[igt-dev] [PATCH i-g-t] runner: Add signal sender name when dying

[Petri Latvala]

On Wed, Sep 18, 2019 at 11:49:57AM +0100, Chris Wilson wrote:
> Quoting Petri Latvala (2019-09-18 11:45:33)
> > On Wed, Sep 18, 2019 at 11:35:18AM +0100, Chris Wilson wrote:
> > > We want to know who sent us the fatal signal, for there are plenty of
> > > fingers to go around.
> > > 
> > > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > > Cc: Petri Latvala <petri.latvala at intel.com>
> > > ---
> > >  runner/executor.c | 29 +++++++++++++++++++++++++++--
> > >  1 file changed, 27 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/runner/executor.c b/runner/executor.c
> > > index c1cfcce83..f7183293d 100644
> > > --- a/runner/executor.c
> > > +++ b/runner/executor.c
> > > @@ -609,6 +609,26 @@ static bool kill_child(int sig, pid_t child)
> > >       return true;
> > >  }
> > >  
> > > +static const char *get_cmdline(pid_t pid, char *buf, size_t len)
> > > +{
> > > +     int fd;
> > > +
> > > +     if (snprintf(buf, len, "/proc/%d/cmdline", pid) > len)
> > > +             return "unknown";
> > > +
> > > +     fd = open(buf, O_RDONLY);
> > > +     if (fd < 0)
> > > +             return "unknown";
> > > +
> > > +     len = read(fd, buf, len - 1);
> > > +     close(fd);
> > > +     if (len < 0)
> > > +             return "unknown";
> > > +
> > > +     buf[len] = '\0';
> > > +     return buf;
> > > +}
> > 
> > Are you intentionally only printing argv[0]? argv items are
> > \0-separated in /proc/*/cmdline.
> 
> Yes, we only want a human friendly process name, or do we need the full
> cmdline because it's java??? If it's java, I'm running away.


Well...


-- 
Petri Latvala