[PATCH] AGP performance fixes

AGP allocation/deallocation is suffering major performance issues due to
the nature of global_flush_tlb() being called on every change_page_attr()
call.

For small allocations this isn't really seen, but when you start allocating
50000 pages of AGP space, for say, texture memory, then things can take
seconds to complete.

In some cases the situation is doubled or even quadrupled in the time due
to SMP, or a deallocation, then a new reallocation.  I've had a case of
upto 20 seconds wait time to deallocate and reallocate AGP space.

This patch fixes the problem by making it the caller's responsibility to
call global_flush_tlb(), and so removes it from every instance of mapping a
page into AGP space until the time that all change_page_attr() changes are
done.

Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>

1 files changed, 4 insertions, 1 deletions

diff --git a/drivers/char/agp/intel-agp.c b/drivers/char/agp/intel-agp.c
index bf4cc9ffd5b1..027161ab88e9 100644
--- a/drivers/char/agp/intel-agp.c
+++ b/drivers/char/agp/intel-agp.c
@@ -270,6 +270,7 @@ static struct agp_memory *alloc_agpphysmem_i8xx(size_t pg_count, int type)
 
         switch (pg_count) {
         case 1: addr = agp_bridge->driver->agp_alloc_page(agp_bridge);
+                global_flush_tlb();
                 break;
         case 4:
                 /* kludge to get 4 physical pages for ARGB cursor */
@@ -330,9 +331,11 @@ static void intel_i810_free_by_type(struct agp_memory *curr)
         if(curr->type == AGP_PHYS_MEMORY) {
                 if (curr->page_count == 4)
                         i8xx_destroy_pages(gart_to_virt(curr->memory[0]));
-                else
+                else {
                         agp_bridge->driver->agp_destroy_page(
                                  gart_to_virt(curr->memory[0]));
+                        global_flush_tlb();
+                }
                 vfree(curr->memory);
         }
         kfree(curr);