linux-kernel - Re: [PATCH] ASoC: fsl_easrc: force cast 'val1' to u32 for do

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251229130211.49b169a0@pumpkin>
Date: Mon, 29 Dec 2025 13:02:11 +0000
From: David Laight <david.laight.linux@...il.com>
To: Shengjiu Wang <shengjiu.wang@....com>
Cc: shengjiu.wang@...il.com, Xiubo.Lee@...il.com, festevam@...il.com,
 nicoleotsuka@...il.com, lgirdwood@...il.com, broonie@...nel.org,
 perex@...ex.cz, tiwai@...e.com, linux-sound@...r.kernel.org,
 linuxppc-dev@...ts.ozlabs.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ASoC: fsl_easrc: force cast 'val1' to u32 for do_div()

On Mon, 29 Dec 2025 16:59:19 +0800
Shengjiu Wang <shengjiu.wang@....com> wrote:

> Fix below cocci warnings, the real width of val1 is 32bit, so use force
> cast to avoid warning.

I just looked at the uses of 'frac_bits' in that file.
Some of the related code looks very dubious.

The 200 has:
	val = (u64)in_rate << frac_bits;
	do_div(val, out_rate);
	r = (uint32_t *)&val;
followed by uses of r[0] and r[1] - that probably isn't going to DTRT big endian
and will generate worse code LE than using val and val >> 32.

Line 2063 has:
	val <<= (frac_bits - 31);
	regmap_write(easrc->regmap, REG_EASRC_RUC(pair->index), EASRC_RSUC_RS_RM(val));
where EASRC_RSUC_RS_RM(v) is (v) << 32
That is a lot of shift lefts on an 'int'.

The code in question can shift 'val1' right and convert to 32 bits
before ctx_priv->ratio_mod is added in.

	David

> 
> cocci warnings: (new ones prefixed by >>)
> >> sound/soc/fsl/fsl_easrc.c:1999:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.  
> 
> Fixes: 27147695aaf7 ("ASoC: fsl_easrc: define functions for memory to memory usage")
> Reported-by: kernel test robot <lkp@...el.com>
> Closes: https://lore.kernel.org/oe-kbuild-all/202512281902.bPsKjjkC-lkp@intel.com/
> Signed-off-by: Shengjiu Wang <shengjiu.wang@....com>
> ---
>  sound/soc/fsl/fsl_easrc.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/sound/soc/fsl/fsl_easrc.c b/sound/soc/fsl/fsl_easrc.c
> index f404a39009e1..16a45a7146e8 100644
> --- a/sound/soc/fsl/fsl_easrc.c
> +++ b/sound/soc/fsl/fsl_easrc.c
> @@ -1996,7 +1996,7 @@ static int fsl_easrc_m2m_calc_out_len(struct fsl_asrc_pair *pair, int input_buff
>  		/* right shift 12 bit to make ratio in 32bit space */
>  		val2 = (u64)in_samples << (frac_bits - 12);
>  		val1 = val1 >> 12;
> -		do_div(val2, val1);
> +		do_div(val2, (u32)val1);
>  		out_samples = val2;
>  
>  		out_length = out_samples * out_width * channels;