(2011/08/04 2:16), Borislav Petkov wrote: > On Sat, Jul 09, 2011 at 03:11:45PM +0900, Masami Hiramatsu wrote: >> BTW, I'm also interested in that structured error events, from the long >> term view and viewpoint of tracers :) > > Let me chime in a "mee too" from the HW errors perspective. Of course, :) and I guess this may be related to Fredrdic's work. >> I think we could expand current TRACE_EVENT macro to define those >> error events. > > Concerning structure, we can use the format file in debugfs which > currently describes the fields of a tracepoint: > > /sys/kernel/debug/tracing/events/../../format > > and expand those to generic kernel events. Userspace parses the format > and knows exactly at what data it is looking at. Also, the idea is to > move that events hierarchy into sysfs so it will be present on every > system. Yeah, that is exactly what I've thought. Current trace events strongly depends on ftrace and perf. And both of them are not enough for handling events as error monitoring. Maybe we have to prepare another monitor for this purpose, because, usually, error monitoring should be run always on the machine, separately from (and concurrently with) performance tuning or debugging tools. Even though, I think the trace-event macros (include/trace/events/) are good starting point. Thank you, -- Masami HIRAMATSU Software Platform Research Dept. Linux Technology Center Hitachi, Ltd., Yokohama Research Laboratory E-mail: masami.hiramatsu.pt@xxxxxxxxxxx -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
