From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8DFD3C25AEF for ; Thu, 22 Jul 2021 20:53:20 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 2C90A60EB6 for ; Thu, 22 Jul 2021 20:53:20 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org 2C90A60EB6 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=kvack.org Received: by kanga.kvack.org (Postfix) id A4B158D0007; Thu, 22 Jul 2021 16:53:03 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 937B2900002; Thu, 22 Jul 2021 16:53:03 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 53E2F8E0001; Thu, 22 Jul 2021 16:53:03 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0179.hostedemail.com [216.40.44.179]) by kanga.kvack.org (Postfix) with ESMTP id 1051B8D0008 for ; Thu, 22 Jul 2021 16:53:03 -0400 (EDT) Received: from smtpin31.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay03.hostedemail.com (Postfix) with ESMTP id B02C8824999B for ; Thu, 22 Jul 2021 20:53:02 +0000 (UTC) X-FDA: 78391423404.31.0C8CE18 Received: from mga04.intel.com (mga04.intel.com [192.55.52.120]) by imf02.hostedemail.com (Postfix) with ESMTP id 3791A70024D9 for ; Thu, 22 Jul 2021 20:53:02 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10053"; a="209845067" X-IronPort-AV: E=Sophos;i="5.84,262,1620716400"; d="scan'208";a="209845067" Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga104.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 Jul 2021 13:52:59 -0700 X-IronPort-AV: E=Sophos;i="5.84,262,1620716400"; d="scan'208";a="502035442" Received: from yyu32-desk.sc.intel.com ([143.183.136.146]) by fmsmga003-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 Jul 2021 13:52:59 -0700 From: Yu-cheng Yu To: x86@kernel.org, "H. Peter Anvin" , Thomas Gleixner , Ingo Molnar , linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org, linux-mm@kvack.org, linux-arch@vger.kernel.org, linux-api@vger.kernel.org, Arnd Bergmann , Andy Lutomirski , Balbir Singh , Borislav Petkov , Cyrill Gorcunov , Dave Hansen , Eugene Syromiatnikov , Florian Weimer , "H.J. Lu" , Jann Horn , Jonathan Corbet , Kees Cook , Mike Kravetz , Nadav Amit , Oleg Nesterov , Pavel Machek , Peter Zijlstra , Randy Dunlap , "Ravi V. Shankar" , Vedvyas Shanbhogue , Dave Martin , Weijiang Yang , Pengfei Xu , Haitao Huang , Rick P Edgecombe Cc: Yu-cheng Yu , "Kirill A . Shutemov" Subject: [PATCH v28 09/32] x86/mm: Introduce _PAGE_COW Date: Thu, 22 Jul 2021 13:51:56 -0700 Message-Id: <20210722205219.7934-10-yu-cheng.yu@intel.com> X-Mailer: git-send-email 2.21.0 In-Reply-To: <20210722205219.7934-1-yu-cheng.yu@intel.com> References: <20210722205219.7934-1-yu-cheng.yu@intel.com> MIME-Version: 1.0 Authentication-Results: imf02.hostedemail.com; dkim=none; spf=none (imf02.hostedemail.com: domain of yu-cheng.yu@intel.com has no SPF policy when checking 192.55.52.120) smtp.mailfrom=yu-cheng.yu@intel.com; dmarc=fail reason="No valid SPF, No valid DKIM" header.from=intel.com (policy=none) X-Rspamd-Server: rspam02 X-Stat-Signature: wfkz45mzx5uitbfwezhem5jq8o1xk5ir X-Rspamd-Queue-Id: 3791A70024D9 X-HE-Tag: 1626987182-80665 Content-Transfer-Encoding: quoted-printable X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: There is essentially no room left in the x86 hardware PTEs on some OSes (not Linux). That left the hardware architects looking for a way to represent a new memory type (shadow stack) within the existing bits. They chose to repurpose a lightly-used state: Write=3D0, Dirty=3D1. The reason it's lightly used is that Dirty=3D1 is normally set by hardwar= e and cannot normally be set by hardware on a Write=3D0 PTE. Software must normally be involved to create one of these PTEs, so software can simply opt to not create them. In places where Linux normally creates Write=3D0, Dirty=3D1, it can use t= he software-defined _PAGE_COW in place of the hardware _PAGE_DIRTY. In othe= r words, whenever Linux needs to create Write=3D0, Dirty=3D1, it instead cr= eates Write=3D0, Cow=3D1, except for shadow stack, which is Write=3D0, Dirty=3D= 1. This clearly separates shadow stack from other data, and results in the following: (a) A modified, copy-on-write (COW) page: (Write=3D0, Cow=3D1) (b) A R/O page that has been COW'ed: (Write=3D0, Cow=3D1) The user page is in a R/O VMA, and get_user_pages() needs a writable copy. The page fault handler creates a copy of the page and sets the new copy's PTE as Write=3D0 and Cow=3D1. (c) A shadow stack PTE: (Write=3D0, Dirty=3D1) (d) A shared shadow stack PTE: (Write=3D0, Cow=3D1) When a shadow stack page is being shared among processes (this happen= s at fork()), its PTE is made Dirty=3D0, so the next shadow stack acces= s causes a fault, and the page is duplicated and Dirty=3D1 is set again= . This is the COW equivalent for shadow stack pages, even though it's copy-on-access rather than copy-on-write. (e) A page where the processor observed a Write=3D1 PTE, started a write,= set Dirty=3D1, but then observed a Write=3D0 PTE. That's possible today,= but will not happen on processors that support shadow stack. Define _PAGE_COW and update pte_*() helpers and apply the same changes to pmd and pud. After this, there are six free bits left in the 64-bit PTE, and no more free bits in the 32-bit PTE (except for PAE) and Shadow Stack is not implemented for the 32-bit kernel. Signed-off-by: Yu-cheng Yu Reviewed-by: Kirill A. Shutemov --- arch/x86/include/asm/pgtable.h | 196 ++++++++++++++++++++++++--- arch/x86/include/asm/pgtable_types.h | 42 +++++- 2 files changed, 217 insertions(+), 21 deletions(-) diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtabl= e.h index 0ddeda0bc0c0..9f1ba76ed79a 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -121,9 +121,20 @@ extern pmdval_t early_pmd_flags; * The following only work if pte_present() is true. * Undefined behaviour if not.. */ -static inline int pte_dirty(pte_t pte) +static inline bool pte_dirty(pte_t pte) { - return pte_flags(pte) & _PAGE_DIRTY; + /* + * A dirty PTE has Dirty=3D1 or Cow=3D1. + */ + return pte_flags(pte) & _PAGE_DIRTY_BITS; +} + +static inline bool pte_shstk(pte_t pte) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return false; + + return (pte_flags(pte) & (_PAGE_RW | _PAGE_DIRTY)) =3D=3D _PAGE_DIRTY; } =20 static inline int pte_young(pte_t pte) @@ -131,9 +142,20 @@ static inline int pte_young(pte_t pte) return pte_flags(pte) & _PAGE_ACCESSED; } =20 -static inline int pmd_dirty(pmd_t pmd) +static inline bool pmd_dirty(pmd_t pmd) +{ + /* + * A dirty PMD has Dirty=3D1 or Cow=3D1. + */ + return pmd_flags(pmd) & _PAGE_DIRTY_BITS; +} + +static inline bool pmd_shstk(pmd_t pmd) { - return pmd_flags(pmd) & _PAGE_DIRTY; + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return false; + + return (pmd_flags(pmd) & (_PAGE_RW | _PAGE_DIRTY)) =3D=3D _PAGE_DIRTY; } =20 static inline int pmd_young(pmd_t pmd) @@ -141,9 +163,12 @@ static inline int pmd_young(pmd_t pmd) return pmd_flags(pmd) & _PAGE_ACCESSED; } =20 -static inline int pud_dirty(pud_t pud) +static inline bool pud_dirty(pud_t pud) { - return pud_flags(pud) & _PAGE_DIRTY; + /* + * A dirty PUD has Dirty=3D1 or Cow=3D1. + */ + return pud_flags(pud) & _PAGE_DIRTY_BITS; } =20 static inline int pud_young(pud_t pud) @@ -153,13 +178,23 @@ static inline int pud_young(pud_t pud) =20 static inline int pte_write(pte_t pte) { - return pte_flags(pte) & _PAGE_RW; + /* + * Shadow stack pages are always writable - but not by normal + * instructions, and only by shadow stack operations. Therefore, + * the W=3D0,D=3D1 test with pte_shstk(). + */ + return (pte_flags(pte) & _PAGE_RW) || pte_shstk(pte); } =20 #define pmd_write pmd_write static inline int pmd_write(pmd_t pmd) { - return pmd_flags(pmd) & _PAGE_RW; + /* + * Shadow stack pages are always writable - but not by normal + * instructions, and only by shadow stack operations. Therefore, + * the W=3D0,D=3D1 test with pmd_shstk(). + */ + return (pmd_flags(pmd) & _PAGE_RW) || pmd_shstk(pmd); } =20 #define pud_write pud_write @@ -297,6 +332,24 @@ static inline pte_t pte_clear_flags(pte_t pte, pteva= l_t clear) return native_make_pte(v & ~clear); } =20 +static inline pte_t pte_mkcow(pte_t pte) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return pte; + + pte =3D pte_clear_flags(pte, _PAGE_DIRTY); + return pte_set_flags(pte, _PAGE_COW); +} + +static inline pte_t pte_clear_cow(pte_t pte) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return pte; + + pte =3D pte_set_flags(pte, _PAGE_DIRTY); + return pte_clear_flags(pte, _PAGE_COW); +} + #ifdef CONFIG_HAVE_ARCH_USERFAULTFD_WP static inline int pte_uffd_wp(pte_t pte) { @@ -316,7 +369,7 @@ static inline pte_t pte_clear_uffd_wp(pte_t pte) =20 static inline pte_t pte_mkclean(pte_t pte) { - return pte_clear_flags(pte, _PAGE_DIRTY); + return pte_clear_flags(pte, _PAGE_DIRTY_BITS); } =20 static inline pte_t pte_mkold(pte_t pte) @@ -326,7 +379,16 @@ static inline pte_t pte_mkold(pte_t pte) =20 static inline pte_t pte_wrprotect(pte_t pte) { - return pte_clear_flags(pte, _PAGE_RW); + pte =3D pte_clear_flags(pte, _PAGE_RW); + + /* + * Blindly clearing _PAGE_RW might accidentally create + * a shadow stack PTE (RW=3D0, Dirty=3D1). Move the hardware + * dirty value to the software bit. + */ + if (pte_dirty(pte)) + pte =3D pte_mkcow(pte); + return pte; } =20 static inline pte_t pte_mkexec(pte_t pte) @@ -336,7 +398,18 @@ static inline pte_t pte_mkexec(pte_t pte) =20 static inline pte_t pte_mkdirty(pte_t pte) { - return pte_set_flags(pte, _PAGE_DIRTY | _PAGE_SOFT_DIRTY); + pteval_t dirty =3D _PAGE_DIRTY; + + /* Avoid creating (HW)Dirty=3D1, Write=3D0 PTEs */ + if (cpu_feature_enabled(X86_FEATURE_SHSTK) && !pte_write(pte)) + dirty =3D _PAGE_COW; + + return pte_set_flags(pte, dirty | _PAGE_SOFT_DIRTY); +} + +static inline pte_t pte_mkwrite_shstk(pte_t pte) +{ + return pte_clear_cow(pte); } =20 static inline pte_t pte_mkyoung(pte_t pte) @@ -346,7 +419,12 @@ static inline pte_t pte_mkyoung(pte_t pte) =20 static inline pte_t pte_mkwrite(pte_t pte) { - return pte_set_flags(pte, _PAGE_RW); + pte =3D pte_set_flags(pte, _PAGE_RW); + + if (pte_dirty(pte)) + pte =3D pte_clear_cow(pte); + + return pte; } =20 static inline pte_t pte_mkhuge(pte_t pte) @@ -393,6 +471,24 @@ static inline pmd_t pmd_clear_flags(pmd_t pmd, pmdva= l_t clear) return native_make_pmd(v & ~clear); } =20 +static inline pmd_t pmd_mkcow(pmd_t pmd) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return pmd; + + pmd =3D pmd_clear_flags(pmd, _PAGE_DIRTY); + return pmd_set_flags(pmd, _PAGE_COW); +} + +static inline pmd_t pmd_clear_cow(pmd_t pmd) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return pmd; + + pmd =3D pmd_set_flags(pmd, _PAGE_DIRTY); + return pmd_clear_flags(pmd, _PAGE_COW); +} + #ifdef CONFIG_HAVE_ARCH_USERFAULTFD_WP static inline int pmd_uffd_wp(pmd_t pmd) { @@ -417,17 +513,36 @@ static inline pmd_t pmd_mkold(pmd_t pmd) =20 static inline pmd_t pmd_mkclean(pmd_t pmd) { - return pmd_clear_flags(pmd, _PAGE_DIRTY); + return pmd_clear_flags(pmd, _PAGE_DIRTY_BITS); } =20 static inline pmd_t pmd_wrprotect(pmd_t pmd) { - return pmd_clear_flags(pmd, _PAGE_RW); + pmd =3D pmd_clear_flags(pmd, _PAGE_RW); + /* + * Blindly clearing _PAGE_RW might accidentally create + * a shadow stack PMD (RW=3D0, Dirty=3D1). Move the hardware + * dirty value to the software bit. + */ + if (pmd_dirty(pmd)) + pmd =3D pmd_mkcow(pmd); + return pmd; } =20 static inline pmd_t pmd_mkdirty(pmd_t pmd) { - return pmd_set_flags(pmd, _PAGE_DIRTY | _PAGE_SOFT_DIRTY); + pmdval_t dirty =3D _PAGE_DIRTY; + + /* Avoid creating (HW)Dirty=3D1, Write=3D0 PMDs */ + if (cpu_feature_enabled(X86_FEATURE_SHSTK) && !pmd_write(pmd)) + dirty =3D _PAGE_COW; + + return pmd_set_flags(pmd, dirty | _PAGE_SOFT_DIRTY); +} + +static inline pmd_t pmd_mkwrite_shstk(pmd_t pmd) +{ + return pmd_clear_cow(pmd); } =20 static inline pmd_t pmd_mkdevmap(pmd_t pmd) @@ -447,7 +562,11 @@ static inline pmd_t pmd_mkyoung(pmd_t pmd) =20 static inline pmd_t pmd_mkwrite(pmd_t pmd) { - return pmd_set_flags(pmd, _PAGE_RW); + pmd =3D pmd_set_flags(pmd, _PAGE_RW); + + if (pmd_dirty(pmd)) + pmd =3D pmd_clear_cow(pmd); + return pmd; } =20 static inline pud_t pud_set_flags(pud_t pud, pudval_t set) @@ -464,6 +583,24 @@ static inline pud_t pud_clear_flags(pud_t pud, pudva= l_t clear) return native_make_pud(v & ~clear); } =20 +static inline pud_t pud_mkcow(pud_t pud) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return pud; + + pud =3D pud_clear_flags(pud, _PAGE_DIRTY); + return pud_set_flags(pud, _PAGE_COW); +} + +static inline pud_t pud_clear_cow(pud_t pud) +{ + if (!cpu_feature_enabled(X86_FEATURE_SHSTK)) + return pud; + + pud =3D pud_set_flags(pud, _PAGE_DIRTY); + return pud_clear_flags(pud, _PAGE_COW); +} + static inline pud_t pud_mkold(pud_t pud) { return pud_clear_flags(pud, _PAGE_ACCESSED); @@ -471,17 +608,32 @@ static inline pud_t pud_mkold(pud_t pud) =20 static inline pud_t pud_mkclean(pud_t pud) { - return pud_clear_flags(pud, _PAGE_DIRTY); + return pud_clear_flags(pud, _PAGE_DIRTY_BITS); } =20 static inline pud_t pud_wrprotect(pud_t pud) { - return pud_clear_flags(pud, _PAGE_RW); + pud =3D pud_clear_flags(pud, _PAGE_RW); + + /* + * Blindly clearing _PAGE_RW might accidentally create + * a shadow stack PUD (RW=3D0, Dirty=3D1). Move the hardware + * dirty value to the software bit. + */ + if (pud_dirty(pud)) + pud =3D pud_mkcow(pud); + return pud; } =20 static inline pud_t pud_mkdirty(pud_t pud) { - return pud_set_flags(pud, _PAGE_DIRTY | _PAGE_SOFT_DIRTY); + pudval_t dirty =3D _PAGE_DIRTY; + + /* Avoid creating (HW)Dirty=3D1, Write=3D0 PUDs */ + if (cpu_feature_enabled(X86_FEATURE_SHSTK) && !pud_write(pud)) + dirty =3D _PAGE_COW; + + return pud_set_flags(pud, dirty | _PAGE_SOFT_DIRTY); } =20 static inline pud_t pud_mkdevmap(pud_t pud) @@ -501,7 +653,11 @@ static inline pud_t pud_mkyoung(pud_t pud) =20 static inline pud_t pud_mkwrite(pud_t pud) { - return pud_set_flags(pud, _PAGE_RW); + pud =3D pud_set_flags(pud, _PAGE_RW); + + if (pud_dirty(pud)) + pud =3D pud_clear_cow(pud); + return pud; } =20 #ifdef CONFIG_HAVE_ARCH_SOFT_DIRTY diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/= pgtable_types.h index 3781a79b6388..1bfab70ff9ac 100644 --- a/arch/x86/include/asm/pgtable_types.h +++ b/arch/x86/include/asm/pgtable_types.h @@ -21,7 +21,8 @@ #define _PAGE_BIT_SOFTW2 10 /* " */ #define _PAGE_BIT_SOFTW3 11 /* " */ #define _PAGE_BIT_PAT_LARGE 12 /* On 2MB or 1GB pages */ -#define _PAGE_BIT_SOFTW4 58 /* available for programmer */ +#define _PAGE_BIT_SOFTW4 57 /* available for programmer */ +#define _PAGE_BIT_SOFTW5 58 /* available for programmer */ #define _PAGE_BIT_PKEY_BIT0 59 /* Protection Keys, bit 1/4 */ #define _PAGE_BIT_PKEY_BIT1 60 /* Protection Keys, bit 2/4 */ #define _PAGE_BIT_PKEY_BIT2 61 /* Protection Keys, bit 3/4 */ @@ -34,6 +35,15 @@ #define _PAGE_BIT_SOFT_DIRTY _PAGE_BIT_SOFTW3 /* software dirty tracking= */ #define _PAGE_BIT_DEVMAP _PAGE_BIT_SOFTW4 =20 +/* + * Indicates a copy-on-write page. + */ +#ifdef CONFIG_X86_SHADOW_STACK +#define _PAGE_BIT_COW _PAGE_BIT_SOFTW5 /* copy-on-write */ +#else +#define _PAGE_BIT_COW 0 +#endif + /* If _PAGE_BIT_PRESENT is clear, we use these: */ /* - if the user mapped it with PROT_NONE; pte_present gives true */ #define _PAGE_BIT_PROTNONE _PAGE_BIT_GLOBAL @@ -115,6 +125,36 @@ #define _PAGE_DEVMAP (_AT(pteval_t, 0)) #endif =20 +/* + * The hardware requires shadow stack to be read-only and Dirty. + * _PAGE_COW is a software-only bit used to separate copy-on-write PTEs + * from shadow stack PTEs: + * (a) A modified, copy-on-write (COW) page: (Write=3D0, Cow=3D1) + * (b) A R/O page that has been COW'ed: (Write=3D0, Cow=3D1) + * The user page is in a R/O VMA, and get_user_pages() needs a + * writable copy. The page fault handler creates a copy of the page + * and sets the new copy's PTE as Write=3D0, Cow=3D1. + * (c) A shadow stack PTE: (Write=3D0, Dirty=3D1) + * (d) A shared (copy-on-access) shadow stack PTE: (Write=3D0, Cow=3D1) + * When a shadow stack page is being shared among processes (this + * happens at fork()), its PTE is cleared of _PAGE_DIRTY, so the nex= t + * shadow stack access causes a fault, and the page is duplicated an= d + * _PAGE_DIRTY is set again. This is the COW equivalent for shadow + * stack pages, even though it's copy-on-access rather than + * copy-on-write. + * (e) A page where the processor observed a Write=3D1 PTE, started a wr= ite, + * set Dirty=3D1, but then observed a Write=3D0 PTE (changed by anot= her + * thread). That's possible today, but will not happen on processor= s + * that support shadow stack. + */ +#ifdef CONFIG_X86_SHADOW_STACK +#define _PAGE_COW (_AT(pteval_t, 1) << _PAGE_BIT_COW) +#else +#define _PAGE_COW (_AT(pteval_t, 0)) +#endif + +#define _PAGE_DIRTY_BITS (_PAGE_DIRTY | _PAGE_COW) + #define _PAGE_PROTNONE (_AT(pteval_t, 1) << _PAGE_BIT_PROTNONE) =20 /* --=20 2.21.0