lkml.org 
[lkml]   [2009]   [May]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 7/7] tracing: add hierarchical enabling of events

On Thu, 7 May 2009, Li Zefan wrote:

> > With the current event directory, you can only enable individual events.
> > The file debugfs/tracing/set_event is used to be able to enable or
> > disable several events at once. But that can still be awkward.
> >
> > This patch adds hierarchical enabling of events. That is, each directory
> > in debugfs/tracing/events has an "enable" file. This file can enable
> > or disable all events within the directory and below.
> >
> > # echo 1 > /debugfs/tracing/events/enable
> >
> > will enable all events.
> >
> > # echo 1 > /debugfs/tracing/events/sched/enable
> >
> > will enable all events in the sched subsystem.
> >
> > # echo 1 > /debugfs/tracing/events/enable
> > # echo 0 > /debugfs/tracing/events/irq/enable
> >
> > will enable all events, but then disable just the irq subsystem events.
> >
> > When reading one of these enable files, there are four results:
> >
> > 0 - all events this file affects are disabled
> > 1 - all events this file affects are enabled
> > X - there is a mixture of events enabled and disabled
> > ? - this file does not affect any event
>
> I would expect reading an enable file will let me know exactly
> which events are disabled and which are enabled.
>
> I think this is useful especially for events/system/enable.
>
> Like this:
>
> $ cat events/irq/enable
> 0 irq_handler_entry
> 0 irq_handler_exit
> 1 softirq_entry
> 1 softirq_exit

I thought about doing something like this, but this idea for the
hierarchical enabling came to me around 11pm, and I had the code written
by 11:15pm ;-)

Which means, I figured I would do it as simple as possible. We do have
"set_event" that gives you a list of enabled events. My thought was still
having a "1" or "0" if all are either enabled or disabled. And when it is
a mixture, I would have a list of enabled events.

Though, it is useful. Maybe in the future. But really, the information is
there, and I did not expect this to be a "what is enabled" file, but
instead a "I want to enable/disable all these events". In other words, I
was much more interested in the "write" ability than the read. But who
knows, maybe this will change in the future.

>
> > +static ssize_t
> > +system_enable_read(struct file *filp, char __user *ubuf, size_t cnt,
> > + loff_t *ppos)
> > +{
> > + const char *system = filp->private_data;
> > + struct ftrace_event_call *call;
> > + char buf[2];
> > + int set = -1;
> > + int all = 0;
> > + int ret;
> > +
> > + if (system[0] == '*')
> > + all = 1;
> > +
> > + mutex_lock(&event_mutex);
> > + list_for_each_entry(call, &ftrace_events, list) {
> > + if (!call->name || !call->regfunc)
> > + continue;
> > +
> > + if (!all && strcmp(call->system, system) != 0)
> > + continue;
> > +
> > + /*
> > + * We need to find out if all the events are set
> > + * or if all events or cleared, or if we have
> > + * a mixture.
> > + */
> > + if (call->enabled) {
> > + switch (set) {
> > + case -1:
> > + set = 1;
> > + break;
> > + case 0:
> > + set = 2;
> > + break;
> > + }
> > + } else {
> > + switch (set) {
> > + case -1:
> > + set = 0;
> > + break;
> > + case 1:
> > + set = 2;
> > + break;
> > + }
> > + }
> > + /*
> > + * If we have a mixture, no need to look further.
> > + */
> > + if (set == 2)
> > + break;
>
> How about:
>
> int set = 0;
>
> ...
> set |= (1 << call->enabled);

* paranoid *

set |= (1 << !!call->enabled);

> ...
>
> set == 0: '?'
> set == 1: '0'
> set == 2: '1'
> set == 3: 'X'
>
> Will this make the code simpler? :)
>
> Or we can go even further:
>
> char result[4] = { '?', '0', '1', 'X' };
> ...
> buf[0] = result[set];

cute, mind sending a patch ;-)


>
> > + }
> > + mutex_unlock(&event_mutex);
> > +
> > + buf[1] = '\n';
> > + switch (set) {
> > + case 0:
> > + buf[0] = '0';
> > + break;
> > + case 1:
> > + buf[0] = '1';
> > + break;
> > + case 2:
> > + buf[0] = 'X';
> > + break;
> > + default:
> > + buf[0] = '?';
> > + }
> > +
> > + ret = simple_read_from_buffer(ubuf, cnt, ppos, buf, 2);
> > +
> > + return ret;
> > +}
> > +
> > +static ssize_t
> > +system_enable_write(struct file *filp, const char __user *ubuf, size_t cnt,
> > + loff_t *ppos)
> > +{
> > + const char *system = filp->private_data;
> > + unsigned long val;
> > + char *command;
> > + char buf[64];
> > + ssize_t ret;
> > +
> > + if (cnt >= sizeof(buf))
> > + return -EINVAL;
> > +
> > + if (copy_from_user(&buf, ubuf, cnt))
> > + return -EFAULT;
> > +
> > + buf[cnt] = 0;
> > +
> > + ret = strict_strtoul(buf, 10, &val);
> > + if (ret < 0)
> > + return ret;
> > +
> > + ret = tracing_update_buffers();
> > + if (ret < 0)
> > + return ret;
> > +
> > + switch (val) {
> > + case 0:
> > + case 1:
> > + break;
> > +
> > + default:
> > + return -EINVAL;
> > + }
> > +
> > + command = kstrdup(system, GFP_KERNEL);
> > + if (!command)
> > + return -ENOMEM;
> > +
> > + ret = ftrace_set_clr_event(command, val);
>
> I think we should pass "sched:" or "sched:*", instead of "sched",
> the comment in ftrace_set_clr_event():
>
> * <name> (no ':') means all events in a subsystem with
> * the name <name> or any event that matches <name>

Yeah, I thought about it too. But writing the patch in 15 minutes, I
decided that a "kstrdup" was easier than adding a ":" ;-)

I will update it.

Thanks!

-- Steve

>
> > + if (ret)
> > + goto out_free;
> > +
> > + ret = cnt;
> > +
> > + out_free:
> > + kfree(command);
> > +
> > + *ppos += cnt;
> > +
> > + return ret;
> > +}
> > +
>
>


\
 
 \ /
  Last update: 2009-05-07 15:25    [W:0.091 / U:0.384 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site