Re: [PATCH v2 1/3] lib: extract 32bit mul/div helpers from nios2 into GENERIC_LIB

[Vineet Gupta <vgupta@kernel.org>]

On 11/6/26 05:07, Simon Schuster wrote:
> Hello Vineet,
>
> On Tue, Nov 03, 2026 at 08:23:23PM -0800, Vineet Gupta wrote:
>>> +/*
>>> + * Software 32-bit multiply for cores built without the hardware multiplier.
>>> + */
>>> +s32 notrace __mulsi3(s32 a, s32 b);
>>> +s32 notrace __divsi3(s32 a, s32 b);
>>> +s32 notrace __udivsi3(s32 a, s32 b);
>>> +s32 notrace __modsi3(s32 a, s32 b);
>>> +s32 notrace __umodsi3(s32 a, s32 b);
>> Shouldn't the 'u' prefixed functions have unsigned int args, result.
>> It seems libgcc/?divmod.c doesn't but ARC does have nint_t for those which
>> is unsigned.
>> Nios it seems was removed from gcc?
> As David has correctly pointed out, nios2 was dropped from gcc with
> release line 14.y being the last one to contain it.
> So at least for nios2, it seems like it has always been the case that
> __udivsi3/__umodsi3 were declared as taking and returning SItypes:
>
> https://github.com/gcc-mirror/gcc/blob/releases/gcc-14/libgcc/config/nios2/lib2-divmod.c#L106
>
> Given that both implementations continue to pass it through to a very
> similarly looking udivmodsi4 as u32, I would assume that this has partly
> relied on 'why should the compiler do any strange upcasts here, it will
> just pass through the registers and be done with it and this is
> implementation-defined-good-enough', so I would not mind rolling a v3
> with u32s for __umodsi3 and __udivsi3. Are these the only two symbols
> you are concerned about, and would these signatures work for you?
>
> u32 notrace __udivsi3(u32 a, u32 b);
> u32 notrace __umodsi3(u32 a, u32 b);

Yep that's what I had in my local version.

Thx,
-Vineet