From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
To: Sven Schnelle <svens@linux.ibm.com>
Cc: Alexei Starovoitov <alexei.starovoitov@gmail.com>,
Steven Rostedt <rostedt@goodmis.org>,
Florent Revest <revest@chromium.org>,
linux-trace-kernel@vger.kernel.org,
LKML <linux-kernel@vger.kernel.org>,
Martin KaFai Lau <martin.lau@linux.dev>,
bpf <bpf@vger.kernel.org>, Alexei Starovoitov <ast@kernel.org>,
Jiri Olsa <jolsa@kernel.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Alan Maguire <alan.maguire@oracle.com>,
Mark Rutland <mark.rutland@arm.com>,
Peter Zijlstra <peterz@infradead.org>,
Thomas Gleixner <tglx@linutronix.de>
Subject: Re: [PATCH v4 4/9] fprobe: rethook: Use ftrace_regs in fprobe exit handler and rethook
Date: Sat, 9 Sep 2023 23:24:35 +0900 [thread overview]
Message-ID: <20230909232435.dfa15f93f1c5eef5b229a7d2@kernel.org> (raw)
In-Reply-To: <yt9dzg1zokyg.fsf@linux.ibm.com>
Hi Sven,
On Wed, 06 Sep 2023 08:49:11 +0200
Sven Schnelle <svens@linux.ibm.com> wrote:
> Hi Masami,
>
> Masami Hiramatsu (Google) <mhiramat@kernel.org> writes:
>
> > Thus, we need to ensure that the ftrace_regs which is saved in the ftrace
> > *without* FTRACE_WITH_REGS flags, can be used for hooking the function
> > return. I saw;
> >
> > void arch_rethook_prepare(struct rethook_node *rh, struct pt_regs *regs, bool mcount)
> > {
> > rh->ret_addr = regs->gprs[14];
> > rh->frame = regs->gprs[15];
> >
> > /* Replace the return addr with trampoline addr */
> > regs->gprs[14] = (unsigned long)&arch_rethook_trampoline;
> > }
> >
> > gprs[15] is a stack pointer, so it is saved in ftrace_regs too, but what about
> > gprs[14]? (I guess it is a link register)
> > We need to read the gprs[14] and ensure that is restored to gpr14 when the
> > ftrace is exit even without FTRACE_WITH_REGS flag.
> >
> > IOW, it is ftrace save regs/restore regs code issue. I need to check how the
> > function_graph implements it.
>
> gpr2-gpr14 are always saved in ftrace_caller/ftrace_regs_caller(),
> regardless of the FTRACE_WITH_REGS flags. The only difference is that
> without the FTRACE_WITH_REGS flag the program status word (psw) is not
> saved because collecting that is a rather expensive operation.
Thanks for checking that! So s390 will recover those saved registers
even if FTRACE_WITH_REGS flag is not set? (I wonder what is the requirement
of the ftrace_regs when returning from ftrace_call() without
FTRACE_WITH_REGS?)
>
> I used the following commands to test rethook (is that the correct
> testcase?)
>
> #!/bin/bash
> cd /sys/kernel/tracing
>
> echo 'r:icmp_rcv icmp_rcv' >kprobe_events
> echo 1 >events/kprobes/icmp_rcv/enable
> ping -c 1 127.0.0.1
> cat trace
No, the kprobe will path pt_regs to rethook.
Cna you run
echo "f:icmp_rcv%return icmp_rcv" >> dynamic_events
instead of kprobe_events?
Thank you,
>
> which gave me:
>
> ping-686 [001] ..s1. 96.890817: icmp_rcv: (ip_protocol_deliver_rcu+0x42/0x218 <- icmp_rcv)
>
> I applied the following patch on top of your patches to make it compile,
> and rethook still seems to work:
>
> commit dab51b0a5b885660630433ac89f8e64a2de0eb86
> Author: Sven Schnelle <svens@linux.ibm.com>
> Date: Wed Sep 6 08:06:23 2023 +0200
>
> rethook wip
>
> Signed-off-by: Sven Schnelle <svens@linux.ibm.com>
>
> diff --git a/arch/s390/kernel/rethook.c b/arch/s390/kernel/rethook.c
> index af10e6bdd34e..4e86c0a1a064 100644
> --- a/arch/s390/kernel/rethook.c
> +++ b/arch/s390/kernel/rethook.c
> @@ -3,8 +3,9 @@
> #include <linux/kprobes.h>
> #include "rethook.h"
>
> -void arch_rethook_prepare(struct rethook_node *rh, struct pt_regs *regs, bool mcount)
> +void arch_rethook_prepare(struct rethook_node *rh, struct ftrace_regs *fregs, bool mcount)
> {
> + struct pt_regs *regs = (struct pt_regs *)fregs;
> rh->ret_addr = regs->gprs[14];
> rh->frame = regs->gprs[15];
>
> @@ -13,10 +14,11 @@ void arch_rethook_prepare(struct rethook_node *rh, struct pt_regs *regs, bool mc
> }
> NOKPROBE_SYMBOL(arch_rethook_prepare);
>
> -void arch_rethook_fixup_return(struct pt_regs *regs,
> +void arch_rethook_fixup_return(struct ftrace_regs *fregs,
> unsigned long correct_ret_addr)
> {
> /* Replace fake return address with real one. */
> + struct pt_regs *regs = (struct pt_regs *)fregs;
> regs->gprs[14] = correct_ret_addr;
> }
> NOKPROBE_SYMBOL(arch_rethook_fixup_return);
> @@ -24,9 +26,9 @@ NOKPROBE_SYMBOL(arch_rethook_fixup_return);
> /*
> * Called from arch_rethook_trampoline
> */
> -unsigned long arch_rethook_trampoline_callback(struct pt_regs *regs)
> +unsigned long arch_rethook_trampoline_callback(struct ftrace_regs *fregs)
> {
> - return rethook_trampoline_handler(regs, regs->gprs[15]);
> + return rethook_trampoline_handler(fregs, fregs->regs.gprs[15]);
> }
> NOKPROBE_SYMBOL(arch_rethook_trampoline_callback);
>
> diff --git a/arch/s390/kernel/rethook.h b/arch/s390/kernel/rethook.h
> index 32f069eed3f3..0fe62424fc78 100644
> --- a/arch/s390/kernel/rethook.h
> +++ b/arch/s390/kernel/rethook.h
> @@ -2,6 +2,6 @@
> #ifndef __S390_RETHOOK_H
> #define __S390_RETHOOK_H
>
> -unsigned long arch_rethook_trampoline_callback(struct pt_regs *regs);
> +unsigned long arch_rethook_trampoline_callback(struct ftrace_regs *fregs);
>
> #endif
>
--
Masami Hiramatsu (Google) <mhiramat@kernel.org>
next prev parent reply other threads:[~2023-09-09 16:25 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-23 15:15 [PATCH v4 0/9] bpf: fprobe: rethook: Use ftrace_regs instead of pt_regs Masami Hiramatsu (Google)
2023-08-23 15:15 ` [PATCH v4 1/9] Documentation: probes: Add a new ret_ip callback parameter Masami Hiramatsu (Google)
2023-08-25 16:11 ` Florent Revest
2023-08-23 15:15 ` [PATCH v4 2/9] fprobe: Use fprobe_regs in fprobe entry handler Masami Hiramatsu (Google)
2023-08-25 16:11 ` Florent Revest
2023-08-23 15:16 ` [PATCH v4 3/9] tracing: Expose ftrace_regs regardless of CONFIG_FUNCTION_TRACER Masami Hiramatsu (Google)
2023-08-23 15:16 ` [PATCH v4 4/9] fprobe: rethook: Use ftrace_regs in fprobe exit handler and rethook Masami Hiramatsu (Google)
2023-08-25 16:12 ` Florent Revest
2023-09-04 13:40 ` Masami Hiramatsu
2023-09-05 7:17 ` Sven Schnelle
2023-09-05 13:36 ` Masami Hiramatsu
2023-09-05 16:30 ` Steven Rostedt
2023-09-06 0:06 ` Masami Hiramatsu
2023-09-06 6:49 ` Sven Schnelle
2023-09-09 14:24 ` Masami Hiramatsu [this message]
2023-09-11 7:55 ` Sven Schnelle
2023-09-11 14:15 ` Masami Hiramatsu
2023-08-23 15:16 ` [PATCH v4 5/9] ftrace: Add ftrace_partial_regs() for converting ftrace_regs to pt_regs Masami Hiramatsu (Google)
2023-08-25 21:49 ` Andrii Nakryiko
2023-08-26 1:56 ` Masami Hiramatsu
2023-09-05 19:50 ` Andrii Nakryiko
2023-09-06 0:28 ` Masami Hiramatsu
2023-09-08 22:56 ` Andrii Nakryiko
2023-08-23 15:16 ` [PATCH v4 6/9] tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH_ARGS Masami Hiramatsu (Google)
2023-08-25 16:12 ` Florent Revest
2023-08-26 3:38 ` Masami Hiramatsu
2023-08-30 7:20 ` Masami Hiramatsu
2023-08-23 15:16 ` [PATCH v4 7/9] bpf: Enable kprobe_multi feature if CONFIG_FPROBE is enabled Masami Hiramatsu (Google)
2023-08-23 15:16 ` [PATCH v4 8/9] Documentations: probes: Update fprobe document to use ftrace_regs Masami Hiramatsu (Google)
2023-08-23 15:17 ` [PATCH v4 9/9] Documentation: tracing: Add a note about argument and retval access Masami Hiramatsu (Google)
2023-08-25 16:12 ` Florent Revest
2023-08-25 16:11 ` [PATCH v4 0/9] bpf: fprobe: rethook: Use ftrace_regs instead of pt_regs Florent Revest
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=20230909232435.dfa15f93f1c5eef5b229a7d2@kernel.org \
--to=mhiramat@kernel.org \
--cc=acme@kernel.org \
--cc=alan.maguire@oracle.com \
--cc=alexei.starovoitov@gmail.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-trace-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=martin.lau@linux.dev \
--cc=peterz@infradead.org \
--cc=revest@chromium.org \
--cc=rostedt@goodmis.org \
--cc=svens@linux.ibm.com \
--cc=tglx@linutronix.de \
/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