_512_zmm_uppers_state); + XCHECK_SZ(sz, nr, XFEATURE_NR_Hi16_ZMM, struct avx_512_hi16_state); + + /* + * Make *SURE* to add any feature numbers in below if + * there are "holes" in the xsave state component + * numbers. + */ + if ((nr < XFEATURE_NR_YMM) || + (nr >= XFEATURES_NR_MAX) || + (nr == XFEATURE_NR_PT_UNIMPLEMENTED_SO_FAR)) { + WARN_ONCE(1, "no structure for xstate: %d\n", nr); + XSTATE_WARN_ON(1); + } +} + /* * This essentially double-checks what the cpu told us about * how large the XSAVE buffer needs to be. We are recalculating @@ -420,6 +468,8 @@ static void do_extra_xstate_size_checks( for (i = FIRST_EXTENDED_XFEATURE_NR; i < XFEATURES_NR_MAX; i++) { if (!xfeature_nr_enabled(i)) continue; + + check_xstate_against_struct(i); /* * Supervisor state components can be managed only by * XSAVES, which is compacted-format only. @@ -445,6 +495,14 @@ static void do_extra_xstate_size_checks( paranoid_xstate_size += xfeature_size(i); } XSTATE_WARN_ON(paranoid_xstate_size != xstate_size); + /* + * Basically, make sure that XSTATE_RESERVE has forced + * xregs_state to be large enough. This is not fatal + * because we reserve a *lot* of extra room in the init + * task struct, but we should at least know we got it + * wrong. + */ + XSTATE_WARN_ON(xstate_size > sizeof(struct xregs_state)); } /* _