Re: [PATCH] rs6000: Rework option -mpowerpc64 handling [PR106680]

[Iain Sandoe <iain@sandoe.co.uk>]

Hi Folks,

> On 28 Sep 2022, at 22:30, Segher Boessenkool <segher@kernel.crashing.org> wrote:

> On Wed, Sep 28, 2022 at 05:18:47PM +0100, Iain Sandoe wrote:
>>> On 28 Sep 2022, at 07:37, Iain Sandoe <iain@sandoe.co.uk> wrote:
>>>> On 28 Sep 2022, at 06:30, Kewen.Lin via Gcc-patches <gcc-patches@gcc.gnu.org> wrote:

>> powerpc-apple-darwin, is perhaps somewhat unusual in that it is nominally a 32b kernel, but the OS supports 64b processes on suitable hardware
> 
> Just like Linux was before there was powerpc64-linux.  I think it should
> still work even?
> 
>> (and the OS does preserve the upper bits of 64b regs in the context).
> 
> That works on Linux as well.  What still does not work is user-mode
> context switches in 32-bit processes (so setjmp and getcontext stuff).

AFAIU the Darwin impl. it is the same - the user context only contains 32b
register images.

Since one can only use the feature between function calls, I guess that the
setjmp/longjmp stuff is not so critical on Darwin***. However, even being able
to use 64b insns between calls could give a massive win in allowing, for
example, lock-free 64b atomics.

Sometime, I need to spend some time with this and make a set of ppc970
library slices (the dynamic loader should pick the right arch for the resident
cpu).

>> I will try to take a look at this this evening see if I can throw
>> any more light on it.
> 
> Thanks!

adding —with-tune=G5 to the configure line .. the cross-build then succeeded
(at "-O1 -g" as I was building to debug) - maybe that will provide a clue, but I’m
out of time for today.

Iain.

===
*** revisiting this topic, did make me wonder about non-call exceptions tho, not
sure if they were considered in the original recipes.