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=-9.1 required=3.0 tests=DKIMWL_WL_HIGH,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, 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 5BB50C43381 for ; Sat, 16 Mar 2019 06:16:24 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 0FE1B218FE for ; Sat, 16 Mar 2019 06:16:23 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1552716984; bh=apc524QGyJcH6bLeD9FBZzBw5aMRBczK4i6qo35kH6A=; h=From:To:Cc:Subject:Date:List-ID:From; b=q/ickIZldyDXmpJFRYpmNORI9wlccFVizN3+sGiEJDMFsfQXVViRYAKk/RRyydpkz Ca6xp2x1KFTYybcOIJ4wRg8bFWWghzk3G7v6DPi5eYzsZdbvtkUcJD8wRMDzplaSMu EhJ/uxlV7ow+ZgLPhsXxTUewdAlWs6wgOi/STQFI= Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726222AbfCPGQW (ORCPT ); Sat, 16 Mar 2019 02:16:22 -0400 Received: from mail.kernel.org ([198.145.29.99]:35358 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725935AbfCPGQV (ORCPT ); Sat, 16 Mar 2019 02:16:21 -0400 Received: from localhost.localdomain (unknown [220.191.37.162]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id EF1CF218E0; Sat, 16 Mar 2019 06:16:18 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1552716980; bh=apc524QGyJcH6bLeD9FBZzBw5aMRBczK4i6qo35kH6A=; h=From:To:Cc:Subject:Date:From; b=WKBQoN4hr/EVsQTRWKPzFzQ5icDJEt9uT4o7fwsaiLqUt84rOUDdlSSxPjV34iDWx LJMZIRftkAwEMBt5GZCxT0H6jNJ/LIBJ5BZ+eV0NlHCRnLuqUuS/bu513Ssv/iAmEN /xYS4A0WFLQpLdylGLAiqXRyfi/bdpG19ddGf9ao= From: guoren@kernel.org To: arnd@arndb.de Cc: linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, ren_guo@c-sky.com Subject: [PATCH] csky/ftrace: Add dynamic function tracer (include graph tracer) Date: Sat, 16 Mar 2019 14:16:12 +0800 Message-Id: <1552716972-28009-1-git-send-email-guoren@kernel.org> X-Mailer: git-send-email 2.7.4 Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Guo Ren Support dynamic ftrace including dynamic graph tracer. Gcc-csky with -pg will produce call site in every function prologue and we can use these call site to hook trace function. gcc with -pg origin call site: push lr jbsr _mcount nop nop nop nop If the (callee - caller)'s offset is in range of bsr instruction, we'll modify code with: push lr bsr _mcount nop nop nop nop Else if the (callee - caller)'s offset is out of bsr instrunction, we'll modify code with: push lr movih r26, ... ori r26, ... jsr r26 (r26 is reserved for jsr link reg in csky abiv2 spec.) Signed-off-by: Guo Ren --- arch/csky/Kconfig | 3 +- arch/csky/abiv2/mcount.S | 43 +++++++++++- arch/csky/include/asm/ftrace.h | 18 ++++- arch/csky/kernel/ftrace.c | 148 ++++++++++++++++++++++++++++++++++++++++- scripts/recordmcount.pl | 3 + 5 files changed, 209 insertions(+), 6 deletions(-) diff --git a/arch/csky/Kconfig b/arch/csky/Kconfig index 398113c..114cb2f 100644 --- a/arch/csky/Kconfig +++ b/arch/csky/Kconfig @@ -28,14 +28,15 @@ config CSKY select GENERIC_SCHED_CLOCK select GENERIC_SMP_IDLE_THREAD select HAVE_ARCH_TRACEHOOK + select HAVE_DYNAMIC_FTRACE select HAVE_FUNCTION_TRACER select HAVE_FUNCTION_GRAPH_TRACER + select HAVE_FTRACE_MCOUNT_RECORD select HAVE_GENERIC_DMA_COHERENT select HAVE_KERNEL_GZIP select HAVE_KERNEL_LZO select HAVE_KERNEL_LZMA select HAVE_PERF_EVENTS - select HAVE_C_RECORDMCOUNT select HAVE_DMA_API_DEBUG select HAVE_DMA_CONTIGUOUS select MAY_HAVE_SPARSE_IRQ diff --git a/arch/csky/abiv2/mcount.S b/arch/csky/abiv2/mcount.S index c633379..7a66fed 100644 --- a/arch/csky/abiv2/mcount.S +++ b/arch/csky/abiv2/mcount.S @@ -61,10 +61,20 @@ addi sp, 16 .endm +.macro nop_stub_6 + nop + nop + nop + nop + nop + nop +.endm + ENTRY(ftrace_stub) jmp lr END(ftrace_stub) +#ifndef CONFIG_DYNAMIC_FTRACE ENTRY(_mcount) mcount_enter @@ -76,7 +86,7 @@ ENTRY(_mcount) bf skip_ftrace mov a0, lr - subi a0, MCOUNT_INSN_SIZE + subi a0, 4 ldw a1, (sp, 24) jsr r26 @@ -101,13 +111,42 @@ skip_ftrace: mcount_exit #endif END(_mcount) +#else /* CONFIG_DYNAMIC_FTRACE */ +ENTRY(_mcount) + mov t1, lr + ldw lr, (sp, 0) + addi sp, 4 + jmp t1 +ENDPROC(_mcount) + +ENTRY(ftrace_caller) + mcount_enter + + ldw a0, (sp, 16) + subi a0, 4 + ldw a1, (sp, 24) + + nop +GLOBAL(ftrace_call) + nop_stub_6 + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + nop +GLOBAL(ftrace_graph_call) + nop_stub_6 +#endif + + mcount_exit +ENDPROC(ftrace_caller) +#endif /* CONFIG_DYNAMIC_FTRACE */ #ifdef CONFIG_FUNCTION_GRAPH_TRACER ENTRY(ftrace_graph_caller) +GLOBAL(ftrace_graph_caller) mov a0, sp addi a0, 24 ldw a1, (sp, 16) - subi a1, MCOUNT_INSN_SIZE + subi a1, 4 mov a2, r8 lrw r26, prepare_ftrace_return jsr r26 diff --git a/arch/csky/include/asm/ftrace.h b/arch/csky/include/asm/ftrace.h index 7547c45..ba35d93 100644 --- a/arch/csky/include/asm/ftrace.h +++ b/arch/csky/include/asm/ftrace.h @@ -4,10 +4,26 @@ #ifndef __ASM_CSKY_FTRACE_H #define __ASM_CSKY_FTRACE_H -#define MCOUNT_INSN_SIZE 4 +#define MCOUNT_INSN_SIZE 14 #define HAVE_FUNCTION_GRAPH_FP_TEST #define HAVE_FUNCTION_GRAPH_RET_ADDR_PTR +#define MCOUNT_ADDR ((unsigned long)_mcount) + +#ifndef __ASSEMBLY__ + +extern void _mcount(unsigned long); + +extern void ftrace_graph_call(void); + +static inline unsigned long ftrace_call_adjust(unsigned long addr) +{ + return addr; +} + +struct dyn_arch_ftrace { +}; +#endif /* !__ASSEMBLY__ */ #endif /* __ASM_CSKY_FTRACE_H */ diff --git a/arch/csky/kernel/ftrace.c b/arch/csky/kernel/ftrace.c index 274c431..ae3870e 100644 --- a/arch/csky/kernel/ftrace.c +++ b/arch/csky/kernel/ftrace.c @@ -3,6 +3,137 @@ #include #include +#include + +#ifdef CONFIG_DYNAMIC_FTRACE + +#define NOP 0x4000 +#define PUSH_LR 0x14d0 +#define MOVIH_LINK 0xea3a +#define ORI_LINK 0xef5a +#define JSR_LINK 0xe8fa +#define BSR_LINK 0xe000 + +/* + * Gcc-csky with -pg will insert stub in function prologue: + * push lr + * jbsr _mcount + * nop + * nop + * nop + * nop + * + * If the (callee - current_pc) is less then 64MB, we'll use bsr: + * push lr + * bsr _mcount + * nop + * nop + * nop + * nop + * else we'll use (movih + ori + jsr): + * push lr + * movih r26, ... + * ori r26, ... + * jsr r26 + * + * (r26 is our reserved link-reg) + * + */ +static inline void make_jbsr(unsigned long callee, unsigned long pc, + uint16_t *call, bool nolr) +{ + long offset; + + call[0] = nolr ? NOP : PUSH_LR; + + offset = (long) callee - (long) pc; + + if (unlikely(offset < -67108864 || offset > 67108864)) { + call[1] = MOVIH_LINK; + call[2] = callee >> 16; + call[3] = ORI_LINK; + call[4] = callee & 0xffff; + call[5] = JSR_LINK; + call[6] = 0; + } else { + offset = offset >> 1; + + call[1] = BSR_LINK | + ((uint16_t)((unsigned long) offset >> 16) & 0x3ff); + call[2] = (uint16_t)((unsigned long) offset & 0xffff); + call[3] = call[4] = call[5] = call[6] = NOP; + } +} + +static int ftrace_check_current_nop(unsigned long hook) +{ + uint16_t olds[7]; + uint16_t nops[7] = {NOP, NOP, NOP, NOP, NOP, NOP, NOP}; + unsigned long hook_pos = hook - 2; + + if (probe_kernel_read((void *)olds, (void *)hook_pos, sizeof(nops))) + return -EFAULT; + + if (memcmp((void *)nops, (void *)olds, sizeof(nops))) { + pr_err("%p: nop but get (%04x %04x %04x %04x %04x %04x %04x)\n", + (void *)hook_pos, + olds[0], olds[1], olds[2], olds[3], olds[4], olds[5], + olds[6]); + + return -EINVAL; + } + + return 0; +} + +static int ftrace_modify_code(unsigned long hook, unsigned long target, + bool enable, bool nolr) +{ + uint16_t call[7]; + uint16_t nops[7] = {NOP, NOP, NOP, NOP, NOP, NOP, NOP}; + unsigned long hook_pos = hook - 2; + int ret = 0; + + make_jbsr(target, hook, call, nolr); + + ret = probe_kernel_write((void *)hook_pos, enable ? call : nops, + sizeof(nops)); + if (ret) + return -EPERM; + + flush_icache_range(hook_pos, hook_pos + MCOUNT_INSN_SIZE); + + return 0; +} + +int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr) +{ + int ret = ftrace_check_current_nop(rec->ip); + + if (ret) + return ret; + + return ftrace_modify_code(rec->ip, addr, true, false); +} + +int ftrace_make_nop(struct module *mod, struct dyn_ftrace *rec, + unsigned long addr) +{ + return ftrace_modify_code(rec->ip, addr, false, false); +} + +int ftrace_update_ftrace_func(ftrace_func_t func) +{ + int ret = ftrace_modify_code((unsigned long)&ftrace_call, + (unsigned long)func, true, true); + return ret; +} + +int __init ftrace_dyn_arch_init(void) +{ + return 0; +} +#endif /* CONFIG_DYNAMIC_FTRACE */ #ifdef CONFIG_FUNCTION_GRAPH_TRACER void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr, @@ -43,8 +174,21 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr, *(unsigned long *)frame_pointer = return_hooker; } } -#endif + +#ifdef CONFIG_DYNAMIC_FTRACE +int ftrace_enable_ftrace_graph_caller(void) +{ + return ftrace_modify_code((unsigned long)&ftrace_graph_call, + (unsigned long)&ftrace_graph_caller, true, true); +} + +int ftrace_disable_ftrace_graph_caller(void) +{ + return ftrace_modify_code((unsigned long)&ftrace_graph_call, + (unsigned long)&ftrace_graph_caller, false, true); +} +#endif /* CONFIG_DYNAMIC_FTRACE */ +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ /* _mcount is defined in abi's mcount.S */ -extern void _mcount(void); EXPORT_SYMBOL(_mcount); diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 68841d0..f716668 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -397,6 +397,9 @@ if ($arch eq "x86_64") { } elsif ($arch eq "nds32") { $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_NDS32_HI20_RELA\\s+_mcount\$"; $alignment = 2; +} elsif ($arch eq "csky") { + $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_CKCORE_PCREL_JSR_IMM26BY2\\s+_mcount\$"; + $alignment = 2; } else { die "Arch $arch is not supported with CONFIG_FTRACE_MCOUNT_RECORD"; } -- 2.7.4