Re: [PATCH] regmap: Check for register readability before checking cache during read

[Konrad Dybcio]

On 13.06.2023 22:07, Mark Brown wrote:
> Ensure that we don't return a spurious cache hit for unreadable registers
> (eg, with the flat cache which doesn't understand sparseness) by checking
> for readability before we do a cache lookup.
> 
> Signed-off-by: Mark Brown <broonie@xxxxxxxxxx>
> ---
(+CC Bjorn)

Hi Mark,

this patch breaks using regmap_field_force_write() on fields that are
parts of registers marked as write-only (e.g. by regmap_access_table.no_ranges)

Is that intended?

What's the recommended fix?

FWIW this breaks soc/qcom/icc-bwmon.c, causing an interrupt storm at boot due
to the "clear the counters" register not being taken care of, so it'd be
appreciated if we can sort this out quickly.

Konrad
>  drivers/base/regmap/regmap.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/base/regmap/regmap.c b/drivers/base/regmap/regmap.c
> index fa2d3fba6ac9..3efbe59ca1a7 100644
> --- a/drivers/base/regmap/regmap.c
> +++ b/drivers/base/regmap/regmap.c
> @@ -2897,6 +2897,9 @@ static int _regmap_read(struct regmap *map, unsigned int reg,
>  	int ret;
>  	void *context = _regmap_map_get_context(map);
>  
> +	if (!regmap_readable(map, reg))
> +		return -EIO;
> +
>  	if (!map->cache_bypass) {
>  		ret = regcache_read(map, reg, val);
>  		if (ret == 0)
> @@ -2906,9 +2909,6 @@ static int _regmap_read(struct regmap *map, unsigned int reg,
>  	if (map->cache_only)
>  		return -EBUSY;
>  
> -	if (!regmap_readable(map, reg))
> -		return -EIO;
> -
>  	ret = map->reg_read(context, reg, val);
>  	if (ret == 0) {
>  		if (regmap_should_log(map))
> 
> ---
> base-commit: 858fd168a95c5b9669aac8db6c14a9aeab446375
> change-id: 20230613-b4-regmap-check-readability-before-cache-9f658338a5c1
> 
> Best regards,