From: tom.leiming@gmail.com
To: acme@ghostprotocols.net
Cc: linux-kernel@vger.kernel.org, Ming Lei <tom.leiming@gmail.com>,
Ian Munsie <imunsie@au1.ibm.com>, Ingo Molnar <mingo@elte.hu>,
Paul Mackerras <paulus@samba.org>,
Peter Zijlstra <a.p.zijlstra@chello.nl>,
Thomas Gleixner <tglx@linutronix.de>,
Tom Zanussi <tzanussi@gmail.com>
Subject: [PATCH 1/2] perf:tools: avoid to create much more maps for kernel symbols on ARM
Date: Wed, 24 Nov 2010 19:35:02 +0800 [thread overview]
Message-ID: <1290598502-31727-1-git-send-email-tom.leiming@gmail.com> (raw)
From: Ming Lei <tom.leiming@gmail.com>
On ARM, module addresss space is ahead of kernel space,
so the module symbols are handled before kernel symbol
in dso__split_kallsyms, then cause one map is created
for each kernel symbol.
This patch fixes the issue by restoring to original kernel
map in dso__split_kallsyms() to avoid create unnecessary maps
for kernel symbols when starting to handle kenel symbol maps
but after module symbol maps are handled over.
Cc: Ian Munsie <imunsie@au1.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tom Zanussi <tzanussi@gmail.com>
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
---
tools/perf/util/symbol.c | 58 ++++++++++++++++++++++++++++++++++++++++++++++
1 files changed, 58 insertions(+), 0 deletions(-)
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 0500895..c5b4ccb 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -520,6 +520,53 @@ static int dso__load_all_kallsyms(struct dso *self, const char *filename,
return kallsyms__parse(filename, &args, map__process_kallsym_symbol);
}
+static int map_for_module(struct map *map, struct machine *self)
+{
+ char *line = NULL;
+ FILE *file;
+ const char *modules;
+ char name[PATH_MAX];
+ int ret = 0;
+
+ if (machine__is_default_guest(self))
+ modules = symbol_conf.default_guest_modules;
+ else {
+ sprintf(name, "%s/proc/modules", self->root_dir);
+ modules = name;
+ }
+
+ file = fopen(modules, "r");
+ if (file == NULL)
+ return ret;
+
+ /*strip [] of map->dso->short_name*/
+ strncpy(name, map->dso->short_name + 1,
+ strlen(map->dso->short_name) - 2);
+ name[strlen(map->dso->short_name) - 3] = '\0';
+
+ while (!feof(file)) {
+ int line_len;
+ size_t n;
+
+ line_len = getline(&line, &n, file);
+ if (line_len < 0)
+ break;
+
+ if (!line)
+ break;
+
+ line[line_len-1] = '\0'; /* \n */
+
+ if (strstr(line, name)) {
+ ret = 1;
+ break;
+ }
+ }
+
+ free(line);
+ fclose(file);
+ return ret;
+}
/*
* Split the symbols into maps, making sure there are no overlaps, i.e. the
* kernel range is broken in several maps, named [kernel].N, as we don't have
@@ -590,6 +637,16 @@ static int dso__split_kallsyms(struct dso *self, struct map *map,
char dso_name[PATH_MAX];
struct dso *dso;
+ /*
+ * If module symbol is ahead of kernel symbol like ARM,
+ * restore to original kernel map to avoid create many
+ * unnecessary maps for kernel symbols.
+ */
+ if (map_for_module(curr_map, machine)) {
+ curr_map = map;
+ goto operate_map;
+ }
+
if (self->kernel == DSO_TYPE_GUEST_KERNEL)
snprintf(dso_name, sizeof(dso_name),
"[guest.kernel].%d",
@@ -616,6 +673,7 @@ static int dso__split_kallsyms(struct dso *self, struct map *map,
++kernel_range;
}
+operate_map:
if (filter && filter(curr_map, pos)) {
discard_symbol: rb_erase(&pos->rb_node, root);
symbol__delete(pos);
--
1.7.3
next reply other threads:[~2010-11-24 11:35 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-24 11:35 tom.leiming [this message]
2010-11-24 14:45 ` [PATCH 1/2] perf:tools: avoid to create much more maps for kernel symbols on ARM Arnaldo Carvalho de Melo
2010-11-25 11:04 ` Ming Lei
2010-11-29 4:34 ` Ming Lei
2010-11-30 15:40 ` [tip:perf/urgent] perf symbols: Fix kallsyms kernel/module map splitting tip-bot for 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=1290598502-31727-1-git-send-email-tom.leiming@gmail.com \
--to=tom.leiming@gmail.com \
--cc=a.p.zijlstra@chello.nl \
--cc=acme@ghostprotocols.net \
--cc=imunsie@au1.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=paulus@samba.org \
--cc=tglx@linutronix.de \
--cc=tzanussi@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.