Re: [PATCH 0/3] livepatch: cleanup kpl_patch kobject release

On Sun, Oct 31, 2021 at 10:26:18PM -0400, Joe Lawrence wrote:

On Sat, Oct 30, 2021 at 12:36:28AM +0800, Ming Lei wrote:

quoted

On Fri, Oct 29, 2021 at 09:51:54AM -0400, Joe Lawrence wrote:

quoted

On Thu, Oct 28, 2021 at 08:57:31PM +0800, Ming Lei wrote:

quoted

Hello,

The 1st patch moves module_put() to release handler of klp_patch
kobject.

The 2nd patch changes to free klp_patch and other kobjects without
klp_mutex.

The 3rd patch switches to synchronous kobject release for klp_patch.

Hi Ming,

I gave the patchset a spin on top of linus tree @ 1fc596a56b33 and ended
up with a stuck task:

Test
----
Enable the livepatch selftests:
  $ grep CONFIG_TEST_LIVEPATCH .config
  CONFIG_TEST_LIVEPATCH=m

Run a continuous kernel build in the background:
  $ while (true); do make clean && make -j$(nproc); done

While continuously executing the selftests:
  $ while (true); do make -C tools/testing/selftests/livepatch/ run_tests; done

Results
-------

Hello Joe,

Thanks for the test!

Can you replace the 3rd patch with the following one then running the test again?

From 599e96f79aebc388ef3854134312c6039a7884bf Mon Sep 17 00:00:00 2001
From: Ming Lei <redacted>
Date: Thu, 28 Oct 2021 20:11:23 +0800
Subject: [PATCH 3/3] livepatch: free klp_patch object synchronously

klp_mutex isn't acquired before calling kobject_put(klp_patch), so it is
fine to free klp_patch object synchronously.

One issue is that enabled store() method, in which the klp_patch kobject
itself is deleted & released. However, sysfs has provided APIs for dealing
with this corner case, so use sysfs_break_active_protection() and
sysfs_unbreak_active_protection() for releasing klp_patch kobject from
enabled_store().

Signed-off-by: Ming Lei <redacted>
---
 include/linux/livepatch.h     |  1 -
 kernel/livepatch/core.c       | 32 +++++++++++++-------------------
 kernel/livepatch/core.h       |  2 +-
 kernel/livepatch/transition.c |  2 +-
 4 files changed, 15 insertions(+), 22 deletions(-)

diff --git a/include/linux/livepatch.h b/include/linux/livepatch.h
index 9712818997c5..4dcebf52fac5 100644
--- a/include/linux/livepatch.h
+++ b/include/linux/livepatch.h

@@ -169,7 +169,6 @@ struct klp_patch {
 	struct list_head obj_list;
 	bool enabled;
 	bool forced;
-	struct work_struct free_work;
 };
 
 #define klp_for_each_object_static(patch, obj) \

diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c
index 9ede093d699a..6cfc54f6bdcc 100644
--- a/kernel/livepatch/core.c
+++ b/kernel/livepatch/core.c

@@ -337,6 +337,7 @@ static ssize_t enabled_store(struct kobject *kobj, struct kobj_attribute *attr,
 	int ret;
 	bool enabled;
 	LIST_HEAD(to_free);
+	struct kernfs_node *kn = NULL;
 
 	ret = kstrtobool(buf, &enabled);
 	if (ret)

@@ -369,7 +370,14 @@ static ssize_t enabled_store(struct kobject *kobj, struct kobj_attribute *attr,
 out:
 	mutex_unlock(&klp_mutex);
 
-	klp_free_patches_async(&to_free);
+	if (list_empty(&to_free)) {
+		kn = sysfs_break_active_protection(kobj, &attr->attr);
+		WARN_ON_ONCE(!kn);
+		sysfs_remove_file(kobj, &attr->attr);
+		klp_free_patches(&to_free);
+		if (kn)
+			sysfs_unbreak_active_protection(kn);
+	}
 
 	if (ret)
 		return ret;

@@ -684,32 +692,19 @@ static void klp_free_patch_finish(struct klp_patch *patch)
 	kobject_put(&patch->kobj);
 }
 
-/*
- * The livepatch might be freed from sysfs interface created by the patch.
- * This work allows to wait until the interface is destroyed in a separate
- * context.
- */
-static void klp_free_patch_work_fn(struct work_struct *work)
-{
-	struct klp_patch *patch =
-		container_of(work, struct klp_patch, free_work);
-
-	klp_free_patch_finish(patch);
-}
-
-static void klp_free_patch_async(struct klp_patch *patch)
+static void klp_free_patch(struct klp_patch *patch)
 {
 	klp_free_patch_start(patch);
-	schedule_work(&patch->free_work);
+	klp_free_patch_finish(patch);
 }
 
-void klp_free_patches_async(struct list_head *to_free)
+void klp_free_patches(struct list_head *to_free)
 {
 	struct klp_patch *patch, *tmp_patch;
 
 	list_for_each_entry_safe(patch, tmp_patch, to_free, list) {
 		list_del_init(&patch->list);
-		klp_free_patch_async(patch);
+		klp_free_patch(patch);
 	}
 }
 

@@ -873,7 +868,6 @@ static int klp_init_patch_early(struct klp_patch *patch)
 	kobject_init(&patch->kobj, &klp_ktype_patch);
 	patch->enabled = false;
 	patch->forced = false;
-	INIT_WORK(&patch->free_work, klp_free_patch_work_fn);
 
 	klp_for_each_object_static(patch, obj) {
 		if (!obj->funcs)

diff --git a/kernel/livepatch/core.h b/kernel/livepatch/core.h
index 8ff97745ba40..ea593f370049 100644
--- a/kernel/livepatch/core.h
+++ b/kernel/livepatch/core.h

@@ -13,7 +13,7 @@ extern struct list_head klp_patches;
 #define klp_for_each_patch(patch)	\
 	list_for_each_entry(patch, &klp_patches, list)
 
-void klp_free_patches_async(struct list_head *to_free);
+void klp_free_patches(struct list_head *to_free);
 void klp_unpatch_replaced_patches(struct klp_patch *new_patch);
 void klp_discard_nops(struct klp_patch *new_patch);
 

diff --git a/kernel/livepatch/transition.c b/kernel/livepatch/transition.c
index a9ebc9c5db02..3eff5fc0deee 100644
--- a/kernel/livepatch/transition.c
+++ b/kernel/livepatch/transition.c

@@ -41,7 +41,7 @@ static void klp_transition_work_fn(struct work_struct *work)
 
 	mutex_unlock(&klp_mutex);
 
-	klp_free_patches_async(&to_free);
+	klp_free_patches(&to_free);
 }
 static DECLARE_DELAYED_WORK(klp_transition_work, klp_transition_work_fn);
 

-- 
2.31.1

Hi Ming,

The previous test runs without hung tasks, but I noticed that is
probably because the selftests quickly wedge.  A simple reproducer for
that:

# (background load)
% while(true); do make clean && make -j$(nproc); done
vs.
# (selftests)
% while(true); do ./tools/testing/selftests/livepatch/test-livepatch.sh || break; done
TEST: basic function patching ... ok
TEST: multiple livepatches ... ok
TEST: atomic replace livepatch ... ok
TEST: basic function patching ... ok
TEST: multiple livepatches ... ok
TEST: atomic replace livepatch ... ok
TEST: basic function patching ... ok
TEST: multiple livepatches ... ERROR: failed to disable livepatch test_klp_livepatch

% lsmod | grep test_klp
test_klp_livepatch     16384  1

% cat /sys/kernel/livepatch/test_klp_livepatch/enabled 
0

% rmmod test_klp_livepatch
rmmod: ERROR: Module test_klp_livepatch is in use

[  249.870587] ===== TEST: multiple livepatches =====
[  249.885520] % modprobe test_klp_livepatch
[  249.916987] livepatch: enabling patch 'test_klp_livepatch'
[  249.923131] livepatch: 'test_klp_livepatch': initializing patching transition
[  249.925575] livepatch: 'test_klp_livepatch': starting patching transition
[  249.934215] livepatch: 'test_klp_livepatch': completing patching transition
[  249.934337] livepatch: 'test_klp_livepatch': patching complete
[  249.946294] test_klp_livepatch: this has been live patched
[  249.963337] % modprobe test_klp_atomic_replace replace=0
[  249.996371] livepatch: enabling patch 'test_klp_atomic_replace'
[  250.002998] livepatch: 'test_klp_atomic_replace': initializing patching transition
[  250.005333] livepatch: 'test_klp_atomic_replace': starting patching transition
[  250.014364] livepatch: 'test_klp_atomic_replace': completing patching transition
[  250.014471] livepatch: 'test_klp_atomic_replace': patching complete
[  250.027259] test_klp_livepatch: this has been live patched
[  250.036050] test_klp_atomic_replace: this has been live patched
[  250.046347] % echo 0 > /sys/kernel/livepatch/test_klp_atomic_replace/enabled
[  250.054403] livepatch: 'test_klp_atomic_replace': initializing unpatching transition
[  250.054574] livepatch: 'test_klp_atomic_replace': starting unpatching transition
[  251.764849] livepatch: 'test_klp_atomic_replace': completing unpatching transition
[  251.799266] livepatch: 'test_klp_atomic_replace': unpatching complete
[  251.911706] % rmmod test_klp_atomic_replace
[  251.970438] test_klp_livepatch: this has been live patched
[  251.980347] % echo 0 > /sys/kernel/livepatch/test_klp_livepatch/enabled
[  251.987936] livepatch: 'test_klp_livepatch': initializing unpatching transition
[  251.988115] livepatch: 'test_klp_livepatch': starting unpatching transition
[  251.997033] livepatch: 'test_klp_livepatch': completing unpatching transition
[  252.027090] livepatch: 'test_klp_livepatch': unpatching complete
[  313.289932] ERROR: failed to disable livepatch test_klp_livepatch

In this case, the "failed to disable" msg occurs because the sysfs
interface / module remain present.

Hi Joe,

Thanks for your test!

Hammm, the check of list_empty() in enabled_store() is inverted, please try the
V3:

https://lore.kernel.org/lkml/20211101112548.3364086-4-ming.lei@redhat.com/T/#u (local)


Thanks,
Ming