Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-16 Thread Pekka Paalanen
On Fri, 13 Dec 2019 11:54:33 -0500 Sean Paul wrote: > On Fri, Dec 13, 2019 at 01:34:46PM +0200, Pekka Paalanen wrote: > > On Thu, 12 Dec 2019 15:32:35 -0500 > > Sean Paul wrote: > > > > > From: Sean Paul > > > > > > For a long while now, we (ChromeOS) have been struggling getting any > > >

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Sean Paul
On Fri, Dec 13, 2019 at 11:54:36AM -0500, Steven Rostedt wrote: > On Fri, 13 Dec 2019 11:47:03 -0500 > Sean Paul wrote: > > > > Why is there a separate trace event for each of these? > > > > > > > To make it easier on userspace to enable just a single drm category. > > > > But trace events

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Sean Paul
On Fri, Dec 13, 2019 at 01:34:46PM +0200, Pekka Paalanen wrote: > On Thu, 12 Dec 2019 15:32:35 -0500 > Sean Paul wrote: > > > From: Sean Paul > > > > For a long while now, we (ChromeOS) have been struggling getting any > > value out of user feedback reports of display failures (notably external

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Steven Rostedt
On Fri, 13 Dec 2019 11:47:03 -0500 Sean Paul wrote: > > Why is there a separate trace event for each of these? > > > > To make it easier on userspace to enable just a single drm category. > But trace events can easily be filtered. Say you have a field called category, you could then just se

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Sean Paul
Hi Steven, Thanks for the feedback! On Thu, Dec 12, 2019 at 10:20:50PM -0500, Steven Rostedt wrote: > On Thu, 12 Dec 2019 15:32:35 -0500 > Sean Paul wrote: > > > > > What about trace_printk()? It doesn't give us the control we get from using > > tracepoints and it's not meant to be left sprin

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Steven Rostedt
On Fri, 13 Dec 2019 13:34:46 +0200 Pekka Paalanen wrote: > Is there a way to pull out messages > from /sys/kernel/debug/tracing/trace and filter them on reading instead > of on writing? FYI, you can mount the tracing directory without having to mount the debug directory. mount -t tracefs nodev

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Ville Syrjälä
On Fri, Dec 13, 2019 at 05:34:25PM +0200, Jani Nikula wrote: > On Thu, 12 Dec 2019, Sean Paul wrote: > > From: Sean Paul > > > > For a long while now, we (ChromeOS) have been struggling getting any > > value out of user feedback reports of display failures (notably external > > displays not worki

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Jani Nikula
On Thu, 12 Dec 2019, Sean Paul wrote: > From: Sean Paul > > For a long while now, we (ChromeOS) have been struggling getting any > value out of user feedback reports of display failures (notably external > displays not working). The problem is that all logging, even fatal > errors (well, fatal in

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-13 Thread Pekka Paalanen
On Thu, 12 Dec 2019 15:32:35 -0500 Sean Paul wrote: > From: Sean Paul > > For a long while now, we (ChromeOS) have been struggling getting any > value out of user feedback reports of display failures (notably external > displays not working). The problem is that all logging, even fatal > errors

Re: [PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-12 Thread Steven Rostedt
On Thu, 12 Dec 2019 15:32:35 -0500 Sean Paul wrote: > > What about trace_printk()? It doesn't give us the control we get from using > tracepoints and it's not meant to be left sprinkled around in code. Not to mention that trace_printk() is not for production use (only for developers debugging p

[PATCH v3] drm: Funnel drm logs to tracepoints

2019-12-12 Thread Sean Paul
From: Sean Paul For a long while now, we (ChromeOS) have been struggling getting any value out of user feedback reports of display failures (notably external displays not working). The problem is that all logging, even fatal errors (well, fatal in the sense that a display won't light up) are logg