On Fri, 2009-09-11 at 12:43 -0400, Mike Frysinger wrote:
> On Mon, Aug 24, 2009 at 22:29,  <[email protected]> wrote:
> > Log Message
> >
> > Fix typo
> >
> > Modified: trunk/arch/blackfin/mach-bf561/include/mach/mem_map.h (7231 =>
> > 7232)
> >
> >  # define COREB_L1_DATA_A_LENGTH   L1_DATA_A_LENGTH
> >  # define COREB_L1_DATA_B_LENGTH   L1_DATA_B_LENGTH
> >  #else
> > -# define COREB_L1_CODE_LENGTH     0x14000
> > +# define COREB_L1_CODE_LENGTH     0x8000
> >  # define COREB_L1_DATA_A_LENGTH   0x8000
> >  # define COREB_L1_DATA_B_LENGTH   0x8000
> >  #endif
> 
> this isnt a typo.  you cant assume cache is enabled in core b which
> means we have to grant access to it.  thus we use a value that covers
> L1 inst sram as well as L1 inst sram/cache, and a small hole of memory
Thanks, you are right, I missed the small hole.

> that doesnt actually exists.  it keeps code that does access checking
> much simpler.  in fact, the comment right above these defines explains
> exactly why the values are what they are.
> 
> what exactly was the problem that made you change this ?
> -mike

_______________________________________________
Linux-kernel-commits mailing list
[email protected]
https://blackfin.uclinux.org/mailman/listinfo/linux-kernel-commits

Reply via email to