Re: [PATCH perf/core v10 09/23] perf probe: Show all cached probes

From: Masami Hiramatsu
Date: Fri Jun 10 2016 - 20:28:34 EST


On Thu, 9 Jun 2016 11:22:55 -0300
Arnaldo Carvalho de Melo <acme@xxxxxxxxxx> wrote:

> Em Wed, Jun 08, 2016 at 06:30:30PM +0900, Masami Hiramatsu escreveu:
> > +char *build_id_cache__origname(const char *sbuild_id)
> > +{
> > + char *linkname;
> > + char buf[PATH_MAX];
> > + char *ret = NULL, *p;
> > + size_t offs = 5; /* == strlen("../..") */
> > +
> > + linkname = build_id_cache__linkname(sbuild_id, NULL, 0);
> > + if (!linkname)
> > + return NULL;
> > +
> > + if (readlink(linkname, buf, PATH_MAX) < 0)
> > + goto out;
> > + /* The link should be "../..<origpath>/<sbuild_id>" */
> > + p = strrchr(buf, '/'); /* Cut off the "/<sbuild_id>" */
> > + if (p && (p > buf + offs)) {
> > + *p = '\0';
> > + if (buf[offs + 1] == '[')
> > + offs++; /*
> > + * This is a DSO name, like [kernel.kallsyms].
> > + * Skip the first '/', since this is not the
> > + * cache of a regular file.
> > + */
> > + ret = strdup(buf + offs); /* Skip "../..[/]" */
>
> strdup can fail.

Ah, right.

>
> > + }
> > +out:
> > + free(linkname);
> > + return ret;
> > +}
> > +
> > static const char *build_id_cache__basename(bool is_kallsyms, bool is_vdso)
> > {
> > return is_kallsyms ? "kallsyms" : (is_vdso ? "vdso" : "elf");
> > @@ -390,6 +419,59 @@ void disable_buildid_cache(void)
> > no_buildid_cache = true;
> > }
> >
> > +int build_id_cache__list_all(struct strlist **result)
> > +{
> > + struct strlist *toplist, *list, *bidlist;
> > + struct str_node *nd, *nd2;
> > + char *topdir, *linkdir;
> > + char sbuild_id[SBUILD_ID_SIZE];
> > + int ret = 0;
> > +
> > + /* Open the top-level directory */
> > + if (asprintf(&topdir, "%s/.build-id/", buildid_dir) < 0)
> > + return -errno;
>
> Wouldn't be better to just return -1 and leave the caller to look at
> errno? I know that there is code that does that, but better stick to the
> errno convention?

Hmm, IMHO, that may depends on case. This case is possible,
but in below case, pr_debug() also calls libc function and
it can override the errno. This can also happen when we run
some cleanup code before returning.

So, I think it will be good for enough small functions, but
not good for complex bigger functions.

>From the viewpoint of debugging, maybe it is also possible
to return -N, where N is based on the place where the error
happens :)

> > + toplist = lsdir(topdir, lsdir_no_dot_filter);
> > + if (!toplist) {
> > + pr_debug("Failed to opendir %s\n", topdir);
> > + ret = -errno;
>
> ditto
>
> > + goto out;
> > + }
> > + bidlist = strlist__new(NULL, NULL);
> > + strlist__for_each(nd, toplist) {
> > + if (asprintf(&linkdir, "%s/%s", topdir, nd->s) < 0) {
> > + ret = -errno;
> > + goto out;
> > + }
> > + /* Open the lower-level directory */
> > + list = lsdir(linkdir, lsdir_no_dot_filter);
> > + if (!list) {
> > + pr_debug("Failed to open %s: %d\n", linkdir, -errno);
> > + goto next;
> > + }
> > + strlist__for_each(nd2, list) {
> > + ret = snprintf(sbuild_id, SBUILD_ID_SIZE, "%s%s",
> > + nd->s, nd2->s);
> > + if (ret != SBUILD_ID_SIZE - 1) {
> > + pr_debug("%s/%s is not buildid cache\n",
> > + nd->s, nd2->s);
> > + continue;
> > + }
> > + strlist__add(bidlist, sbuild_id);
>
> strlist__add() can fail, please check its result.

Ah, I forgot to update that...

>
> > + }
> > + strlist__delete(list);
> > +next:
> > + free(linkdir);
> > + }
> > +
> > + *result = bidlist;
> > +out:
> > + if (toplist)
> > + strlist__delete(toplist);
>
> No need for checking toplist.

OK.

>
> > + free(topdir);
> > +
> > + return ret;
> > +}
>
> <SNIP>
>
> - Arnaldo

Thank you!

--
Masami Hiramatsu <mhiramat@xxxxxxxxxx>