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=-3.8 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS 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 0144BC0044B for ; Mon, 1 Oct 2018 14:16:54 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id B1DC9213A2 for ; Mon, 1 Oct 2018 14:16:53 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org B1DC9213A2 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=lst.de Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1729523AbeJAUyx (ORCPT ); Mon, 1 Oct 2018 16:54:53 -0400 Received: from verein.lst.de ([213.95.11.211]:37900 "EHLO newverein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1729295AbeJAUyw (ORCPT ); Mon, 1 Oct 2018 16:54:52 -0400 Received: by newverein.lst.de (Postfix, from userid 2005) id 1DBED68BDF; Mon, 1 Oct 2018 16:16:48 +0200 (CEST) To: Will Deacon , Catalin Marinas , Julien Thierry , Steven Rostedt , Josh Poimboeuf , Ingo Molnar , Ard Biesheuvel , Arnd Bergmann , AKASHI Takahiro Cc: linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, live-patching@vger.kernel.org Subject: [PATCH v3 2/4] arm64: implement ftrace with regs In-Reply-To: <20181001140910.086E768BC7@newverein.lst.de> References: <20181001140910.086E768BC7@newverein.lst.de> Message-Id: <20181001141648.1DBED68BDF@newverein.lst.de> Date: Mon, 1 Oct 2018 16:16:48 +0200 (CEST) From: duwe@lst.de (Torsten Duwe) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Check for compiler support of -fpatchable-function-entry and use it to intercept functions immediately on entry, saving the LR in x9. patchable-function-entry in GCC disables IPA-RA, which means ABI register calling conventions are obeyed *and* scratch registers are available. Disable ftracing in efi/libstub, because this triggers cross-section linker errors now (-pg is disabled already for those files). Add an ftrace_caller which can handle LR in x9, as well as an ftrace_regs_caller that additionally writes out a set of pt_regs for inspection. Introduce and handle an ftrace_regs_trampoline for module PLTs. Signed-off-by: Torsten Duwe --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -110,6 +110,7 @@ config ARM64 select HAVE_DEBUG_KMEMLEAK select HAVE_DMA_CONTIGUOUS select HAVE_DYNAMIC_FTRACE + select HAVE_DYNAMIC_FTRACE_WITH_REGS select HAVE_EFFICIENT_UNALIGNED_ACCESS select HAVE_FTRACE_MCOUNT_RECORD select HAVE_FUNCTION_TRACER --- a/arch/arm64/Makefile +++ b/arch/arm64/Makefile @@ -78,6 +78,15 @@ ifeq ($(CONFIG_ARM64_MODULE_PLTS),y) KBUILD_LDFLAGS_MODULE += -T $(srctree)/arch/arm64/kernel/module.lds endif +ifeq ($(CONFIG_DYNAMIC_FTRACE_WITH_REGS),y) + CC_FLAGS_FTRACE := -fpatchable-function-entry=2 + KBUILD_CPPFLAGS += -DCC_USING_PATCHABLE_FUNCTION_ENTRY + ifeq ($(call cc-option,-fpatchable-function-entry=2),) + $(error Cannot use CONFIG_DYNAMIC_FTRACE_WITH_REGS: \ + -fpatchable-function-entry not supported by compiler) + endif +endif + # Default value head-y := arch/arm64/kernel/head.o --- a/arch/arm64/include/asm/ftrace.h +++ b/arch/arm64/include/asm/ftrace.h @@ -16,6 +16,17 @@ #define MCOUNT_ADDR ((unsigned long)_mcount) #define MCOUNT_INSN_SIZE AARCH64_INSN_SIZE +/* DYNAMIC_FTRACE_WITH_REGS is implemented by adding 2 NOPs at the beginning + of each function, with the second NOP actually calling ftrace. In contrary + to a classic _mcount call, the call instruction to be modified is thus + the second one, and not the only one. */ +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS +#define ARCH_SUPPORTS_FTRACE_OPS 1 +#define REC_IP_BRANCH_OFFSET AARCH64_INSN_SIZE +#else +#define REC_IP_BRANCH_OFFSET 0 +#endif + #ifndef __ASSEMBLY__ #include --- a/arch/arm64/kernel/Makefile +++ b/arch/arm64/kernel/Makefile @@ -7,9 +7,9 @@ CPPFLAGS_vmlinux.lds := -DTEXT_OFFSET=$( AFLAGS_head.o := -DTEXT_OFFSET=$(TEXT_OFFSET) CFLAGS_armv8_deprecated.o := -I$(src) -CFLAGS_REMOVE_ftrace.o = -pg -CFLAGS_REMOVE_insn.o = -pg -CFLAGS_REMOVE_return_address.o = -pg +CFLAGS_REMOVE_ftrace.o = -pg $(CC_FLAGS_FTRACE) +CFLAGS_REMOVE_insn.o = -pg $(CC_FLAGS_FTRACE) +CFLAGS_REMOVE_return_address.o = -pg $(CC_FLAGS_FTRACE) # Object file lists. arm64-obj-y := debug-monitors.o entry.o irq.o fpsimd.o \ --- a/drivers/firmware/efi/libstub/Makefile +++ b/drivers/firmware/efi/libstub/Makefile @@ -11,7 +11,8 @@ cflags-$(CONFIG_X86) += -m$(BITS) -D__K -fPIC -fno-strict-aliasing -mno-red-zone \ -mno-mmx -mno-sse -fshort-wchar -cflags-$(CONFIG_ARM64) := $(subst -pg,,$(KBUILD_CFLAGS)) -fpie +cflags-$(CONFIG_ARM64) := $(filter-out -pg $(CC_FLAGS_FTRACE)\ + ,$(KBUILD_CFLAGS)) -fpie cflags-$(CONFIG_ARM) := $(subst -pg,,$(KBUILD_CFLAGS)) \ -fno-builtin -fpic -mno-single-pic-base --- a/arch/arm64/kernel/entry-ftrace.S +++ b/arch/arm64/kernel/entry-ftrace.S @@ -13,6 +13,8 @@ #include #include #include +#include +#include /* * Gcc with -pg will put the following code in the beginning of each function: @@ -123,6 +125,7 @@ skip_ftrace_call: // } ENDPROC(_mcount) #else /* CONFIG_DYNAMIC_FTRACE */ +#ifndef CC_USING_PATCHABLE_FUNCTION_ENTRY /* * _mcount() is used to build the kernel with -pg option, but all the branch * instructions to _mcount() are replaced to NOP initially at kernel start up, @@ -162,6 +165,92 @@ ftrace_graph_call: // ftrace_graph_cal mcount_exit ENDPROC(ftrace_caller) +#else /* CC_USING_PATCHABLE_FUNCTION_ENTRY */ + +/* Since no -pg or similar compiler flag is used, there should really be + no reference to _mcount; so do not define one. Only a function address + is needed in order to refer to it. */ +ENTRY(_mcount) + ret /* just in case, prevent any fall through. */ +ENDPROC(_mcount) + +ENTRY(ftrace_regs_caller) + sub sp, sp, #S_FRAME_SIZE + stp x29, x9, [sp, #-16] /* FP/LR link */ + + stp x10, x11, [sp, #S_X10] + stp x12, x13, [sp, #S_X12] + stp x14, x15, [sp, #112] + stp x16, x17, [sp, #128] + stp x18, x19, [sp, #144] + stp x20, x21, [sp, #160] + stp x22, x23, [sp, #176] + stp x24, x25, [sp, #192] + stp x26, x27, [sp, #208] + + b ftrace_common +ENDPROC(ftrace_regs_caller) + +ENTRY(ftrace_caller) + sub sp, sp, #S_FRAME_SIZE + stp x29, x9, [sp, #-16] /* FP/LR link */ + +ftrace_common: + stp x28, x29, [sp, #224] /* FP in pt_regs + "our" x28 */ + + /* save function arguments */ + stp x0, x1, [sp] + stp x2, x3, [sp, #16] + stp x4, x5, [sp, #32] + stp x6, x7, [sp, #48] + stp x8, x9, [sp, #64] + + /* The link Register at callee entry */ + str x9, [sp, #S_LR] /* to pt_regs.r[30] */ + /* The program counter just after the ftrace call site */ + str lr, [sp, #S_PC] + /* The stack pointer as it was on ftrace_caller entry... */ + add x29, sp, #S_FRAME_SIZE + str x29, [sp, #S_SP] + + ldr_l x2, function_trace_op, x0 + mov x1, x9 /* saved LR == parent IP */ + sub x0, lr, #8 /* function entry == IP */ + mov x3, sp /* pt_regs are @sp */ + sub sp, sp, #16 /* skip over FP/LR link */ + + .global ftrace_call +ftrace_call: + + bl ftrace_stub + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + .global ftrace_graph_call +ftrace_graph_call: // ftrace_graph_caller(); + nop // If enabled, this will be replaced + // "b ftrace_graph_caller" +#endif + +ftrace_common_return: + add sp, sp, #16 /* advance to pt_regs for restore */ + + ldp x0, x1, [sp] + ldp x2, x3, [sp, #16] + ldp x4, x5, [sp, #32] + ldp x6, x7, [sp, #48] + ldp x8, x9, [sp, #64] + + ldp x28, x29, [sp, #224] + + ldr x9, [sp, #S_PC] + ldr lr, [sp, #S_LR] + add sp, sp, #S_FRAME_SIZE + + ret x9 + +ENDPROC(ftrace_caller) + +#endif /* CC_USING_PATCHABLE_FUNCTION_ENTRY */ #endif /* CONFIG_DYNAMIC_FTRACE */ ENTRY(ftrace_stub) @@ -197,12 +286,20 @@ ENDPROC(ftrace_stub) * and run return_to_handler() later on its exit. */ ENTRY(ftrace_graph_caller) +#ifndef CC_USING_PATCHABLE_FUNCTION_ENTRY mcount_get_lr_addr x0 // pointer to function's saved lr mcount_get_pc x1 // function's pc mcount_get_parent_fp x2 // parent's fp bl prepare_ftrace_return // prepare_ftrace_return(&lr, pc, fp) mcount_exit +#else + add x0, sp, #S_LR+16 /* address of (LR pointing into caller) */ + ldr x1, [sp, #S_PC+16] + ldr x2, [sp, #S_X28 + 8+16] /* caller's frame pointer */ + bl prepare_ftrace_return + b ftrace_common_return +#endif ENDPROC(ftrace_graph_caller) /* --- a/arch/arm64/kernel/ftrace.c +++ b/arch/arm64/kernel/ftrace.c @@ -65,18 +65,66 @@ int ftrace_update_ftrace_func(ftrace_fun return ftrace_modify_code(pc, 0, new, false); } +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS +/* Have the assembler generate a known "mov x9,x30" at compile time. */ +static void notrace noinline __attribute__((used)) mov_x9_x30(void) +{ + asm(" .global insn_mov_x9_x30\n" + "insn_mov_x9_x30: mov x9,x30\n" : : : "x9"); +} +#endif + +#ifdef CONFIG_ARM64_MODULE_PLTS +int use_ftrace_trampoline(struct module *mod, unsigned long *addr) +{ + struct plt_entry trampoline; + trampoline = get_plt_entry(*addr); + if (*addr == FTRACE_ADDR) { + if (!plt_entries_equal(mod->arch.ftrace_trampoline, + &trampoline)) { + + /* point the trampoline to our ftrace entry point */ + module_disable_ro(mod); + *mod->arch.ftrace_trampoline = trampoline; + module_enable_ro(mod, true); + + /* update trampoline before patching in the branch */ + smp_wmb(); + } + *addr = (unsigned long)(void *)mod->arch.ftrace_trampoline; + } + else if (*addr == FTRACE_REGS_ADDR) { + if (!plt_entries_equal(mod->arch.ftrace_regs_trampoline, + &trampoline)) { + + /* point the trampoline to our ftrace entry point */ + module_disable_ro(mod); + *mod->arch.ftrace_regs_trampoline = trampoline; + module_enable_ro(mod, true); + + /* update trampoline before patching in the branch */ + smp_wmb(); + } + *addr = (unsigned long)(void *)mod->arch.ftrace_regs_trampoline; + } + else + return -EINVAL; + return 0; +} +#endif + /* * Turn on the call to ftrace_caller() in instrumented function */ int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr) { - unsigned long pc = rec->ip; + unsigned long pc = rec->ip + REC_IP_BRANCH_OFFSET; + int ret; u32 old, new; long offset = (long)pc - (long)addr; if (offset < -SZ_128M || offset >= SZ_128M) { #ifdef CONFIG_ARM64_MODULE_PLTS - struct plt_entry trampoline; struct module *mod; /* @@ -96,54 +144,67 @@ int ftrace_make_call(struct dyn_ftrace * if (WARN_ON(!mod)) return -EINVAL; - /* - * There is only one ftrace trampoline per module. For now, - * this is not a problem since on arm64, all dynamic ftrace - * invocations are routed via ftrace_caller(). This will need - * to be revisited if support for multiple ftrace entry points - * is added in the future, but for now, the pr_err() below - * deals with a theoretical issue only. - */ - trampoline = get_plt_entry(addr); - if (!plt_entries_equal(mod->arch.ftrace_trampoline, - &trampoline)) { - if (!plt_entries_equal(mod->arch.ftrace_trampoline, - &(struct plt_entry){})) { - pr_err("ftrace: far branches to multiple entry points unsupported inside a single module\n"); - return -EINVAL; - } - - /* point the trampoline to our ftrace entry point */ - module_disable_ro(mod); - *mod->arch.ftrace_trampoline = trampoline; - module_enable_ro(mod, true); - - /* update trampoline before patching in the branch */ - smp_wmb(); + /* Check against our well-known list of ftrace entry points */ + if (addr == FTRACE_ADDR || addr == FTRACE_REGS_ADDR) { + ret = use_ftrace_trampoline(mod, &addr); + if (ret < 0) + return ret; } - addr = (unsigned long)(void *)mod->arch.ftrace_trampoline; + else + return -EINVAL; + #else /* CONFIG_ARM64_MODULE_PLTS */ return -EINVAL; #endif /* CONFIG_ARM64_MODULE_PLTS */ } old = aarch64_insn_gen_nop(); + if (IS_ENABLED(CONFIG_DYNAMIC_FTRACE_WITH_REGS)) { + new = *(u32*)&mov_x9_x30; + ret = ftrace_modify_code(pc - REC_IP_BRANCH_OFFSET, + old, new, true); + if (ret) + return ret; + smp_wmb(); /* ensure LR saver is in place before ftrace call */ + } new = aarch64_insn_gen_branch_imm(pc, addr, AARCH64_INSN_BRANCH_LINK); return ftrace_modify_code(pc, old, new, true); } +int ftrace_modify_call(struct dyn_ftrace *rec, unsigned long old_addr, + unsigned long addr) +{ + unsigned long pc = rec->ip + REC_IP_BRANCH_OFFSET; + u32 old, new; + + old = aarch64_insn_gen_branch_imm(pc, old_addr, true); + new = aarch64_insn_gen_branch_imm(pc, addr, true); + + return ftrace_modify_code(pc, old, new, true); +} + + /* * Turn off the call to ftrace_caller() in instrumented function */ int ftrace_make_nop(struct module *mod, struct dyn_ftrace *rec, unsigned long addr) { - unsigned long pc = rec->ip; + unsigned long pc = rec->ip + REC_IP_BRANCH_OFFSET; bool validate = true; + int ret; u32 old = 0, new; long offset = (long)pc - (long)addr; +#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY + /* -fpatchable-function-entry= does not generate a profiling call + * initially; the NOPs are already there. + */ + if (addr == MCOUNT_ADDR) + return 0; +#endif + if (offset < -SZ_128M || offset >= SZ_128M) { #ifdef CONFIG_ARM64_MODULE_PLTS u32 replaced; @@ -188,7 +249,16 @@ int ftrace_make_nop(struct module *mod, new = aarch64_insn_gen_nop(); - return ftrace_modify_code(pc, old, new, validate); + ret = ftrace_modify_code(pc, old, new, validate); + if (ret) + return ret; + if (IS_ENABLED(CONFIG_DYNAMIC_FTRACE_WITH_REGS)) { + smp_wmb(); /* ftrace call must not remain without LR saver. */ + old = *(u32*)&mov_x9_x30; + ret = ftrace_modify_code(pc - REC_IP_BRANCH_OFFSET, + old, new, true); + } + return ret; } void arch_ftrace_update_code(int command) --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -115,6 +115,7 @@ #define MCOUNT_REC() . = ALIGN(8); \ __start_mcount_loc = .; \ KEEP(*(__mcount_loc)) \ + KEEP(*(__patchable_function_entries)) \ __stop_mcount_loc = .; #else #define MCOUNT_REC() --- a/include/linux/compiler_types.h +++ b/include/linux/compiler_types.h @@ -61,8 +61,12 @@ extern void __chk_io_ptr(const volatile #if defined(CC_USING_HOTPATCH) && !defined(__CHECKER__) #define notrace __attribute__((hotpatch(0,0))) #else +#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY +#define notrace __attribute__((patchable_function_entry(0))) +#else #define notrace __attribute__((no_instrument_function)) #endif +#endif /* Intel compiler defines __GNUC__. So we will overwrite implementations * coming from above header files here --- a/arch/arm64/include/asm/module.h +++ b/arch/arm64/include/asm/module.h @@ -33,6 +33,7 @@ struct mod_arch_specific { /* for CONFIG_DYNAMIC_FTRACE */ struct plt_entry *ftrace_trampoline; + struct plt_entry *ftrace_regs_trampoline; }; #endif --- a/arch/arm64/kernel/module.c +++ b/arch/arm64/kernel/module.c @@ -454,6 +454,10 @@ int module_finalize(const Elf_Ehdr *hdr, if (IS_ENABLED(CONFIG_DYNAMIC_FTRACE) && !strcmp(".text.ftrace_trampoline", secstrs + s->sh_name)) me->arch.ftrace_trampoline = (void *)s->sh_addr; + if (IS_ENABLED(CONFIG_DYNAMIC_FTRACE_WITH_REGS) && + !strcmp(".text.ftrace_regs_trampoline", + secstrs + s->sh_name)) + me->arch.ftrace_regs_trampoline = (void *)s->sh_addr; #endif }