Skip to contentSkip to conversation

Fix some issues at gcov parse tool and add a new table on reports

PATCH In review Mauro Carvalho Chehab 2 participants 8 messages Sep 22, 2022 – Oct 5, 2022

lore
Share & tools
lore
#0 lore raw

cover Fix some issues at gcov parse tool and add a new table on reports

From: Mauro Carvalho Chehab <[email protected]>

Patches 1 to 4 fix some bugs at the script, some of them causing it to crash.

Patch 5 address a GCC optimization-related issue, where the function is renamed at the object file, gaining a .irda.<number> extension.

Patch 6 adds a new functionality at the tool when producing an HTML output. It basically calculates the interception of all code coverage sets, and shows how many functions are common to all sets and with one are unique to an specific tests. It helps to discover cases where tests or test sets are not needed to be taken into account when producing a report.

Mauro Carvalho Chehab (6):
scripts/code_cov_parse_info: avoid warning messages on parsing scripts/code_cov_parse_info: fix --only-drm filtering rules scripts/code_cov_parse_info: fix brf record handling scripts/code_cov_parse_info: Fix indentation at the output html file scripts/code_cov_parse_info: take gcc IPA optimizations into account scripts/code_cov_parse_info: print common coverage and extra per
function

 scripts/code_cov_parse_info | 111 ++++++++++++++++++++++++++++++++----
 1 file changed, 99 insertions(+), 12 deletions(-)
Signature · 1 line
2.37.2
#1 lore raw

PATCH 1/6 scripts/code_cov_parse_info: avoid warning messages on parsing

From: Mauro Carvalho Chehab <[email protected]>

Some fields (FNH, BRH, LH) produce warning messages when they're found at the info file.

Fix it.

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
scripts/code_cov_parse_info +3 -3
index 09f9aba5760b..78b7e2dada5e 100755
@@ -195,7 +195,7 @@ sub parse_info_data($)
 		# FNH:<number of function hit>
 		if (m/^FNH:(-?\d+)/) {
 			my $hits = $1;
-			if ($record{$source}{$func}{fnh} < $hits) {
+			if (!defined($record{$source}{$func}{fnh}) || $record{$source}{$func}{fnh} < $hits) {
 				$record{$source}{$func}{fnh} = $hits;
 			}
 			next;
@@ -234,7 +234,7 @@ sub parse_info_data($)
 		# BRH:<number of branches hit>
 		if (m/^BRH:(-?\d+)/) {
 			my $hits = $1;
-			if ($record{$source}{$func}{brh} < $hits) {
+			if (!defined($record{$source}{$func}{brh}) || $record{$source}{$func}{brh} < $hits) {
 				$record{$source}{$func}{brh} = $hits;
 			}
 			next;
@@ -268,7 +268,7 @@ sub parse_info_data($)
 		# LH:<number of lines with a non-zero execution count>
 		if (m/^LH:(-?\d+)/) {
 			my $hits = $1;
-			if ($record{$source}{$func}{lh} < $hits) {
+			if (!defined($record{$source}{$func}{lh}) || $record{$source}{$func}{lh} < $hits) {
 				$record{$source}{$func}{lh} = $hits;
 			}
 			next;
Signature · 1 line
2.37.2
#2 lore raw

PATCH 2/6 scripts/code_cov_parse_info: fix --only-drm filtering rules

In reply to Mauro Carvalho Chehab

From: Mauro Carvalho Chehab <[email protected]>

The --only-drm filtering rules were filtering too much, ending to exclude functions on files with "h" on their names. Fix it.

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
scripts/code_cov_parse_info +5 -3
index 78b7e2dada5e..037a07e26b1d 100755
@@ -857,8 +857,8 @@ if ($only_i915) {
 
 if ($only_drm) {
 	# Please keep in sync with the documentation
-	push @src_exclude_regexes, "trace.*\.h";
-	push @src_exclude_regexes, "drm.*\.h";
+	push @src_exclude_regexes, "trace.*\.h\$";
+	push @src_exclude_regexes, "^/drm/";
 }
 
 $str = open_filter_file($func_filters, \@func_regexes, \@func_exclude_regexes);
@@ -1051,7 +1051,9 @@ E. g. it will exclude *.h files that match the following regular expressions:
 
 And *.h files that don't match:
 
-	- drm
+	- /drm/
+
+(e. g. macros and other code outside DRM subsystem)
 
 =item B<--only-i915> or B<--only_i915>
 
Signature · 1 line
2.37.2
#3 lore raw

PATCH 3/6 scripts/code_cov_parse_info: fix brf record handling

In reply to Mauro Carvalho Chehab

From: Mauro Carvalho Chehab <[email protected]>

Don't override $record{source} due to BRF, or the script will crash.

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
scripts/code_cov_parse_info +1 -1
index 037a07e26b1d..ef26192f2456 100755
@@ -228,7 +228,7 @@ sub parse_info_data($)
 
 		# BRF:<number of branches found>
 		if (m/^BRF:(-?\d+)/) {
-			$record{$source}{brf} = $1;
+			$record{$source}{$func}{brf} = $1;
 			next;
 		}
 		# BRH:<number of branches hit>
Signature · 1 line
2.37.2
#4 lore raw

PATCH 4/6 scripts/code_cov_parse_info: Fix indentation at the output html file

In reply to Mauro Carvalho Chehab

From: Mauro Carvalho Chehab <[email protected]>

The indentation is not relevant for the output, but it makes easier when debugging issues at the produced file. So, fix it.

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
scripts/code_cov_parse_info +3 -3
index ef26192f2456..e8bac730e88b 100755
@@ -650,7 +650,7 @@ sub generate_report($)
 	} else {
 		print OUT "    <td>N. A.</td>\n";
 	}
-	print OUT "  <td>" . $stats{"func_count"} . "</td>";
+	print OUT "    <td>" . $stats{"func_count"} . "</td>\n";
 	print OUT "  </tr><tr>\n";
 
 	print OUT "    <td><b>Branches</b></td>\n";
@@ -671,7 +671,7 @@ sub generate_report($)
 	} else {
 		print OUT "    <td>N. A.</td>\n";
 	}
-	print OUT "  <td>" . $stats{"branch_count"} . "</td>";
+	print OUT "    <td>" . $stats{"branch_count"} . "</td>\n";
 	print OUT "  </tr><tr>\n";
 
 	print OUT "    <td><b>Lines</b></td>\n";
@@ -693,7 +693,7 @@ sub generate_report($)
 	} else {
 		print OUT "    <td>N. A.</td>\n";
 	}
-	print OUT "  <td>" . $stats{"line_count"} . "</td>";
+	print OUT "    <td>" . $stats{"line_count"} . "</td>\n";
 
 	# If there are more than one tests per file, report them
 	my $total = scalar(keys %test_names);
Signature · 1 line
2.37.2
Reviewed-by on v1 4/6
#5 lore raw

Re: [igt-dev] [PATCH 4/6] scripts/code_cov_parse_info: Fix indentation at the output html file

Hi Mauro,

Quoted 4 lines · Mauro Carvalho Chehab

From: Mauro Carvalho Chehab <[email protected]>

The indentation is not relevant for the output, but it makes easier when debugging issues at the produced file. So, fix it.

--------------------------------------------- ^^^^^^^^^^ Please drop this from commit message,
rest looks good so

Quoted 44 lines

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
scripts/code_cov_parse_info +3 -3
index ef26192f2456..e8bac730e88b 100755
@@ -650,7 +650,7 @@ sub generate_report($)
 	} else {
 		print OUT "    <td>N. A.</td>\n";
 	}
-	print OUT "  <td>" . $stats{"func_count"} . "</td>";
+	print OUT "    <td>" . $stats{"func_count"} . "</td>\n";
 	print OUT "  </tr><tr>\n";
 
 	print OUT "    <td><b>Branches</b></td>\n";
@@ -671,7 +671,7 @@ sub generate_report($)
 	} else {
 		print OUT "    <td>N. A.</td>\n";
 	}
-	print OUT "  <td>" . $stats{"branch_count"} . "</td>";
+	print OUT "    <td>" . $stats{"branch_count"} . "</td>\n";
 	print OUT "  </tr><tr>\n";
 
 	print OUT "    <td><b>Lines</b></td>\n";
@@ -693,7 +693,7 @@ sub generate_report($)
 	} else {
 		print OUT "    <td>N. A.</td>\n";
 	}
-	print OUT "  <td>" . $stats{"line_count"} . "</td>";
+	print OUT "    <td>" . $stats{"line_count"} . "</td>\n";
 
 	# If there are more than one tests per file, report them
 	my $total = scalar(keys %test_names);
Signature · 1 line
2.37.2
#6 lore raw

PATCH 5/6 scripts/code_cov_parse_info: take gcc IPA optimizations into account

In reply to Mauro Carvalho Chehab

From: Mauro Carvalho Chehab <[email protected]>

When -O2/-O3 is used with GCC, IPA will produce function variants at the object code. So, for a function or macro foo(), it may produce several actual functions at the object code, like:

	foo.isra.3
	foo.isra.4

For the purposes of code coverage, all of those are the same function, so use a regex with "\w+", which will exclude the .isra.<number>, considering all variants as the same function.

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
scripts/code_cov_parse_info +6 -2
index e8bac730e88b..7f3e1f9ef116 100755
@@ -136,7 +136,9 @@ sub parse_info_data($)
 		# Function coverage
 
 		# FN:<line number of function start>,<function name>
-		if (m/^FN:(-?\d+),(.*)/) {
+		# Note: /w+ intentionally removes IPA gcc-optimization names, as
+		# this is more related to branch coverage
+		if (m/^FN:(-?\d+),(\w+)/) {
 			my $ln = $1;
 
 			$func = $2;
@@ -156,7 +158,9 @@ sub parse_info_data($)
 
 		# Parse functions that were actually used
 		# FNDA:<execution count>,<function name>
-		if (m/^FNDA:(-?\d+),(.*)/) {
+		# Note: /w+ intentionally removes IPA gcc-optimization names, as
+		# this is more related to branch coverage
+		if (m/^FNDA:(-?\d+),(\w+)/) {
 			my $count = $1;
 
 			# Negative gcov results are possible, as reported at:
Signature · 1 line
2.37.2
#7 lore raw

PATCH 6/6 scripts/code_cov_parse_info: print common coverage and extra per function

In reply to Mauro Carvalho Chehab

From: Mauro Carvalho Chehab <[email protected]>

While analyzing the data, it is nice to have a quick summary about how much each set of the analyzed data contributes to the total.

So, generate a function summary report with the interception of all the sets ("common function coverage"), plus the difference that an individual test group contributes.

This is useful to provide a quick feedback on several situations:

- When comparing FULL with BAT:
- how many functions are added on FULL?
- are there functions that are only on BAT?

- When comparing different machines:
- How many functions each individual machine has over the common group?

- When comparing a group of tests:
- How many functions an individual test adds for the coverage that weren't covered at the common group yet?

---

To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/6] at: https://lore.kernel.org/all/[email protected]/

 scripts/code_cov_parse_info | 81 +++++++++++++++++++++++++++++++++++++
 1 file changed, 81 insertions(+)
scripts/code_cov_parse_info +81 -0
index 7f3e1f9ef116..72a1e8c2b0ad 100755
@@ -574,6 +574,8 @@ sub generate_report($)
 	%used_source = ();
 	%files = ();
 	%test_names = ();
+	my @all_func_keys;
+	my %info_files_with_func;
 
 	foreach my $f (@info_files) {
 		foreach my $source (keys(%{$report{$f}{"all_line"}})) {
@@ -586,6 +588,14 @@ sub generate_report($)
 			foreach my $file (keys(%{$report{$f}{"all_func"}{$func}})) {
 				$all_func{$func}{$file}->{ln} = $report{$f}{"all_func"}{$func}{$file}->{ln};
 				$used_func{$func}->{$file} = 1 if ($report{$f}{"used_func"}{$func}->{$file});
+				if ($report{$f}{"used_func"}{$func}->{$file}) {
+					$used_func{$func}->{$file} = 1;
+					if (!$info_files_with_func{"$file $func"}) {
+						$info_files_with_func{"$file $func"} = 1;
+					} else {
+						$info_files_with_func{"$file $func"}++;
+					}
+				}
 			}
 		}
 		foreach my $source (keys(%{$report{$f}{"all_branch"}})) {
@@ -603,6 +613,25 @@ sub generate_report($)
 	}
 	gen_stats();
 
+	# Gen code coverage set comparision counters
+	my $common_func_count = 0;
+	foreach my $k (keys %info_files_with_func) {
+		$common_func_count++ if ($info_files_with_func{$k} == scalar (@info_files));
+	}
+	foreach my $f (@info_files) {
+		$report{$f}{"more_func"} = 0;
+		$report{$f}{"uniq_func"} = 0;
+		foreach my $func (keys(%{$report{$f}{"all_func"}})) {
+			foreach my $file (keys(%{$report{$f}{"all_func"}{$func}})) {
+				next if (!$report{$f}{"used_func"}{$func}->{$file});
+				next if ($info_files_with_func{"$file $func"} == scalar (@info_files));
+
+				$report{$f}{"more_func"}++;
+				$report{$f}{"uniq_func"}++ if ($info_files_with_func{"$file $func"} == 1);
+			}
+		}
+	}
+
 	# Colors for the html output
 
 	my $red    = "style=\"background-color:#ffb3b3\"";
@@ -720,6 +749,58 @@ sub generate_report($)
 	}
 	print OUT "  </tr>\n</table><p/>\n\n";
 
+	# Print function diff
+	print OUT "  <h2>Differences on function coverage</h2>\n";
+	print OUT "<table width=\"100%\" border=1 cellspacing=0 cellpadding=0>\n  <tr>\n";
+	print OUT "    <th></th>\n";
+	print OUT "    <th>Common to all</th>\n";
+	foreach my $f (@info_files) {
+		print OUT "    <th>$f</th>\n";
+	}
+	print OUT "    <th>TOTAL</th>\n";
+
+	print OUT "  </tr><tr>\n";
+	print OUT "    <td>#Functions per category</td>\n";
+
+	printf OUT "    <td>%d</td>\n", $common_func_count;
+
+	foreach my $f (@info_files) {
+		my %st = %{$report{$f}{"stats"}};
+		if ($st{"func_count"}) {
+			printf OUT "    <td>%d</td>\n", $st{"func_used"};
+
+		}
+	}
+	print OUT "    <td>" . $stats{"func_count"} . "</td>\n";
+
+	print OUT "  </tr><tr>\n";
+	print OUT "    <td>#functions not in common</td>\n";
+	print OUT "    <td></td>\n";
+	foreach my $f (@info_files) {
+		my %st = %{$report{$f}{"stats"}};
+		if ($st{"func_count"}) {
+			printf OUT "    <td>%d</td>\n", $report{$f}{"more_func"};
+
+		}
+	}
+	print OUT "    <td></td>\n";
+
+	print OUT "  </tr><tr>\n";
+	print OUT "    <td>Unique functions</td>\n";
+	print OUT "    <td></td>\n";
+	foreach my $f (@info_files) {
+		my %st = %{$report{$f}{"stats"}};
+		if ($st{"func_count"}) {
+			printf OUT "    <td>%d</td>\n", $report{$f}{"uniq_func"};
+
+		}
+	}
+	print OUT "    <td></td>\n";
+
+	print OUT "  </tr>\n</table><p/>\n\n";
+
+	# Print the filters applied when generating the report
+
 	if ($filter_str ne "") {
 		printf OUT "<p>Filters: %s.</p>\n", $filter_str;
 	} else {
Signature · 1 line
2.37.2