/*
*/
/*
* Copyright (c) 2009, 2013, Intel Corporation.
* All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*
* Authors:
* Eric Anholt <eric@anholt.net>
*
*/
#include "drmP.h"
#include "drm.h"
#include "i915_drm.h"
#include "i915_drv.h"
/** @file i915_gem_tiling.c
*
* Support for managing tiling state of buffer objects.
*
* The idea behind tiling is to increase cache hit rates by rearranging
* pixel data so that a group of pixel accesses are in the same cacheline.
* the order of 30%.
*
* Intel architectures make this somewhat more complicated, though, by
* adjustments made to addressing of data when the memory is in interleaved
* mode (matched pairs of DIMMS) to improve memory bandwidth.
* For interleaved memory, the CPU sends every sequential 64 bytes
* to an alternate memory channel so it can get the bandwidth from both.
*
* The GPU also rearranges its accesses for increased bandwidth to interleaved
* memory, and it matches what the CPU does for non-tiled. However, when tiled
* it does it a little differently, since one walks addresses not just in the
* X direction but also Y. So, along with alternating channels when bit
* 6 of the address flips, it also alternates when other bits flip -- Bits 9
* (every 512 bytes, an X tile scanline) and 10 (every two X tile scanlines)
* are common to both the 915 and 965-class hardware.
*
* The CPU also sometimes XORs in higher bits as well, to improve
* bandwidth doing strided access like we do so frequently in graphics. This
* is called "Channel XOR Randomization" in the MCH documentation. The result
* is that the CPU is XORing in either bit 11 or bit 17 to bit 6 of its address
* decode.
*
* All of this bit 6 XORing has an effect on our memory management,
* as we need to make sure that the 3d driver can correctly address object
* contents.
*
* If we don't have interleaved memory, all tiling is safe and no swizzling is
* required.
*
* When bit 17 is XORed in, we simply refuse to tile at all. Bit
* 17 is not just a page offset, so as we page an objet out and back in,
* individual pages in it will have different bit 17 addresses, resulting in
* each 64 bytes being swapped with its neighbor!
*
* Otherwise, if interleaved, we have to tell the 3d driver what the address
* swizzling it needs to do is, since it's writing with the CPU to the pages
* (bit 6 and potentially bit 11 XORed in), and the GPU is reading from the
* pages (bit 6, 9, and 10 XORed in), resulting in a cumulative bit swizzling
* required by the CPU of XORing in bit 6, 9, 10, and potentially 11, in order
* to match what the GPU expects.
*/
/**
* Detects bit 6 swizzling of address lookup between IGD access and CPU
* access through main memory.
*/
void
{
if (IS_VALLEYVIEW(dev)) {
/* Enable swizzling when the channels are populated with
* identically sized dimms. We don't need to check the 3rd
* channel because no cpu with gpu attached ships in that
* configuration. Also, swizzling only makes sense for 2
* channels anyway. */
} else {
}
/* On IRONLAKE whatever DRAM config, GPU always do
* same swizzling setup.
*/
/* As far as we know, the 865 doesn't have these bit 6
* swizzling issues.
*/
/* On mobile 9xx chipsets, channel interleave by the CPU is
* determined by DCC. For single-channel, neither the CPU
* nor the GPU do swizzling. For dual channel interleaved,
* the GPU's interleave is bit 9 and 10 for X tiled, and bit
* 9 for Y tiled. The CPU's interleave is independent, and
* can be based on either bit 11 (haven't seen this yet) or
* bit 17 (common).
*/
switch (dcc & DCC_ADDRESSING_MODE_MASK) {
break;
if (dcc & DCC_CHANNEL_XOR_DISABLE) {
/* This is the base swizzling by the GPU for
* tiled buffers.
*/
} else if ((dcc & DCC_CHANNEL_XOR_BIT_17) == 0) {
/* Work around: for fixing 965GM flickering issue on OpenArena */
} else {
/* Bit 17 swizzling by the CPU in addition. */
}
break;
}
if (dcc == 0xffffffff) {
DRM_ERROR("Couldn't read from MCHBAR. "
"Disabling tiling.\n");
}
} else {
/* The 965, G33, and newer, have a very flexible memory
* configuration. It will enable dual-channel mode
* (interleaving) on as much memory as it can, and the GPU
* will additionally sometimes enable different bit 6
* swizzling for tiled objects from the CPU.
*
* Here's what I found on the G965:
* slot fill memory size swizzling
* 0A 0B 1A 1B 1-ch 2-ch
* 512 0 0 0 512 0 O
* 512 0 512 0 16 1008 X
* 512 0 0 512 16 1008 X
* 0 512 0 512 16 1008 X
* 1024 1024 1024 0 2048 1024 O
*
* We could probably detect this based on either the DRB
* matching, which was the case for the swizzling required in
* the table above, or from the 1-ch value being less than
* the minimum size of a rank.
*/
} else {
}
}
}
/* Check pitch constriants for all chips & tiling formats */
static bool
{
int tile_width;
/* Linear is always fine */
if (tiling_mode == I915_TILING_NONE)
return true;
tile_width = 128;
else
tile_width = 512;
/* check maximum stride & object size */
/* i965+ stores the end address of the gtt mapping in the fence
* reg, so dont bother to check the size */
return false;
return false;
} else {
if (stride > 8192)
return false;
return false;
} else {
return false;
}
}
if (stride < tile_width)
return false;
/* 965+ just needs multiples of tile width */
return false;
return true;
}
/* Pre-965 needs power of two tile widths */
return false;
return true;
}
/* Is the current GTT allocation valid for the change in tiling? */
static bool
{
if (tiling_mode == I915_TILING_NONE)
return true;
return true;
return false;
} else {
return false;
}
return false;
return false;
return true;
}
/**
* Sets the tiling mode of an object, returning the required swizzling of
* bit 6 of addresses in the object.
*/
int
/* LINTED */
{
int ret = 0;
return -ENOENT;
if (!i915_tiling_ok(dev,
return -EINVAL;
}
return -EBUSY;
}
} else {
else
/* Hide bit 17 swizzling from the user. This prevents old Mesa
* from aborting the application on sw fallbacks to bit 17,
* If there was a user that was relying on the swizzle
* break it, but we don't have any of those.
*/
/* If we can't handle the swizzling, make it untiled. */
}
}
/* We need to rebind the object if its current allocation
* no longer meets the alignment restrictions for its new
* tiling mode. Otherwise we can just leave it alone, but
* need to ensure that any fence register is cleared.
* the next fenced (either through the GTT or by the BLT unit
* on older GPUs) access.
*
* After updating the tiling parameters, we then flag whether
* we need to update an associated fence register. Note this
* has to also include the unfenced register the GPU uses
* whilst executing a fenced command for an untiled object.
*/
/* Rebind if we need a change of alignment */
if (!obj->map_and_fenceable) {
false);
}
if (ret == 0) {
obj->fence_dirty =
obj->fenced_gpu_access ||
/* Force the fence to be reacquired for GTT access */
}
}
/* we have to maintain this existing ABI... */
/* Try to preallocate memory required to save swizzling on put-pages */
sizeof(long), GFP_KERNEL);
}
} else {
sizeof(long));
}
}
return ret;
}
/**
* Returns the current tiling mode and required bit 6 swizzling for the object.
*/
int
/* LINTED */
{
return -ENOENT;
switch (obj->tiling_mode) {
case I915_TILING_X:
break;
case I915_TILING_Y:
break;
case I915_TILING_NONE:
break;
default:
DRM_ERROR("unknown tiling mode\n");
}
/* Hide bit 17 from the user -- see comment in i915_gem_set_tiling */
return 0;
}
/**
* Swap every 64 bytes of this page around, to account for it having a new
* bit 17 of its physical address and therefore being interpreted differently
* by the GPU.
*/
static void
{
char *vaddr;
int i;
for (i = 0; i < PAGE_SIZE; i += 128) {
}
}
void
{
int i;
return;
for (i = 0; i < page_count; i++) {
if ((new_bit_17 & 0x1) !=
}
}
}
void
{
int i;
sizeof(long), GFP_KERNEL);
DRM_ERROR("Failed to allocate memory for bit 17 "
"record\n");
return;
}
}
for (i = 0; i < page_count; i++) {
else
}
}