linux_dsm_epyc7002/drivers/gpu/drm/vc4/vc4_plane.c

1013 lines
29 KiB
C
Raw Normal View History

/*
* Copyright (C) 2015 Broadcom
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
/**
* DOC: VC4 plane module
*
* Each DRM plane is a layer of pixels being scanned out by the HVS.
*
* At atomic modeset check time, we compute the HVS display element
* state that would be necessary for displaying the plane (giving us a
* chance to figure out if a plane configuration is invalid), then at
* atomic flush time the CRTC will ask us to write our element state
* into the region of the HVS that it has allocated for us.
*/
#include <drm/drm_atomic.h>
#include <drm/drm_atomic_helper.h>
#include <drm/drm_fb_cma_helper.h>
#include <drm/drm_plane_helper.h>
#include "uapi/drm/vc4_drm.h"
#include "vc4_drv.h"
#include "vc4_regs.h"
static const struct hvs_format {
u32 drm; /* DRM_FORMAT_* */
u32 hvs; /* HVS_FORMAT_* */
u32 pixel_order;
} hvs_formats[] = {
{
.drm = DRM_FORMAT_XRGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
.pixel_order = HVS_PIXEL_ORDER_ABGR,
},
{
.drm = DRM_FORMAT_ARGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
.pixel_order = HVS_PIXEL_ORDER_ABGR,
},
{
.drm = DRM_FORMAT_ABGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
.pixel_order = HVS_PIXEL_ORDER_ARGB,
},
{
.drm = DRM_FORMAT_XBGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
.pixel_order = HVS_PIXEL_ORDER_ARGB,
},
{
.drm = DRM_FORMAT_RGB565, .hvs = HVS_PIXEL_FORMAT_RGB565,
.pixel_order = HVS_PIXEL_ORDER_XRGB,
},
{
.drm = DRM_FORMAT_BGR565, .hvs = HVS_PIXEL_FORMAT_RGB565,
.pixel_order = HVS_PIXEL_ORDER_XBGR,
},
{
.drm = DRM_FORMAT_ARGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551,
.pixel_order = HVS_PIXEL_ORDER_ABGR,
},
{
.drm = DRM_FORMAT_XRGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551,
.pixel_order = HVS_PIXEL_ORDER_ABGR,
},
{
.drm = DRM_FORMAT_RGB888, .hvs = HVS_PIXEL_FORMAT_RGB888,
.pixel_order = HVS_PIXEL_ORDER_XRGB,
},
{
.drm = DRM_FORMAT_BGR888, .hvs = HVS_PIXEL_FORMAT_RGB888,
.pixel_order = HVS_PIXEL_ORDER_XBGR,
},
{
.drm = DRM_FORMAT_YUV422,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
},
{
.drm = DRM_FORMAT_YVU422,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
},
{
.drm = DRM_FORMAT_YUV420,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
},
{
.drm = DRM_FORMAT_YVU420,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
},
{
.drm = DRM_FORMAT_NV12,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
},
{
.drm = DRM_FORMAT_NV21,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
},
{
.drm = DRM_FORMAT_NV16,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
},
{
.drm = DRM_FORMAT_NV61,
.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE,
.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
},
};
static const struct hvs_format *vc4_get_hvs_format(u32 drm_format)
{
unsigned i;
for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) {
if (hvs_formats[i].drm == drm_format)
return &hvs_formats[i];
}
return NULL;
}
static enum vc4_scaling_mode vc4_get_scaling_mode(u32 src, u32 dst)
{
if (dst > src)
return VC4_SCALING_PPF;
else if (dst < src)
return VC4_SCALING_TPZ;
else
return VC4_SCALING_NONE;
}
static bool plane_enabled(struct drm_plane_state *state)
{
return state->fb && state->crtc;
}
static struct drm_plane_state *vc4_plane_duplicate_state(struct drm_plane *plane)
{
struct vc4_plane_state *vc4_state;
if (WARN_ON(!plane->state))
return NULL;
vc4_state = kmemdup(plane->state, sizeof(*vc4_state), GFP_KERNEL);
if (!vc4_state)
return NULL;
memset(&vc4_state->lbm, 0, sizeof(vc4_state->lbm));
__drm_atomic_helper_plane_duplicate_state(plane, &vc4_state->base);
if (vc4_state->dlist) {
vc4_state->dlist = kmemdup(vc4_state->dlist,
vc4_state->dlist_count * 4,
GFP_KERNEL);
if (!vc4_state->dlist) {
kfree(vc4_state);
return NULL;
}
vc4_state->dlist_size = vc4_state->dlist_count;
}
return &vc4_state->base;
}
static void vc4_plane_destroy_state(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct vc4_dev *vc4 = to_vc4_dev(plane->dev);
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
if (vc4_state->lbm.allocated) {
unsigned long irqflags;
spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags);
drm_mm_remove_node(&vc4_state->lbm);
spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags);
}
kfree(vc4_state->dlist);
__drm_atomic_helper_plane_destroy_state(&vc4_state->base);
kfree(state);
}
/* Called during init to allocate the plane's atomic state. */
static void vc4_plane_reset(struct drm_plane *plane)
{
struct vc4_plane_state *vc4_state;
WARN_ON(plane->state);
vc4_state = kzalloc(sizeof(*vc4_state), GFP_KERNEL);
if (!vc4_state)
return;
plane->state = &vc4_state->base;
plane->state->alpha = DRM_BLEND_ALPHA_OPAQUE;
vc4_state->base.plane = plane;
}
static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val)
{
if (vc4_state->dlist_count == vc4_state->dlist_size) {
u32 new_size = max(4u, vc4_state->dlist_count * 2);
treewide: kmalloc() -> kmalloc_array() The kmalloc() function has a 2-factor argument form, kmalloc_array(). This patch replaces cases of: kmalloc(a * b, gfp) with: kmalloc_array(a * b, gfp) as well as handling cases of: kmalloc(a * b * c, gfp) with: kmalloc(array3_size(a, b, c), gfp) as it's slightly less ugly than: kmalloc_array(array_size(a, b), c, gfp) This does, however, attempt to ignore constant size factors like: kmalloc(4 * 1024, gfp) though any constants defined via macros get caught up in the conversion. Any factors with a sizeof() of "unsigned char", "char", and "u8" were dropped, since they're redundant. The tools/ directory was manually excluded, since it has its own implementation of kmalloc(). The Coccinelle script used for this was: // Fix redundant parens around sizeof(). @@ type TYPE; expression THING, E; @@ ( kmalloc( - (sizeof(TYPE)) * E + sizeof(TYPE) * E , ...) | kmalloc( - (sizeof(THING)) * E + sizeof(THING) * E , ...) ) // Drop single-byte sizes and redundant parens. @@ expression COUNT; typedef u8; typedef __u8; @@ ( kmalloc( - sizeof(u8) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(__u8) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(char) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(unsigned char) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(u8) * COUNT + COUNT , ...) | kmalloc( - sizeof(__u8) * COUNT + COUNT , ...) | kmalloc( - sizeof(char) * COUNT + COUNT , ...) | kmalloc( - sizeof(unsigned char) * COUNT + COUNT , ...) ) // 2-factor product with sizeof(type/expression) and identifier or constant. @@ type TYPE; expression THING; identifier COUNT_ID; constant COUNT_CONST; @@ ( - kmalloc + kmalloc_array ( - sizeof(TYPE) * (COUNT_ID) + COUNT_ID, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * COUNT_ID + COUNT_ID, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * (COUNT_CONST) + COUNT_CONST, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * COUNT_CONST + COUNT_CONST, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * (COUNT_ID) + COUNT_ID, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * COUNT_ID + COUNT_ID, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * (COUNT_CONST) + COUNT_CONST, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * COUNT_CONST + COUNT_CONST, sizeof(THING) , ...) ) // 2-factor product, only identifiers. @@ identifier SIZE, COUNT; @@ - kmalloc + kmalloc_array ( - SIZE * COUNT + COUNT, SIZE , ...) // 3-factor product with 1 sizeof(type) or sizeof(expression), with // redundant parens removed. @@ expression THING; identifier STRIDE, COUNT; type TYPE; @@ ( kmalloc( - sizeof(TYPE) * (COUNT) * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(TYPE) * (COUNT) * STRIDE + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(TYPE) * COUNT * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(TYPE) * COUNT * STRIDE + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(THING) * (COUNT) * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | kmalloc( - sizeof(THING) * (COUNT) * STRIDE + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | kmalloc( - sizeof(THING) * COUNT * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | kmalloc( - sizeof(THING) * COUNT * STRIDE + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) ) // 3-factor product with 2 sizeof(variable), with redundant parens removed. @@ expression THING1, THING2; identifier COUNT; type TYPE1, TYPE2; @@ ( kmalloc( - sizeof(TYPE1) * sizeof(TYPE2) * COUNT + array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2)) , ...) | kmalloc( - sizeof(TYPE1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2)) , ...) | kmalloc( - sizeof(THING1) * sizeof(THING2) * COUNT + array3_size(COUNT, sizeof(THING1), sizeof(THING2)) , ...) | kmalloc( - sizeof(THING1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(THING1), sizeof(THING2)) , ...) | kmalloc( - sizeof(TYPE1) * sizeof(THING2) * COUNT + array3_size(COUNT, sizeof(TYPE1), sizeof(THING2)) , ...) | kmalloc( - sizeof(TYPE1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(TYPE1), sizeof(THING2)) , ...) ) // 3-factor product, only identifiers, with redundant parens removed. @@ identifier STRIDE, SIZE, COUNT; @@ ( kmalloc( - (COUNT) * STRIDE * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * (STRIDE) * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * STRIDE * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - (COUNT) * (STRIDE) * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * (STRIDE) * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - (COUNT) * STRIDE * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - (COUNT) * (STRIDE) * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * STRIDE * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) ) // Any remaining multi-factor products, first at least 3-factor products, // when they're not all constants... @@ expression E1, E2, E3; constant C1, C2, C3; @@ ( kmalloc(C1 * C2 * C3, ...) | kmalloc( - (E1) * E2 * E3 + array3_size(E1, E2, E3) , ...) | kmalloc( - (E1) * (E2) * E3 + array3_size(E1, E2, E3) , ...) | kmalloc( - (E1) * (E2) * (E3) + array3_size(E1, E2, E3) , ...) | kmalloc( - E1 * E2 * E3 + array3_size(E1, E2, E3) , ...) ) // And then all remaining 2 factors products when they're not all constants, // keeping sizeof() as the second factor argument. @@ expression THING, E1, E2; type TYPE; constant C1, C2, C3; @@ ( kmalloc(sizeof(THING) * C2, ...) | kmalloc(sizeof(TYPE) * C2, ...) | kmalloc(C1 * C2 * C3, ...) | kmalloc(C1 * C2, ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * (E2) + E2, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * E2 + E2, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * (E2) + E2, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * E2 + E2, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - (E1) * E2 + E1, E2 , ...) | - kmalloc + kmalloc_array ( - (E1) * (E2) + E1, E2 , ...) | - kmalloc + kmalloc_array ( - E1 * E2 + E1, E2 , ...) ) Signed-off-by: Kees Cook <keescook@chromium.org>
2018-06-13 03:55:00 +07:00
u32 *new_dlist = kmalloc_array(new_size, 4, GFP_KERNEL);
if (!new_dlist)
return;
memcpy(new_dlist, vc4_state->dlist, vc4_state->dlist_count * 4);
kfree(vc4_state->dlist);
vc4_state->dlist = new_dlist;
vc4_state->dlist_size = new_size;
}
vc4_state->dlist[vc4_state->dlist_count++] = val;
}
/* Returns the scl0/scl1 field based on whether the dimensions need to
* be up/down/non-scaled.
*
* This is a replication of a table from the spec.
*/
static u32 vc4_get_scl_field(struct drm_plane_state *state, int plane)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
switch (vc4_state->x_scaling[plane] << 2 | vc4_state->y_scaling[plane]) {
case VC4_SCALING_PPF << 2 | VC4_SCALING_PPF:
return SCALER_CTL0_SCL_H_PPF_V_PPF;
case VC4_SCALING_TPZ << 2 | VC4_SCALING_PPF:
return SCALER_CTL0_SCL_H_TPZ_V_PPF;
case VC4_SCALING_PPF << 2 | VC4_SCALING_TPZ:
return SCALER_CTL0_SCL_H_PPF_V_TPZ;
case VC4_SCALING_TPZ << 2 | VC4_SCALING_TPZ:
return SCALER_CTL0_SCL_H_TPZ_V_TPZ;
case VC4_SCALING_PPF << 2 | VC4_SCALING_NONE:
return SCALER_CTL0_SCL_H_PPF_V_NONE;
case VC4_SCALING_NONE << 2 | VC4_SCALING_PPF:
return SCALER_CTL0_SCL_H_NONE_V_PPF;
case VC4_SCALING_NONE << 2 | VC4_SCALING_TPZ:
return SCALER_CTL0_SCL_H_NONE_V_TPZ;
case VC4_SCALING_TPZ << 2 | VC4_SCALING_NONE:
return SCALER_CTL0_SCL_H_TPZ_V_NONE;
default:
case VC4_SCALING_NONE << 2 | VC4_SCALING_NONE:
/* The unity case is independently handled by
* SCALER_CTL0_UNITY.
*/
return 0;
}
}
static int vc4_plane_setup_clipping_and_scaling(struct drm_plane_state *state)
{
struct drm_plane *plane = state->plane;
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
struct drm_framebuffer *fb = state->fb;
struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0);
u32 subpixel_src_mask = (1 << 16) - 1;
drm: Nuke fb->pixel_format Replace uses of fb->pixel_format with fb->format->format. Less duplicated information is a good thing. Note that coccinelle failed to eliminate the "/* fourcc format */" comment from drm_framebuffer.h, so I had to do that part manually. @@ struct drm_framebuffer *FB; expression E; @@ drm_helper_mode_fill_fb_struct(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *FB; expression E; @@ i9xx_get_initial_plane_config(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *FB; expression E; @@ ironlake_get_initial_plane_config(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *FB; expression E; @@ skylake_get_initial_plane_config(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *a; struct drm_framebuffer b; @@ ( - a->pixel_format + a->format->format | - b.pixel_format + b.format->format ) @@ struct drm_plane_state *a; struct drm_plane_state b; @@ ( - a->fb->pixel_format + a->fb->format->format | - b.fb->pixel_format + b.fb->format->format ) @@ struct drm_crtc *CRTC; @@ ( - CRTC->primary->fb->pixel_format + CRTC->primary->fb->format->format | - CRTC->primary->state->fb->pixel_format + CRTC->primary->state->fb->format->format ) @@ struct drm_mode_set *set; @@ ( - set->fb->pixel_format + set->fb->format->format | - set->crtc->primary->fb->pixel_format + set->crtc->primary->fb->format->format ) @@ @@ struct drm_framebuffer { ... - uint32_t pixel_format; ... }; v2: Fix commit message (Laurent) Rebase due to earlier removal of many fb->pixel_format uses, including the 'fb->format = drm_format_info(fb->format->format);' snafu v3: Adjusted the semantic patch a bit and regenerated due to code changes Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: http://patchwork.freedesktop.org/patch/msgid/1481751175-18463-1-git-send-email-ville.syrjala@linux.intel.com
2016-12-15 04:32:55 +07:00
u32 format = fb->format->format;
int num_planes = fb->format->num_planes;
u32 h_subsample = 1;
u32 v_subsample = 1;
int i;
for (i = 0; i < num_planes; i++)
vc4_state->offsets[i] = bo->paddr + fb->offsets[i];
/* We don't support subpixel source positioning for scaling. */
if ((state->src_x & subpixel_src_mask) ||
(state->src_y & subpixel_src_mask) ||
(state->src_w & subpixel_src_mask) ||
(state->src_h & subpixel_src_mask)) {
return -EINVAL;
}
vc4_state->src_x = state->src_x >> 16;
vc4_state->src_y = state->src_y >> 16;
vc4_state->src_w[0] = state->src_w >> 16;
vc4_state->src_h[0] = state->src_h >> 16;
vc4_state->crtc_x = state->crtc_x;
vc4_state->crtc_y = state->crtc_y;
vc4_state->crtc_w = state->crtc_w;
vc4_state->crtc_h = state->crtc_h;
vc4_state->x_scaling[0] = vc4_get_scaling_mode(vc4_state->src_w[0],
vc4_state->crtc_w);
vc4_state->y_scaling[0] = vc4_get_scaling_mode(vc4_state->src_h[0],
vc4_state->crtc_h);
if (num_planes > 1) {
vc4_state->is_yuv = true;
h_subsample = drm_format_horz_chroma_subsampling(format);
v_subsample = drm_format_vert_chroma_subsampling(format);
vc4_state->src_w[1] = vc4_state->src_w[0] / h_subsample;
vc4_state->src_h[1] = vc4_state->src_h[0] / v_subsample;
vc4_state->x_scaling[1] =
vc4_get_scaling_mode(vc4_state->src_w[1],
vc4_state->crtc_w);
vc4_state->y_scaling[1] =
vc4_get_scaling_mode(vc4_state->src_h[1],
vc4_state->crtc_h);
/* YUV conversion requires that scaling be enabled,
* even on a plane that's otherwise 1:1. Choose TPZ
* for simplicity.
*/
if (vc4_state->x_scaling[0] == VC4_SCALING_NONE)
vc4_state->x_scaling[0] = VC4_SCALING_TPZ;
if (vc4_state->y_scaling[0] == VC4_SCALING_NONE)
vc4_state->y_scaling[0] = VC4_SCALING_TPZ;
} else {
vc4_state->x_scaling[1] = VC4_SCALING_NONE;
vc4_state->y_scaling[1] = VC4_SCALING_NONE;
}
vc4_state->is_unity = (vc4_state->x_scaling[0] == VC4_SCALING_NONE &&
vc4_state->y_scaling[0] == VC4_SCALING_NONE &&
vc4_state->x_scaling[1] == VC4_SCALING_NONE &&
vc4_state->y_scaling[1] == VC4_SCALING_NONE);
/* No configuring scaling on the cursor plane, since it gets
non-vblank-synced updates, and scaling requires requires
LBM changes which have to be vblank-synced.
*/
if (plane->type == DRM_PLANE_TYPE_CURSOR && !vc4_state->is_unity)
return -EINVAL;
/* Clamp the on-screen start x/y to 0. The hardware doesn't
* support negative y, and negative x wastes bandwidth.
*/
if (vc4_state->crtc_x < 0) {
for (i = 0; i < num_planes; i++) {
u32 cpp = fb->format->cpp[i];
u32 subs = ((i == 0) ? 1 : h_subsample);
vc4_state->offsets[i] += (cpp *
(-vc4_state->crtc_x) / subs);
}
vc4_state->src_w[0] += vc4_state->crtc_x;
vc4_state->src_w[1] += vc4_state->crtc_x / h_subsample;
vc4_state->crtc_x = 0;
}
if (vc4_state->crtc_y < 0) {
for (i = 0; i < num_planes; i++) {
u32 subs = ((i == 0) ? 1 : v_subsample);
vc4_state->offsets[i] += (fb->pitches[i] *
(-vc4_state->crtc_y) / subs);
}
vc4_state->src_h[0] += vc4_state->crtc_y;
vc4_state->src_h[1] += vc4_state->crtc_y / v_subsample;
vc4_state->crtc_y = 0;
}
return 0;
}
static void vc4_write_tpz(struct vc4_plane_state *vc4_state, u32 src, u32 dst)
{
u32 scale, recip;
scale = (1 << 16) * src / dst;
/* The specs note that while the reciprocal would be defined
* as (1<<32)/scale, ~0 is close enough.
*/
recip = ~0 / scale;
vc4_dlist_write(vc4_state,
VC4_SET_FIELD(scale, SCALER_TPZ0_SCALE) |
VC4_SET_FIELD(0, SCALER_TPZ0_IPHASE));
vc4_dlist_write(vc4_state,
VC4_SET_FIELD(recip, SCALER_TPZ1_RECIP));
}
static void vc4_write_ppf(struct vc4_plane_state *vc4_state, u32 src, u32 dst)
{
u32 scale = (1 << 16) * src / dst;
vc4_dlist_write(vc4_state,
SCALER_PPF_AGC |
VC4_SET_FIELD(scale, SCALER_PPF_SCALE) |
VC4_SET_FIELD(0, SCALER_PPF_IPHASE));
}
static u32 vc4_lbm_size(struct drm_plane_state *state)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
/* This is the worst case number. One of the two sizes will
* be used depending on the scaling configuration.
*/
u32 pix_per_line = max(vc4_state->src_w[0], (u32)vc4_state->crtc_w);
u32 lbm;
if (!vc4_state->is_yuv) {
if (vc4_state->is_unity)
return 0;
else if (vc4_state->y_scaling[0] == VC4_SCALING_TPZ)
lbm = pix_per_line * 8;
else {
/* In special cases, this multiplier might be 12. */
lbm = pix_per_line * 16;
}
} else {
/* There are cases for this going down to a multiplier
* of 2, but according to the firmware source, the
* table in the docs is somewhat wrong.
*/
lbm = pix_per_line * 16;
}
lbm = roundup(lbm, 32);
return lbm;
}
static void vc4_write_scaling_parameters(struct drm_plane_state *state,
int channel)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
/* Ch0 H-PPF Word 0: Scaling Parameters */
if (vc4_state->x_scaling[channel] == VC4_SCALING_PPF) {
vc4_write_ppf(vc4_state,
vc4_state->src_w[channel], vc4_state->crtc_w);
}
/* Ch0 V-PPF Words 0-1: Scaling Parameters, Context */
if (vc4_state->y_scaling[channel] == VC4_SCALING_PPF) {
vc4_write_ppf(vc4_state,
vc4_state->src_h[channel], vc4_state->crtc_h);
vc4_dlist_write(vc4_state, 0xc0c0c0c0);
}
/* Ch0 H-TPZ Words 0-1: Scaling Parameters, Recip */
if (vc4_state->x_scaling[channel] == VC4_SCALING_TPZ) {
vc4_write_tpz(vc4_state,
vc4_state->src_w[channel], vc4_state->crtc_w);
}
/* Ch0 V-TPZ Words 0-2: Scaling Parameters, Recip, Context */
if (vc4_state->y_scaling[channel] == VC4_SCALING_TPZ) {
vc4_write_tpz(vc4_state,
vc4_state->src_h[channel], vc4_state->crtc_h);
vc4_dlist_write(vc4_state, 0xc0c0c0c0);
}
}
/* Writes out a full display list for an active plane to the plane's
* private dlist state.
*/
static int vc4_plane_mode_set(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct vc4_dev *vc4 = to_vc4_dev(plane->dev);
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
struct drm_framebuffer *fb = state->fb;
u32 ctl0_offset = vc4_state->dlist_count;
drm: Nuke fb->pixel_format Replace uses of fb->pixel_format with fb->format->format. Less duplicated information is a good thing. Note that coccinelle failed to eliminate the "/* fourcc format */" comment from drm_framebuffer.h, so I had to do that part manually. @@ struct drm_framebuffer *FB; expression E; @@ drm_helper_mode_fill_fb_struct(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *FB; expression E; @@ i9xx_get_initial_plane_config(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *FB; expression E; @@ ironlake_get_initial_plane_config(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *FB; expression E; @@ skylake_get_initial_plane_config(...) { ... - FB->pixel_format = E; ... } @@ struct drm_framebuffer *a; struct drm_framebuffer b; @@ ( - a->pixel_format + a->format->format | - b.pixel_format + b.format->format ) @@ struct drm_plane_state *a; struct drm_plane_state b; @@ ( - a->fb->pixel_format + a->fb->format->format | - b.fb->pixel_format + b.fb->format->format ) @@ struct drm_crtc *CRTC; @@ ( - CRTC->primary->fb->pixel_format + CRTC->primary->fb->format->format | - CRTC->primary->state->fb->pixel_format + CRTC->primary->state->fb->format->format ) @@ struct drm_mode_set *set; @@ ( - set->fb->pixel_format + set->fb->format->format | - set->crtc->primary->fb->pixel_format + set->crtc->primary->fb->format->format ) @@ @@ struct drm_framebuffer { ... - uint32_t pixel_format; ... }; v2: Fix commit message (Laurent) Rebase due to earlier removal of many fb->pixel_format uses, including the 'fb->format = drm_format_info(fb->format->format);' snafu v3: Adjusted the semantic patch a bit and regenerated due to code changes Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: http://patchwork.freedesktop.org/patch/msgid/1481751175-18463-1-git-send-email-ville.syrjala@linux.intel.com
2016-12-15 04:32:55 +07:00
const struct hvs_format *format = vc4_get_hvs_format(fb->format->format);
u64 base_format_mod = fourcc_mod_broadcom_mod(fb->modifier);
int num_planes = drm_format_num_planes(format->drm);
bool mix_plane_alpha;
bool covers_screen;
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
u32 scl0, scl1, pitch0;
u32 lbm_size, tiling;
unsigned long irqflags;
u32 hvs_format = format->hvs;
int ret, i;
ret = vc4_plane_setup_clipping_and_scaling(state);
if (ret)
return ret;
/* Allocate the LBM memory that the HVS will use for temporary
* storage due to our scaling/format conversion.
*/
lbm_size = vc4_lbm_size(state);
if (lbm_size) {
if (!vc4_state->lbm.allocated) {
spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags);
drm: Improve drm_mm search (and fix topdown allocation) with rbtrees The drm_mm range manager claimed to support top-down insertion, but it was neither searching for the top-most hole that could fit the allocation request nor fitting the request to the hole correctly. In order to search the range efficiently, we create a secondary index for the holes using either their size or their address. This index allows us to find the smallest hole or the hole at the bottom or top of the range efficiently, whilst keeping the hole stack to rapidly service evictions. v2: Search for holes both high and low. Rename flags to mode. v3: Discover rb_entry_safe() and use it! v4: Kerneldoc for enum drm_mm_insert_mode. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: "Christian König" <christian.koenig@amd.com> Cc: David Airlie <airlied@linux.ie> Cc: Russell King <rmk+kernel@armlinux.org.uk> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Sean Paul <seanpaul@chromium.org> Cc: Lucas Stach <l.stach@pengutronix.de> Cc: Christian Gmeiner <christian.gmeiner@gmail.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Thierry Reding <thierry.reding@gmail.com> Cc: Stephen Warren <swarren@wwwdotorg.org> Cc: Alexandre Courbot <gnurou@gmail.com> Cc: Eric Anholt <eric@anholt.net> Cc: Sinclair Yeh <syeh@vmware.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Sinclair Yeh <syeh@vmware.com> # vmwgfx Reviewed-by: Lucas Stach <l.stach@pengutronix.de> #etnaviv Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20170202210438.28702-1-chris@chris-wilson.co.uk
2017-02-03 04:04:38 +07:00
ret = drm_mm_insert_node_generic(&vc4->hvs->lbm_mm,
&vc4_state->lbm,
lbm_size, 32, 0, 0);
spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags);
} else {
WARN_ON_ONCE(lbm_size != vc4_state->lbm.size);
}
}
if (ret)
return ret;
/* SCL1 is used for Cb/Cr scaling of planar formats. For RGB
* and 4:4:4, scl1 should be set to scl0 so both channels of
* the scaler do the same thing. For YUV, the Y plane needs
* to be put in channel 1 and Cb/Cr in channel 0, so we swap
* the scl fields here.
*/
if (num_planes == 1) {
scl0 = vc4_get_scl_field(state, 0);
scl1 = scl0;
} else {
scl0 = vc4_get_scl_field(state, 1);
scl1 = vc4_get_scl_field(state, 0);
}
switch (base_format_mod) {
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
case DRM_FORMAT_MOD_LINEAR:
tiling = SCALER_CTL0_TILING_LINEAR;
pitch0 = VC4_SET_FIELD(fb->pitches[0], SCALER_SRC_PITCH);
break;
case DRM_FORMAT_MOD_BROADCOM_VC4_T_TILED: {
/* For T-tiled, the FB pitch is "how many bytes from
* one row to the next, such that pitch * tile_h ==
* tile_size * tiles_per_row."
*/
u32 tile_size_shift = 12; /* T tiles are 4kb */
u32 tile_h_shift = 5; /* 16 and 32bpp are 32 pixels high */
u32 tiles_w = fb->pitches[0] >> (tile_size_shift - tile_h_shift);
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
tiling = SCALER_CTL0_TILING_256B_OR_T;
pitch0 = (VC4_SET_FIELD(0, SCALER_PITCH0_TILE_Y_OFFSET) |
VC4_SET_FIELD(0, SCALER_PITCH0_TILE_WIDTH_L) |
VC4_SET_FIELD(tiles_w, SCALER_PITCH0_TILE_WIDTH_R));
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
break;
}
case DRM_FORMAT_MOD_BROADCOM_SAND64:
case DRM_FORMAT_MOD_BROADCOM_SAND128:
case DRM_FORMAT_MOD_BROADCOM_SAND256: {
uint32_t param = fourcc_mod_broadcom_param(fb->modifier);
/* Column-based NV12 or RGBA.
*/
if (fb->format->num_planes > 1) {
if (hvs_format != HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE) {
DRM_DEBUG_KMS("SAND format only valid for NV12/21");
return -EINVAL;
}
hvs_format = HVS_PIXEL_FORMAT_H264;
} else {
if (base_format_mod == DRM_FORMAT_MOD_BROADCOM_SAND256) {
DRM_DEBUG_KMS("SAND256 format only valid for H.264");
return -EINVAL;
}
}
switch (base_format_mod) {
case DRM_FORMAT_MOD_BROADCOM_SAND64:
tiling = SCALER_CTL0_TILING_64B;
break;
case DRM_FORMAT_MOD_BROADCOM_SAND128:
tiling = SCALER_CTL0_TILING_128B;
break;
case DRM_FORMAT_MOD_BROADCOM_SAND256:
tiling = SCALER_CTL0_TILING_256B_OR_T;
break;
default:
break;
}
if (param > SCALER_TILE_HEIGHT_MASK) {
DRM_DEBUG_KMS("SAND height too large (%d)\n", param);
return -EINVAL;
}
pitch0 = VC4_SET_FIELD(param, SCALER_TILE_HEIGHT);
break;
}
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
default:
DRM_DEBUG_KMS("Unsupported FB tiling flag 0x%16llx",
(long long)fb->modifier);
return -EINVAL;
}
/* Control word */
vc4_dlist_write(vc4_state,
SCALER_CTL0_VALID |
VC4_SET_FIELD(SCALER_CTL0_RGBA_EXPAND_ROUND, SCALER_CTL0_RGBA_EXPAND) |
(format->pixel_order << SCALER_CTL0_ORDER_SHIFT) |
(hvs_format << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
VC4_SET_FIELD(tiling, SCALER_CTL0_TILING) |
(vc4_state->is_unity ? SCALER_CTL0_UNITY : 0) |
VC4_SET_FIELD(scl0, SCALER_CTL0_SCL0) |
VC4_SET_FIELD(scl1, SCALER_CTL0_SCL1));
/* Position Word 0: Image Positions and Alpha Value */
vc4_state->pos0_offset = vc4_state->dlist_count;
vc4_dlist_write(vc4_state,
VC4_SET_FIELD(state->alpha >> 8, SCALER_POS0_FIXED_ALPHA) |
VC4_SET_FIELD(vc4_state->crtc_x, SCALER_POS0_START_X) |
VC4_SET_FIELD(vc4_state->crtc_y, SCALER_POS0_START_Y));
/* Position Word 1: Scaled Image Dimensions. */
if (!vc4_state->is_unity) {
vc4_dlist_write(vc4_state,
VC4_SET_FIELD(vc4_state->crtc_w,
SCALER_POS1_SCL_WIDTH) |
VC4_SET_FIELD(vc4_state->crtc_h,
SCALER_POS1_SCL_HEIGHT));
}
/* Don't waste cycles mixing with plane alpha if the set alpha
* is opaque or there is no per-pixel alpha information.
* In any case we use the alpha property value as the fixed alpha.
*/
mix_plane_alpha = state->alpha != DRM_BLEND_ALPHA_OPAQUE &&
fb->format->has_alpha;
/* Position Word 2: Source Image Size, Alpha */
vc4_state->pos2_offset = vc4_state->dlist_count;
vc4_dlist_write(vc4_state,
VC4_SET_FIELD(fb->format->has_alpha ?
SCALER_POS2_ALPHA_MODE_PIPELINE :
SCALER_POS2_ALPHA_MODE_FIXED,
SCALER_POS2_ALPHA_MODE) |
(mix_plane_alpha ? SCALER_POS2_ALPHA_MIX : 0) |
(fb->format->has_alpha ? SCALER_POS2_ALPHA_PREMULT : 0) |
VC4_SET_FIELD(vc4_state->src_w[0], SCALER_POS2_WIDTH) |
VC4_SET_FIELD(vc4_state->src_h[0], SCALER_POS2_HEIGHT));
/* Position Word 3: Context. Written by the HVS. */
vc4_dlist_write(vc4_state, 0xc0c0c0c0);
/* Pointer Word 0/1/2: RGB / Y / Cb / Cr Pointers
*
* The pointers may be any byte address.
*/
vc4_state->ptr0_offset = vc4_state->dlist_count;
for (i = 0; i < num_planes; i++)
vc4_dlist_write(vc4_state, vc4_state->offsets[i]);
/* Pointer Context Word 0/1/2: Written by the HVS */
for (i = 0; i < num_planes; i++)
vc4_dlist_write(vc4_state, 0xc0c0c0c0);
drm/vc4: Add T-format scanout support. The T tiling format is what V3D uses for textures, with no raster support at all until later revisions of the hardware (and always at a large 3D performance penalty). If we can't scan out V3D's format, then we often need to do a relayout at some stage of the pipeline, either right before texturing from the scanout buffer (common in X11 without a compositor) or between a tiled screen buffer right before scanout (an option I've considered in trying to resolve this inconsistency, but which means needing to use the dirty fb ioctl and having some update policy). T-format scanout lets us avoid either of those shadow copies, for a massive, obvious performance improvement to X11 window dragging without a compositor. Unfortunately, enabling a compositor to work around the discrepancy has turned out to be too costly in memory consumption for the Raspbian distribution. Because the HVS operates a scanline at a time, compositing from T does increase the memory bandwidth cost of scanout. On my 1920x1080@32bpp display on a RPi3, we go from about 15% of system memory bandwidth with linear to about 20% with tiled. However, for X11 this still ends up being a huge performance win in active usage. This patch doesn't yet handle src_x/src_y offsetting within the tiled buffer. However, we fail to do so for untiled buffers already. Signed-off-by: Eric Anholt <eric@anholt.net> Link: http://patchwork.freedesktop.org/patch/msgid/20170608001336.12842-1-eric@anholt.net Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
2017-06-08 07:13:35 +07:00
/* Pitch word 0 */
vc4_dlist_write(vc4_state, pitch0);
/* Pitch word 1/2 */
for (i = 1; i < num_planes; i++) {
if (hvs_format != HVS_PIXEL_FORMAT_H264) {
vc4_dlist_write(vc4_state,
VC4_SET_FIELD(fb->pitches[i],
SCALER_SRC_PITCH));
} else {
vc4_dlist_write(vc4_state, pitch0);
}
}
/* Colorspace conversion words */
if (vc4_state->is_yuv) {
vc4_dlist_write(vc4_state, SCALER_CSC0_ITR_R_601_5);
vc4_dlist_write(vc4_state, SCALER_CSC1_ITR_R_601_5);
vc4_dlist_write(vc4_state, SCALER_CSC2_ITR_R_601_5);
}
if (!vc4_state->is_unity) {
/* LBM Base Address. */
if (vc4_state->y_scaling[0] != VC4_SCALING_NONE ||
vc4_state->y_scaling[1] != VC4_SCALING_NONE) {
vc4_dlist_write(vc4_state, vc4_state->lbm.start);
}
if (num_planes > 1) {
/* Emit Cb/Cr as channel 0 and Y as channel
* 1. This matches how we set up scl0/scl1
* above.
*/
vc4_write_scaling_parameters(state, 1);
}
vc4_write_scaling_parameters(state, 0);
/* If any PPF setup was done, then all the kernel
* pointers get uploaded.
*/
if (vc4_state->x_scaling[0] == VC4_SCALING_PPF ||
vc4_state->y_scaling[0] == VC4_SCALING_PPF ||
vc4_state->x_scaling[1] == VC4_SCALING_PPF ||
vc4_state->y_scaling[1] == VC4_SCALING_PPF) {
u32 kernel = VC4_SET_FIELD(vc4->hvs->mitchell_netravali_filter.start,
SCALER_PPF_KERNEL_OFFSET);
/* HPPF plane 0 */
vc4_dlist_write(vc4_state, kernel);
/* VPPF plane 0 */
vc4_dlist_write(vc4_state, kernel);
/* HPPF plane 1 */
vc4_dlist_write(vc4_state, kernel);
/* VPPF plane 1 */
vc4_dlist_write(vc4_state, kernel);
}
}
vc4_state->dlist[ctl0_offset] |=
VC4_SET_FIELD(vc4_state->dlist_count, SCALER_CTL0_SIZE);
/* crtc_* are already clipped coordinates. */
covers_screen = vc4_state->crtc_x == 0 && vc4_state->crtc_y == 0 &&
vc4_state->crtc_w == state->crtc->mode.hdisplay &&
vc4_state->crtc_h == state->crtc->mode.vdisplay;
/* Background fill might be necessary when the plane has per-pixel
* alpha content or a non-opaque plane alpha and could blend from the
* background or does not cover the entire screen.
*/
vc4_state->needs_bg_fill = fb->format->has_alpha || !covers_screen ||
state->alpha != DRM_BLEND_ALPHA_OPAQUE;
return 0;
}
/* If a modeset involves changing the setup of a plane, the atomic
* infrastructure will call this to validate a proposed plane setup.
* However, if a plane isn't getting updated, this (and the
* corresponding vc4_plane_atomic_update) won't get called. Thus, we
* compute the dlist here and have all active plane dlists get updated
* in the CRTC's flush.
*/
static int vc4_plane_atomic_check(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
vc4_state->dlist_count = 0;
if (plane_enabled(state))
return vc4_plane_mode_set(plane, state);
else
return 0;
}
static void vc4_plane_atomic_update(struct drm_plane *plane,
struct drm_plane_state *old_state)
{
/* No contents here. Since we don't know where in the CRTC's
* dlist we should be stored, our dlist is uploaded to the
* hardware with vc4_plane_write_dlist() at CRTC atomic_flush
* time.
*/
}
u32 vc4_plane_write_dlist(struct drm_plane *plane, u32 __iomem *dlist)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(plane->state);
int i;
vc4_state->hw_dlist = dlist;
/* Can't memcpy_toio() because it needs to be 32-bit writes. */
for (i = 0; i < vc4_state->dlist_count; i++)
writel(vc4_state->dlist[i], &dlist[i]);
return vc4_state->dlist_count;
}
u32 vc4_plane_dlist_size(const struct drm_plane_state *state)
{
const struct vc4_plane_state *vc4_state =
container_of(state, typeof(*vc4_state), base);
return vc4_state->dlist_count;
}
/* Updates the plane to immediately (well, once the FIFO needs
* refilling) scan out from at a new framebuffer.
*/
void vc4_plane_async_set_fb(struct drm_plane *plane, struct drm_framebuffer *fb)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(plane->state);
struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0);
uint32_t addr;
/* We're skipping the address adjustment for negative origin,
* because this is only called on the primary plane.
*/
WARN_ON_ONCE(plane->state->crtc_x < 0 || plane->state->crtc_y < 0);
addr = bo->paddr + fb->offsets[0];
/* Write the new address into the hardware immediately. The
* scanout will start from this address as soon as the FIFO
* needs to refill with pixels.
*/
writel(addr, &vc4_state->hw_dlist[vc4_state->ptr0_offset]);
/* Also update the CPU-side dlist copy, so that any later
* atomic updates that don't do a new modeset on our plane
* also use our updated address.
*/
vc4_state->dlist[vc4_state->ptr0_offset] = addr;
}
static void vc4_plane_atomic_async_update(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct vc4_plane_state *vc4_state = to_vc4_plane_state(plane->state);
if (plane->state->fb != state->fb) {
vc4_plane_async_set_fb(plane, state->fb);
drm_atomic_set_fb_for_plane(plane->state, state->fb);
}
/* Set the cursor's position on the screen. This is the
* expected change from the drm_mode_cursor_universal()
* helper.
*/
plane->state->crtc_x = state->crtc_x;
plane->state->crtc_y = state->crtc_y;
/* Allow changing the start position within the cursor BO, if
* that matters.
*/
plane->state->src_x = state->src_x;
plane->state->src_y = state->src_y;
/* Update the display list based on the new crtc_x/y. */
vc4_plane_atomic_check(plane, plane->state);
/* Note that we can't just call vc4_plane_write_dlist()
* because that would smash the context data that the HVS is
* currently using.
*/
writel(vc4_state->dlist[vc4_state->pos0_offset],
&vc4_state->hw_dlist[vc4_state->pos0_offset]);
writel(vc4_state->dlist[vc4_state->pos2_offset],
&vc4_state->hw_dlist[vc4_state->pos2_offset]);
writel(vc4_state->dlist[vc4_state->ptr0_offset],
&vc4_state->hw_dlist[vc4_state->ptr0_offset]);
}
static int vc4_plane_atomic_async_check(struct drm_plane *plane,
struct drm_plane_state *state)
{
/* No configuring new scaling in the fast path. */
if (plane->state->crtc_w != state->crtc_w ||
plane->state->crtc_h != state->crtc_h ||
plane->state->src_w != state->src_w ||
plane->state->src_h != state->src_h)
return -EINVAL;
return 0;
}
static int vc4_prepare_fb(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct vc4_bo *bo;
struct dma_fence *fence;
int ret;
if (!state->fb)
return 0;
bo = to_vc4_bo(&drm_fb_cma_get_gem_obj(state->fb, 0)->base);
fence = reservation_object_get_excl_rcu(bo->resv);
drm_atomic_set_fence_for_plane(state, fence);
if (plane->state->fb == state->fb)
return 0;
ret = vc4_bo_inc_usecnt(bo);
if (ret)
return ret;
return 0;
}
static void vc4_cleanup_fb(struct drm_plane *plane,
struct drm_plane_state *state)
{
struct vc4_bo *bo;
if (plane->state->fb == state->fb || !state->fb)
return;
bo = to_vc4_bo(&drm_fb_cma_get_gem_obj(state->fb, 0)->base);
vc4_bo_dec_usecnt(bo);
}
static const struct drm_plane_helper_funcs vc4_plane_helper_funcs = {
.atomic_check = vc4_plane_atomic_check,
.atomic_update = vc4_plane_atomic_update,
.prepare_fb = vc4_prepare_fb,
.cleanup_fb = vc4_cleanup_fb,
.atomic_async_check = vc4_plane_atomic_async_check,
.atomic_async_update = vc4_plane_atomic_async_update,
};
static void vc4_plane_destroy(struct drm_plane *plane)
{
drm_plane_helper_disable(plane, NULL);
drm_plane_cleanup(plane);
}
static bool vc4_format_mod_supported(struct drm_plane *plane,
uint32_t format,
uint64_t modifier)
{
/* Support T_TILING for RGB formats only. */
switch (format) {
case DRM_FORMAT_XRGB8888:
case DRM_FORMAT_ARGB8888:
case DRM_FORMAT_ABGR8888:
case DRM_FORMAT_XBGR8888:
case DRM_FORMAT_RGB565:
case DRM_FORMAT_BGR565:
case DRM_FORMAT_ARGB1555:
case DRM_FORMAT_XRGB1555:
switch (fourcc_mod_broadcom_mod(modifier)) {
case DRM_FORMAT_MOD_LINEAR:
case DRM_FORMAT_MOD_BROADCOM_VC4_T_TILED:
case DRM_FORMAT_MOD_BROADCOM_SAND64:
case DRM_FORMAT_MOD_BROADCOM_SAND128:
return true;
default:
return false;
}
case DRM_FORMAT_NV12:
case DRM_FORMAT_NV21:
switch (fourcc_mod_broadcom_mod(modifier)) {
case DRM_FORMAT_MOD_LINEAR:
case DRM_FORMAT_MOD_BROADCOM_SAND64:
case DRM_FORMAT_MOD_BROADCOM_SAND128:
case DRM_FORMAT_MOD_BROADCOM_SAND256:
return true;
default:
return false;
}
case DRM_FORMAT_YUV422:
case DRM_FORMAT_YVU422:
case DRM_FORMAT_YUV420:
case DRM_FORMAT_YVU420:
case DRM_FORMAT_NV16:
case DRM_FORMAT_NV61:
default:
return (modifier == DRM_FORMAT_MOD_LINEAR);
}
}
static const struct drm_plane_funcs vc4_plane_funcs = {
.update_plane = drm_atomic_helper_update_plane,
.disable_plane = drm_atomic_helper_disable_plane,
.destroy = vc4_plane_destroy,
.set_property = NULL,
.reset = vc4_plane_reset,
.atomic_duplicate_state = vc4_plane_duplicate_state,
.atomic_destroy_state = vc4_plane_destroy_state,
.format_mod_supported = vc4_format_mod_supported,
};
struct drm_plane *vc4_plane_init(struct drm_device *dev,
enum drm_plane_type type)
{
struct drm_plane *plane = NULL;
struct vc4_plane *vc4_plane;
u32 formats[ARRAY_SIZE(hvs_formats)];
u32 num_formats = 0;
int ret = 0;
unsigned i;
static const uint64_t modifiers[] = {
DRM_FORMAT_MOD_BROADCOM_VC4_T_TILED,
DRM_FORMAT_MOD_BROADCOM_SAND128,
DRM_FORMAT_MOD_BROADCOM_SAND64,
DRM_FORMAT_MOD_BROADCOM_SAND256,
DRM_FORMAT_MOD_LINEAR,
DRM_FORMAT_MOD_INVALID
};
vc4_plane = devm_kzalloc(dev->dev, sizeof(*vc4_plane),
GFP_KERNEL);
if (!vc4_plane)
return ERR_PTR(-ENOMEM);
for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) {
/* Don't allow YUV in cursor planes, since that means
* tuning on the scaler, which we don't allow for the
* cursor.
*/
if (type != DRM_PLANE_TYPE_CURSOR ||
hvs_formats[i].hvs < HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE) {
formats[num_formats++] = hvs_formats[i].drm;
}
}
plane = &vc4_plane->base;
ret = drm_universal_plane_init(dev, plane, 0,
&vc4_plane_funcs,
formats, num_formats,
modifiers, type, NULL);
drm_plane_helper_add(plane, &vc4_plane_helper_funcs);
drm_plane_create_alpha_property(plane);
return plane;
}