From: mhiramat at kernel.org (Masami Hiramatsu)
Subject: [PATCH 2/4] x86/kprobes: Fix frame pointer annotations
Date: Fri, 10 May 2019 12:21:03 +0900 [thread overview]
Message-ID: <20190510122103.5a7bc5416b7af96b27d4fab4@kernel.org> (raw)
In-Reply-To: <20190509174316.pzuakeu657g3fnlm@home.goodmis.org>
On Thu, 9 May 2019 13:43:16 -0400
Steven Rostedt <rostedt at goodmis.org> wrote:
> On Thu, May 09, 2019 at 09:20:06AM -0700, Andy Lutomirski wrote:
> > > +END(call_to_exception_trampoline)
> > > --- a/arch/x86/kernel/kprobes/core.c
> > > +++ b/arch/x86/kernel/kprobes/core.c
> > > @@ -731,29 +731,8 @@ asm(
> > > ".global kretprobe_trampoline\n"
> > > ".type kretprobe_trampoline, @function\n"
> > > "kretprobe_trampoline:\n"
> > > - /* We don't bother saving the ss register */
> > > -#ifdef CONFIG_X86_64
> > > - " pushq %rsp\n"
> > > - " pushfq\n"
> > > - SAVE_REGS_STRING
> > > - " movq %rsp, %rdi\n"
> > > - " call trampoline_handler\n"
> > > - /* Replace saved sp with true return address. */
> > > - " movq %rax, 19*8(%rsp)\n"
> > > - RESTORE_REGS_STRING
> > > - " popfq\n"
> > > -#else
> > > - " pushl %esp\n"
> > > - " pushfl\n"
> > > - SAVE_REGS_STRING
> > > - " movl %esp, %eax\n"
> > > - " call trampoline_handler\n"
> > > - /* Replace saved sp with true return address. */
> > > - " movl %eax, 15*4(%esp)\n"
> > > - RESTORE_REGS_STRING
> > > - " popfl\n"
> > > -#endif
> > > - " ret\n"
> > > + "push trampoline_handler\n"
> > > + "jmp call_to_exception_trampoline\n"
> > > ".size kretprobe_trampoline, .-kretprobe_trampoline\n"
> > > );
> >
> >
> > Potentially minor nit: you’re doing popfl, but you’re not doing TRACE_IRQ_whatever. This makes me think that you should either add the tracing (ugh!) or you should maybe just skip the popfl.
>
>
> Note, kprobes (and ftrace for that matter) are not saving flags for
> interrupts, but because it must not modify the sign, zero and carry flags.
Yes, optprobe also has to save and restore the flags.
Above trampline is for kretprobe, which is placed at the function return, so
we don't have to care about flags.
Thanks,
--
Masami Hiramatsu <mhiramat at kernel.org>
WARNING: multiple messages have this Message-ID (diff)
From: mhiramat@kernel.org (Masami Hiramatsu)
Subject: [PATCH 2/4] x86/kprobes: Fix frame pointer annotations
Date: Fri, 10 May 2019 12:21:03 +0900 [thread overview]
Message-ID: <20190510122103.5a7bc5416b7af96b27d4fab4@kernel.org> (raw)
Message-ID: <20190510032103.5l2pplOV1fxGiMqlEQQUvJcYOvEpYDYmqBbyPkina6Y@z> (raw)
In-Reply-To: <20190509174316.pzuakeu657g3fnlm@home.goodmis.org>
On Thu, 9 May 2019 13:43:16 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:
> On Thu, May 09, 2019@09:20:06AM -0700, Andy Lutomirski wrote:
> > > +END(call_to_exception_trampoline)
> > > --- a/arch/x86/kernel/kprobes/core.c
> > > +++ b/arch/x86/kernel/kprobes/core.c
> > > @@ -731,29 +731,8 @@ asm(
> > > ".global kretprobe_trampoline\n"
> > > ".type kretprobe_trampoline, @function\n"
> > > "kretprobe_trampoline:\n"
> > > - /* We don't bother saving the ss register */
> > > -#ifdef CONFIG_X86_64
> > > - " pushq %rsp\n"
> > > - " pushfq\n"
> > > - SAVE_REGS_STRING
> > > - " movq %rsp, %rdi\n"
> > > - " call trampoline_handler\n"
> > > - /* Replace saved sp with true return address. */
> > > - " movq %rax, 19*8(%rsp)\n"
> > > - RESTORE_REGS_STRING
> > > - " popfq\n"
> > > -#else
> > > - " pushl %esp\n"
> > > - " pushfl\n"
> > > - SAVE_REGS_STRING
> > > - " movl %esp, %eax\n"
> > > - " call trampoline_handler\n"
> > > - /* Replace saved sp with true return address. */
> > > - " movl %eax, 15*4(%esp)\n"
> > > - RESTORE_REGS_STRING
> > > - " popfl\n"
> > > -#endif
> > > - " ret\n"
> > > + "push trampoline_handler\n"
> > > + "jmp call_to_exception_trampoline\n"
> > > ".size kretprobe_trampoline, .-kretprobe_trampoline\n"
> > > );
> >
> >
> > Potentially minor nit: you’re doing popfl, but you’re not doing TRACE_IRQ_whatever. This makes me think that you should either add the tracing (ugh!) or you should maybe just skip the popfl.
>
>
> Note, kprobes (and ftrace for that matter) are not saving flags for
> interrupts, but because it must not modify the sign, zero and carry flags.
Yes, optprobe also has to save and restore the flags.
Above trampline is for kretprobe, which is placed at the function return, so
we don't have to care about flags.
Thanks,
--
Masami Hiramatsu <mhiramat at kernel.org>
next prev parent reply other threads:[~2019-05-10 3:21 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-08 7:49 [PATCH 0/4] x86: int3 fallout peterz
2019-05-08 7:49 ` Peter Zijlstra
2019-05-08 7:49 ` [PATCH 1/4] x86/entry/32: Clean up return from interrupt preemption path peterz
2019-05-08 7:49 ` Peter Zijlstra
2019-05-08 7:49 ` [PATCH 2/4] x86/kprobes: Fix frame pointer annotations peterz
2019-05-08 7:49 ` Peter Zijlstra
2019-05-08 11:54 ` jpoimboe
2019-05-08 11:54 ` Josh Poimboeuf
2019-05-08 12:04 ` peterz
2019-05-08 12:04 ` Peter Zijlstra
[not found] ` <20190508124248.u5ukpbhnh4wpiccq@treble>
[not found] ` <20190508153907.GM2589@hirez.programming.kicks-ass.net>
2019-05-08 18:48 ` jpoimboe
2019-05-08 18:48 ` Josh Poimboeuf
2019-05-09 1:20 ` mhiramat
2019-05-09 1:20 ` Masami Hiramatsu
2019-05-09 8:14 ` peterz
2019-05-09 8:14 ` Peter Zijlstra
2019-05-09 9:27 ` peterz
2019-05-09 9:27 ` Peter Zijlstra
2019-05-09 14:00 ` jpoimboe
2019-05-09 14:00 ` Josh Poimboeuf
2019-05-09 14:01 ` mhiramat
2019-05-09 14:01 ` Masami Hiramatsu
2019-05-09 17:14 ` peterz
2019-05-09 17:14 ` Peter Zijlstra
2019-05-10 4:58 ` mhiramat
2019-05-10 4:58 ` Masami Hiramatsu
2019-05-10 12:31 ` peterz
2019-05-10 12:31 ` Peter Zijlstra
2019-05-11 0:52 ` mhiramat
2019-05-11 0:52 ` Masami Hiramatsu
2019-05-10 12:40 ` peterz
2019-05-10 12:40 ` Peter Zijlstra
2019-05-11 0:56 ` mhiramat
2019-05-11 0:56 ` Masami Hiramatsu
2019-05-13 8:15 ` peterz
2019-05-13 8:15 ` Peter Zijlstra
2019-05-09 16:20 ` luto
2019-05-09 16:20 ` Andy Lutomirski
2019-05-09 17:18 ` peterz
2019-05-09 17:18 ` Peter Zijlstra
2019-05-09 17:43 ` rostedt
2019-05-09 17:43 ` Steven Rostedt
2019-05-10 3:21 ` mhiramat [this message]
2019-05-10 3:21 ` Masami Hiramatsu
2019-05-10 12:14 ` peterz
2019-05-10 12:14 ` Peter Zijlstra
2019-05-10 12:17 ` peterz
2019-05-10 12:17 ` Peter Zijlstra
2019-05-10 14:54 ` rostedt
2019-05-10 14:54 ` Steven Rostedt
2019-05-09 17:37 ` rostedt
2019-05-09 17:37 ` Steven Rostedt
2019-05-09 18:26 ` peterz
2019-05-09 18:26 ` Peter Zijlstra
2019-05-09 18:36 ` rostedt
2019-05-09 18:36 ` Steven Rostedt
2019-05-08 7:49 ` [PATCH 3/4] x86/ftrace: Add pt_regs frame annotations peterz
2019-05-08 7:49 ` Peter Zijlstra
2019-05-08 7:49 ` [RFC][PATCH 4/4] x86_32: Provide consistent pt_regs peterz
2019-05-08 7:49 ` Peter Zijlstra
2019-05-08 11:57 ` jpoimboe
2019-05-08 11:57 ` Josh Poimboeuf
2019-05-08 20:58 ` torvalds
2019-05-08 20:58 ` Linus Torvalds
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=20190510122103.5a7bc5416b7af96b27d4fab4@kernel.org \
--to=linux-kselftest@vger.kernel.org \
/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