Sorry. I'm a fresh. This is my first patch commit. And still now *mutt*
does not configured properly.
I will study howto and commit patch again. Sorry.
-------------------
On Wed, Jan 16, 2013 at 03:04:25PM +0800, yanyg wrote:
@sysfs_update_group@ always fails and borns a WARNING CALL-trace
if the group exists. BUT still now, no *reference* of this routine at all.
I don't understand.
OK, sysfs routine @sysfs_update_group@ *never* works properly
>From fe40df7d716658a1fbdab79768628c81d4363428 Mon Sep 17 00:00:00 2001
From: yanyg <yany...@gmail.com>
Date: Wed, 16 Jan 2013 11:25:15 +0800
Subject: [PATCH] sysfs-fixes:
@sysfs_update_group@ always fail and generates WARNING CALL trace.
fix @internal_create_group@ to avoid it.
Why did you embed the patch here, with the words above it? I would have
to edit it to apply it :(
It's also not in a format I can apply (no signed-off-by: line.)
So sorry.
---
fs/sysfs/group.c | 14 ++++++++++----
1 files changed, 10 insertions(+), 4 deletions(-)
diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c
index 2df555c..51fcf62 100644
--- a/fs/sysfs/group.c
+++ b/fs/sysfs/group.c
@@ -60,7 +60,7 @@ static int internal_create_group(struct kobject *kobj, int
update,
const struct attribute_group *grp)
{
struct sysfs_dirent *sd;
- int error;
+ int error, do_put = 0;
BUG_ON(!kobj || (!update && !kobj->sd));
@@ -73,9 +73,14 @@ static int internal_create_group(struct kobject *kobj, int update,
return -EINVAL;
}
if (grp->name) {
- error = sysfs_create_subdir(kobj, grp->name, &sd);
- if (error)
- return error;
+ sd = sysfs_get_dirent(kobj->sd, NULL, grp->name);
+ if (sd) {
+ do_put = 1;
+ } else {
+ error = sysfs_create_subdir(kobj, grp->name, &sd);
+ if (error)
+ return error;
+ }
} else
sd = kobj->sd;
sysfs_get(sd);
@@ -85,6 +90,7 @@ static int internal_create_group(struct kobject *kobj, int
update,
sysfs_remove_subdir(sd);
}
sysfs_put(sd);
+ if (do_put) sysfs_put(sd);
Always run your patches through scripts/checkpatch.pl and fix the
problems it tells you about.
What code in the kernel is hitting this today?
I write a simply virtual block device driver which calls
@sysfs_update_group@ to updates attributes and modes, it fails.
YOU can reproduce the bug by append @sysfs_update_group@ at below any of
@sysfs_create_group@ call.
thanks,
greg k-h
So too sorry for my silly.
yanyg02
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
