[PATCH] ASoC: mediatek: mt8365: Fix out-of-bounds read in get_iir_coef()
Priyanka Mani
priyankamani2100 at gmail.com
Sun Oct 11 22:31:19 PDT 2026
get_iir_coef() looks up a resampling coefficient set by reading an
index "k" from iir_coef_tbl_matrix[][] and using it to index
iir_coef_tbl_list[], which has 8 entries (valid indices 0..7).
Before the lookup, "k" is only rejected when it is >= IIR_NO_NEED (11),
or equal to IIR_RATIOVER (9) or IIR_INV_COEF (10). The value 8 is
neither a sentinel nor a valid table index, so it falls through to the
else branch and iir_coef_tbl_list[8] is read one element past the end
of the array. The bogus coef pointer and count are then returned to
mt8365_dai_set_config(), which dereferences coef.
The current matrix never produces 8, so this cannot be triggered today,
but the bounds check is wrong: it relies on the matrix contents instead
of the size of the array it guards. Bound the lookup with the actual
array size so only valid indices reach the table.
This was found by Smatch:
sound/soc/mediatek/mt8365/mt8365-dai-i2s.c:271 get_iir_coef()
error: buffer overflow 'iir_coef_tbl_list' 8 <= 8
Fixes: 402bbb13a195 ("ASoC: mediatek: mt8365: Add I2S DAI support")
Reported-by: Dan Carpenter <dan.carpenter at linaro.org>
Closes: https://lore.kernel.org/all/67af3580-a907-403d-8bfd-76b5c304ab99@stanley.mountain/
Signed-off-by: Priyanka Mani <priyankamani2100 at gmail.com>
---
sound/soc/mediatek/mt8365/mt8365-dai-i2s.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/mediatek/mt8365/mt8365-dai-i2s.c b/sound/soc/mediatek/mt8365/mt8365-dai-i2s.c
index a058973662b3..2e678f4aa4b1 100644
--- a/sound/soc/mediatek/mt8365/mt8365-dai-i2s.c
+++ b/sound/soc/mediatek/mt8365/mt8365-dai-i2s.c
@@ -267,7 +267,7 @@ static const u32 *get_iir_coef(unsigned int input_fs,
if (k >= IIR_NO_NEED) {
} else if (k == IIR_RATIOVER) {
} else if (k == IIR_INV_COEF) {
- } else {
+ } else if (k < ARRAY_SIZE(iir_coef_tbl_list)) {
coef = iir_coef_tbl_list[k].coef;
cnt = iir_coef_tbl_list[k].cnt;
}
--
2.43.0
More information about the linux-arm-kernel
mailing list