Messages in this thread | | | Subject | Re: [PATCH v6] perf stat: Fix wrong skipping for per-die aggregation | From | "Jin, Yao" <> | Date | Mon, 18 Jan 2021 12:16:34 +0800 |
| |
Hi Arnaldo, Jiri,
On 1/16/2021 4:31 AM, Arnaldo Carvalho de Melo wrote: > Em Fri, Jan 15, 2021 at 05:28:14PM -0300, Arnaldo Carvalho de Melo escreveu: >> Em Thu, Jan 14, 2021 at 08:00:32PM +0100, Jiri Olsa escreveu: >>> On Thu, Jan 14, 2021 at 09:27:55AM +0800, Jin Yao wrote: >>> >>> SNIP >>> >>>> 2.003776312 S1-D0 1 855616 Bytes llc_misses.mem_read >>>> 2.003776312 S1-D1 1 949376 Bytes llc_misses.mem_read >>>> 3.006512788 S0-D0 1 1338880 Bytes llc_misses.mem_read >>>> 3.006512788 S0-D1 1 920064 Bytes llc_misses.mem_read >>>> 3.006512788 S1-D0 1 877184 Bytes llc_misses.mem_read >>>> 3.006512788 S1-D1 1 1020736 Bytes llc_misses.mem_read >>>> 4.008895291 S0-D0 1 926592 Bytes llc_misses.mem_read >>>> 4.008895291 S0-D1 1 906368 Bytes llc_misses.mem_read >>>> 4.008895291 S1-D0 1 892224 Bytes llc_misses.mem_read >>>> 4.008895291 S1-D1 1 987712 Bytes llc_misses.mem_read >>>> 5.001590993 S0-D0 1 962624 Bytes llc_misses.mem_read >>>> 5.001590993 S0-D1 1 912512 Bytes llc_misses.mem_read >>>> 5.001590993 S1-D0 1 891200 Bytes llc_misses.mem_read >>>> 5.001590993 S1-D1 1 978432 Bytes llc_misses.mem_read >>>> >>>> On no-die system, die_id is 0, actually it's hashmap(socket,0), original behavior >>>> is not changed. >>>> >>>> Reported-by: Huang Ying <ying.huang@intel.com> >>>> Signed-off-by: Jin Yao <yao.jin@linux.intel.com> >>>> --- >>>> v6: >>>> Fix the perf test python failure by adding hashmap.c to python-ext-sources. >>>> >>>> root@kbl-ppc:~# ./perf test python >>>> 19: 'import perf' in python : Ok >>> >>> Acked-by: Jiri Olsa <jolsa@redhat.com> >> >> Jin, this is breaking the build in some 32-bit system, can you please >> take a look to validate these warnings? > > One such system: > > 28 13.75 debian:experimental-x-mipsel : FAIL mipsel-linux-gnu-gcc (Debian 10.2.1-3) 10.2.1 20201224 > > >> CC /tmp/build/perf/util/srcline.o >> util/stat.c: In function 'pkg_id_hash': >> util/stat.c:285:9: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] >> return (int64_t)key & 0xffffffff; >> ^ >> util/stat.c: In function 'pkg_id_equal': >> util/stat.c:291:9: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] >> return (int64_t)key1 == (int64_t)key2; >> ^ >> util/stat.c:291:26: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] >> return (int64_t)key1 == (int64_t)key2; >> ^ >> util/stat.c: In function 'check_per_pkg': >> util/stat.c:342:26: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] >> if (hashmap__find(mask, (void *)key, NULL)) >> ^ >> util/stat.c:345:28: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] >> ret = hashmap__add(mask, (void *)key, (void *)1); >> ^ >> CC /tmp/build/perf/tests/expand-cgroup.o >
Thanks for reporting this build issue on 32 bit system.
In v7, I use size_t to replace uint64_t and change the hash key to 'die_id << 16 | socket_id'. I assume that 16 bits is enough for socket id, is it true? But if it's not true, we have to use a more complicated way such as allocate a structure which has die_id and socket_id and add it to hashmap.
But I'm not sure if that's necessary because I can't imagine a system which has socket id > 65535.
Thanks Jin Yao
| |