Re: kern/46734: pool not drained

[Martin Husemann <martin%duskware.de@localhost>]

The following reply was made to PR kern/46734; it has been noted by GNATS.

From: Martin Husemann <martin%duskware.de@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/46734: pool not drained
Date: Tue, 24 Jul 2012 10:01:58 +0200

 Runing the tests in /usr/tests/sbin/resize_ffs is enough to trigger this
 problem (at least on sparc64 SMP, might be 8k page size and recent poll
 changes uncovered an old bug).
 
 Given this backtrace:
 pool_destroy(73c3200, 0, 1, 6, 175c310, 17985e8) at netbsd:pool_destroy+0x320
 pool_cache_destroy(18ade40, 73c3200, 0, 175c2d8, 175c310, 1797610) at 
netbsd:poo
 l_cache_destroy+0x10                                                           
 puffs_vfsop_done(1c4b800, 0, 4f72, 6744e10, 0, 8840dc08) at 
puffs:puffs_vfsop_do
 ne+0x4                                                                         
 vfs_detach(1c4b660, 0, 675e000, 186e9d0, 1794458, 1c14000) at 
netbsd:vfs_detach+
 0x80                                                                           
 module_do_unload(18a4000, 0, 675e000, 6, 0, 1c14000) at 
netbsd:module_do_unload+
 0x74                                                                           
 
 I did:
 db{1}> sh pool 73c3200
 POOL puffprkl: size 112, align 16, ioff 0, roflags 0x00000040
         alloc 0x1875d10                                      
         minitems 0, minpages 0, maxpages 4294967295, npages 16
         itemsperpage 72, nitems 1151, nout 1, hardlimit 4294967295
         nget 1135, nfail 0, nput 1134                             
         npagealloc 16, npagefree 0, hiwat 16, nidle 15
 
 Which code is supposed to drain this pool before module unload?
 
 Martin