[2.6 patch] drivers/input/misc/wistron_btns.c: section fixes

[2.6 patch] drivers/input/misc/wistron_btns.c: section fixes

[Adrian Bunk]

This patch contains the following fixes:
- it doesn't make sense to mark a variable on the stack as __initdata
- struct dmi_ids is using the __init dmi_matched()
  since the only user of struct dmi_ids is the __init select_keymap(),
  the solution is to make struct dmi_ids __initdata

Signed-off-by: Adrian Bunk <bunk@stusta.de>

---

 drivers/input/misc/wistron_btns.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

--- linux-2.6.17-mm2-full/drivers/input/misc/wistron_btns.c.old	2006-06-26 02:03:20.000000000 +0200
+++ linux-2.6.17-mm2-full/drivers/input/misc/wistron_btns.c	2006-06-26 02:07:58.000000000 +0200
@@ -94,7 +94,7 @@
 
 static ssize_t __init locate_wistron_bios(void __iomem *base)
 {
-	static const unsigned char __initdata signature[] =
+	static const unsigned char signature[] =
 		{ 0x42, 0x21, 0x55, 0x30 };
 	ssize_t offset;
 
@@ -333,7 +333,7 @@
  * a list of buttons and their key codes (reported when loading this module
  * with force=1) and the output of dmidecode to $MODULE_AUTHOR.
  */
-static struct dmi_system_id dmi_ids[] = {
+static struct dmi_system_id __initdata dmi_ids[] = {
 	{
 		.callback = dmi_matched,
 		.ident = "Fujitsu-Siemens Amilo Pro V2000",

Re: [2.6 patch] drivers/input/misc/wistron_btns.c: section fixes

[Dmitry Torokhov]

On Monday 26 June 2006 06:35, Adrian Bunk wrote:
> This patch contains the following fixes:
> - it doesn't make sense to mark a variable on the stack as __initdata
> - struct dmi_ids is using the __init dmi_matched()

Since when did static variables become allocated on stack?

-- 
Dmitry

Re: [2.6 patch] drivers/input/misc/wistron_btns.c: section fixes

[Dmitry Torokhov]

On 6/26/06, Dmitry Torokhov <dtor_core@ameritech.net> wrote:
> On Monday 26 June 2006 06:35, Adrian Bunk wrote:
> > This patch contains the following fixes:
> > - it doesn't make sense to mark a variable on the stack as __initdata
> > - struct dmi_ids is using the __init dmi_matched()
>
> Since when did static variables become allocated on stack?
>

BTW, if I add __intidata to dmi_ids array GCC (3.4.4) bitches at me:

  CC      drivers/input/misc/wistron_btns.o
drivers/input/misc/wistron_btns.c:345: error: dmi_ids causes a section
type conflict
make[1]: *** [drivers/input/misc/wistron_btns.o] Error 1
make: *** [drivers/input/misc/] Error 2

I have to declare it as "static const struct dmi_system_id __initdata
dmi_ids[] = {..." for it to compile successfully, but there is a
comment that we should not use const with __initdata. Although I
checked the assembly oputput and const __initdata endes up in
init.data section - exactly where we want it. Wierd...

-- 
Dmitry

Re: [2.6 patch] drivers/input/misc/wistron_btns.c: section fixes

[Adrian Bunk]

On Mon, Jun 26, 2006 at 07:50:31AM -0400, Dmitry Torokhov wrote:
> On Monday 26 June 2006 06:35, Adrian Bunk wrote:
> > This patch contains the following fixes:
> > - it doesn't make sense to mark a variable on the stack as __initdata
> > - struct dmi_ids is using the __init dmi_matched()
> 
> Since when did static variables become allocated on stack?

Only when I miss the static...

> Dmitry

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed