Re: [PATCH v2 2/2] perf tool: Fix ppid for synthesized fork events

[Arnaldo Carvalho de Melo]

Em Mon, Mar 30, 2015 at 01:53:04PM -0600, David Ahern escreveu:
> On 3/30/15 1:40 PM, Arnaldo Carvalho de Melo wrote:
> >>@@ -81,6 +84,7 @@ static pid_t perf_event__get_comm_tgid(pid_t pid, char *comm, size_t len)
> >>
> >>  	name = strstr(bf, "Name:");
> >>  	tgids = strstr(bf, "Tgid:");
> >>+	ppids = strstr(bf, "PPid:");
> >
> >can't we make this:
> >
> >	ppids = strstr(tgids, "PPid:");
> >
> >To speed it up a teeny little bit? 8-)
> 
> Sure, I thought about that as well, but it puts an assumption on order of
> the data in the file. Why have the assumption?

Good question, perhaps we should leave it as is, should be in the noise,
right?
 
> >>  	if (name) {
> >>  		name += 5;  /* strlen("Name:") */
> >>@@ -109,32 +113,51 @@ static pid_t perf_event__get_comm_tgid(pid_t pid, char *comm, size_t len)
> >>  		if (nl)
> >>  			*nl = '\0';
> >>
> >>-		tgid = atoi(tgids);
> >>+		*tgid = atoi(tgids);
> >>
> >>  	} else
> >>  		pr_debug("Tgid: string not found for pid %d\n", pid);
> >>
> >>-	return tgid;
> >>+	if (ppids) {
> >>+		ppids += 5;  /* strlen("PPid:") */
> >>+
> >>+		while (*ppids && isspace(*ppids))
> >>+			++ppids;
> >
> >The above could be simplified to:
> >
> >		while (isspace(*ppids))
> >			++ppids;
> 
> sure.

And then, the this loop can be elided as well, as you can see below...

> >$ cat isspace.c
> >#include <ctype.h>
> >#include <stdio.h>
> >int main(void) { return printf("isspace('\\0')=%d\n", isspace('\0')); }
> >$ ./isspace
> >isspace('\0')=0
> >$
> >
> >>+		nl = strchr(ppids, '\n');
> >>+		if (nl)
> >>+			 *nl = '\0';
> >
> >We also don't need to find and zero this '\n', as:
> >
> >$ cat atoi.c
> >#include <string.h>
> >#include <stdio.h>
> >int main(void) { return printf("atoi(\"1234\\n\")=%d\n",
> >atoi("1234\n")); }
> >$ ./atoi
> >atoi("1234\n")=1234
> >$
> 
> ok. another assumption on implementation. fine with taking it out.

Not really, POSIX sayz:

http://pubs.opengroup.org/onlinepubs/9699919799/functions/atoi.html

The call atoi(str) shall be equivalent to:

(int) strtol(str, (char **)NULL, 10)

----

And:

http://pubs.opengroup.org/onlinepubs/9699919799/functions/strtol.html

These functions shall convert the initial portion of the string pointed
to by str to a type long and long long representation, respectively.
First, they decompose the input string into three parts:

An initial, possibly empty, sequence of white-space characters (as
specified by isspace())

A subject sequence interpreted as an integer represented in some radix
determined by the value of base

A final string of one or more unrecognized characters, including the
terminating NUL character of the input string.

----------

So even that isspace loop can be ditched, no?

> ><SNIP>
> >
> >>+	if (machine__is_host(machine)) {
> >>+		if (perf_event__get_comm_ids(pid, event->comm.comm,
> >>+					     sizeof(event->comm.comm),
> >>+					     tgid, ppid) != 0) {
> >>+			return -1;
> >>+		}
> >>+	} else
> >>+		*tgid = machine->pid;
> >
> >Somebody, I think PeterZ and also Ingo, routinely asks for having {} on
> >the else part of an if that has {}, please do so.
> 
> ack

Thanks!

- Arnaldo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/