Re: [PATCH] lib/genaloc: Fix allocation of aligned buffer from non-aligned chunk

[Andrew Morton]

On Wed, 7 Nov 2018 08:27:31 +0200 Alexey Skidanov <alexey.skidanov@xxxxxxxxx> wrote:

> 
> 
> On 11/7/18 12:15 AM, Andrew Morton wrote:
> > On Tue,  6 Nov 2018 14:20:53 +0200 Alexey Skidanov <alexey.skidanov@xxxxxxxxx> wrote:
> > 
> >> On success, gen_pool_first_fit_align() returns the bit number such that
> >> chunk_start_addr + (bit << order) is properly aligned. On failure,
> >> the bitmap size parameter is returned.
> >>
> >> When the chunk_start_addr isn't aligned properly, the
> >> chunk_start_addr + (bit << order) isn't aligned too.
> >>
> >> To fix this, gen_pool_first_fit_align() takes into account
> >> the chunk_start_addr alignment and returns the bit value such that
> >> chunk_start_addr + (bit << order) is properly aligned
> >> (exactly as it done in CMA).
> >>
> >> ...
> >>
> >> --- a/include/linux/genalloc.h
> >> +++ b/include/linux/genalloc.h
> >>
> >> ...
> >>
> >> +		struct gen_pool *pool, unsigned long start_add)
> >>
> >> ...
> >>
> >> +		struct gen_pool *pool, unsigned long start_add)
> >>
> >> ...
> >>
> >> +		struct gen_pool *pool, unsigned long start_add)
> >>
> >> ...
> >>
> > 
> > We have three typos here.  Which makes me wonder why we're passing the
> > new argument and then not using it?
> > 
> genpool uses allocation callbacks function that implement some
> allocation strategy - bes fit, first fit, ... All of them has the same
> type. The added chunk start_addr is used only in one of them -
> gen_pool_first_fit_align()

OK, but the argument name here is start_add, not start_addr.