Messages in this thread | | | Date | Tue, 10 May 2011 10:09:18 +0200 | From | Ingo Molnar <> | Subject | Re: Fix powerTOP regression with 2.6.39-rc5 |
| |
* Steven Rostedt <rostedt@goodmis.org> wrote:
> > > > PowerTop uses the perf ABI because it's a rather convenient and unified > > > > method to get a rich selection of events via the same facility, same > > > > ring-buffer, using a system call ABI, etc. > > > > > > It seams that Arjan read the perf kernel code to see how the raw data was > > > laid out and read it directly. > > > > Yes, of course - i also have code that uses the syscall directly, it's > > easier to code up ad-hoc than to parse some XML-lookalike descriptor from > > /debug/tracing/events/. > > So you admit this is a ad-hoc way of doing things. [...]
There's nothing wrong with being able to code and make use of instrumentation ad-hoc, using a simple interface.
The *worst* possible thing is to force user-space to use something complex. So we kept the perf ABI simple.
> [...] Thus a library is a perfect solution. [...]
That's a non sequitor.
> [...] And the event formats are far from XML-look-a-like. You really think > this: > > name: sched_switch > ID: 57 > format: > field:unsigned short common_type; offset:0; size:2; > field:unsigned char common_flags; offset:2; size:1; > field:unsigned char common_preempt_count; offset:3; size:1; > field:int common_pid; offset:4; size:4; > field:int common_lock_depth; offset:8; size:4; > > field:char prev_comm[TASK_COMM_LEN]; offset:12; size:16; > field:pid_t prev_pid; offset:28; size:4; > field:int prev_prio; offset:32; size:4; > field:long prev_state; offset:40; size:8; > field:char next_comm[TASK_COMM_LEN]; offset:48; size:16; > field:pid_t next_pid; offset:64; size:4; > field:int next_prio; offset:68; size:4; > > Is equivalent to this: > > <?xml version="1.0" encoding="UTF-8"?> > <KernelShark><CaptureSettings><Events><CaptureType>Events</CaptureType><System>sched</System></Events><Plugin>function_graph</Plugin><File>/tmp/trace.dat</File></CaptureSettings></KernelShark>
I did not say that it's equivalent, i said it's XML look-alike.
Steve, we even joked about that, that if we continue like this we'll end up with an XML parser ... I requested several changes to the description format so that it becomes more human readable.
Thanks,
Ingo
| |