Re: [PATCH 1/2] alpha: use .data.init_task instead of .data.init_thread.

[Richard Henderson <rth@twiddle.net>]

On 07/31/2009 02:23 PM, Richard Henderson wrote:
> Similarly it is *not* a bug that the page_aligned sections were before
> data, because we already knew we had 2 page alignment from the end of
> init + 2 pages of init_thread.

Indeed, I'll go further and say that the common definition of 
RW_DATA_SECTION is buggy.

> #define RW_DATA_SECTION(cacheline, nosave, pagealigned, inittask)       \
>         . = ALIGN(PAGE_SIZE);                                           \
>         .data : AT(ADDR(.data) - LOAD_OFFSET) {                         \
>                 INIT_TASK(inittask)                                     \
>                 CACHELINE_ALIGNED_DATA(cacheline)                       \
>                 READ_MOSTLY_DATA(cacheline)                             \
>                 DATA_DATA                                               \
>                 CONSTRUCTORS                                            \
>                 NOSAVE_DATA(nosave)                                     \
>                 PAGE_ALIGNED_DATA(pagealigned)                          \
>         }

Given that we align the entire .data section why have interior padding 
to re-align for page-aligned data?  Surely a better ordering would be

	. = ALIGN(PAGE_SIZE);
	.data : AT(ADDR(.data) - LOAD_OFFSET) {
		INIT_TASK(inittask)
		NOSAVE_DATA
		PAGE_ALIGNED_DATA(pagealigned)
		CACHELINE_ALIGNED_DATA(cacheline)
		READ_MOSTLY_DATA(cacheline)
		DATA_DATA
		CONSTRUCTORS
	}

with that change, and without

> -	. = ALIGN(2 * PAGE_SIZE);
> +	. = ALIGN(PAGE_SIZE);
>  	__init_end = .;

the patch looks like it'd be ok.  All you're doing with this fragment is 
failing to free a page of padding.  You could change it to

	. = ALIGN(THREAD_SIZE)

if it makes you feel better.


r~