public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
From: "Keijiro Yano" <keijiro_yano@yahoo.co.jp>
To: <linux-mtd@lists.infradead.org>
Subject: jffs2 mutex problem
Date: Sat, 12 Nov 2005 00:43:32 +0900	[thread overview]
Message-ID: <000c01c5e6d6$abf22490$0b01a8c0@YANO> (raw)

Hello,

I find out the problem related to the mutex lock in JFFS2.
I always use Linux 2.6.10, but Linux 2.6.14 also has a same
problem, I think.

jffs2_create() calls jffs2_new_inode() that allocates new
"jffs2_inode_info" with jffs2_alloc_inode().
Generally, in this case, jffs2_i_init_once(), which is the
constructor for jffs2_inode_cachep, will be called and
initialize the mutex with lock state by the below code.

static void jffs2_i_init_once(void * foo, kmem_cache_t * cachep, unsigned 
long flags)
{
struct jffs2_inode_info *ei = (struct jffs2_inode_info *) foo;

if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
    SLAB_CTOR_CONSTRUCTOR) {
init_MUTEX_LOCKED(&ei->sem);

But, the constructor is called by kmem_cache_alloc()
only when no active objects are left in a cache.

Therefore, sometimes jffs2_i_init_once() will not be
called after jffs2_alloc_inode() is done.

So, jffs2_do_create(), which is called by jffs2_create(), will
do "up(&f->sem)" in unlocking state.

I try to fix this problem with the following patch. Please
review it and/or give me your comments.

--- fs/jffs2/os-linux.h 10 Feb 2005 10:15:41 -0000 1.1.1.1
+++ fs/jffs2/os-linux.h 9 Nov 2005 08:36:13 -0000
@@ -93,8 +93,8 @@
  f->usercompr = 0;
 #else
  memset(f, 0, sizeof(*f));
- init_MUTEX_LOCKED(&f->sem);
 #endif
+ init_MUTEX_LOCKED(&f->sem);
 }

 #define jffs2_is_readonly(c) (OFNI_BS_2SFFJ(c)->s_flags & MS_RDONLY)

--- fs/jffs2/super.c 10 Feb 2005 10:15:41 -0000 1.1.1.1
+++ fs/jffs2/super.c 9 Nov 2005 08:36:13 -0000
@@ -51,7 +51,7 @@

  if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
      SLAB_CTOR_CONSTRUCTOR) {
- init_MUTEX_LOCKED(&ei->sem);
  inode_init_once(&ei->vfs_inode);
  }
 }

Thanks,
Keijiro Yano

--------------------------------------
Know more about Breast Cancer
http://pr.mail.yahoo.co.jp/pinkribbon/

             reply	other threads:[~2005-11-11 15:43 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-11-11 15:43 Keijiro Yano [this message]
2005-11-12 13:11 ` jffs2 mutex problem Artem B. Bityutskiy
2005-11-14 12:55   ` Keijiro Yano
2005-11-14 14:55     ` Artem B. Bityutskiy
2005-11-15 13:46       ` Keijiro Yano
2005-11-15 13:54         ` Josh Boyer
2005-11-15 13:57         ` Artem B. Bityutskiy
2005-11-24 16:28         ` Artem B. Bityutskiy
2005-11-25 15:17           ` Keijiro Yano

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='000c01c5e6d6$abf22490$0b01a8c0@YANO' \
    --to=keijiro_yano@yahoo.co.jp \
    --cc=linux-mtd@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox