Re: kern/38707: scheduler related deadlock during build.sh

[Andrew Doran <andrew%hairylemon.org@localhost>]

The following reply was made to PR kern/38707; it has been noted by GNATS.

From: Andrew Doran <andrew%hairylemon.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: kern/38707: scheduler related deadlock during build.sh
Date: Fri, 23 May 2008 16:13:35 +0100

 Will close the following as a dup.
 
 ----- Forwarded message from martin%duskware.de@localhost -----
 
 Subject: kern/38730: recursive lock in scheduler
 From: martin%duskware.de@localhost
 Date: Thu, 22 May 2008 19:55:00 +0000 (UTC)
 To: 
kern-bug-people%netbsd.org@localhost,gnats-admin%netbsd.org@localhost,netbsd-bugs%netbsd.org@localhost
 Cc: 
 
 >Number:         38730
 >Category:       kern
 >Synopsis:       recursive lock in scheduler
 >Confidential:   no
 >Severity:       critical
 >Priority:       high
 >Responsible:    kern-bug-people
 >State:          open
 >Class:          sw-bug
 >Submitter-Id:   net
 >Arrival-Date:   Thu May 22 19:55:00 +0000 2008
 >Originator:     Martin Husemann
 >Release:        NetBSD 4.99.63
 >Organization:
 The NetBSD Foundation, Inc.
 >Environment:
 System: NetBSD sunny-weather.duskware.de 4.99.63 NetBSD 4.99.63 (SUNNY.MP) 
#48: Thu May 22 17:50:52 CEST 2008 
martin%sunny-weather.duskware.de@localhost:/usr/src/sys/arch/sparc64/compile/SUNNY.MP
 sparc64
 Architecture: sparc64
 Machine: sparc64
 >Description:
 
 Mutex error: lockdebug_wantlock: locking against myself
                                                        
 lock address : 0x000000000d047e80 type     :               spin
 shared holds :                  0 exclusive:                  1
 shares wanted:                  0 exclusive:                  1
 current cpu  :                  1 last held:                  1
 current lwp  : 0x000000000f89a120 last held: 0x000000000f89a120
 last locked  : 0x0000000001160bf8 unlocked : 0x0000000001160610
 initialized  : 0x000000000114d744                              
 owner field  : 0x00ff0a0000000000 wait/spin:                0/1
                                                                
 panic: LOCKDEBUG
 Stopped in pid 1022.1 (cc1) at  netbsd:cpu_Debugger+0x4:        nop
 db{1}> bt                                                          
 lockdebug_abort1(d803900, a, 127c598, 12c2480, 1, 20) at 
netbsd:lockdebug_abort1
 +0x7c                                                                          
 mutex_vector_enter(d047e80, 1a, d047d80, 12c1a18, de98000, 1814000) at 
netbsd:mu
 tex_vector_enter+0x230                                                         
 sched_nextlwp(d812040, 1, 127c400, f89a120, de98000, f89a170) at 
netbsd:sched_ne
 xtlwp+0x288                                                                    
 mi_switch(f89a120, 12c3d70, 14f, 12c1c48, deb40c8, ffffffffffffffff) at 
netbsd:m
 i_switch+0x40c                                                                 
 preempt(f89a120, 17e798, 31dedb1234, 8000000000000000, 31dedb1234, 4ce800) at 
ne
 tbsd:preempt+0xa0                                                              
 trap(f795ed0, fffffffffffffffe, 17e798, 44820092, 121a9a0, ffffffffffff9d68) 
at n
 etbsd:trap+0x8b4
 ?(1, 80, 40c12600, c, 20, 5fdf30) at 0x10090b8
 db{1}> mach cpu 0                             
 db{0}> bt        
 mi_switch(f9cac80, 12c3d70, 14f, 12c1c48, 1455980, 42c54a10) at 
netbsd:mi_switch
 +0x308                                                                         
 preempt(f9cac80, 3d566c, fffffffffffffff8, 0, 42c54a10, 0) at 
netbsd:preempt+0xa
 0                                                                              
 trap(f7bfed0, fffffffffffffffe, 3d566c, 820092, 121aa00, 0) at 
netbsd:trap+0x8b4
                                                                                
 ?(42c00980, 42b93800, 18, 0, 30, 2) at 0x10090b8
 db{0}> x/i 0x000000000114d744                   
 netbsd:mutex_obj_alloc+0x24:    call            netbsd:mutex_init
 db{0}> x/i 0x0000000001160bf8                                    
 netbsd:mi_switch+0x758: call            netbsd:mutex_vector_enter
 db{0}> x/i 0x0000000001160610                                    
 netbsd:mi_switch+0x170: call            netbsd:mutex_vector_exit
 0x114d744 is in mutex_obj_alloc (../../../../kern/kern_mutex.c:970).
 965     mutex_obj_alloc(kmutex_type_t type, int ipl)
 966     {
 967             struct kmutexobj *mo;
 968     
 969             mo = pool_cache_get(mutex_obj_cache, PR_WAITOK);
 970             mutex_init(&mo->mo_lock, type, ipl);
 971             mo->mo_refcnt = 1;
 972     
 973             return (kmutex_t *)mo;
 974     }
 0x1160bf8 is in mi_switch (../../../../sys/lwp.h:407).
 402             KASSERT(ci1 != ci2);
 403             if (ci1 < ci2) {
 404                     mutex_spin_enter(spc1->spc_mutex);
 405                     mutex_spin_enter(spc2->spc_mutex);
 406             } else {
 407                     mutex_spin_enter(spc2->spc_mutex);
 408                     mutex_spin_enter(spc1->spc_mutex);
 409             }
 410     }
 0x1160610 is in mi_switch (../../../../sys/lwp.h:344).
 339      * Unlock an LWP. XXXLKM
 340      */
 341     static inline void
 342     lwp_unlock(lwp_t *l)
 343     {
 344             mutex_spin_exit(l->l_mutex);
 345     }
 346     
 347     static inline void
 348     lwp_changepri(lwp_t *l, pri_t pri)
 
 
 >How-To-Repeat:
 
 No idea - I did a "make -j 4" in a kernel compile directory over NFS
 (expecting another bug to show up)
 
 >Fix:
 n/a
 
 
 ----- End forwarded message -----