Re: Strange KASSERTs triggering on (some?) shark in -current

[Nick Hudson <nick.hudson%gmx.co.uk@localhost>]

On 12/04/2021 17:59, Martin Husemann wrote:
> Hey folks,
>
> I have a very strange issue: I recently ressurected one of my sharks
> after its spinning disk died. After unsuccessfully trying various ATA<->SD
> adapters I ended up with a working ATA<->CF adapter, and running some slightly
> older -current on that machine works fine (9.99.71 from August 24 last year).
>
> Trying a -current kernel (with old userland) fails spectacularily with
> random KASSERT like this:
>
> [ 107.6750871] panic: kernel diagnostic assertion "VALID_PSR(tf->tf_spsr)" failed: file "../../../../arch/arm/arm/syscall.c", line 104 spsr: 0x13
> [ 107.6750871] cpu0: Begin traceback...
> [ 107.6750871] 0xf422bdcc: netbsd:db_panic+0xc
> [ 107.6750871] 0xf422bde4: netbsd:vpanic+0x11c
> [ 107.6750871] 0xf422bdfc: netbsd:kern_assert+0x3c
> [ 107.6750871] 0xf422bebc: netbsd:swi_handler+0xc8
> [ 107.6750871] 0xf422bf64: netbsd:swi_entry+0x5c

I change the KASSERT at the beginning of Feb, but I don't think this is
the problem.

Dump the trapframe by dropping into ddb and using 'mach frame'?

Nick