inotify bug?

inotify bug?

[Lee Revell]

I got this in dmesg with 2.6.13-rc7-rt3 the other day.  Maybe it's
useful.  There were no visible effects.

idr_remove called for id=1024 which is not allocated.
 [sub_remove+157/272] sub_remove+0x9d/0x110 (8)
 [__wake_up+44/80] __wake_up+0x2c/0x50 (28) 
 [idr_remove+24/128] idr_remove+0x18/0x80 (28)
 [remove_watch_no_event+77/208] remove_watch_no_event+0x4d/0xd0 (12)
 [inotify_inode_is_dead+88/128] inotify_inode_is_dead+0x58/0x80 (12)
 [prune_dcache+222/480] prune_dcache+0xde/0x1e0 (28)
 [shrink_dcache_memory+56/64] shrink_dcache_memory+0x38/0x40 (28)
 [shrink_slab+277/416] shrink_slab+0x115/0x1a0 (4)
 [balance_pgdat+686/1024] balance_pgdat+0x2ae/0x400 (48)
 [kswapd+168/240] kswapd+0xa8/0xf0 (96)
 [autoremove_wake_function+0/48] autoremove_wake_function+0x0/0x30 (12)
 [kswapd+0/240] kswapd+0x0/0xf0 (16)
 [kernel_thread_helper+5/24] kernel_thread_helper+0x5/0x18 (16)

Lee

Re: inotify bug?

[Chris Wright]

* Lee Revell (rlrevell@joe-job.com) wrote:
> I got this in dmesg with 2.6.13-rc7-rt3 the other day.  Maybe it's
> useful.  There were no visible effects.
> 
> idr_remove called for id=1024 which is not allocated.

This should already be fixed in 2.6.13.

http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=7c657f2f25d50c602df9291bc6242b98fc090759

thanks,
-chris

Re: inotify bug?

[Lee Revell]

On Thu, 2005-09-15 at 14:48 -0700, Chris Wright wrote:
> * Lee Revell (rlrevell@joe-job.com) wrote:
> > I got this in dmesg with 2.6.13-rc7-rt3 the other day.  Maybe it's
> > useful.  There were no visible effects.
> > 
> > idr_remove called for id=1024 which is not allocated.
> 
> This should already be fixed in 2.6.13.
> 
> http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=7c657f2f25d50c602df9291bc6242b98fc090759
> 

Thanks, nice to see that my suspicion of an off by one error was right
(the 1024 was a dead giveaway).

BTW the rate of kernel development is just insane these days, even
compared to a year ago.  It's quite encouraging.  At this rate we're
going to leave the "competition" in the dust.

Lee