All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 1/2] perf tools: Fix crash in perf_hpp__reset_output_field
@ 2017-06-01 11:17 Jiri Olsa
  2017-06-01 11:17 ` [PATCH 2/2] perf tools: Remove extra list_del_init calls in list reset Jiri Olsa
  0 siblings, 1 reply; 4+ messages in thread
From: Jiri Olsa @ 2017-06-01 11:17 UTC (permalink / raw)
  To: Arnaldo Carvalho de Melo
  Cc: lkml, Ingo Molnar, Peter Zijlstra, Namhyung Kim, David Ahern, Du,
	Changbin

Du Changbin reported crash [1] when calling perf_hpp__reset_output_field
after unregistering field via perf_hpp__column_unregister.

This ends up in calling following list_del* sequence on
the same format:

  perf_hpp__column_unregister:
    list_del(&format->list);
  perf_hpp__reset_output_field:
    list_del_init(&fmt->list);

where the later list_del_init might touch already
freed formats.

Fixing this by replacing list_del with list_del_init
in perf_hpp__column_unregister.

[1] http://marc.info/?l=linux-kernel&m=149059595826019&w=2

Reported-by: "Du, Changbin" <changbin.du@intel.com>
Link: http://lkml.kernel.org/n/tip-8umo89ntt3kawmfwsivav43t@git.kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 tools/perf/ui/hist.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index 59addd52d9cd..feb2174ddd1f 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -530,7 +530,7 @@ void perf_hpp_list__prepend_sort_field(struct perf_hpp_list *list,
 
 void perf_hpp__column_unregister(struct perf_hpp_fmt *format)
 {
-	list_del(&format->list);
+	list_del_init(&format->list);
 }
 
 void perf_hpp__cancel_cumulate(void)
-- 
2.9.4

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* [PATCH 2/2] perf tools: Remove extra list_del_init calls in list reset
  2017-06-01 11:17 [PATCH 1/2] perf tools: Fix crash in perf_hpp__reset_output_field Jiri Olsa
@ 2017-06-01 11:17 ` Jiri Olsa
  2017-06-01 13:02   ` Arnaldo Carvalho de Melo
  0 siblings, 1 reply; 4+ messages in thread
From: Jiri Olsa @ 2017-06-01 11:17 UTC (permalink / raw)
  To: Arnaldo Carvalho de Melo
  Cc: lkml, Ingo Molnar, Peter Zijlstra, Namhyung Kim, David Ahern, Du,
	Changbin

We only needs to remove the format from the currently
iterated list. The other removals/inits are superfluous
because we free the format in any case.

Link: http://lkml.kernel.org/n/tip-8umo89ntt3kawmfwsivav43t@git.kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 tools/perf/ui/hist.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index feb2174ddd1f..a0fee2ac8599 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -614,15 +614,15 @@ void perf_hpp__reset_output_field(struct perf_hpp_list *list)
 
 	/* reset output fields */
 	perf_hpp_list__for_each_format_safe(list, fmt, tmp) {
-		list_del_init(&fmt->list);
-		list_del_init(&fmt->sort_list);
+		list_del(&fmt->list);
+		/* Remove the fmt from next loop processing. */
+		list_del(&fmt->sort_list);
 		fmt_free(fmt);
 	}
 
 	/* reset sort keys */
 	perf_hpp_list__for_each_sort_list_safe(list, fmt, tmp) {
-		list_del_init(&fmt->list);
-		list_del_init(&fmt->sort_list);
+		list_del(&fmt->sort_list);
 		fmt_free(fmt);
 	}
 }
-- 
2.9.4

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH 2/2] perf tools: Remove extra list_del_init calls in list reset
  2017-06-01 11:17 ` [PATCH 2/2] perf tools: Remove extra list_del_init calls in list reset Jiri Olsa
@ 2017-06-01 13:02   ` Arnaldo Carvalho de Melo
  2017-06-01 13:12     ` Jiri Olsa
  0 siblings, 1 reply; 4+ messages in thread
From: Arnaldo Carvalho de Melo @ 2017-06-01 13:02 UTC (permalink / raw)
  To: Jiri Olsa
  Cc: lkml, Ingo Molnar, Peter Zijlstra, Namhyung Kim, David Ahern, Du,
	Changbin

Em Thu, Jun 01, 2017 at 01:17:44PM +0200, Jiri Olsa escreveu:
> We only needs to remove the format from the currently
> iterated list. The other removals/inits are superfluous
> because we free the format in any case.
> 
> Link: http://lkml.kernel.org/n/tip-8umo89ntt3kawmfwsivav43t@git.kernel.org
> Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> ---
>  tools/perf/ui/hist.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
> index feb2174ddd1f..a0fee2ac8599 100644
> --- a/tools/perf/ui/hist.c
> +++ b/tools/perf/ui/hist.c
> @@ -614,15 +614,15 @@ void perf_hpp__reset_output_field(struct perf_hpp_list *list)
>  
>  	/* reset output fields */
>  	perf_hpp_list__for_each_format_safe(list, fmt, tmp) {
> -		list_del_init(&fmt->list);
> -		list_del_init(&fmt->sort_list);
> +		list_del(&fmt->list);
> +		/* Remove the fmt from next loop processing. */
> +		list_del(&fmt->sort_list);

Why not just add the comment and leave it as list_del_init(), then, in
fmt_free() -> fmt->free() -> hse_free() (for instance), have:

	BUG_ON(!list_empty(&fmt->list));
	BUG_ON(!list_empty(&fmt->sort_list));

The patch would be smaller and overall the code would be more robust?

>  		fmt_free(fmt);
>  	}
>  
>  	/* reset sort keys */
>  	perf_hpp_list__for_each_sort_list_safe(list, fmt, tmp) {
> -		list_del_init(&fmt->list);
> -		list_del_init(&fmt->sort_list);
> +		list_del(&fmt->sort_list);

Ditto, just ditch what you said you would do in the cset comment, i.e.
ditch the extra "list_del_init(&fmt->list);" call, no?

- Arnaldo

>  		fmt_free(fmt);
>  	}
>  }
> -- 
> 2.9.4

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH 2/2] perf tools: Remove extra list_del_init calls in list reset
  2017-06-01 13:02   ` Arnaldo Carvalho de Melo
@ 2017-06-01 13:12     ` Jiri Olsa
  0 siblings, 0 replies; 4+ messages in thread
From: Jiri Olsa @ 2017-06-01 13:12 UTC (permalink / raw)
  To: Arnaldo Carvalho de Melo
  Cc: Jiri Olsa, lkml, Ingo Molnar, Peter Zijlstra, Namhyung Kim,
	David Ahern, Du, Changbin

On Thu, Jun 01, 2017 at 10:02:26AM -0300, Arnaldo Carvalho de Melo wrote:
> Em Thu, Jun 01, 2017 at 01:17:44PM +0200, Jiri Olsa escreveu:
> > We only needs to remove the format from the currently
> > iterated list. The other removals/inits are superfluous
> > because we free the format in any case.
> > 
> > Link: http://lkml.kernel.org/n/tip-8umo89ntt3kawmfwsivav43t@git.kernel.org
> > Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> > ---
> >  tools/perf/ui/hist.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
> > index feb2174ddd1f..a0fee2ac8599 100644
> > --- a/tools/perf/ui/hist.c
> > +++ b/tools/perf/ui/hist.c
> > @@ -614,15 +614,15 @@ void perf_hpp__reset_output_field(struct perf_hpp_list *list)
> >  
> >  	/* reset output fields */
> >  	perf_hpp_list__for_each_format_safe(list, fmt, tmp) {
> > -		list_del_init(&fmt->list);
> > -		list_del_init(&fmt->sort_list);
> > +		list_del(&fmt->list);
> > +		/* Remove the fmt from next loop processing. */
> > +		list_del(&fmt->sort_list);
> 
> Why not just add the comment and leave it as list_del_init(), then, in
> fmt_free() -> fmt->free() -> hse_free() (for instance), have:
> 
> 	BUG_ON(!list_empty(&fmt->list));
> 	BUG_ON(!list_empty(&fmt->sort_list));
> 
> The patch would be smaller and overall the code would be more robust?

ok, that could catch something.. will change

thanks,
jirka

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2017-06-01 13:12 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-06-01 11:17 [PATCH 1/2] perf tools: Fix crash in perf_hpp__reset_output_field Jiri Olsa
2017-06-01 11:17 ` [PATCH 2/2] perf tools: Remove extra list_del_init calls in list reset Jiri Olsa
2017-06-01 13:02   ` Arnaldo Carvalho de Melo
2017-06-01 13:12     ` Jiri Olsa

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.