From patchwork Mon Apr 15 13:43:19 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Linus Walleij X-Patchwork-Id: 13630081 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 292E6C4345F for ; Mon, 15 Apr 2024 13:44:04 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:Cc:To:MIME-Version:Message-Id:Date: Subject:From:Reply-To:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:In-Reply-To:References: List-Owner; bh=uqMhsb4ZvM7gZImYKBrgAIxClA7OzI0wtqRBaQcvJBk=; b=zjGNWzzUqIMhAa ISm/vVXK3877C7UoWUBWMH8WOsiJl+W5hCq4Qm4v/EB4qihztikNSyukvZzHcTgnaDRS4ouNw7Z4U Np8qYwL9Pqo+i3ijfAwUQhn8LesEj3XjCqdCS5jV+so89BKtcOoBMwBQAfJuuM61sJr+3w3LMPx8G LxKneRQPtS3BR9aHpnm34mA5fwnRVv8cTuH/MqaaXWa6+R1Giuk3GeFB/BcYkUGh9hRnFUWo7BB0o 5BHAN9TULQX+xFK3WMKPzYe930ZOCjvaX5mWkEV9jDcXYi56awAUt9sVhr2ozDExNBR/WHruNXxzM /L9v2tdaSGf2xl7vZVpg==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.97.1 #2 (Red Hat Linux)) id 1rwMcf-00000008XV7-3B0k; Mon, 15 Apr 2024 13:43:49 +0000 Received: from mail-lj1-x22d.google.com ([2a00:1450:4864:20::22d]) by bombadil.infradead.org with esmtps (Exim 4.97.1 #2 (Red Hat Linux)) id 1rwMcS-00000008XN1-2dWl for linux-arm-kernel@lists.infradead.org; Mon, 15 Apr 2024 13:43:39 +0000 Received: by mail-lj1-x22d.google.com with SMTP id 38308e7fff4ca-2da0f8f7b24so23849371fa.1 for ; Mon, 15 Apr 2024 06:43:34 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1713188612; x=1713793412; darn=lists.infradead.org; h=cc:to:content-transfer-encoding:mime-version:message-id:date :subject:from:from:to:cc:subject:date:message-id:reply-to; bh=4t1Ya7U4PIOxTugO2wR7OUmFXO0mNdioXZKNSh+xxSI=; b=gbHVfpkuINNeFzwFQvC2bntP2YJIgd4vO/Uey8xJPWUPvfyYo5T+eiH77Y5m/Jo1t2 QGKAcNFoNqQ1QMbqJGZAu64JOn6KY9pWhnKf7tY9+IqkfALHz2tkwegKX3N+gPiolOXj I8ce7VSu1eUXwaJHhPxo3UvqOaMEWHWq/k8lWLEg/62+saK67zotJOh0IMJyj0KUaK0m FjY9WZr93E2K5eaGbz+XLQ0cXaUV1hTAZQf13A+S97NGBqtKfNbo0TCrkcXp1mszoEgb kb0mQGZZkJJkHY5HLkoDhWrUwwwzSR68RmEhH84rMh/KAkdpXW8iFk6ZmOoZD8VIk2HP DIpA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1713188612; x=1713793412; h=cc:to:content-transfer-encoding:mime-version:message-id:date :subject:from:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=4t1Ya7U4PIOxTugO2wR7OUmFXO0mNdioXZKNSh+xxSI=; b=OBqzyyiXRlaimfUUJUXR4O30g4EBhcZ/9GoGfrFk+tOmh22unnt9cIlM0hLUi4UrBU MhVH1BEkt1kErHiIunk2mxZkKGjVpNR9oExWcod0g1I0xJ8T+WIKTEmuYQczOy01yqIE KfFij4T2pQtePRxS/pGjVQNtbY2W8NZEFcMFHzqC2lmKFwNewbpqteqh6Agfo0+MJO8w +wf7yosGxnjDNsqklOM1IIT2mq7EMzDc+dJPtW1qm7Dk/yDV3e60CfGutgGGO+4pM+oq 8g3SEF+3SBB5FxrQM0MCFY8Nj4reCN0Y0FYunIq3eQr8KKlP4ICzRxbTDsPOLtN/R0D6 yp+g== X-Gm-Message-State: AOJu0YxJnIgfeOQqgs911FL8wqktzCXXHqPBdcVgrCuLNNa7somLNY0F uMa9Bkf+bpAtrY/nf0K3AZtL3qQzgSHiujMt0N3WGpRaQRu/2/Naa9ALSUDZVPi0z99UappwrRk j X-Google-Smtp-Source: AGHT+IHn/NNt2ymv8+0fGiNGYOpotEuf2vZCPpdp93l4jofNI/23GYBzJGrEugbgh5uL2PTjgezmfw== X-Received: by 2002:a2e:be06:0:b0:2d6:f8d7:ab95 with SMTP id z6-20020a2ebe06000000b002d6f8d7ab95mr3939155ljq.21.1713188612344; Mon, 15 Apr 2024 06:43:32 -0700 (PDT) Received: from [192.168.1.140] ([85.235.12.238]) by smtp.gmail.com with ESMTPSA id p11-20020a2eb98b000000b002d43737e2d7sm1297988ljp.30.2024.04.15.06.43.31 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 Apr 2024 06:43:31 -0700 (PDT) From: Linus Walleij Subject: [PATCH v5 00/10] CFI for ARM32 using LLVM Date: Mon, 15 Apr 2024 15:43:19 +0200 Message-Id: <20240415-arm32-cfi-v5-0-ff11093eeccc@linaro.org> MIME-Version: 1.0 X-B4-Tracking: v=1; b=H4sIAPguHWYC/2XPTQ7CIBAF4Ks0rMUMw4/WlfcwLigFJdHWgCGap nd31IWYLt/MfC+ZiWWfos9s10ws+RJzHAcKetUwd7bDyfPYU2YIqEAIzW26SuQuRG50byAgTWH L6P6WfIiPT9fhSPkc831Mz091Ee/ptwWxbimCAzetkhJMAKO7/SUONo3rMZ3Yu6bgj0rY1BSJO rdR3tLGdHJBZUWFqKkkiqgsBLBKO1xQVVHc1lQRtfS0MkK2AvUfnef5Bew56E5UAQAA To: Russell King , Sami Tolvanen , Kees Cook , Nathan Chancellor , Nick Desaulniers , Ard Biesheuvel , Arnd Bergmann Cc: linux-arm-kernel@lists.infradead.org, llvm@lists.linux.dev, Linus Walleij X-Mailer: b4 0.13.0 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20240415_064336_770909_0AC8E23A X-CRM114-Status: GOOD ( 27.67 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org This is a first patch set to support CLANG CFI (Control Flow Integrity) on ARM32. For information about what CFI is, see: https://clang.llvm.org/docs/ControlFlowIntegrity.html For the kernel KCFI flavor, see: https://lwn.net/Articles/898040/ The base changes required to bring up KCFI on ARM32 was mostly related to the use of custom vtables in the kernel, combined with defines to call into these vtable members directly from sites where they are used. We annotate all assembly calls that are called directly from C with SYM_TYPED_FUNC_START()/SYM_FUNC_END() so it is easy to see while reading the assembly that these functions are called from C and can have CFI prototype information prefixed to them. As protype prefix information is just some random bytes, it is not possible to "fall through" into an assembly function that is tagged with SYM_TYPED_FUNC_START(): there will be some binary noise in front of the function so this design pattern needs to be explicitly avoided at each site where it occurred. The approach to binding the calls to C is two-fold: - Either convert the affected vtable struct to C and provide per-CPU prototypes for all the calls (done for TLB, cache) or: - Provide prototypes in a special files just for CFI and tag all these functions addressable. The permissive mode handles the new breakpoint type (0x03) that LLVM CLANG is emitting. To runtime-test the patches: - Enable CONFIG_LKDTM - echo CFI_FORWARD_PROTO > /sys/kernel/debug/provoke-crash/DIRECT The patch set has been booted to userspace on the following test platforms: - Arm Versatile (QEMU) - Arm Versatile Express (QEMU) - multi_v7 booted on Versatile Express (QEMU) - Footbridge Netwinder (SA110 ARMv4) - Ux500 (ARMv7 SMP) - Gemini (FA526) I am not saying there will not be corner cases that we need to fix in addition to this, but it is enough to get started. Looking at what was fixed for arm64 I am a bit weary that e.g. BPF might need something to trampoline properly. But hopefullt people can get to testing it and help me fix remaining issues before the final version, or we can fix it in-tree. Signed-off-by: Linus Walleij --- Changes in v5: - I started to put the patches into the patch tracker and it rightfully complained that the patches tagging all assembly with CFI symbol type macros and adding C prototypes were too large. - Split the two patches annotating assembly into one patch doing the annotation and one patch adding the C prototypes. This is a good split anyway. - The first patches from the series are unchanged and in the patch tracker, I resend them anyway and will soon populate the patch tracker with the split patches from this series unless there are more comments. - Link to v4: https://lore.kernel.org/r/20240328-arm32-cfi-v4-0-a11046139125@linaro.org Changes in v4: - Rebase on v6.9-rc1 - Use Ard's patch for converting TLB operation vtables to C - Rewrite the cache vtables in C and use SYM_SYM_TYPED_FUNC in the assembly to make CFI work all the way down. - Instead of tagging all the delay functions as __nocfi get to the root cause and annotate the loop delay code with SYM_TYPED_FUNC_START() and rewrite it using explicit branches so we get CFI all the way down. - Drop the patch turning highmem page accesses into static inlines: this was probably a development artifact since this code does a lot of cache and TLB flusing, and that assembly is now properly annotated. - Do not define static inlines tagged __nocfi for all the proc functions, instead provide proper C prototypes in a separate CFI-only file and make these explicitly addressable. - Link to v3: https://lore.kernel.org/r/20240311-arm32-cfi-v3-0-224a0f0a45c2@linaro.org Changes in v3: - Use report_cfi_failure() like everyone else in the breakpoint handler. - I think we cannot implement target and type for the report callback without operand bundling compiler extensions, so just leaving these as zero. - Link to v2: https://lore.kernel.org/r/20240307-arm32-cfi-v2-0-cc74ea0306b3@linaro.org Changes in v2: - Add the missing ftrace graph tracer stub. - Enable permissive mode using a breakpoint handler. - Link to v1: https://lore.kernel.org/r/20240225-arm32-cfi-v1-0-6943306f065b@linaro.org --- Ard Biesheuvel (1): ARM: mm: Make tlbflush routines CFI safe Linus Walleij (9): ARM: bugs: Check in the vtable instead of defined aliases ARM: ftrace: Define ftrace_stub_graph ARM: mm: Type-annotate all cache assembly routines ARM: mm: Rewrite cacheflush vtables in CFI safe C ARM: mm Type-annotate all per-processor assembly routines ARM: mm: Define prototypes for all per-processor calls ARM: lib: Annotate loop delay instructions for CFI ARM: hw_breakpoint: Handle CFI breakpoints ARM: Support CLANG CFI arch/arm/Kconfig | 1 + arch/arm/include/asm/glue-cache.h | 28 +- arch/arm/include/asm/hw_breakpoint.h | 1 + arch/arm/kernel/bugs.c | 2 +- arch/arm/kernel/entry-ftrace.S | 4 + arch/arm/kernel/hw_breakpoint.c | 30 ++ arch/arm/lib/delay-loop.S | 16 +- arch/arm/mm/Makefile | 3 + arch/arm/mm/cache-b15-rac.c | 1 + arch/arm/mm/cache-fa.S | 47 +-- arch/arm/mm/cache-nop.S | 57 +-- arch/arm/mm/cache-v4.S | 55 +-- arch/arm/mm/cache-v4wb.S | 47 ++- arch/arm/mm/cache-v4wt.S | 55 +-- arch/arm/mm/cache-v6.S | 49 ++- arch/arm/mm/cache-v7.S | 74 ++-- arch/arm/mm/cache-v7m.S | 53 ++- arch/arm/mm/cache.c | 663 +++++++++++++++++++++++++++++++++++ arch/arm/mm/proc-arm1020.S | 69 ++-- arch/arm/mm/proc-arm1020e.S | 70 ++-- arch/arm/mm/proc-arm1022.S | 69 ++-- arch/arm/mm/proc-arm1026.S | 70 ++-- arch/arm/mm/proc-arm720.S | 25 +- arch/arm/mm/proc-arm740.S | 26 +- arch/arm/mm/proc-arm7tdmi.S | 34 +- arch/arm/mm/proc-arm920.S | 76 ++-- arch/arm/mm/proc-arm922.S | 69 ++-- arch/arm/mm/proc-arm925.S | 66 ++-- arch/arm/mm/proc-arm926.S | 75 ++-- arch/arm/mm/proc-arm940.S | 69 ++-- arch/arm/mm/proc-arm946.S | 65 ++-- arch/arm/mm/proc-arm9tdmi.S | 26 +- arch/arm/mm/proc-fa526.S | 24 +- arch/arm/mm/proc-feroceon.S | 105 +++--- arch/arm/mm/proc-macros.S | 33 -- arch/arm/mm/proc-mohawk.S | 74 ++-- arch/arm/mm/proc-sa110.S | 23 +- arch/arm/mm/proc-sa1100.S | 31 +- arch/arm/mm/proc-v6.S | 31 +- arch/arm/mm/proc-v7-2level.S | 8 +- arch/arm/mm/proc-v7-3level.S | 8 +- arch/arm/mm/proc-v7-bugs.c | 4 +- arch/arm/mm/proc-v7.S | 66 ++-- arch/arm/mm/proc-v7m.S | 41 +-- arch/arm/mm/proc-xsc3.S | 75 ++-- arch/arm/mm/proc-xscale.S | 127 +++---- arch/arm/mm/proc.c | 500 ++++++++++++++++++++++++++ arch/arm/mm/tlb-fa.S | 12 +- arch/arm/mm/tlb-v4.S | 15 +- arch/arm/mm/tlb-v4wb.S | 12 +- arch/arm/mm/tlb-v4wbi.S | 12 +- arch/arm/mm/tlb-v6.S | 12 +- arch/arm/mm/tlb-v7.S | 14 +- arch/arm/mm/tlb.c | 84 +++++ 54 files changed, 2334 insertions(+), 972 deletions(-) --- base-commit: 4cece764965020c22cff7665b18a012006359095 change-id: 20240115-arm32-cfi-65d60f201108 Best regards,