Bug 232616 - [PowerPC64] Reset from db> does not reboot
Summary: [PowerPC64] Reset from db> does not reboot
Status: New
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: CURRENT
Hardware: powerpc Any
: --- Affects Some People
Assignee: freebsd-bugs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-10-24 14:04 UTC by Sean Bruno
Modified: 2018-10-26 13:46 UTC (History)
8 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Sean Bruno freebsd_committer 2018-10-24 14:04:07 UTC
It appears that the calls to OPAL to reset the box from the debugger prompt don't succeed in rebooting the box.  If you drop to debugger on the Tyan Power8 hosts, and issue a reset command, the machine will not reboot.


fatal user trap:

   exception       = 0x1500 (unknown)
   srr0            = 0x8102a9ae4 (0x400000080e629ae4)
   srr1            = 0x900000000000f032
   current msr     = 0x9000000000009032
   lr              = 0x10000504 (0x400000000e380504)
   curthread       = 0xc00000001cf58000
          pid = 61703, comm = 60519.exe

[ thread pid 61703 tid 101117 ]
Stopped at      0x8102a9ae4
db> [-- sbruno@conserver.nyi.freebsd.org attached -- Wed Oct 24 13:27:45 2018]
bt
Tracing pid 61703 tid 101117 td 0xc00000001cf58000
KDB: reentering
KDB: stack backtrace:
0xe0000000f8c108c0: at .kdb_backtrace+0x5c
0xe0000000f8c109f0: at .kdb_reenter+0x44
0xe0000000f8c10a70: at .trap+0x84
0xe0000000f8c10c30: at .powerpc_interrupt+0x290
0xe0000000f8c10cd0: kernel DSI read trap @ 0x3fffffffffffd1d0 by .db_write_bytes+0x1dc: srr1=0x9000000000001032
            r1=0xe0000000f8c10f80 cr=0x2000f024 xer=0x20000000 ctr=0 r2=0xc000000002d4f5b0 sr=0x40000000
0xe0000000f8c10f80: at -0x4
0xe0000000f8c11010: at .db_trace_thread+0x38
0xe0000000f8c110a0: at .db_command_init+0x910
0xe0000000f8c11130: at .db_skip_to_eol+0x6cc
0xe0000000f8c11260: at .db_command_loop+0x74
0xe0000000f8c112e0: at .db_fetch_ksymtab+0x1fc
0xe0000000f8c11460: at .kdb_trap+0x18c
0xe0000000f8c11530: at .cpu_fetch_syscall_args+0x69c
0xe0000000f8c115c0: at .trap+0xe1c
0xe0000000f8c11780: at .powerpc_interrupt+0x290
0xe0000000f8c11820: user trap 0x1500 by 0x8102a9ae4: srr1=0x900000000000f032
            r1=0x3fffffffffffd1d0 cr=0x24000028 xer=0 ctr=0x10000818 r2=0x10018be0
db> reset
[157832.066391229,5] OPAL: Reboot request...
[157832.066469165,5] RESET: Initiating fast reboot 8...
[157833.068031022,5] RESET: Fast reboot timed out waiting for secondaries to call in
Comment 1 Leandro Lupori 2018-10-24 14:24:45 UTC
Reset works for me on Power8 machines.
I frequently use on Habanero P8 machines, such as ppcports and ppcdevref.

So probably this is not a general issue with reset, but maybe only under certain circumstances or OPAL versions.

IIRC, reset didn't work on pseries, running FreeBSD with QEMU. But so far it worked fine for me on powernv.
Comment 2 Mark Millard 2018-10-24 14:32:47 UTC
I've never had reset at the db> prompt work on a PowerMac G5.
I always end up power cycling the machine. I do not remember
any messages after the command was entered.

I'm not so much after a G5 fix as noting that the one line
description might want to specify a more specific context
for this report. Fixing one context may well not help the
other one.