[PATCH 4.9] perf probe: Fix --funcs to show correct symbols for offline module

[PATCH 4.9] perf probe: Fix --funcs to show correct symbols for offline module

[Krister Johansen]

Hi Greg,
There's a bit more fallout from pulling in the patches we discussed in
this thread:

https://lkml.org/lkml/2017/7/5/537

While those patches fixed the build, I didn't get a chance to run a full
set of tests on the result.  Last week, when attempting to set a
kretprobe in a kernel module using 'perf probe' I hit an assertion that
triggers when you attempt to delete an object that's still in a tree.

It turned out that the patch below was also pushed around the same time
as the others that had been pulled into 4.9, and so we'd missed a case
where a refcount was not getting correctly incremented.  The patch below
is just a git cherry-pick of the commit from HEAD to the stable 4.9
branch.  This addressed the refcount assertion failure I was seeing.

Thanks,

-K

-------------------------------------------------

From: Masami Hiramatsu <mhiramat@kernel.org>

commit eebc509b20881b92d62e317b2c073e57c5f200f0 upstream.

Fix --funcs (-F) option to show correct symbols for offline module.
Since previous perf-probe uses machine__findnew_module_map() for offline
module, even if user passes a module file (with full path) which is for
other architecture, perf-probe always tries to load symbol map for
current kernel module.

This fix uses dso__new_map() to load the map from given binary as same
as a map for user applications.

Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/148350053478.19001.15435255244512631545.stgit@devbox
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/probe-event.c | 16 ++++------------
 1 file changed, 4 insertions(+), 12 deletions(-)

diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 6c50d9f..b7c7f42 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -163,7 +163,7 @@ static struct map *kernel_get_module_map(const char *module)
 
 	/* A file path -- this is an offline module */
 	if (module && strchr(module, '/'))
-		return machine__findnew_module_map(host_machine, 0, module);
+		return dso__new_map(module);
 
 	if (!module)
 		module = "kernel";
@@ -173,6 +173,7 @@ static struct map *kernel_get_module_map(const char *module)
 		if (strncmp(pos->dso->short_name + 1, module,
 			    pos->dso->short_name_len - 2) == 0 &&
 		    module[pos->dso->short_name_len - 2] == '\0') {
+			map__get(pos);
 			return pos;
 		}
 	}
@@ -188,15 +189,6 @@ struct map *get_target_map(const char *target, bool user)
 		return kernel_get_module_map(target);
 }
 
-static void put_target_map(struct map *map, bool user)
-{
-	if (map && user) {
-		/* Only the user map needs to be released */
-		map__put(map);
-	}
-}
-
-
 static int convert_exec_to_group(const char *exec, char **result)
 {
 	char *ptr1, *ptr2, *exec_copy;
@@ -412,7 +404,7 @@ static int find_alternative_probe_point(struct debuginfo *dinfo,
 	}
 
 out:
-	put_target_map(map, uprobes);
+	map__put(map);
 	return ret;
 
 }
@@ -2944,7 +2936,7 @@ static int find_probe_trace_events_from_map(struct perf_probe_event *pev,
 	}
 
 out:
-	put_target_map(map, pev->uprobes);
+	map__put(map);
 	free(syms);
 	return ret;
 
-- 
2.10.1 (Apple Git-78)

Re: [PATCH 4.9] perf probe: Fix --funcs to show correct symbols for offline module

[Greg KH]

On Fri, Aug 25, 2017 at 04:05:35PM -0700, Krister Johansen wrote:
> Hi Greg,
> There's a bit more fallout from pulling in the patches we discussed in
> this thread:
> 
> https://lkml.org/lkml/2017/7/5/537
> 
> While those patches fixed the build, I didn't get a chance to run a full
> set of tests on the result.  Last week, when attempting to set a
> kretprobe in a kernel module using 'perf probe' I hit an assertion that
> triggers when you attempt to delete an object that's still in a tree.
> 
> It turned out that the patch below was also pushed around the same time
> as the others that had been pulled into 4.9, and so we'd missed a case
> where a refcount was not getting correctly incremented.  The patch below
> is just a git cherry-pick of the commit from HEAD to the stable 4.9
> branch.  This addressed the refcount assertion failure I was seeing.

Thanks, now queued up.

greg k-h