The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: Andy Tai <lichengtai@yahoo.com>
To: atai@atai.org, Matthew Wilcox <willy@debian.org>
Cc: linux-kernel@vger.kernel.org
Subject: Re: file locking (fcntl) bug in 2.4.19
Date: Fri, 30 Aug 2002 21:13:18 -0700 (PDT)	[thread overview]
Message-ID: <20020831041318.26193.qmail@web10501.mail.yahoo.com> (raw)
In-Reply-To: <20020830205006.80088.qmail@web10501.mail.yahoo.com>

Hi, I just figured out the problem is not with the
locking code in the kernel, which works with your
patch, but in the user space application setting the
file permission in such a way as to use mandatory
locks, that surely fails on a NFS volume...

The user space application in this case is Samba
2.2.5.

Thanks for your help.

Andy

--- Andy Tai <lichengtai@yahoo.com> wrote:
> Mr. Wilcox, thanks for the reply.  I checked the
> source of the stock 2.4.19 kernel and it seems to
> have
> part of the patch incorporated.  I checked to make
> sure the patch is fully applied and recompiled the
> kernel, and the bug still shows up (less frequently,
> but the test program lockbug.c can still trigger the
> bug). It seems the problem is not totally gone yet.
> 
> Thanks for any help.
> 
> Andy  
> --- Matthew Wilcox <willy@debian.org> wrote:
> > 
> > Yep, 2.4 & 2.5 are broken.  Here's a patch which
> > both marcelo and alan
> > refuse to apply, but fixes the problem.
> > 
> > diff -urNX dontdiff linux-2418/fs/locks.c
> > linux-2418-acct/fs/locks.c
> > --- linux-2418/fs/locks.c	Thu Oct 11 08:52:18 2001
> > +++ linux-2418-acct/fs/locks.c	Mon Jul  1 16:23:36
> > 2002
> > @@ -134,15 +134,9 @@
> >  static kmem_cache_t *filelock_cache;
> >  
> >  /* Allocate an empty lock structure. */
> > -static struct file_lock *locks_alloc_lock(int
> > account)
> > +static struct file_lock *locks_alloc_lock(void)
> >  {
> > -	struct file_lock *fl;
> > -	if (account && current->locks >=
> > current->rlim[RLIMIT_LOCKS].rlim_cur)
> > -		return NULL;
> > -	fl = kmem_cache_alloc(filelock_cache,
> > SLAB_KERNEL);
> > -	if (fl)
> > -		current->locks++;
> > -	return fl;
> > +	return kmem_cache_alloc(filelock_cache,
> > SLAB_KERNEL);
> >  }
> >  
> >  /* Free a lock which is not in use. */
> > @@ -152,7 +146,6 @@
> >  		BUG();
> >  		return;
> >  	}
> > -	current->locks--;
> >  	if (waitqueue_active(&fl->fl_wait))
> >  		panic("Attempting to free lock with active wait
> > queue");
> >  
> > @@ -219,7 +212,7 @@
> >  /* Fill in a file_lock structure with an
> > appropriate FLOCK lock. */
> >  static struct file_lock *flock_make_lock(struct
> > file *filp, unsigned int type)
> >  {
> > -	struct file_lock *fl = locks_alloc_lock(1);
> > +	struct file_lock *fl = locks_alloc_lock();
> >  	if (fl == NULL)
> >  		return NULL;
> >  
> > @@ -348,7 +341,7 @@
> >  /* Allocate a file_lock initialised to this type
> of
> > lease */
> >  static int lease_alloc(struct file *filp, int
> type,
> > struct file_lock **flp)
> >  {
> > -	struct file_lock *fl = locks_alloc_lock(1);
> > +	struct file_lock *fl = locks_alloc_lock();
> >  	if (fl == NULL)
> >  		return -ENOMEM;
> >  
> > @@ -712,7 +705,7 @@
> >  			 size_t count)
> >  {
> >  	struct file_lock *fl;
> > -	struct file_lock *new_fl = locks_alloc_lock(0);
> > +	struct file_lock *new_fl = locks_alloc_lock();
> >  	int error;
> >  
> >  	if (new_fl == NULL)
> > @@ -872,8 +865,8 @@
> >  	 * We may need two file_lock structures for this
> > operation,
> >  	 * so we get them in advance to avoid races.
> >  	 */
> > -	new_fl = locks_alloc_lock(0);
> > -	new_fl2 = locks_alloc_lock(0);
> > +	new_fl = locks_alloc_lock();
> > +	new_fl2 = locks_alloc_lock();
> >  	error = -ENOLCK; /* "no luck" */
> >  	if (!(new_fl && new_fl2))
> >  		goto out_nolock;
> > @@ -1426,7 +1419,7 @@
> >  int fcntl_setlk(unsigned int fd, unsigned int
> cmd,
> > struct flock *l)
> >  {
> >  	struct file *filp;
> > -	struct file_lock *file_lock =
> locks_alloc_lock(0);
> > +	struct file_lock *file_lock =
> locks_alloc_lock();
> >  	struct flock flock;
> >  	struct inode *inode;
> >  	int error;
> > @@ -1582,7 +1575,7 @@
> >  int fcntl_setlk64(unsigned int fd, unsigned int
> > cmd, struct flock64 *l)
> >  {
> >  	struct file *filp;
> > -	struct file_lock *file_lock =
> locks_alloc_lock(0);
> > +	struct file_lock *file_lock =
> locks_alloc_lock();
> >  	struct flock64 flock;
> >  	struct inode *inode;
> >  	int error;
> > 
> > -- 
> > Revolutions do not require corporate support.
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Finance - Get real-time stock quotes
> http://finance.yahoo.com


__________________________________________________
Do You Yahoo!?
Yahoo! Finance - Get real-time stock quotes
http://finance.yahoo.com

  reply	other threads:[~2002-08-31  4:08 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-08-30 12:25 file locking (fcntl) bug in 2.4.19 Matthew Wilcox
2002-08-30 20:50 ` Andy Tai
2002-08-31  4:13   ` Andy Tai [this message]
  -- strict thread matches above, loose matches on Subject: below --
2002-08-30  3:52 Andy Tai

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=20020831041318.26193.qmail@web10501.mail.yahoo.com \
    --to=lichengtai@yahoo.com \
    --cc=atai@atai.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=willy@debian.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