ASoC: fsl_easrc: Use div64_u64 for 64-by-64 division

Fix a coccinelle warning about do_div() truncating a 64-bit divisor:

sound/soc/fsl/fsl_easrc.c:2061:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.

In fsl_easrc_m2m_calc_out_len(), val1 is computed as:

  val1 = (u64)in_rate << frac_bits;   // frac_bits up to 39
  do_div(val1, out_rate);
  val1 += (s64)ctx_priv->ratio_mod << (frac_bits - 31);
  val1 = val1 >> 12;

In the worst case (in_rate=384000, out_rate=8000, frac_bits=39):
  val1 = 384000 << 39 / 8000 = 26,388,279,068,672
  val1 >> 12 = 6,440,497,829  (33 bits, exceeds 32-bit range)

val1 is then used as the divisor in do_div(val2, val1), where
do_div() silently truncates it to 32 bits, producing incorrect
results. Use div64_u64() to perform a proper 64-by-64 division.

Fixes: 955ac62405 ("ASoC: fsl_easrc: Add EASRC ASoC CPU DAI drivers")
Cc: stable@vger.kernel.org
Signed-off-by: wangdicheng <wangdicheng@kylinos.cn>
Link: https://patch.msgid.link/20260717091542.721877-4-wangdich9700@163.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This commit is contained in:
wangdicheng 2026-07-17 17:15:42 +08:00 committed by Mark Brown
parent fcf2a365da
commit a46ccc7187
No known key found for this signature in database
GPG Key ID: 24D68B725D5487D0

View File

@ -2050,7 +2050,7 @@ 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);
val2 = div64_u64(val2, val1);
out_samples = val2;
out_length = out_samples * out_width * channels;