highmem handling again

highmem handling again

[Pavel Machek]

Hi!

I agree that swsusp_free is not the right place to restore_highmem(),
but I can't find "right" place to do it... Best I could come up is
with is:

It did not work at the end of swsusp_resume, or at the end of
swsusp_restore, IIRC.

								Pavel

--- clean-mm/kernel/power/disk.c	2004-07-28 23:39:49.000000000 +0200
+++ linux-mm/kernel/power/disk.c	2004-08-08 21:11:38.000000000 +0200
@@ -184,8 +187,11 @@
 			error = power_down(pm_disk_mode);
 			pr_debug("PM: Power down failed.\n");
 		}
-	} else
+	} else {
+		extern int restore_highmem(void);
+		restore_highmem();
 		pr_debug("PM: Image restored successfully.\n");
+	}
 	swsusp_free();
  Done:
 	finish();
--- clean-mm/kernel/power/swsusp.c	2004-07-28 23:39:49.000000000 +0200
+++ linux-mm/kernel/power/swsusp.c	2004-08-08 20:55:59.000000000 +0200
@@ -523,7 +523,7 @@
 	return 0;
 }
 
-static int restore_highmem(void)
+int restore_highmem(void)
 {
 	while (highmem_copy) {
 		struct highmem_page *save = highmem_copy;




-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!

Re: highmem handling again

[Sam Ravnborg]

On Sun, Aug 08, 2004 at 09:23:00PM +0200, Pavel Machek wrote:
>  		}
> -	} else
> +	} else {
> +		extern int restore_highmem(void);
> +		restore_highmem();

Prototype in .h files...

	Sam

Re: highmem handling again

[Pavel Machek]

Hi!

> >  		}
> > -	} else
> > +	} else {
> > +		extern int restore_highmem(void);
> > +		restore_highmem();
> 
> Prototype in .h files...

I know, this is proof of concept, only. I do not want another
cross-file reference.
							Pavel
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!

Re: highmem handling again

[Nigel Cunningham]

Hi.

On Mon, 2004-08-09 at 05:23, Pavel Machek wrote:
> Hi!
> 
> I agree that swsusp_free is not the right place to restore_highmem(),
> but I can't find "right" place to do it... Best I could come up is
> with is:
> 
> It did not work at the end of swsusp_resume, or at the end of
> swsusp_restore, IIRC.

I'm doing: copy back lowmem, restore local processor context, flush
local tlb, restore highmem, allow other processors to restore their
processor contexts and flush local tlbs. It works fine here.

Nigel

Re: highmem handling again

[Pavel Machek]

Hi!

> > I agree that swsusp_free is not the right place to restore_highmem(),
> > but I can't find "right" place to do it... Best I could come up is
> > with is:
> > 
> > It did not work at the end of swsusp_resume, or at the end of
> > swsusp_restore, IIRC.
> 
> I'm doing: copy back lowmem, restore local processor context, flush
> local tlb, restore highmem, allow other processors to restore their
> processor contexts and flush local tlbs. It works fine here.

Yes, I'd like to do the same, I just have problems finding that place
in new structure.
								Pavel
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!