Current state of regmap.

Current state of regmap.

[Jonathan Cameron]

Hi Mark,

As my previous email said I've been messing around with regmap today.
I've run into a couple of bugs and was wondering if I'm running
behind the current version.  I've pulled what is up on kernel.org
for linux-next at the moment.

As I'm emailing anyway.  Everywhere other than regmap-debugfs max_register
is treated as the max valid.  In there we have a loop that is
for (i = 0; i < map->max_register; i++) so that treats it as one
larger than the max valid register index.

Thanks,

Jonathan.

p.s. Nice work by the way.

Re: Current state of regmap.

[Mark Brown]

On Mon, Sep 05, 2011 at 03:14:14PM +0100, Jonathan Cameron wrote:

> As my previous email said I've been messing around with regmap today.
> I've run into a couple of bugs and was wondering if I'm running
> behind the current version.  I've pulled what is up on kernel.org
> for linux-next at the moment.

That's the current state, I'm running development systems here where all
the I2C and SPI devices on the system go through regmap and anyone
running ASoC will be using it.  Though there's a few patches floating
around now and I'm just testing the module_put() fix so it'll not be the
current state soonish.  kernel.org is closed for updates right now, if I
do much i'll probably show up on opensource.wolfsonmicro.com in the
short term.

> As I'm emailing anyway.  Everywhere other than regmap-debugfs max_register
> is treated as the max valid.  In there we have a loop that is
> for (i = 0; i < map->max_register; i++) so that treats it as one
> larger than the max valid register index.

That's an off by one in the debugfs code, will fix.  You can check this
sort of stuff out by looking at the users.