From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Ingo Molnar <mingo@kernel.org>, Thomas Gleixner <tglx@linutronix.de>
Cc: Jiri Olsa <jolsa@kernel.org>, Namhyung Kim <namhyung@kernel.org>,
Clark Williams <williams@redhat.com>,
linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
Masami Hiramatsu <mhiramat@kernel.org>,
Arnaldo Carvalho de Melo <acme@redhat.com>,
Ravi Bangoria <ravi.bangoria@linux.ibm.com>,
Steven Rostedt <rostedt@goodmis.org>,
Tom Zanussi <tom.zanussi@linux.intel.com>
Subject: [PATCH 18/25] perf probe: Show correct statement line number by perf probe -l
Date: Tue, 19 Nov 2019 08:32:38 -0300 [thread overview]
Message-ID: <20191119113245.19593-19-acme@kernel.org> (raw)
In-Reply-To: <20191119113245.19593-1-acme@kernel.org>
From: Masami Hiramatsu <mhiramat@kernel.org>
The dwarf_getsrc_die() can return the line which is not a statement nor
the least line number among the lines which shares same address.
This can lead perf probe --list shows incorrect line number for probed
address.
To fix this, this introduces cu_getsrc_die() which returns only a
statement line and which is the least line number (we call it the
representive line for an address), and use it in cu_find_lineinfo().
Also, if the given address is the entry address of a real function,
cu_find_lineinfo() returns the function declared line number instead of
the start line number of the function body.
For example, without this change perf probe -l shows incorrect line as
below.
# perf probe -a kernel_read:2
Added new event:
probe:kernel_read (on kernel_read:2)
You can now use it in all perf tools, such as:
perf record -e probe:kernel_read -aR sleep 1
# perf probe -l
probe:kernel_read (on kernel_read:1@linux-5.0.0/fs/read_write.c)
With this fix, it shows correct line number as below;
# perf probe -l
probe:kernel_read (on kernel_read:2@linux-5.0.0/fs/read_write.c)
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Ravi Bangoria <ravi.bangoria@linux.ibm.com>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Tom Zanussi <tom.zanussi@linux.intel.com>
Link: http://lore.kernel.org/lkml/157406471067.24476.17463149618465494448.stgit@devnote2
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/util/dwarf-aux.c | 62 ++++++++++++++++++++++++++++++++++---
1 file changed, 58 insertions(+), 4 deletions(-)
diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
index 5544bfbd0f6c..aa898014ad12 100644
--- a/tools/perf/util/dwarf-aux.c
+++ b/tools/perf/util/dwarf-aux.c
@@ -59,6 +59,51 @@ const char *cu_get_comp_dir(Dwarf_Die *cu_die)
return dwarf_formstring(&attr);
}
+/* Unlike dwarf_getsrc_die(), cu_getsrc_die() only returns statement line */
+static Dwarf_Line *cu_getsrc_die(Dwarf_Die *cu_die, Dwarf_Addr addr)
+{
+ Dwarf_Addr laddr;
+ Dwarf_Lines *lines;
+ Dwarf_Line *line;
+ size_t nlines, l, u, n;
+ bool flag;
+
+ if (dwarf_getsrclines(cu_die, &lines, &nlines) != 0 ||
+ nlines == 0)
+ return NULL;
+
+ /* Lines are sorted by address, use binary search */
+ l = 0; u = nlines - 1;
+ while (l < u) {
+ n = u - (u - l) / 2;
+ line = dwarf_onesrcline(lines, n);
+ if (!line || dwarf_lineaddr(line, &laddr) != 0)
+ return NULL;
+ if (addr < laddr)
+ u = n - 1;
+ else
+ l = n;
+ }
+ /* Going backward to find the lowest line */
+ do {
+ line = dwarf_onesrcline(lines, --l);
+ if (!line || dwarf_lineaddr(line, &laddr) != 0)
+ return NULL;
+ } while (laddr == addr);
+ l++;
+ /* Going foward to find the statement line */
+ do {
+ line = dwarf_onesrcline(lines, l++);
+ if (!line || dwarf_lineaddr(line, &laddr) != 0 ||
+ dwarf_linebeginstatement(line, &flag) != 0)
+ return NULL;
+ if (laddr > addr)
+ return NULL;
+ } while (!flag);
+
+ return line;
+}
+
/**
* cu_find_lineinfo - Get a line number and file name for given address
* @cu_die: a CU DIE
@@ -72,17 +117,26 @@ int cu_find_lineinfo(Dwarf_Die *cu_die, unsigned long addr,
const char **fname, int *lineno)
{
Dwarf_Line *line;
- Dwarf_Addr laddr;
+ Dwarf_Die die_mem;
+ Dwarf_Addr faddr;
- line = dwarf_getsrc_die(cu_die, (Dwarf_Addr)addr);
- if (line && dwarf_lineaddr(line, &laddr) == 0 &&
- addr == (unsigned long)laddr && dwarf_lineno(line, lineno) == 0) {
+ if (die_find_realfunc(cu_die, (Dwarf_Addr)addr, &die_mem)
+ && die_entrypc(&die_mem, &faddr) == 0 &&
+ faddr == addr) {
+ *fname = dwarf_decl_file(&die_mem);
+ dwarf_decl_line(&die_mem, lineno);
+ goto out;
+ }
+
+ line = cu_getsrc_die(cu_die, (Dwarf_Addr)addr);
+ if (line && dwarf_lineno(line, lineno) == 0) {
*fname = dwarf_linesrc(line, NULL, NULL);
if (!*fname)
/* line number is useless without filename */
*lineno = 0;
}
+out:
return *lineno ?: -ENOENT;
}
--
2.21.0
next prev parent reply other threads:[~2019-11-19 11:34 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-19 11:32 [GIT PULL] perf/core improvements and fixes Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 01/25] perf vendor events arm64: Fix commas so PMU event files are valid JSON Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 02/25] perf vendor events power8: " Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 03/25] perf vendor events power9: " Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 04/25] perf scripts python: exported-sql-viewer.py: Fix use of TRUE with SQLite Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 05/25] perf maps: Purge the entries from maps->names in __maps__purge() Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 06/25] perf maps: Do not use an rbtree to sort by map name Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 07/25] perf map_groups: Add a front end cache for map lookups by name Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 08/25] perf map: No need to adjust the long name of modules Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 09/25] perf record: No need to process the synthesized MMAP events twice Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 10/25] perf machine: No need to check if kernel module maps pre-exist Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 11/25] perf map_groups: Auto sort maps by name, if needed Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 12/25] perf callchain: Fix segfault in thread__resolve_callchain_sample() Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 13/25] libtraceevent: Fix parsing of event %o and %X argument types Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 14/25] perf map: Use bitmap for booleans Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 15/25] perf map: Move seldom used ->flags field to second cacheline Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 16/25] x86/insn: perf tools: Add some instructions to the new instructions test Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 17/25] x86/insn: Add some Intel instructions to the opcode map Arnaldo Carvalho de Melo
2019-11-19 11:32 ` Arnaldo Carvalho de Melo [this message]
2019-11-19 11:32 ` [PATCH 19/25] perf probe: Verify given line is a representive line Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 20/25] perf probe: Do not show non representive lines by perf-probe -L Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 21/25] perf probe: Generate event name with line number Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 22/25] perf probe: Support multiprobe event Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 23/25] perf probe: Support DW_AT_const_value constant value Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 24/25] perf probe: Trace a magic number if variable is not found Arnaldo Carvalho de Melo
2019-11-19 11:32 ` [PATCH 25/25] perf parse: Report initial event parsing error Arnaldo Carvalho de Melo
2019-11-19 12:00 ` [GIT PULL] perf/core improvements and fixes Ingo Molnar
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=20191119113245.19593-19-acme@kernel.org \
--to=acme@kernel.org \
--cc=acme@redhat.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mhiramat@kernel.org \
--cc=mingo@kernel.org \
--cc=namhyung@kernel.org \
--cc=ravi.bangoria@linux.ibm.com \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=tom.zanussi@linux.intel.com \
--cc=williams@redhat.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).