[PATCH 5/8] math: Fix UB on cbrtf

[Carlos O'Donell]

On 4/25/25 4:54 PM, Adhemerval Zanella wrote:
> The left shift overflows for 'int64_t', use unsigned instead.  It syncs
> with CORE-MATH commit f7c7408d1749ec2859ea249495af699359ae559b.
> 
> Checked on aarch64-linux-gnu, x86_64-linux-gnu, and i686-linux-gnu.

LGTM.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>

> ---
>   sysdeps/ieee754/flt-32/s_cbrtf.c | 6 +++---
>   1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/sysdeps/ieee754/flt-32/s_cbrtf.c b/sysdeps/ieee754/flt-32/s_cbrtf.c
> index 5a7a9a952d..df9e888c1f 100644
> --- a/sysdeps/ieee754/flt-32/s_cbrtf.c
> +++ b/sysdeps/ieee754/flt-32/s_cbrtf.c
> @@ -3,7 +3,7 @@
>   Copyright (c) 2023, 2024 Alexei Sibidanov.
>   
>   The original version of this file was copied from the CORE-MATH
> -project (file src/binary32/cbrt/cbrtf.c, revision bc385c2).
> +project (file src/binary32/cbrt/cbrtf.c, revision f7c7408d).

OK.

>   
>   Permission is hereby granted, free of charge, to any person obtaining a copy
>   of this software and associated documentation files (the "Software"), to deal
> @@ -61,8 +61,8 @@ __cbrtf (float x)
>     e += 899;
>     uint32_t et = e / 3, it = e % 3;
>     uint64_t isc = escale[it].u;
> -  isc += (int64_t) (et - 342) << 52;
> -  isc |= (int64_t) sgn << 63;
> +  isc += (uint64_t) (et - 342) << 52;
> +  isc |= (uint64_t) sgn << 63;
>     double cvt2 = asdouble (isc);
>     static const double c[] =
>       {

OK.

commit f7c7408d1749ec2859ea249495af699359ae559b
Author: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Date:   Thu Mar 13 14:37:02 2025 +0100

     [cbrtf] fixed undefined behavior


-- 
Cheers,
Carlos.