Re: [PATCH] lib: count_zeros: fix 32/64-bit inconsistency in count_trailing_zeros()

[Andy Shevchenko]

On Thu, Mar 12, 2026 at 07:08:16PM -0400, Yury Norov wrote:
> Based on 'sizeof(x) == 4' condition, in 32-bit case the function is wired
> to ffs(), while in 64-bit case to __ffs(). The difference is substantial:
> ffs(x) == __ffs(x) + 1. Also, ffs(0) == 0, while __ffs(0) is undefined.
> 
> The 32-bit behaviour is inconsistent with the function description, so it
> needs to get fixed.
> 
> There are 9 individual users for the function in 6 different subsystems.
> Some arches and drivers are 64-bit only:
>  - arch/loongarch/kvm/intc/eiointc.c;
>  - drivers/hv/mshv_vtl_main.c;
>  - kernel/liveupdate/kexec_handover.c;
> 
> The others are:
>  - ib_umem_find_best_pgsz(): as per comment, __ffs() should be correct;
>  - rzv2m_csi_reg_write_bit(): ARCH_RENESAS only, unclear;
>  - lz77_match_len(): CIFS_COMPRESSION only, unclear, experimental;
> 
> None of them explicitly tweak their code for a word length, or x == 0.
> 
> Requesting comments from the corresponding maintainers on how to proceed
> with this.
> 
> The attached patch gets rid of 32-bit explicit support, so that both
> 32- and 64-bit versions rely on __ffs().

> CC: "K. Y. Srinivasan" <kys@xxxxxxxxxxxxx> (hyperv)
> CC: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx> (hyperv)
> CC: Jason Gunthorpe <jgg@xxxxxxxx> (infiniband)
> CC: Leon Romanovsky <leon@xxxxxxxxxx> (infiniband)
> CC: Mark Brown <broonie@xxxxxxxxxx> (spi)
> CC: Steve French <sfrench@xxxxxxxxx> (smb)
> CC: Alexander Graf <graf@xxxxxxxxxx> (kexec)
> CC: Mike Rapoport <rppt@xxxxxxxxxx> (kexec)
> CC: Pasha Tatashin <pasha.tatashin@xxxxxxxxxx> (kexec)

Please, move the Cc: list to the...

> Signed-off-by: Yury Norov <ynorov@xxxxxxxxxx>
> ---

...comments block. It will have the same effect on emails, but drastically
reduces unneeded noise in the commit message in the Git history.

You may also read this subthread (patch 18) on how to handle it locally:
https://lore.kernel.org/linux-iio/20260123113708.416727-19-bigeasy@xxxxxxxxxxxxx/

>  include/linux/count_zeros.h | 9 +++------

...

> +#define COUNT_TRAILING_ZEROS_0 (-1)

Shouldn't we also saturate this to BITS_PER_LONG?

-- 
With Best Regards,
Andy Shevchenko