On Tue, Sep 09, 2025 at 12:24:06AM +0200, Riccardo Mottola wrote:
> Hi all!
> 
> not new to OpenBSD, not new to sparc, but new to it on sparc64.
> 
> My Netra T1 was running Solaris 8, was used less and less until it stopped
> booting due to a faulty RAM module. Found out which one, bought one used...
> and it runs again, but with no more real use: time to put OpenBSD on it! I
> already have NetBSD/sparc64 (and also sparc32, but OpenBSD/sparc32 is sadly
> no more...) so a good BigEndian testing system comes, much more capable than
> the 32bit ones!
> 
> essentially before Holidays the system booted, worked, I installed packages,
> did some first compilation attempts, no big issues, run for a couple of
> hours a couple of days before "summer pause".
> 
> Today I needed to test some updated code... boot, git update && make and
> after a while it locks.
> Out of chance, I still had the serial console attached with minicom on my
> workstation running.
> 
> Here the output:
> 
> login: splassert: assertwaitok: want 0 have 1
> panic: mtx 0x1c1e6d0: locking against myself
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> panic: kernel diagnostic assertion "p->p_wchan == NULL" failed: file
> "/usr/src/
> sys/kern/kern_sched.c", line 370
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> panic: sleep: not SONPROC but 2
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> panic: sleep: not SONPROC but 2
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> panic: sleep: not SONPROC but 2
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> panic: sleep: not SONPROC but 2
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> panic: sleep: not SONPROC but 2
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> --db_more--
> LOM event: +12h26m58s host FAULT: watchdog triggered
> 
> LOM event: +12h26m58s Fault LED ON
> 
> 
> Apart that the orange led was really on...
> 
> I try to reboot it and:
> WARNING: CHECK AND RESET THE DATE!
> panic: kernel data fault: pc=1585a24 addr=c6c000
> Stopped at      db_enter+0x8:   nop
>     TID    PID    UID     PRFLAGS     PFLAGS  CPU  COMMAND
> *247958      1      0         0x2          0    0  init
> data_access_fault(4002efbf690, 30, 1585a24, c6c000, c6cfa0, 80080f) at
> data_acc
> ess_fault+0x2f0
> Ldatafault_internal(1c7cf50, 0, 19663b8, 400042f1a44, 400042f1a44, 3b9ac800)
> at
>  Ldatafault_internal+0xcc
> pool_do_put(1c7cf50, 400042b9b90, 1, 1c84ed0, 192850c, 400042f1a40) at
> pool_do_
> put+0x1fc
> pool_put(1c7cf50, 400042b9b90, 0, 0, 0, 0) at pool_put+0x3c
> process_zap(400042e7270, 4002effc8d0, 198b8e8, 0, 0, 0) at process_zap+0xac
> dowait6(4000000000000000, 0, 0, 4000040000000000, 2, 40000) at dowait6+0x588
> sys_wait4(16, 4002efbff40, 4002efbfe00, 1887710, 124dc60, 1) at
> sys_wait4+0xa4
> syscall(4002efbfee0, 19bdf40, 4002efbff40, 1794008, 1c00, 4002efc3858) at
> sysca
> ll+0x3ac
> syscall_setup(ffffffffffffffff, ffffffcafdafaa74, 2, 0, 0, 1ffd9b8c0f8) at
> sysc
> all_setup+0x124
> https://www.openbsd.org/ddb.html describes the minimum info required in bug
> reports.  Insufficient info makes it difficult to find and fix bugs.
> ddb>
> 
> 
> I wonder if the new memory module is faulty again in a more subtle way? or
> perhaps another one is failing in a way that startup is not rejecting the
> DIMM immediatly?

Hard to judge, it could be a hardware issue. If all you see is random
crashes with little pattern then I would check that the memory is really
OK (including the sockets).
The netra T1 is a single core USIIi so I doubt this is any of the ghosts
we're hunting on fast SMP boxes like the M10-1.

-- 
:wq Claudio

Reply via email to