SIGSEGV on perf's 'Basic parsing test'

SIGSEGV on perf's 'Basic parsing test'

[Guilherme Amadio]

Dear Arnaldo, Namhyung, Ian,

I noticed that on perf-tools-next, some new tests fail on my machine
with a segmentation fault. I am on perf-tools-next at the following
commit: 85c60a01b85ee956adf8d46c253fc50f75363e8f.

After compiling perf with DEBUG=1 and running the test on gdb, this is
the stack trace that I get from the crash:

Reading symbols from ./perf...
(gdb) run
Starting program: /home/amadio/src/linux/tools/perf/perf test -F 11
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib64/libthread_db.so.1".
 11.1: Basic parsing test                                            : Ok
event syntax error: 'temp_test_hwmon_event1'
                     \___ Bad event name

Unable to find event on a PMU of 'temp_test_hwmon_event1'

Program received signal SIGSEGV, Segmentation fault.
__GI___libc_free (mem=0x1c0b4908070c0a7a) at malloc.c:3375
3375      if (chunk_is_mmapped (p))                       /* release mmapped memory. */
(gdb) bt
#0  __GI___libc_free (mem=0x1c0b4908070c0a7a) at malloc.c:3375
#1  0x000055555564fbc8 in __zfree (ptr=0x7ffff6c00031) at ../../lib/zalloc.c:13
#2  0x000055555581997c in evsel__free_stat_priv (evsel=0x5555560d8ca0) at util/stat.c:148
#3  0x0000555555819c90 in evlist__free_stats (evlist=0x5555560d7d30) at util/stat.c:215
#4  0x000055555575d0dc in evlist__delete (evlist=0x5555560d7d30) at util/evlist.c:187
#5  0x00005555557124d9 in do_test (i=0, with_pmu=false, with_alias=false) at tests/hwmon_pmu.c:198
#6  0x0000555555712587 in test__hwmon_pmu (with_pmu=false) at tests/hwmon_pmu.c:212
#7  0x000055555571263b in test__hwmon_pmu_without_pmu (test=0x55555605ae40 <suite.hwmon_pmu>, subtest=1) at tests/hwmon_pmu.c:229
#8  0x00005555556c8e9d in start_test (test=0x55555605ae40 <suite.hwmon_pmu>, i=10, subi=1, child=0x5555560d7818, width=64, pass=1) at tests/builtin-test.c:424
#9  0x00005555556c9588 in __cmd_test (suites=0x5555560d73e0, argc=1, argv=0x7fffffffe200, skiplist=0x0) at tests/builtin-test.c:571
#10 0x00005555556ca0b5 in cmd_test (argc=1, argv=0x7fffffffe200) at tests/builtin-test.c:773
#11 0x0000555555645459 in run_builtin (p=0x555556047050 <commands+624>, argc=3, argv=0x7fffffffe200) at perf.c:351
#12 0x0000555555645704 in handle_internal_command (argc=3, argv=0x7fffffffe200) at perf.c:404
#13 0x0000555555645865 in run_argv (argcp=0x7fffffffe00c, argv=0x7fffffffe000) at perf.c:448
#14 0x0000555555645bb6 in main (argc=3, argv=0x7fffffffe200) at perf.c:560
(gdb) quit


Please let me know what other information you'd like to have from my system.
The CPU is an AMD 3950X and the motherboard is an ASUS Pro WS X570-ACE if that
helps.

$ uname -a
Linux gentoo.cern.ch 6.12.0-gentoo #1 SMP PREEMPT_DYNAMIC Mon Nov 18 16:50:03 CET 2024 x86_64 AMD Ryzen 9 3950X 16-Core Processor AuthenticAMD GNU/Linux
$ dmesg | sed -ne '/DMI/{p; q}'
[    0.000000] DMI: ASUS System Product Name/Pro WS X570-ACE, BIOS 4902 08/29/2024

Best regards,
-Guilherme

Re: SIGSEGV on perf's 'Basic parsing test'

[Namhyung Kim]

On Fri, Nov 22, 2024 at 10:49:31PM +0100, Guilherme Amadio wrote:
> Dear Arnaldo, Namhyung, Ian,
> 
> I noticed that on perf-tools-next, some new tests fail on my machine
> with a segmentation fault. I am on perf-tools-next at the following
> commit: 85c60a01b85ee956adf8d46c253fc50f75363e8f.

I've added a couple of fixes for the HWMON tests.  Can you please test
it again with the latest branch?

Thanks,
Namhyung

> 
> After compiling perf with DEBUG=1 and running the test on gdb, this is
> the stack trace that I get from the crash:
> 
> Reading symbols from ./perf...
> (gdb) run
> Starting program: /home/amadio/src/linux/tools/perf/perf test -F 11
> [Thread debugging using libthread_db enabled]
> Using host libthread_db library "/usr/lib64/libthread_db.so.1".
>  11.1: Basic parsing test                                            : Ok
> event syntax error: 'temp_test_hwmon_event1'
>                      \___ Bad event name
> 
> Unable to find event on a PMU of 'temp_test_hwmon_event1'
> 
> Program received signal SIGSEGV, Segmentation fault.
> __GI___libc_free (mem=0x1c0b4908070c0a7a) at malloc.c:3375
> 3375      if (chunk_is_mmapped (p))                       /* release mmapped memory. */
> (gdb) bt
> #0  __GI___libc_free (mem=0x1c0b4908070c0a7a) at malloc.c:3375
> #1  0x000055555564fbc8 in __zfree (ptr=0x7ffff6c00031) at ../../lib/zalloc.c:13
> #2  0x000055555581997c in evsel__free_stat_priv (evsel=0x5555560d8ca0) at util/stat.c:148
> #3  0x0000555555819c90 in evlist__free_stats (evlist=0x5555560d7d30) at util/stat.c:215
> #4  0x000055555575d0dc in evlist__delete (evlist=0x5555560d7d30) at util/evlist.c:187
> #5  0x00005555557124d9 in do_test (i=0, with_pmu=false, with_alias=false) at tests/hwmon_pmu.c:198
> #6  0x0000555555712587 in test__hwmon_pmu (with_pmu=false) at tests/hwmon_pmu.c:212
> #7  0x000055555571263b in test__hwmon_pmu_without_pmu (test=0x55555605ae40 <suite.hwmon_pmu>, subtest=1) at tests/hwmon_pmu.c:229
> #8  0x00005555556c8e9d in start_test (test=0x55555605ae40 <suite.hwmon_pmu>, i=10, subi=1, child=0x5555560d7818, width=64, pass=1) at tests/builtin-test.c:424
> #9  0x00005555556c9588 in __cmd_test (suites=0x5555560d73e0, argc=1, argv=0x7fffffffe200, skiplist=0x0) at tests/builtin-test.c:571
> #10 0x00005555556ca0b5 in cmd_test (argc=1, argv=0x7fffffffe200) at tests/builtin-test.c:773
> #11 0x0000555555645459 in run_builtin (p=0x555556047050 <commands+624>, argc=3, argv=0x7fffffffe200) at perf.c:351
> #12 0x0000555555645704 in handle_internal_command (argc=3, argv=0x7fffffffe200) at perf.c:404
> #13 0x0000555555645865 in run_argv (argcp=0x7fffffffe00c, argv=0x7fffffffe000) at perf.c:448
> #14 0x0000555555645bb6 in main (argc=3, argv=0x7fffffffe200) at perf.c:560
> (gdb) quit
> 
> 
> Please let me know what other information you'd like to have from my system.
> The CPU is an AMD 3950X and the motherboard is an ASUS Pro WS X570-ACE if that
> helps.
> 
> $ uname -a
> Linux gentoo.cern.ch 6.12.0-gentoo #1 SMP PREEMPT_DYNAMIC Mon Nov 18 16:50:03 CET 2024 x86_64 AMD Ryzen 9 3950X 16-Core Processor AuthenticAMD GNU/Linux
> $ dmesg | sed -ne '/DMI/{p; q}'
> [    0.000000] DMI: ASUS System Product Name/Pro WS X570-ACE, BIOS 4902 08/29/2024
> 
> Best regards,
> -Guilherme
> 

Re: SIGSEGV on perf's 'Basic parsing test'

[Guilherme Amadio]

Dear Namhyung, Ian,

On Mon, Nov 25, 2024 at 09:46:48PM -0800, Namhyung Kim wrote:
> On Fri, Nov 22, 2024 at 10:49:31PM +0100, Guilherme Amadio wrote:
> > Dear Arnaldo, Namhyung, Ian,
> > 
> > I noticed that on perf-tools-next, some new tests fail on my machine
> > with a segmentation fault. I am on perf-tools-next at the following
> > commit: 85c60a01b85ee956adf8d46c253fc50f75363e8f.
> 
> I've added a couple of fixes for the HWMON tests.  Can you please test
> it again with the latest branch?

Thank you. The test no longer crashes, which is good. However, it still
fails as shown below:

gentoo perf $ sudo ./perf test 11
 11: Hwmon PMU                                                       :
 11.1: Basic parsing test                                            : Ok
 11.2: Parsing without PMU name                                      : FAILED!
 11.3: Parsing with PMU name                                         : FAILED!
gentoo perf $ sudo ./perf test -vF 11
--- start ---
---- end ----
 11.1: Basic parsing test                                            : Ok
--- start ---
Testing 'temp_test_hwmon_event1'
Using CPUID AuthenticAMD-23-71-0
hwmon_pmu: removing event 'pwm3' that has no input file
hwmon_pmu: removing event 'intrusion0' that has no input file
hwmon_pmu: removing event 'pwm5' that has no input file
hwmon_pmu: removing event 'pwm2' that has no input file
hwmon_pmu: removing event 'pwm7' that has no input file
hwmon_pmu: removing event 'pwm4' that has no input file
hwmon_pmu: removing event 'pwm1' that has no input file
hwmon_pmu: removing event 'intrusion1' that has no input file
hwmon_pmu: removing event 'fan5' that has no input file
FAILED tests/hwmon_pmu.c:159 failed to parse event 'temp_test_hwmon_event1', err 1
event syntax error: 'temp_test_hwmon_event1'
                     \___ Bad event name

Unable to find event on a PMU of 'temp_test_hwmon_event1'
---- end ----
 11.2: Parsing without PMU name                                      : FAILED!
--- start ---
Testing 'hwmon_a_test_hwmon_pmu/temp_test_hwmon_event1/'
FAILED tests/hwmon_pmu.c:159 failed to parse event 'hwmon_a_test_hwmon_pmu/temp_test_hwmon_event1/', err 1
event syntax error: 'hwmon_a_test_hwmon_pmu/temp_test_hwmon_event1/'
                     \___ Bad event or PMU

Unable to find PMU or event on a PMU of 'hwmon_a_test_hwmon_pmu'
---- end ----
 11.3: Parsing with PMU name                                         : FAILED!

Best regards,
-Guilherme

> Thanks,
> Namhyung
> 
> > 
> > After compiling perf with DEBUG=1 and running the test on gdb, this is
> > the stack trace that I get from the crash:
> > 
> > Reading symbols from ./perf...
> > (gdb) run
> > Starting program: /home/amadio/src/linux/tools/perf/perf test -F 11
> > [Thread debugging using libthread_db enabled]
> > Using host libthread_db library "/usr/lib64/libthread_db.so.1".
> >  11.1: Basic parsing test                                            : Ok
> > event syntax error: 'temp_test_hwmon_event1'
> >                      \___ Bad event name
> > 
> > Unable to find event on a PMU of 'temp_test_hwmon_event1'
> > 
> > Program received signal SIGSEGV, Segmentation fault.
> > __GI___libc_free (mem=0x1c0b4908070c0a7a) at malloc.c:3375
> > 3375      if (chunk_is_mmapped (p))                       /* release mmapped memory. */
> > (gdb) bt
> > #0  __GI___libc_free (mem=0x1c0b4908070c0a7a) at malloc.c:3375
> > #1  0x000055555564fbc8 in __zfree (ptr=0x7ffff6c00031) at ../../lib/zalloc.c:13
> > #2  0x000055555581997c in evsel__free_stat_priv (evsel=0x5555560d8ca0) at util/stat.c:148
> > #3  0x0000555555819c90 in evlist__free_stats (evlist=0x5555560d7d30) at util/stat.c:215
> > #4  0x000055555575d0dc in evlist__delete (evlist=0x5555560d7d30) at util/evlist.c:187
> > #5  0x00005555557124d9 in do_test (i=0, with_pmu=false, with_alias=false) at tests/hwmon_pmu.c:198
> > #6  0x0000555555712587 in test__hwmon_pmu (with_pmu=false) at tests/hwmon_pmu.c:212
> > #7  0x000055555571263b in test__hwmon_pmu_without_pmu (test=0x55555605ae40 <suite.hwmon_pmu>, subtest=1) at tests/hwmon_pmu.c:229
> > #8  0x00005555556c8e9d in start_test (test=0x55555605ae40 <suite.hwmon_pmu>, i=10, subi=1, child=0x5555560d7818, width=64, pass=1) at tests/builtin-test.c:424
> > #9  0x00005555556c9588 in __cmd_test (suites=0x5555560d73e0, argc=1, argv=0x7fffffffe200, skiplist=0x0) at tests/builtin-test.c:571
> > #10 0x00005555556ca0b5 in cmd_test (argc=1, argv=0x7fffffffe200) at tests/builtin-test.c:773
> > #11 0x0000555555645459 in run_builtin (p=0x555556047050 <commands+624>, argc=3, argv=0x7fffffffe200) at perf.c:351
> > #12 0x0000555555645704 in handle_internal_command (argc=3, argv=0x7fffffffe200) at perf.c:404
> > #13 0x0000555555645865 in run_argv (argcp=0x7fffffffe00c, argv=0x7fffffffe000) at perf.c:448
> > #14 0x0000555555645bb6 in main (argc=3, argv=0x7fffffffe200) at perf.c:560
> > (gdb) quit
> > 
> > 
> > Please let me know what other information you'd like to have from my system.
> > The CPU is an AMD 3950X and the motherboard is an ASUS Pro WS X570-ACE if that
> > helps.
> > 
> > $ uname -a
> > Linux gentoo.cern.ch 6.12.0-gentoo #1 SMP PREEMPT_DYNAMIC Mon Nov 18 16:50:03 CET 2024 x86_64 AMD Ryzen 9 3950X 16-Core Processor AuthenticAMD GNU/Linux
> > $ dmesg | sed -ne '/DMI/{p; q}'
> > [    0.000000] DMI: ASUS System Product Name/Pro WS X570-ACE, BIOS 4902 08/29/2024
> > 
> > Best regards,
> > -Guilherme
> >