Re: [PATCH v4 1/3] regcache: Move HW readback after cache initialisation

[Andy Shevchenko]

On Mon, Mar 02, 2026 at 02:49:49PM +0000, Mark Brown wrote:
> On Mon, Mar 02, 2026 at 10:56:55AM +0100, Andy Shevchenko wrote:

...

> > +	/*
> > +	 * Some devices such as PMICs don't have cache defaults,
> > +	 * we cope with this by reading back the HW registers and
> > +	 * crafting the cache defaults by hand.
> > +	 */
> > +	ret = regcache_hw_init(map, count);
> > +	if (ret)
> > +		goto err_exit;
> > +
> 
> We need to delete the free of reg_defaults from regcache_hw_init() when
> it fails, the error handling for err_exit also frees that so we'd end up
> with a double free.  Which is awfully error prone and part of what
> you're trying to fix here :/

Ah, looking again at it I see your point.

Indeed, in the error path of regcache_hw_init() we free the reg_defaults,
but the same is also done at err_free label of regcache_init().

NULLifying it at regcache_hw_init() doesn't sound like a good idea as
you pointed out as error prone case. I will think more about it, thanks
for catching this (I dunno why I missed that one)!

-- 
With Best Regards,
Andy Shevchenko