[3.8.y.z extended stable] Patch "drm/ttm: Fix memory type compatibility check" has been added to staging queue
Kamal Mostafa
kamal at canonical.com
Fri Dec 6 23:08:20 UTC 2013
This is a note to let you know that I have just added a patch titled
drm/ttm: Fix memory type compatibility check
to the linux-3.8.y-queue branch of the 3.8.y.z extended stable tree
which can be found at:
http://kernel.ubuntu.com/git?p=ubuntu/linux.git;a=shortlog;h=refs/heads/linux-3.8.y-queue
This patch is scheduled to be released in version 3.8.13.14.
If you, or anyone else, feels it should not be added to this tree, please
reply to this email.
For more information about the 3.8.y.z tree, see
https://wiki.ubuntu.com/Kernel/Dev/ExtendedStable
Thanks.
-Kamal
------
>From 04623382c8e829413c2b69f740a9455e2e764ab7 Mon Sep 17 00:00:00 2001
From: Thomas Hellstrom <thellstrom at vmware.com>
Date: Mon, 28 Oct 2013 02:02:19 -0700
Subject: drm/ttm: Fix memory type compatibility check
commit 59c8e66378fb78adbcd05f0d09783dde6fef282b upstream.
Also check the busy placements before deciding to move a buffer object.
Failing to do this may result in a completely unneccessary move within a
single memory type.
Signed-off-by: Thomas Hellstrom <thellstrom at vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob at vmware.com>
Signed-off-by: Kamal Mostafa <kamal at canonical.com>
---
drivers/gpu/drm/ttm/ttm_bo.c | 32 ++++++++++++++++++++------------
1 file changed, 20 insertions(+), 12 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 52b20b1..ba60a89 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1090,24 +1090,32 @@ out_unlock:
return ret;
}
-static int ttm_bo_mem_compat(struct ttm_placement *placement,
- struct ttm_mem_reg *mem)
+static bool ttm_bo_mem_compat(struct ttm_placement *placement,
+ struct ttm_mem_reg *mem,
+ uint32_t *new_flags)
{
int i;
if (mem->mm_node && placement->lpfn != 0 &&
(mem->start < placement->fpfn ||
mem->start + mem->num_pages > placement->lpfn))
- return -1;
+ return false;
for (i = 0; i < placement->num_placement; i++) {
- if ((placement->placement[i] & mem->placement &
- TTM_PL_MASK_CACHING) &&
- (placement->placement[i] & mem->placement &
- TTM_PL_MASK_MEM))
- return i;
+ *new_flags = placement->placement[i];
+ if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
+ (*new_flags & mem->placement & TTM_PL_MASK_MEM))
+ return true;
+ }
+
+ for (i = 0; i < placement->num_busy_placement; i++) {
+ *new_flags = placement->busy_placement[i];
+ if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
+ (*new_flags & mem->placement & TTM_PL_MASK_MEM))
+ return true;
}
- return -1;
+
+ return false;
}
int ttm_bo_validate(struct ttm_buffer_object *bo,
@@ -1116,6 +1124,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
bool no_wait_gpu)
{
int ret;
+ uint32_t new_flags;
BUG_ON(!ttm_bo_is_reserved(bo));
/* Check that range is valid */
@@ -1126,8 +1135,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
/*
* Check whether we need to move buffer.
*/
- ret = ttm_bo_mem_compat(placement, &bo->mem);
- if (ret < 0) {
+ if (!ttm_bo_mem_compat(placement, &bo->mem, &new_flags)) {
ret = ttm_bo_move_buffer(bo, placement, interruptible,
no_wait_gpu);
if (ret)
@@ -1137,7 +1145,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
* Use the access and other non-mapping-related flag bits from
* the compatible memory placement flags to the active flags
*/
- ttm_flag_masked(&bo->mem.placement, placement->placement[ret],
+ ttm_flag_masked(&bo->mem.placement, new_flags,
~TTM_PL_MASK_MEMTYPE);
}
/*
--
1.8.3.2
More information about the kernel-team
mailing list