From patchwork Wed Dec 12 00:03:54 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Edgecombe, Rick P" X-Patchwork-Id: 10725341 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id A101B6C5 for ; Wed, 12 Dec 2018 00:13:00 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 8D86F2B587 for ; Wed, 12 Dec 2018 00:13:00 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 7EE452B591; Wed, 12 Dec 2018 00:13:00 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.2 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_MED autolearn=ham version=3.3.1 Received: from mother.openwall.net (mother.openwall.net [195.42.179.200]) by mail.wl.linuxfoundation.org (Postfix) with SMTP id 480E42B587 for ; Wed, 12 Dec 2018 00:12:59 +0000 (UTC) Received: (qmail 30218 invoked by uid 550); 12 Dec 2018 00:12:26 -0000 Mailing-List: contact kernel-hardening-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-ID: Delivered-To: mailing list kernel-hardening@lists.openwall.com Received: (qmail 30105 invoked from network); 12 Dec 2018 00:12:24 -0000 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.56,343,1539673200"; d="scan'208";a="282839408" From: Rick Edgecombe To: akpm@linux-foundation.org, luto@kernel.org, will.deacon@arm.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org, kernel-hardening@lists.openwall.com, naveen.n.rao@linux.vnet.ibm.com, anil.s.keshavamurthy@intel.com, davem@davemloft.net, mhiramat@kernel.org, rostedt@goodmis.org, mingo@redhat.com, ast@kernel.org, daniel@iogearbox.net, jeyu@kernel.org, namit@vmware.com, netdev@vger.kernel.org, ard.biesheuvel@linaro.org, jannh@google.com Cc: kristen@linux.intel.com, dave.hansen@intel.com, deneen.t.dock@intel.com, Rick Edgecombe Subject: [PATCH v2 4/4] x86/vmalloc: Add TLB efficient x86 arch_vunmap Date: Tue, 11 Dec 2018 16:03:54 -0800 Message-Id: <20181212000354.31955-5-rick.p.edgecombe@intel.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20181212000354.31955-1-rick.p.edgecombe@intel.com> References: <20181212000354.31955-1-rick.p.edgecombe@intel.com> X-Virus-Scanned: ClamAV using ClamSMTP This adds a more efficient x86 architecture specific implementation of arch_vunmap, that can free any type of special permission memory with only 1 TLB flush. In order to enable this, _set_pages_p and _set_pages_np are made non-static and renamed set_pages_p_noflush and set_pages_np_noflush to better communicate their different (non-flushing) behavior from the rest of the set_pages_* functions. The method for doing this with only 1 TLB flush was suggested by Andy Lutomirski. Suggested-by: Andy Lutomirski Signed-off-by: Rick Edgecombe --- arch/x86/include/asm/set_memory.h | 2 + arch/x86/mm/Makefile | 3 +- arch/x86/mm/pageattr.c | 11 +++-- arch/x86/mm/vmalloc.c | 71 +++++++++++++++++++++++++++++++ 4 files changed, 80 insertions(+), 7 deletions(-) create mode 100644 arch/x86/mm/vmalloc.c diff --git a/arch/x86/include/asm/set_memory.h b/arch/x86/include/asm/set_memory.h index 07a25753e85c..70ee81e8914b 100644 --- a/arch/x86/include/asm/set_memory.h +++ b/arch/x86/include/asm/set_memory.h @@ -84,6 +84,8 @@ int set_pages_x(struct page *page, int numpages); int set_pages_nx(struct page *page, int numpages); int set_pages_ro(struct page *page, int numpages); int set_pages_rw(struct page *page, int numpages); +int set_pages_np_noflush(struct page *page, int numpages); +int set_pages_p_noflush(struct page *page, int numpages); extern int kernel_set_to_readonly; void set_kernel_text_rw(void); diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile index 4b101dd6e52f..189681f863a6 100644 --- a/arch/x86/mm/Makefile +++ b/arch/x86/mm/Makefile @@ -13,7 +13,8 @@ CFLAGS_REMOVE_mem_encrypt_identity.o = -pg endif obj-y := init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \ - pat.o pgtable.o physaddr.o setup_nx.o tlb.o cpu_entry_area.o + pat.o pgtable.o physaddr.o setup_nx.o tlb.o cpu_entry_area.o \ + vmalloc.o # Make sure __phys_addr has no stackprotector nostackp := $(call cc-option, -fno-stack-protector) diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c index db7a10082238..db0a4dfb5a7f 100644 --- a/arch/x86/mm/pageattr.c +++ b/arch/x86/mm/pageattr.c @@ -2248,9 +2248,7 @@ int set_pages_rw(struct page *page, int numpages) return set_memory_rw(addr, numpages); } -#ifdef CONFIG_DEBUG_PAGEALLOC - -static int __set_pages_p(struct page *page, int numpages) +int set_pages_p_noflush(struct page *page, int numpages) { unsigned long tempaddr = (unsigned long) page_address(page); struct cpa_data cpa = { .vaddr = &tempaddr, @@ -2269,7 +2267,7 @@ static int __set_pages_p(struct page *page, int numpages) return __change_page_attr_set_clr(&cpa, 0); } -static int __set_pages_np(struct page *page, int numpages) +int set_pages_np_noflush(struct page *page, int numpages) { unsigned long tempaddr = (unsigned long) page_address(page); struct cpa_data cpa = { .vaddr = &tempaddr, @@ -2288,6 +2286,7 @@ static int __set_pages_np(struct page *page, int numpages) return __change_page_attr_set_clr(&cpa, 0); } +#ifdef CONFIG_DEBUG_PAGEALLOC void __kernel_map_pages(struct page *page, int numpages, int enable) { if (PageHighMem(page)) @@ -2303,9 +2302,9 @@ void __kernel_map_pages(struct page *page, int numpages, int enable) * and hence no memory allocations during large page split. */ if (enable) - __set_pages_p(page, numpages); + set_pages_p_noflush(page, numpages); else - __set_pages_np(page, numpages); + set_pages_np_noflush(page, numpages); /* * We should perform an IPI and flush all tlbs, diff --git a/arch/x86/mm/vmalloc.c b/arch/x86/mm/vmalloc.c new file mode 100644 index 000000000000..be9ea42c3dfe --- /dev/null +++ b/arch/x86/mm/vmalloc.c @@ -0,0 +1,71 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * vmalloc.c: x86 arch version of vmalloc.c + * + * (C) Copyright 2018 Intel Corporation + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; version 2 + * of the License. + */ + +#include +#include +#include + +static void set_area_direct_np(struct vm_struct *area) +{ + int i; + + for (i = 0; i < area->nr_pages; i++) + set_pages_np_noflush(area->pages[i], 1); +} + +static void set_area_direct_prw(struct vm_struct *area) +{ + int i; + + for (i = 0; i < area->nr_pages; i++) + set_pages_p_noflush(area->pages[i], 1); +} + +void arch_vunmap(struct vm_struct *area, int deallocate_pages) +{ + int immediate = area->flags & VM_IMMEDIATE_UNMAP; + int special = area->flags & VM_HAS_SPECIAL_PERMS; + + /* Unmap from vmalloc area */ + remove_vm_area(area->addr); + + /* If no need to reset directmap perms, just check if need to flush */ + if (!(deallocate_pages || special)) { + if (immediate) + vm_unmap_aliases(); + return; + } + + /* From here we need to make sure to reset the direct map perms */ + + /* + * If the area being freed does not have any extra capabilities, we can + * just reset the directmap to RW before freeing. + */ + if (!immediate) { + set_area_direct_prw(area); + vm_unmap_aliases(); + return; + } + + /* + * If the vm being freed has security sensitive capabilities such as + * executable we need to make sure there is no W window on the directmap + * before removing the X in the TLB. So we set not present first so we + * can flush without any other CPU picking up the mapping. Then we reset + * RW+P without a flush, since NP prevented it from being cached by + * other cpus. + */ + set_area_direct_np(area); + vm_unmap_aliases(); + set_area_direct_prw(area); +}