From: Arnaldo Carvalho de Melo <acme@infradead.org>
To: Ingo Molnar <mingo@kernel.org>
Cc: linux-kernel@vger.kernel.org,
Arnaldo Carvalho de Melo <acme@redhat.com>,
David Ahern <dsahern@gmail.com>,
Frederic Weisbecker <fweisbec@gmail.com>,
Jiri Olsa <jolsa@redhat.com>, Mike Galbraith <efault@gmx.de>,
Namhyung Kim <namhyung@gmail.com>,
Paul Mackerras <paulus@samba.org>,
Peter Zijlstra <peterz@infradead.org>,
Stephane Eranian <eranian@google.com>
Subject: [PATCH 08/74] perf evsel: Introduce event fallback method
Date: Thu, 24 Jan 2013 17:07:17 -0300 [thread overview]
Message-ID: <1359058103-31645-9-git-send-email-acme@infradead.org> (raw)
In-Reply-To: <1359058103-31645-1-git-send-email-acme@infradead.org>
From: Arnaldo Carvalho de Melo <acme@redhat.com>
The only fallback right now is for HW cpu-cycles -> SW cpu-clock, that
was done in the same way in both 'top' and 'record'.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-58l1mgibh9oa9m0pd3fasxa5@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/builtin-record.c | 23 +++--------------------
tools/perf/builtin-top.c | 20 +++-----------------
tools/perf/util/evsel.c | 28 ++++++++++++++++++++++++++++
tools/perf/util/evsel.h | 3 +++
4 files changed, 37 insertions(+), 37 deletions(-)
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index a4b9726..b8d0a39 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -224,6 +224,7 @@ static bool perf_evlist__equal(struct perf_evlist *evlist,
static int perf_record__open(struct perf_record *rec)
{
+ char msg[128];
struct perf_evsel *pos;
struct perf_evlist *evlist = rec->evlist;
struct perf_session *session = rec->session;
@@ -249,27 +250,9 @@ try_again:
goto out;
}
- /*
- * If it's cycles then fall back to hrtimer
- * based cpu-clock-tick sw counter, which
- * is always available even if no PMU support.
- *
- * PPC returns ENXIO until 2.6.37 (behavior changed
- * with commit b0a873e).
- */
- if ((err == ENOENT || err == ENXIO)
- && attr->type == PERF_TYPE_HARDWARE
- && attr->config == PERF_COUNT_HW_CPU_CYCLES) {
-
+ if (perf_evsel__fallback(pos, err, msg, sizeof(msg))) {
if (verbose)
- ui__warning("The cycles event is not supported, "
- "trying to fall back to cpu-clock-ticks\n");
- attr->type = PERF_TYPE_SOFTWARE;
- attr->config = PERF_COUNT_SW_CPU_CLOCK;
- if (pos->name) {
- free(pos->name);
- pos->name = NULL;
- }
+ ui__warning("%s\n", msg);
goto try_again;
}
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 74fca61..8d41d0b 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -892,6 +892,7 @@ static void perf_top__mmap_read(struct perf_top *top)
static void perf_top__start_counters(struct perf_top *top)
{
+ char msg[128];
struct perf_evsel *counter;
struct perf_evlist *evlist = top->evlist;
struct perf_record_opts *opts = &top->record_opts;
@@ -909,25 +910,10 @@ try_again:
ui__error_paranoid();
goto out_err;
}
- /*
- * If it's cycles then fall back to hrtimer
- * based cpu-clock-tick sw counter, which
- * is always available even if no PMU support:
- */
- if ((err == ENOENT || err == ENXIO) &&
- (attr->type == PERF_TYPE_HARDWARE) &&
- (attr->config == PERF_COUNT_HW_CPU_CYCLES)) {
+ if (perf_evsel__fallback(counter, err, msg, sizeof(msg))) {
if (verbose)
- ui__warning("Cycles event not supported,\n"
- "trying to fall back to cpu-clock-ticks\n");
-
- attr->type = PERF_TYPE_SOFTWARE;
- attr->config = PERF_COUNT_SW_CPU_CLOCK;
- if (counter->name) {
- free(counter->name);
- counter->name = NULL;
- }
+ ui__warning("%s\n", msg);
goto try_again;
}
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index ee6ee3f..0c88e5c 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -1378,3 +1378,31 @@ int perf_evsel__fprintf(struct perf_evsel *evsel,
fputc('\n', fp);
return ++printed;
}
+
+bool perf_evsel__fallback(struct perf_evsel *evsel, int err,
+ char *msg, size_t msgsize)
+{
+ if ((err == ENOENT || err == ENXIO) &&
+ evsel->attr.type == PERF_TYPE_HARDWARE &&
+ evsel->attr.config == PERF_COUNT_HW_CPU_CYCLES) {
+ /*
+ * If it's cycles then fall back to hrtimer based
+ * cpu-clock-tick sw counter, which is always available even if
+ * no PMU support.
+ *
+ * PPC returns ENXIO until 2.6.37 (behavior changed with commit
+ * b0a873e).
+ */
+ scnprintf(msg, msgsize, "%s",
+"The cycles event is not supported, trying to fall back to cpu-clock-ticks");
+
+ evsel->attr.type = PERF_TYPE_SOFTWARE;
+ evsel->attr.config = PERF_COUNT_SW_CPU_CLOCK;
+
+ free(evsel->name);
+ evsel->name = NULL;
+ return true;
+ }
+
+ return false;
+}
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 9cb8a02..26d8cfd 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -251,4 +251,7 @@ struct perf_attr_details {
int perf_evsel__fprintf(struct perf_evsel *evsel,
struct perf_attr_details *details, FILE *fp);
+
+bool perf_evsel__fallback(struct perf_evsel *evsel, int err,
+ char *msg, size_t msgsize);
#endif /* __PERF_EVSEL_H */
--
1.8.1.1.361.gec3ae6e
next prev parent reply other threads:[~2013-01-24 20:33 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <1359058103-31645-1-git-send-email-acme@infradead.org>
2013-01-24 20:07 ` [PATCH 01/74] perf hists: Exchange order of comparing items when collapsing hists Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 02/74] perf hists: Link hist entries before inserting to an output tree Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 03/74] perf diff: Use internal rb tree for compute resort Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 04/74] perf test: Add a test case for hists__{match,link} Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 05/74] perf test: Remove leftover temp file left by one of the attr tests Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 06/74] perf tests: Adjust some message log levels to help diagnosing problems in " Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 07/74] perf evsel: Do missing feature fallbacks in just one place Arnaldo Carvalho de Melo
2013-01-24 20:07 ` Arnaldo Carvalho de Melo [this message]
2013-01-24 20:07 ` [PATCH 09/74] perf evsel: Introduce perf_evsel__open_strerror method Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 10/74] perf test: Check for linking problems in the python binding Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 11/74] perf python: Fix breakage introduced by the test_attr infrastructure Arnaldo Carvalho de Melo
2013-01-29 4:06 ` Thomas Backlund
2013-01-24 20:07 ` [PATCH 12/74] perf tools: Add missing closedir in multi tracepoint processing Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 13/74] perf tools: Add support for wildcard in tracepoint system name Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 14/74] perf tests: Add event parsing test for '*:*' tracepoints Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 15/74] perf tests: Check python path on attr and binding test Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 16/74] perf header: Ensure read/write finished successfully Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 17/74] perf record: Don't pass host machine to guest synthesizer Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 18/74] perf hists: Rename hists__fprintf_nr_events to events_stats__fprintf Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 19/74] perf session: There is no need for a per session hists instance Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 20/74] perf machine: Introduce struct machines Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 21/74] perf tests: Fix PYTHONPATH for python-use test tracepoints Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 22/74] perf machine: Simplify accessing the host machine Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 23/74] perf kvm: Initialize file_name var to fix segfault Arnaldo Carvalho de Melo
2013-01-25 3:00 ` Xiao Guangrong
2013-01-24 20:07 ` [PATCH 24/74] perf tests: Add return states enum for tests Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 25/74] perf tests: Don't fail if a matching vmlinux isn't found, skip that test Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 26/74] perf tools: remove redundant checks from _sort__sym_cmp Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 27/74] perf kmem: use ARRAY_SIZE instead of reinventing it Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 28/74] perf script: " Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 29/74] uprobes: remove redundant check Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 30/74] perf ui/gtk: Factor out common browser routines Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 31/74] perf ui/gtk: Setup browser window early Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 32/74] tools lib traceevent: test correct variable after allocation Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 33/74] tools lib traceevent: Update FSF postal address to be URL's Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 34/74] tools lib traceevent: Add copyright header Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 35/74] perf tools: Fix GNU make v3.80 compatibility issue Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 36/74] perf tools: Fix possible (unlikely) buffer overflow Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 37/74] perf symbols: Include elf.h header regardless LIBELF_SUPPORT Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 38/74] perf bench: Flush stdout before starting bench suite Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 39/74] perf tools: Add anonymous huge page recognition Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 40/74] perf: Missing field in PERF_RECORD_SAMPLE documentation Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 41/74] perf probe: Allow of casting an array of char to string Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 42/74] perf sort: Move misplaced sort entry functions Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 43/74] perf sort: Get rid of unnecessary __maybe_unused Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 44/74] perf sort: Fix --sort pid output Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 45/74] perf sort: Align cpu column to right Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 46/74] perf sort: Calculate parent column width too Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 47/74] perf sort: Clean up sort__first_dimension setting Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 48/74] perf sort: Separate out branch stack specific sort keys Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 49/74] perf report: Update documentation for " Arnaldo Carvalho de Melo
2013-01-24 20:07 ` [PATCH 50/74] perf symbols: Move name malloc to when needed in dso__load Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 51/74] perf symbols: Mark vmlinux filename as allocated Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 52/74] perf tools: Move get_term_dimensions from top to util.c Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 53/74] perf annotate browser: Fix segfault when drawing out-of-bounds jumps Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 54/74] perf tools: Mark branch_info maps as referenced Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 55/74] perf tools: Remove unused 'unset' parameter from parse_events Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 56/74] tools lib traceevent: Fix warning on '>=' operator Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 57/74] perf tools: Get rid of unused include of config.mak Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 58/74] perf tools: Do not include PERF-VERSION-FILE to Makefile Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 59/74] perf tools: Fix PMU format parsing test failure Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 60/74] perf tools: Move ltrim() to util/string.c Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 61/74] perf tools: Fix usage of __ in parse_events_term struct Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 62/74] perf pmu: Fix usage of __ in struct names Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 63/74] perf ui browsers: " Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 64/74] perf tools: Fix usage of __ in event parsing " Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 65/74] perf tests: Use ARRAY_SIZE() were applicable Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 66/74] perf pmu: Privatize perf_pmu_{format,alias} structs Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 67/74] perf tools: Remove some needless die() calls from the main routine Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 68/74] perf tools: Reinstate 'signed' field flag for tracepoints Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 69/74] perf script: Don't display trace info when invoking scripts Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 70/74] perf script: hook up perf_scripting_context->pevent Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 71/74] perf script: Remove workqueue-stats script Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 72/74] perf tools: Allow passing NULL to intlist__find Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 73/74] perf tools: Allow passing a list to intlist__new Arnaldo Carvalho de Melo
2013-01-24 20:08 ` [PATCH 74/74] perf test: Allow skipping tests Arnaldo Carvalho de Melo
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=1359058103-31645-9-git-send-email-acme@infradead.org \
--to=acme@infradead.org \
--cc=acme@redhat.com \
--cc=dsahern@gmail.com \
--cc=efault@gmx.de \
--cc=eranian@google.com \
--cc=fweisbec@gmail.com \
--cc=jolsa@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=namhyung@gmail.com \
--cc=paulus@samba.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).