From: Josh Poimboeuf <jpoimboe@redhat.com>
To: Petr Mladek <pmladek@suse.com>
Cc: Jessica Yu <jeyu@redhat.com>, Jiri Kosina <jikos@kernel.org>,
Miroslav Benes <mbenes@suse.cz>,
linux-kernel@vger.kernel.org, live-patching@vger.kernel.org,
Michael Ellerman <mpe@ellerman.id.au>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
x86@kernel.org, linuxppc-dev@lists.ozlabs.org,
linux-s390@vger.kernel.org, Vojtech Pavlik <vojtech@suse.com>,
Jiri Slaby <jslaby@suse.cz>,
Chris J Arges <chris.j.arges@canonical.com>,
Andy Lutomirski <luto@kernel.org>, Ingo Molnar <mingo@kernel.org>,
Peter Zijlstra <peterz@infradead.org>
Subject: Re: [PATCH v3 01/15] stacktrace/x86: add function for detecting reliable stack traces
Date: Fri, 16 Dec 2016 16:09:47 -0600 [thread overview]
Message-ID: <20161216220947.j7bzbttaqklqic37@treble> (raw)
In-Reply-To: <20161216130739.GB393@pathway.suse.cz>
On Fri, Dec 16, 2016 at 02:07:39PM +0100, Petr Mladek wrote:
> On Thu 2016-12-08 12:08:26, Josh Poimboeuf wrote:
> > For live patching and possibly other use cases, a stack trace is only
> > useful if it can be assured that it's completely reliable. Add a new
> > save_stack_trace_tsk_reliable() function to achieve that.
> >
> > Scenarios which indicate that a stack trace may be unreliable:
> >
> > - running task
>
> It seems that this has to be enforced by save_stack_trace_tsk_reliable()
> caller. It should be mentioned in the function description.
Agreed.
> > - interrupt stack
>
> I guess that it is detected by saved regs on the stack. And it covers
> also dynamic changes like kprobes. Do I get it correctly, please?
Right.
> What about ftrace? Is ftrace without regs safe and detected?
Yes, it's safe because the mcount code does the right thing with respect
to frame pointers. See save_mcount_regs().
> > - preemption
>
> I wonder if some very active kthreads might almost always be
> preempted using irq in preemptive kernel. Then they block
> the conversion with the non-reliable stacks. Have you noticed
> such problems, please?
I haven't seen such a case and I think it would be quite rare for a
kthread to be CPU-bound like that.
> > - corrupted stack data
> > - stack grows the wrong way
>
> This is detected in unwind_next_frame() and passed via state->error.
> Am I right?
Right. I'll add more details to the commit message for all of these.
>
>
> > - stack walk doesn't reach the bottom
> > - user didn't provide a large enough entries array
> >
> > Also add CONFIG_HAVE_RELIABLE_STACKTRACE so arch-independent code can
> > determine at build time whether the function is implemented.
> >
> > diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c
> > index 0653788..3e0cf5e 100644
> > --- a/arch/x86/kernel/stacktrace.c
> > +++ b/arch/x86/kernel/stacktrace.c
> > @@ -74,6 +74,64 @@ void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace)
> > }
> > EXPORT_SYMBOL_GPL(save_stack_trace_tsk);
> >
> > +#ifdef CONFIG_HAVE_RELIABLE_STACKTRACE
> > +static int __save_stack_trace_reliable(struct stack_trace *trace,
> > + struct task_struct *task)
> > +{
> > + struct unwind_state state;
> > + struct pt_regs *regs;
> > + unsigned long addr;
> > +
> > + for (unwind_start(&state, task, NULL, NULL); !unwind_done(&state);
> > + unwind_next_frame(&state)) {
> > +
> > + regs = unwind_get_entry_regs(&state);
> > + if (regs) {
> > + /*
> > + * Preemption and page faults on the stack can make
> > + * frame pointers unreliable.
> > + */
> > + if (!user_mode(regs))
> > + return -1;
>
> By other words, it we find regs on the stack, it almost always mean
> a non-reliable stack. The only exception is when we are in the
> userspace mode. Do I get it correctly, please?
Right.
> > +
> > + /*
> > + * This frame contains the (user mode) pt_regs at the
> > + * end of the stack. Finish the unwind.
> > + */
> > + unwind_next_frame(&state);
> > + break;
> > + }
> > +
> > + addr = unwind_get_return_address(&state);
> > + if (!addr || save_stack_address(trace, addr, false))
> > + return -1;
> > + }
> > +
> > + if (!unwind_done(&state) || unwind_error(&state))
> > + return -1;
> > +
> > + if (trace->nr_entries < trace->max_entries)
> > + trace->entries[trace->nr_entries++] = ULONG_MAX;
> > +
> > + return 0;
> > +}
>
> Great work! I am surprised that it looks so straightforward.
>
> I still have to think and investigate it more. But it looks
> very promissing.
>
> Best Regards,
> Petr
--
Josh
next prev parent reply other threads:[~2016-12-16 22:10 UTC|newest]
Thread overview: 78+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-08 18:08 [PATCH v3 00/15] livepatch: hybrid consistency model Josh Poimboeuf
2016-12-08 18:08 ` [PATCH v3 01/15] stacktrace/x86: add function for detecting reliable stack traces Josh Poimboeuf
2016-12-16 13:07 ` Petr Mladek
2016-12-16 22:09 ` Josh Poimboeuf [this message]
2016-12-19 16:25 ` Miroslav Benes
2016-12-19 17:25 ` Josh Poimboeuf
2016-12-19 18:23 ` Miroslav Benes
2016-12-20 9:39 ` Petr Mladek
2016-12-20 21:21 ` Josh Poimboeuf
2016-12-08 18:08 ` [PATCH v3 02/15] x86/entry: define _TIF_ALLWORK_MASK flags explicitly Josh Poimboeuf
2016-12-16 14:17 ` Petr Mladek
2016-12-16 22:13 ` Josh Poimboeuf
2016-12-19 16:39 ` Miroslav Benes
2017-01-10 8:49 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 03/15] livepatch: temporary stubs for klp_patch_pending() and klp_update_patch_state() Josh Poimboeuf
2016-12-16 14:41 ` Petr Mladek
2016-12-16 22:15 ` Josh Poimboeuf
2016-12-08 18:08 ` [PATCH v3 04/15] livepatch/x86: add TIF_PATCH_PENDING thread flag Josh Poimboeuf
2016-12-08 18:27 ` Andy Lutomirski
2016-12-16 15:39 ` Petr Mladek
2016-12-21 13:54 ` Miroslav Benes
2017-01-11 7:06 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 05/15] livepatch/powerpc: " Josh Poimboeuf
2016-12-16 16:00 ` Petr Mladek
2016-12-21 14:30 ` Miroslav Benes
2017-01-10 8:29 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 06/15] livepatch/s390: reorganize TIF thread flag bits Josh Poimboeuf
2016-12-21 15:29 ` Miroslav Benes
2016-12-08 18:08 ` [PATCH v3 07/15] livepatch/s390: add TIF_PATCH_PENDING thread flag Josh Poimboeuf
2016-12-08 18:08 ` [PATCH v3 08/15] livepatch: separate enabled and patched states Josh Poimboeuf
2016-12-16 16:21 ` Petr Mladek
2016-12-23 12:54 ` Miroslav Benes
2017-01-10 9:10 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 09/15] livepatch: remove unnecessary object loaded check Josh Poimboeuf
2016-12-16 16:26 ` Petr Mladek
2016-12-23 12:58 ` Miroslav Benes
2017-01-10 9:14 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 10/15] livepatch: move patching functions into patch.c Josh Poimboeuf
2016-12-16 16:49 ` Petr Mladek
2016-12-23 13:06 ` Miroslav Benes
2017-01-10 9:15 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 11/15] livepatch: use kstrtobool() in enabled_store() Josh Poimboeuf
2016-12-16 16:55 ` Petr Mladek
2016-12-16 22:19 ` Josh Poimboeuf
2016-12-23 13:13 ` Miroslav Benes
2016-12-08 18:08 ` [PATCH v3 12/15] livepatch: store function sizes Josh Poimboeuf
2016-12-19 13:10 ` Petr Mladek
2016-12-23 13:40 ` Miroslav Benes
2017-01-11 10:09 ` Kamalesh Babulal
2016-12-08 18:08 ` [PATCH v3 13/15] livepatch: change to a per-task consistency model Josh Poimboeuf
2016-12-20 17:32 ` Petr Mladek
2016-12-21 21:25 ` Josh Poimboeuf
2016-12-22 14:34 ` Petr Mladek
2016-12-22 18:31 ` Josh Poimboeuf
2017-01-10 13:00 ` Petr Mladek
2017-01-10 20:46 ` Josh Poimboeuf
2017-01-11 15:18 ` Petr Mladek
2017-01-11 15:26 ` Josh Poimboeuf
2016-12-23 9:24 ` Miroslav Benes
2016-12-23 10:18 ` Petr Mladek
2017-01-06 20:07 ` Josh Poimboeuf
2017-01-10 10:40 ` Petr Mladek
2017-01-04 13:44 ` Miroslav Benes
2017-01-06 21:01 ` Josh Poimboeuf
2017-01-10 10:45 ` Miroslav Benes
2017-01-05 9:34 ` Miroslav Benes
2017-01-06 21:04 ` Josh Poimboeuf
2016-12-08 18:08 ` [PATCH v3 14/15] livepatch: add /proc/<pid>/patch_state Josh Poimboeuf
2016-12-21 11:20 ` Petr Mladek
2017-01-04 14:50 ` Miroslav Benes
2016-12-08 18:08 ` [PATCH v3 15/15] livepatch: allow removal of a disabled patch Josh Poimboeuf
2016-12-21 14:44 ` Petr Mladek
2017-01-04 14:57 ` Miroslav Benes
2017-01-06 21:04 ` Josh Poimboeuf
2016-12-10 5:46 ` [PATCH v3 00/15] livepatch: hybrid consistency model Balbir Singh
2016-12-10 17:17 ` Josh Poimboeuf
2016-12-11 2:08 ` Balbir Singh
2016-12-12 14:04 ` Josh Poimboeuf
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=20161216220947.j7bzbttaqklqic37@treble \
--to=jpoimboe@redhat.com \
--cc=chris.j.arges@canonical.com \
--cc=heiko.carstens@de.ibm.com \
--cc=jeyu@redhat.com \
--cc=jikos@kernel.org \
--cc=jslaby@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=live-patching@vger.kernel.org \
--cc=luto@kernel.org \
--cc=mbenes@suse.cz \
--cc=mingo@kernel.org \
--cc=mpe@ellerman.id.au \
--cc=peterz@infradead.org \
--cc=pmladek@suse.com \
--cc=vojtech@suse.com \
--cc=x86@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;
as well as URLs for NNTP newsgroup(s).