Re: [PATCH] iio: magnetometer: rm3100: check TMRC in rm3100_get_samp_freq()

From: Andy Shevchenko

Date: Thu Nov 19 2026 - 04:49:13 EST


On Thu, Nov 19, 2026 at 12:41:06AM +0800, Maoyi Xie wrote:
> rm3100_get_samp_freq() reads TMRC from the sensor and uses the value
> minus RM3100_TMRC_OFFSET as an index into rm3100_samp_rates[] without a
> check. A value outside 0x92 to 0x9f reads outside the table. The driver
> only writes values in that range, so this needs a chip or bus that
> misbehaves after probe.
>
> Commit 792595bab492 ("iio: magnetometer: rm3100: add boundary check for
> the value read from RM3100_REG_TMRC") added the check to
> rm3100_common_probe() after a crash on failing hardware. Do the same
> here and return -EINVAL.
>
> Found with a CodeQL-based static analysis tool that we are developing.
> Reproduced in a userspace harness under ASan and UBSan, which report
> nothing with this patch. Not tested on hardware, only compile-tested
> with W=1 on next-20261116.

...

> static int rm3100_get_samp_freq(struct rm3100_data *data, int *val, int *val2)
> {
> unsigned int tmp;
> int ret;
> + int samp_rate_index;

Preserve reversed xmas tree order.

...

> - *val = rm3100_samp_rates[tmp - RM3100_TMRC_OFFSET][0];
> - *val2 = rm3100_samp_rates[tmp - RM3100_TMRC_OFFSET][1];
> + samp_rate_index = tmp - RM3100_TMRC_OFFSET;
> + if (samp_rate_index < 0 || samp_rate_index >= RM3100_SAMP_NUM)

Wouldn't be better to use ARRAY_SIZE() if possible?

> + return -EINVAL;

Maybe -ERANGE?

> + *val = rm3100_samp_rates[samp_rate_index][0];
> + *val2 = rm3100_samp_rates[samp_rate_index][1];

--
With Best Regards,
Andy Shevchenko