From: Yang Jihong <yangjihong1@huawei.com>
To: Adrian Hunter <adrian.hunter@intel.com>, <peterz@infradead.org>,
<mingo@redhat.com>, <acme@kernel.org>, <mark.rutland@arm.com>,
<alexander.shishkin@linux.intel.com>, <jolsa@kernel.org>,
<namhyung@kernel.org>, <irogers@google.com>,
<anshuman.khandual@arm.com>, <jesussanp@google.com>,
<linux-perf-users@vger.kernel.org>,
<linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2 4/4] perf tools: Add printing perf_event_attr config symbol in perf_event_attr__fprintf()
Date: Mon, 15 May 2023 16:52:04 +0800 [thread overview]
Message-ID: <d598ec43-5729-33a4-0b2a-976459803bac@huawei.com> (raw)
In-Reply-To: <aa9f1c70-eb6c-8c09-8af6-f1a0a4b60831@intel.com>
Hello,
On 2023/5/12 18:34, Adrian Hunter wrote:
> On 11/05/23 10:51, Yang Jihong wrote:
>> When printing perf_event_attr, always display perf_event_attr config and its symbol
>> to improve the readability of debugging information.
>>
>> Before:
>>
>> # perf --debug verbose=2 record -e cycles,cpu-clock,sched:sched_switch,branch-load-misses,r101,mem:0x0 -C 0 true
>> <SNIP>
>> ------------------------------------------------------------
>> perf_event_attr:
>> size 136
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 5
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 1
>> size 136
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 6
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 2
>> size 136
>> config 0x143
>> { sample_period, sample_freq } 1
>> sample_type IP|TID|TIME|CPU|PERIOD|RAW|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 7
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 3
>> size 136
>> config 0x10005
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 9
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 4
>> size 136
>> config 0x101
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 10
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 5
>> size 136
>> { sample_period, sample_freq } 1
>> sample_type IP|TID|TIME|CPU|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> sample_id_all 1
>> exclude_guest 1
>> bp_type 3
>> { bp_len, config2 } 0x4
>> ------------------------------------------------------------
>> <SNIP>
>>
>> After:
>>
>> # perf --debug verbose=2 record -e cycles,cpu-clock,sched:sched_switch,branch-load-misses,r101,mem:0x0 -C 0 true
>> <SNIP>
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 0 (PERF_TYPE_HARDWARE)
>> size 136
>> config 0 (PERF_COUNT_HW_CPU_CYCLES)
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 5
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 1 (PERF_TYPE_SOFTWARE)
>> size 136
>> config 0 (PERF_COUNT_SW_CPU_CLOCK)
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 6
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 2 (PERF_TYPE_TRACEPOINT)
>> size 136
>> config 0x143 (sched:sched_switch)
>> { sample_period, sample_freq } 1
>> sample_type IP|TID|TIME|CPU|PERIOD|RAW|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 7
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 3 (PERF_TYPE_HW_CACHE)
>> size 136
>> config 0x10005 (PERF_COUNT_HW_CACHE_RESULT_MISS | PERF_COUNT_HW_CACHE_OP_READ | PERF_COUNT_HW_CACHE_BPU)
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 9
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 4 (PERF_TYPE_RAW)
>> size 136
>> config 0x101
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> freq 1
>> sample_id_all 1
>> exclude_guest 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 10
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 5 (PERF_TYPE_BREAKPOINT)
>> size 136
>> config 0
>> { sample_period, sample_freq } 1
>> sample_type IP|TID|TIME|CPU|IDENTIFIER
>> read_format ID
>> disabled 1
>> inherit 1
>> sample_id_all 1
>> exclude_guest 1
>> bp_type 3
>> { bp_len, config2 } 0x4
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 11
>> ------------------------------------------------------------
>> perf_event_attr:
>> type 1 (PERF_TYPE_SOFTWARE)
>> size 136
>> config 9 (PERF_COUNT_SW_DUMMY)
>> { sample_period, sample_freq } 4000
>> sample_type IP|TID|TIME|CPU|PERIOD|IDENTIFIER
>> read_format ID
>> inherit 1
>> mmap 1
>> comm 1
>> freq 1
>> task 1
>> sample_id_all 1
>> mmap2 1
>> comm_exec 1
>> ksymbol 1
>> bpf_event 1
>> ------------------------------------------------------------
>> sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 12
>> <SNIP>
>>
>> Signed-off-by: Yang Jihong <yangjihong1@huawei.com>
>> ---
>> tools/perf/util/perf_event_attr_fprintf.c | 159 +++++++++++++++++++++-
>> 1 file changed, 158 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/util/perf_event_attr_fprintf.c b/tools/perf/util/perf_event_attr_fprintf.c
>> index cd0905d8cb7a..7730f0d7c28a 100644
>> --- a/tools/perf/util/perf_event_attr_fprintf.c
>> +++ b/tools/perf/util/perf_event_attr_fprintf.c
>> @@ -7,6 +7,11 @@
>> #include <linux/perf_event.h>
>> #include "util/evsel_fprintf.h"
>>
>> +#ifdef HAVE_LIBTRACEEVENT
>> +#include <stdlib.h> // for free
>
> stdlib has too much ordinary stuff for the "// for free" comment
> to be interesting, and it is very unlikely to be updated if another
> dependency on stdlib were added.
>
>> +#include "trace-event.h"
>> +#endif
>
> No where else is "#ifdef HAVE_LIBTRACEEVENT" put around
> "#include "trace-event.h"
>
OK, will remove "// for free" and "#ifdef HAVE_LIBTRACEEVENT" in v3.
>> +
>> struct bit_names {
>> int bit;
>> const char *name;
>> @@ -86,6 +91,85 @@ static const char *stringify_perf_type_id(u64 value)
>> return NULL;
>> }
>> }
>> +
>> +static const char *stringify_perf_hw_id(u64 value)
>> +{
>> + /* sync with enum perf_hw_id in perf_event.h */
>
> As with patch 3, this comment and same ones below could be left out
OK, will remove in v3.
>
>> + switch (value) {
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CPU_CYCLES)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_INSTRUCTIONS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_REFERENCES)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_MISSES)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_BRANCH_MISSES)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_BUS_CYCLES)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_STALLED_CYCLES_FRONTEND)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_STALLED_CYCLES_BACKEND)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_REF_CPU_CYCLES)
>> + default:
>> + return NULL;
>> + }
>> +}
>> +
>> +static const char *stringify_perf_hw_cache_id(u64 value)
>> +{
>> + /* sync with enum perf_hw_cache_id in perf_event.h */
>> + switch (value) {
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_L1D)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_L1I)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_LL)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_DTLB)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_ITLB)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_BPU)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_NODE)
>> + default:
>> + return NULL;
>> + }
>> +}
>> +
>> +static const char *stringify_perf_hw_cache_op_id(u64 value)
>> +{
>> + /* sync with enum perf_hw_cache_op_id in perf_event.h */
>> + switch (value) {
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_OP_READ)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_OP_WRITE)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_OP_PREFETCH)
>> + default:
>> + return NULL;
>> + }
>> +}
>> +
>> +static const char *stringify_perf_hw_cache_op_result_id(u64 value)
>> +{
>> + /* sync with enum perf_hw_cache_op_result_id in perf_event.h */
>> + switch (value) {
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_RESULT_ACCESS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_HW_CACHE_RESULT_MISS)
>> + default:
>> + return NULL;
>> + }
>> +}
>> +
>> +static const char *stringify_perf_sw_id(u64 value)
>> +{
>> + /* sync with enum perf_sw_ids in perf_event.h */
>> + switch (value) {
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_CPU_CLOCK)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_TASK_CLOCK)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_PAGE_FAULTS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_CONTEXT_SWITCHES)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_CPU_MIGRATIONS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_PAGE_FAULTS_MIN)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_PAGE_FAULTS_MAJ)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_ALIGNMENT_FAULTS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_EMULATION_FAULTS)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_DUMMY)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_BPF_OUTPUT)
>> + ENUM_ID_TO_STR_CASE(PERF_COUNT_SW_CGROUP_SWITCHES)
>> + default:
>> + return NULL;
>> + }
>> +}
>> #undef ENUM_ID_TO_STR_CASE
>>
>> static void __p_type_id(char *buf, size_t size, u64 value)
>> @@ -98,6 +182,78 @@ static void __p_type_id(char *buf, size_t size, u64 value)
>> snprintf(buf, size, "%"PRIu64" (%s)", value, str);
>> }
>>
>> +static void __p_config_hw_id(char *buf, size_t size, u64 value)
>> +{
>> + const char *str = stringify_perf_hw_id(value);
>> +
>> + if (str == NULL)
>> + snprintf(buf, size, "%"PRIx64, value);
>> + else
>> + snprintf(buf, size, "%"PRIx64" (%s)", value, str);
>
> This could use the print_id() or equivalent helper mentioned in patch 3
OK, will fix in v3.
>
>> +}
>> +
>> +static void __p_config_sw_id(char *buf, size_t size, u64 value)
>> +{
>> + const char *str = stringify_perf_sw_id(value);
>> +
>> + if (str == NULL)
>> + snprintf(buf, size, "%"PRIx64, value);
>> + else
>> + snprintf(buf, size, "%"PRIx64" (%s)", value, str);
>
> Ditto
OK, ditto
>
>> +}
>> +
>> +static void __p_config_hw_cache_id(char *buf, size_t size, u64 value)
>> +{
>> + const char *hw_cache_str = stringify_perf_hw_cache_id(value & 0xff);
>> + const char *hw_cache_op_str =
>> + stringify_perf_hw_cache_op_id((value & 0xff00) >> 8);
>> + const char *hw_cache_op_result_str =
>> + stringify_perf_hw_cache_op_result_id((value & 0xff0000) >> 16);
>> +
>> + if (hw_cache_str == NULL || hw_cache_op_str == NULL ||
>> + hw_cache_op_result_str == NULL) {
>> + snprintf(buf, size, "%#"PRIx64, value);
>> + } else {
>> + snprintf(buf, size, "%#"PRIx64" (%s | %s | %s)", value,
>> + hw_cache_op_result_str, hw_cache_op_str, hw_cache_str);
>> + }
>> +}
>> +
>> +#ifdef HAVE_LIBTRACEEVENT
>> +static void __p_config_tracepoint_id(char *buf, size_t size, u64 value)
>> +{
>> + char *str = tracepoint_id_to_name(value);
>> +
>> + if (str != NULL) {
>> + snprintf(buf, size, "%#"PRIx64" (%s)", value, str);
>> + free(str);
>> + } else {
>> + snprintf(buf, size, "%#"PRIx64, value);
>> + }
>
> Ditto
>
> Note, can just call free() on 'str' whether or not it is NULL.
>
OK, ditto
Thanks for such a detailed suggestion.
Thanks,
Yang
prev parent reply other threads:[~2023-05-15 8:52 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-11 7:51 [PATCH v2 0/4] perf tools: Add printing perf_event_attr `config` and `id` symbol in perf_event_attr__fprintf() Yang Jihong
2023-05-11 7:51 ` [PATCH v2 1/4] perf trace-event-info: Add tracepoint_id_to_name() helper Yang Jihong
2023-05-12 10:33 ` Adrian Hunter
2023-05-15 8:42 ` Yang Jihong
2023-05-11 7:51 ` [PATCH v2 2/4] perf tools: Extend PRINT_ATTRf to support printing of members with a value of 0 Yang Jihong
2023-05-11 7:51 ` [PATCH v2 3/4] perf tools: Add printing perf_event_attr type symbol in perf_event_attr__fprintf() Yang Jihong
2023-05-12 10:34 ` Adrian Hunter
2023-05-15 8:48 ` Yang Jihong
2023-05-15 8:56 ` Adrian Hunter
2023-05-15 9:23 ` Yang Jihong
2023-05-16 7:01 ` Yang Jihong
2023-05-11 7:51 ` [PATCH v2 4/4] perf tools: Add printing perf_event_attr config " Yang Jihong
2023-05-12 10:34 ` Adrian Hunter
2023-05-15 8:52 ` Yang Jihong [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=d598ec43-5729-33a4-0b2a-976459803bac@huawei.com \
--to=yangjihong1@huawei.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=anshuman.khandual@arm.com \
--cc=irogers@google.com \
--cc=jesussanp@google.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).