Re: port-alpha/38335 (kernel freeze on alpha MP system)

["Michael L. Hitch" <mhitch%lightning.msu.montana.edu@localhost>]

The following reply was made to PR port-alpha/38335; it has been noted by GNATS.

From: "Michael L. Hitch" <mhitch%lightning.msu.montana.edu@localhost>
To: Jarle Greipsland <jarle%uninett.no@localhost>
Cc: gnats-bugs%NetBSD.org@localhost, gnats-admin%netbsd.org@localhost
Subject: Re: port-alpha/38335 (kernel freeze on alpha MP system)
Date: Sun, 11 Oct 2009 12:13:57 -0600 (MDT)

 On Thu, 8 Oct 2009, Jarle Greipsland wrote:
 
 > The system then keeps on running.  Though for every "build -j4"
 > job I run I get approximately one
 >  "Whoa!  pool_cache_get returned an in-use entry!"
 > message on the console.  I guess this means we have a
 > papered-over bug now?
 
    Yes - that patch just attempts to work around a duplicate pool cache 
 entry.  I have not been able to replicate this on my ES40, and don't have 
 remote power capability on my CS20 which makes it harder to debug when 
 things can hang hard.
 
    I'm wondering if using IPL_HIGH for the mutex changes anything:
 
 Index: sys/arch/alpha/alpha/pmap.c
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/alpha/alpha/pmap.c,v
 retrieving revision 1.243
 diff -u -p -r1.243 pmap.c
 --- sys/arch/alpha/alpha/pmap.c 4 Oct 2009 17:00:31 -0000       1.243
 +++ sys/arch/alpha/alpha/pmap.c 11 Oct 2009 18:01:59 -0000
 @@ -962,7 +962,7 @@ pmap_bootstrap(paddr_t ptaddr, u_int max
          for (i = 0; i < ALPHA_MAXPROCS; i++) {
                  TAILQ_INIT(&pmap_tlb_shootdown_q[i].pq_head);
                  mutex_init(&pmap_tlb_shootdown_q[i].pq_lock, MUTEX_DEFAULT,
 -                   IPL_SCHED);
 +                   IPL_HIGH);
          }
   #endif
 
 
    If that doesn't change anything, then I'd try increasing the IPL level 
 of the pool_cache entry:
 
 Index: sys/arch/alpha/alpha/pmap.c
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/alpha/alpha/pmap.c,v
 retrieving revision 1.243
 diff -u -p -r1.243 pmap.c
 --- sys/arch/alpha/alpha/pmap.c 4 Oct 2009 17:00:31 -0000       1.243
 +++ sys/arch/alpha/alpha/pmap.c 11 Oct 2009 17:59:33 -0000
 @@ -958,7 +958,7 @@ pmap_bootstrap(paddr_t ptaddr, u_int max
           */
          pool_cache_bootstrap(&pmap_tlb_shootdown_job_cache,
              sizeof(struct pmap_tlb_shootdown_job), CACHE_LINE_SIZE,
 -            0, PR_LARGECACHE, "pmaptlb", NULL, IPL_VM, NULL, NULL, NULL);
 +            0, PR_LARGECACHE, "pmaptlb", NULL, IPL_SCHED, NULL, NULL, NULL);
          for (i = 0; i < ALPHA_MAXPROCS; i++) {
                  TAILQ_INIT(&pmap_tlb_shootdown_q[i].pq_head);
                  mutex_init(&pmap_tlb_shootdown_q[i].pq_lock, MUTEX_DEFAULT,
 
    And:
 
 Index: sys/arch/alpha/alpha/pmap.c
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/alpha/alpha/pmap.c,v
 retrieving revision 1.243
 diff -u -p -r1.243 pmap.c
 --- sys/arch/alpha/alpha/pmap.c 4 Oct 2009 17:00:31 -0000       1.243
 +++ sys/arch/alpha/alpha/pmap.c 11 Oct 2009 17:59:56 -0000
 @@ -958,7 +958,7 @@ pmap_bootstrap(paddr_t ptaddr, u_int max
           */
          pool_cache_bootstrap(&pmap_tlb_shootdown_job_cache,
              sizeof(struct pmap_tlb_shootdown_job), CACHE_LINE_SIZE,
 -            0, PR_LARGECACHE, "pmaptlb", NULL, IPL_VM, NULL, NULL, NULL);
 +            0, PR_LARGECACHE, "pmaptlb", NULL, IPL_HIGH, NULL, NULL, NULL);
          for (i = 0; i < ALPHA_MAXPROCS; i++) {
                  TAILQ_INIT(&pmap_tlb_shootdown_q[i].pq_head);
                  mutex_init(&pmap_tlb_shootdown_q[i].pq_lock, MUTEX_DEFAULT,
 
 
 --
 Michael L. Hitch                       mhitch%montana.edu@localhost
 Computer Consultant
 Information Technology Center
 Montana State University       Bozeman, MT     USA