some clean up to intel-gtt.c
authorJan Beulich <JBeulich@novell.com>
Fri, 24 Sep 2010 12:25:30 +0000 (13:25 +0100)
committerChris Wilson <chris@chris-wilson.co.uk>
Fri, 24 Sep 2010 13:22:12 +0000 (14:22 +0100)
In commit e517a5e97080bbe52857bd0d7df9b66602d53c4d the call to
map_page_into_agp() got removed from intel_i830_setup_flush(), but the
counterpart call from intel_i830_fini_flush() to unmap_page_from_agp()
was left in place.

Additionally, the page allocated here never gets its physical address
used for sending to hardware, so there's no need to allocate it with
GFP_DMA32. Nor is __GFP_ZERO really necessary, as the page is used
only to store data to force flushing of some internal processor state.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Eric Anholt <eric@anholt.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
drivers/char/agp/intel-gtt.c

index 9a03815483c78a4df3a88e29893b18df95270ae3..0c8ff6d8824bdbe3e37e9d7e337a23735b4227b8 100644 (file)
@@ -814,7 +814,6 @@ static void i830_cleanup(void)
 {
        kunmap(intel_private.i8xx_page);
        intel_private.i8xx_flush_page = NULL;
-       unmap_page_from_agp(intel_private.i8xx_page);
 
        __free_page(intel_private.i8xx_page);
        intel_private.i8xx_page = NULL;
@@ -826,7 +825,7 @@ static void intel_i830_setup_flush(void)
        if (intel_private.i8xx_page)
                return;
 
-       intel_private.i8xx_page = alloc_page(GFP_KERNEL | __GFP_ZERO | GFP_DMA32);
+       intel_private.i8xx_page = alloc_page(GFP_KERNEL);
        if (!intel_private.i8xx_page)
                return;