From: Wang Nan <wangnan0@huawei.com>
To: <acme@kernel.org>
Cc: <linux-kernel@vger.kernel.org>, <pi3orama@163.com>,
<lizefan@huawei.com>, <netdev@vger.kernel.org>,
<davem@davemloft.net>, Wang Nan <wangnan0@huawei.com>,
He Kuang <hekuang@huawei.com>,
"Arnaldo Carvalho de Melo" <acme@redhat.com>,
Jiri Olsa <jolsa@kernel.org>,
Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>,
Namhyung Kim <namhyung@kernel.org>
Subject: [PATCH 51/53] perf record: Read from tailsize ring buffer
Date: Mon, 11 Jan 2016 13:48:42 +0000 [thread overview]
Message-ID: <1452520124-2073-52-git-send-email-wangnan0@huawei.com> (raw)
In-Reply-To: <1452520124-2073-1-git-send-email-wangnan0@huawei.com>
tailsize_rb_find_start() introduced to find the first available event
from a tailsize ring buffer through tailsize. event with '/overwrite/'
setting is able to be read. record__mmap_should_read() is changed
accordingly.
Reading a active tailsize ring buffer is unsafe. A global tailsize ring
buffer director is introduced into 'struct record'
record__mmap_should_read() returns true if tailsize_evt_stopped is true.
Following patch whould turn off events attached to tailsize ring buffer
and toggle this director.
Signed-off-by: Wang Nan <wangnan0@huawei.com>
Signed-off-by: He Kuang <hekuang@huawei.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Zefan Li <lizefan@huawei.com>
Cc: pi3orama@163.com
---
tools/perf/builtin-record.c | 69 ++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 68 insertions(+), 1 deletion(-)
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 8e56f92..6c8905b 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -56,6 +56,7 @@ struct record {
bool no_buildid_cache_set;
bool timestamp_filename;
bool switch_output;
+ bool tailsize_evt_stopped;
unsigned long long samples;
};
@@ -79,6 +80,63 @@ static int process_synthesized_event(struct perf_tool *tool,
return record__write(rec, event, event->header.size);
}
+static int
+tailsize_rb_find_start(void *buf, u64 head, int mask, u64 *p_evt_head)
+{
+ int buf_size = mask + 1;
+ u64 evt_head = head;
+ u64 *pevt_size;
+
+ pr_debug("start reading tailsize, head=%"PRId64"\n", head);
+ while (true) {
+ struct perf_event_header *pheader;
+
+ pevt_size = buf + ((evt_head - sizeof(*pevt_size)) & mask);
+ pr_debug4("read tailsize: size: %"PRId64"\n", *pevt_size);
+
+ if (*pevt_size % sizeof(u64) != 0) {
+ pr_warning("Tailsize ring buffer corrupted: unaligned\n");
+ return -1;
+ }
+
+ if (!*pevt_size) {
+ if (evt_head) {
+ pr_warning("Tailsize ring buffer corrupted: size is 0 but evt_head (0x%"PRIx64") is not 0\n",
+ (unsigned long)evt_head);
+ return -1;
+ }
+ *p_evt_head = evt_head;
+ return 0;
+ }
+
+ if (evt_head < *pevt_size) {
+ pr_warning("Tailsize ring buffer corrupted: head (%"PRId64") < size (%"PRId64")\n",
+ evt_head, *pevt_size);
+ return -1;
+ }
+
+ evt_head -= *pevt_size;
+
+ if (evt_head + buf_size < head) {
+ evt_head += *pevt_size;
+ pr_debug("Finish reading tailsize buffer, evt_head=%"PRIx64", head=%"PRIx64"\n",
+ evt_head, head);
+ *p_evt_head = evt_head;
+ return 0;
+ }
+
+ pheader = (struct perf_event_header *)(buf + (evt_head & mask));
+ if (pheader->size != *pevt_size) {
+ pr_warning("Tailsize ring buffer corrupted: found size mismatch: %d vs %"PRId64"\n",
+ pheader->size, *pevt_size);
+ return -1;
+ }
+ }
+
+ pr_warning("ERROR: shouldn't get there\n");
+ return -1;
+}
+
static int record__mmap_read(struct record *rec, int idx)
{
struct perf_mmap *md = &rec->evlist->mmap[idx];
@@ -88,10 +146,17 @@ static int record__mmap_read(struct record *rec, int idx)
unsigned long size;
void *buf;
int rc = 0;
+ int channel;
if (old == head)
return 0;
+ channel = perf_evlist__idx_channel(rec->evlist, idx);
+ if (perf_evlist__channel_check(rec->evlist, channel, TAILSIZE)) {
+ if (tailsize_rb_find_start(data, head, md->mask, &old))
+ return -1;
+ }
+
rec->samples++;
size = head - old;
@@ -462,7 +527,8 @@ static bool record__mmap_should_read(struct record *rec, int idx)
if (perf_evlist__channel_idx(rec->evlist, &channel, &idx))
return false;
if (perf_evlist__channel_check(rec->evlist, channel, RDONLY))
- return false;
+ if (perf_evlist__channel_check(rec->evlist, channel, TAILSIZE))
+ return rec->tailsize_evt_stopped;
return true;
}
@@ -1226,6 +1292,7 @@ static struct record record = {
.mmap2 = perf_event__process_mmap2,
.ordered_events = true,
},
+ .tailsize_evt_stopped = false,
};
const char record_callchain_help[] = CALLCHAIN_RECORD_HELP
--
1.8.3.4
next prev parent reply other threads:[~2016-01-11 13:48 UTC|newest]
Thread overview: 90+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-11 13:47 [PATCH 00/53] perf tools: Bugfix, BPF improvement and perf record flight record mode Wang Nan
2016-01-11 13:47 ` [PATCH 01/53] perf tools: Add -lutil in python lib list for broken python-config Wang Nan
2016-01-12 9:43 ` Jiri Olsa
2016-01-11 13:47 ` [PATCH 02/53] perf tools: Fix phony build target for build-test Wang Nan
2016-01-11 13:47 ` [PATCH 03/53] perf tools: Set parallel making options build-test Wang Nan
2016-01-11 13:47 ` [PATCH 04/53] perf tools: Pass O option to Makefile.perf in build-test Wang Nan
2016-01-11 13:47 ` [PATCH 05/53] perf tools: Test correct path of perf " Wang Nan
2016-01-11 15:24 ` Arnaldo Carvalho de Melo
2016-01-11 22:06 ` Arnaldo Carvalho de Melo
2016-01-11 22:39 ` Arnaldo Carvalho de Melo
2016-01-11 22:39 ` Arnaldo Carvalho de Melo
2016-01-12 7:16 ` Wangnan (F)
2016-01-12 14:08 ` Arnaldo Carvalho de Melo
2016-01-11 13:47 ` [PATCH 06/53] perf tools: Fix PowerPC native building Wang Nan
2016-01-11 13:47 ` [PATCH 07/53] tools: Move Makefile.arch from perf/config to tools/scripts Wang Nan
2016-01-11 13:52 ` Wangnan (F)
2016-01-11 14:10 ` Arnaldo Carvalho de Melo
2016-01-11 13:47 ` [PATCH 08/53] perf tools: Add missing sources in perf's MANIFEST Wang Nan
2016-01-11 13:48 ` [PATCH 09/53] perf: bpf: Fix build breakage due to libbpf Wang Nan
2016-01-11 13:48 ` [PATCH 10/53] tools build: Add BPF feature check to test-all Wang Nan
2016-01-11 13:48 ` [PATCH 11/53] perf test: Fix false TEST_OK result for 'perf test hist' Wang Nan
2016-01-11 14:25 ` Sergei Shtylyov
2016-01-11 14:58 ` Arnaldo Carvalho de Melo
2016-01-11 15:32 ` Arnaldo Carvalho de Melo
2016-01-11 13:48 ` [PATCH 12/53] perf test: Reset err after using it hold errcode in hist testcases Wang Nan
2016-01-11 13:48 ` [PATCH 13/53] perf tools: Prevent calling machine__delete() on non-allocated machine Wang Nan
2016-01-11 15:42 ` Arnaldo Carvalho de Melo
2016-01-12 7:03 ` Wangnan (F)
2016-01-12 14:07 ` Arnaldo Carvalho de Melo
2016-01-11 13:48 ` [PATCH 14/53] perf test: Check environment before start real BPF test Wang Nan
2016-01-11 21:55 ` Arnaldo Carvalho de Melo
2016-01-12 7:40 ` Wangnan (F)
2016-01-12 14:10 ` Arnaldo Carvalho de Melo
2016-01-11 13:48 ` [PATCH 15/53] perf tools: Fix symbols searching for offline module in buildid-cache Wang Nan
2016-01-11 13:48 ` [PATCH 16/53] perf tools: Fix mmap2 event allocation in synthesize code Wang Nan
2016-01-11 13:48 ` [PATCH 17/53] perf test: Improve bp_signal Wang Nan
2016-01-11 21:37 ` Arnaldo Carvalho de Melo
2016-01-12 4:13 ` Wangnan (F)
2016-01-12 9:21 ` Jiri Olsa
2016-01-12 14:11 ` Arnaldo Carvalho de Melo
2016-01-12 14:17 ` Will Deacon
2016-01-11 13:48 ` [PATCH 18/53] perf tools: Add API to config maps in bpf object Wang Nan
2016-01-11 13:48 ` [PATCH 19/53] perf tools: Enable BPF object configure syntax Wang Nan
2016-01-11 13:48 ` [PATCH 20/53] perf record: Apply config to BPF objects before recording Wang Nan
2016-01-11 13:48 ` [PATCH 21/53] perf tools: Enable passing event to BPF object Wang Nan
2016-01-11 13:48 ` [PATCH 22/53] perf tools: Support perf event alias name Wang Nan
2016-01-11 13:48 ` [PATCH 23/53] perf tools: Support setting different slots in a BPF map separately Wang Nan
2016-01-11 13:48 ` [PATCH 24/53] perf tools: Enable indices setting syntax for BPF maps Wang Nan
2016-01-11 13:48 ` [PATCH 25/53] perf tools: Introduce bpf-output event Wang Nan
2016-01-11 13:48 ` [PATCH 26/53] perf data: Support converting data from bpf_perf_event_output() Wang Nan
2016-01-11 13:48 ` [PATCH 27/53] perf/core: Put size of a sample at the end of it by PERF_SAMPLE_TAILSIZE Wang Nan
2016-01-11 18:09 ` Alexei Starovoitov
2016-01-12 5:33 ` Wangnan (F)
2016-01-12 6:11 ` Alexei Starovoitov
2016-01-12 12:36 ` Wangnan (F)
2016-01-12 19:56 ` Alexei Starovoitov
2016-01-13 4:34 ` Wangnan (F)
2016-01-13 5:14 ` Alexei Starovoitov
2016-01-12 14:05 ` Peter Zijlstra
2016-01-12 14:14 ` Peter Zijlstra
2016-01-11 13:48 ` [PATCH 28/53] perf tools: Move timestamp creation to util Wang Nan
2016-01-11 13:48 ` [PATCH 29/53] perf tools: Make ordered_events reusable Wang Nan
2016-01-11 21:33 ` Arnaldo Carvalho de Melo
2016-01-11 13:48 ` [PATCH 30/53] perf record: Extract synthesize code to record__synthesize() Wang Nan
2016-01-11 13:48 ` [PATCH 31/53] perf tools: Add perf_data_file__switch() helper Wang Nan
2016-01-11 13:48 ` [PATCH 32/53] perf record: Turns auxtrace_snapshot_enable into 3 states Wang Nan
2016-01-11 13:48 ` [PATCH 33/53] perf record: Introduce record__finish_output() to finish a perf.data Wang Nan
2016-01-11 13:48 ` [PATCH 34/53] perf record: Use OPT_BOOLEAN_SET for buildid cache related options Wang Nan
2016-01-11 13:48 ` [PATCH 35/53] perf record: Add '--timestamp-filename' option to append timestamp to output filename Wang Nan
2016-01-11 13:48 ` [PATCH 36/53] perf record: Split output into multiple files via '--switch-output' Wang Nan
2016-01-11 13:48 ` [PATCH 37/53] perf record: Force enable --timestamp-filename when --switch-output is provided Wang Nan
2016-01-11 13:48 ` [PATCH 38/53] perf record: Disable buildid cache options by default in switch output mode Wang Nan
2016-01-11 13:48 ` [PATCH 39/53] perf record: Re-synthesize tracking events after output switching Wang Nan
2016-01-11 13:48 ` [PATCH 40/53] perf record: Generate tracking events for process forked by perf Wang Nan
2016-01-11 13:48 ` [PATCH 41/53] perf record: Ensure return non-zero rc when mmap fail Wang Nan
2016-01-11 13:48 ` [PATCH 42/53] perf record: Prevent reading invalid data in record__mmap_read Wang Nan
2016-01-11 14:21 ` Sergei Shtylyov
2016-01-11 15:00 ` Arnaldo Carvalho de Melo
2016-01-11 15:01 ` Arnaldo Carvalho de Melo
2016-01-11 13:48 ` [PATCH 43/53] perf tools: Add evlist channel helpers Wang Nan
2016-01-11 13:48 ` [PATCH 44/53] perf tools: Automatically add new channel according to evlist Wang Nan
2016-01-11 13:48 ` [PATCH 45/53] perf tools: Operate multiple channels Wang Nan
2016-01-11 13:48 ` [PATCH 46/53] perf tools: Squash overwrite setting into channel Wang Nan
2016-01-11 13:48 ` [PATCH 47/53] perf record: Don't read from and poll overwrite channel Wang Nan
2016-01-11 13:48 ` [PATCH 48/53] perf tools: Enable overwrite settings Wang Nan
2016-01-11 13:48 ` [PATCH 49/53] perf tools: Consider TAILSIZE bit when caclulate is_pos Wang Nan
2016-01-11 13:48 ` [PATCH 50/53] perf tools: Set tailsize attribut bit for overwrite events Wang Nan
2016-01-11 13:48 ` Wang Nan [this message]
2016-01-11 13:48 ` [PATCH 52/53] perf record: Toggle tailsize ring buffer for reading Wang Nan
2016-01-11 13:48 ` [PATCH 53/53] perf record: Allow generate tracking events at the end of output Wang Nan
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=1452520124-2073-52-git-send-email-wangnan0@huawei.com \
--to=wangnan0@huawei.com \
--cc=acme@kernel.org \
--cc=acme@redhat.com \
--cc=davem@davemloft.net \
--cc=hekuang@huawei.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lizefan@huawei.com \
--cc=masami.hiramatsu.pt@hitachi.com \
--cc=namhyung@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pi3orama@163.com \
/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).