From: Steven Rostedt <rostedt@goodmis.org>
To: linux-kernel@vger.kernel.org
Cc: Ingo Molnar <mingo@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Frederic Weisbecker <fweisbec@gmail.com>
Subject: [for-next][PATCH 04/20] ftrace: Fix function probe to only enable needed functions
Date: Fri, 15 Mar 2013 15:39:39 -0400 [thread overview]
Message-ID: <20130315195122.681528384@goodmis.org> (raw)
In-Reply-To: 20130315193935.359219613@goodmis.org
[-- Attachment #1: Type: text/plain, Size: 4644 bytes --]
From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
Currently the function probe enables all functions and runs a "hash"
against every function call to see if it should call a probe. This
is extremely wasteful.
Note, a probe is something like:
echo schedule:traceoff > /debug/tracing/set_ftrace_filter
When schedule is called, the probe will disable tracing. But currently,
it has a call back for *all* functions, and checks to see if the
called function is the probe that is needed.
The probe function has been created before ftrace was rewritten to
allow for more than one "op" to be registered by the function tracer.
When probes were created, it couldn't limit the functions without also
limiting normal function calls. But now we can, it's about time
to update the probe code.
Todo, have separate ops for different entries. That is, assign
a ftrace_ops per probe, instead of one op for all probes. But
as there's not many probes assigned, this may not be that urgent.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
kernel/trace/ftrace.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 46 insertions(+), 2 deletions(-)
diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index e6effd0..dab031f 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -2988,18 +2988,20 @@ static void ftrace_free_entry_rcu(struct rcu_head *rhp)
kfree(entry);
}
-
int
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
void *data)
{
struct ftrace_func_probe *entry;
+ struct ftrace_hash **orig_hash = &trace_probe_ops.filter_hash;
+ struct ftrace_hash *hash;
struct ftrace_page *pg;
struct dyn_ftrace *rec;
int type, len, not;
unsigned long key;
int count = 0;
char *search;
+ int ret;
type = filter_parse_regex(glob, strlen(glob), &search, ¬);
len = strlen(search);
@@ -3010,8 +3012,16 @@ register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
mutex_lock(&ftrace_lock);
- if (unlikely(ftrace_disabled))
+ hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);
+ if (!hash) {
+ count = -ENOMEM;
+ goto out_unlock;
+ }
+
+ if (unlikely(ftrace_disabled)) {
+ count = -ENODEV;
goto out_unlock;
+ }
do_for_each_ftrace_rec(pg, rec) {
@@ -3043,6 +3053,13 @@ register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
}
}
+ ret = enter_record(hash, rec, 0);
+ if (ret < 0) {
+ kfree(entry);
+ count = ret;
+ goto out_unlock;
+ }
+
entry->ops = ops;
entry->ip = rec->ip;
@@ -3050,10 +3067,16 @@ register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
hlist_add_head_rcu(&entry->node, &ftrace_func_hash[key]);
} while_for_each_ftrace_rec();
+
+ ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
+ if (ret < 0)
+ count = ret;
+
__enable_ftrace_function_probe();
out_unlock:
mutex_unlock(&ftrace_lock);
+ free_ftrace_hash(hash);
return count;
}
@@ -3067,7 +3090,10 @@ static void
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
void *data, int flags)
{
+ struct ftrace_func_entry *rec_entry;
struct ftrace_func_probe *entry;
+ struct ftrace_hash **orig_hash = &trace_probe_ops.filter_hash;
+ struct ftrace_hash *hash;
struct hlist_node *n, *tmp;
char str[KSYM_SYMBOL_LEN];
int type = MATCH_FULL;
@@ -3088,6 +3114,12 @@ __unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
}
mutex_lock(&ftrace_lock);
+
+ hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);
+ if (!hash)
+ /* Hmm, should report this somehow */
+ goto out_unlock;
+
for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
struct hlist_head *hhd = &ftrace_func_hash[i];
@@ -3108,12 +3140,24 @@ __unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
continue;
}
+ rec_entry = ftrace_lookup_ip(hash, entry->ip);
+ /* It is possible more than one entry had this ip */
+ if (rec_entry)
+ free_hash_entry(hash, rec_entry);
+
hlist_del_rcu(&entry->node);
call_rcu_sched(&entry->rcu, ftrace_free_entry_rcu);
}
}
__disable_ftrace_function_probe();
+ /*
+ * Remove after the disable is called. Otherwise, if the last
+ * probe is removed, a null hash means *all enabled*.
+ */
+ ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
+ out_unlock:
mutex_unlock(&ftrace_lock);
+ free_ftrace_hash(hash);
}
void
--
1.7.10.4
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 490 bytes --]
next prev parent reply other threads:[~2013-03-15 19:55 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-15 19:39 [for-next][PATCH 00/20] tracing: function triggers, stack tracer fixes, clocks and documenation Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 01/20] tracing: Consolidate updating of count for traceon/off Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 02/20] tracing: Consolidate ftrace_trace_onoff_unreg() into callback Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 03/20] ftrace: Separate unlimited probes from count limited probes Steven Rostedt
2013-03-15 19:39 ` Steven Rostedt [this message]
2013-03-15 19:39 ` [for-next][PATCH 05/20] tracing: Add alloc/free_snapshot() to replace duplicate code Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 06/20] tracing: Add snapshot trigger to function probes Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 07/20] tracing: Fix comments for ftrace_event_file/call flags Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 08/20] ftrace: Clean up function probe methods Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 09/20] ftrace: Use manual free after synchronize_sched() not call_rcu_sched() Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 10/20] tracing: Add a way to soft disable trace events Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 11/20] tracing: Add function probe triggers to enable/disable events Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 12/20] tracing: Add skip argument to trace_dump_stack() Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 13/20] tracing: Add function probe to trigger stack traces Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 14/20] tracing: Use stack of calling function for stack tracer Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 15/20] tracing: Fix stack tracer with fentry use Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 16/20] tracing: Remove most or all of stack tracer stack size from stack_max_size Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 17/20] tracing: Add function-trace option to disable function tracing of latency tracers Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 18/20] tracing: Add "uptime" trace clock that uses jiffies Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 19/20] tracing: Add "perf" trace_clock Steven Rostedt
2013-03-15 19:39 ` [for-next][PATCH 20/20] tracing: Bring Documentation/trace/ftrace.txt up to date Steven Rostedt
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=20130315195122.681528384@goodmis.org \
--to=rostedt@goodmis.org \
--cc=akpm@linux-foundation.org \
--cc=fweisbec@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.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