>Number:         12427
>Category:       port-sparc
>Synopsis:       pagedaemon: deadlock detected!
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    port-sparc-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Mar 17 06:24:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     Brad Spencer
>Release:        1.5S updated via cvs 3/15/2001 +- day
>Organization:
	Home
>Environment:
	
System: NetBSD frodo.eldar.org 1.5S NetBSD 1.5S (THE_SHIRE_4M_NOMP) #0: Thu Mar 15 20:19:05 EST 2001     brad@frodo.eldar.org:/usr/src/sys/arch/sparc/compile/THE_SHIRE_4M_NOMP sparc


>Description:

I have a Sparc 10 with dual 100MHZ Ross Hypersparc processors and 32MB
of ram.  The machine is running completely diskless.  While running
the rc5des client, and doing the final linking of two NetBSD sparc
kernels, all in a screen session, the machine would start to spew the
following message:

pagedaemon: deadlock detected!

over and over again.  It was possible to break into DDB and do a 'ps'.

Stopped in pid 243 (rc5des) at  cpu_Debugger+0x4:       jmpl            [%o7 + 0x8], %g0
db> ps
 PID             PPID       PGRP        UID S   FLAGS          COMMAND    WAIT
 349                1        349          0 3  0x4082            getty   ttyin
 337              335        267       3000 3  0x4006               ld uvn_fp1
 336              334        273       3000 3  0x4006               ld uvn_fp1
 335              267        267       3000 3  0x4082               sh    wait
 334              273        273       3000 3  0x4082               sh    wait
 273              262        273       3000 3  0x4082             make    wait
 267              257        267       3000 3  0x4082             make    wait
 262              230        262       3000 3  0x4082             tcsh   pause
 257              230        257       3000 3  0x4082             tcsh   pause
 256              225        256       3000 3  0x4106     screen-3.9.5 flt_nor
>How-To-Repeat:

Run rc5des, and two links of a netbsd kernel in screen.

I suspect that this bug can be prompted to occur on demand.

>Fix:

Don't know, but I may be able to provide more information, if
requested to do so.
>Release-Note:
>Audit-Trail:
>Unformatted:
 >243              235        243       3000 7  0x4006           rc5des
  235              230        235       3000 3  0x4082             tcsh   pause
  230                1        230       3000 3   0x180     screen-3.9.5  select
  225              207        225       3000 3  0x4082             tcsh   pause
  207              206        206       3000 2  0x4180            xterm
  206              205        206       3000 3  0x4080             tcsh   pause
  205              199        199          0 3  0x4180             rshd  select
  202                1        202          0 2   0x480             cron
  199                1        199          0 3    0x80            inetd  select
  192                1        192          0 3    0x80             sshd  select
  173                1        173          0 3    0x84             ntpd   pause
  163                1        163          0 3     0x4            rwhod uvn_fp1
  110                1        110          0 3     0x4           ypbind flt_nor
  105                1        105          0 3    0x80          rpcbind  select
  93                 1         93          0 3     0x4          syslogd flt_nor
  9                  0          0          0 3 0x20204         aiodoned aiodone
  8                  0          0          0 3 0x20204          ioflush  syncer
  7                  0          0          0 3 0x20204           reaper  reaper
  6                  0          0          0 3 0x20204       pagedaemon km_getw
  5                  0          0          0 3 0x20284            nfsio  nfsidl
  4                  0          0          0 3 0x20284            nfsio  nfsidl
  3                  0          0          0 3 0x20284            nfsio  nfsidl
  2                  0          0          0 3 0x20284            nfsio  nfsidl
  1                  0          1          0 3  0x4080             init    wait
  0                 -1          0          0 3 0x20204          swapper schedpw
 
 I am not sure that a diskless sparc can take a crash dump???