linux_dsm_epyc7002/drivers/gpu/drm/tegra/plane.c
Thierry Reding 273da5a046 drm/tegra: Reuse IOVA mapping where possible
This partially reverts the DMA API support that was recently merged
because it was causing performance regressions on older Tegra devices.
Unfortunately, the cache maintenance performed by dma_map_sg() and
dma_unmap_sg() causes performance to drop by a factor of 10.

The right solution for this would be to cache mappings for buffers per
consumer device, but that's a bit involved. Instead, we simply revert to
the old behaviour of sharing IOVA mappings when we know that devices can
do so (i.e. they share the same IOMMU domain).

Cc: <stable@vger.kernel.org> # v5.5
Reported-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
2020-02-06 18:21:55 +01:00

602 lines
13 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (C) 2017 NVIDIA CORPORATION. All rights reserved.
*/
#include <linux/iommu.h>
#include <drm/drm_atomic.h>
#include <drm/drm_atomic_helper.h>
#include <drm/drm_fourcc.h>
#include <drm/drm_gem_framebuffer_helper.h>
#include <drm/drm_plane_helper.h>
#include "dc.h"
#include "plane.h"
static void tegra_plane_destroy(struct drm_plane *plane)
{
struct tegra_plane *p = to_tegra_plane(plane);
drm_plane_cleanup(plane);
kfree(p);
}
static void tegra_plane_reset(struct drm_plane *plane)
{
struct tegra_plane *p = to_tegra_plane(plane);
struct tegra_plane_state *state;
unsigned int i;
if (plane->state)
__drm_atomic_helper_plane_destroy_state(plane->state);
kfree(plane->state);
plane->state = NULL;
state = kzalloc(sizeof(*state), GFP_KERNEL);
if (state) {
plane->state = &state->base;
plane->state->plane = plane;
plane->state->zpos = p->index;
plane->state->normalized_zpos = p->index;
for (i = 0; i < 3; i++)
state->iova[i] = DMA_MAPPING_ERROR;
}
}
static struct drm_plane_state *
tegra_plane_atomic_duplicate_state(struct drm_plane *plane)
{
struct tegra_plane_state *state = to_tegra_plane_state(plane->state);
struct tegra_plane_state *copy;
unsigned int i;
copy = kmalloc(sizeof(*copy), GFP_KERNEL);
if (!copy)
return NULL;
__drm_atomic_helper_plane_duplicate_state(plane, &copy->base);
copy->tiling = state->tiling;
copy->format = state->format;
copy->swap = state->swap;
copy->bottom_up = state->bottom_up;
copy->opaque = state->opaque;
for (i = 0; i < 2; i++)
copy->blending[i] = state->blending[i];
for (i = 0; i < 3; i++) {
copy->iova[i] = DMA_MAPPING_ERROR;
copy->sgt[i] = NULL;
}
return &copy->base;
}
static void tegra_plane_atomic_destroy_state(struct drm_plane *plane,
struct drm_plane_state *state)
{
__drm_atomic_helper_plane_destroy_state(state);
kfree(state);
}
static bool tegra_plane_format_mod_supported(struct drm_plane *plane,
uint32_t format,
uint64_t modifier)
{
const struct drm_format_info *info = drm_format_info(format);
if (modifier == DRM_FORMAT_MOD_LINEAR)
return true;
if (info->num_planes == 1)
return true;
return false;
}
const struct drm_plane_funcs tegra_plane_funcs = {
.update_plane = drm_atomic_helper_update_plane,
.disable_plane = drm_atomic_helper_disable_plane,
.destroy = tegra_plane_destroy,
.reset = tegra_plane_reset,
.atomic_duplicate_state = tegra_plane_atomic_duplicate_state,
.atomic_destroy_state = tegra_plane_atomic_destroy_state,
.format_mod_supported = tegra_plane_format_mod_supported,
};
static int tegra_dc_pin(struct tegra_dc *dc, struct tegra_plane_state *state)
{
struct iommu_domain *domain = iommu_get_domain_for_dev(dc->dev);
unsigned int i;
int err;
for (i = 0; i < state->base.fb->format->num_planes; i++) {
struct tegra_bo *bo = tegra_fb_get_plane(state->base.fb, i);
dma_addr_t phys_addr, *phys;
struct sg_table *sgt;
if (!domain || dc->client.group)
phys = &phys_addr;
else
phys = NULL;
sgt = host1x_bo_pin(dc->dev, &bo->base, phys);
if (IS_ERR(sgt)) {
err = PTR_ERR(sgt);
goto unpin;
}
if (sgt) {
err = dma_map_sg(dc->dev, sgt->sgl, sgt->nents,
DMA_TO_DEVICE);
if (err == 0) {
err = -ENOMEM;
goto unpin;
}
/*
* The display controller needs contiguous memory, so
* fail if the buffer is discontiguous and we fail to
* map its SG table to a single contiguous chunk of
* I/O virtual memory.
*/
if (err > 1) {
err = -EINVAL;
goto unpin;
}
state->iova[i] = sg_dma_address(sgt->sgl);
state->sgt[i] = sgt;
} else {
state->iova[i] = phys_addr;
}
}
return 0;
unpin:
dev_err(dc->dev, "failed to map plane %u: %d\n", i, err);
while (i--) {
struct tegra_bo *bo = tegra_fb_get_plane(state->base.fb, i);
struct sg_table *sgt = state->sgt[i];
if (sgt)
dma_unmap_sg(dc->dev, sgt->sgl, sgt->nents,
DMA_TO_DEVICE);
host1x_bo_unpin(dc->dev, &bo->base, sgt);
state->iova[i] = DMA_MAPPING_ERROR;
state->sgt[i] = NULL;
}
return err;
}
static void tegra_dc_unpin(struct tegra_dc *dc, struct tegra_plane_state *state)
{
unsigned int i;
for (i = 0; i < state->base.fb->format->num_planes; i++) {
struct tegra_bo *bo = tegra_fb_get_plane(state->base.fb, i);
struct sg_table *sgt = state->sgt[i];
if (sgt)
dma_unmap_sg(dc->dev, sgt->sgl, sgt->nents,
DMA_TO_DEVICE);
host1x_bo_unpin(dc->dev, &bo->base, sgt);
state->iova[i] = DMA_MAPPING_ERROR;
state->sgt[i] = NULL;
}
}
int tegra_plane_prepare_fb(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct tegra_dc *dc = to_tegra_dc(state->crtc);
if (!state->fb)
return 0;
drm_gem_fb_prepare_fb(plane, state);
return tegra_dc_pin(dc, to_tegra_plane_state(state));
}
void tegra_plane_cleanup_fb(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct tegra_dc *dc = to_tegra_dc(state->crtc);
if (dc)
tegra_dc_unpin(dc, to_tegra_plane_state(state));
}
int tegra_plane_state_add(struct tegra_plane *plane,
struct drm_plane_state *state)
{
struct drm_crtc_state *crtc_state;
struct tegra_dc_state *tegra;
int err;
/* Propagate errors from allocation or locking failures. */
crtc_state = drm_atomic_get_crtc_state(state->state, state->crtc);
if (IS_ERR(crtc_state))
return PTR_ERR(crtc_state);
/* Check plane state for visibility and calculate clipping bounds */
err = drm_atomic_helper_check_plane_state(state, crtc_state,
0, INT_MAX, true, true);
if (err < 0)
return err;
tegra = to_dc_state(crtc_state);
tegra->planes |= WIN_A_ACT_REQ << plane->index;
return 0;
}
int tegra_plane_format(u32 fourcc, u32 *format, u32 *swap)
{
/* assume no swapping of fetched data */
if (swap)
*swap = BYTE_SWAP_NOSWAP;
switch (fourcc) {
case DRM_FORMAT_ARGB4444:
*format = WIN_COLOR_DEPTH_B4G4R4A4;
break;
case DRM_FORMAT_ARGB1555:
*format = WIN_COLOR_DEPTH_B5G5R5A1;
break;
case DRM_FORMAT_RGB565:
*format = WIN_COLOR_DEPTH_B5G6R5;
break;
case DRM_FORMAT_RGBA5551:
*format = WIN_COLOR_DEPTH_A1B5G5R5;
break;
case DRM_FORMAT_ARGB8888:
*format = WIN_COLOR_DEPTH_B8G8R8A8;
break;
case DRM_FORMAT_ABGR8888:
*format = WIN_COLOR_DEPTH_R8G8B8A8;
break;
case DRM_FORMAT_ABGR4444:
*format = WIN_COLOR_DEPTH_R4G4B4A4;
break;
case DRM_FORMAT_ABGR1555:
*format = WIN_COLOR_DEPTH_R5G5B5A;
break;
case DRM_FORMAT_BGRA5551:
*format = WIN_COLOR_DEPTH_AR5G5B5;
break;
case DRM_FORMAT_XRGB1555:
*format = WIN_COLOR_DEPTH_B5G5R5X1;
break;
case DRM_FORMAT_RGBX5551:
*format = WIN_COLOR_DEPTH_X1B5G5R5;
break;
case DRM_FORMAT_XBGR1555:
*format = WIN_COLOR_DEPTH_R5G5B5X1;
break;
case DRM_FORMAT_BGRX5551:
*format = WIN_COLOR_DEPTH_X1R5G5B5;
break;
case DRM_FORMAT_BGR565:
*format = WIN_COLOR_DEPTH_R5G6B5;
break;
case DRM_FORMAT_BGRA8888:
*format = WIN_COLOR_DEPTH_A8R8G8B8;
break;
case DRM_FORMAT_RGBA8888:
*format = WIN_COLOR_DEPTH_A8B8G8R8;
break;
case DRM_FORMAT_XRGB8888:
*format = WIN_COLOR_DEPTH_B8G8R8X8;
break;
case DRM_FORMAT_XBGR8888:
*format = WIN_COLOR_DEPTH_R8G8B8X8;
break;
case DRM_FORMAT_UYVY:
*format = WIN_COLOR_DEPTH_YCbCr422;
break;
case DRM_FORMAT_YUYV:
if (!swap)
return -EINVAL;
*format = WIN_COLOR_DEPTH_YCbCr422;
*swap = BYTE_SWAP_SWAP2;
break;
case DRM_FORMAT_YUV420:
*format = WIN_COLOR_DEPTH_YCbCr420P;
break;
case DRM_FORMAT_YUV422:
*format = WIN_COLOR_DEPTH_YCbCr422P;
break;
default:
return -EINVAL;
}
return 0;
}
bool tegra_plane_format_is_yuv(unsigned int format, bool *planar)
{
switch (format) {
case WIN_COLOR_DEPTH_YCbCr422:
case WIN_COLOR_DEPTH_YUV422:
if (planar)
*planar = false;
return true;
case WIN_COLOR_DEPTH_YCbCr420P:
case WIN_COLOR_DEPTH_YUV420P:
case WIN_COLOR_DEPTH_YCbCr422P:
case WIN_COLOR_DEPTH_YUV422P:
case WIN_COLOR_DEPTH_YCbCr422R:
case WIN_COLOR_DEPTH_YUV422R:
case WIN_COLOR_DEPTH_YCbCr422RA:
case WIN_COLOR_DEPTH_YUV422RA:
if (planar)
*planar = true;
return true;
}
if (planar)
*planar = false;
return false;
}
static bool __drm_format_has_alpha(u32 format)
{
switch (format) {
case DRM_FORMAT_ARGB1555:
case DRM_FORMAT_RGBA5551:
case DRM_FORMAT_ABGR8888:
case DRM_FORMAT_ARGB8888:
return true;
}
return false;
}
static int tegra_plane_format_get_alpha(unsigned int opaque,
unsigned int *alpha)
{
if (tegra_plane_format_is_yuv(opaque, NULL)) {
*alpha = opaque;
return 0;
}
switch (opaque) {
case WIN_COLOR_DEPTH_B5G5R5X1:
*alpha = WIN_COLOR_DEPTH_B5G5R5A1;
return 0;
case WIN_COLOR_DEPTH_X1B5G5R5:
*alpha = WIN_COLOR_DEPTH_A1B5G5R5;
return 0;
case WIN_COLOR_DEPTH_R8G8B8X8:
*alpha = WIN_COLOR_DEPTH_R8G8B8A8;
return 0;
case WIN_COLOR_DEPTH_B8G8R8X8:
*alpha = WIN_COLOR_DEPTH_B8G8R8A8;
return 0;
case WIN_COLOR_DEPTH_B5G6R5:
*alpha = opaque;
return 0;
}
return -EINVAL;
}
/*
* This is applicable to Tegra20 and Tegra30 only where the opaque formats can
* be emulated using the alpha formats and alpha blending disabled.
*/
static int tegra_plane_setup_opacity(struct tegra_plane *tegra,
struct tegra_plane_state *state)
{
unsigned int format;
int err;
switch (state->format) {
case WIN_COLOR_DEPTH_B5G5R5A1:
case WIN_COLOR_DEPTH_A1B5G5R5:
case WIN_COLOR_DEPTH_R8G8B8A8:
case WIN_COLOR_DEPTH_B8G8R8A8:
state->opaque = false;
break;
default:
err = tegra_plane_format_get_alpha(state->format, &format);
if (err < 0)
return err;
state->format = format;
state->opaque = true;
break;
}
return 0;
}
static int tegra_plane_check_transparency(struct tegra_plane *tegra,
struct tegra_plane_state *state)
{
struct drm_plane_state *old, *plane_state;
struct drm_plane *plane;
old = drm_atomic_get_old_plane_state(state->base.state, &tegra->base);
/* check if zpos / transparency changed */
if (old->normalized_zpos == state->base.normalized_zpos &&
to_tegra_plane_state(old)->opaque == state->opaque)
return 0;
/* include all sibling planes into this commit */
drm_for_each_plane(plane, tegra->base.dev) {
struct tegra_plane *p = to_tegra_plane(plane);
/* skip this plane and planes on different CRTCs */
if (p == tegra || p->dc != tegra->dc)
continue;
plane_state = drm_atomic_get_plane_state(state->base.state,
plane);
if (IS_ERR(plane_state))
return PTR_ERR(plane_state);
}
return 1;
}
static unsigned int tegra_plane_get_overlap_index(struct tegra_plane *plane,
struct tegra_plane *other)
{
unsigned int index = 0, i;
WARN_ON(plane == other);
for (i = 0; i < 3; i++) {
if (i == plane->index)
continue;
if (i == other->index)
break;
index++;
}
return index;
}
static void tegra_plane_update_transparency(struct tegra_plane *tegra,
struct tegra_plane_state *state)
{
struct drm_plane_state *new;
struct drm_plane *plane;
unsigned int i;
for_each_new_plane_in_state(state->base.state, plane, new, i) {
struct tegra_plane *p = to_tegra_plane(plane);
unsigned index;
/* skip this plane and planes on different CRTCs */
if (p == tegra || p->dc != tegra->dc)
continue;
index = tegra_plane_get_overlap_index(tegra, p);
if (new->fb && __drm_format_has_alpha(new->fb->format->format))
state->blending[index].alpha = true;
else
state->blending[index].alpha = false;
if (new->normalized_zpos > state->base.normalized_zpos)
state->blending[index].top = true;
else
state->blending[index].top = false;
/*
* Missing framebuffer means that plane is disabled, in this
* case mark B / C window as top to be able to differentiate
* windows indices order in regards to zPos for the middle
* window X / Y registers programming.
*/
if (!new->fb)
state->blending[index].top = (index == 1);
}
}
static int tegra_plane_setup_transparency(struct tegra_plane *tegra,
struct tegra_plane_state *state)
{
struct tegra_plane_state *tegra_state;
struct drm_plane_state *new;
struct drm_plane *plane;
int err;
/*
* If planes zpos / transparency changed, sibling planes blending
* state may require adjustment and in this case they will be included
* into this atom commit, otherwise blending state is unchanged.
*/
err = tegra_plane_check_transparency(tegra, state);
if (err <= 0)
return err;
/*
* All planes are now in the atomic state, walk them up and update
* transparency state for each plane.
*/
drm_for_each_plane(plane, tegra->base.dev) {
struct tegra_plane *p = to_tegra_plane(plane);
/* skip planes on different CRTCs */
if (p->dc != tegra->dc)
continue;
new = drm_atomic_get_new_plane_state(state->base.state, plane);
tegra_state = to_tegra_plane_state(new);
/*
* There is no need to update blending state for the disabled
* plane.
*/
if (new->fb)
tegra_plane_update_transparency(p, tegra_state);
}
return 0;
}
int tegra_plane_setup_legacy_state(struct tegra_plane *tegra,
struct tegra_plane_state *state)
{
int err;
err = tegra_plane_setup_opacity(tegra, state);
if (err < 0)
return err;
err = tegra_plane_setup_transparency(tegra, state);
if (err < 0)
return err;
return 0;
}