On Wed, May 28, 2008 at 03:02:14PM +0100, Jeremy Fitzhardinge wrote:
quoted text
> 
> The use of __section(.data.page_aligned) (or worse 
> __attribute__((section(".data.page_aligned"))) is fairly verbose and 
> brittle. I've got a (totally untested) proposed patch below, to 
> introduce __page_aligned_data|bss which sets the section and the 
> alignment. This will work, but it requires that all page-aligned 
> variables also have an alignment associated with them, so that mis-sized 
> ones don't push the others around.
> 
> There aren't very many users of .data|bss.page_aligned, so it should be 
> easy enough to fix them all up.
> 
> A link-time warning would be good too, of course.

I cooked up this:
diff --git a/arch/x86/kernel/vmlinux_32.lds.S b/arch/x86/kernel/vmlinux_32.lds.S
index ce5ed08..963b2ae 100644
--- a/arch/x86/kernel/vmlinux_32.lds.S
+++ b/arch/x86/kernel/vmlinux_32.lds.S
@@ -40,6 +40,7 @@ SECTIONS
   .text : AT(ADDR(.text) - LOAD_OFFSET) {
        . = ALIGN(PAGE_SIZE); /* not really needed, already page aligned */
        *(.text.page_aligned)
+       end_text_page_aligned = .;
        TEXT_TEXT
        SCHED_TEXT
        LOCK_TEXT
@@ -49,6 +50,9 @@ SECTIONS
        _etext = .;                     /* End of text section */
   } :text = 0x9090

+       ASSERT((end_text_page_aligned == ALIGN((end_text_page_aligned), PAGE_SIZE)),
+              "Text in .text.page_aligned are not modulo PAGE_SIZE")
+
   . = ALIGN(16);               /* Exception table */
   __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) {
        __start___ex_table = .;
@@ -89,6 +93,8 @@ SECTIONS
        *(.data.page_aligned)
        *(.data.idt)
   }
+       ASSERT((. == ALIGN(PAGE_SIZE)),
+              "Data in .data.page_aligned are not modulo PAGE_SIZE")

   . = ALIGN(32);
   .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {


But we should try to do it so all archs can benefit.
And it failed in the second ASSERT - I dunno why.

Soccer duties - so I have to run.

	Sam
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/