kern/56730: Lock error panic in entropy_enter_early()

[gson%gson.org@localhost (Andreas Gustafsson)]

>Number:         56730
>Category:       kern
>Synopsis:       Lock error panic in entropy_enter_early()
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Feb 24 11:25:01 +0000 2022
>Originator:     Andreas Gustafsson
>Release:        NetBSD-current, source date 2022.02.23.12.17.06
>Organization:
  
>Environment:
System: NetBSD
Architecture: mips
Machine: pmax
>Description:

The TNF pmax testbed recently paniced during boot:

[   1.0000000] entropy: WARNING: extracting entropy too early
[   1.0078165] Mutex error: mutex_vector_enter,479: locking against myself

[   1.0078165] lock address : 0x00000000805a5380
[   1.0078165] current cpu  :                  0
[   1.0078165] current lwp  : 0x000000008059c600
[   1.0078165] owner field  : 000000000000000000 wait/spin:                0/1

[   1.0078165] panic: lock error: Mutex: mutex_vector_enter,479: locking against myself: lock 0x805a5380 cpu 0 lwp 0x8059c600
[   1.0078165] cpu0: Begin traceback...
[   1.0078165] pid -2140824880 not found
[   1.0078165] cpu0: End traceback...
[   1.0078165] kernel: breakpoint trap
Stopped in pid 0.0 (system) at  netbsd:cpu_Debugger+0x4:        jr      ra
                bdslot: nop
0x80659af8: cpu_Debugger+0x4 (0,0x8000,0x805d34fc,0x8058ed40) ra 0x80353a0c sz 0
0x80659af8: vpanic+0x1a4 (0,0x8000,0x805d34fc,0x8058ed40) ra 0x80353a88 sz 48
0x80659b28: panic+0x24 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x80346a50 sz 32
0x80659b48: lockdebug_abort+0xe4 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x802f6b14 sz 56
0x80659b80: mutex_abort+0x48 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x802d4760 sz 32
0x80659ba0: entropy_enter_early+0x110 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x802d5204 sz 40
0x80659bc8: rnd_add_data_1+0x4f0 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x802d71f0 sz 88
0x80659c20: rnd_add_data+0x120 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x802d7400 sz 96
0x80659c80: rnd_add_uint32+0x20 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0x802c980c sz 32
0x80659ca0: clockrnd_sample+0x40 (0,0x804fcc94,0x804bc5c4,0x1df) ra 0 sz 0
User-level: pid 0.0
at          1
v0          1
v1          8058ed40    cpu_info_store

The full console log is at

  http://releng.netbsd.org/b5reports/pmax/2022/2022.02.23.12.17.06/test.log

Grepping the logs shows two earlier instances:

  http://releng.netbsd.org/b5reports/pmax/2021/2021.06.04.06.28.42/test.log
  http://releng.netbsd.org/b5reports/pmax/2021/2021.10.04.07.04.39/test.log

This hpcmips panic also looks like it could be the same issue:

  http://releng.netbsd.org/b5reports/hpcmips/2021/2021.01.17.00.13.49/test.log

And it has also happened once on sparc, so it's not mips specific:

  http://releng.netbsd.org/b5reports/sparc/2021/2021.04.20.10.39.58/test.log

>How-To-Repeat:

>Fix: