[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: kernel: invalidate: busy buffer



Ralf Hildebrandt wrote:
> 
> Am 01.12.2001 um 12:30:05 +0100 schrieb Wiktor Wodecki folgendes:
> 
> > this was discussed a couple of times, look in the archives
> 
> a) It was discussed ONCE in the last 6 months -- at least according to the
>    Subject: lines.
> 
> b) if it was discussed more often, I can't find it, since the archive has no
>    search facility (hint, hint).
> 
> c) The answer doesn't apply to my case:
>    I never umount something, especially not at Dec  1 00:42:05 (weekend, in
>    the middle of the night!)
> 
> d) I don't see this at bootup time, but in mid-flight.
> 

There are numerous ways in which this can happen, and they
are all on the "undoing something" path.  Unmounting a cdrom,
unmounting a filesystem in general.  `losetup -d' will also
do it.  There must be some sort of irregular activity which 
is causing this.  But it seems to be harmless.

If you're curious, you could rebuild your kernel with this patch:

--- linux-2.4.17-pre2/fs/buffer.c	Fri Nov 30 14:32:14 2001
+++ linux-akpm/fs/buffer.c	Sat Dec  1 13:54:33 2001
@@ -692,7 +692,7 @@ void invalidate_bdev(struct block_device
 					remove_inode_queue(bh);
 				}
 			} else
-				printk("invalidate: busy buffer\n");
+				printk("invalidate: busy buffer(%s)\n", current->comm);
 
 			write_unlock(&hash_table_lock);
 			if (slept)

which will tell you which program is causing it.





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]