Re: [PATCH 3/4] perf timechart: Fix memory leaks in draw_wakeups()

[Namhyung Kim <namhyung@kernel.org>]

On Fri, Jun 05, 2026 at 11:17:33PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] The modified fallback loop condition `while (c && !from && !to)` prematurely aborts the search for task coordinates if either `from` or `to` has already been found.
> --
> 
> commit 757fad46c9b9ecd50deaba0cc7cc1ec56200ab8e
> Author: Namhyung Kim <namhyung@kernel.org>
> 
> perf timechart: Fix memory leaks in draw_wakeups()
> 
> This commit modifies the fallback coordinate search in draw_wakeups()
> to only run when it doesn't have from and to tasks, and ensures c->Y is
> set before copying the comm string to fix memory leaks.
> 
> > diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
> > index fded4be93de0d..9ac4662ef16e9 100644
> > --- a/tools/perf/builtin-timechart.c
> > +++ b/tools/perf/builtin-timechart.c
> 
> [ ... ]
> 
> > @@ -1071,12 +1071,12 @@ static void draw_wakeups(struct timechart *tchart)
> >  					c = c->next;
> >  				}
> >  				c = p->all;
> > -				while (c) {
> > -					if (p->pid == we->waker && !from) {
> > +				while (c && !from && !to) {
> 
> [Severity: Medium]
> Will this condition prematurely abort the search for the second task's
> coordinates?
> 
> Because the waker and wakee are typically two different PIDs, they are
> processed in different iterations of the outer per_pid loop. When the first
> one is found, either from or to becomes non-zero.
> 
> When the outer loop reaches the second PID, the condition !from && !to
> evaluates to false immediately, which skips the fallback coordinate search
> entirely for that second task.
> 
> Should this condition instead be !from || !to so the loop continues as long
> as either coordinate is still missing?

Oops, right.  I meant it.  Will send v2 soon.

Thanks,
Namhyung