Re: [PATCH v2] perf list: Display hybrid pmu events with cpu type

From: John Garry <john.garry@huawei.com>
To: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: <jolsa@kernel.org>, <peterz@infradead.org>, <mingo@redhat.com>,
	<alexander.shishkin@linux.intel.com>,
	<Linux-kernel@vger.kernel.org>,
	<linux-perf-users@vger.kernel.org>, <ak@linux.intel.com>,
	<kan.liang@intel.com>, "irogers@google.com" <irogers@google.com>
Subject: Re: [PATCH v2] perf list: Display hybrid pmu events with cpu type
Date: Wed, 15 Dec 2021 17:34:32 +0000	[thread overview]
Message-ID: <2bf7048d-2c20-8d3f-1b37-18a4056490a3@huawei.com> (raw)
In-Reply-To: <be6e0b56-ba29-a2ee-c153-d5aec102b587@huawei.com>

On 15/12/2021 16:18, John Garry wrote:

- yao.jin@linux.intel.com, yao.jin@intel.com

Both these author addresses bounce for me :(

And it's not just my arm64 platform which is damaged, but also my x86 
broadwell machine - uncore aliasing for perf list is broken

Before snippet:
unc_cbo_cache_lookup.any_es
[Unit: uncore_cbox L3 Lookup any request that access cache and found 
line in E or S-state]
unc_cbo_cache_lookup.any_i
[Unit: uncore_cbox L3 Lookup any request that access cache and found 
line in I-state]

After snippet:
unc_cbo_cache_lookup.any_es
[Unit: uncore_cbox L3 Lookup any request that access cache and found 
line in E or S-state]
unc_cbo_cache_lookup.any_es
[Unit: uncore_cbox L3 Lookup any request that access cache and found 
line in E or S-state]
unc_cbo_cache_lookup.any_i
[Unit: uncore_cbox L3 Lookup any request that access cache and found 
line in I-state]
unc_cbo_cache_lookup.any_i
[Unit: uncore_cbox L3 Lookup any request that access cache and found 
line in I-state]

Notice how the events are repeated (twice, for each cbox PMU) after, 
when they should not be.

This seems to be the broken code added in print_pmu_events():

 > qsort(aliases, len, sizeof(struct sevent), cmp_sevent);
 > for (j = 0; j < len; j++) {
 > /* Skip duplicates */
 > - if (j > 0 && !strcmp(aliases[j].name, aliases[j - 1].name))
 > - continue;
 > + if (j > 0 && !strcmp(aliases[j].name, aliases[j - 1].name)) {
 > + if (!aliases[j].pmu || !aliases[j - 1].pmu ||
 > + !strcmp(aliases[j].pmu, aliases[j - 1].pmu)) {
 > + continue;
 > + }
 > + }

Anyone an idea on the !strcmp(aliases[j].pmu, aliases[j - 1].pmu) check 
or how to fix it?

Thanks,
John