From mboxrd@z Thu Jan 1 00:00:00 1970 From: Josh Poimboeuf Date: Thu, 17 Jun 2021 04:40:32 +0000 Subject: Re: [PATCH -tip v7 09/13] kprobes: Setup instruction pointer in __kretprobe_trampoline_handler Message-Id: <20210617044032.txng4enhiduacvt6@treble> List-Id: References: <162209754288.436794.3904335049560916855.stgit@devnote2> <162209762943.436794.874947392889792501.stgit@devnote2> <20210617043909.fgu2lhnkxflmy5mk@treble> In-Reply-To: <20210617043909.fgu2lhnkxflmy5mk@treble> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Masami Hiramatsu Cc: Steven Rostedt , Ingo Molnar , X86 ML , Daniel Xu , 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, linux-ia64@vger.kernel.org, Abhishek Sagar , Andrii Nakryiko On Wed, Jun 16, 2021 at 11:39:11PM -0500, Josh Poimboeuf wrote: > On Thu, May 27, 2021 at 03:40:29PM +0900, Masami Hiramatsu wrote: > > To simplify the stacktrace with pt_regs from kretprobe handler, > > set the correct return address to the instruction pointer in > > the pt_regs before calling kretprobe handlers. > > > > Suggested-by: Josh Poimboeuf > > Signed-off-by: Masami Hiramatsu > > Tested-by: Andrii Nakryik > > --- > > Changes in v3: > > - Cast the correct_ret_addr to unsigned long. > > --- > > kernel/kprobes.c | 3 +++ > > 1 file changed, 3 insertions(+) > > > > diff --git a/kernel/kprobes.c b/kernel/kprobes.c > > index 54e5b89aad67..1598aca375c9 100644 > > --- a/kernel/kprobes.c > > +++ b/kernel/kprobes.c > > @@ -1914,6 +1914,9 @@ unsigned long __kretprobe_trampoline_handler(struct pt_regs *regs, > > BUG_ON(1); > > } > > > > + /* Set the instruction pointer to the correct address */ > > + instruction_pointer_set(regs, (unsigned long)correct_ret_addr); > > + > > /* Run them. */ > > first = current->kretprobe_instances.first; > > while (first) { > > > > Hi Masami, > > I know I suggested this patch, but I believe it would only be useful in > combination with the use of UNWIND_HINT_REGS in SAVE_REGS_STRING. But I > think that would be tricky to pull off correctly. Instead, we have > UNWIND_HINT_FUNC, which is working fine. > > So I'd suggest dropping this patch, as the unwinder isn't actually > reading regs->ip after all. ... and I guess this means patches 6-8 are no longer necessary. -- Josh