Skip to contentSkip to conversation

localmodconfig: Fix missing depends of config files included in if statements

GIT PULL New Steven Rostedt 2 participants 10 messages Apr 29, 2013 – Apr 30, 2013

Subsystems (4)
lore
Share & tools
lore

Steven Rostedt

author maintainer
#0 lore raw

Linus,

A bug was recently found (today) in the make localmodconfig where it would miss dependencies of config files that are included in other config files inside an if statement.

Also added a debug print that helped in solving this bug.

Please pull the latest ktest-v3.10 tree, which can be found at:

git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-ktest.git
ktest-v3.10

Head SHA1: 7714fe5ac0cf752eae7a20ef23408aa26271a3eb

Steven Rostedt (2):
localmodconfig: Add debug prints for dependencies of module configs
localmodconfig: Process source kconfig files as they are found

----

 scripts/kconfig/streamline_config.pl |   26 ++++++++++++++++----------
 1 file changed, 16 insertions(+), 10 deletions(-)
author maintainer
#1 lore raw

PATCH 1/2 localmodconfig: Add debug prints for dependencies of module configs

From: Steven Rostedt <[email protected]>

When a config for a module is added to the list to save in the final config file, add a print to show what dependencies are used. This is useful to debug when a config is disabled by the make oldconfig after localmodconfig is finished.

This print only appears if the environment variable LOCALMODCONFIG_DEBUG is defined.

---
 scripts/kconfig/streamline_config.pl |    9 +++++++++
 1 file changed, 9 insertions(+)
scripts/kconfig/streamline_config.pl +9 -0
index 3368939..343a568 100644
@@ -396,6 +396,15 @@ foreach my $module (keys(%modules)) {
 	foreach my $conf (@arr) {
 	    $configs{$conf} = $module;
 	    dprint "$conf added by direct ($module)\n";
+	    if ($debugprint) {
+		my $c=$conf;
+		$c =~ s/^CONFIG_//;
+		if (defined($depends{$c})) {
+		    dprint " deps = $depends{$c}\n";
+		} else {
+		    dprint " no deps\n";
+		}
+	    }
 	}
     } else {
 	# Most likely, someone has a custom (binary?) module loaded.
Signature · 1 line
1.7.10.4
author maintainer
#2 lore raw

PATCH 2/2 localmodconfig: Process source kconfig files as they are found

In reply to Steven Rostedt

From: Steven Rostedt <[email protected]>

A bug was reported that caused localmodconfig to not keep all the dependencies of ATH9K. This was caused by the kconfig file:

In drivers/net/wireless/ath/Kconfig:

---
if ATH_CARDS

config ATH_DEBUG
bool "Atheros wireless debugging"
---help---
Say Y, if you want to debug atheros wireless drivers. Right now only ath9k makes use of this.

source "drivers/net/wireless/ath/ath5k/Kconfig" source "drivers/net/wireless/ath/ath9k/Kconfig" source "drivers/net/wireless/ath/carl9170/Kconfig" source "drivers/net/wireless/ath/ath6kl/Kconfig" source "drivers/net/wireless/ath/ar5523/Kconfig" source "drivers/net/wireless/ath/wil6210/Kconfig"

endif
---

The current way kconfig works, it processes new source files after the first file is completed. It creates an array of new source config files and when the one file is finished, it continues with the next file.

Unfortunately, this means that it loses the fact that the source file is within an "if" statement, and this means that each of these source file's configs will not have the proper dependencies set.

As ATH9K requires ATH_CARDS set, the localmodconfig did not see that dependency, and did not enable ATH_CARDS. When the oldconfig was run, it forced ATH9K to be disabled.

---
 scripts/kconfig/streamline_config.pl |   17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)
scripts/kconfig/streamline_config.pl +7 -10
index 343a568..4606cdf 100644
@@ -156,7 +156,6 @@ sub read_kconfig {
 
     my $state = "NONE";
     my $config;
-    my @kconfigs;
 
     my $cont = 0;
     my $line;
@@ -190,7 +189,13 @@ sub read_kconfig {
 
 	# collect any Kconfig sources
 	if (/^source\s*"(.*)"/) {
-	    $kconfigs[$#kconfigs+1] = $1;
+	    my $kconfig = $1;
+	    # prevent reading twice.
+	    if (!defined($read_kconfigs{$kconfig})) {
+		$read_kconfigs{$kconfig} = 1;
+		read_kconfig($kconfig);
+	    }
+	    next;
 	}
 
 	# configs found
@@ -250,14 +255,6 @@ sub read_kconfig {
 	}
     }
     close($kinfile);
-
-    # read in any configs that were found.
-    foreach my $kconfig (@kconfigs) {
-	if (!defined($read_kconfigs{$kconfig})) {
-	    $read_kconfigs{$kconfig} = 1;
-	    read_kconfig($kconfig);
-	}
-    }
 }
 
 if ($kconfig) {
Signature · 1 line
1.7.10.4
#3 lore raw
In reply to Steven Rostedt
Quoted 6 lines · Steven Rostedt

Please pull the latest ktest-v3.10 tree, which can be found at:

git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-ktest.git
ktest-v3.10

Hmm. This is your ktest tree (which I already pulled the ktest changes from), not your localmodconfig tree. But then all the explanations are about your localmodconfig changes.

I assume you *meant* to ask me to pull your linux-kconfig.git tree? That has a "ktest-v3.10" tag too, but I think that's a mistake too, and you *meant* to make a tag called "localconfig" like you have done in the past.

There's just too much confusion here for me to touch anything at all, so please fix things up.

Linus

#4 lore raw
Quoted 3 lines · Linus Torvalds

There's just too much confusion here for me to touch anything at all, so please fix things up.

Oh, and you might as well check that I resolved the conflicts in the trace pull correctly while you're at it. They looked pretty obvious, and it compiles for me, but ..

Linus

author maintainer
#5 lore raw
Quoted 9 lines · Linus Torvalds
Quoted 3 lines · Linus Torvalds

There's just too much confusion here for me to touch anything at all, so please fix things up.

Oh, and you might as well check that I resolved the conflicts in the trace pull correctly while you're at it. They looked pretty obvious, and it compiles for me, but ..

Almost.

I'll also run my full ftrace test suite on your latest kernel, and see if it finds anything else.

Please apply:

---
tracing: Fix small merge bug

During the 3.10 merge, a conflict happened and the resolution was almost, but not quite, correct. An if statement was reversed.

kernel/trace/trace.c +1 -1
index 581630a..ae6fa2d 100644
@@ -904,7 +904,7 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
 		return;
 
 	WARN_ON_ONCE(!irqs_disabled());
-	if (tr->allocated_snapshot) {
+	if (!tr->allocated_snapshot) {
 		/* Only the nop tracer should hit this when disabling */
 		WARN_ON_ONCE(tr->current_trace != &nop_trace);
 		return;
author maintainer
#6 lore raw
Quoted 4 lines · Steven Rostedt
Quoted 3 lines · Linus Torvalds
Quoted 1 line · Linus Torvalds
Quoted 9 lines
Quoted 6 lines
Quoted 2 lines

There's just too much confusion here for me to touch anything at all, so please fix things up.

Oh, and you might as well check that I resolved the conflicts in the trace pull correctly while you're at it. They looked pretty obvious, and it compiles for me, but ..

Almost.

BTW, what's the preferred method for this. I already posted a lot of work to linux-next when I found bugs that required going into your tree. I backported the fixes knowing that it will cause conflicts when you merge.

IIRC, you stated that you don't mind doing conflict resolutions yourself, so I did not try to fix it a head of time, as the conflicts were rather minor.

Should I have merged your tree and done the conflict resolutions myself, or was it OK to do what I did, and let you do the conflict resolution and send you any fixes that needed to be done afterward?

-- Steve

#7 lore raw
Quoted 6 lines · Steven Rostedt

BTW, what's the preferred method for this. I already posted a lot of work to linux-next when I found bugs that required going into your tree. I backported the fixes knowing that it will cause conflicts when you merge.

Quoted 7 lines

IIRC, you stated that you don't mind doing conflict resolutions yourself, so I did not try to fix it a head of time, as the conflicts were rather minor.

Should I have merged your tree and done the conflict resolutions myself, or was it OK to do what I did, and let you do the conflict resolution and send you any fixes that needed to be done afterward?

You did the right thing. In general, if you know there will be conflicts, it's nice if you mention then in the pull request, but for simple stuff like this it's really not a big deal. The fact that I screwed up and then missed a "!" when editing it all is embarrassing, but it wasn't because the conflict was really *complicated*, it was just stupid editing error.

If the conflicts are really complex, at some point I really enjoy getting a "here's a pre-merged branch if you prefer it", and if people send that, I still tend to do the merge myself, but then I often just compare against the pre-merged one afterwards to verify. But that's actually extra work, so I'd suggest doing that only for things that really warrant it. It's generally a bad thing if it happens, because it means that we had some bad workflow and people stepped on each others toes (or we had unlucky backports etc that ended up being in the same area as much more work).

Linus

author maintainer
#8 lore raw
In reply to Steven Rostedt
Quoted 3 lines · Steven Rostedt

I'll also run my full ftrace test suite on your latest kernel, and see if it finds anything else.

With the fix I sent, everything passed.

Thanks,

-- Steve

author maintainer
#9 lore raw
In reply to Linus Torvalds
Quoted 5 lines · Linus Torvalds
Quoted 4 lines · Steven Rostedt

Please pull the latest ktest-v3.10 tree, which can be found at:

Quoted 8 lines
Quoted 2 lines

git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-ktest.git
ktest-v3.10

Hmm. This is your ktest tree (which I already pulled the ktest changes from), not your localmodconfig tree. But then all the explanations are about your localmodconfig changes.

I assume you *meant* to ask me to pull your linux-kconfig.git tree?

Face palm!

Quoted 6 lines

That has a "ktest-v3.10" tag too, but I think that's a mistake too, and you *meant* to make a tag called "localconfig" like you have done in the past.

There's just too much confusion here for me to touch anything at all, so please fix things up.

Yeah, sorry. I have no idea what I was thinking. I was doing my double check of ktest changes and must have still had that on my mind when I switched over to the localmodconfig stuff.

I'll clean it up.

-- Steve