Skip to contentSkip to conversation

tracing: recordmcount.pl Amend the documentation according to the implementation

PATCH Superseded Li Hong 2 participants 18 messages Oct 27, 2009 – Oct 30, 2009

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Li Hong

author
#0 lore raw

PATCH 1/9 tracing: recordmcount.pl Amend the documentation according to the implementation

In documentation, we says we will use the first function in a section as a reference. Actually, the algorithm is: choose the first global function we meet as a reference. If there is none, choose the first local one. So let the documentation consistent to the code.

Also add several clarifications.

scripts/recordmcount.pl +48 -35
index 090d300..ac908b3 100755
@@ -6,73 +6,88 @@
 #                   all the offsets to the calls to mcount.
 #
 #
-# What we want to end up with is a section in vmlinux called
-# __mcount_loc that contains a list of pointers to all the
-# call sites in the kernel that call mcount. Later on boot up, the kernel
-# will read this list, save the locations and turn them into nops.
-# When tracing or profiling is later enabled, these locations will then
-# be converted back to pointers to some function.
+# What we want to end up with this is a data area in .init.data in vmlinux 
+# that contains a list of pointers to all the call sites in the kernel that
+# call mcount. Later on boot up, the kernel will read this list, save the 
+# locations and turn them into nops. When tracing or profiling is later 
+# enabled, these locations will then be converted back to pointers to 
+# some function.
 #
 # This is no easy feat. This script is called just after the original
 # object is compiled and before it is linked.
 #
-# The references to the call sites are offsets from the section of text
-# that the call site is in. Hence, all functions in a section that
-# has a call site to mcount, will have the offset from the beginning of
-# the section and not the beginning of the function.
+# When parse this object file using 'objdump', the references to the call 
+# sites are offsets from the section that the call site is in. Hence, all 
+# functions in a section that  has a call site to mcount, will have the 
+# offset from the beginning of the section and not the beginning of the 
+# function.
 #
-# The trick is to find a way to record the beginning of the section.
-# The way we do this is to look at the first function in the section
-# which will also be the location of that section after final link.
+# But where this section will reside finally in vmlinx is undetermined at 
+# this point. So we can't use this kind of offsets to record the final
+# address of this call site.
+#
+# The trick is to change the call offset refering the start of a section to
+# refering a function symbol in this section. During the link step, 'ld' will
+# compute the final address according to the information we record.
+# 
 # e.g.
 #
 #  .section ".sched.text", "ax"
-#  .globl my_func
-#  my_func:
 #        [...]
-#        call mcount  (offset: 0x5)
+#  func1:
+#        [...]
+#        call mcount  (offset: 0x10)
+#        [...]
+#        ret
+#  .globl fun2
+#  func2:             (offset: 0x20)
+#        [...]
 #        [...]
 #        ret
-#  other_func:
+#  func3:
 #        [...]
-#        call mcount (offset: 0x1b)
+#        call mcount (offset: 0x30)
 #        [...]
 #
 # Both relocation offsets for the mcounts in the above example will be
-# offset from .sched.text. If we make another file called tmp.s with:
+# offset from .sched.text. If we choose global symbol func2 as a reference and
+# make another file called tmp.s with the new offsets:
 #
 #  .section __mcount_loc
-#  .quad  my_func + 0x5
-#  .quad  my_func + 0x1b
+#  .quad  func2 - 0x10
+#  .quad  func2 + 0x10
 #
-# We can then compile this tmp.s into tmp.o, and link it to the original
+# We can then compile this tmp.s into tmp.o, and link it back to the original
 # object.
 #
-# But this gets hard if my_func is not globl (a static function).
-# In such a case we have:
+# In our algorithm, we will choose the first global function we meet in this 
+# section as the reference. But this gets hard if there is no global functions 
+# in this section. In such a case we have to select a local one. E.g. func1:
 #
 #  .section ".sched.text", "ax"
-#  my_func:
+#  func1:
 #        [...]
-#        call mcount  (offset: 0x5)
+#        call mcount  (offset: 0x10)
 #        [...]
 #        ret
-#  other_func:
+#  func2:
 #        [...]
-#        call mcount (offset: 0x1b)
+#        call mcount (offset: 0x20)
 #        [...]
+#  .section "other.section"
 #
 # If we make the tmp.s the same as above, when we link together with
-# the original object, we will end up with two symbols for my_func:
+# the original object, we will end up with two symbols for func1:
 # one local, one global.  After final compile, we will end up with
-# an undefined reference to my_func.
+# an undefined reference to func1 or a wrong reference to another global
+# func1 in other files.
 #
 # Since local objects can reference local variables, we need to find
 # a way to make tmp.o reference the local objects of the original object
-# file after it is linked together. To do this, we convert the my_func
+# file after it is linked together. To do this, we convert func1
 # into a global symbol before linking tmp.o. Then after we link tmp.o
-# we will only have a single symbol for my_func that is global.
-# We can convert my_func back into a local symbol and we are done.
+# we will only have a single symbol for func1 that is global.
+# We can convert func1 back into a local symbol and we are done.
 #
 # Here are the steps we take:
 #
@@ -86,10 +101,8 @@
 # 6) Link together this new object with the list object.
 # 7) Convert the local functions back to local symbols and rename
 #    the result as the original object.
-#    End.
 # 8) Link the object with the list object.
 # 9) Move the result back to the original object.
-#    End.
 #
 
 use strict;
Signature · 1 line
1.6.0.4
author
#1 lore raw

PATCH 2/9 tracing: recordmcount.pl Correct the check on the number of parameters

Quoted 1 line

From 2ee0d7914a95461a19f174085a081bad345cddd5 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:27:32 +0800
Subject: [PATCH] tracing: recordmcount.pl Correct the check on the number of parameters

scripts/recordmcount.pl +1 -1
index ac908b3..0850b83 100755
@@ -112,7 +112,7 @@ $P =~ s@.*/@@g;
 
 my $V = '0.1';
 
-if ($#ARGV < 7) {
+if ($#ARGV != 10) {
 	print "usage: $P arch bits objdump objcopy cc ld nm rm mv is_module inputfile\n";
 	print "version: $V\n";
 	exit(1);
Signature · 1 line
1.6.0.4
#2 lore raw

[tip:tracing/core] tracing: Correct the check for number of arguments in recordmcount.pl

Commit-ID:  e2d753fac5b3954a3b6001f98479f0435fe7c868
Gitweb:     http://git.kernel.org/tip/e2d753fac5b3954a3b6001f98479f0435fe7c868
Author:     Li Hong <[email protected]>
AuthorDate: Tue, 27 Oct 2009 14:57:33 +0800
Committer:  Steven Rostedt <[email protected]>
CommitDate: Thu, 29 Oct 2009 15:11:41 -0400

tracing: Correct the check for number of arguments in recordmcount.pl

The number of arguments passed into recordmcount.pl is 10, but the code checks if only 7 are passed in.

LKML-Reference: <20091027065733.GB22032@uhli>

---
 scripts/recordmcount.pl |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
scripts/recordmcount.pl +1 -1
index a569be7..a512af1 100755
@@ -113,7 +113,7 @@ $P =~ s@.*/@@g;
 
 my $V = '0.1';
 
-if ($#ARGV < 7) {
+if ($#ARGV != 10) {
 	print "usage: $P arch bits objdump objcopy cc ld nm rm mv is_module inputfile\n";
 	print "version: $V\n";
 	exit(1);
author
#3 lore raw

Re: [PATCH 3/9] tracing: recordmcount.pl Support absolute path check on $inputfile

In reply to Li Hong
Quoted 1 line

From fdf89709ddf4a6ff2f2d21ff8a964e949a069fe4 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:32:18 +0800
Subject: [PATCH] tracing: recordmcount.pl Support absolute path check on $inputfile

scripts/recordmcount.pl +1 -3
index 0850b83..94daf9e 100755
@@ -122,9 +122,7 @@ my ($arch, $bits, $objdump, $objcopy, $cc,
     $ld, $nm, $rm, $mv, $is_module, $inputfile) = @ARGV;
 
 # This file refers to mcount and shouldn't be ftraced, so lets' ignore it
-if ($inputfile eq "kernel/trace/ftrace.o") {
-    exit(0);
-}
+exit (0) if ($inputfile =~ "kernel/trace/ftrace.o");
 
 # Acceptable sections to record.
 my %text_sections = (
Signature · 1 line
1.6.0.4
#4 lore raw

Re: [PATCH 3/9] tracing: recordmcount.pl Support absolute path check on $inputfile

Quoted 12 lines · Li Hong
Quoted 1 line

From fdf89709ddf4a6ff2f2d21ff8a964e949a069fe4 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:32:18 +0800
Subject: [PATCH] tracing: recordmcount.pl Support absolute path check on $inputfile

scripts/recordmcount.pl +0 -0
index 0850b83..94daf9e 100755

scripts/recordmcount.pl my ($arch, $bits, $objdump, $objcopy, $cc, · line 125

Quoted 7 lines
scripts/recordmcount.pl +1 -3
@@ -122,9 +122,7 @@ my ($arch, $bits, $objdump, $objcopy, $cc,
     $ld, $nm, $rm, $mv, $is_module, $inputfile) = @ARGV;
 
 # This file refers to mcount and shouldn't be ftraced, so lets' ignore it
-if ($inputfile eq "kernel/trace/ftrace.o") {
-    exit(0);
-}
+exit (0) if ($inputfile =~ "kernel/trace/ftrace.o");

I prefer to keep the C like notation for error paths. Less to scare C programers from perl the better.

Also, if you are using a regular expression to match, might as well do it fully:

if ($inputfile =~ m,kernel/tracel/ftrace\.o$,) {
	exit(0);
}

-- Steve

author
#5 lore raw

PATCH 4/9 tracing: recordmcount.pl Objcopy check should disable local reference correctly

In reply to Li Hong
Quoted 1 line

From 580ea035d04d4ca58300423db3dd5b4c73c8d61c Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:41:22 +0800
Subject: [PATCH] tracing: recordmcount.pl Objcopy check should disable local reference correctly

Use a function to do objcopy version check. Disable the local reference if copy can't support it. Remove some unused variables.

scripts/recordmcount.pl +24 -29
index 94daf9e..970c6d6 100755
@@ -155,6 +155,28 @@ my $function_regex;	# Find the name of a function
 my $mcount_regex;	# Find the call site to mcount (return offset)
 my $alignment;		# The .align value to use for $mcount_section
 my $section_type;	# Section header plus possible alignment command
+my $can_use_local = 0; 	# If we can use local function references
+
+##
+# check_objcopy - whether objcopy supports --globalize-symbols
+#
+#  --globalize-symbols came out in 2.17, we must test the version
+#  of objcopy, and if it is less than 2.17, then we can not
+#  record local functions.
+sub check_objcopy
+{
+    open (IN, "$objcopy --version |") or die "error running $objcopy";
+    while (<IN>) {
+        if (/objcopy.*\s(\d+)\.(\d+)/) {
+            $can_use_local = 1 if ($1 > 2 || ($1 == 2 && $2 >= 17)); 
+            last;
+        }
+    }
+    close (IN);
+
+    print STDERR "WARNING: could not find objcopy version or version is less than 2.17.\n" .
+	"\tLocal function references is disabled.\n" if !$can_use_local; 
+}
 
 if ($arch eq "x86") {
     if ($bits == 64) {
@@ -289,34 +311,7 @@ if ($filename =~ m,^(.*)(\.\S),) {
 my $mcount_s = $dirname . "/.tmp_mc_" . $prefix . ".s";
 my $mcount_o = $dirname . "/.tmp_mc_" . $prefix . ".o";
 
-#
-# --globalize-symbols came out in 2.17, we must test the version
-# of objcopy, and if it is less than 2.17, then we can not
-# record local functions.
-my $use_locals = 01;
-my $local_warn_once = 0;
-my $found_version = 0;
-
-open (IN, "$objcopy --version |") || die "error running $objcopy";
-while (<IN>) {
-    if (/objcopy.*\s(\d+)\.(\d+)/) {
-	my $major = $1;
-	my $minor = $2;
-
-	$found_version = 1;
-	if ($major < 2 ||
-	    ($major == 2 && $minor < 17)) {
-	    $use_locals = 0;
-	}
-	last;
-    }
-}
-close (IN);
-
-if (!$found_version) {
-    print STDERR "WARNING: could not find objcopy version.\n" .
-	"\tDisabling local function references.\n";
-}
+check_objcopy();
 
 #
 # Step 1: find all the local (static functions) and weak symbols.
@@ -363,7 +358,7 @@ sub update_funcs
     if (defined $locals{$ref_func}) {
 
 	# only use locals if objcopy supports globalize-symbols
-	if (!$use_locals) {
+	if (!$can_use_local) {
 	    return;
 	}
 	$convert{$ref_func} = 1;
Signature · 1 line
1.6.0.4
#6 lore raw

Re: [PATCH 4/9] tracing: recordmcount.pl Objcopy check should disable local reference correctly

Quoted 29 lines · Li Hong
Quoted 1 line

From 580ea035d04d4ca58300423db3dd5b4c73c8d61c Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:41:22 +0800
Subject: [PATCH] tracing: recordmcount.pl Objcopy check should disable local reference correctly

Use a function to do objcopy version check. Disable the local reference if copy can't support it. Remove some unused variables.

scripts/recordmcount.pl +11 -0
index 94daf9e..970c6d6 100755
@@ -155,6 +155,28 @@ my $function_regex;	# Find the name of a function
 my $mcount_regex;	# Find the call site to mcount (return offset)
 my $alignment;		# The .align value to use for $mcount_section
 my $section_type;	# Section header plus possible alignment command
+my $can_use_local = 0; 	# If we can use local function references
+
+##
+# check_objcopy - whether objcopy supports --globalize-symbols
+#
+#  --globalize-symbols came out in 2.17, we must test the version
+#  of objcopy, and if it is less than 2.17, then we can not
+#  record local functions.
+sub check_objcopy
+{
+    open (IN, "$objcopy --version |") or die "error running $objcopy";

scripts/recordmcount.pl my $function_regex; # Find the name of a function · line 178

Quoted 10 lines
scripts/recordmcount.pl +10 -0
+    while (<IN>) {
+        if (/objcopy.*\s(\d+)\.(\d+)/) {
+            $can_use_local = 1 if ($1 > 2 || ($1 == 2 && $2 >= 17)); 
+            last;
+        }
+    }
+    close (IN);
+
+    print STDERR "WARNING: could not find objcopy version or version is less than 2.17.\n" .
+	"\tLocal function references is disabled.\n" if !$can_use_local; 

Please use C like if's. This tricked even me. I looked at it, and said to myself, "Wait, this prints an error every time... oh wait!".

There are some cases where I don't mind the perl if style, and used it for assigning defaults. But this just makes it more difficult to understand.

The rest of the patch looks good.

-- Steve

Quoted 49 lines
scripts/recordmcount.pl +3 -29
+}
 
 if ($arch eq "x86") {
     if ($bits == 64) {
@@ -289,34 +311,7 @@ if ($filename =~ m,^(.*)(\.\S),) {
 my $mcount_s = $dirname . "/.tmp_mc_" . $prefix . ".s";
 my $mcount_o = $dirname . "/.tmp_mc_" . $prefix . ".o";
 
-#
-# --globalize-symbols came out in 2.17, we must test the version
-# of objcopy, and if it is less than 2.17, then we can not
-# record local functions.
-my $use_locals = 01;
-my $local_warn_once = 0;
-my $found_version = 0;
-
-open (IN, "$objcopy --version |") || die "error running $objcopy";
-while (<IN>) {
-    if (/objcopy.*\s(\d+)\.(\d+)/) {
-	my $major = $1;
-	my $minor = $2;
-
-	$found_version = 1;
-	if ($major < 2 ||
-	    ($major == 2 && $minor < 17)) {
-	    $use_locals = 0;
-	}
-	last;
-    }
-}
-close (IN);
-
-if (!$found_version) {
-    print STDERR "WARNING: could not find objcopy version.\n" .
-	"\tDisabling local function references.\n";
-}
+check_objcopy();
 
 #
 # Step 1: find all the local (static functions) and weak symbols.
@@ -363,7 +358,7 @@ sub update_funcs
     if (defined $locals{$ref_func}) {
 
 	# only use locals if objcopy supports globalize-symbols
-	if (!$use_locals) {
+	if (!$can_use_local) {
 	    return;
 	}
 	$convert{$ref_func} = 1;
author
#7 lore raw

PATCH 5/9 tracing: recordmcount.pl Clarify the logic on mcount section check

In reply to Li Hong
Quoted 1 line

From eb8a2d53bc2484be223e4fa0df8804389e969b72 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:53:52 +0800
Subject: [PATCH] tracing: recordmcount.pl Clarify the logic on mcount section check

Move the mcount section check to the beginning of the objdump read loop. It is clearer, because mcount section check uses headers dump part of objdump to identify a mcount section, which goes before the section parts.

scripts/recordmcount.pl +19 -16
index 970c6d6..a6585b6 100755
@@ -384,9 +384,26 @@ open(IN, "$objdump -hdr $inputfile|") || die "error running $objdump";
 
 my $text;
 
-my $read_headers = 1;
 
 while (<IN>) {
+    # read headers first
+    my $read_headers = 1;
+    
+    if ($read_headers && /$mcount_section/) {
+	#
+	# Somehow the make process can execute this script on an
+	# object twice. If it does, we would duplicate the mcount
+	# section and it will cause the function tracer self test
+	# to fail. Check if the mcount section exists, and if it does,
+	# warn and exit.
+	#
+	print STDERR "ERROR: $mcount_section already in $inputfile\n" .
+	    "\tThis may be an indication that your build is corrupted.\n" .
+	    "\tDelete $inputfile and try again. If the same object file\n" .
+	    "\tstill causes an issue, then disable CONFIG_DYNAMIC_FTRACE.\n";
+	exit(-1);
+    }
+
     # is it a section?
     if (/$section_regex/) {
 	$read_headers = 0;
@@ -427,21 +444,7 @@ while (<IN>) {
 		$offset = hex $1;
 	    }
 	}
-    } elsif ($read_headers && /$mcount_section/) {
-	#
-	# Somehow the make process can execute this script on an
-	# object twice. If it does, we would duplicate the mcount
-	# section and it will cause the function tracer self test
-	# to fail. Check if the mcount section exists, and if it does,
-	# warn and exit.
-	#
-	print STDERR "ERROR: $mcount_section already in $inputfile\n" .
-	    "\tThis may be an indication that your build is corrupted.\n" .
-	    "\tDelete $inputfile and try again. If the same object file\n" .
-	    "\tstill causes an issue, then disable CONFIG_DYNAMIC_FTRACE.\n";
-	exit(-1);
-    }
-
+    } 
     # is this a call site to mcount? If so, record it to print later
     if ($text_found && /$mcount_regex/) {
 	$offsets[$#offsets + 1] = hex $1;
Signature · 1 line
1.6.0.4
#8 lore raw

Re: [PATCH 5/9] tracing: recordmcount.pl Clarify the logic on mcount section check

Quoted 16 lines · Li Hong
Quoted 1 line

From eb8a2d53bc2484be223e4fa0df8804389e969b72 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 12:53:52 +0800
Subject: [PATCH] tracing: recordmcount.pl Clarify the logic on mcount section check

Move the mcount section check to the beginning of the objdump read loop. It is clearer, because mcount section check uses headers dump part of objdump to identify a mcount section, which goes before the section parts.

scripts/recordmcount.pl +0 -0
index 970c6d6..a6585b6 100755

scripts/recordmcount.pl open(IN, "$objdump -hdr $inputfile|") || die "error running $objdump"; · line 391

Quoted 9 lines
scripts/recordmcount.pl +3 -1
@@ -384,9 +384,26 @@ open(IN, "$objdump -hdr $inputfile|") || die "error running $objdump";
 
 my $text;
 
-my $read_headers = 1;
 
 while (<IN>) {
+    # read headers first
+    my $read_headers = 1;
+    

Please keep the $read_headers initialization outside of the while loop. It makes it easier for C programers to understand. Otherwise it looks like it gets reset every time.

-- Steve

Quoted 41 lines
scripts/recordmcount.pl +16 -15
+    if ($read_headers && /$mcount_section/) {
+	#
+	# Somehow the make process can execute this script on an
+	# object twice. If it does, we would duplicate the mcount
+	# section and it will cause the function tracer self test
+	# to fail. Check if the mcount section exists, and if it does,
+	# warn and exit.
+	#
+	print STDERR "ERROR: $mcount_section already in $inputfile\n" .
+	    "\tThis may be an indication that your build is corrupted.\n" .
+	    "\tDelete $inputfile and try again. If the same object file\n" .
+	    "\tstill causes an issue, then disable CONFIG_DYNAMIC_FTRACE.\n";
+	exit(-1);
+    }
+
     # is it a section?
     if (/$section_regex/) {
 	$read_headers = 0;
@@ -427,21 +444,7 @@ while (<IN>) {
 		$offset = hex $1;
 	    }
 	}
-    } elsif ($read_headers && /$mcount_section/) {
-	#
-	# Somehow the make process can execute this script on an
-	# object twice. If it does, we would duplicate the mcount
-	# section and it will cause the function tracer self test
-	# to fail. Check if the mcount section exists, and if it does,
-	# warn and exit.
-	#
-	print STDERR "ERROR: $mcount_section already in $inputfile\n" .
-	    "\tThis may be an indication that your build is corrupted.\n" .
-	    "\tDelete $inputfile and try again. If the same object file\n" .
-	    "\tstill causes an issue, then disable CONFIG_DYNAMIC_FTRACE.\n";
-	exit(-1);
-    }
-
+    } 
     # is this a call site to mcount? If so, record it to print later
     if ($text_found && /$mcount_regex/) {
 	$offsets[$#offsets + 1] = hex $1;
author
#9 lore raw

PATCH 6/9 tracing: recordmcount.pl Exit early if no work to do

In reply to Li Hong
Quoted 1 line

From 1589563d8113db41fa77dd657459b563dcecd389 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:13:37 +0800
Subject: [PATCH] tracing: recordmcount.pl Exit early if no work to do

Also keep the global symbols and use it to check if no work to do, exit early.

scripts/recordmcount.pl +19 -8
index a6585b6..d750da8 100755
@@ -91,7 +91,7 @@
 #
 # Here are the steps we take:
 #
-# 1) Record all the local symbols by using 'nm'
+# 1) Record all the global, local and weak symbols by using 'nm'
 # 2) Use objdump to find all the call site offsets and sections for
 #    mcount.
 # 3) Compile the list into its own object.
@@ -143,12 +143,15 @@ $mv = "mv" if ((length $mv) == 0);
 #print STDERR "running: $P '$arch' '$objdump' '$objcopy' '$cc' '$ld' " .
 #    "'$nm' '$rm' '$mv' '$inputfile'\n";
 
+my %globals;		# List of global functions
 my %locals;		# List of local (static) functions
 my %weak;		# List of weak functions
 my %convert;		# List of local functions used that needs conversion
 
 my $type;
-my $nm_regex;		# Find the local functions (return function)
+my $global_regex;	# Match a global function (return function)
+my $local_regex;	# Match a local function (return function)
+my $weak_regex; 	# Match a weak function (return function)
 my $section_regex;	# Find the start of a section
 my $function_regex;	# Find the name of a function
 			#    (return offset and func name)
@@ -190,7 +193,9 @@ if ($arch eq "x86") {
 # We base the defaults off of i386, the other archs may
 # feel free to change them in the below if statements.
 #
-$nm_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\S+)";
+$global_regex = "^[0-9a-fA-F]+\\s+T\\s+(\\S+)";
+$local_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\S+)";
+$weak_regex = "^[0-9a-fA-F]+\\s+([wW])\\s+(\\S+)";
 $section_regex = "Disassembly of section\\s+(\\S+):";
 $function_regex = "^([0-9a-fA-F]+)\\s+<(.*?)>:";
 $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\smcount\$";
@@ -239,7 +244,8 @@ if ($arch eq "x86_64") {
     $cc .= " -m32";
 
 } elsif ($arch eq "powerpc") {
-    $nm_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\.?\\S+)";
+    $global_regex = "^[0-9a-fA-F]+\\s+T\\s+(\\.?\\S+)";
+    $local_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\.?\\S+)";
     $function_regex = "^([0-9a-fA-F]+)\\s+<(\\.?.*?)>:";
     $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\s\\.?_mcount\$";
 
@@ -314,19 +320,24 @@ my $mcount_o = $dirname . "/.tmp_mc_" . $prefix . ".o";
 check_objcopy();
 
 #
-# Step 1: find all the local (static functions) and weak symbols.
-#        't' is local, 'w/W' is weak (we never use a weak function)
+# Step 1: find all the global and local (static functions) and weak symbols.
+#        'T' is global 't' is local, 'w/W' is weak
 #
 open (IN, "$nm $inputfile|") || die "error running $nm";
 while (<IN>) {
-    if (/$nm_regex/) {
+    if (/$global_regex/) {
+        $globals{$1} = 1;
+    } elsif (/$local_regex/) {
 	$locals{$1} = 1;
-    } elsif (/^[0-9a-fA-F]+\s+([wW])\s+(\S+)/) {
+    } elsif (/$weak_regex/) {
 	$weak{$2} = $1;
     }
 }
 close(IN);
 
+# Exit early if no work to do 
+exit(0) unless (%globals or (%locals and $can_use_local)); 
+
 my @offsets;		# Array of offsets of mcount callers
 my $ref_func;		# reference function to use for offsets
 my $offset = 0;		# offset of ref_func to section beginning
Signature · 1 line
1.6.0.4
#10 lore raw

Re: [PATCH 6/9] tracing: recordmcount.pl Exit early if no work to do

Quoted 7 lines · Li Hong
Quoted 1 line

From 1589563d8113db41fa77dd657459b563dcecd389 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:13:37 +0800
Subject: [PATCH] tracing: recordmcount.pl Exit early if no work to do

Also keep the global symbols and use it to check if no work to do, exit early.

How many files does this fall out of? This looks like added complexity for no real gain.

-- Steve

Quoted 83 lines
scripts/recordmcount.pl +19 -8
index a6585b6..d750da8 100755
@@ -91,7 +91,7 @@
 #
 # Here are the steps we take:
 #
-# 1) Record all the local symbols by using 'nm'
+# 1) Record all the global, local and weak symbols by using 'nm'
 # 2) Use objdump to find all the call site offsets and sections for
 #    mcount.
 # 3) Compile the list into its own object.
@@ -143,12 +143,15 @@ $mv = "mv" if ((length $mv) == 0);
 #print STDERR "running: $P '$arch' '$objdump' '$objcopy' '$cc' '$ld' " .
 #    "'$nm' '$rm' '$mv' '$inputfile'\n";
 
+my %globals;		# List of global functions
 my %locals;		# List of local (static) functions
 my %weak;		# List of weak functions
 my %convert;		# List of local functions used that needs conversion
 
 my $type;
-my $nm_regex;		# Find the local functions (return function)
+my $global_regex;	# Match a global function (return function)
+my $local_regex;	# Match a local function (return function)
+my $weak_regex; 	# Match a weak function (return function)
 my $section_regex;	# Find the start of a section
 my $function_regex;	# Find the name of a function
 			#    (return offset and func name)
@@ -190,7 +193,9 @@ if ($arch eq "x86") {
 # We base the defaults off of i386, the other archs may
 # feel free to change them in the below if statements.
 #
-$nm_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\S+)";
+$global_regex = "^[0-9a-fA-F]+\\s+T\\s+(\\S+)";
+$local_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\S+)";
+$weak_regex = "^[0-9a-fA-F]+\\s+([wW])\\s+(\\S+)";
 $section_regex = "Disassembly of section\\s+(\\S+):";
 $function_regex = "^([0-9a-fA-F]+)\\s+<(.*?)>:";
 $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\smcount\$";
@@ -239,7 +244,8 @@ if ($arch eq "x86_64") {
     $cc .= " -m32";
 
 } elsif ($arch eq "powerpc") {
-    $nm_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\.?\\S+)";
+    $global_regex = "^[0-9a-fA-F]+\\s+T\\s+(\\.?\\S+)";
+    $local_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\.?\\S+)";
     $function_regex = "^([0-9a-fA-F]+)\\s+<(\\.?.*?)>:";
     $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\s\\.?_mcount\$";
 
@@ -314,19 +320,24 @@ my $mcount_o = $dirname . "/.tmp_mc_" . $prefix . ".o";
 check_objcopy();
 
 #
-# Step 1: find all the local (static functions) and weak symbols.
-#        't' is local, 'w/W' is weak (we never use a weak function)
+# Step 1: find all the global and local (static functions) and weak symbols.
+#        'T' is global 't' is local, 'w/W' is weak
 #
 open (IN, "$nm $inputfile|") || die "error running $nm";
 while (<IN>) {
-    if (/$nm_regex/) {
+    if (/$global_regex/) {
+        $globals{$1} = 1;
+    } elsif (/$local_regex/) {
 	$locals{$1} = 1;
-    } elsif (/^[0-9a-fA-F]+\s+([wW])\s+(\S+)/) {
+    } elsif (/$weak_regex/) {
 	$weak{$2} = $1;
     }
 }
 close(IN);
 
+# Exit early if no work to do 
+exit(0) unless (%globals or (%locals and $can_use_local)); 
+
 my @offsets;		# Array of offsets of mcount callers
 my $ref_func;		# reference function to use for offsets
 my $offset = 0;		# offset of ref_func to section beginning
author
#11 lore raw

PATCH 7/9 tracing: recordmcount.pl Combine the condition validation in update_funcs

In reply to Li Hong
Quoted 1 line

From d0aa9b99e4ffd13fada96bfe1651055b459e9988 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:20:48 +0800
Subject: [PATCH] tracing: recordmcount.pl Combine the condition validation in update_funcs

Move all the condition validation into function update_funcs. Also update_funcs shouldn't die if $ref_func is undefined for there may be more than one valid sections in an object file.

scripts/recordmcount.pl +3 -5
index d750da8..490b4cd 100755
@@ -350,9 +350,7 @@ my $offset = 0;		# offset of ref_func to section beginning
 #
 sub update_funcs
 {
-    return if ($#offsets < 0);
-
-    defined($ref_func) || die "No function to reference";
+    return unless ($ref_func and @offsets);
 
     # A section only had a weak function, to represent it.
     # Unfortunately, a weak function may be overwritten by another
@@ -426,7 +424,7 @@ while (<IN>) {
 	    $read_function = 0;
 	}
 	# print out any recorded offsets
-	update_funcs() if (defined($ref_func));
+	update_funcs();
 
 	# reset all markers and arrays
 	$text_found = 0;
@@ -463,7 +461,7 @@ while (<IN>) {
 }
 
 # dump out anymore offsets that may have been found
-update_funcs() if (defined($ref_func));
+update_funcs();
 
 # If we did not find any mcount callers, we are done (do nothing).
 if (!$opened) {
Signature · 1 line
1.6.0.4
#12 lore raw

Re: [PATCH 7/9] tracing: recordmcount.pl Combine the condition validation in update_funcs

Quoted 9 lines · Li Hong
Quoted 1 line

From d0aa9b99e4ffd13fada96bfe1651055b459e9988 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:20:48 +0800
Subject: [PATCH] tracing: recordmcount.pl Combine the condition validation in update_funcs

Move all the condition validation into function update_funcs. Also update_funcs shouldn't die if $ref_func is undefined for there may be more than one valid sections in an object file.

Nice clean up!

-- Steve

Quoted 35 lines
scripts/recordmcount.pl +3 -5
index d750da8..490b4cd 100755
@@ -350,9 +350,7 @@ my $offset = 0;		# offset of ref_func to section beginning
 #
 sub update_funcs
 {
-    return if ($#offsets < 0);
-
-    defined($ref_func) || die "No function to reference";
+    return unless ($ref_func and @offsets);
 
     # A section only had a weak function, to represent it.
     # Unfortunately, a weak function may be overwritten by another
@@ -426,7 +424,7 @@ while (<IN>) {
 	    $read_function = 0;
 	}
 	# print out any recorded offsets
-	update_funcs() if (defined($ref_func));
+	update_funcs();
 
 	# reset all markers and arrays
 	$text_found = 0;
@@ -463,7 +461,7 @@ while (<IN>) {
 }
 
 # dump out anymore offsets that may have been found
-update_funcs() if (defined($ref_func));
+update_funcs();
 
 # If we did not find any mcount callers, we are done (do nothing).
 if (!$opened) {
author
#13 lore raw

PATCH 8/9 tracing: recordmcount.pl We won't use weak function as reference, remove the check

In reply to Li Hong
Quoted 1 line

From 4433c78ca3c0319fd75d1fd6d64929fa101e38a4 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:25:39 +0800
Subject: [PATCH] tracing: recordmcount.pl We won't use weak function as reference, remove the check

scripts/recordmcount.pl +0 -11
index 490b4cd..867c24a 100755
@@ -352,17 +352,6 @@ sub update_funcs
 {
     return unless ($ref_func and @offsets);
 
-    # A section only had a weak function, to represent it.
-    # Unfortunately, a weak function may be overwritten by another
-    # function of the same name, making all these offsets incorrect.
-    # To be safe, we simply print a warning and bail.
-    if (defined $weak{$ref_func}) {
-	print STDERR
-	    "$inputfile: WARNING: referencing weak function" .
-	    " $ref_func for mcount\n";
-	return;
-    }
-
     # is this function static? If so, note this fact.
     if (defined $locals{$ref_func}) {
 
Signature · 1 line
1.6.0.4
#14 lore raw

Re: [PATCH 8/9] tracing: recordmcount.pl We won't use weak function as reference, remove the check

Quoted 16 lines · Li Hong
Quoted 1 line

From 4433c78ca3c0319fd75d1fd6d64929fa101e38a4 Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:25:39 +0800
Subject: [PATCH] tracing: recordmcount.pl We won't use weak function as reference, remove the check

scripts/recordmcount.pl +0 -1
index 490b4cd..867c24a 100755
@@ -352,17 +352,6 @@ sub update_funcs
 {
     return unless ($ref_func and @offsets);
 
-    # A section only had a weak function, to represent it.

scripts/recordmcount.pl sub update_funcs · line 365

Quoted 10 lines
scripts/recordmcount.pl +0 -10
-    # Unfortunately, a weak function may be overwritten by another
-    # function of the same name, making all these offsets incorrect.
-    # To be safe, we simply print a warning and bail.
-    if (defined $weak{$ref_func}) {
-	print STDERR
-	    "$inputfile: WARNING: referencing weak function" .
-	    " $ref_func for mcount\n";
-	return;
-    }
-

I consider this a sanity check. It may be triggered if someone modifies the rest of the script and somehow a weak function gets through. I'd like to keep it. It does not hurt to have it. Maybe we should make it "die" instead of just returning.

-- Steve

Quoted 3 lines
scripts/recordmcount.pl +0 -0
     # is this function static? If so, note this fact.
     if (defined $locals{$ref_func}) {
 
author
#15 lore raw

PATCH 9/9 tracing: recordmcount.pl Remove the redundant code

In reply to Li Hong
Quoted 1 line

From d0aa71d523b6315fd3ea0ee66fddb020a625856f Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:28:53 +0800
Subject: [PATCH] tracing: recordmcount.pl Remove the redundant code

If an object file has some local symbols and objcopy doesn't support local reference, we have exited before the real work starts. So remove the redundant check in update_funcs.

scripts/recordmcount.pl +1 -8
index 867c24a..ff3a9eb 100755
@@ -353,14 +353,7 @@ sub update_funcs
     return unless ($ref_func and @offsets);
 
     # is this function static? If so, note this fact.
-    if (defined $locals{$ref_func}) {
-
-	# only use locals if objcopy supports globalize-symbols
-	if (!$can_use_local) {
-	    return;
-	}
-	$convert{$ref_func} = 1;
-    }
+    $convert{$ref_func} = 1 if $locals{$ref_func}; 
 
     # Loop through all the mcount caller offsets and print a reference
     # to the caller based from the ref_func.
Signature · 1 line
1.6.0.4
#16 lore raw

Re: [PATCH 9/9] tracing: recordmcount.pl Remove the redundant code

Quoted 8 lines · Li Hong
Quoted 1 line

From d0aa71d523b6315fd3ea0ee66fddb020a625856f Mon Sep 17 00:00:00 2001

From: Li Hong <[email protected]>
Date: Tue, 27 Oct 2009 13:28:53 +0800
Subject: [PATCH] tracing: recordmcount.pl Remove the redundant code

If an object file has some local symbols and objcopy doesn't support local reference, we have exited before the real work starts. So remove the redundant check in update_funcs.

Are you sure? Where does it exit? Even if I add patch 6, you can still have a section that does not have any globals in it.

-- Steve

Quoted 23 lines
scripts/recordmcount.pl +1 -8
index 867c24a..ff3a9eb 100755
@@ -353,14 +353,7 @@ sub update_funcs
     return unless ($ref_func and @offsets);
 
     # is this function static? If so, note this fact.
-    if (defined $locals{$ref_func}) {
-
-	# only use locals if objcopy supports globalize-symbols
-	if (!$can_use_local) {
-	    return;
-	}
-	$convert{$ref_func} = 1;
-    }
+    $convert{$ref_func} = 1 if $locals{$ref_func}; 
 
     # Loop through all the mcount caller offsets and print a reference
     # to the caller based from the ref_func.
#17 lore raw
In reply to Li Hong
Quoted 19 lines · Li Hong

In documentation, we says we will use the first function in a section as a reference. Actually, the algorithm is: choose the first global function we meet as a reference. If there is none, choose the first local one. So let the documentation consistent to the code.

Also add several clarifications.

scripts/recordmcount.pl +0 -2
index 090d300..ac908b3 100755
@@ -6,73 +6,88 @@
 #                   all the offsets to the calls to mcount.
 #
 #
-# What we want to end up with is a section in vmlinux called
-# __mcount_loc that contains a list of pointers to all the

scripts/recordmcount.pl line 14

Quoted 10 lines
scripts/recordmcount.pl +6 -4
-# call sites in the kernel that call mcount. Later on boot up, the kernel
-# will read this list, save the locations and turn them into nops.
-# When tracing or profiling is later enabled, these locations will then
-# be converted back to pointers to some function.
+# What we want to end up with this is a data area in .init.data in vmlinux 
+# that contains a list of pointers to all the call sites in the kernel that
+# call mcount. Later on boot up, the kernel will read this list, save the 
+# locations and turn them into nops. When tracing or profiling is later 
+# enabled, these locations will then be converted back to pointers to 
+# some function.

Yes it is in .init.data, but the point is that we called it __mcount_loc in the object files. I don't want to blindly throw that name away.

Something to the lines of this.

Each object file will have a section called __mcount_loc that will hold the list of pointers to mcount callers. After final linking, the vmlinux will have within .init.data the list of all callers to mcount between _start_mcount_loc and __stop_mcount_loc.

scripts/recordmcount.pl line 21

Quoted 11 lines
scripts/recordmcount.pl +3 -4
 #
 # This is no easy feat. This script is called just after the original
 # object is compiled and before it is linked.
 #
-# The references to the call sites are offsets from the section of text
-# that the call site is in. Hence, all functions in a section that
-# has a call site to mcount, will have the offset from the beginning of
-# the section and not the beginning of the function.
+# When parse this object file using 'objdump', the references to the call 
+# sites are offsets from the section that the call site is in. Hence, all 
+# functions in a section that  has a call site to mcount, will have the 
                                ^^
                            double space

scripts/recordmcount.pl line 29

Quoted 11 lines
scripts/recordmcount.pl +7 -3
+# offset from the beginning of the section and not the beginning of the 
+# function.
 #
-# The trick is to find a way to record the beginning of the section.
-# The way we do this is to look at the first function in the section
-# which will also be the location of that section after final link.
+# But where this section will reside finally in vmlinx is undetermined at 
+# this point. So we can't use this kind of offsets to record the final
+# address of this call site.
+#
+# The trick is to change the call offset refering the start of a section to

referring

scripts/recordmcount.pl line 30

Quoted 1 line
scripts/recordmcount.pl +1 -0
+# refering a function symbol in this section. During the link step, 'ld' will

referring

The rest looks good.

-- Steve

Quoted 6 lines
scripts/recordmcount.pl +2 -1
+# compute the final address according to the information we record.
+# 
 # e.g.
 #
 #  .section ".sched.text", "ax"
-#