bpf.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Masami Hiramatsu <mhiramat@kernel.org>
To: Steven Rostedt <rostedt@goodmis.org>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	Ingo Molnar <mingo@kernel.org>
Cc: X86 ML <x86@kernel.org>, Masami Hiramatsu <mhiramat@kernel.org>,
	Daniel Xu <dxu@dxuuu.xyz>,
	linux-kernel@vger.kernel.org, bpf@vger.kernel.org,
	kuba@kernel.org, mingo@redhat.com, ast@kernel.org,
	Thomas Gleixner <tglx@linutronix.de>,
	Borislav Petkov <bp@alien8.de>,
	Peter Zijlstra <peterz@infradead.org>,
	kernel-team@fb.com, yhs@fb.com, linux-ia64@vger.kernel.org,
	Abhishek Sagar <sagar.abhishek@gmail.com>,
	Andrii Nakryiko <andrii.nakryiko@gmail.com>,
	Paul McKenney <paulmck@kernel.org>
Subject: [PATCH -tip v11 27/27] x86/kprobes: Fixup return address in generic trampoline handler
Date: Tue, 14 Sep 2021 23:42:51 +0900	[thread overview]
Message-ID: <163163057094.489837.9044470370440745866.stgit@devnote2> (raw)
In-Reply-To: <163163030719.489837.2236069935502195491.stgit@devnote2>

In x86, the fake return address on the stack saved by
__kretprobe_trampoline() will be replaced with the real return
address after returning from trampoline_handler(). Before fixing
the return address, the real return address can be found in the
'current->kretprobe_instances'.

However, since there is a window between updating the
'current->kretprobe_instances' and fixing the address on the stack,
if an interrupt happens at that timing and the interrupt handler
does stacktrace, it may fail to unwind because it can not get
the correct return address from 'current->kretprobe_instances'.

This will eliminate that window by fixing the return address
right before updating 'current->kretprobe_instances'.

Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Tested-by: Andrii Nakryiko <andrii@kernel.org>
---
 Changes in v9:
  - Fixes the changelog. This can eliminate the window.
  - Add more comment how it works.
 Changes in v7:
  - Add a prototype for arch_kretprobe_fixup_return()
---
 arch/x86/kernel/kprobes/core.c |   18 ++++++++++++++++--
 include/linux/kprobes.h        |    3 +++
 kernel/kprobes.c               |   11 +++++++++++
 3 files changed, 30 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
index 7e1111c19605..fce99e249d61 100644
--- a/arch/x86/kernel/kprobes/core.c
+++ b/arch/x86/kernel/kprobes/core.c
@@ -1065,6 +1065,16 @@ NOKPROBE_SYMBOL(__kretprobe_trampoline);
  */
 STACK_FRAME_NON_STANDARD_FP(__kretprobe_trampoline);
 
+/* This is called from kretprobe_trampoline_handler(). */
+void arch_kretprobe_fixup_return(struct pt_regs *regs,
+				 kprobe_opcode_t *correct_ret_addr)
+{
+	unsigned long *frame_pointer = &regs->sp + 1;
+
+	/* Replace fake return address with real one. */
+	*frame_pointer = (unsigned long)correct_ret_addr;
+}
+
 /*
  * Called from __kretprobe_trampoline
  */
@@ -1082,8 +1092,12 @@ __used __visible void trampoline_handler(struct pt_regs *regs)
 	regs->sp += sizeof(long);
 	frame_pointer = &regs->sp + 1;
 
-	/* Replace fake return address with real one. */
-	*frame_pointer = kretprobe_trampoline_handler(regs, frame_pointer);
+	/*
+	 * The return address at 'frame_pointer' is recovered by the
+	 * arch_kretprobe_fixup_return() which called from the
+	 * kretprobe_trampoline_handler().
+	 */
+	kretprobe_trampoline_handler(regs, frame_pointer);
 
 	/*
 	 * Copy FLAGS to 'pt_regs::sp' so that __kretprobe_trapmoline()
diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
index 6d47a9da1e0a..e974caf39d3e 100644
--- a/include/linux/kprobes.h
+++ b/include/linux/kprobes.h
@@ -188,6 +188,9 @@ extern void arch_prepare_kretprobe(struct kretprobe_instance *ri,
 				   struct pt_regs *regs);
 extern int arch_trampoline_kprobe(struct kprobe *p);
 
+void arch_kretprobe_fixup_return(struct pt_regs *regs,
+				 kprobe_opcode_t *correct_ret_addr);
+
 void __kretprobe_trampoline(void);
 /*
  * Since some architecture uses structured function pointer,
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index ebc587b9a346..b62af9fc3607 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -1922,6 +1922,15 @@ unsigned long kretprobe_find_ret_addr(struct task_struct *tsk, void *fp,
 }
 NOKPROBE_SYMBOL(kretprobe_find_ret_addr);
 
+void __weak arch_kretprobe_fixup_return(struct pt_regs *regs,
+					kprobe_opcode_t *correct_ret_addr)
+{
+	/*
+	 * Do nothing by default. Please fill this to update the fake return
+	 * address on the stack with the correct one on each arch if possible.
+	 */
+}
+
 unsigned long __kretprobe_trampoline_handler(struct pt_regs *regs,
 					     void *frame_pointer)
 {
@@ -1967,6 +1976,8 @@ unsigned long __kretprobe_trampoline_handler(struct pt_regs *regs,
 		first = first->next;
 	}
 
+	arch_kretprobe_fixup_return(regs, correct_ret_addr);
+
 	/* Unlink all nodes for this frame. */
 	first = current->kretprobe_instances.first;
 	current->kretprobe_instances.first = node->next;


  parent reply	other threads:[~2021-09-14 14:43 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-14 14:38 [PATCH -tip v11 00/27] kprobes: Fix stacktrace with kretprobes on x86 Masami Hiramatsu
2021-09-14 14:38 ` [PATCH -tip v11 01/27] kprobes: Do not use local variable when creating debugfs file Masami Hiramatsu
2021-09-14 14:38 ` [PATCH -tip v11 02/27] kprobes: Use helper to parse boolean input from userspace Masami Hiramatsu
2021-09-14 14:38 ` [PATCH -tip v11 03/27] kprobe: Simplify prepare_kprobe() by dropping redundant version Masami Hiramatsu
2021-09-14 14:39 ` [PATCH -tip v11 04/27] csky: ftrace: Drop duplicate implementation of arch_check_ftrace_location() Masami Hiramatsu
2021-09-14 14:39 ` [PATCH -tip v11 05/27] kprobes: Make arch_check_ftrace_location static Masami Hiramatsu
2021-09-14 14:39 ` [PATCH -tip v11 06/27] kprobes: treewide: Cleanup the error messages for kprobes Masami Hiramatsu
2021-09-14 14:39 ` [PATCH -tip v11 07/27] kprobes: Fix coding style issues Masami Hiramatsu
2021-09-14 14:39 ` [PATCH -tip v11 08/27] kprobes: Use IS_ENABLED() instead of kprobes_built_in() Masami Hiramatsu
2021-09-14 14:39 ` [PATCH -tip v11 09/27] kprobes: Add assertions for required lock Masami Hiramatsu
2021-09-14 14:40 ` [PATCH -tip v11 10/27] kprobes: treewide: Use 'kprobe_opcode_t *' for the code address in get_optimized_kprobe() Masami Hiramatsu
2021-09-14 14:40 ` [PATCH -tip v11 11/27] kprobes: Use bool type for functions which returns boolean value Masami Hiramatsu
2021-09-14 14:40 ` [PATCH -tip v11 12/27] ia64: kprobes: Fix to pass correct trampoline address to the handler Masami Hiramatsu
2021-09-14 14:40 ` [PATCH -tip v11 13/27] kprobes: treewide: Replace arch_deref_entry_point() with dereference_symbol_descriptor() Masami Hiramatsu
2021-09-14 14:40 ` [PATCH -tip v11 14/27] kprobes: treewide: Remove trampoline_address from kretprobe_trampoline_handler() Masami Hiramatsu
2021-09-14 14:40 ` [PATCH -tip v11 15/27] kprobes: treewide: Make it harder to refer kretprobe_trampoline directly Masami Hiramatsu
2021-09-14 14:41 ` [PATCH -tip v11 16/27] kprobes: Add kretprobe_find_ret_addr() for searching return address Masami Hiramatsu
2021-09-14 14:41 ` [PATCH -tip v11 17/27] objtool: Add frame-pointer-specific function ignore Masami Hiramatsu
2021-09-14 14:41 ` [PATCH -tip v11 18/27] objtool: Ignore unwind hints for ignored functions Masami Hiramatsu
2021-09-14 14:41 ` [PATCH -tip v11 19/27] x86/kprobes: Add UNWIND_HINT_FUNC on kretprobe_trampoline() Masami Hiramatsu
2021-09-14 14:41 ` [PATCH -tip v11 20/27] ARC: Add instruction_pointer_set() API Masami Hiramatsu
2021-09-14 14:41 ` [PATCH -tip v11 21/27] ia64: " Masami Hiramatsu
2021-09-14 14:42 ` [PATCH -tip v11 22/27] arm: kprobes: Make space for instruction pointer on stack Masami Hiramatsu
2021-09-14 14:42 ` [PATCH -tip v11 23/27] kprobes: Enable stacktrace from pt_regs in kretprobe handler Masami Hiramatsu
2021-09-14 14:42 ` [PATCH -tip v11 24/27] x86/kprobes: Push a fake return address at kretprobe_trampoline Masami Hiramatsu
2021-09-14 14:42 ` [PATCH -tip v11 25/27] x86/unwind: Recover kretprobe trampoline entry Masami Hiramatsu
2021-09-14 14:42 ` [PATCH -tip v11 26/27] tracing: Show kretprobe unknown indicator only for kretprobe_trampoline Masami Hiramatsu
2021-09-14 14:42 ` Masami Hiramatsu [this message]
2021-09-14 22:55 ` [PATCH -tip v11 00/27] kprobes: Fix stacktrace with kretprobes on x86 Andrii Nakryiko
2021-09-29  2:24 ` Masami Hiramatsu
2021-09-30 18:17   ` Alexei Starovoitov
2021-09-30 19:34     ` Thomas Gleixner
2021-09-30 21:22       ` Steven Rostedt
2021-09-30 23:11         ` Thomas Gleixner
2021-09-30 23:27           ` Masami Hiramatsu
2021-09-30 23:37             ` Steven Rostedt
2021-10-01  0:35               ` Masami Hiramatsu
2021-09-30 23:54       ` Masami Hiramatsu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=163163057094.489837.9044470370440745866.stgit@devnote2 \
    --to=mhiramat@kernel.org \
    --cc=andrii.nakryiko@gmail.com \
    --cc=ast@kernel.org \
    --cc=bp@alien8.de \
    --cc=bpf@vger.kernel.org \
    --cc=dxu@dxuuu.xyz \
    --cc=jpoimboe@redhat.com \
    --cc=kernel-team@fb.com \
    --cc=kuba@kernel.org \
    --cc=linux-ia64@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=mingo@redhat.com \
    --cc=paulmck@kernel.org \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    --cc=sagar.abhishek@gmail.com \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.org \
    --cc=yhs@fb.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).