From: Masami Hiramatsu <mhiramat@kernel.org>
To: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Daniel Xu <dxu@dxuuu.xyz>, Steven Rostedt <rostedt@goodmis.org>,
Ingo Molnar <mingo@kernel.org>, X86 ML <x86@kernel.org>,
linux-kernel@vger.kernel.org, bpf@vger.kernel.org,
kuba@kernel.org, mingo@redhat.com, ast@kernel.org,
tglx@linutronix.de, kernel-team@fb.com, yhs@fb.com
Subject: Re: [PATCH -tip 0/5] kprobes: Fix stacktrace in kretprobes
Date: Wed, 10 Mar 2021 18:57:34 +0900 [thread overview]
Message-ID: <20210310185734.332d9d52a26780ba02d09197@kernel.org> (raw)
In-Reply-To: <20210309011945.ky7v3pnbdpxhmxkh@treble>
Hi Josh,
On Mon, 8 Mar 2021 19:19:45 -0600
Josh Poimboeuf <jpoimboe@redhat.com> wrote:
> On Mon, Mar 08, 2021 at 11:52:10AM +0900, Masami Hiramatsu wrote:
> > So at the kretprobe handler, we have 2 issues.
> > 1) the return address (caller_func+0x15) is not on the stack.
> > this can be solved by searching from current->kretprobe_instances.
> > 2) the stack frame size of kretprobe_trampoline is unknown
> > Since the stackframe is fixed, the fixed number (0x98) can be used.
> >
> > However, those solutions are only for the kretprobe handler. The stacktrace
> > from interrupt handler hit in the kretprobe_trampoline still doesn't work.
> >
> > So, here is my idea;
> >
> > 1) Change the trampline code to prepare stack frame at first and save
> > registers on it, instead of "push". This will makes ORC easy to setup
> > stackframe information for this code.
> > 2) change the return addres fixup timing. Instead of using return value
> > of trampoline handler, before removing the real return address from
> > current->kretprobe_instances.
> > 3) Then, if orc_find() finds the ip is in the kretprobe_trampoline, it
> > checks the contents of the end of stackframe (at the place of regs->sp)
> > is same as the address of it. If it is, it can find the correct address
> > from current->kretprobe_instances. If not, there is a correct address.
> >
> > I need to find how the ORC info is prepared for 1), maybe UNWIND_HINT macro
> > may help?
>
> Hi Masami,
>
> If I understand correctly, for #1 you need an unwind hint which treats
> the instruction *after* the "pushq %rsp" as the beginning of the
> function.
Thanks for the patch. In that case, should I still change the stack allocation?
Or can I continue to use a series of "push/pop" ?
>
> I'm thinking this should work:
OK, Let me test it.
Thanks!
>
>
> diff --git a/arch/x86/include/asm/unwind_hints.h b/arch/x86/include/asm/unwind_hints.h
> index 8e574c0afef8..8b33674288ea 100644
> --- a/arch/x86/include/asm/unwind_hints.h
> +++ b/arch/x86/include/asm/unwind_hints.h
> @@ -52,6 +52,11 @@
> UNWIND_HINT sp_reg=ORC_REG_SP sp_offset=8 type=UNWIND_HINT_TYPE_FUNC
> .endm
>
> +#else
> +
> +#define UNWIND_HINT_FUNC \
> + UNWIND_HINT(ORC_REG_SP, 8, UNWIND_HINT_TYPE_FUNC, 0)
> +
> #endif /* __ASSEMBLY__ */
>
> #endif /* _ASM_X86_UNWIND_HINTS_H */
> diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
> index df776cdca327..38ff1387f95d 100644
> --- a/arch/x86/kernel/kprobes/core.c
> +++ b/arch/x86/kernel/kprobes/core.c
> @@ -767,6 +767,7 @@ asm(
> /* We don't bother saving the ss register */
> #ifdef CONFIG_X86_64
> " pushq %rsp\n"
> + UNWIND_HINT_FUNC
> " pushfq\n"
> SAVE_REGS_STRING
> " movq %rsp, %rdi\n"
> @@ -790,7 +791,6 @@ asm(
> ".size kretprobe_trampoline, .-kretprobe_trampoline\n"
> );
> NOKPROBE_SYMBOL(kretprobe_trampoline);
> -STACK_FRAME_NON_STANDARD(kretprobe_trampoline);
>
>
> /*
>
--
Masami Hiramatsu <mhiramat@kernel.org>
next prev parent reply other threads:[~2021-03-10 9:58 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-05 15:38 [PATCH -tip 0/5] kprobes: Fix stacktrace in kretprobes Masami Hiramatsu
2021-03-05 15:39 ` [PATCH -tip 1/5] ia64: kprobes: Fix to pass correct trampoline address to the handler Masami Hiramatsu
2021-03-05 15:39 ` [PATCH -tip 2/5] kprobes: treewide: Replace arch_deref_entry_point() with dereference_function_descriptor() Masami Hiramatsu
2021-03-05 15:39 ` [PATCH -tip 3/5] kprobes: treewide: Remove trampoline_address from kretprobe_trampoline_handler() Masami Hiramatsu
2021-03-10 14:21 ` Miroslav Benes
2021-03-10 15:42 ` Masami Hiramatsu
2021-03-11 13:37 ` Masami Hiramatsu
2021-03-05 15:39 ` [PATCH -tip 4/5] kprobes: stacktrace: Recover the address changed by kretprobe Masami Hiramatsu
2021-03-05 15:39 ` [PATCH -tip 5/5] tracing: Remove kretprobe unknown indicator from stacktrace Masami Hiramatsu
2021-03-05 15:44 ` Steven Rostedt
2021-03-05 19:16 ` [PATCH -tip 0/5] kprobes: Fix stacktrace in kretprobes Daniel Xu
2021-03-06 1:13 ` Masami Hiramatsu
2021-03-07 21:23 ` Daniel Xu
2021-03-08 2:52 ` Masami Hiramatsu
2021-03-08 13:05 ` Masami Hiramatsu
2021-03-09 1:19 ` Josh Poimboeuf
2021-03-10 9:57 ` Masami Hiramatsu [this message]
2021-03-10 15:08 ` Josh Poimboeuf
2021-03-10 15:55 ` Masami Hiramatsu
2021-03-10 18:31 ` Josh Poimboeuf
2021-03-11 0:20 ` Masami Hiramatsu
2021-03-11 1:06 ` Josh Poimboeuf
2021-03-11 1:54 ` Masami Hiramatsu
2021-03-11 16:51 ` Josh Poimboeuf
2021-03-12 3:22 ` Masami Hiramatsu
2021-03-10 22:46 ` Daniel Xu
2021-03-09 21:34 ` Daniel Xu
2021-03-10 10:50 ` 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=20210310185734.332d9d52a26780ba02d09197@kernel.org \
--to=mhiramat@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=dxu@dxuuu.xyz \
--cc=jpoimboe@redhat.com \
--cc=kernel-team@fb.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=mingo@redhat.com \
--cc=rostedt@goodmis.org \
--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).