linux-perf-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Casey Chen <cachen@purestorage.com>
To: Namhyung Kim <namhyung@kernel.org>
Cc: linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org,
	 yzhong@purestorage.com
Subject: Re: [PATCH 1/1] perf tool: fix handling NULL al->maps returned from thread__find_map
Date: Mon, 15 Jul 2024 15:19:00 -0700	[thread overview]
Message-ID: <CALCePG0gUqXqz3nsK4JU3CdunzbXRKN0QnL1jinkKjKffKyCQw@mail.gmail.com> (raw)
In-Reply-To: <ZpGFfjebaW9LGNBo@google.com>

On Fri, Jul 12, 2024 at 12:35 PM Namhyung Kim <namhyung@kernel.org> wrote:
>
> Hello,
>
> On Wed, Jul 10, 2024 at 03:29:27PM -0700, Casey Chen wrote:
> > On Mon, Jul 8, 2024 at 10:01 PM Namhyung Kim <namhyung@kernel.org> wrote:
> > >
> > > Hello,
> > >
> > > On Mon, Jul 8, 2024 at 4:23 PM Casey Chen <cachen@purestorage.com> wrote:
> > > >
> > > > With 0dd5041c9a0e ("perf addr_location: Add init/exit/copy functions"),
> > > > thread__find_map() would return with al->maps or al->map being NULL
> > > > when cpumode is 3 (macro PERF_RECORD_MISC_HYPERVISOR),
> > > > later deferencing on it would crash.
> > > >
> > > > Fix callers of thread__find_map() or thread__find_symbol() to handle
> > > > this.
> > >
> > > It looks like you drop the callchain if it doesn't find a map/symbol.
> > > Can we keep the entries with raw hex numbers instead?
> > >
> > In add_callchain_ip(), my change let it return if either al.maps is
> > NULL or al.map is NULL after thread__find_symbol(), I'm not sure what
> > else can add_callchain_ip() could do to keep raw hex numbers. If it
> > proceeds, al.sym is NULL, the code inside 'if (al.sym != NULL)' would
> > skip. callchain_srcline() would return NULL. chain_cursor_append()
> > would append a node whose ms.maps/ ms.map are NULL. Later
> > dereferencing them would cause trouble. But we could add other
> > information to the node, like ip, branch, nr_loop_iter, iter_cycles,
> > branch_from, are these information good to have ? but how to avoid
> > dereferencing NULL maps/map later.
>
> By checking if it's NULL?  I think it's normal to have NULL map or sym
> due to missing events, stripped binaries and so on.  The callchain code
> used to print raw ip address when it doesn't have symbols.  And srcline
> can/should do the same.
>

Could you help make a fix ? Actually I don't quite understand how it works.

> Thanks,
> Namhyung
>
> > >
> > > > ---
> > > >  tools/perf/arch/powerpc/util/skip-callchain-idx.c | 10 ++++++----
> > > >  tools/perf/util/machine.c                         |  5 +++++
> > > >  tools/perf/util/unwind-libdw.c                    |  6 ++++--
> > > >  3 files changed, 15 insertions(+), 6 deletions(-)
> > > >
> > > > diff --git a/tools/perf/arch/powerpc/util/skip-callchain-idx.c b/tools/perf/arch/powerpc/util/skip-callchain-idx.c
> > > > index 5f3edb3004d8..25b0804df4c4 100644
> > > > --- a/tools/perf/arch/powerpc/util/skip-callchain-idx.c
> > > > +++ b/tools/perf/arch/powerpc/util/skip-callchain-idx.c
> > > > @@ -255,13 +255,14 @@ int arch_skip_callchain_idx(struct thread *thread, struct ip_callchain *chain)
> > > >
> > > >         thread__find_symbol(thread, PERF_RECORD_MISC_USER, ip, &al);
> > > >
> > > > -       if (al.map)
> > > > -               dso = map__dso(al.map);
> > > > +       if (!al.map)
> > > > +               goto out;
> > > > +
> > > > +       dso = map__dso(al.map);
> > > >
> > > >         if (!dso) {
> > > >                 pr_debug("%" PRIx64 " dso is NULL\n", ip);
> > > > -               addr_location__exit(&al);
> > > > -               return skip_slot;
> > > > +               goto out;
> > > >         }
> > > >
> > > >         rc = check_return_addr(dso, map__start(al.map), ip);
> > > > @@ -282,6 +283,7 @@ int arch_skip_callchain_idx(struct thread *thread, struct ip_callchain *chain)
> > > >                 skip_slot = 3;
> > > >         }
> > > >
> > > > +out:
> > > >         addr_location__exit(&al);
> > > >         return skip_slot;
> > > >  }
> > > > diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
> > > > index 8477edefc299..fa4037d7f3d4 100644
> > > > --- a/tools/perf/util/machine.c
> > > > +++ b/tools/perf/util/machine.c
> > > > @@ -2098,7 +2098,12 @@ static int add_callchain_ip(struct thread *thread,
> > > >                         }
> > > >                         goto out;
> > > >                 }
> > > > +
> > > >                 thread__find_symbol(thread, *cpumode, ip, &al);
> > > > +               if (!al.maps || !al.map) {
> > > > +                       err = 1;
> > > > +                       goto out;
> > > > +               }
> > > >         }
> > > >
> > > >         if (al.sym != NULL) {
> > > > diff --git a/tools/perf/util/unwind-libdw.c b/tools/perf/util/unwind-libdw.c
> > > > index b38d322734b4..fb038ef55be2 100644
> > > > --- a/tools/perf/util/unwind-libdw.c
> > > > +++ b/tools/perf/util/unwind-libdw.c
> > > > @@ -53,8 +53,10 @@ static int __report_module(struct addr_location *al, u64 ip,
> > > >          */
> > > >         thread__find_symbol(ui->thread, PERF_RECORD_MISC_USER, ip, al);
> > > >
> > > > -       if (al->map)
> > > > -               dso = map__dso(al->map);
> > > > +       if (!al->map)
> > > > +               return -1;
> > > > +
> > > > +       dso = map__dso(al->map);
> > > >
> > > >         if (!dso)
> > > >                 return 0;
> > > > --
> > > > 2.45.2
> > > >
> > > >

  reply	other threads:[~2024-07-15 22:19 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-07-08 23:23 [PATCH 0/1] perf tool: fix handling NULL al->maps returned from thread__find_map Casey Chen
2024-07-08 23:23 ` [PATCH 1/1] " Casey Chen
2024-07-09  5:01   ` Namhyung Kim
2024-07-10 22:29     ` Casey Chen
2024-07-12 19:35       ` Namhyung Kim
2024-07-15 22:19         ` Casey Chen [this message]
2024-07-08 23:23 ` [PATCH] debug patch for perf report segfault Casey Chen

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=CALCePG0gUqXqz3nsK4JU3CdunzbXRKN0QnL1jinkKjKffKyCQw@mail.gmail.com \
    --to=cachen@purestorage.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=namhyung@kernel.org \
    --cc=yzhong@purestorage.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).