Re: [PATCH v2] m68k: save extra registers on more syscall entry points

[Michael Schmitz <schmitzmic@gmail.com>]

Hi Linus,

Am 19.06.2021 um 11:38 schrieb Linus Torvalds:
> On Fri, Jun 18, 2021 at 3:34 PM Michael Schmitz <schmitzmic@gmail.com> wrote:
>>
>> Is your patch to copy_thread() to add the extra stack frame still needed?
>
> So it's been a long time since I did any m68k assembly, but I think
> the m68k patch for the PF_IO_WORKER thread case should look something
> like the attached.
>
> Note: my only m68k work was ever on the 68008, and used the Motorola
> syntax, not the odd Sun assembler syntax, so my m68k asm skills really
> aren't good.
>
> Put another way: I'd be surprised if the attached patch actually
> works, but I think it's fairly close. I tried to add comments to
> explain the code at least a bit.

That went well:

*** FORMAT ERROR ***   FORMAT=0
Current process id is 1
BAD KERNEL TRAP: 00000000
Modules linked in:
PC: [<00002af0>] resume_userspace+0x14/0x16
SR: 2204  SP: (ptrval)  a2: 00000000
d0: 00000000    d1: 00000000    d2: 00000000    d3: 00000000
d4: 00000000    d5: 00000000    a0: 00000000    a1: 00000000
Process init (pid: 1, task=(ptrval))
Frame format=0
Stack from 0081bffc:
         19bc0000
Call Trace:
Code: 1029 0007 660c 4cdf 073e 201f 588f dfdf <4e73> 254f 03ec e308 660a 
487a ffe0 60ff 002a f6ba 598f 48e7 031e 486f 001c 61ff
Disabling lock debugging due to kernel taint
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b

Looks like the zeroed frame was restored where we'd have expected an 
actual save frame?

I'll next try and apply your solution to IO worker threads only ...

Cheers,

	Michael

>
> Hmm?
>
>          Linus
>