Subject: [failures] kobject-dont-block-for-each-kobject_uevent.patch removed from -mm tree
To: vdavydov@xxxxxxxxxxxxx,cl@xxxxxxxxx,greg@xxxxxxxxx,penberg@xxxxxxxxxx,mm-commits@xxxxxxxxxxxxxxx
From: akpm@xxxxxxxxxxxxxxxxxxxx
Date: Thu, 13 Feb 2014 11:53:58 -0800
The patch titled
Subject: kobject: don't block for each kobject_uevent
has been removed from the -mm tree. Its filename was
kobject-dont-block-for-each-kobject_uevent.patch
This patch was dropped because it had testing failures
------------------------------------------------------
From: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx>
Subject: kobject: don't block for each kobject_uevent
Currently kobject_uevent has somewhat unpredictable semantics. The point
is, since it may call a usermode helper and wait for it to execute
(UMH_WAIT_EXEC), it is impossible to say for sure what lock dependencies
it will introduce for the caller - strictly speaking it depends on what fs
the binary is located on and the set of locks fork may take. There are
quite a few kobject_uevent's users that do not take this into account and
call it with various mutexes taken, e.g. rtnl_mutex, net_mutex, which
might potentially lead to a deadlock.
Since there is actually no reason to wait for the usermode helper to
execute there, let's make kobject_uevent start the helper asynchronously
with the aid of the UMH_NO_WAIT flag.
Personally, I'm interested in this, because I really want kobject_uevent
to be called under the slab_mutex in the slub implementation as it used to
be some time ago, because it greatly simplifies synchronization and
automatically fixes a kmemcg-related race. However, there was a deadlock
detected on an attempt to call kobject_uevent under the slab_mutex (see
https://lkml.org/lkml/2012/1/14/45), which was reported to be fixed by
releasing the slab_mutex for kobject_uevent. Unfortunately, there was no
information about who exactly blocked on the slab_mutex causing the
usermode helper to stall, neither have I managed to find this out or
reproduce the issue.
BTW, this is not the first attempt to make kobject_uevent use UMH_NO_WAIT.
Previous one was made by commit f520360d93c ("kobject: don't block for
each kobject_uevent"), but it was wrong (it passed arguments allocated on
stack to async thread) so it was reverted in 05f54c13cd0c ("Revert
"kobject: don't block for each kobject_uevent"."). It targeted on
speeding up the boot process though.
Signed-off-by: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx>
Cc: Greg KH <greg@xxxxxxxxx>
Cc: Christoph Lameter <cl@xxxxxxxxx>
Cc: Pekka Enberg <penberg@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/kobject.h | 1
lib/kobject_uevent.c | 42 ++++++++++++++++++++++++++++++++------
2 files changed, 37 insertions(+), 6 deletions(-)
diff -puN include/linux/kobject.h~kobject-dont-block-for-each-kobject_uevent include/linux/kobject.h
--- a/include/linux/kobject.h~kobject-dont-block-for-each-kobject_uevent
+++ a/include/linux/kobject.h
@@ -119,6 +119,7 @@ struct kobj_type {
};
struct kobj_uevent_env {
+ char *argv[3];
char *envp[UEVENT_NUM_ENVP];
int envp_idx;
char buf[UEVENT_BUFFER_SIZE];
diff -puN lib/kobject_uevent.c~kobject-dont-block-for-each-kobject_uevent lib/kobject_uevent.c
--- a/lib/kobject_uevent.c~kobject-dont-block-for-each-kobject_uevent
+++ a/lib/kobject_uevent.c
@@ -124,6 +124,30 @@ static int kobj_usermode_filter(struct k
return 0;
}
+static int init_uevent_argv(struct kobj_uevent_env *env, const char *subsystem)
+{
+ int len;
+
+ len = strlcpy(&env->buf[env->buflen], subsystem,
+ sizeof(env->buf) - env->buflen);
+ if (len >= (sizeof(env->buf) - env->buflen)) {
+ WARN(1, KERN_ERR "init_uevent_argv: buffer size too small\n");
+ return -ENOMEM;
+ }
+
+ env->argv[0] = uevent_helper;
+ env->argv[1] = &env->buf[env->buflen];
+ env->argv[2] = NULL;
+
+ env->buflen += len + 1;
+ return 0;
+}
+
+static void cleanup_uevent_env(struct subprocess_info *info)
+{
+ kfree(info->data);
+}
+
/**
* kobject_uevent_env - send an uevent with environmental data
*
@@ -301,11 +325,8 @@ int kobject_uevent_env(struct kobject *k
/* call uevent_helper, usually only enabled during early boot */
if (uevent_helper[0] && !kobj_usermode_filter(kobj)) {
- char *argv [3];
+ struct subprocess_info *info;
- argv [0] = uevent_helper;
- argv [1] = (char *)subsystem;
- argv [2] = NULL;
retval = add_uevent_var(env, "HOME=/");
if (retval)
goto exit;
@@ -313,9 +334,18 @@ int kobject_uevent_env(struct kobject *k
"PATH=/sbin:/bin:/usr/sbin:/usr/bin");
if (retval)
goto exit;
+ retval = init_uevent_argv(env, subsystem);
+ if (retval)
+ goto exit;
- retval = call_usermodehelper(argv[0], argv,
- env->envp, UMH_WAIT_EXEC);
+ retval = -ENOMEM;
+ info = call_usermodehelper_setup(env->argv[0], env->argv,
+ env->envp, GFP_KERNEL,
+ NULL, cleanup_uevent_env, env);
+ if (info)
+ retval = call_usermodehelper_exec(info, UMH_NO_WAIT);
+ if (!retval)
+ env = NULL; /* will be freed by cleanup_uevent_env */
}
exit:
_
Patches currently in -mm which might be from vdavydov@xxxxxxxxxxxxx are
origin.patch
mm-vmscan-respect-numa-policy-mask-when-shrinking-slab-on-direct-reclaim.patch
mm-vmscan-move-call-to-shrink_slab-to-shrink_zones.patch
mm-vmscan-remove-shrink_control-arg-from-do_try_to_free_pages.patch
mm-vmscan-shrink_slab-rename-max_pass-freeable.patch
slub-do-not-drop-slab_mutex-for-sysfs_slab_add.patch
linux-next.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
