* [PATCH] perf report: Skip symbols for idle symbols
@ 2022-06-08 5:51 Rafał Miłecki
2022-10-18 12:54 ` Rafał Miłecki
0 siblings, 1 reply; 3+ messages in thread
From: Rafał Miłecki @ 2022-06-08 5:51 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo
Cc: Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
linux-perf-users, linux-kernel, Rafał Miłecki
From: Rafał Miłecki <rafal@milecki.pl>
Displaying idle symbols is rather useless and makes reviewing "real"
symbols harder. perf top already skips them so perf report should
probably do the same.
BEFORE
# Overhead Command Shared Object Symbol
# ........ ............... ..................... ...................................
#
75.04% swapper [kernel.kallsyms] [k] arch_cpu_idle
1.05% ksoftirqd/0 [kernel.kallsyms] [k] __pi___inval_dcache_area
0.78% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb_core
0.75% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup
0.58% ksoftirqd/0 [kernel.kallsyms] [k] __dev_queue_xmit
0.56% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_start_xmit
0.55% ksoftirqd/0 [kernel.kallsyms] [k] __skb_flow_dissect
0.53% ksoftirqd/0 [kernel.kallsyms] [k] __pi___clean_dcache_area_poc
AFTER
# Overhead Command Shared Object Symbol
# ........ ............... ..................... ...................................
#
4.21% ksoftirqd/0 [kernel.kallsyms] [k] __pi___inval_dcache_area
3.13% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb_core
3.01% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup
2.32% ksoftirqd/0 [kernel.kallsyms] [k] __dev_queue_xmit
2.23% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_start_xmit
2.21% ksoftirqd/0 [kernel.kallsyms] [k] __skb_flow_dissect
2.13% ksoftirqd/0 [kernel.kallsyms] [k] __pi___clean_dcache_area_poc
2.00% ksoftirqd/0 [kernel.kallsyms] [k] eth_type_trans
1.86% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_poll_rx
1.73% ksoftirqd/0 [kernel.kallsyms] [k] __local_bh_enable_ip
1.63% ksoftirqd/0 [kernel.kallsyms] [k] fib_rules_lookup
1.61% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_poll_tx
(...)
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
---
tools/perf/builtin-report.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index afe4a5539ecc..7e31af009d44 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -316,9 +316,12 @@ static int process_sample_event(struct perf_tool *tool,
&rep->total_cycles);
}
- ret = hist_entry_iter__add(&iter, &al, rep->max_stack, rep);
- if (ret < 0)
- pr_debug("problem adding hist entry, skipping event\n");
+ if (al.sym == NULL || !al.sym->idle) {
+ ret = hist_entry_iter__add(&iter, &al, rep->max_stack, rep);
+ if (ret < 0)
+ pr_debug("problem adding hist entry, skipping event\n");
+ }
+
out_put:
addr_location__put(&al);
return ret;
--
2.34.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] perf report: Skip symbols for idle symbols
2022-06-08 5:51 [PATCH] perf report: Skip symbols for idle symbols Rafał Miłecki
@ 2022-10-18 12:54 ` Rafał Miłecki
2022-10-18 19:18 ` Arnaldo Carvalho de Melo
0 siblings, 1 reply; 3+ messages in thread
From: Rafał Miłecki @ 2022-10-18 12:54 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo
Cc: Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
linux-perf-users, linux-kernel, Rafał Miłecki
On 8.06.2022 07:51, Rafał Miłecki wrote:
> From: Rafał Miłecki <rafal@milecki.pl>
>
> Displaying idle symbols is rather useless and makes reviewing "real"
> symbols harder. perf top already skips them so perf report should
> probably do the same.
>
> BEFORE
>
> # Overhead Command Shared Object Symbol
> # ........ ............... ..................... ...................................
> #
> 75.04% swapper [kernel.kallsyms] [k] arch_cpu_idle
> 1.05% ksoftirqd/0 [kernel.kallsyms] [k] __pi___inval_dcache_area
> 0.78% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb_core
> 0.75% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup
> 0.58% ksoftirqd/0 [kernel.kallsyms] [k] __dev_queue_xmit
> 0.56% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_start_xmit
> 0.55% ksoftirqd/0 [kernel.kallsyms] [k] __skb_flow_dissect
> 0.53% ksoftirqd/0 [kernel.kallsyms] [k] __pi___clean_dcache_area_poc
>
> AFTER
>
> # Overhead Command Shared Object Symbol
> # ........ ............... ..................... ...................................
> #
> 4.21% ksoftirqd/0 [kernel.kallsyms] [k] __pi___inval_dcache_area
> 3.13% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb_core
> 3.01% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup
> 2.32% ksoftirqd/0 [kernel.kallsyms] [k] __dev_queue_xmit
> 2.23% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_start_xmit
> 2.21% ksoftirqd/0 [kernel.kallsyms] [k] __skb_flow_dissect
> 2.13% ksoftirqd/0 [kernel.kallsyms] [k] __pi___clean_dcache_area_poc
> 2.00% ksoftirqd/0 [kernel.kallsyms] [k] eth_type_trans
> 1.86% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_poll_rx
> 1.73% ksoftirqd/0 [kernel.kallsyms] [k] __local_bh_enable_ip
> 1.63% ksoftirqd/0 [kernel.kallsyms] [k] fib_rules_lookup
> 1.61% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_poll_tx
> (...)
>
> Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Ping? Does it look OK to you?
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] perf report: Skip symbols for idle symbols
2022-10-18 12:54 ` Rafał Miłecki
@ 2022-10-18 19:18 ` Arnaldo Carvalho de Melo
0 siblings, 0 replies; 3+ messages in thread
From: Arnaldo Carvalho de Melo @ 2022-10-18 19:18 UTC (permalink / raw)
To: Rafał Miłecki
Cc: Peter Zijlstra, Ingo Molnar, Mark Rutland, Alexander Shishkin,
Jiri Olsa, Namhyung Kim, linux-perf-users, linux-kernel,
Rafał Miłecki
Em Tue, Oct 18, 2022 at 02:54:07PM +0200, Rafał Miłecki escreveu:
> On 8.06.2022 07:51, Rafał Miłecki wrote:
> > From: Rafał Miłecki <rafal@milecki.pl>
> >
> > Displaying idle symbols is rather useless and makes reviewing "real"
> > symbols harder. perf top already skips them so perf report should
> > probably do the same.
> >
> > BEFORE
> >
> > # Overhead Command Shared Object Symbol
> > # ........ ............... ..................... ...................................
> > #
> > 75.04% swapper [kernel.kallsyms] [k] arch_cpu_idle
> > 1.05% ksoftirqd/0 [kernel.kallsyms] [k] __pi___inval_dcache_area
> > 0.78% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb_core
> > 0.75% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup
> > 0.58% ksoftirqd/0 [kernel.kallsyms] [k] __dev_queue_xmit
> > 0.56% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_start_xmit
> > 0.55% ksoftirqd/0 [kernel.kallsyms] [k] __skb_flow_dissect
> > 0.53% ksoftirqd/0 [kernel.kallsyms] [k] __pi___clean_dcache_area_poc
> >
> > AFTER
> >
> > # Overhead Command Shared Object Symbol
> > # ........ ............... ..................... ...................................
> > #
> > 4.21% ksoftirqd/0 [kernel.kallsyms] [k] __pi___inval_dcache_area
> > 3.13% ksoftirqd/0 [kernel.kallsyms] [k] __netif_receive_skb_core
> > 3.01% ksoftirqd/0 [kernel.kallsyms] [k] fib_table_lookup
> > 2.32% ksoftirqd/0 [kernel.kallsyms] [k] __dev_queue_xmit
> > 2.23% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_start_xmit
> > 2.21% ksoftirqd/0 [kernel.kallsyms] [k] __skb_flow_dissect
> > 2.13% ksoftirqd/0 [kernel.kallsyms] [k] __pi___clean_dcache_area_poc
> > 2.00% ksoftirqd/0 [kernel.kallsyms] [k] eth_type_trans
> > 1.86% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_poll_rx
> > 1.73% ksoftirqd/0 [kernel.kallsyms] [k] __local_bh_enable_ip
> > 1.63% ksoftirqd/0 [kernel.kallsyms] [k] fib_rules_lookup
> > 1.61% ksoftirqd/0 [kernel.kallsyms] [k] bcm4908_enet_poll_tx
> > (...)
> >
> > Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
>
> Ping? Does it look OK to you?
Its ok, but I wonder if after all those years changing this default
won't confuse users, so perhaps adding a 'I' hotkey to toggle showing
idle symbols?
Additionally having a 'perf config report.hide_idle=true' (for 'perf
top' as well), so that users can just switch to their liking?
Also add an entry to tools/perf/Documentation/tips.txt so that it
appears from time to time on the last line on the screen?
- Arnaldo
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-10-19 1:14 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-06-08 5:51 [PATCH] perf report: Skip symbols for idle symbols Rafał Miłecki
2022-10-18 12:54 ` Rafał Miłecki
2022-10-18 19:18 ` Arnaldo Carvalho de Melo
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).