[PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Marc Kleine-Budde]

In commit "403567217d3f perf symbols: Do not read symbols/data from
device files" a check to uninitialzied memory was added. This leads to
the following valgrind output:

==24515== Syscall param stat(file_name) points to uninitialised byte(s)
==24515==    at 0x75B26D5: _xstat (in /lib/x86_64-linux-gnu/libc-2.22.so)
==24515==    by 0x4E548D: stat (stat.h:454)
==24515==    by 0x4E548D: is_regular_file (util.c:687)
==24515==    by 0x4A5BEE: dso__load (symbol.c:1435)
==24515==    by 0x4BB1AE: map__load (map.c:289)
==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
==24515==    by 0x441B3E: process_event (builtin-script.c:795)
==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
==24515==  Address 0x807c6a0 is 0 bytes inside a block of size 4,096 alloc'd
==24515==    at 0x4C29C0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==24515==    by 0x4A5BCB: dso__load (symbol.c:1421)
==24515==    by 0x4BB1AE: map__load (map.c:289)
==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
==24515==    by 0x441B3E: process_event (builtin-script.c:795)
==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
==24515==    by 0x4C0EAC: __perf_session__process_events (session.c:1804)
==24515==    by 0x4C0EAC: perf_session__process_events (session.c:1858)

This patch fixes the problem by removing the check.

Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
---
 tools/perf/util/symbol.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 8a747dc6cf86..b4070daeb17f 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
 	 * Read the build id if possible. This is required for
 	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
 	 */
-	if (is_regular_file(name) &&
-	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
+	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
 		dso__set_build_id(dso, build_id);
 
 	/*
-- 
2.8.1

Re: [PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Jiri Olsa]

On Thu, Jun 23, 2016 at 04:20:21PM +0200, Marc Kleine-Budde wrote:
> In commit "403567217d3f perf symbols: Do not read symbols/data from
> device files" a check to uninitialzied memory was added. This leads to
> the following valgrind output:
> 
> ==24515== Syscall param stat(file_name) points to uninitialised byte(s)
> ==24515==    at 0x75B26D5: _xstat (in /lib/x86_64-linux-gnu/libc-2.22.so)
> ==24515==    by 0x4E548D: stat (stat.h:454)
> ==24515==    by 0x4E548D: is_regular_file (util.c:687)
> ==24515==    by 0x4A5BEE: dso__load (symbol.c:1435)
> ==24515==    by 0x4BB1AE: map__load (map.c:289)
> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
> ==24515==  Address 0x807c6a0 is 0 bytes inside a block of size 4,096 alloc'd
> ==24515==    at 0x4C29C0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==24515==    by 0x4A5BCB: dso__load (symbol.c:1421)
> ==24515==    by 0x4BB1AE: map__load (map.c:289)
> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
> ==24515==    by 0x4C0EAC: __perf_session__process_events (session.c:1804)
> ==24515==    by 0x4C0EAC: perf_session__process_events (session.c:1858)
> 
> This patch fixes the problem by removing the check.
> 
> Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
> ---
>  tools/perf/util/symbol.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index 8a747dc6cf86..b4070daeb17f 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
>  	 * Read the build id if possible. This is required for
>  	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
>  	 */
> -	if (is_regular_file(name) &&
> -	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
> +	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>  		dso__set_build_id(dso, build_id);

ouch copy&paste error.. we better fix it properly,
could you please check attached patch?

thanks,
jirka


---
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index b044f1a32d16..37e8d20ae03e 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -1430,7 +1430,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
 	 * Read the build id if possible. This is required for
 	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
 	 */
-	if (is_regular_file(name) &&
+	if (is_regular_file(dso->long_name) &&
 	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
 		dso__set_build_id(dso, build_id);
 

Re: [PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Marc Kleine-Budde]

[-- Attachment #1.1: Type: text/plain, Size: 5712 bytes --]

On 06/24/2016 08:43 AM, Jiri Olsa wrote:
> On Thu, Jun 23, 2016 at 04:20:21PM +0200, Marc Kleine-Budde wrote:
>> In commit "403567217d3f perf symbols: Do not read symbols/data from
>> device files" a check to uninitialzied memory was added. This leads to
>> the following valgrind output:
>>
>> ==24515== Syscall param stat(file_name) points to uninitialised byte(s)
>> ==24515==    at 0x75B26D5: _xstat (in /lib/x86_64-linux-gnu/libc-2.22.so)
>> ==24515==    by 0x4E548D: stat (stat.h:454)
>> ==24515==    by 0x4E548D: is_regular_file (util.c:687)
>> ==24515==    by 0x4A5BEE: dso__load (symbol.c:1435)
>> ==24515==    by 0x4BB1AE: map__load (map.c:289)
>> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
>> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
>> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
>> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
>> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
>> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
>> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
>> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
>> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
>> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
>> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
>> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
>> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
>> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
>> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
>> ==24515==  Address 0x807c6a0 is 0 bytes inside a block of size 4,096 alloc'd
>> ==24515==    at 0x4C29C0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
>> ==24515==    by 0x4A5BCB: dso__load (symbol.c:1421)
>> ==24515==    by 0x4BB1AE: map__load (map.c:289)
>> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
>> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
>> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
>> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
>> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
>> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
>> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
>> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
>> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
>> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
>> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
>> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
>> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
>> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
>> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
>> ==24515==    by 0x4C0EAC: __perf_session__process_events (session.c:1804)
>> ==24515==    by 0x4C0EAC: perf_session__process_events (session.c:1858)
>>
>> This patch fixes the problem by removing the check.
>>
>> Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
>> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
>> ---
>>  tools/perf/util/symbol.c | 3 +--
>>  1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
>> index 8a747dc6cf86..b4070daeb17f 100644
>> --- a/tools/perf/util/symbol.c
>> +++ b/tools/perf/util/symbol.c
>> @@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
>>  	 * Read the build id if possible. This is required for
>>  	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
>>  	 */
>> -	if (is_regular_file(name) &&
>> -	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>> +	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>>  		dso__set_build_id(dso, build_id);
> 
> ouch copy&paste error.. we better fix it properly,
> could you please check attached patch?

The use of uninitialized memory is gone, but the "--symfs" option is
ignored:

> stat("/lib/libpthread-2.20.so", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)

You'll look for the dso in the root file system not in the sysroot.

> stat("./root//lib/libpthread-2.20.so", {st_mode=S_IFREG|0755, st_size=68048, ...}) = 0
> open("./root//lib/libpthread-2.20.so", O_RDONLY) = 4
> stat("./root//lib/libpthread-2.20.so", {st_mode=S_IFREG|0755, st_size=68048, ...}) = 0
> open("./root//lib/libpthread-2.20.so", O_RDONLY) = 4
> stat("./root//usr/lib/debug/lib/libpthread-2.20.so.debug", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)
> stat("./root//usr/lib/debug/lib/libpthread-2.20.so", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)
> stat("./root//lib/libpthread-2.20.so", {st_mode=S_IFREG|0755, st_size=68048, ...}) = 0
> open("./root//lib/libpthread-2.20.so", O_RDONLY) = 5
> stat("./root//lib/.debug/libpthread-2.20.so", {st_mode=S_IFREG|0644, st_size=266508, ...}) = 0
> open("./root//lib/.debug/libpthread-2.20.so", O_RDONLY) = 5

Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Jiri Olsa]

On Fri, Jun 24, 2016 at 09:54:22AM +0200, Marc Kleine-Budde wrote:
> On 06/24/2016 08:43 AM, Jiri Olsa wrote:
> > On Thu, Jun 23, 2016 at 04:20:21PM +0200, Marc Kleine-Budde wrote:
> >> In commit "403567217d3f perf symbols: Do not read symbols/data from
> >> device files" a check to uninitialzied memory was added. This leads to
> >> the following valgrind output:
> >>
> >> ==24515== Syscall param stat(file_name) points to uninitialised byte(s)
> >> ==24515==    at 0x75B26D5: _xstat (in /lib/x86_64-linux-gnu/libc-2.22.so)
> >> ==24515==    by 0x4E548D: stat (stat.h:454)
> >> ==24515==    by 0x4E548D: is_regular_file (util.c:687)
> >> ==24515==    by 0x4A5BEE: dso__load (symbol.c:1435)
> >> ==24515==    by 0x4BB1AE: map__load (map.c:289)
> >> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
> >> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
> >> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
> >> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
> >> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
> >> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
> >> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
> >> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
> >> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
> >> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
> >> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
> >> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
> >> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
> >> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
> >> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
> >> ==24515==  Address 0x807c6a0 is 0 bytes inside a block of size 4,096 alloc'd
> >> ==24515==    at 0x4C29C0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> >> ==24515==    by 0x4A5BCB: dso__load (symbol.c:1421)
> >> ==24515==    by 0x4BB1AE: map__load (map.c:289)
> >> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
> >> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
> >> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
> >> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
> >> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
> >> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
> >> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
> >> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
> >> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
> >> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
> >> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
> >> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
> >> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
> >> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
> >> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
> >> ==24515==    by 0x4C0EAC: __perf_session__process_events (session.c:1804)
> >> ==24515==    by 0x4C0EAC: perf_session__process_events (session.c:1858)
> >>
> >> This patch fixes the problem by removing the check.
> >>
> >> Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
> >> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
> >> ---
> >>  tools/perf/util/symbol.c | 3 +--
> >>  1 file changed, 1 insertion(+), 2 deletions(-)
> >>
> >> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> >> index 8a747dc6cf86..b4070daeb17f 100644
> >> --- a/tools/perf/util/symbol.c
> >> +++ b/tools/perf/util/symbol.c
> >> @@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
> >>  	 * Read the build id if possible. This is required for
> >>  	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
> >>  	 */
> >> -	if (is_regular_file(name) &&
> >> -	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
> >> +	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
> >>  		dso__set_build_id(dso, build_id);
> > 
> > ouch copy&paste error.. we better fix it properly,
> > could you please check attached patch?
> 
> The use of uninitialized memory is gone, but the "--symfs" option is
> ignored:
> 
> > stat("/lib/libpthread-2.20.so", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)
> 
> You'll look for the dso in the root file system not in the sysroot.

that seems like unrelated to the 'use of uninitialized memory' issue, right?
I can't see how the patch I sent could do that..

thanks,
jirka

Re: [PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Marc Kleine-Budde]

[-- Attachment #1.1: Type: text/plain, Size: 5403 bytes --]

On 06/24/2016 01:51 PM, Jiri Olsa wrote:
> On Fri, Jun 24, 2016 at 09:54:22AM +0200, Marc Kleine-Budde wrote:
>> On 06/24/2016 08:43 AM, Jiri Olsa wrote:
>>> On Thu, Jun 23, 2016 at 04:20:21PM +0200, Marc Kleine-Budde wrote:
>>>> In commit "403567217d3f perf symbols: Do not read symbols/data from
>>>> device files" a check to uninitialzied memory was added. This leads to
>>>> the following valgrind output:
>>>>
>>>> ==24515== Syscall param stat(file_name) points to uninitialised byte(s)
>>>> ==24515==    at 0x75B26D5: _xstat (in /lib/x86_64-linux-gnu/libc-2.22.so)
>>>> ==24515==    by 0x4E548D: stat (stat.h:454)
>>>> ==24515==    by 0x4E548D: is_regular_file (util.c:687)
>>>> ==24515==    by 0x4A5BEE: dso__load (symbol.c:1435)
>>>> ==24515==    by 0x4BB1AE: map__load (map.c:289)
>>>> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
>>>> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
>>>> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
>>>> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
>>>> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
>>>> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
>>>> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
>>>> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
>>>> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
>>>> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
>>>> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
>>>> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
>>>> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
>>>> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
>>>> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
>>>> ==24515==  Address 0x807c6a0 is 0 bytes inside a block of size 4,096 alloc'd
>>>> ==24515==    at 0x4C29C0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
>>>> ==24515==    by 0x4A5BCB: dso__load (symbol.c:1421)
>>>> ==24515==    by 0x4BB1AE: map__load (map.c:289)
>>>> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
>>>> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
>>>> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
>>>> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
>>>> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
>>>> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
>>>> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
>>>> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
>>>> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
>>>> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
>>>> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
>>>> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
>>>> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
>>>> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
>>>> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
>>>> ==24515==    by 0x4C0EAC: __perf_session__process_events (session.c:1804)
>>>> ==24515==    by 0x4C0EAC: perf_session__process_events (session.c:1858)
>>>>
>>>> This patch fixes the problem by removing the check.
>>>>
>>>> Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
>>>> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
>>>> ---
>>>>  tools/perf/util/symbol.c | 3 +--
>>>>  1 file changed, 1 insertion(+), 2 deletions(-)
>>>>
>>>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
>>>> index 8a747dc6cf86..b4070daeb17f 100644
>>>> --- a/tools/perf/util/symbol.c
>>>> +++ b/tools/perf/util/symbol.c
>>>> @@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
>>>>  	 * Read the build id if possible. This is required for
>>>>  	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
>>>>  	 */
>>>> -	if (is_regular_file(name) &&
>>>> -	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>>>> +	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>>>>  		dso__set_build_id(dso, build_id);
>>>
>>> ouch copy&paste error.. we better fix it properly,
>>> could you please check attached patch?
>>
>> The use of uninitialized memory is gone, but the "--symfs" option is
>> ignored:
>>
>>> stat("/lib/libpthread-2.20.so", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)
>>
>> You'll look for the dso in the root file system not in the sysroot.
> 
> that seems like unrelated to the 'use of uninitialized memory' issue, right?
> I can't see how the patch I sent could do that..

Right. That's a different issue. You can add my:

Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>

thanks,
Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Marc Kleine-Budde]

[-- Attachment #1.1: Type: text/plain, Size: 5621 bytes --]

On 06/24/2016 01:56 PM, Marc Kleine-Budde wrote:
> On 06/24/2016 01:51 PM, Jiri Olsa wrote:
>> On Fri, Jun 24, 2016 at 09:54:22AM +0200, Marc Kleine-Budde wrote:
>>> On 06/24/2016 08:43 AM, Jiri Olsa wrote:
>>>> On Thu, Jun 23, 2016 at 04:20:21PM +0200, Marc Kleine-Budde wrote:
>>>>> In commit "403567217d3f perf symbols: Do not read symbols/data from
>>>>> device files" a check to uninitialzied memory was added. This leads to
>>>>> the following valgrind output:
>>>>>
>>>>> ==24515== Syscall param stat(file_name) points to uninitialised byte(s)
>>>>> ==24515==    at 0x75B26D5: _xstat (in /lib/x86_64-linux-gnu/libc-2.22.so)
>>>>> ==24515==    by 0x4E548D: stat (stat.h:454)
>>>>> ==24515==    by 0x4E548D: is_regular_file (util.c:687)
>>>>> ==24515==    by 0x4A5BEE: dso__load (symbol.c:1435)
>>>>> ==24515==    by 0x4BB1AE: map__load (map.c:289)
>>>>> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
>>>>> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
>>>>> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
>>>>> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
>>>>> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
>>>>> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
>>>>> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
>>>>> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
>>>>> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
>>>>> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
>>>>> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
>>>>> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
>>>>> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
>>>>> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
>>>>> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
>>>>> ==24515==  Address 0x807c6a0 is 0 bytes inside a block of size 4,096 alloc'd
>>>>> ==24515==    at 0x4C29C0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
>>>>> ==24515==    by 0x4A5BCB: dso__load (symbol.c:1421)
>>>>> ==24515==    by 0x4BB1AE: map__load (map.c:289)
>>>>> ==24515==    by 0x4BB1AE: map__find_symbol (map.c:333)
>>>>> ==24515==    by 0x4835B3: thread__find_addr_location (event.c:1300)
>>>>> ==24515==    by 0x4B5342: add_callchain_ip (machine.c:1652)
>>>>> ==24515==    by 0x4B5342: thread__resolve_callchain_sample (machine.c:1906)
>>>>> ==24515==    by 0x4B9E7D: thread__resolve_callchain (machine.c:1958)
>>>>> ==24515==    by 0x441B3E: process_event (builtin-script.c:795)
>>>>> ==24515==    by 0x441B3E: process_sample_event (builtin-script.c:920)
>>>>> ==24515==    by 0x4BEE29: perf_evlist__deliver_sample (session.c:1192)
>>>>> ==24515==    by 0x4BEE29: machines__deliver_event (session.c:1229)
>>>>> ==24515==    by 0x4BF770: perf_session__deliver_event (session.c:1286)
>>>>> ==24515==    by 0x4BF770: ordered_events__deliver_event (session.c:114)
>>>>> ==24515==    by 0x4C1D17: __ordered_events__flush (ordered-events.c:207)
>>>>> ==24515==    by 0x4C1D17: ordered_events__flush.part.3 (ordered-events.c:274)
>>>>> ==24515==    by 0x4BF44C: perf_session__process_user_event (session.c:1325)
>>>>> ==24515==    by 0x4BF44C: perf_session__process_event (session.c:1451)
>>>>> ==24515==    by 0x4C0EAC: __perf_session__process_events (session.c:1804)
>>>>> ==24515==    by 0x4C0EAC: perf_session__process_events (session.c:1858)
>>>>>
>>>>> This patch fixes the problem by removing the check.
>>>>>
>>>>> Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
>>>>> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
>>>>> ---
>>>>>  tools/perf/util/symbol.c | 3 +--
>>>>>  1 file changed, 1 insertion(+), 2 deletions(-)
>>>>>
>>>>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
>>>>> index 8a747dc6cf86..b4070daeb17f 100644
>>>>> --- a/tools/perf/util/symbol.c
>>>>> +++ b/tools/perf/util/symbol.c
>>>>> @@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
>>>>>  	 * Read the build id if possible. This is required for
>>>>>  	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
>>>>>  	 */
>>>>> -	if (is_regular_file(name) &&
>>>>> -	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>>>>> +	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
>>>>>  		dso__set_build_id(dso, build_id);
>>>>
>>>> ouch copy&paste error.. we better fix it properly,
>>>> could you please check attached patch?
>>>
>>> The use of uninitialized memory is gone, but the "--symfs" option is
>>> ignored:
>>>
>>>> stat("/lib/libpthread-2.20.so", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)
>>>
>>> You'll look for the dso in the root file system not in the sysroot.
>>
>> that seems like unrelated to the 'use of uninitialized memory' issue, right?
>> I can't see how the patch I sent could do that..
> 
> Right. That's a different issue. You can add my:
> 
> Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>

I don't see this patch on mainline nor on linux-next, or am I missing
something?

Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [PATCH] perf: symbols: dso__load(): remove check of file on uninitialized memory

[Jiri Olsa]

On Thu, Jul 21, 2016 at 09:21:00AM +0200, Marc Kleine-Budde wrote:

SNIP

> >>>>> This patch fixes the problem by removing the check.
> >>>>>
> >>>>> Fixes: 403567217d3f perf symbols: Do not read symbols/data from device files
> >>>>> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
> >>>>> ---
> >>>>>  tools/perf/util/symbol.c | 3 +--
> >>>>>  1 file changed, 1 insertion(+), 2 deletions(-)
> >>>>>
> >>>>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> >>>>> index 8a747dc6cf86..b4070daeb17f 100644
> >>>>> --- a/tools/perf/util/symbol.c
> >>>>> +++ b/tools/perf/util/symbol.c
> >>>>> @@ -1432,8 +1432,7 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
> >>>>>  	 * Read the build id if possible. This is required for
> >>>>>  	 * DSO_BINARY_TYPE__BUILDID_DEBUGINFO to work
> >>>>>  	 */
> >>>>> -	if (is_regular_file(name) &&
> >>>>> -	    filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
> >>>>> +	if (filename__read_build_id(dso->long_name, build_id, BUILD_ID_SIZE) > 0)
> >>>>>  		dso__set_build_id(dso, build_id);
> >>>>
> >>>> ouch copy&paste error.. we better fix it properly,
> >>>> could you please check attached patch?
> >>>
> >>> The use of uninitialized memory is gone, but the "--symfs" option is
> >>> ignored:
> >>>
> >>>> stat("/lib/libpthread-2.20.so", 0x7ffe440c1270) = -1 ENOENT (No such file or directory)
> >>>
> >>> You'll look for the dso in the root file system not in the sysroot.
> >>
> >> that seems like unrelated to the 'use of uninitialized memory' issue, right?
> >> I can't see how the patch I sent could do that..
> > 
> > Right. That's a different issue. You can add my:
> > 
> > Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
> 
> I don't see this patch on mainline nor on linux-next, or am I missing
> something?


sorry for delay, forgot about this one...

I asked Arnaldo to take this version instead:
http://marc.info/?l=linux-kernel&m=146908497508581&w=2

thanks,
jirka