From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id F143EEE49B2 for ; Wed, 23 Aug 2023 15:18:05 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237072AbjHWPSG (ORCPT ); Wed, 23 Aug 2023 11:18:06 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:45566 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S236948AbjHWPSC (ORCPT ); Wed, 23 Aug 2023 11:18:02 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [139.178.84.217]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 89E21171A; Wed, 23 Aug 2023 08:17:40 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 78F5360E03; Wed, 23 Aug 2023 15:15:46 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 728FDC433C8; Wed, 23 Aug 2023 15:15:42 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803745; bh=0drD+LQDmx/AHTKLGdH1kpL/kFKRrBczUa0Lao0mBUc=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=BHa5BB/ygXp14fJ5mlCzLHnp2z4FiKXUQhn0X611gMkzK0gRcKI25E7Ait3VmQpnF sEc8LFG2fmdqrcsXdRH1QKJStOcgrHyCUVEmy25sJppLkiwv6pdGS2tiSS9gI1bCrN 1MlvLL3MGN8oSSD18gHH9xqd6pIH79L51VL78zeon27+Nhq/vhXTNSuBWw2zAVbmNg nI8dPLPSnPshg70iA4RjBdVG0nml2s+LGyxuDEkYhK24dN5J9bSTzZ8cuUFD3QCfGQ sMXoF6ewCGSHdSeeh3KjBPriZwUeQagYnxgFxfj40Uh5gNSIxfHf3nAGvc9Yix0XDW R5qDiyorawPjw== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 1/9] Documentation: probes: Add a new ret_ip callback parameter Date: Thu, 24 Aug 2023 00:15:40 +0900 Message-Id: <169280373992.282662.14835192462715188987.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Add a new ret_ip callback parameter description. Fixes: cb16330d1274 ("fprobe: Pass return address to the handlers") Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Changes in v4: - Update ret_ip description (Thanks Florent!) --- Documentation/trace/fprobe.rst | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/Documentation/trace/fprobe.rst b/Documentation/trace/fprobe.rst index 40dd2fbce861..5851a14eb893 100644 --- a/Documentation/trace/fprobe.rst +++ b/Documentation/trace/fprobe.rst @@ -91,9 +91,9 @@ The prototype of the entry/exit callback function are as = follows: =20 .. code-block:: c =20 - int entry_callback(struct fprobe *fp, unsigned long entry_ip, struct pt_r= egs *regs, void *entry_data); + int entry_callback(struct fprobe *fp, unsigned long entry_ip, unsigned lo= ng ret_ip, struct pt_regs *regs, void *entry_data); =20 - void exit_callback(struct fprobe *fp, unsigned long entry_ip, struct pt_r= egs *regs, void *entry_data); + void exit_callback(struct fprobe *fp, unsigned long entry_ip, unsigned lo= ng ret_ip, struct pt_regs *regs, void *entry_data); =20 Note that the @entry_ip is saved at function entry and passed to exit hand= ler. If the entry callback function returns !0, the corresponding exit callback= will be cancelled. @@ -108,6 +108,10 @@ If the entry callback function returns !0, the corresp= onding exit callback will Note that this may not be the actual entry address of the function= but the address where the ftrace is instrumented. =20 +@ret_ip + This is the return address that the traced function will return to, + somewhere in the caller. This can be used at both entry and exit. + @regs This is the `pt_regs` data structure at the entry and exit. Note t= hat the instruction pointer of @regs may be different from the @entry_= ip From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 21DD0EE49B2 for ; Wed, 23 Aug 2023 15:19:37 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237050AbjHWPTh (ORCPT ); Wed, 23 Aug 2023 11:19:37 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:48862 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S236997AbjHWPRh (ORCPT ); Wed, 23 Aug 2023 11:17:37 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C2A02101; Wed, 23 Aug 2023 08:17:14 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 4F70160B54; Wed, 23 Aug 2023 15:15:58 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id E5378C433C7; Wed, 23 Aug 2023 15:15:53 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803757; bh=orZA9xUFVqFpdLvkaKF8ReMXYbDCtCeDSHjhlJ2TwQ8=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=OOsgWt3EMGwTrbLLgm8nSA75EMAMh2PXbjmzYKmX5K0her4Vi1X4rOkYhhBb9swbI 8bmRe3M2VMlCxz49OMtqysnOETx5T5YMdmBrZxv0TmxNE6EeuoEdyj9aQ5l5oBa4Qa +0UaoZil4es06/8TYxgYR5k05W/RCfiV6K7s1O9EvnSavpj31dzq1sH6R5Pklp2/kz 0TymQxNhf8bB0hsGWpleS0iPx1buM4M+6aU2JpJq5/piHrOZfaGiMmT0wkboemsfkS hJgMpUvTWu/AUo9soYj2WR9vUiq3ZzOZp/Lv2NsSGiJdmRnhMg9hYvZEGVgOUi+NoO YUw/i4iB+HmCA== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 2/9] fprobe: Use fprobe_regs in fprobe entry handler Date: Thu, 24 Aug 2023 00:15:51 +0900 Message-Id: <169280375109.282662.4109179404470188137.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) This allows fprobes to be available with CONFIG_DYNAMIC_FTRACE_WITH_ARGS instead of CONFIG_DYNAMIC_FTRACE_WITH_REGS, then we can enable fprobe on arm64. Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Changes in v3: - Use FTRACE_OPS_FL_SAVE_ARGS instead of FTRACE_OPS_FL_SAVE_REGS. Changes in v4: - Keep CONFIG_FPROBE check for multi-kprobe because this depends on FPROBE API. --- include/linux/fprobe.h | 2 +- kernel/trace/Kconfig | 3 ++- kernel/trace/bpf_trace.c | 10 +++++++--- kernel/trace/fprobe.c | 4 ++-- kernel/trace/trace_fprobe.c | 6 +++++- lib/test_fprobe.c | 4 ++-- samples/fprobe/fprobe_example.c | 2 +- 7 files changed, 20 insertions(+), 11 deletions(-) diff --git a/include/linux/fprobe.h b/include/linux/fprobe.h index 3e03758151f4..36c0595f7b93 100644 --- a/include/linux/fprobe.h +++ b/include/linux/fprobe.h @@ -35,7 +35,7 @@ struct fprobe { int nr_maxactive; =20 int (*entry_handler)(struct fprobe *fp, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *regs, void *entry_data); void (*exit_handler)(struct fprobe *fp, unsigned long entry_ip, unsigned long ret_ip, struct pt_regs *regs, diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 61c541c36596..976fd594b446 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -287,7 +287,7 @@ config DYNAMIC_FTRACE_WITH_ARGS config FPROBE bool "Kernel Function Probe (fprobe)" depends on FUNCTION_TRACER - depends on DYNAMIC_FTRACE_WITH_REGS + depends on DYNAMIC_FTRACE_WITH_REGS || DYNAMIC_FTRACE_WITH_ARGS depends on HAVE_RETHOOK select RETHOOK default n @@ -672,6 +672,7 @@ config FPROBE_EVENTS select TRACING select PROBE_EVENTS select DYNAMIC_EVENTS + depends on DYNAMIC_FTRACE_WITH_REGS default y help This allows user to add tracing events on the function entry and diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index bd1a42b23f3f..22d00c817f1a 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -2467,7 +2467,7 @@ static int __init bpf_event_init(void) fs_initcall(bpf_event_init); #endif /* CONFIG_MODULES */ =20 -#ifdef CONFIG_FPROBE +#if defined(CONFIG_FPROBE) && defined(CONFIG_DYNAMIC_FTRACE_WITH_REGS) struct bpf_kprobe_multi_link { struct bpf_link link; struct fprobe fp; @@ -2659,10 +2659,14 @@ kprobe_multi_link_prog_run(struct bpf_kprobe_multi_= link *link, =20 static int kprobe_multi_link_handler(struct fprobe *fp, unsigned long fentry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *fregs, void *data) { struct bpf_kprobe_multi_link *link; + struct pt_regs *regs =3D ftrace_get_regs(fregs); + + if (!regs) + return 0; =20 link =3D container_of(fp, struct bpf_kprobe_multi_link, fp); kprobe_multi_link_prog_run(link, get_entry_ip(fentry_ip), regs); @@ -2917,7 +2921,7 @@ int bpf_kprobe_multi_link_attach(const union bpf_attr= *attr, struct bpf_prog *pr kvfree(cookies); return err; } -#else /* !CONFIG_FPROBE */ +#else /* !CONFIG_FPROBE || !CONFIG_DYNAMIC_FTRACE_WITH_REGS */ int bpf_kprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_pr= og *prog) { return -EOPNOTSUPP; diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c index 3b21f4063258..07deb52df44a 100644 --- a/kernel/trace/fprobe.c +++ b/kernel/trace/fprobe.c @@ -46,7 +46,7 @@ static inline void __fprobe_handler(unsigned long ip, uns= igned long parent_ip, } =20 if (fp->entry_handler) - ret =3D fp->entry_handler(fp, ip, parent_ip, ftrace_get_regs(fregs), ent= ry_data); + ret =3D fp->entry_handler(fp, ip, parent_ip, fregs, entry_data); =20 /* If entry_handler returns !0, nmissed is not counted. */ if (rh) { @@ -182,7 +182,7 @@ static void fprobe_init(struct fprobe *fp) fp->ops.func =3D fprobe_kprobe_handler; else fp->ops.func =3D fprobe_handler; - fp->ops.flags |=3D FTRACE_OPS_FL_SAVE_REGS; + fp->ops.flags |=3D FTRACE_OPS_FL_SAVE_ARGS; } =20 static int fprobe_init_rethook(struct fprobe *fp, int num) diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c index 8bfe23af9c73..71bf38d698f1 100644 --- a/kernel/trace/trace_fprobe.c +++ b/kernel/trace/trace_fprobe.c @@ -320,12 +320,16 @@ NOKPROBE_SYMBOL(fexit_perf_func); #endif /* CONFIG_PERF_EVENTS */ =20 static int fentry_dispatcher(struct fprobe *fp, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *fregs, void *entry_data) { struct trace_fprobe *tf =3D container_of(fp, struct trace_fprobe, fp); + struct pt_regs *regs =3D ftrace_get_regs(fregs); int ret =3D 0; =20 + if (!regs) + return 0; + if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE)) fentry_trace_func(tf, entry_ip, regs); #ifdef CONFIG_PERF_EVENTS diff --git a/lib/test_fprobe.c b/lib/test_fprobe.c index 24de0e5ff859..ff607babba18 100644 --- a/lib/test_fprobe.c +++ b/lib/test_fprobe.c @@ -40,7 +40,7 @@ static noinline u32 fprobe_selftest_nest_target(u32 value= , u32 (*nest)(u32)) =20 static notrace int fp_entry_handler(struct fprobe *fp, unsigned long ip, unsigned long ret_ip, - struct pt_regs *regs, void *data) + struct ftrace_regs *fregs, void *data) { KUNIT_EXPECT_FALSE(current_test, preemptible()); /* This can be called on the fprobe_selftest_target and the fprobe_selfte= st_target2 */ @@ -81,7 +81,7 @@ static notrace void fp_exit_handler(struct fprobe *fp, un= signed long ip, =20 static notrace int nest_entry_handler(struct fprobe *fp, unsigned long ip, unsigned long ret_ip, - struct pt_regs *regs, void *data) + struct ftrace_regs *fregs, void *data) { KUNIT_EXPECT_FALSE(current_test, preemptible()); return 0; diff --git a/samples/fprobe/fprobe_example.c b/samples/fprobe/fprobe_exampl= e.c index 64e715e7ed11..1545a1aac616 100644 --- a/samples/fprobe/fprobe_example.c +++ b/samples/fprobe/fprobe_example.c @@ -50,7 +50,7 @@ static void show_backtrace(void) =20 static int sample_entry_handler(struct fprobe *fp, unsigned long ip, unsigned long ret_ip, - struct pt_regs *regs, void *data) + struct ftrace_regs *fregs, void *data) { if (use_trace) /* From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 364F6EE49B5 for ; Wed, 23 Aug 2023 15:18:09 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237085AbjHWPSJ (ORCPT ); Wed, 23 Aug 2023 11:18:09 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:45602 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237052AbjHWPSC (ORCPT ); Wed, 23 Aug 2023 11:18:02 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [139.178.84.217]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 2B64510F8; Wed, 23 Aug 2023 08:17:41 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id BC273646A1; Wed, 23 Aug 2023 15:16:09 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id AEE86C433C8; Wed, 23 Aug 2023 15:16:05 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803768; bh=XUnWIvG2GF1W9eYoGRGfvmiM9wn5WNYnK+ukiZ/XFnE=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=L4bTLvYAis9FElX7EW5jKDI/MZ768XYa656tzXf75Q5JWfna2O26b72HYOQ2+Fy0Z X/NYFUXeMEGwoPLeFrkySgxkc0O3dEWIiu5l3AffrhksWPQ1Ten4Apa1ozEOt2H2hy WTJJLmpMw9QcdlHPPu2TSy/6hIzod9RDfOaULiEduJkwNuw3TCGqeNU1vP/uAift+T lt3YOQIZ3ozlunqffM3qDXRj4gKY4+Y+bJZp2/n7uVGZoUWxwaVRhStob4InVYGUv4 ep1CjWxl4XJCAp8Nw+GSdHe/VqFDOLTvUgXVUx9MyOYMlREk3E+FAjMXIBXgwL/E2L H6yeu9UaMcs6Q== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 3/9] tracing: Expose ftrace_regs regardless of CONFIG_FUNCTION_TRACER Date: Thu, 24 Aug 2023 00:16:03 +0900 Message-Id: <169280376296.282662.3179053699894615088.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) In order to be able to use ftrace_regs even from features unrelated to function tracer (e.g. kretprobe), expose ftrace_regs structures and APIs even if the CONFIG_FUNCTION_TRACER=3Dn. Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Changes in v3: - arch/s390/include/asm/ftrace.h: hide ftrace_regs parts if CONFIG_FUNCTION_TRACER=3Dn to avoid conflict. - Fixed typo. - Define accessor macros only if HAVE_REGS_AND_STACK_ACCESS_API=3Dy --- arch/s390/include/asm/ftrace.h | 4 +++ include/linux/ftrace.h | 51 +++++++++++++++++++++++-------------= ---- 2 files changed, 33 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/ftrace.h b/arch/s390/include/asm/ftrace.h index e5c5cb1207e2..74cd2e8bf660 100644 --- a/arch/s390/include/asm/ftrace.h +++ b/arch/s390/include/asm/ftrace.h @@ -41,6 +41,8 @@ static inline unsigned long ftrace_call_adjust(unsigned l= ong addr) return addr; } =20 +#ifdef CONFIG_FUNCTION_TRACER + struct ftrace_regs { struct pt_regs regs; }; @@ -80,6 +82,8 @@ ftrace_regs_set_instruction_pointer(struct ftrace_regs *f= regs, #define ftrace_regs_query_register_offset(name) \ regs_query_register_offset(name) =20 +#endif /* CONFIG_FUNCTION_TRACER */ + #ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS /* * When an ftrace registered caller is tracing a function that is diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index aad9cf8876b5..fe335d861f08 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -112,24 +112,44 @@ static inline int ftrace_mod_get_kallsym(unsigned int= symnum, unsigned long *val } #endif =20 -#ifdef CONFIG_FUNCTION_TRACER - -extern int ftrace_enabled; - -#ifndef CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS +/* + * If the architecture doesn't support FTRACE_WITH_ARGS or disables functi= on + * tracer, define the default(pt_regs compatible) ftrace_regs. + */ +#if !defined(CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS) || !defined(CONFIG_FUNC= TION_TRACER) =20 struct ftrace_regs { struct pt_regs regs; }; #define arch_ftrace_get_regs(fregs) (&(fregs)->regs) =20 +#ifdef CONFIG_HAVE_REGS_AND_STACK_ACCESS_API + /* * ftrace_regs_set_instruction_pointer() is to be defined by the architect= ure * if to allow setting of the instruction pointer from the ftrace_regs when * HAVE_DYNAMIC_FTRACE_WITH_ARGS is set and it supports live kernel patchi= ng. */ #define ftrace_regs_set_instruction_pointer(fregs, ip) do { } while (0) -#endif /* CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS */ + +#define ftrace_regs_get_instruction_pointer(fregs) \ + instruction_pointer(ftrace_get_regs(fregs)) +#define ftrace_regs_get_argument(fregs, n) \ + regs_get_kernel_argument(ftrace_get_regs(fregs), n) +#define ftrace_regs_get_stack_pointer(fregs) \ + kernel_stack_pointer(ftrace_get_regs(fregs)) +#define ftrace_regs_return_value(fregs) \ + regs_return_value(ftrace_get_regs(fregs)) +#define ftrace_regs_set_return_value(fregs, ret) \ + regs_set_return_value(ftrace_get_regs(fregs), ret) +#define ftrace_override_function_with_return(fregs) \ + override_function_with_return(ftrace_get_regs(fregs)) +#define ftrace_regs_query_register_offset(name) \ + regs_query_register_offset(name) + +#endif /* CONFIG_HAVE_REGS_AND_STACK_ACCESS_API */ + +#endif /* !CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS || !CONFIG_FUNCTION_TRACER= */ =20 static __always_inline struct pt_regs *ftrace_get_regs(struct ftrace_regs = *fregs) { @@ -151,22 +171,9 @@ static __always_inline bool ftrace_regs_has_args(struc= t ftrace_regs *fregs) return ftrace_get_regs(fregs) !=3D NULL; } =20 -#ifndef CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS -#define ftrace_regs_get_instruction_pointer(fregs) \ - instruction_pointer(ftrace_get_regs(fregs)) -#define ftrace_regs_get_argument(fregs, n) \ - regs_get_kernel_argument(ftrace_get_regs(fregs), n) -#define ftrace_regs_get_stack_pointer(fregs) \ - kernel_stack_pointer(ftrace_get_regs(fregs)) -#define ftrace_regs_return_value(fregs) \ - regs_return_value(ftrace_get_regs(fregs)) -#define ftrace_regs_set_return_value(fregs, ret) \ - regs_set_return_value(ftrace_get_regs(fregs), ret) -#define ftrace_override_function_with_return(fregs) \ - override_function_with_return(ftrace_get_regs(fregs)) -#define ftrace_regs_query_register_offset(name) \ - regs_query_register_offset(name) -#endif +#ifdef CONFIG_FUNCTION_TRACER + +extern int ftrace_enabled; =20 typedef void (*ftrace_func_t)(unsigned long ip, unsigned long parent_ip, struct ftrace_ops *op, struct ftrace_regs *fregs); From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id DF8A5EE49A3 for ; Wed, 23 Aug 2023 15:17:44 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237001AbjHWPRo (ORCPT ); Wed, 23 Aug 2023 11:17:44 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:48932 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S236948AbjHWPRi (ORCPT ); Wed, 23 Aug 2023 11:17:38 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 4BD2510D0; Wed, 23 Aug 2023 08:17:15 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 617F364C08; Wed, 23 Aug 2023 15:16:21 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 2D49CC433C8; Wed, 23 Aug 2023 15:16:16 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803780; bh=cpfpfNevjBqZFBdnMIu7QDAs5/Kw/wypH4ND8Zd1+xI=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=IcBUjtyJ9jSoV9RtJfpkX5+XABoLJ7Qo/CgxH53U5YOz7iIac3ywTRSZMFc2ZuORl 1bmfZlSQMiKdpN8iJtiic6iV247cw/yCl3/GGm5+EY8Kw7lrGkqFomAmXeXYCPIxEE 7LO04i7C4Bvv7fVfqmrZZKJ4uil3/TNpv4lNRZhwgjcrHuqrh7z1kJJHoxag2Dyi1L PmGM1tGwcV8OxNRtFlwZYLT1SX6oC6PCI0ZWiUr5jIbEj2IohaBIvKXkxEUBJOyiBt AAkz0OfooaADROA4EUMBtIckGWRlRspaYCTFaZ5jAGlTBQqYEN30Q40OzyQUo/VbKm KfiFQd0AZHwgg== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 4/9] fprobe: rethook: Use ftrace_regs in fprobe exit handler and rethook Date: Thu, 24 Aug 2023 00:16:14 +0900 Message-Id: <169280377434.282662.7610009313268953247.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Change the fprobe exit handler and rethook to use ftrace_regs data structure instead of pt_regs. This also introduce HAVE_PT_REGS_TO_FTRACE_REGS_CAST which means the ftrace_regs's memory layout is equal to the pt_regs so that those are able to cast. Only if it is enabled, kretprobe will use rethook since kretprobe requires pt_regs for backward compatibility. This means the archs which currently implement rethook for kretprobes needs= to set that flag and it must ensure struct ftrace_regs is same as pt_regs. If not, it must be either disabling kretprobe or implementing kretprobe trampoline separately from rethook trampoline. Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Changes in v3: - Config rename to HAVE_PT_REGS_TO_FTRACE_REGS_CAST - Use ftrace_regs_* APIs instead of ftrace_get_regs(). Changes in v4: - Add static_assert() to ensure at least the size of pt_regs and ftrace_regs are same if HAVE_PT_REGS_TO_FTRACE_REGS_CAST=3Dy. --- arch/Kconfig | 1 + arch/loongarch/Kconfig | 1 + arch/s390/Kconfig | 1 + arch/x86/Kconfig | 1 + arch/x86/kernel/rethook.c | 13 +++++++------ include/linux/fprobe.h | 2 +- include/linux/ftrace.h | 6 ++++++ include/linux/rethook.h | 11 ++++++----- kernel/kprobes.c | 10 ++++++++-- kernel/trace/Kconfig | 7 +++++++ kernel/trace/bpf_trace.c | 6 +++++- kernel/trace/fprobe.c | 6 +++--- kernel/trace/rethook.c | 16 ++++++++-------- kernel/trace/trace_fprobe.c | 6 +++++- lib/test_fprobe.c | 6 +++--- samples/fprobe/fprobe_example.c | 2 +- 16 files changed, 64 insertions(+), 31 deletions(-) diff --git a/arch/Kconfig b/arch/Kconfig index aff2746c8af2..e41a270c30bb 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -201,6 +201,7 @@ config KRETPROBE_ON_RETHOOK def_bool y depends on HAVE_RETHOOK depends on KRETPROBES + depends on HAVE_PT_REGS_TO_FTRACE_REGS_CAST || !HAVE_DYNAMIC_FTRACE_WITH_= ARGS select RETHOOK =20 config USER_RETURN_NOTIFIER diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig index e71d5bf2cee0..33c3a4598ae0 100644 --- a/arch/loongarch/Kconfig +++ b/arch/loongarch/Kconfig @@ -103,6 +103,7 @@ config LOONGARCH select HAVE_DMA_CONTIGUOUS select HAVE_DYNAMIC_FTRACE select HAVE_DYNAMIC_FTRACE_WITH_ARGS + select HAVE_PT_REGS_TO_FTRACE_REGS_CAST select HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS select HAVE_DYNAMIC_FTRACE_WITH_REGS select HAVE_EBPF_JIT diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index 5b39918b7042..ef06c3c2b06d 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -165,6 +165,7 @@ config S390 select HAVE_DMA_CONTIGUOUS select HAVE_DYNAMIC_FTRACE select HAVE_DYNAMIC_FTRACE_WITH_ARGS + select HAVE_PT_REGS_TO_FTRACE_REGS_CAST select HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS select HAVE_DYNAMIC_FTRACE_WITH_REGS select HAVE_EBPF_JIT if HAVE_MARCH_Z196_FEATURES diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index e36261b4ea14..7c1f3194e209 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -207,6 +207,7 @@ config X86 select HAVE_DYNAMIC_FTRACE select HAVE_DYNAMIC_FTRACE_WITH_REGS select HAVE_DYNAMIC_FTRACE_WITH_ARGS if X86_64 + select HAVE_PT_REGS_TO_FTRACE_REGS_CAST if X86_64 select HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS select HAVE_SAMPLE_FTRACE_DIRECT if X86_64 select HAVE_SAMPLE_FTRACE_DIRECT_MULTI if X86_64 diff --git a/arch/x86/kernel/rethook.c b/arch/x86/kernel/rethook.c index 8a1c0111ae79..d714d0276c93 100644 --- a/arch/x86/kernel/rethook.c +++ b/arch/x86/kernel/rethook.c @@ -83,7 +83,8 @@ __used __visible void arch_rethook_trampoline_callback(st= ruct pt_regs *regs) * arch_rethook_fixup_return() which called from this * rethook_trampoline_handler(). */ - rethook_trampoline_handler(regs, (unsigned long)frame_pointer); + rethook_trampoline_handler((struct ftrace_regs *)regs, + (unsigned long)frame_pointer); =20 /* * Copy FLAGS to 'pt_regs::ss' so that arch_rethook_trapmoline() @@ -104,22 +105,22 @@ NOKPROBE_SYMBOL(arch_rethook_trampoline_callback); STACK_FRAME_NON_STANDARD_FP(arch_rethook_trampoline); =20 /* This is called from rethook_trampoline_handler(). */ -void arch_rethook_fixup_return(struct pt_regs *regs, +void arch_rethook_fixup_return(struct ftrace_regs *fregs, unsigned long correct_ret_addr) { - unsigned long *frame_pointer =3D (void *)(regs + 1); + unsigned long *frame_pointer =3D (void *)(fregs + 1); =20 /* Replace fake return address with real one. */ *frame_pointer =3D correct_ret_addr; } NOKPROBE_SYMBOL(arch_rethook_fixup_return); =20 -void arch_rethook_prepare(struct rethook_node *rh, struct pt_regs *regs, b= ool mcount) +void arch_rethook_prepare(struct rethook_node *rh, struct ftrace_regs *fre= gs, bool mcount) { - unsigned long *stack =3D (unsigned long *)regs->sp; + unsigned long *stack =3D (unsigned long *)ftrace_regs_get_stack_pointer(f= regs); =20 rh->ret_addr =3D stack[0]; - rh->frame =3D regs->sp; + rh->frame =3D (unsigned long)stack; =20 /* Replace the return addr with trampoline addr */ stack[0] =3D (unsigned long) arch_rethook_trampoline; diff --git a/include/linux/fprobe.h b/include/linux/fprobe.h index 36c0595f7b93..b9c0c216dedb 100644 --- a/include/linux/fprobe.h +++ b/include/linux/fprobe.h @@ -38,7 +38,7 @@ struct fprobe { unsigned long ret_ip, struct ftrace_regs *regs, void *entry_data); void (*exit_handler)(struct fprobe *fp, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *regs, void *entry_data); }; =20 diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index fe335d861f08..c0a42d0860b8 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -151,6 +151,12 @@ struct ftrace_regs { =20 #endif /* !CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS || !CONFIG_FUNCTION_TRACER= */ =20 +#ifdef CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST + +static_assert(sizeof(struct pt_regs) =3D=3D sizeof(struct ftrace_regs)); + +#endif /* CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST */ + static __always_inline struct pt_regs *ftrace_get_regs(struct ftrace_regs = *fregs) { if (!fregs) diff --git a/include/linux/rethook.h b/include/linux/rethook.h index 26b6f3c81a76..138d64c8b67b 100644 --- a/include/linux/rethook.h +++ b/include/linux/rethook.h @@ -7,6 +7,7 @@ =20 #include #include +#include #include #include #include @@ -14,7 +15,7 @@ =20 struct rethook_node; =20 -typedef void (*rethook_handler_t) (struct rethook_node *, void *, unsigned= long, struct pt_regs *); +typedef void (*rethook_handler_t) (struct rethook_node *, void *, unsigned= long, struct ftrace_regs *); =20 /** * struct rethook - The rethook management data structure. @@ -64,12 +65,12 @@ void rethook_free(struct rethook *rh); void rethook_add_node(struct rethook *rh, struct rethook_node *node); struct rethook_node *rethook_try_get(struct rethook *rh); void rethook_recycle(struct rethook_node *node); -void rethook_hook(struct rethook_node *node, struct pt_regs *regs, bool mc= ount); +void rethook_hook(struct rethook_node *node, struct ftrace_regs *regs, boo= l mcount); unsigned long rethook_find_ret_addr(struct task_struct *tsk, unsigned long= frame, struct llist_node **cur); =20 /* Arch dependent code must implement arch_* and trampoline code */ -void arch_rethook_prepare(struct rethook_node *node, struct pt_regs *regs,= bool mcount); +void arch_rethook_prepare(struct rethook_node *node, struct ftrace_regs *r= egs, bool mcount); void arch_rethook_trampoline(void); =20 /** @@ -84,11 +85,11 @@ static inline bool is_rethook_trampoline(unsigned long = addr) } =20 /* If the architecture needs to fixup the return address, implement it. */ -void arch_rethook_fixup_return(struct pt_regs *regs, +void arch_rethook_fixup_return(struct ftrace_regs *regs, unsigned long correct_ret_addr); =20 /* Generic trampoline handler, arch code must prepare asm stub */ -unsigned long rethook_trampoline_handler(struct pt_regs *regs, +unsigned long rethook_trampoline_handler(struct ftrace_regs *regs, unsigned long frame); =20 #ifdef CONFIG_RETHOOK diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 0c6185aefaef..821dff656149 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -2132,7 +2132,11 @@ static int pre_handler_kretprobe(struct kprobe *p, s= truct pt_regs *regs) if (rp->entry_handler && rp->entry_handler(ri, regs)) rethook_recycle(rhn); else - rethook_hook(rhn, regs, kprobe_ftrace(p)); + /* + * We can cast pt_regs to ftrace_regs because this depends on + * HAVE_PT_REGS_TO_FTRACE_REGS_CAST. + */ + rethook_hook(rhn, (struct ftrace_regs *)regs, kprobe_ftrace(p)); =20 return 0; } @@ -2140,9 +2144,11 @@ NOKPROBE_SYMBOL(pre_handler_kretprobe); =20 static void kretprobe_rethook_handler(struct rethook_node *rh, void *data, unsigned long ret_addr, - struct pt_regs *regs) + struct ftrace_regs *fregs) { struct kretprobe *rp =3D (struct kretprobe *)data; + /* Ditto, this depends on HAVE_PT_REGS_TO_FTRACE_REGS_CAST. */ + struct pt_regs *regs =3D (struct pt_regs *)fregs; struct kretprobe_instance *ri; struct kprobe_ctlblk *kcb; =20 diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 976fd594b446..d56304276318 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -57,6 +57,13 @@ config HAVE_DYNAMIC_FTRACE_WITH_ARGS This allows for use of ftrace_regs_get_argument() and ftrace_regs_get_stack_pointer(). =20 +config HAVE_PT_REGS_TO_FTRACE_REGS_CAST + bool + help + If this is set, the memory layout of the ftrace_regs data structure + is the same as the pt_regs. So the pt_regs is possible to be casted + to ftrace_regs. + config HAVE_DYNAMIC_FTRACE_NO_PATCHABLE bool help diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index 22d00c817f1a..c4d57c7cdc7c 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -2675,10 +2675,14 @@ kprobe_multi_link_handler(struct fprobe *fp, unsign= ed long fentry_ip, =20 static void kprobe_multi_link_exit_handler(struct fprobe *fp, unsigned long fentry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *fregs, void *data) { struct bpf_kprobe_multi_link *link; + struct pt_regs *regs =3D ftrace_get_regs(fregs); + + if (!regs) + return; =20 link =3D container_of(fp, struct bpf_kprobe_multi_link, fp); kprobe_multi_link_prog_run(link, get_entry_ip(fentry_ip), regs); diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c index 07deb52df44a..dfddc7e8424e 100644 --- a/kernel/trace/fprobe.c +++ b/kernel/trace/fprobe.c @@ -53,7 +53,7 @@ static inline void __fprobe_handler(unsigned long ip, uns= igned long parent_ip, if (ret) rethook_recycle(rh); else - rethook_hook(rh, ftrace_get_regs(fregs), true); + rethook_hook(rh, fregs, true); } } =20 @@ -120,7 +120,7 @@ static void fprobe_kprobe_handler(unsigned long ip, uns= igned long parent_ip, } =20 static void fprobe_exit_handler(struct rethook_node *rh, void *data, - unsigned long ret_ip, struct pt_regs *regs) + unsigned long ret_ip, struct ftrace_regs *fregs) { struct fprobe *fp =3D (struct fprobe *)data; struct fprobe_rethook_node *fpr; @@ -141,7 +141,7 @@ static void fprobe_exit_handler(struct rethook_node *rh= , void *data, return; } =20 - fp->exit_handler(fp, fpr->entry_ip, ret_ip, regs, + fp->exit_handler(fp, fpr->entry_ip, ret_ip, fregs, fp->entry_data_size ? (void *)fpr->data : NULL); ftrace_test_recursion_unlock(bit); } diff --git a/kernel/trace/rethook.c b/kernel/trace/rethook.c index 5eb9b598f4e9..7c5cf9d5910c 100644 --- a/kernel/trace/rethook.c +++ b/kernel/trace/rethook.c @@ -189,7 +189,7 @@ NOKPROBE_SYMBOL(rethook_try_get); /** * rethook_hook() - Hook the current function return. * @node: The struct rethook node to hook the function return. - * @regs: The struct pt_regs for the function entry. + * @fregs: The struct ftrace_regs for the function entry. * @mcount: True if this is called from mcount(ftrace) context. * * Hook the current running function return. This must be called when the @@ -199,9 +199,9 @@ NOKPROBE_SYMBOL(rethook_try_get); * from the real function entry (e.g. kprobes) @mcount must be set false. * This is because the way to hook the function return depends on the cont= ext. */ -void rethook_hook(struct rethook_node *node, struct pt_regs *regs, bool mc= ount) +void rethook_hook(struct rethook_node *node, struct ftrace_regs *fregs, bo= ol mcount) { - arch_rethook_prepare(node, regs, mcount); + arch_rethook_prepare(node, fregs, mcount); __llist_add(&node->llist, ¤t->rethooks); } NOKPROBE_SYMBOL(rethook_hook); @@ -269,7 +269,7 @@ unsigned long rethook_find_ret_addr(struct task_struct = *tsk, unsigned long frame } NOKPROBE_SYMBOL(rethook_find_ret_addr); =20 -void __weak arch_rethook_fixup_return(struct pt_regs *regs, +void __weak arch_rethook_fixup_return(struct ftrace_regs *fregs, unsigned long correct_ret_addr) { /* @@ -281,7 +281,7 @@ void __weak arch_rethook_fixup_return(struct pt_regs *r= egs, } =20 /* This function will be called from each arch-defined trampoline. */ -unsigned long rethook_trampoline_handler(struct pt_regs *regs, +unsigned long rethook_trampoline_handler(struct ftrace_regs *fregs, unsigned long frame) { struct llist_node *first, *node =3D NULL; @@ -295,7 +295,7 @@ unsigned long rethook_trampoline_handler(struct pt_regs= *regs, BUG_ON(1); } =20 - instruction_pointer_set(regs, correct_ret_addr); + ftrace_regs_set_instruction_pointer(fregs, correct_ret_addr); =20 /* * These loops must be protected from rethook_free_rcu() because those @@ -315,7 +315,7 @@ unsigned long rethook_trampoline_handler(struct pt_regs= *regs, handler =3D READ_ONCE(rhn->rethook->handler); if (handler) handler(rhn, rhn->rethook->data, - correct_ret_addr, regs); + correct_ret_addr, fregs); =20 if (first =3D=3D node) break; @@ -323,7 +323,7 @@ unsigned long rethook_trampoline_handler(struct pt_regs= *regs, } =20 /* Fixup registers for returning to correct address. */ - arch_rethook_fixup_return(regs, correct_ret_addr); + arch_rethook_fixup_return(fregs, correct_ret_addr); =20 /* Unlink used shadow stack */ first =3D current->rethooks.first; diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c index 71bf38d698f1..c60d0d9f1a95 100644 --- a/kernel/trace/trace_fprobe.c +++ b/kernel/trace/trace_fprobe.c @@ -341,10 +341,14 @@ static int fentry_dispatcher(struct fprobe *fp, unsig= ned long entry_ip, NOKPROBE_SYMBOL(fentry_dispatcher); =20 static void fexit_dispatcher(struct fprobe *fp, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *fregs, void *entry_data) { struct trace_fprobe *tf =3D container_of(fp, struct trace_fprobe, fp); + struct pt_regs *regs =3D ftrace_get_regs(fregs); + + if (!regs) + return; =20 if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE)) fexit_trace_func(tf, entry_ip, ret_ip, regs); diff --git a/lib/test_fprobe.c b/lib/test_fprobe.c index ff607babba18..d1e80653bf0c 100644 --- a/lib/test_fprobe.c +++ b/lib/test_fprobe.c @@ -59,9 +59,9 @@ static notrace int fp_entry_handler(struct fprobe *fp, un= signed long ip, =20 static notrace void fp_exit_handler(struct fprobe *fp, unsigned long ip, unsigned long ret_ip, - struct pt_regs *regs, void *data) + struct ftrace_regs *fregs, void *data) { - unsigned long ret =3D regs_return_value(regs); + unsigned long ret =3D ftrace_regs_return_value(fregs); =20 KUNIT_EXPECT_FALSE(current_test, preemptible()); if (ip !=3D target_ip) { @@ -89,7 +89,7 @@ static notrace int nest_entry_handler(struct fprobe *fp, = unsigned long ip, =20 static notrace void nest_exit_handler(struct fprobe *fp, unsigned long ip, unsigned long ret_ip, - struct pt_regs *regs, void *data) + struct ftrace_regs *fregs, void *data) { KUNIT_EXPECT_FALSE(current_test, preemptible()); KUNIT_EXPECT_EQ(current_test, ip, target_nest_ip); diff --git a/samples/fprobe/fprobe_example.c b/samples/fprobe/fprobe_exampl= e.c index 1545a1aac616..d476d1f07538 100644 --- a/samples/fprobe/fprobe_example.c +++ b/samples/fprobe/fprobe_example.c @@ -67,7 +67,7 @@ static int sample_entry_handler(struct fprobe *fp, unsign= ed long ip, } =20 static void sample_exit_handler(struct fprobe *fp, unsigned long ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *regs, void *data) { unsigned long rip =3D ret_ip; From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id AA098EE49B0 for ; Wed, 23 Aug 2023 15:18:13 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237093AbjHWPSN (ORCPT ); Wed, 23 Aug 2023 11:18:13 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:45736 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237054AbjHWPSC (ORCPT ); Wed, 23 Aug 2023 11:18:02 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [139.178.84.217]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 79F40172C; Wed, 23 Aug 2023 08:17:41 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id CFCF565204; Wed, 23 Aug 2023 15:16:32 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 861ECC433C8; Wed, 23 Aug 2023 15:16:28 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803792; bh=auDMVayuMUnIghRCjZjrPWMDbCTvhrmJ85LBeBWaOGI=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=idnQDC5l9kLB2hyhfoXnPf1loh2XKxbpGcY8DJZ5SLBavLSuj44HdsBXOcjDpDwQ5 ZI5t0CfuyBLSm5i5uZSoIBAuXpp63LOfObzv1TS2ycufkhRNqkQGp1OM3IBRPXbpRy ixQaeLTcPKpLAdKdCD8SR3ECsMM80aEYGhKyHF0L7BMUypixqQGRXnWhbryp3t9ep7 IITfB5m6Z/FrWF52KlTtKT1mL/EF4t2Yqn/p6upUmw+yAMo5s0HdcMnDFT16+PeVnK 1H6m654UZlKF8iRVbA5k1+AvunRXEwP0WpU7E2DhXWFOlocJIPGHit1xf9uZXF1Bzu 8hFXdKjx12CHg== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 5/9] ftrace: Add ftrace_partial_regs() for converting ftrace_regs to pt_regs Date: Thu, 24 Aug 2023 00:16:26 +0900 Message-Id: <169280378611.282662.4078983611827223131.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Add ftrace_partial_regs() which converts the ftrace_regs to pt_regs. If the architecture defines its own ftrace_regs, this copies partial registers to pt_regs and returns it. If not, ftrace_regs is the same as pt_regs and ftrace_partial_regs() will return ftrace_regs::regs. Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Changes in v3: - Fix to use pt_regs::regs instead of x. - Return ftrace_regs::regs forcibly if HAVE_PT_REGS_COMPAT_FTRACE_REGS=3D= y. - Fix typo. - Fix to copy correct registers to the pt_regs on arm64. Changes in v4: - Change the patch order in the series so that fprobe event can use this. --- arch/arm64/include/asm/ftrace.h | 11 +++++++++++ include/linux/ftrace.h | 17 +++++++++++++++++ 2 files changed, 28 insertions(+) diff --git a/arch/arm64/include/asm/ftrace.h b/arch/arm64/include/asm/ftrac= e.h index ab158196480c..5ad24f315d52 100644 --- a/arch/arm64/include/asm/ftrace.h +++ b/arch/arm64/include/asm/ftrace.h @@ -137,6 +137,17 @@ ftrace_override_function_with_return(struct ftrace_reg= s *fregs) fregs->pc =3D fregs->lr; } =20 +static __always_inline struct pt_regs * +ftrace_partial_regs(const struct ftrace_regs *fregs, struct pt_regs *regs) +{ + memcpy(regs->regs, fregs->regs, sizeof(u64) * 9); + regs->sp =3D fregs->sp; + regs->pc =3D fregs->pc; + regs->regs[29] =3D fregs->fp; + regs->regs[30] =3D fregs->lr; + return regs; +} + int ftrace_regs_query_register_offset(const char *name); =20 int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec); diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index c0a42d0860b8..a6ed2aa71efc 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -165,6 +165,23 @@ static __always_inline struct pt_regs *ftrace_get_regs= (struct ftrace_regs *fregs return arch_ftrace_get_regs(fregs); } =20 +#if !defined(CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS) || \ + defined(CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST) + +static __always_inline struct pt_regs * +ftrace_partial_regs(struct ftrace_regs *fregs, struct pt_regs *regs) +{ + /* + * If CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST=3Dy, ftrace_regs memory + * layout is the same as pt_regs. So always returns that address. + * Since arch_ftrace_get_regs() will check some members and may return + * NULL, we can not use it. + */ + return &fregs->regs; +} + +#endif /* !CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS || CONFIG_HAVE_PT_REGS_TO_= FTRACE_REGS_CAST */ + /* * When true, the ftrace_regs_{get,set}_*() functions may be used on fregs. * Note: this can be true even when ftrace_get_regs() cannot provide a pt_= regs. From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 72BABEE4993 for ; Wed, 23 Aug 2023 15:17:43 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S236983AbjHWPRl (ORCPT ); Wed, 23 Aug 2023 11:17:41 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:48862 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S236912AbjHWPRi (ORCPT ); Wed, 23 Aug 2023 11:17:38 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 5709710F0; Wed, 23 Aug 2023 08:17:16 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id A2F2F645F7; Wed, 23 Aug 2023 15:16:42 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 8675CC433C8; Wed, 23 Aug 2023 15:16:39 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803801; bh=NSihKHbTD4Jn9DldCU6pu28QTieR1+AN7Z/GsZZggpU=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=J27RbJJkXKeecA2nGKIjPkL2Ggsmo/Q0ci1RFtlGoi1twA6JhjWgO3/Eot4V/vULe oysmPNgWYYzHFjVDy5fTTc1GgE+wwExYeHdjlQbkmAoySWc8cXeFwzC6RbIfRcA922 0l1WUSu2hTywu7/vyASLrVhSM2JHojNa4/OGijx3DccLvcAasdR2GVgYkLzesTVCTe 8cpuq8+0eL6hn+Uz8W0pIb3TRrzOs7QG8PfQPt3VXiTCRwS/RbI8rYI+7W4u4HIDXI RXoIlbbPhl54/7yRCnP5lyQdzv4SBGBNP2Lxx8jcbfC0CzuDOlJdjh6013CNp3KbDK cxbEjp31yTvQw== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 6/9] tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH_ARGS Date: Thu, 24 Aug 2023 00:16:37 +0900 Message-Id: <169280379741.282662.12221517584561036597.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Allow fprobe events to be enabled with CONFIG_DYNAMIC_FTRACE_WITH_ARGS. With this change, fprobe events mostly use ftrace_regs instead of pt_regs. Note that if the arch doesn't enable HAVE_PT_REGS_COMPAT_FTRACE_REGS, fprobe events will not be able to be used from perf. Signed-off-by: Masami Hiramatsu (Google) --- Changes in v3: - introduce ftrace_regs_get_kernel_stack_nth(). - fix typo. Changes in v4: - Use per-cpu pt_regs stack and ftrace_partial_regs() for perf event. --- include/linux/ftrace.h | 15 ++++ kernel/trace/Kconfig | 1=20 kernel/trace/trace_fprobe.c | 135 ++++++++++++++++++++++++++++-------= ---- kernel/trace/trace_probe_tmpl.h | 2 - 4 files changed, 112 insertions(+), 41 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index a6ed2aa71efc..fb0f87d19d35 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -194,6 +194,21 @@ static __always_inline bool ftrace_regs_has_args(struc= t ftrace_regs *fregs) return ftrace_get_regs(fregs) !=3D NULL; } =20 +#ifdef CONFIG_HAVE_REGS_AND_STACK_ACCESS_API +static __always_inline unsigned long +ftrace_regs_get_kernel_stack_nth(struct ftrace_regs *fregs, unsigned int n= th) +{ + unsigned long *stackp; + + stackp =3D (unsigned long *)ftrace_regs_get_stack_pointer(fregs); + if (((unsigned long)(stackp + nth) & ~(THREAD_SIZE - 1)) =3D=3D + ((unsigned long)stackp & ~(THREAD_SIZE - 1))) + return *(stackp + nth); + + return 0; +} +#endif /* CONFIG_HAVE_REGS_AND_STACK_ACCESS_API */ + #ifdef CONFIG_FUNCTION_TRACER =20 extern int ftrace_enabled; diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index d56304276318..6fb4ecf8767d 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -679,7 +679,6 @@ config FPROBE_EVENTS select TRACING select PROBE_EVENTS select DYNAMIC_EVENTS - depends on DYNAMIC_FTRACE_WITH_REGS default y help This allows user to add tracing events on the function entry and diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c index c60d0d9f1a95..90ad28260a9f 100644 --- a/kernel/trace/trace_fprobe.c +++ b/kernel/trace/trace_fprobe.c @@ -132,7 +132,7 @@ static int process_fetch_insn(struct fetch_insn *code, void *rec, void *dest, void *base) { - struct pt_regs *regs =3D rec; + struct ftrace_regs *fregs =3D rec; unsigned long val; int ret; =20 @@ -140,17 +140,17 @@ process_fetch_insn(struct fetch_insn *code, void *rec= , void *dest, /* 1st stage: get value from context */ switch (code->op) { case FETCH_OP_STACK: - val =3D regs_get_kernel_stack_nth(regs, code->param); + val =3D ftrace_regs_get_kernel_stack_nth(fregs, code->param); break; case FETCH_OP_STACKP: - val =3D kernel_stack_pointer(regs); + val =3D ftrace_regs_get_stack_pointer(fregs); break; case FETCH_OP_RETVAL: - val =3D regs_return_value(regs); + val =3D ftrace_regs_return_value(fregs); break; #ifdef CONFIG_HAVE_FUNCTION_ARG_ACCESS_API case FETCH_OP_ARG: - val =3D regs_get_kernel_argument(regs, code->param); + val =3D ftrace_regs_get_argument(fregs, code->param); break; #endif case FETCH_NOP_SYMBOL: /* Ignore a place holder */ @@ -170,7 +170,7 @@ NOKPROBE_SYMBOL(process_fetch_insn) /* function entry handler */ static nokprobe_inline void __fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip, - struct pt_regs *regs, + struct ftrace_regs *fregs, struct trace_event_file *trace_file) { struct fentry_trace_entry_head *entry; @@ -184,36 +184,36 @@ __fentry_trace_func(struct trace_fprobe *tf, unsigned= long entry_ip, if (trace_trigger_soft_disabled(trace_file)) return; =20 - dsize =3D __get_data_size(&tf->tp, regs); + dsize =3D __get_data_size(&tf->tp, fregs); =20 entry =3D trace_event_buffer_reserve(&fbuffer, trace_file, sizeof(*entry) + tf->tp.size + dsize); if (!entry) return; =20 - fbuffer.regs =3D regs; + fbuffer.regs =3D ftrace_get_regs(fregs); entry =3D fbuffer.entry =3D ring_buffer_event_data(fbuffer.event); entry->ip =3D entry_ip; - store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize); + store_trace_args(&entry[1], &tf->tp, fregs, sizeof(*entry), dsize); =20 trace_event_buffer_commit(&fbuffer); } =20 static void fentry_trace_func(struct trace_fprobe *tf, unsigned long entry_ip, - struct pt_regs *regs) + struct ftrace_regs *fregs) { struct event_file_link *link; =20 trace_probe_for_each_link_rcu(link, &tf->tp) - __fentry_trace_func(tf, entry_ip, regs, link->file); + __fentry_trace_func(tf, entry_ip, fregs, link->file); } NOKPROBE_SYMBOL(fentry_trace_func); =20 /* Kretprobe handler */ static nokprobe_inline void __fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs, + unsigned long ret_ip, struct ftrace_regs *fregs, struct trace_event_file *trace_file) { struct fexit_trace_entry_head *entry; @@ -227,94 +227,157 @@ __fexit_trace_func(struct trace_fprobe *tf, unsigned= long entry_ip, if (trace_trigger_soft_disabled(trace_file)) return; =20 - dsize =3D __get_data_size(&tf->tp, regs); + dsize =3D __get_data_size(&tf->tp, fregs); =20 entry =3D trace_event_buffer_reserve(&fbuffer, trace_file, sizeof(*entry) + tf->tp.size + dsize); if (!entry) return; =20 - fbuffer.regs =3D regs; + fbuffer.regs =3D ftrace_get_regs(fregs); entry =3D fbuffer.entry =3D ring_buffer_event_data(fbuffer.event); entry->func =3D entry_ip; entry->ret_ip =3D ret_ip; - store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize); + store_trace_args(&entry[1], &tf->tp, fregs, sizeof(*entry), dsize); =20 trace_event_buffer_commit(&fbuffer); } =20 static void fexit_trace_func(struct trace_fprobe *tf, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs) + unsigned long ret_ip, struct ftrace_regs *fregs) { struct event_file_link *link; =20 trace_probe_for_each_link_rcu(link, &tf->tp) - __fexit_trace_func(tf, entry_ip, ret_ip, regs, link->file); + __fexit_trace_func(tf, entry_ip, ret_ip, fregs, link->file); } NOKPROBE_SYMBOL(fexit_trace_func); =20 #ifdef CONFIG_PERF_EVENTS =20 +#if !defined(CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS) || \ + defined(CONFIG_HAVE_PT_REGS_TO_FTRACE_REGS_CAST) + +static __always_inline +struct pt_regs *perf_fprobe_partial_regs(struct ftrace_regs *fregs) +{ + /* See include/linux/ftrace.h, this returns &fregs->regs */ + return ftrace_partial_regs(fregs, NULL); +} + +#define perf_fprobe_return_regs(regs) do {} while (0) + +#else /* CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS && !CONFIG_HAVE_PT_REGS_TO_F= TRACE_REGS_CAST */ + +/* Since fprobe handlers can be nested, pt_regs buffer need to be a stack = */ +#define PERF_FPROBE_REGS_MAX 4 + +struct pt_regs_stack { + struct pt_regs regs[PERF_FPROBE_REGS_MAX]; + int idx; +}; + +static DEFINE_PER_CPU(struct pt_regs_stack, perf_fprobe_regs); + +static __always_inline +struct pt_regs *perf_fprobe_partial_regs(struct ftrace_regs *fregs) +{ + struct pt_regs_stack *stack =3D this_cpu_ptr(&perf_fprobe_regs); + struct pt_regs *regs; + + if (stack->idx < PERF_FPROBE_REGS_MAX) { + regs =3D stack->regs[stack->idx++]; + return ftrace_partial_regs(fregs, regs); + } + return NULL; +} + +static __always_inline void perf_fprobe_return_regs(struct pt_regs *regs) +{ + struct pt_regs_stack *stack =3D this_cpu_ptr(&perf_fprobe_regs); + + if (WARN_ON_ONCE(regs !=3D stack->regs[stack->idx])) + return; + + --stack->idx; +} + +#endif /* !CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS || CONFIG_HAVE_PT_REGS_TO_= FTRACE_REGS_CAST */ + static int fentry_perf_func(struct trace_fprobe *tf, unsigned long entry_i= p, - struct pt_regs *regs) + struct ftrace_regs *fregs) { struct trace_event_call *call =3D trace_probe_event_call(&tf->tp); struct fentry_trace_entry_head *entry; struct hlist_head *head; int size, __size, dsize; + struct pt_regs *regs; int rctx; =20 + regs =3D perf_fprobe_partial_regs(fregs); + if (!regs) + return -EINVAL; + head =3D this_cpu_ptr(call->perf_events); if (hlist_empty(head)) - return 0; + goto out; =20 - dsize =3D __get_data_size(&tf->tp, regs); + dsize =3D __get_data_size(&tf->tp, fregs); __size =3D sizeof(*entry) + tf->tp.size + dsize; size =3D ALIGN(__size + sizeof(u32), sizeof(u64)); size -=3D sizeof(u32); =20 entry =3D perf_trace_buf_alloc(size, NULL, &rctx); if (!entry) - return 0; + goto out; =20 entry->ip =3D entry_ip; memset(&entry[1], 0, dsize); - store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize); + store_trace_args(&entry[1], &tf->tp, fregs, sizeof(*entry), dsize); perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs, head, NULL); +out: + perf_fprobe_return_regs(regs); return 0; } NOKPROBE_SYMBOL(fentry_perf_func); =20 static void fexit_perf_func(struct trace_fprobe *tf, unsigned long entry_ip, - unsigned long ret_ip, struct pt_regs *regs) + unsigned long ret_ip, struct ftrace_regs *fregs) { struct trace_event_call *call =3D trace_probe_event_call(&tf->tp); struct fexit_trace_entry_head *entry; struct hlist_head *head; int size, __size, dsize; + struct pt_regs *regs; int rctx; =20 + regs =3D perf_fprobe_partial_regs(fregs); + if (!regs) + return; + head =3D this_cpu_ptr(call->perf_events); if (hlist_empty(head)) - return; + goto out; =20 - dsize =3D __get_data_size(&tf->tp, regs); + dsize =3D __get_data_size(&tf->tp, fregs); __size =3D sizeof(*entry) + tf->tp.size + dsize; size =3D ALIGN(__size + sizeof(u32), sizeof(u64)); size -=3D sizeof(u32); =20 entry =3D perf_trace_buf_alloc(size, NULL, &rctx); if (!entry) - return; + goto out; =20 entry->func =3D entry_ip; entry->ret_ip =3D ret_ip; - store_trace_args(&entry[1], &tf->tp, regs, sizeof(*entry), dsize); + store_trace_args(&entry[1], &tf->tp, fregs, sizeof(*entry), dsize); perf_trace_buf_submit(entry, size, rctx, call->event.type, 1, regs, head, NULL); +out: + perf_fprobe_return_regs(regs); } NOKPROBE_SYMBOL(fexit_perf_func); #endif /* CONFIG_PERF_EVENTS */ @@ -324,17 +387,14 @@ static int fentry_dispatcher(struct fprobe *fp, unsig= ned long entry_ip, void *entry_data) { struct trace_fprobe *tf =3D container_of(fp, struct trace_fprobe, fp); - struct pt_regs *regs =3D ftrace_get_regs(fregs); int ret =3D 0; =20 - if (!regs) - return 0; - if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE)) - fentry_trace_func(tf, entry_ip, regs); + fentry_trace_func(tf, entry_ip, fregs); + #ifdef CONFIG_PERF_EVENTS if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE)) - ret =3D fentry_perf_func(tf, entry_ip, regs); + ret =3D fentry_perf_func(tf, entry_ip, fregs); #endif return ret; } @@ -345,16 +405,13 @@ static void fexit_dispatcher(struct fprobe *fp, unsig= ned long entry_ip, void *entry_data) { struct trace_fprobe *tf =3D container_of(fp, struct trace_fprobe, fp); - struct pt_regs *regs =3D ftrace_get_regs(fregs); - - if (!regs) - return; =20 if (trace_probe_test_flag(&tf->tp, TP_FLAG_TRACE)) - fexit_trace_func(tf, entry_ip, ret_ip, regs); + fexit_trace_func(tf, entry_ip, ret_ip, fregs); + #ifdef CONFIG_PERF_EVENTS if (trace_probe_test_flag(&tf->tp, TP_FLAG_PROFILE)) - fexit_perf_func(tf, entry_ip, ret_ip, regs); + fexit_perf_func(tf, entry_ip, ret_ip, fregs); #endif } NOKPROBE_SYMBOL(fexit_dispatcher); diff --git a/kernel/trace/trace_probe_tmpl.h b/kernel/trace/trace_probe_tmp= l.h index 3935b347f874..05445a745a07 100644 --- a/kernel/trace/trace_probe_tmpl.h +++ b/kernel/trace/trace_probe_tmpl.h @@ -232,7 +232,7 @@ process_fetch_insn_bottom(struct fetch_insn *code, unsi= gned long val, =20 /* Sum up total data length for dynamic arrays (strings) */ static nokprobe_inline int -__get_data_size(struct trace_probe *tp, struct pt_regs *regs) +__get_data_size(struct trace_probe *tp, void *regs) { struct probe_arg *arg; int i, len, ret =3D 0; From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 10BDFEE4993 for ; Wed, 23 Aug 2023 15:18:12 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233986AbjHWPSM (ORCPT ); Wed, 23 Aug 2023 11:18:12 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:45790 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237065AbjHWPSD (ORCPT ); Wed, 23 Aug 2023 11:18:03 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [139.178.84.217]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C66101738; Wed, 23 Aug 2023 08:17:43 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id BE6EB64E61; Wed, 23 Aug 2023 15:16:52 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 52F74C433C7; Wed, 23 Aug 2023 15:16:49 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803811; bh=7B4CIzVJLnLCmWiU4haDKVfyzxSXSWysR7Bdn2Pzp1I=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=H7oOGHI+6cNBySWTDApvSYjiXOwCS3lUZ70b9FAR9meElipdFe9RAQ8tbjSRohg6f afZj6zv4mFUwzLIQDPQFoZJOlzIO7uzxjy+ZXT9D0Z25yq0LiQi/aU1lAryknpLJne jpzAUIP3Fy80gUQV4Yuz504zmfBPIjVUwrvgIVZpB8rWPAtIHz1/fgSO7DRtWxSF0d 4ZHOcxSFyvW9cPgBq0wShwAyxDR/qn1xeMX9Q33JqZ/kWWDogvXhTxl9KrnoWBYp9P wi/gA5Xw1ZL0mrSqujBB5YDHwws5aF6pv85zCtctiu1xWGORsqe2NGA7QP8CkYOesK /wYQzPVf9/dIQ== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 7/9] bpf: Enable kprobe_multi feature if CONFIG_FPROBE is enabled Date: Thu, 24 Aug 2023 00:16:47 +0900 Message-Id: <169280380720.282662.17571417296398071399.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Enable kprobe_multi feature if CONFIG_FPROBE is enabled. The pt_regs is converted from ftrace_regs by ftrace_partial_regs(), thus some registers may always returns 0. But it should be enough for function entry (access arguments) and exit (access return value). Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- kernel/trace/bpf_trace.c | 22 +++++++++------------- 1 file changed, 9 insertions(+), 13 deletions(-) diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index c4d57c7cdc7c..ac9bd5e7ec27 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -2467,7 +2467,7 @@ static int __init bpf_event_init(void) fs_initcall(bpf_event_init); #endif /* CONFIG_MODULES */ =20 -#if defined(CONFIG_FPROBE) && defined(CONFIG_DYNAMIC_FTRACE_WITH_REGS) +#ifdef CONFIG_FPROBE struct bpf_kprobe_multi_link { struct bpf_link link; struct fprobe fp; @@ -2489,6 +2489,8 @@ struct user_syms { char *buf; }; =20 +static DEFINE_PER_CPU(struct pt_regs, bpf_kprobe_multi_pt_regs); + static int copy_user_syms(struct user_syms *us, unsigned long __user *usym= s, u32 cnt) { unsigned long __user usymbol; @@ -2630,13 +2632,14 @@ static u64 bpf_kprobe_multi_entry_ip(struct bpf_run= _ctx *ctx) =20 static int kprobe_multi_link_prog_run(struct bpf_kprobe_multi_link *link, - unsigned long entry_ip, struct pt_regs *regs) + unsigned long entry_ip, struct ftrace_regs *fregs) { struct bpf_kprobe_multi_run_ctx run_ctx =3D { .link =3D link, .entry_ip =3D entry_ip, }; struct bpf_run_ctx *old_run_ctx; + struct pt_regs *regs; int err; =20 if (unlikely(__this_cpu_inc_return(bpf_prog_active) !=3D 1)) { @@ -2646,6 +2649,7 @@ kprobe_multi_link_prog_run(struct bpf_kprobe_multi_li= nk *link, =20 migrate_disable(); rcu_read_lock(); + regs =3D ftrace_partial_regs(fregs, this_cpu_ptr(&bpf_kprobe_multi_pt_reg= s)); old_run_ctx =3D bpf_set_run_ctx(&run_ctx.run_ctx); err =3D bpf_prog_run(link->link.prog, regs); bpf_reset_run_ctx(old_run_ctx); @@ -2663,13 +2667,9 @@ kprobe_multi_link_handler(struct fprobe *fp, unsigne= d long fentry_ip, void *data) { struct bpf_kprobe_multi_link *link; - struct pt_regs *regs =3D ftrace_get_regs(fregs); - - if (!regs) - return 0; =20 link =3D container_of(fp, struct bpf_kprobe_multi_link, fp); - kprobe_multi_link_prog_run(link, get_entry_ip(fentry_ip), regs); + kprobe_multi_link_prog_run(link, get_entry_ip(fentry_ip), fregs); return 0; } =20 @@ -2679,13 +2679,9 @@ kprobe_multi_link_exit_handler(struct fprobe *fp, un= signed long fentry_ip, void *data) { struct bpf_kprobe_multi_link *link; - struct pt_regs *regs =3D ftrace_get_regs(fregs); - - if (!regs) - return; =20 link =3D container_of(fp, struct bpf_kprobe_multi_link, fp); - kprobe_multi_link_prog_run(link, get_entry_ip(fentry_ip), regs); + kprobe_multi_link_prog_run(link, get_entry_ip(fentry_ip), fregs); } =20 static int symbols_cmp_r(const void *a, const void *b, const void *priv) @@ -2925,7 +2921,7 @@ int bpf_kprobe_multi_link_attach(const union bpf_attr= *attr, struct bpf_prog *pr kvfree(cookies); return err; } -#else /* !CONFIG_FPROBE || !CONFIG_DYNAMIC_FTRACE_WITH_REGS */ +#else /* !CONFIG_FPROBE */ int bpf_kprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_pr= og *prog) { return -EOPNOTSUPP; From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 83889EE4993 for ; Wed, 23 Aug 2023 15:17:57 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237030AbjHWPR5 (ORCPT ); Wed, 23 Aug 2023 11:17:57 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:34626 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237015AbjHWPRt (ORCPT ); Wed, 23 Aug 2023 11:17:49 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 969A410C9; Wed, 23 Aug 2023 08:17:24 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 672EF65929; Wed, 23 Aug 2023 15:17:04 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 74894C433C8; Wed, 23 Aug 2023 15:16:59 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803823; bh=HSJ8K/iEdi2nQ2WTLzit3GrsQpQ2UPYAXTJjDkiZDzI=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=NA3LsibwzfMp3dlMzWwfKnecOG42Tqh+WNSoe5Dy90DAfrZYOo2Pf92LrJqqvQeNE J+Mn/kZXIiTZKX22TwSTkOvqwTEuTrrzRh5qJrE/TcCwwa9IM07XkXZEOXA9aX11Ad SR0VvTDzvZe/QfxMc6snhmCW0M0SqchieYdSZnJEEDjT61239XmV0+ZJn37RlZCOnK 9zXqc+4nbWaFJ4VTlsKV0j2xomPk9tdgd8Sse0XZRuzpGAbEbfUiTufxi+WvXg5+dO 73pUCpk0s7fgpdKQpQCZbYwJjjjdxtkgXpoAbyvJmQZxxJmr7J5THs60iNU5NPRlEV gh3cu51HkirJQ== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 8/9] Documentations: probes: Update fprobe document to use ftrace_regs Date: Thu, 24 Aug 2023 00:16:57 +0900 Message-Id: <169280381726.282662.9429943163047257398.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Update fprobe document so that the entry/exit handler uses ftrace_regs instead of pt_regs. Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Documentation/trace/fprobe.rst | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/Documentation/trace/fprobe.rst b/Documentation/trace/fprobe.rst index 5851a14eb893..64ef522f7a64 100644 --- a/Documentation/trace/fprobe.rst +++ b/Documentation/trace/fprobe.rst @@ -91,9 +91,9 @@ The prototype of the entry/exit callback function are as = follows: =20 .. code-block:: c =20 - int entry_callback(struct fprobe *fp, unsigned long entry_ip, unsigned lo= ng ret_ip, struct pt_regs *regs, void *entry_data); + int entry_callback(struct fprobe *fp, unsigned long entry_ip, unsigned lo= ng ret_ip, struct ftrace_regs *fregs, void *entry_data); =20 - void exit_callback(struct fprobe *fp, unsigned long entry_ip, unsigned lo= ng ret_ip, struct pt_regs *regs, void *entry_data); + void exit_callback(struct fprobe *fp, unsigned long entry_ip, unsigned lo= ng ret_ip, struct ftrace_regs *fregs, void *entry_data); =20 Note that the @entry_ip is saved at function entry and passed to exit hand= ler. If the entry callback function returns !0, the corresponding exit callback= will be cancelled. @@ -112,12 +112,10 @@ If the entry callback function returns !0, the corres= ponding exit callback will This is the return address that the traced function will return to, somewhere in the caller. This can be used at both entry and exit. =20 -@regs - This is the `pt_regs` data structure at the entry and exit. Note t= hat - the instruction pointer of @regs may be different from the @entry_= ip - in the entry_handler. If you need traced instruction pointer, you = need - to use @entry_ip. On the other hand, in the exit_handler, the inst= ruction - pointer of @regs is set to the currect return address. +@fregs + This is the `ftrace_regs` data structure at the entry and exit. No= te that + the instruction pointer of @fregs may be incorrect in entry handle= r and + exit handler, so you have to use @entry_ip and @ret_ip instead. =20 @entry_data This is a local storage to share the data between entry and exit h= andlers. From nobody Thu Sep 11 16:48:18 2025 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 59827EE49B2 for ; Wed, 23 Aug 2023 15:17:55 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237023AbjHWPRz (ORCPT ); Wed, 23 Aug 2023 11:17:55 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:34532 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237022AbjHWPRt (ORCPT ); Wed, 23 Aug 2023 11:17:49 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C921F10FB; Wed, 23 Aug 2023 08:17:24 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 649D86101C; Wed, 23 Aug 2023 15:17:15 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 5B611C433C8; Wed, 23 Aug 2023 15:17:11 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1692803834; bh=Ylc3TSXa5ifgz7bjH+Ib7u/kImYYyrGh5B2vMyFRLTg=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=QdaCZTMe1fTvB5F//p9urTbKVisS0r5vyJbMOt8UfyK47Y0TDyRf9fUAOtEaO07Hj IABla5AdXYBqAIbwFWhkJB3zElvQhMeeBgT5Dzus6a5rawSTD0NEAZZQEogg96sMue TCuw+1eTlKcrOruAulr9uXJJ9RVoVlBB2WyIMCDblgQ3COklM2O2PCfa6EFOM+1DAl +OFL09Q04N3zSTOeFdbI31vipmVr7r3Ygz2thKSxrAzaTnmMTVRThxeFN9ls1dqDMw RWBzdlwOCycGx/6g21G2nRRDYUdyIXS6UeKUrsl8rBmoRTzFzVOtz+NIgXG/mcTlyA ooPvUYw3CIoqg== From: "Masami Hiramatsu (Google)" To: Alexei Starovoitov , Steven Rostedt , Florent Revest Cc: linux-trace-kernel@vger.kernel.org, LKML , Martin KaFai Lau , bpf , Sven Schnelle , Alexei Starovoitov , Jiri Olsa , Arnaldo Carvalho de Melo , Daniel Borkmann , Alan Maguire , Mark Rutland , Peter Zijlstra , Thomas Gleixner Subject: [PATCH v4 9/9] Documentation: tracing: Add a note about argument and retval access Date: Thu, 24 Aug 2023 00:17:09 +0900 Message-Id: <169280382895.282662.14910495061790007288.stgit@devnote2> X-Mailer: git-send-email 2.34.1 In-Reply-To: <169280372795.282662.9784422934484459769.stgit@devnote2> References: <169280372795.282662.9784422934484459769.stgit@devnote2> User-Agent: StGit/0.19 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Masami Hiramatsu (Google) Add a note about the argument and return value accecss will be best effort. Depending on the type, it will be passed via stack or a pair of the registers, but $argN and $retval only support the single register access. Suggested-by: Alexei Starovoitov Signed-off-by: Masami Hiramatsu (Google) Acked-by: Florent Revest --- Documentation/trace/fprobetrace.rst | 8 ++++++-- Documentation/trace/kprobetrace.rst | 8 ++++++-- 2 files changed, 12 insertions(+), 4 deletions(-) diff --git a/Documentation/trace/fprobetrace.rst b/Documentation/trace/fpro= betrace.rst index 8e9bebcf0a2e..e35e6b18df40 100644 --- a/Documentation/trace/fprobetrace.rst +++ b/Documentation/trace/fprobetrace.rst @@ -59,8 +59,12 @@ Synopsis of fprobe-events and bitfield are supported. =20 (\*1) This is available only when BTF is enabled. - (\*2) only for the probe on function entry (offs =3D=3D 0). - (\*3) only for return probe. + (\*2) only for the probe on function entry (offs =3D=3D 0). Note, this a= rgument access + is best effort, because depending on the argument type, it may be = passed on + the stack. But this only support the arguments via registers. + (\*3) only for return probe. Note that this is also best effort. Dependi= ng on the + return value type, it might be passed via a pair of registers. But= this only + accesses one register. (\*4) this is useful for fetching a field of data structures. (\*5) "u" means user-space dereference. =20 diff --git a/Documentation/trace/kprobetrace.rst b/Documentation/trace/kpro= betrace.rst index 8a2dfee38145..bf9cecb69fc9 100644 --- a/Documentation/trace/kprobetrace.rst +++ b/Documentation/trace/kprobetrace.rst @@ -61,8 +61,12 @@ Synopsis of kprobe_events (x8/x16/x32/x64), "char", "string", "ustring", "symbol", "symstr" and bitfield are supported. =20 - (\*1) only for the probe on function entry (offs =3D=3D 0). - (\*2) only for return probe. + (\*1) only for the probe on function entry (offs =3D=3D 0). Note, this a= rgument access + is best effort, because depending on the argument type, it may be = passed on + the stack. But this only support the arguments via registers. + (\*2) only for return probe. Note that this is also best effort. Dependi= ng on the + return value type, it might be passed via a pair of registers. But= this only + accesses one register. (\*3) this is useful for fetching a field of data structures. (\*4) "u" means user-space dereference. See :ref:`user_mem_access`.