[PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[Cyrill Gorcunov]

This patch does clean up relocate_kernel_(32|64).S a bit by getting rid
of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE instead

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
---

 relocate_kernel_32.S |    3 +--
 relocate_kernel_64.S |    3 +--
 2 files changed, 2 insertions(+), 4 deletions(-)

Index: linux-2.6.git/arch/x86/kernel/relocate_kernel_32.S
===================================================================
--- linux-2.6.git.orig/arch/x86/kernel/relocate_kernel_32.S	2008-03-15 11:24:51.000000000 +0300
+++ linux-2.6.git/arch/x86/kernel/relocate_kernel_32.S	2008-03-21 23:00:10.000000000 +0300
@@ -15,12 +15,11 @@
  */
 
 #define PTR(x) (x << 2)
-#define PAGE_ALIGNED (1 << PAGE_SHIFT)
 #define PAGE_ATTR 0x63 /* _PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_DIRTY */
 #define PAE_PGD_ATTR 0x01 /* _PAGE_PRESENT */
 
 	.text
-	.align PAGE_ALIGNED
+	.align PAGE_SIZE
 	.globl relocate_kernel
 relocate_kernel:
 	movl	8(%esp), %ebp /* list of pages */
Index: linux-2.6.git/arch/x86/kernel/relocate_kernel_64.S
===================================================================
--- linux-2.6.git.orig/arch/x86/kernel/relocate_kernel_64.S	2008-03-15 11:24:51.000000000 +0300
+++ linux-2.6.git/arch/x86/kernel/relocate_kernel_64.S	2008-03-21 22:59:08.000000000 +0300
@@ -15,11 +15,10 @@
  */
 
 #define PTR(x) (x << 3)
-#define PAGE_ALIGNED (1 << PAGE_SHIFT)
 #define PAGE_ATTR 0x63 /* _PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_DIRTY */
 
 	.text
-	.align PAGE_ALIGNED
+	.align PAGE_SIZE
 	.code64
 	.globl relocate_kernel
 relocate_kernel:

Re: [PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[H. Peter Anvin]

Cyrill Gorcunov wrote:
> This patch does clean up relocate_kernel_(32|64).S a bit by getting rid
> of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE instead
> 
> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>

Looks good!

	-hpa

Re: [PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[Cyrill Gorcunov]

[H. Peter Anvin - Fri, Mar 21, 2008 at 01:18:13PM -0700]
> Cyrill Gorcunov wrote:
>> This patch does clean up relocate_kernel_(32|64).S a bit by getting rid
>> of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE instead
>> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
>
> Looks good!
>
> 	-hpa
>

Thanks Peter for review, btw we have in relocate_kernel_32.S

188:	movl	%cr4, %eax
189:	andl	$0, %eax
190:	movl	%eax, %cr4

but maybe would be better just replace it with

	xorl	%eax, %eax
	movl	%eax, %cr4

or there is some nonobvious reason not to do so?

		- Cyrill -

Re: [PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[H. Peter Anvin]

Cyrill Gorcunov wrote:
> [H. Peter Anvin - Fri, Mar 21, 2008 at 01:18:13PM -0700]
>> Cyrill Gorcunov wrote:
>>> This patch does clean up relocate_kernel_(32|64).S a bit by getting rid
>>> of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE instead
>>> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
>> Looks good!
>>
>> 	-hpa
>>
> 
> Thanks Peter for review, btw we have in relocate_kernel_32.S
> 
> 188:	movl	%cr4, %eax
> 189:	andl	$0, %eax
> 190:	movl	%eax, %cr4
> 
> but maybe would be better just replace it with
> 
> 	xorl	%eax, %eax
> 	movl	%eax, %cr4
> 
> or there is some nonobvious reason not to do so?
> 

The only case I can think of would be if the code expects to patch the 
zero, but even then it'd be cantankerous, because gas will optimize that 
zero down to a single byte.

So, no :)

	-hpa

Re: [PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[Cyrill Gorcunov]

[H. Peter Anvin - Fri, Mar 21, 2008 at 01:50:33PM -0700]
> Cyrill Gorcunov wrote:
>> [H. Peter Anvin - Fri, Mar 21, 2008 at 01:18:13PM -0700]
>>> Cyrill Gorcunov wrote:
>>>> This patch does clean up relocate_kernel_(32|64).S a bit by getting rid
>>>> of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE instead
>>>> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
>>> Looks good!
>>>
>>> 	-hpa
>>>
>> Thanks Peter for review, btw we have in relocate_kernel_32.S
>> 188:	movl	%cr4, %eax
>> 189:	andl	$0, %eax
>> 190:	movl	%eax, %cr4
>> but maybe would be better just replace it with
>> 	xorl	%eax, %eax
>> 	movl	%eax, %cr4
>> or there is some nonobvious reason not to do so?
>
> The only case I can think of would be if the code expects to patch the 
> zero, but even then it'd be cantankerous, because gas will optimize that 
> zero down to a single byte.
>
> So, no :)
>
> 	-hpa
>
>

ah! thanks ;)

		- Cyrill -

Re: [PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[Ingo Molnar]

* Cyrill Gorcunov <gorcunov@gmail.com> wrote:

> This patch does clean up relocate_kernel_(32|64).S a bit by getting 
> rid of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE 
> instead

thanks Cyrill, applied.

	Ingo

Re: [PATCH] x86: relocate_kernel - use predefined PAGE_SIZE instead of own alias

[Cyrill Gorcunov]

[H. Peter Anvin - Fri, Mar 21, 2008 at 01:50:33PM -0700]
> Cyrill Gorcunov wrote:
>> [H. Peter Anvin - Fri, Mar 21, 2008 at 01:18:13PM -0700]
>>> Cyrill Gorcunov wrote:
>>>> This patch does clean up relocate_kernel_(32|64).S a bit by getting rid
>>>> of local PAGE_ALIGNED macro. We should use well-known PAGE_SIZE instead
>>>> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
>>> Looks good!
>>>
>>> 	-hpa
>>>
>> Thanks Peter for review, btw we have in relocate_kernel_32.S
>> 188:	movl	%cr4, %eax
>> 189:	andl	$0, %eax
>> 190:	movl	%eax, %cr4
>> but maybe would be better just replace it with
>> 	xorl	%eax, %eax
>> 	movl	%eax, %cr4
>> or there is some nonobvious reason not to do so?
>
> The only case I can think of would be if the code expects to patch the 
> zero, but even then it'd be cantankerous, because gas will optimize that 
> zero down to a single byte.
>
> So, no :)
>
> 	-hpa
>
well, Peter, it seems I'm too tired on this week and misinterpred
your message, my prev message with text "ah..." is wrong, i'll patch
this code later (almost sleeping now ;)

		- Cyrill -