From patchwork Thu Sep 8 14:01:03 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Kevin Brodsky X-Patchwork-Id: 9321475 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 046FA607D3 for ; Thu, 8 Sep 2016 14:09:40 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id E9F2329815 for ; Thu, 8 Sep 2016 14:09:39 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id DF02629833; Thu, 8 Sep 2016 14:09:39 +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=-4.2 required=2.0 tests=BAYES_00, RCVD_IN_DNSWL_MED autolearn=ham version=3.3.1 Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.9]) (using TLSv1.2 with cipher AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 016D929815 for ; Thu, 8 Sep 2016 14:09:38 +0000 (UTC) Received: from localhost ([127.0.0.1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.85_2 #1 (Red Hat Linux)) id 1bhzzc-0003j9-UX; Thu, 08 Sep 2016 14:07:52 +0000 Received: from foss.arm.com ([217.140.101.70]) by bombadil.infradead.org with esmtp (Exim 4.85_2 #1 (Red Hat Linux)) id 1bhzuS-0006SI-Qw for linux-arm-kernel@lists.infradead.org; Thu, 08 Sep 2016 14:02:51 +0000 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.72.51.249]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 739EB5CAB; Thu, 8 Sep 2016 07:01:38 -0700 (PDT) Received: from e107154-lin.arm.com (e107154-lin.cambridge.arm.com [10.2.131.184]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id B01B83F21A; Thu, 8 Sep 2016 07:01:37 -0700 (PDT) From: Kevin Brodsky To: linux-arm-kernel@lists.infradead.org Subject: [RFC PATCH 8/8] arm64: Wire up and expose the new compat vDSO Date: Thu, 8 Sep 2016 15:01:03 +0100 Message-Id: <20160908140103.7468-9-kevin.brodsky@arm.com> X-Mailer: git-send-email 2.9.3 In-Reply-To: <20160908140103.7468-1-kevin.brodsky@arm.com> References: <20160908140103.7468-1-kevin.brodsky@arm.com> X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20160908_070233_407651_005DC7AE X-CRM114-Status: GOOD ( 13.47 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Kevin Brodsky , Will Deacon , Dave Martin MIME-Version: 1.0 Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+patchwork-linux-arm=patchwork.kernel.org@lists.infradead.org X-Virus-Scanned: ClamAV using ClamSMTP Expose the new compat vDSO via the COMPAT_VDSO config option. The option is not enabled by default for two reasons: * The vDSO page replaces the vector page. The vDSO provides its own sigreturn trampolines, replacing those in the vector page, but the kuser helpers are gone. As a result enabling the compat vDSO will break userspace programs relying on the kuser helpers. * We really need a 32-bit compiler this time, and we rely on the user to provide it themselves by setting CROSS_COMPILE_ARM32. Therefore enabling the option by default would make little sense, since the user must explicitly set an environment variable anyway. CONFIG_COMPAT_VDSO is not directly used in the code, because we want to ignore it (build as if it were not set) if the user didn't set CROSS_COMPILE_ARM32 properly. If the variable has been set to a valid prefix, CONFIG_VDSO32 will be set; this is the option that the code and Makefiles will test. Signed-off-by: Kevin Brodsky --- arch/arm64/Kconfig | 20 ++++++++++++++++++++ arch/arm64/Makefile | 20 ++++++++++++++++++-- arch/arm64/kernel/Makefile | 8 ++++++-- 3 files changed, 44 insertions(+), 4 deletions(-) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index bc3f00f586f1..10afbd17b158 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -1012,6 +1012,26 @@ config SYSVIPC_COMPAT def_bool y depends on COMPAT && SYSVIPC +config COMPAT_VDSO + bool "32-bit vDSO" + depends on COMPAT + default n + help + Warning: this completely removes the compat vector page, including + kuser helpers, which may break 32-bit processes. + + Warning: a 32-bit toolchain is necessary to build the vDSO. You + must explicitly define which toolchain should be used by setting + CROSS_COMPILE_ARM32 to the prefix of the 32-bit toolchain (same format + as CROSS_COMPILE). If a 32-bit compiler cannot be found, a warning + will be printed and the kernel will be built as if COMPAT_VDSO had not + been set. + + Provide a vDSO to 32-bit processes. It includes the symbols provided + by the vDSO from the 32-bit kernel, so that a 32-bit libc can use + the compat vDSO without modification. It also provides sigreturn + trampolines, and replaces the vector page. + endmenu menu "Power management options" diff --git a/arch/arm64/Makefile b/arch/arm64/Makefile index 5b54f8c021d8..88c18fbb550b 100644 --- a/arch/arm64/Makefile +++ b/arch/arm64/Makefile @@ -29,10 +29,24 @@ $(warning LSE atomics not supported by binutils) endif endif -KBUILD_CFLAGS += -mgeneral-regs-only $(lseinstr) +ifeq ($(CONFIG_COMPAT_VDSO), y) + # Check that the user has provided a valid prefix for arm32 (we allow + # CROSS_COMPILE_ARM32 to be defined but empty, in case the system gcc should + # be used) + ifeq ($(shell [ -v CROSS_COMPILE_ARM32 ] && echo y),) + $(warning CROSS_COMPILE_ARM32 not defined, the compat vDSO will not be built) + else ifeq ($(shell which $(CROSS_COMPILE_ARM32)gcc 2> /dev/null),) + $(warning $(CROSS_COMPILE_ARM32)gcc not found, the compat vDSO will not be built) + else + export CONFIG_VDSO32 := y + vdso32 := -DCONFIG_VDSO32=1 + endif +endif + +KBUILD_CFLAGS += -mgeneral-regs-only $(lseinstr) $(vdso32) KBUILD_CFLAGS += -fno-asynchronous-unwind-tables KBUILD_CFLAGS += $(call cc-option, -mpc-relative-literal-loads) -KBUILD_AFLAGS += $(lseinstr) +KBUILD_AFLAGS += $(lseinstr) $(vdso32) ifeq ($(CONFIG_CPU_BIG_ENDIAN), y) KBUILD_CPPFLAGS += -mbig-endian @@ -129,6 +143,8 @@ archclean: prepare: vdso_prepare vdso_prepare: prepare0 $(Q)$(MAKE) $(build)=arch/arm64/kernel/vdso include/generated/vdso-offsets.h + $(if $(CONFIG_VDSO32),$(Q)$(MAKE) $(build)=arch/arm64/kernel/vdso32 \ + include/generated/vdso32-offsets.h) define archhelp echo '* Image.gz - Compressed kernel image (arch/$(ARCH)/boot/Image.gz)' diff --git a/arch/arm64/kernel/Makefile b/arch/arm64/kernel/Makefile index 14f7b651c787..509d0adc9bf0 100644 --- a/arch/arm64/kernel/Makefile +++ b/arch/arm64/kernel/Makefile @@ -25,8 +25,11 @@ OBJCOPYFLAGS := --prefix-symbols=__efistub_ $(obj)/%.stub.o: $(obj)/%.o FORCE $(call if_changed,objcopy) -arm64-obj-$(CONFIG_COMPAT) += sys32.o kuser32.o signal32.o \ - sys_compat.o entry32.o +arm64-obj-$(CONFIG_COMPAT) += sys32.o signal32.o sys_compat.o \ + entry32.o +ifneq ($(CONFIG_VDSO32),y) +arm64-obj-y += kuser32.o +endif arm64-obj-$(CONFIG_FUNCTION_TRACER) += ftrace.o entry-ftrace.o arm64-obj-$(CONFIG_MODULES) += arm64ksyms.o module.o arm64-obj-$(CONFIG_ARM64_MODULE_PLTS) += module-plts.o @@ -50,6 +53,7 @@ arm64-obj-$(CONFIG_KEXEC) += machine_kexec.o relocate_kernel.o \ cpu-reset.o obj-y += $(arm64-obj-y) vdso/ probes/ +obj-$(CONFIG_VDSO32) += vdso32/ obj-m += $(arm64-obj-m) head-y := head.o extra-y += $(head-y) vmlinux.lds