Message ID | 20210722094551.15255-13-nancy.lin@mediatek.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | Add MediaTek SoC DRM (vdosys1) support for mt8195 | expand |
Hi, Nancy: Nancy.Lin <nancy.lin@mediatek.com> 於 2021年7月22日 週四 下午5:46寫道: > > Add pseudo ovl module files: My English is not good. The word 'pseudo' seems like 'looks like but indeed not the same'. I think the 'real' ovl also has rdma and mixer inside it, so I prefer to treat these two ovl as different kind of ovl, not a real one and a pseudo one. Does I misunderstanding the word 'pseudo'? > Pseudo ovl is an encapsulated module and designed for simplified > DRM control flow. This module is composed of 8 RDMAs, 4 MERGEs and > an ETHDR. Two RDMAs merge into one layer, so this module support 4 > layers. > > Signed-off-by: Nancy.Lin <nancy.lin@mediatek.com> > --- > drivers/gpu/drm/mediatek/Makefile | 4 +- > .../gpu/drm/mediatek/mtk_disp_pseudo_ovl.c | 593 ++++++++++++++++++ > .../gpu/drm/mediatek/mtk_disp_pseudo_ovl.h | 23 + > drivers/gpu/drm/mediatek/mtk_mdp_rdma.c | 456 ++++++++++++++ > drivers/gpu/drm/mediatek/mtk_mdp_rdma.h | 109 ++++ > drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h | 160 +++++ Seperate mtk_dmp_rdma driver to an independent patch. > 6 files changed, 1344 insertions(+), 1 deletion(-) > create mode 100644 drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > create mode 100644 drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > > diff --git a/drivers/gpu/drm/mediatek/Makefile b/drivers/gpu/drm/mediatek/Makefile > index 27c89847d43b..31613564f499 100644 > --- a/drivers/gpu/drm/mediatek/Makefile > +++ b/drivers/gpu/drm/mediatek/Makefile > @@ -13,7 +13,9 @@ mediatek-drm-y := mtk_disp_ccorr.o \ > mtk_drm_gem.o \ > mtk_drm_plane.o \ > mtk_dsi.o \ > - mtk_dpi.o > + mtk_dpi.o \ > + mtk_disp_pseudo_ovl.o \ > + mtk_mdp_rdma.o > > obj-$(CONFIG_DRM_MEDIATEK) += mediatek-drm.o > > diff --git a/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > new file mode 100644 > index 000000000000..0446fa99dd0a > --- /dev/null > +++ b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > @@ -0,0 +1,593 @@ > +// SPDX-License-Identifier: GPL-2.0-only > +/* > + * Copyright (c) 2021 MediaTek Inc. > + */ > + > +#include <drm/drm_fourcc.h> > +#include <linux/clk.h> > +#include <linux/reset.h> > +#include <linux/component.h> > +#include <linux/of_device.h> > +#include <linux/of_address.h> > +#include <linux/platform_device.h> > +#include <linux/pm_runtime.h> > +#include <linux/soc/mediatek/mtk-mmsys.h> > +#include <linux/soc/mediatek/mtk-cmdq.h> > + > +#include "mtk_drm_drv.h" > +#include "mtk_drm_crtc.h" > +#include "mtk_drm_ddp_comp.h" > +#include "mtk_mdp_rdma.h" > +#include "mtk_disp_pseudo_ovl.h" > + > +#define DISP_MERGE_ENABLE 0x0 > + #define MERGE_ENABLE BIT(0) > +#define DISP_MERGE_CFG_0 0x10 > +#define DISP_MERGE_CFG_1 0x14 > +#define DISP_MERGE_CFG_4 0x20 > +#define DISP_MERGE_CFG_5 0x24 > +#define DISP_MERGE_CFG_10 0x38 > + #define CFG_10_NO_SWAP 0 > +#define DISP_MERGE_CFG_12 0x40 > + #define CFG12_10_10_1PI_2PO_BUF_MODE 6 > + #define CFG12_11_10_1PI_2PO_MERGE 18 > +#define DISP_MERGE_CFG_24 0x70 > +#define DISP_MERGE_CFG_25 0x74 > +#define DISP_MERGE_CFG_26 0x78 > +#define DISP_MERGE_CFG_27 0x7c > +#define DISP_MERGE_MUTE_0 0xf00 > + > +#define MTK_PSEUDO_OVL_SINGLE_PIPE_MAX_WIDTH 1920 Which hardware has this limitation? I would like this naming to reflect which hardware has this limitation. > + > +enum mtk_pseudo_ovl_comp_type { > + PSEUDO_OVL_TYPE_RDMA = 0, > + PSEUDO_OVL_TYPE_MERGE, > + PSEUDO_OVL_TYPE_NUM, > +}; > + > +enum mtk_pseudo_ovl_comp_id { > + PSEUDO_OVL_RDMA_BASE = 0, > + PSEUDO_OVL_MDP_RDMA0 = PSEUDO_OVL_RDMA_BASE, I think you could directly use PSEUDO_OVL_MDP_RDMA0 and remove PSEUDO_OVL_RDMA_BASE. > + PSEUDO_OVL_MDP_RDMA1, > + PSEUDO_OVL_MDP_RDMA2, > + PSEUDO_OVL_MDP_RDMA3, > + PSEUDO_OVL_MDP_RDMA4, > + PSEUDO_OVL_MDP_RDMA5, > + PSEUDO_OVL_MDP_RDMA6, > + PSEUDO_OVL_MDP_RDMA7, > + PSEUDO_OVL_MERGE_BASE, > + PSEUDO_OVL_MERGE0 = PSEUDO_OVL_MERGE_BASE, Ditto. > + PSEUDO_OVL_MERGE1, > + PSEUDO_OVL_MERGE2, > + PSEUDO_OVL_MERGE3, > + PSEUDO_OVL_ID_MAX > +}; > + > +struct pseudo_ovl_data { > + unsigned int layer_nr; > + struct mtk_mdp_rdma_fifo fifo; > +}; > + > +struct pseudo_ovl_comp_match { > + enum mtk_pseudo_ovl_comp_type type; > + int alias_id; > +}; > + > +struct pseudo_ovl_merge_config { > + unsigned int fmt; > + unsigned int merge_mode; > + unsigned int in_w[2]; > + unsigned int out_w[2]; > + unsigned int in_h; > +}; > + > +struct mtk_pseudo_ovl_comp { > + struct device *dev; > + struct clk *clks[2]; > + struct cmdq_client_reg cmdq_base; > + void __iomem *regs; > +}; > + > +struct mtk_disp_pseudo_ovl { > + struct mtk_pseudo_ovl_comp pseudo_ovl_comp[PSEUDO_OVL_ID_MAX]; > + const struct pseudo_ovl_data *data; > + struct device *mmsys_dev; > +}; > + > +static const char * const pseudo_ovl_comp_str[] = { > + "PSEUDO_OVL_MDP_RDMA0", > + "PSEUDO_OVL_MDP_RDMA1", > + "PSEUDO_OVL_MDP_RDMA2", > + "PSEUDO_OVL_MDP_RDMA3", > + "PSEUDO_OVL_MDP_RDMA4", > + "PSEUDO_OVL_MDP_RDMA5", > + "PSEUDO_OVL_MDP_RDMA6", > + "PSEUDO_OVL_MDP_RDMA7", > + "PSEUDO_OVL_MERGE0", > + "PSEUDO_OVL_MERGE1", > + "PSEUDO_OVL_MERGE2", > + "PSEUDO_OVL_MERGE3", > + "PSEUDO_OVL_ID_MAX" > +}; > + > +static const char * const private_comp_stem[PSEUDO_OVL_TYPE_NUM] = { > + [PSEUDO_OVL_TYPE_RDMA] = "vdo1_rdma", > + [PSEUDO_OVL_TYPE_MERGE] = "merge", > +}; > + > +static const struct pseudo_ovl_comp_match comp_matches[PSEUDO_OVL_ID_MAX] = { > + [PSEUDO_OVL_MDP_RDMA0] = { PSEUDO_OVL_TYPE_RDMA, 0 }, > + [PSEUDO_OVL_MDP_RDMA1] = { PSEUDO_OVL_TYPE_RDMA, 1 }, > + [PSEUDO_OVL_MDP_RDMA2] = { PSEUDO_OVL_TYPE_RDMA, 2 }, > + [PSEUDO_OVL_MDP_RDMA3] = { PSEUDO_OVL_TYPE_RDMA, 3 }, > + [PSEUDO_OVL_MDP_RDMA4] = { PSEUDO_OVL_TYPE_RDMA, 4 }, > + [PSEUDO_OVL_MDP_RDMA5] = { PSEUDO_OVL_TYPE_RDMA, 5 }, > + [PSEUDO_OVL_MDP_RDMA6] = { PSEUDO_OVL_TYPE_RDMA, 6 }, > + [PSEUDO_OVL_MDP_RDMA7] = { PSEUDO_OVL_TYPE_RDMA, 7 }, > + [PSEUDO_OVL_MERGE0] = { PSEUDO_OVL_TYPE_MERGE, 1 }, > + [PSEUDO_OVL_MERGE1] = { PSEUDO_OVL_TYPE_MERGE, 2 }, > + [PSEUDO_OVL_MERGE2] = { PSEUDO_OVL_TYPE_MERGE, 3 }, > + [PSEUDO_OVL_MERGE3] = { PSEUDO_OVL_TYPE_MERGE, 4 }, > +}; > + > +static int mtk_pseudo_ovl_fifo_setting(struct mtk_disp_pseudo_ovl *pseudo_ovl, > + struct cmdq_pkt *handle) > +{ > + struct mtk_pseudo_ovl_comp *rdma = NULL; > + const struct pseudo_ovl_data *data = pseudo_ovl->data; > + const struct mtk_mdp_rdma_fifo *fifo = &data->fifo; > + int i; > + > + for (i = PSEUDO_OVL_MDP_RDMA0; i <= PSEUDO_OVL_MDP_RDMA7; i++) { > + rdma = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + i]; > + mtk_mdp_rdma_fifo_config(rdma->regs, handle, &rdma->cmdq_base, fifo); > + } > + > + return 0; > +} > + > +static void mtk_pseudo_ovl_merge_config(struct mtk_pseudo_ovl_comp *comp, > + struct pseudo_ovl_merge_config *merge_cfg, > + struct cmdq_pkt *cmdq_pkt) > +{ > + switch (merge_cfg->merge_mode) { > + case 6: > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_0); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_4); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_24); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_25); > + break; > + case 18: > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_0); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_1); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_4); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_24); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_25); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_26); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_27); > + break; > + default: > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_0); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_1); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[0]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_4); > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[1]), > + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_5); > + break; > + } > + > + mtk_ddp_write(cmdq_pkt, merge_cfg->merge_mode, &comp->cmdq_base, > + comp->regs, DISP_MERGE_CFG_12); > + mtk_ddp_write(cmdq_pkt, CFG_10_NO_SWAP, &comp->cmdq_base, > + comp->regs, DISP_MERGE_CFG_10); > + mtk_ddp_write_mask(cmdq_pkt, 1, &comp->cmdq_base, comp->regs, > + DISP_MERGE_ENABLE, MERGE_ENABLE); > +} > + > +void mtk_pseudo_ovl_layer_config(struct device *dev, unsigned int idx, > + struct mtk_plane_state *state, > + struct cmdq_pkt *cmdq_pkt) > +{ > + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); > + struct mtk_plane_pending_state *pending = &state->pending; > + struct pseudo_ovl_merge_config merge_cfg = {0}; > + struct mtk_mdp_rdma_cfg rdma_config = {0}; > + struct mtk_pseudo_ovl_comp *rdma_l; > + struct mtk_pseudo_ovl_comp *rdma_r; > + struct mtk_pseudo_ovl_comp *merge; > + const struct drm_format_info *fmt_info = drm_format_info(pending->format); > + bool use_dual_pipe = false; > + > + dev_dbg(dev, "%s+ idx:%d, enable:%d, fmt:0x%x\n", __func__, idx, > + pending->enable, pending->format); > + dev_dbg(dev, "addr 0x%lx, fb w:%d, {%d,%d,%d,%d}\n", > + pending->addr, (pending->pitch / fmt_info->cpp[0]), > + pending->x, pending->y, pending->width, pending->height); > + > + rdma_l = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * idx]; > + rdma_r = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * idx + 1]; > + merge = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_MERGE_BASE + idx]; > + > + if (!pending->enable) { > + mtk_ddp_write_mask(cmdq_pkt, 0x0, &merge->cmdq_base, merge->regs, > + DISP_MERGE_ENABLE, MERGE_ENABLE); > + mtk_mdp_rdma_stop(rdma_l->regs, cmdq_pkt, &rdma_l->cmdq_base); > + mtk_mdp_rdma_stop(rdma_r->regs, cmdq_pkt, &rdma_r->cmdq_base); > + return; > + } > + > + if (pending->width > MTK_PSEUDO_OVL_SINGLE_PIPE_MAX_WIDTH) > + use_dual_pipe = true; > + > + merge_cfg.out_w[0] = pending->width; > + merge_cfg.in_h = pending->height; > + merge_cfg.fmt = pending->format; > + if (use_dual_pipe) { > + merge_cfg.merge_mode = CFG12_11_10_1PI_2PO_MERGE; > + merge_cfg.in_w[0] = (pending->width / 2) + ((pending->width / 2) % 2); > + merge_cfg.in_w[1] = (pending->width / 2) - ((pending->width / 2) % 2); > + } else { > + merge_cfg.merge_mode = CFG12_10_10_1PI_2PO_BUF_MODE; > + merge_cfg.in_w[0] = pending->width; > + } > + > + mtk_pseudo_ovl_merge_config(merge, &merge_cfg, cmdq_pkt); > + > + mtk_mmsys_ddp_config(pseudo_ovl->mmsys_dev, MMSYS_CONFIG_MERGE_ASYNC_WIDTH, > + idx, pending->width / 2, cmdq_pkt); > + mtk_mmsys_ddp_config(pseudo_ovl->mmsys_dev, MMSYS_CONFIG_MERGE_ASYNC_HEIGHT, > + idx, pending->height, cmdq_pkt); > + > + rdma_config.source_width = pending->pitch / fmt_info->cpp[0]; > + rdma_config.csc_enable = fmt_info->is_yuv ? true : false; > + rdma_config.profile = RDMA_CSC_FULL709_TO_RGB; > + rdma_config.encode_type = RDMA_ENCODE_NONE; > + rdma_config.block_size = RDMA_BLOCK_NONE; > + rdma_config.width = merge_cfg.in_w[0]; > + rdma_config.height = pending->height; > + rdma_config.addr0 = pending->addr; > + rdma_config.fmt = pending->format; > + mtk_mdp_rdma_config(rdma_l->regs, &rdma_config, cmdq_pkt, &rdma_l->cmdq_base); > + > + rdma_config.x_left = merge_cfg.in_w[0]; > + rdma_config.width = merge_cfg.in_w[1]; > + mtk_mdp_rdma_config(rdma_r->regs, &rdma_config, cmdq_pkt, &rdma_r->cmdq_base); > + > + mtk_ddp_write_mask(cmdq_pkt, 0x1, &merge->cmdq_base, merge->regs, > + DISP_MERGE_ENABLE, MERGE_ENABLE); > + mtk_ddp_write_mask(cmdq_pkt, 0x0, &merge->cmdq_base, merge->regs, > + DISP_MERGE_MUTE_0, 0x1); > + > + mtk_mdp_rdma_start(rdma_l->regs, cmdq_pkt, &rdma_l->cmdq_base); > + if (use_dual_pipe) > + mtk_mdp_rdma_start(rdma_r->regs, cmdq_pkt, &rdma_r->cmdq_base); > + else > + mtk_mdp_rdma_stop(rdma_r->regs, cmdq_pkt, &rdma_r->cmdq_base); > +} > + > +void mtk_pseudo_ovl_config(struct device *dev, unsigned int w, unsigned int h, > + unsigned int vrefresh, unsigned int bpc, > + struct cmdq_pkt *cmdq_pkt) > +{ > + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); > + > + dev_info(dev, "%s w:%d, h:%d\n", __func__, w, h); > + > + mtk_pseudo_ovl_fifo_setting(pseudo_ovl, cmdq_pkt); > +} > + > +void mtk_pseudo_ovl_start(struct device *dev) > +{ > +} > + > +void mtk_pseudo_ovl_stop(struct device *dev) > +{ > + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); > + struct mtk_pseudo_ovl_comp *rdma_l; > + struct mtk_pseudo_ovl_comp *rdma_r; > + struct mtk_pseudo_ovl_comp *merge; > + unsigned int reg; > + u32 i; > + > + for (i = 0; i < pseudo_ovl->data->layer_nr; i++) { > + rdma_l = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * i]; > + rdma_r = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * i + 1]; > + merge = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_MERGE_BASE + i]; > + > + mtk_mdp_rdma_stop(rdma_l->regs, NULL, &rdma_l->cmdq_base); > + mtk_mdp_rdma_stop(rdma_r->regs, NULL, &rdma_r->cmdq_base); > + > + reg = readl(merge->regs + DISP_MERGE_ENABLE); > + reg = reg & ~MERGE_ENABLE; > + writel_relaxed(reg, merge->regs + DISP_MERGE_ENABLE); > + > + device_reset_optional(merge->dev); > + } > +} > + > +int mtk_pseudo_ovl_clk_enable(struct device *dev) > +{ > + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); > + struct mtk_pseudo_ovl_comp *comp; > + int ret; > + int i; > + int j; > + > + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_ID_MAX; i++) { > + comp = &pseudo_ovl->pseudo_ovl_comp[i]; > + if (!comp->dev) > + continue; > + > + /* Need to power on for private rdma devices */ > + if (i < PSEUDO_OVL_MERGE_BASE) { > + ret = pm_runtime_get_sync(comp->dev); > + if (ret < 0) > + dev_err(dev, > + "Failed to power on, err %d-%s\n", > + ret, pseudo_ovl_comp_str[i]); > + } > + > + for (j = 0; j < ARRAY_SIZE(comp->clks); j++) { > + if (IS_ERR(comp->clks[j])) > + break; > + > + ret = clk_prepare_enable(comp->clks[j]); > + if (ret) > + dev_err(dev, > + "Failed to enable clock %d, err %d-%s\n", > + i, ret, pseudo_ovl_comp_str[i]); > + } > + } > + > + return ret; > +} > + > +void mtk_pseudo_ovl_clk_disable(struct device *dev) > +{ > + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); > + struct mtk_pseudo_ovl_comp *comp; > + int ret; > + int i; > + int j; > + > + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_ID_MAX; i++) { > + comp = &pseudo_ovl->pseudo_ovl_comp[i]; > + if (!comp->dev) > + continue; > + > + for (j = 0; i < ARRAY_SIZE(comp->clks); j++) { > + if (IS_ERR(comp->clks[j])) > + break; > + clk_disable_unprepare(comp->clks[j]); > + } > + > + /* Need to power off for private rdma devices */ > + if (i < PSEUDO_OVL_MERGE_BASE) { > + ret = pm_runtime_put(comp->dev); > + if (ret < 0) > + dev_err(dev, > + "Failed to power off, err-%s\n", > + ret, pseudo_ovl_comp_str[i]); > + } > + } > +} > + > +static int pseudo_ovl_comp_get_id(struct device *dev, struct device_node *node, > + enum mtk_pseudo_ovl_comp_type type) > +{ > + int alias_id = of_alias_get_id(node, private_comp_stem[type]); > + int ret; > + int i; > + > + for (i = 0; i < ARRAY_SIZE(comp_matches); i++) > + if (comp_matches[i].type == type && > + comp_matches[i].alias_id == alias_id) > + return i; > + > + dev_err(dev, "Failed to get id. type: %d, alias: %d\n", type, alias_id); > + return -EINVAL; > +} > + > +static int private_comp_init(struct device *dev, struct device_node *node, > + struct mtk_pseudo_ovl_comp *comp, > + enum mtk_pseudo_ovl_comp_id id) > +{ > + struct platform_device *comp_pdev; > + int ret; > + int i; > + > + if (id < 0 || id >= PSEUDO_OVL_ID_MAX) { > + dev_err(dev, "Invalid component id %d\n", id); > + return -EINVAL; > + } > + > + comp_pdev = of_find_device_by_node(node); > + if (!comp_pdev) { > + dev_warn(dev, "can't find platform device of node:%s\n", > + node->name); > + return -ENODEV; > + } > + comp->dev = &comp_pdev->dev; > + comp->regs = of_iomap(node, 0); > + > + for (i = 0; i < ARRAY_SIZE(comp->clks); i++) { > + comp->clks[i] = of_clk_get(node, i); > + if (IS_ERR(comp->clks[i])) > + break; > + } > + > +#if IS_REACHABLE(CONFIG_MTK_CMDQ) > + ret = cmdq_dev_get_client_reg(comp->dev, &comp->cmdq_base, 0); > + if (ret) > + dev_info(dev, "get mediatek,gce-client-reg fail!\n"); > +#endif > + > + if (id < PSEUDO_OVL_MERGE_BASE) > + pm_runtime_enable(comp->dev); > + > + dev_info(dev, "[DRM]regs:0x%p, node:%s\n", comp->regs, pseudo_ovl_comp_str[id]); > + > + return 0; > +} > + > +static int mtk_disp_pseudo_ovl_comp_probe(struct platform_device *pdev) > +{ > + return 0; > +} > + > +static int mtk_disp_pseudo_ovl_comp_remove(struct platform_device *pdev) > +{ > + return 0; > +} > + > +static const struct of_device_id mtk_pseudo_ovl_comp_dt_ids[] = { > + { > + .compatible = "mediatek,mt8195-vdo1-rdma", > + .data = (void *)PSEUDO_OVL_TYPE_RDMA, > + }, { > + .compatible = "mediatek,mt8195-vdo1-merge", > + .data = (void *)PSEUDO_OVL_TYPE_MERGE, > + }, > + {}, > +}; > + > +static struct platform_driver mtk_disp_pseudo_ovl_comp_driver = { > + .probe = mtk_disp_pseudo_ovl_comp_probe, > + .remove = mtk_disp_pseudo_ovl_comp_remove, > + .driver = { > + .name = "mediatek-disp-pseudo-ovl-comp", > + .owner = THIS_MODULE, > + .of_match_table = mtk_pseudo_ovl_comp_dt_ids, > + }, > +}; > +module_platform_driver(mtk_disp_pseudo_ovl_comp_driver); > + > +static int pseudo_ovl_comp_init(struct device *dev) > +{ > + struct mtk_disp_pseudo_ovl *priv = dev_get_drvdata(dev); > + struct device_node *node, *parent; > + int i, ret; > + > + parent = dev->parent->of_node->parent; > + > + for_each_child_of_node(parent, node) { > + const struct of_device_id *of_id; > + enum mtk_pseudo_ovl_comp_type type; > + struct mtk_pseudo_ovl_comp *comp; > + int id; > + > + of_id = of_match_node(mtk_pseudo_ovl_comp_dt_ids, node); > + if (!of_id) > + continue; > + > + if (!of_device_is_available(node)) { > + dev_info(dev, "Skipping disabled component %pOF\n", > + node); > + continue; > + } > + > + type = (enum mtk_pseudo_ovl_comp_type)of_id->data; > + id = pseudo_ovl_comp_get_id(dev, node, type); > + if (id < 0) { > + dev_warn(dev, "Skipping unknown component %pOF\n", > + node); > + continue; > + } > + > + ret = private_comp_init(dev, node, &priv->pseudo_ovl_comp[id], id); > + if (ret) > + return ret; > + } > + > + return 0; > +} > + > +static const struct pseudo_ovl_data mt8195_pseudo_ovl_driver_data = { > + .layer_nr = 4, > + .fifo.read_request_type = 7, > + .fifo.command_div = 1, > + .fifo.ext_preutra_en = 1, > + .fifo.ultra_en = 0, > + .fifo.pre_ultra_en = 1, > + .fifo.ext_ultra_en = 1, > + .fifo.extrd_arb_max_0 = 3, > + .fifo.buf_resv_size_0 = 0, > + .fifo.issue_req_th_0 = 0, > + .fifo.ultra_h_con_0 = 156, > + .fifo.ultra_l_con_0 = 104, > +}; Now only support one SoC, so remove driver data and change these to definition instead of variable. > + > +static const struct of_device_id pseudo_ovl_driver_dt_match[] = { > + { .compatible = "mediatek,mt8195-disp-ethdr", > + .data = &mt8195_pseudo_ovl_driver_data}, > + {}, > +}; > +MODULE_DEVICE_TABLE(of, mtk_disp_pseudo_ovl_driver_dt_match); > + > +static int mtk_disp_pseudo_ovl_probe(struct platform_device *pdev) > +{ > + struct mtk_disp_pseudo_ovl *priv; > + struct device *dev = &pdev->dev; > + struct device_node *phandle = dev->parent->of_node; > + const struct of_device_id *of_id; > + int ret; > + int i; > + > + dev_info(dev, "%s+\n", __func__); > + > + of_id = of_match_node(pseudo_ovl_driver_dt_match, phandle); > + if (!of_id) > + return -ENODEV; > + > + priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL); > + if (!priv) > + return -ENOMEM; > + > + priv->mmsys_dev = pdev->dev.platform_data; > + > + priv->data = of_id->data; > + platform_set_drvdata(pdev, priv); > + > + ret = pseudo_ovl_comp_init(dev); > + if (ret) { > + dev_notice(dev, "pseudo_ovl comp init fail\n"); > + return ret; > + } > + > + dev_info(dev, "%s-\n", __func__); > + return ret; > +} > + > +static int mtk_disp_pseudo_ovl_remove(struct platform_device *pdev) > +{ > + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(&pdev->dev); > + int i; > + > + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_MERGE_BASE; i++) > + pm_runtime_disable(pseudo_ovl->pseudo_ovl_comp[i].dev); > + > + return 0; > +} > + > +struct platform_driver mtk_disp_pseudo_ovl_driver = { > + .probe = mtk_disp_pseudo_ovl_probe, > + .remove = mtk_disp_pseudo_ovl_remove, > + .driver = { > + .name = "mediatek-disp-pseudo-ovl", > + .owner = THIS_MODULE, > + }, > +}; > +module_platform_driver(mtk_disp_pseudo_ovl_driver); > diff --git a/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > new file mode 100644 > index 000000000000..b3fe1e1702b8 > --- /dev/null > +++ b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > @@ -0,0 +1,23 @@ > +/* SPDX-License-Identifier: GPL-2.0-only */ > +/* > + * Copyright (c) 2021 MediaTek Inc. > + */ > + > +#ifndef __MTK_DISP_PSEUDO_OVL_H__ > +#define __MTK_DISP_PSEUDO_OVL_H__ > + > +#include <drm/mediatek_drm.h> > + > +void mtk_pseudo_ovl_start(struct device *dev); > +void mtk_pseudo_ovl_stop(struct device *dev); > +int mtk_pseudo_ovl_clk_enable(struct device *dev); > +void mtk_pseudo_ovl_clk_disable(struct device *dev); > +void mtk_pseudo_ovl_config(struct device *dev, unsigned int w, unsigned int h, > + unsigned int vrefresh, unsigned int bpc, > + struct cmdq_pkt *cmdq_pkt); > +void mtk_pseudo_ovl_layer_config(struct device *dev, unsigned int idx, > + struct mtk_plane_state *state, > + struct cmdq_pkt *cmdq_pkt); > + > +#endif > + > diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > new file mode 100644 > index 000000000000..81d3cc4872eb > --- /dev/null > +++ b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > @@ -0,0 +1,456 @@ > +// SPDX-License-Identifier: GPL-2.0-only > +/* > + * Copyright (c) 2021 MediaTek Inc. > + */ > + > +#include <drm/drm_fourcc.h> > +#include "mtk_drm_drv.h" > +#include "mtk_mdp_reg_rdma.h" > +#include "mtk_mdp_rdma.h" > + > +#define RDMA_INPUT_SWAP BIT(14) > +#define RDMA_INPUT_10BIT BIT(18) > +#define IRQ_INT_EN_ALL \ > + (REG_FLD_MASK(FLD_UNDERRUN_INT_EN) |\ > + REG_FLD_MASK(FLD_REG_UPDATE_INT_EN) |\ > + REG_FLD_MASK(FLD_FRAME_COMPLETE_INT_EN)) > + > +static unsigned int rdma_get_y_pitch(unsigned int fmt, unsigned int width) > +{ > + switch (fmt) { > + default: > + case DRM_FORMAT_RGB565: > + case DRM_FORMAT_BGR565: > + return 2 * width; > + case DRM_FORMAT_RGB888: > + case DRM_FORMAT_BGR888: > + return 3 * width; > + case DRM_FORMAT_RGBX8888: > + case DRM_FORMAT_RGBA8888: > + case DRM_FORMAT_BGRX8888: > + case DRM_FORMAT_BGRA8888: > + case DRM_FORMAT_XRGB8888: > + case DRM_FORMAT_ARGB8888: > + case DRM_FORMAT_XBGR8888: > + case DRM_FORMAT_ABGR8888: > + case DRM_FORMAT_ARGB2101010: > + case DRM_FORMAT_ABGR2101010: > + case DRM_FORMAT_RGBA1010102: > + case DRM_FORMAT_BGRA1010102: > + return 4 * width; > + case DRM_FORMAT_UYVY: > + case DRM_FORMAT_YUYV: > + return 2 * width; > + case DRM_FORMAT_NV12: > + case DRM_FORMAT_NV21: > + return 1 * width; > + } > +} > + > +static unsigned int rdma_get_uv_pitch(unsigned int fmt, unsigned int width) > +{ > + switch (fmt) { > + default: > + return 0; > + case DRM_FORMAT_NV12: > + case DRM_FORMAT_NV21: > + return 4 * width; > + } > +} > + > +static unsigned int rdma_get_block_h(unsigned int mode) > +{ > + switch (mode) { > + default: > + return 0; > + case RDMA_BLOCK_8x8: > + case RDMA_BLOCK_16x8: > + case RDMA_BLOCK_32x8: > + return 8; > + case RDMA_BLOCK_8x16: > + case RDMA_BLOCK_16x16: > + case RDMA_BLOCK_32x16: > + return 16; > + case RDMA_BLOCK_8x32: > + case RDMA_BLOCK_16x32: > + case RDMA_BLOCK_32x32: > + return 32; > + } > +} > + > +static unsigned int rdma_get_horizontal_shift_uv(unsigned int fmt) > +{ > + switch (fmt) { > + default: > + return 0; > + case DRM_FORMAT_NV12: > + case DRM_FORMAT_NV21: > + return 1; > + } > +} > + > +static unsigned int rdma_get_vertical_shift_uv(unsigned int fmt) > +{ > + switch (fmt) { > + default: > + return 0; > + case DRM_FORMAT_NV12: > + case DRM_FORMAT_NV21: > + return 1; > + } > +} > + > +static unsigned int rdma_get_bits_per_pixel_y(unsigned int fmt) > +{ > + switch (fmt) { > + default: > + case DRM_FORMAT_RGB565: > + case DRM_FORMAT_BGR565: > + return 16; > + case DRM_FORMAT_RGB888: > + case DRM_FORMAT_BGR888: > + return 24; > + case DRM_FORMAT_RGBX8888: > + case DRM_FORMAT_RGBA8888: > + case DRM_FORMAT_BGRX8888: > + case DRM_FORMAT_BGRA8888: > + case DRM_FORMAT_XRGB8888: > + case DRM_FORMAT_ARGB8888: > + case DRM_FORMAT_XBGR8888: > + case DRM_FORMAT_ABGR8888: > + return 32; > + case DRM_FORMAT_UYVY: > + case DRM_FORMAT_YUYV: > + return 16; > + case DRM_FORMAT_NV12: > + case DRM_FORMAT_NV21: > + return 8; > + } > +} > + > +static unsigned int rdma_get_bits_per_pixel_uv(unsigned int fmt) > +{ > + switch (fmt) { > + default: > + return 0; > + case DRM_FORMAT_NV12: > + case DRM_FORMAT_NV21: > + return 16; > + } > +} > + > +static bool with_alpha(uint32_t format) > +{ > + const struct drm_format_info *fmt_info = drm_format_info(format); > + > + return fmt_info->has_alpha; > +} > + > +static unsigned int rdma_fmt_convert(unsigned int fmt) > +{ > + switch (fmt) { > + default: > + case DRM_FORMAT_RGB565: > + return RDMA_INPUT_FORMAT_RGB565; > + case DRM_FORMAT_BGR565: > + return RDMA_INPUT_FORMAT_RGB565 | RDMA_INPUT_SWAP; > + case DRM_FORMAT_RGB888: > + return RDMA_INPUT_FORMAT_RGB888; > + case DRM_FORMAT_BGR888: > + return RDMA_INPUT_FORMAT_RGB888 | RDMA_INPUT_SWAP; > + case DRM_FORMAT_RGBX8888: > + case DRM_FORMAT_RGBA8888: > + return RDMA_INPUT_FORMAT_ARGB8888; > + case DRM_FORMAT_BGRX8888: > + case DRM_FORMAT_BGRA8888: > + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_SWAP; > + case DRM_FORMAT_XRGB8888: > + case DRM_FORMAT_ARGB8888: > + return RDMA_INPUT_FORMAT_RGBA8888; > + case DRM_FORMAT_XBGR8888: > + case DRM_FORMAT_ABGR8888: > + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_SWAP; > + case DRM_FORMAT_ABGR2101010: > + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_SWAP | > + RDMA_INPUT_10BIT; > + case DRM_FORMAT_ARGB2101010: > + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_10BIT; > + case DRM_FORMAT_RGBA1010102: > + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_SWAP | > + RDMA_INPUT_10BIT; > + case DRM_FORMAT_BGRA1010102: > + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_10BIT; > + case DRM_FORMAT_UYVY: > + return RDMA_INPUT_FORMAT_UYVY; > + case DRM_FORMAT_YUYV: > + return RDMA_INPUT_FORMAT_YUY2; > + } > +} > + > +void mtk_mdp_rdma_start(void __iomem *base, struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base) > +{ > + unsigned int inten = IRQ_INT_EN_ALL; > + > + mtk_ddp_write_mask(cmdq_pkt, inten, cmdq_base, base, > + MDP_RDMA_INTERRUPT_ENABLE, IRQ_INT_EN_ALL); Why do you enable rdma interrupt? > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_ROT_ENABLE, 1), cmdq_base, > + base, MDP_RDMA_EN, REG_FLD_MASK(FLD_ROT_ENABLE)); > +} > + > +void mtk_mdp_rdma_stop(void __iomem *base, struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base) > +{ > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_ROT_ENABLE, 0), cmdq_base, > + base, MDP_RDMA_EN, REG_FLD_MASK(FLD_ROT_ENABLE)); > + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, > + MDP_RDMA_INTERRUPT_ENABLE, IRQ_INT_EN_ALL); > + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, > + MDP_RDMA_INTERRUPT_STATUS, IRQ_INT_EN_ALL); > + mtk_ddp_write_mask(cmdq_pkt, 1, cmdq_base, base, MDP_RDMA_RESET, ~0); > + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, MDP_RDMA_RESET, ~0); > +} > + > +void mtk_mdp_rdma_fifo_config(void __iomem *base, struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base, > + const struct mtk_mdp_rdma_fifo *fifo) Why do you pass fifo config from ovl to rdma? I think just rdma know how to set these fifo parameter. > +{ > + int reg; > + int reg_val; > + int reg_mask; > + > + reg = MDP_RDMA_GMCIF_CON; > + reg_val = REG_FLD_VAL(FLD_RD_REQ_TYPE, fifo->read_request_type) | > + REG_FLD_VAL(FLD_COMMAND_DIV, fifo->command_div) | > + REG_FLD_VAL(FLD_EXT_PREULTRA_EN, fifo->ext_preutra_en) | > + REG_FLD_VAL(FLD_ULTRA_EN, fifo->ultra_en) | > + REG_FLD_VAL(PRE_ULTRA_EN, fifo->pre_ultra_en) | > + REG_FLD_VAL(FLD_EXT_ULTRA_EN, fifo->ext_ultra_en); > + reg_mask = REG_FLD_MASK(FLD_RD_REQ_TYPE) | > + REG_FLD_MASK(FLD_COMMAND_DIV) | > + REG_FLD_MASK(FLD_EXT_PREULTRA_EN) | > + REG_FLD_MASK(FLD_ULTRA_EN) | > + REG_FLD_MASK(PRE_ULTRA_EN) | > + REG_FLD_MASK(FLD_EXT_ULTRA_EN); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_DMA_CON_0; > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX, fifo->extrd_arb_max_0) | > + REG_FLD_VAL(FLD_BUF_RESV_SIZE, fifo->buf_resv_size_0) | > + REG_FLD_VAL(FLD_ISSUE_REQ_TH, fifo->issue_req_th_0); > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX) | > + REG_FLD_MASK(FLD_BUF_RESV_SIZE) | > + REG_FLD_MASK(FLD_ISSUE_REQ_TH); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_UTRA_H_CON_0; > + reg_val = REG_FLD_VAL(FLD_PREUTRA_H_OFS_0, fifo->ultra_h_con_0); > + reg_mask = REG_FLD_MASK(FLD_PREUTRA_H_OFS_0); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_UTRA_L_CON_0; > + reg_val = REG_FLD_VAL(FLD_PREUTRA_L_OFS_0, fifo->ultra_l_con_0); > + reg_mask = REG_FLD_MASK(FLD_PREUTRA_L_OFS_0); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_DMABUF_CON_1; > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_1, 0) | > + REG_FLD_VAL(FLD_BUF_RESV_SIZE_1, 0) | > + REG_FLD_VAL(FLD_ISSUE_REQ_TH_1, 0); > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_1) | > + REG_FLD_MASK(FLD_BUF_RESV_SIZE_1) | > + REG_FLD_MASK(FLD_ISSUE_REQ_TH_1); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_ULTRA_TH_HIGH_CON_1; > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_1, 0); > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_1); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_ULTRA_TH_LOW_CON_1; > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_1, 0); > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_1); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_DMABUF_CON_2; > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_2, 0) | > + REG_FLD_VAL(FLD_BUF_RESV_SIZE_2, 0) | > + REG_FLD_VAL(FLD_ISSUE_REQ_TH_2, 0); > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_2) | > + REG_FLD_MASK(FLD_BUF_RESV_SIZE_2) | > + REG_FLD_MASK(FLD_ISSUE_REQ_TH_2); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_UTRA_H_CON_2; > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_2, 0); > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_2); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_ULTRA_TH_LOW_CON_2; > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_2, 0); > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_2); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_DMABUF_CON_3; > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_3, 0) | > + REG_FLD_VAL(FLD_BUF_RESV_SIZE_3, 0) | > + REG_FLD_VAL(FLD_ISSUE_REQ_TH_3, 0); > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_3) | > + REG_FLD_MASK(FLD_BUF_RESV_SIZE_3) | > + REG_FLD_MASK(FLD_ISSUE_REQ_TH_3); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_UTRA_H_CON_3; > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_3, 0); > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_3); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > + > + reg = MDP_RDMA_ULTRA_TH_LOW_CON_3; > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_3, 0); > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_3); > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); > +} > + > +void mtk_mdp_rdma_config(void __iomem *base, struct mtk_mdp_rdma_cfg *cfg, > + struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base) > +{ > + unsigned int src_pitch_uv = rdma_get_uv_pitch(cfg->fmt, cfg->source_width); > + unsigned int src_pitch_y = rdma_get_y_pitch(cfg->fmt, cfg->source_width); This is got from mtk_plane_atomic_update(). > + unsigned int h_shift_uv = rdma_get_horizontal_shift_uv(cfg->fmt); > + unsigned int v_shift_uv = rdma_get_vertical_shift_uv(cfg->fmt); > + unsigned int bpp_uv = rdma_get_bits_per_pixel_uv(cfg->fmt); > + unsigned int block_h = rdma_get_block_h(cfg->block_size); block_h = 0; > + unsigned int bpp_y = rdma_get_bits_per_pixel_y(cfg->fmt); drm_format_info(). > + unsigned int y_start_line = 0; > + unsigned int offset_y = 0; > + unsigned int offset_u = 0; > + unsigned int offset_v = 0; > + > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_UNIFORM_CONFIG, 1), > + cmdq_base, base, MDP_RDMA_SRC_CON, > + REG_FLD_MASK(FLD_UNIFORM_CONFIG)); > + mtk_ddp_write_mask(cmdq_pkt, rdma_fmt_convert(cfg->fmt), cmdq_base, > + base, MDP_RDMA_SRC_CON, REG_FLD_MASK(FLD_SWAP) | > + REG_FLD_MASK(FLD_SRC_FORMAT) | > + REG_FLD_MASK(FLD_BIT_NUMBER)); > + > + if (!cfg->csc_enable && with_alpha(cfg->fmt)) > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_ARGB, 1), > + cmdq_base, base, MDP_RDMA_SRC_CON, > + REG_FLD_MASK(FLD_OUTPUT_ARGB)); > + else > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_ARGB, 0), > + cmdq_base, base, MDP_RDMA_SRC_CON, > + REG_FLD_MASK(FLD_OUTPUT_ARGB)); > + > + mtk_ddp_write_mask(cmdq_pkt, cfg->addr0, cmdq_base, base, > + MDP_RDMA_SRC_BASE_0, REG_FLD_MASK(FLD_SRC_BASE_0)); > + mtk_ddp_write_mask(cmdq_pkt, cfg->addr1, cmdq_base, base, > + MDP_RDMA_SRC_BASE_1, REG_FLD_MASK(FLD_SRC_BASE_1)); > + mtk_ddp_write_mask(cmdq_pkt, cfg->addr2, cmdq_base, base, > + MDP_RDMA_SRC_BASE_2, REG_FLD_MASK(FLD_SRC_BASE_2)); > + > + mtk_ddp_write_mask(cmdq_pkt, src_pitch_y, cmdq_base, base, > + MDP_RDMA_MF_BKGD_SIZE_IN_BYTE, > + REG_FLD_MASK(FLD_MF_BKGD_WB)); > + mtk_ddp_write_mask(cmdq_pkt, src_pitch_uv, cmdq_base, base, > + MDP_RDMA_SF_BKGD_SIZE_IN_BYTE, > + REG_FLD_MASK(FLD_SF_BKGD_WB)); > + > + if (cfg->encode_type == RDMA_ENCODE_AFBC) { Never be here, so remove. > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_BKGD_WP, cfg->source_width), > + cmdq_base, base, MDP_RDMA_MF_BKGD_SIZE_IN_PIXEL, > + REG_FLD_MASK(FLD_MF_BKGD_WP)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_BKGD_HP, cfg->height), > + cmdq_base, base, MDP_RDMA_MF_BKGD_H_SIZE_IN_PIXEL, > + REG_FLD_MASK(FLD_BKGD_HP)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_YUV_TRANSFORM, 1), > + cmdq_base, base, MDP_RDMA_COMP_CON, > + REG_FLD_MASK(FLD_AFBC_YUV_TRANSFORM)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_UFBDC_EN, 1), cmdq_base, > + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_UFBDC_EN)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_EN, 1), cmdq_base, > + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_AFBC_EN)); > + } else { > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_YUV_TRANSFORM, 0), > + cmdq_base, base, MDP_RDMA_COMP_CON, > + REG_FLD_MASK(FLD_AFBC_YUV_TRANSFORM)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_UFBDC_EN, 0), cmdq_base, > + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_UFBDC_EN)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_EN, 0), cmdq_base, > + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_AFBC_EN)); > + } > + > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_10B, 1), cmdq_base, > + base, MDP_RDMA_CON, REG_FLD_MASK(FLD_OUTPUT_10B)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SIMPLE_MODE, 1), cmdq_base, > + base, MDP_RDMA_CON, REG_FLD_MASK(FLD_SIMPLE_MODE)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_TRANS_EN, cfg->csc_enable), > + cmdq_base, base, MDP_RDMA_TRANSFORM_0, > + REG_FLD_MASK(FLD_TRANS_EN)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_INT_MATRIX_SEL, cfg->profile), > + cmdq_base, base, MDP_RDMA_TRANSFORM_0, > + REG_FLD_MASK(FLD_INT_MATRIX_SEL)); > + > + if (cfg->block_size == RDMA_BLOCK_NONE) { > + y_start_line = cfg->y_top; > + > + offset_y = (cfg->x_left * bpp_y >> 3) + y_start_line * src_pitch_y; > + offset_u = ((cfg->x_left >> h_shift_uv) * bpp_uv >> 3) + > + (y_start_line >> v_shift_uv) * src_pitch_uv; > + offset_v = ((cfg->x_left >> h_shift_uv) * bpp_uv >> 3) + > + (y_start_line >> v_shift_uv) * src_pitch_uv; > + } else { Never be here, so remove. > + offset_y = (cfg->x_left * block_h * bpp_y >> 3) + > + (cfg->y_top) * src_pitch_y; > + offset_u = ((cfg->x_left >> h_shift_uv) * (block_h >> v_shift_uv) * > + bpp_uv >> 3) + (cfg->y_top) * src_pitch_uv; > + offset_v = ((cfg->x_left >> h_shift_uv) * (block_h >> v_shift_uv) * > + bpp_uv >> 3) + (cfg->y_top) * src_pitch_uv; > + } > + > + if (cfg->encode_type == RDMA_ENCODE_AFBC) { Never be here, so remove. > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_WP, cfg->x_left), > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_WP, > + REG_FLD_MASK(FLD_SRC_OFFSET_WP)); > + > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_HP, cfg->y_top), > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_HP, > + REG_FLD_MASK(FLD_SRC_OFFSET_HP)); > + } > + > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_0, offset_y), > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_0, > + REG_FLD_MASK(FLD_SRC_OFFSET_0)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_1, offset_u), > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_1, > + REG_FLD_MASK(FLD_SRC_OFFSET_1)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_2, offset_v), > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_2, > + REG_FLD_MASK(FLD_SRC_OFFSET_2)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_SRC_W, cfg->width), > + cmdq_base, base, MDP_RDMA_MF_SRC_SIZE, > + REG_FLD_MASK(FLD_MF_SRC_W)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_SRC_H, cfg->height), > + cmdq_base, base, MDP_RDMA_MF_SRC_SIZE, > + REG_FLD_MASK(FLD_MF_SRC_H)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_CLIP_W, cfg->width), > + cmdq_base, base, MDP_RDMA_MF_CLIP_SIZE, > + REG_FLD_MASK(FLD_MF_CLIP_W)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_CLIP_H, cfg->height), > + cmdq_base, base, MDP_RDMA_MF_CLIP_SIZE, > + REG_FLD_MASK(FLD_MF_CLIP_H)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_OFFSET_W_1, 0), > + cmdq_base, base, MDP_RDMA_MF_OFFSET_1, > + REG_FLD_MASK(FLD_MF_OFFSET_W_1)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_OFFSET_H_1, 0), > + cmdq_base, base, MDP_RDMA_MF_OFFSET_1, > + REG_FLD_MASK(FLD_MF_OFFSET_H_1)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_LINE_THRESHOLD, cfg->height), > + cmdq_base, base, MDP_RDMA_TARGET_LINE, > + REG_FLD_MASK(FLD_LINE_THRESHOLD)); > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_TARGET_LINE_EN, 1), > + cmdq_base, base, MDP_RDMA_TARGET_LINE, > + REG_FLD_MASK(FLD_TARGET_LINE_EN)); > +} > diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > new file mode 100644 > index 000000000000..c16bfb716610 > --- /dev/null > +++ b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > @@ -0,0 +1,109 @@ > +/* SPDX-License-Identifier: GPL-2.0-only */ > +/* > + * Copyright (c) 2021 MediaTek Inc. > + */ > + > +#ifndef __MTK_MDP_RDMA_H__ > +#define __MTK_MDP_RDMA_H__ > + > +enum rdma_format { > + RDMA_INPUT_FORMAT_RGB565 = 0, > + RDMA_INPUT_FORMAT_RGB888 = 1, > + RDMA_INPUT_FORMAT_RGBA8888 = 2, > + RDMA_INPUT_FORMAT_ARGB8888 = 3, > + RDMA_INPUT_FORMAT_UYVY = 4, > + RDMA_INPUT_FORMAT_YUY2 = 5, > + RDMA_INPUT_FORMAT_Y8 = 7, > + RDMA_INPUT_FORMAT_YV12 = 8, > + RDMA_INPUT_FORMAT_UYVY_3PL = 9, > + RDMA_INPUT_FORMAT_NV12 = 12, > + RDMA_INPUT_FORMAT_UYVY_2PL = 13, > + RDMA_INPUT_FORMAT_Y410 = 14 > +}; > + > +enum rdma_profile { > + RDMA_CSC_RGB_TO_JPEG = 0, > + RDMA_CSC_RGB_TO_FULL709 = 1, > + RDMA_CSC_RGB_TO_BT601 = 2, > + RDMA_CSC_RGB_TO_BT709 = 3, > + RDMA_CSC_JPEG_TO_RGB = 4, > + RDMA_CSC_FULL709_TO_RGB = 5, > + RDMA_CSC_BT601_TO_RGB = 6, > + RDMA_CSC_BT709_TO_RGB = 7, > + RDMA_CSC_JPEG_TO_BT601 = 8, > + RDMA_CSC_JPEG_TO_BT709 = 9, > + RDMA_CSC_BT601_TO_JPEG = 10, > + RDMA_CSC_BT709_TO_BT601 = 11, > + RDMA_CSC_BT601_TO_BT709 = 12 > +}; > + > +enum rdma_encode { > + RDMA_ENCODE_NONE = 0, > + RDMA_ENCODE_AFBC = 1, > + RDMA_ENCODE_HYFBC = 2, > + RDMA_ENCODE_UFO_DCP = 3 > +}; > + > +enum rdma_block { > + RDMA_BLOCK_NONE = 0, > + RDMA_BLOCK_8x8 = 1, > + RDMA_BLOCK_8x16 = 2, > + RDMA_BLOCK_8x32 = 3, > + RDMA_BLOCK_16x8 = 4, > + RDMA_BLOCK_16x16 = 5, > + RDMA_BLOCK_16x32 = 6, > + RDMA_BLOCK_32x8 = 7, > + RDMA_BLOCK_32x16 = 8, > + RDMA_BLOCK_32x32 = 9 > +}; > + > +struct mtk_mdp_rdma_cfg { > + enum rdma_encode encode_type; > + enum rdma_block block_size; > + enum rdma_profile profile; > + unsigned int source_width; > + unsigned int addr0; > + unsigned int addr1; > + unsigned int addr2; > + unsigned int width; > + unsigned int height; > + unsigned int x_left; > + unsigned int y_top; > + bool csc_enable; > + int fmt; > +}; > + > +struct mtk_mdp_rdma_fifo { > + int read_request_type; > + int command_div; > + int ext_preutra_en; > + int ultra_en; > + int pre_ultra_en; > + int ext_ultra_en; > + int extrd_arb_max_0; > + int buf_resv_size_0; > + int issue_req_th_0; > + int ultra_h_con_0; > + int ultra_l_con_0; > +}; Move rdma internal definition in .c file. > + > +void mtk_mdp_rdma_start(void __iomem *base, > + struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base); > + > +void mtk_mdp_rdma_stop(void __iomem *base, > + struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base); > + > +void mtk_mdp_rdma_fifo_config(void __iomem *base, > + struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base, > + const struct mtk_mdp_rdma_fifo *fifo); > + > +void mtk_mdp_rdma_config(void __iomem *base, > + struct mtk_mdp_rdma_cfg *cfg, > + struct cmdq_pkt *cmdq_pkt, > + struct cmdq_client_reg *cmdq_base); Move these prototype to mtk_disp_drv.h > + > +#endif // __MTK_MDP_RDMA_H__ > + > diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h b/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > new file mode 100644 > index 000000000000..08abd9f39bd8 > --- /dev/null > +++ b/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > @@ -0,0 +1,160 @@ > +/* SPDX-License-Identifier: GPL-2.0-only */ > +/* > + * Copyright (c) 2021 MediaTek Inc. > + */ > + > +#ifndef __MDP_RDMA_REGS_H__ > +#define __MDP_RDMA_REGS_H__ > + > +#define REG_FLD(width, shift) \ > + ((unsigned int)((((width) & 0xff) << 16) | ((shift) & 0xff))) > + > +#define REG_FLD_WIDTH(field) ((unsigned int)(((field) >> 16) & 0xff)) > + > +#define REG_FLD_SHIFT(field) ((unsigned int)((field) & 0xff)) > + > +#define REG_FLD_MASK(field) \ > + ((unsigned int)((1ULL << REG_FLD_WIDTH(field)) - 1) \ > + << REG_FLD_SHIFT(field)) > + > +#define REG_FLD_VAL(field, val) \ > + (((val) << REG_FLD_SHIFT(field)) & REG_FLD_MASK(field)) > + > +#define MDP_RDMA_EN 0x000 > +#define FLD_ROT_ENABLE REG_FLD(1, 0) > + > +#define MDP_RDMA_RESET 0x008 > + > +#define MDP_RDMA_INTERRUPT_ENABLE 0x010 > +#define FLD_UNDERRUN_INT_EN REG_FLD(1, 2) > +#define FLD_REG_UPDATE_INT_EN REG_FLD(1, 1) > +#define FLD_FRAME_COMPLETE_INT_EN REG_FLD(1, 0) > + > +#define MDP_RDMA_INTERRUPT_STATUS 0x018 > + > +#define MDP_RDMA_CON 0x020 > +#define FLD_OUTPUT_10B REG_FLD(1, 5) > +#define FLD_SIMPLE_MODE REG_FLD(1, 4) > + > +#define MDP_RDMA_GMCIF_CON 0x028 > +#define FLD_EXT_ULTRA_EN REG_FLD(1, 18) > +#define PRE_ULTRA_EN REG_FLD(2, 16) > +#define FLD_ULTRA_EN REG_FLD(2, 12) > +#define FLD_RD_REQ_TYPE REG_FLD(4, 4) > +#define FLD_EXT_PREULTRA_EN REG_FLD(1, 3) > +#define FLD_COMMAND_DIV REG_FLD(1, 0) > + > +#define MDP_RDMA_SRC_CON 0x030 > +#define FLD_OUTPUT_ARGB REG_FLD(1, 25) > +#define FLD_BIT_NUMBER REG_FLD(2, 18) > +#define FLD_UNIFORM_CONFIG REG_FLD(1, 17) > +#define FLD_SWAP REG_FLD(1, 14) > +#define FLD_SRC_FORMAT REG_FLD(4, 0) > + > +#define MDP_RDMA_COMP_CON 0x038 > +#define FLD_AFBC_EN REG_FLD(1, 22) > +#define FLD_AFBC_YUV_TRANSFORM REG_FLD(1, 21) > +#define FLD_UFBDC_EN REG_FLD(1, 12) > + > +#define MDP_RDMA_MF_BKGD_SIZE_IN_BYTE 0x060 > +#define FLD_MF_BKGD_WB REG_FLD(23, 0) > + > +#define MDP_RDMA_MF_BKGD_SIZE_IN_PIXEL 0x068 > +#define FLD_MF_BKGD_WP REG_FLD(23, 0) > + > +#define MDP_RDMA_MF_SRC_SIZE 0x070 > +#define FLD_MF_SRC_H REG_FLD(15, 16) > +#define FLD_MF_SRC_W REG_FLD(15, 0) > + > +#define MDP_RDMA_MF_CLIP_SIZE 0x078 > +#define FLD_MF_CLIP_H REG_FLD(15, 16) > +#define FLD_MF_CLIP_W REG_FLD(15, 0) > + > +#define MDP_RDMA_MF_OFFSET_1 0x080 > +#define FLD_MF_OFFSET_H_1 REG_FLD(6, 16) > +#define FLD_MF_OFFSET_W_1 REG_FLD(5, 0) > + > +#define MDP_RDMA_SF_BKGD_SIZE_IN_BYTE 0x090 > +#define FLD_SF_BKGD_WB REG_FLD(23, 0) > + > +#define MDP_RDMA_MF_BKGD_H_SIZE_IN_PIXEL 0x098 > +#define FLD_BKGD_HP REG_FLD(23, 0) > + > +#define MDP_RDMA_TARGET_LINE 0x0a0 > +#define FLD_LINE_THRESHOLD REG_FLD(15, 17) > +#define FLD_TARGET_LINE_EN REG_FLD(1, 16) > + > +#define MDP_RDMA_SRC_OFFSET_0 0x118 > +#define FLD_SRC_OFFSET_0 REG_FLD(32, 0) > + > +#define MDP_RDMA_SRC_OFFSET_1 0x120 > +#define FLD_SRC_OFFSET_1 REG_FLD(32, 0) > + > +#define MDP_RDMA_SRC_OFFSET_2 0x128 > +#define FLD_SRC_OFFSET_2 REG_FLD(32, 0) > + > +#define MDP_RDMA_SRC_OFFSET_WP 0x148 > +#define FLD_SRC_OFFSET_WP REG_FLD(32, 0) > + > +#define MDP_RDMA_SRC_OFFSET_HP 0x150 > +#define FLD_SRC_OFFSET_HP REG_FLD(32, 0) > + > +#define MDP_RDMA_TRANSFORM_0 0x200 > +#define FLD_INT_MATRIX_SEL REG_FLD(5, 23) > +#define FLD_TRANS_EN REG_FLD(1, 16) > + > +#define MDP_RDMA_DMA_CON_0 0x240 > +#define FLD_EXTRD_ARB_MAX REG_FLD(4, 24) > +#define FLD_BUF_RESV_SIZE REG_FLD(8, 16) > +#define FLD_ISSUE_REQ_TH REG_FLD(8, 0) > + > +#define MDP_RDMA_UTRA_H_CON_0 0x248 > +#define FLD_PREUTRA_H_OFS_0 REG_FLD(10, 10) > + > +#define MDP_RDMA_UTRA_L_CON_0 0x250 > +#define FLD_PREUTRA_L_OFS_0 REG_FLD(10, 10) > + > +#define MDP_RDMA_DMABUF_CON_1 0x258 > +#define FLD_EXTRD_ARB_MAX_1 REG_FLD(4, 24) > +#define FLD_BUF_RESV_SIZE_1 REG_FLD(7, 16) > +#define FLD_ISSUE_REQ_TH_1 REG_FLD(7, 0) > + > +#define MDP_RDMA_ULTRA_TH_HIGH_CON_1 0x260 > +#define FLD_PRE_ULTRA_TH_HIGH_OFS_1 REG_FLD(10, 10) > + > +#define MDP_RDMA_ULTRA_TH_LOW_CON_1 0x268 > +#define FLD_PRE_ULTRA_TH_LOW_OFS_1 REG_FLD(10, 10) > + > +#define MDP_RDMA_DMABUF_CON_2 0x270 > +#define FLD_EXTRD_ARB_MAX_2 REG_FLD(4, 24) > +#define FLD_BUF_RESV_SIZE_2 REG_FLD(6, 16) > +#define FLD_ISSUE_REQ_TH_2 REG_FLD(6, 0) > + > +#define MDP_RDMA_UTRA_H_CON_2 0x278 > +#define FLD_PRE_ULTRA_TH_HIGH_OFS_2 REG_FLD(10, 10) > + > +#define MDP_RDMA_ULTRA_TH_LOW_CON_2 0x280 > +#define FLD_PRE_ULTRA_TH_LOW_OFS_2 REG_FLD(10, 10) > + > +#define MDP_RDMA_DMABUF_CON_3 0x288 > +#define FLD_EXTRD_ARB_MAX_3 REG_FLD(4, 24) > +#define FLD_BUF_RESV_SIZE_3 REG_FLD(6, 16) > +#define FLD_ISSUE_REQ_TH_3 REG_FLD(6, 0) > + > +#define MDP_RDMA_UTRA_H_CON_3 0x290 > +#define FLD_PRE_ULTRA_TH_HIGH_OFS_3 REG_FLD(10, 10) > + > +#define MDP_RDMA_ULTRA_TH_LOW_CON_3 0x298 > +#define FLD_PRE_ULTRA_TH_LOW_OFS_3 REG_FLD(10, 10) > + > +#define MDP_RDMA_SRC_BASE_0 0xf00 > +#define FLD_SRC_BASE_0 REG_FLD(32, 0) > + > +#define MDP_RDMA_SRC_BASE_1 0xf08 > +#define FLD_SRC_BASE_1 REG_FLD(32, 0) > + > +#define MDP_RDMA_SRC_BASE_2 0xf10 > +#define FLD_SRC_BASE_2 REG_FLD(32, 0) > + > +#endif /* __MDP_RDMA_REGS_H__ */ Only one file need these, so move register definiton in .c file. Regards, Chun-Kuang. > + > -- > 2.18.0 >
Hi Chun-Kuang, Thanks for your review. On Sun, 2021-07-25 at 09:57 +0800, Chun-Kuang Hu wrote: > Hi, Nancy: > > Nancy.Lin <nancy.lin@mediatek.com> 於 2021年7月22日 週四 下午5:46寫道: > > > > Add pseudo ovl module files: > > My English is not good. The word 'pseudo' seems like 'looks like but > indeed not the same'. > I think the 'real' ovl also has rdma and mixer inside it, so I prefer > to treat these two ovl as different kind of ovl, not a real one and a > pseudo one. > Does I misunderstanding the word 'pseudo'? > I was thinking of encapsulating several engines in this component. The prefix "pseudo" was meant to distinguish with the origin ovl. I will change pseudo_ovl to ovl_adaptor. > > Pseudo ovl is an encapsulated module and designed for simplified > > DRM control flow. This module is composed of 8 RDMAs, 4 MERGEs and > > an ETHDR. Two RDMAs merge into one layer, so this module support 4 > > layers. > > > > Signed-off-by: Nancy.Lin <nancy.lin@mediatek.com> > > --- > > drivers/gpu/drm/mediatek/Makefile | 4 +- > > .../gpu/drm/mediatek/mtk_disp_pseudo_ovl.c | 593 > > ++++++++++++++++++ > > .../gpu/drm/mediatek/mtk_disp_pseudo_ovl.h | 23 + > > drivers/gpu/drm/mediatek/mtk_mdp_rdma.c | 456 ++++++++++++++ > > drivers/gpu/drm/mediatek/mtk_mdp_rdma.h | 109 ++++ > > drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h | 160 +++++ > > Seperate mtk_dmp_rdma driver to an independent patch. > OK, I will separate it. > > 6 files changed, 1344 insertions(+), 1 deletion(-) > > create mode 100644 drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > > create mode 100644 drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > > create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > > create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > > create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > > > > diff --git a/drivers/gpu/drm/mediatek/Makefile > > b/drivers/gpu/drm/mediatek/Makefile > > index 27c89847d43b..31613564f499 100644 > > --- a/drivers/gpu/drm/mediatek/Makefile > > +++ b/drivers/gpu/drm/mediatek/Makefile > > @@ -13,7 +13,9 @@ mediatek-drm-y := mtk_disp_ccorr.o \ > > mtk_drm_gem.o \ > > mtk_drm_plane.o \ > > mtk_dsi.o \ > > - mtk_dpi.o > > + mtk_dpi.o \ > > + mtk_disp_pseudo_ovl.o \ > > + mtk_mdp_rdma.o > > > > obj-$(CONFIG_DRM_MEDIATEK) += mediatek-drm.o > > > > diff --git a/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > > b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > > new file mode 100644 > > index 000000000000..0446fa99dd0a > > --- /dev/null > > +++ b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c > > @@ -0,0 +1,593 @@ > > +// SPDX-License-Identifier: GPL-2.0-only > > +/* > > + * Copyright (c) 2021 MediaTek Inc. > > + */ > > + > > +#include <drm/drm_fourcc.h> > > +#include <linux/clk.h> > > +#include <linux/reset.h> > > +#include <linux/component.h> > > +#include <linux/of_device.h> > > +#include <linux/of_address.h> > > +#include <linux/platform_device.h> > > +#include <linux/pm_runtime.h> > > +#include <linux/soc/mediatek/mtk-mmsys.h> > > +#include <linux/soc/mediatek/mtk-cmdq.h> > > + > > +#include "mtk_drm_drv.h" > > +#include "mtk_drm_crtc.h" > > +#include "mtk_drm_ddp_comp.h" > > +#include "mtk_mdp_rdma.h" > > +#include "mtk_disp_pseudo_ovl.h" > > + > > +#define DISP_MERGE_ENABLE 0x0 > > + #define MERGE_ENABLE BIT(0) > > +#define DISP_MERGE_CFG_0 0x10 > > +#define DISP_MERGE_CFG_1 0x14 > > +#define DISP_MERGE_CFG_4 0x20 > > +#define DISP_MERGE_CFG_5 0x24 > > +#define DISP_MERGE_CFG_10 0x38 > > + #define CFG_10_NO_SWAP 0 > > +#define DISP_MERGE_CFG_12 0x40 > > + #define CFG12_10_10_1PI_2PO_BUF_MODE 6 > > + #define CFG12_11_10_1PI_2PO_MERGE 18 > > +#define DISP_MERGE_CFG_24 0x70 > > +#define DISP_MERGE_CFG_25 0x74 > > +#define DISP_MERGE_CFG_26 0x78 > > +#define DISP_MERGE_CFG_27 0x7c > > +#define DISP_MERGE_MUTE_0 0xf00 > > + > > +#define MTK_PSEUDO_OVL_SINGLE_PIPE_MAX_WIDTH 1920 > > Which hardware has this limitation? I would like this naming to > reflect which hardware has this limitation. > RDMA has the limitation, I will modify its naming. > > + > > +enum mtk_pseudo_ovl_comp_type { > > + PSEUDO_OVL_TYPE_RDMA = 0, > > + PSEUDO_OVL_TYPE_MERGE, > > + PSEUDO_OVL_TYPE_NUM, > > +}; > > + > > +enum mtk_pseudo_ovl_comp_id { > > + PSEUDO_OVL_RDMA_BASE = 0, > > + PSEUDO_OVL_MDP_RDMA0 = PSEUDO_OVL_RDMA_BASE, > > I think you could directly use PSEUDO_OVL_MDP_RDMA0 and remove > PSEUDO_OVL_RDMA_BASE. > OK, I will remove it. > > + PSEUDO_OVL_MDP_RDMA1, > > + PSEUDO_OVL_MDP_RDMA2, > > + PSEUDO_OVL_MDP_RDMA3, > > + PSEUDO_OVL_MDP_RDMA4, > > + PSEUDO_OVL_MDP_RDMA5, > > + PSEUDO_OVL_MDP_RDMA6, > > + PSEUDO_OVL_MDP_RDMA7, > > + PSEUDO_OVL_MERGE_BASE, > > + PSEUDO_OVL_MERGE0 = PSEUDO_OVL_MERGE_BASE, > > Ditto. OK, I will remove it. > > > + PSEUDO_OVL_MERGE1, > > + PSEUDO_OVL_MERGE2, > > + PSEUDO_OVL_MERGE3, > > + PSEUDO_OVL_ID_MAX > > +}; > > + > > +struct pseudo_ovl_data { > > + unsigned int layer_nr; > > + struct mtk_mdp_rdma_fifo fifo; > > +}; > > + > > +struct pseudo_ovl_comp_match { > > + enum mtk_pseudo_ovl_comp_type type; > > + int alias_id; > > +}; > > + > > +struct pseudo_ovl_merge_config { > > + unsigned int fmt; > > + unsigned int merge_mode; > > + unsigned int in_w[2]; > > + unsigned int out_w[2]; > > + unsigned int in_h; > > +}; > > + > > +struct mtk_pseudo_ovl_comp { > > + struct device *dev; > > + struct clk *clks[2]; > > + struct cmdq_client_reg cmdq_base; > > + void __iomem *regs; > > +}; > > + > > +struct mtk_disp_pseudo_ovl { > > + struct mtk_pseudo_ovl_comp > > pseudo_ovl_comp[PSEUDO_OVL_ID_MAX]; > > + const struct pseudo_ovl_data *data; > > + struct device *mmsys_dev; > > +}; > > + > > +static const char * const pseudo_ovl_comp_str[] = { > > + "PSEUDO_OVL_MDP_RDMA0", > > + "PSEUDO_OVL_MDP_RDMA1", > > + "PSEUDO_OVL_MDP_RDMA2", > > + "PSEUDO_OVL_MDP_RDMA3", > > + "PSEUDO_OVL_MDP_RDMA4", > > + "PSEUDO_OVL_MDP_RDMA5", > > + "PSEUDO_OVL_MDP_RDMA6", > > + "PSEUDO_OVL_MDP_RDMA7", > > + "PSEUDO_OVL_MERGE0", > > + "PSEUDO_OVL_MERGE1", > > + "PSEUDO_OVL_MERGE2", > > + "PSEUDO_OVL_MERGE3", > > + "PSEUDO_OVL_ID_MAX" > > +}; > > + > > +static const char * const private_comp_stem[PSEUDO_OVL_TYPE_NUM] = > > { > > + [PSEUDO_OVL_TYPE_RDMA] = "vdo1_rdma", > > + [PSEUDO_OVL_TYPE_MERGE] = "merge", > > +}; > > + > > +static const struct pseudo_ovl_comp_match > > comp_matches[PSEUDO_OVL_ID_MAX] = { > > + [PSEUDO_OVL_MDP_RDMA0] = { PSEUDO_OVL_TYPE_RDMA, 0 > > }, > > + [PSEUDO_OVL_MDP_RDMA1] = { PSEUDO_OVL_TYPE_RDMA, 1 > > }, > > + [PSEUDO_OVL_MDP_RDMA2] = { PSEUDO_OVL_TYPE_RDMA, 2 > > }, > > + [PSEUDO_OVL_MDP_RDMA3] = { PSEUDO_OVL_TYPE_RDMA, 3 > > }, > > + [PSEUDO_OVL_MDP_RDMA4] = { PSEUDO_OVL_TYPE_RDMA, 4 > > }, > > + [PSEUDO_OVL_MDP_RDMA5] = { PSEUDO_OVL_TYPE_RDMA, 5 > > }, > > + [PSEUDO_OVL_MDP_RDMA6] = { PSEUDO_OVL_TYPE_RDMA, 6 > > }, > > + [PSEUDO_OVL_MDP_RDMA7] = { PSEUDO_OVL_TYPE_RDMA, 7 > > }, > > + [PSEUDO_OVL_MERGE0] = { PSEUDO_OVL_TYPE_MERGE, 1 }, > > + [PSEUDO_OVL_MERGE1] = { PSEUDO_OVL_TYPE_MERGE, 2 }, > > + [PSEUDO_OVL_MERGE2] = { PSEUDO_OVL_TYPE_MERGE, 3 }, > > + [PSEUDO_OVL_MERGE3] = { PSEUDO_OVL_TYPE_MERGE, 4 }, > > +}; > > + > > +static int mtk_pseudo_ovl_fifo_setting(struct mtk_disp_pseudo_ovl > > *pseudo_ovl, > > + struct cmdq_pkt *handle) > > +{ > > + struct mtk_pseudo_ovl_comp *rdma = NULL; > > + const struct pseudo_ovl_data *data = pseudo_ovl->data; > > + const struct mtk_mdp_rdma_fifo *fifo = &data->fifo; > > + int i; > > + > > + for (i = PSEUDO_OVL_MDP_RDMA0; i <= PSEUDO_OVL_MDP_RDMA7; > > i++) { > > + rdma = &pseudo_ovl- > > >pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + i]; > > + mtk_mdp_rdma_fifo_config(rdma->regs, handle, &rdma- > > >cmdq_base, fifo); > > + } > > + > > + return 0; > > +} > > + > > +static void mtk_pseudo_ovl_merge_config(struct mtk_pseudo_ovl_comp > > *comp, > > + struct > > pseudo_ovl_merge_config *merge_cfg, > > + struct cmdq_pkt *cmdq_pkt) > > +{ > > + switch (merge_cfg->merge_mode) { > > + case 6: > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_0); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->out_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_4); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_24); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_25); > > + break; > > + case 18: > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_0); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[1]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_1); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->out_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_4); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_24); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[1]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_25); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_26); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[1]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_27); > > + break; > > + default: > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_0); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->in_w[1]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_1); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->out_w[0]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_4); > > + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | > > merge_cfg->out_w[1]), > > + &comp->cmdq_base, comp->regs, > > DISP_MERGE_CFG_5); > > + break; > > + } > > + > > + mtk_ddp_write(cmdq_pkt, merge_cfg->merge_mode, &comp- > > >cmdq_base, > > + comp->regs, DISP_MERGE_CFG_12); > > + mtk_ddp_write(cmdq_pkt, CFG_10_NO_SWAP, &comp->cmdq_base, > > + comp->regs, DISP_MERGE_CFG_10); > > + mtk_ddp_write_mask(cmdq_pkt, 1, &comp->cmdq_base, comp- > > >regs, > > + DISP_MERGE_ENABLE, MERGE_ENABLE); > > +} > > + > > +void mtk_pseudo_ovl_layer_config(struct device *dev, unsigned int > > idx, > > + struct mtk_plane_state *state, > > + struct cmdq_pkt *cmdq_pkt) > > +{ > > + struct mtk_disp_pseudo_ovl *pseudo_ovl = > > dev_get_drvdata(dev); > > + struct mtk_plane_pending_state *pending = &state->pending; > > + struct pseudo_ovl_merge_config merge_cfg = {0}; > > + struct mtk_mdp_rdma_cfg rdma_config = {0}; > > + struct mtk_pseudo_ovl_comp *rdma_l; > > + struct mtk_pseudo_ovl_comp *rdma_r; > > + struct mtk_pseudo_ovl_comp *merge; > > + const struct drm_format_info *fmt_info = > > drm_format_info(pending->format); > > + bool use_dual_pipe = false; > > + > > + dev_dbg(dev, "%s+ idx:%d, enable:%d, fmt:0x%x\n", __func__, > > idx, > > + pending->enable, pending->format); > > + dev_dbg(dev, "addr 0x%lx, fb w:%d, {%d,%d,%d,%d}\n", > > + pending->addr, (pending->pitch / fmt_info->cpp[0]), > > + pending->x, pending->y, pending->width, pending- > > >height); > > + > > + rdma_l = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE > > + 2 * idx]; > > + rdma_r = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE > > + 2 * idx + 1]; > > + merge = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_MERGE_BASE > > + idx]; > > + > > + if (!pending->enable) { > > + mtk_ddp_write_mask(cmdq_pkt, 0x0, &merge- > > >cmdq_base, merge->regs, > > + DISP_MERGE_ENABLE, > > MERGE_ENABLE); > > + mtk_mdp_rdma_stop(rdma_l->regs, cmdq_pkt, &rdma_l- > > >cmdq_base); > > + mtk_mdp_rdma_stop(rdma_r->regs, cmdq_pkt, &rdma_r- > > >cmdq_base); > > + return; > > + } > > + > > + if (pending->width > MTK_PSEUDO_OVL_SINGLE_PIPE_MAX_WIDTH) > > + use_dual_pipe = true; > > + > > + merge_cfg.out_w[0] = pending->width; > > + merge_cfg.in_h = pending->height; > > + merge_cfg.fmt = pending->format; > > + if (use_dual_pipe) { > > + merge_cfg.merge_mode = CFG12_11_10_1PI_2PO_MERGE; > > + merge_cfg.in_w[0] = (pending->width / 2) + > > ((pending->width / 2) % 2); > > + merge_cfg.in_w[1] = (pending->width / 2) - > > ((pending->width / 2) % 2); > > + } else { > > + merge_cfg.merge_mode = > > CFG12_10_10_1PI_2PO_BUF_MODE; > > + merge_cfg.in_w[0] = pending->width; > > + } > > + > > + mtk_pseudo_ovl_merge_config(merge, &merge_cfg, cmdq_pkt); > > + > > + mtk_mmsys_ddp_config(pseudo_ovl->mmsys_dev, > > MMSYS_CONFIG_MERGE_ASYNC_WIDTH, > > + idx, pending->width / 2, cmdq_pkt); > > + mtk_mmsys_ddp_config(pseudo_ovl->mmsys_dev, > > MMSYS_CONFIG_MERGE_ASYNC_HEIGHT, > > + idx, pending->height, cmdq_pkt); > > + > > + rdma_config.source_width = pending->pitch / fmt_info- > > >cpp[0]; > > + rdma_config.csc_enable = fmt_info->is_yuv ? true : false; > > + rdma_config.profile = RDMA_CSC_FULL709_TO_RGB; > > + rdma_config.encode_type = RDMA_ENCODE_NONE; > > + rdma_config.block_size = RDMA_BLOCK_NONE; > > + rdma_config.width = merge_cfg.in_w[0]; > > + rdma_config.height = pending->height; > > + rdma_config.addr0 = pending->addr; > > + rdma_config.fmt = pending->format; > > + mtk_mdp_rdma_config(rdma_l->regs, &rdma_config, cmdq_pkt, > > &rdma_l->cmdq_base); > > + > > + rdma_config.x_left = merge_cfg.in_w[0]; > > + rdma_config.width = merge_cfg.in_w[1]; > > + mtk_mdp_rdma_config(rdma_r->regs, &rdma_config, cmdq_pkt, > > &rdma_r->cmdq_base); > > + > > + mtk_ddp_write_mask(cmdq_pkt, 0x1, &merge->cmdq_base, merge- > > >regs, > > + DISP_MERGE_ENABLE, MERGE_ENABLE); > > + mtk_ddp_write_mask(cmdq_pkt, 0x0, &merge->cmdq_base, merge- > > >regs, > > + DISP_MERGE_MUTE_0, 0x1); > > + > > + mtk_mdp_rdma_start(rdma_l->regs, cmdq_pkt, &rdma_l- > > >cmdq_base); > > + if (use_dual_pipe) > > + mtk_mdp_rdma_start(rdma_r->regs, cmdq_pkt, &rdma_r- > > >cmdq_base); > > + else > > + mtk_mdp_rdma_stop(rdma_r->regs, cmdq_pkt, &rdma_r- > > >cmdq_base); > > +} > > + > > +void mtk_pseudo_ovl_config(struct device *dev, unsigned int w, > > unsigned int h, > > + unsigned int vrefresh, unsigned int bpc, > > + struct cmdq_pkt *cmdq_pkt) > > +{ > > + struct mtk_disp_pseudo_ovl *pseudo_ovl = > > dev_get_drvdata(dev); > > + > > + dev_info(dev, "%s w:%d, h:%d\n", __func__, w, h); > > + > > + mtk_pseudo_ovl_fifo_setting(pseudo_ovl, cmdq_pkt); > > +} > > + > > +void mtk_pseudo_ovl_start(struct device *dev) > > +{ > > +} > > + > > +void mtk_pseudo_ovl_stop(struct device *dev) > > +{ > > + struct mtk_disp_pseudo_ovl *pseudo_ovl = > > dev_get_drvdata(dev); > > + struct mtk_pseudo_ovl_comp *rdma_l; > > + struct mtk_pseudo_ovl_comp *rdma_r; > > + struct mtk_pseudo_ovl_comp *merge; > > + unsigned int reg; > > + u32 i; > > + > > + for (i = 0; i < pseudo_ovl->data->layer_nr; i++) { > > + rdma_l = &pseudo_ovl- > > >pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * i]; > > + rdma_r = &pseudo_ovl- > > >pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * i + 1]; > > + merge = &pseudo_ovl- > > >pseudo_ovl_comp[PSEUDO_OVL_MERGE_BASE + i]; > > + > > + mtk_mdp_rdma_stop(rdma_l->regs, NULL, &rdma_l- > > >cmdq_base); > > + mtk_mdp_rdma_stop(rdma_r->regs, NULL, &rdma_r- > > >cmdq_base); > > + > > + reg = readl(merge->regs + DISP_MERGE_ENABLE); > > + reg = reg & ~MERGE_ENABLE; > > + writel_relaxed(reg, merge->regs + > > DISP_MERGE_ENABLE); > > + > > + device_reset_optional(merge->dev); > > + } > > +} > > + > > +int mtk_pseudo_ovl_clk_enable(struct device *dev) > > +{ > > + struct mtk_disp_pseudo_ovl *pseudo_ovl = > > dev_get_drvdata(dev); > > + struct mtk_pseudo_ovl_comp *comp; > > + int ret; > > + int i; > > + int j; > > + > > + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_ID_MAX; i++) > > { > > + comp = &pseudo_ovl->pseudo_ovl_comp[i]; > > + if (!comp->dev) > > + continue; > > + > > + /* Need to power on for private rdma devices */ > > + if (i < PSEUDO_OVL_MERGE_BASE) { > > + ret = pm_runtime_get_sync(comp->dev); > > + if (ret < 0) > > + dev_err(dev, > > + "Failed to power on, err > > %d-%s\n", > > + ret, > > pseudo_ovl_comp_str[i]); > > + } > > + > > + for (j = 0; j < ARRAY_SIZE(comp->clks); j++) { > > + if (IS_ERR(comp->clks[j])) > > + break; > > + > > + ret = clk_prepare_enable(comp->clks[j]); > > + if (ret) > > + dev_err(dev, > > + "Failed to enable clock %d, > > err %d-%s\n", > > + i, ret, > > pseudo_ovl_comp_str[i]); > > + } > > + } > > + > > + return ret; > > +} > > + > > +void mtk_pseudo_ovl_clk_disable(struct device *dev) > > +{ > > + struct mtk_disp_pseudo_ovl *pseudo_ovl = > > dev_get_drvdata(dev); > > + struct mtk_pseudo_ovl_comp *comp; > > + int ret; > > + int i; > > + int j; > > + > > + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_ID_MAX; i++) > > { > > + comp = &pseudo_ovl->pseudo_ovl_comp[i]; > > + if (!comp->dev) > > + continue; > > + > > + for (j = 0; i < ARRAY_SIZE(comp->clks); j++) { > > + if (IS_ERR(comp->clks[j])) > > + break; > > + clk_disable_unprepare(comp->clks[j]); > > + } > > + > > + /* Need to power off for private rdma devices */ > > + if (i < PSEUDO_OVL_MERGE_BASE) { > > + ret = pm_runtime_put(comp->dev); > > + if (ret < 0) > > + dev_err(dev, > > + "Failed to power off, err- > > %s\n", > > + ret, > > pseudo_ovl_comp_str[i]); > > + } > > + } > > +} > > + > > +static int pseudo_ovl_comp_get_id(struct device *dev, struct > > device_node *node, > > + enum mtk_pseudo_ovl_comp_type > > type) > > +{ > > + int alias_id = of_alias_get_id(node, > > private_comp_stem[type]); > > + int ret; > > + int i; > > + > > + for (i = 0; i < ARRAY_SIZE(comp_matches); i++) > > + if (comp_matches[i].type == type && > > + comp_matches[i].alias_id == alias_id) > > + return i; > > + > > + dev_err(dev, "Failed to get id. type: %d, alias: %d\n", > > type, alias_id); > > + return -EINVAL; > > +} > > + > > +static int private_comp_init(struct device *dev, struct > > device_node *node, > > + struct mtk_pseudo_ovl_comp *comp, > > + enum mtk_pseudo_ovl_comp_id id) > > +{ > > + struct platform_device *comp_pdev; > > + int ret; > > + int i; > > + > > + if (id < 0 || id >= PSEUDO_OVL_ID_MAX) { > > + dev_err(dev, "Invalid component id %d\n", id); > > + return -EINVAL; > > + } > > + > > + comp_pdev = of_find_device_by_node(node); > > + if (!comp_pdev) { > > + dev_warn(dev, "can't find platform device of > > node:%s\n", > > + node->name); > > + return -ENODEV; > > + } > > + comp->dev = &comp_pdev->dev; > > + comp->regs = of_iomap(node, 0); > > + > > + for (i = 0; i < ARRAY_SIZE(comp->clks); i++) { > > + comp->clks[i] = of_clk_get(node, i); > > + if (IS_ERR(comp->clks[i])) > > + break; > > + } > > + > > +#if IS_REACHABLE(CONFIG_MTK_CMDQ) > > + ret = cmdq_dev_get_client_reg(comp->dev, &comp->cmdq_base, > > 0); > > + if (ret) > > + dev_info(dev, "get mediatek,gce-client-reg > > fail!\n"); > > +#endif > > + > > + if (id < PSEUDO_OVL_MERGE_BASE) > > + pm_runtime_enable(comp->dev); > > + > > + dev_info(dev, "[DRM]regs:0x%p, node:%s\n", comp->regs, > > pseudo_ovl_comp_str[id]); > > + > > + return 0; > > +} > > + > > +static int mtk_disp_pseudo_ovl_comp_probe(struct platform_device > > *pdev) > > +{ > > + return 0; > > +} > > + > > +static int mtk_disp_pseudo_ovl_comp_remove(struct platform_device > > *pdev) > > +{ > > + return 0; > > +} > > + > > +static const struct of_device_id mtk_pseudo_ovl_comp_dt_ids[] = { > > + { > > + .compatible = "mediatek,mt8195-vdo1-rdma", > > + .data = (void *)PSEUDO_OVL_TYPE_RDMA, > > + }, { > > + .compatible = "mediatek,mt8195-vdo1-merge", > > + .data = (void *)PSEUDO_OVL_TYPE_MERGE, > > + }, > > + {}, > > +}; > > + > > +static struct platform_driver mtk_disp_pseudo_ovl_comp_driver = { > > + .probe = mtk_disp_pseudo_ovl_comp_probe, > > + .remove = mtk_disp_pseudo_ovl_comp_remove, > > + .driver = { > > + .name = "mediatek-disp-pseudo-ovl-comp", > > + .owner = THIS_MODULE, > > + .of_match_table = mtk_pseudo_ovl_comp_dt_ids, > > + }, > > +}; > > +module_platform_driver(mtk_disp_pseudo_ovl_comp_driver); > > + > > +static int pseudo_ovl_comp_init(struct device *dev) > > +{ > > + struct mtk_disp_pseudo_ovl *priv = dev_get_drvdata(dev); > > + struct device_node *node, *parent; > > + int i, ret; > > + > > + parent = dev->parent->of_node->parent; > > + > > + for_each_child_of_node(parent, node) { > > + const struct of_device_id *of_id; > > + enum mtk_pseudo_ovl_comp_type type; > > + struct mtk_pseudo_ovl_comp *comp; > > + int id; > > + > > + of_id = of_match_node(mtk_pseudo_ovl_comp_dt_ids, > > node); > > + if (!of_id) > > + continue; > > + > > + if (!of_device_is_available(node)) { > > + dev_info(dev, "Skipping disabled component > > %pOF\n", > > + node); > > + continue; > > + } > > + > > + type = (enum mtk_pseudo_ovl_comp_type)of_id->data; > > + id = pseudo_ovl_comp_get_id(dev, node, type); > > + if (id < 0) { > > + dev_warn(dev, "Skipping unknown component > > %pOF\n", > > + node); > > + continue; > > + } > > + > > + ret = private_comp_init(dev, node, &priv- > > >pseudo_ovl_comp[id], id); > > + if (ret) > > + return ret; > > + } > > + > > + return 0; > > +} > > + > > +static const struct pseudo_ovl_data mt8195_pseudo_ovl_driver_data > > = { > > + .layer_nr = 4, > > + .fifo.read_request_type = 7, > > + .fifo.command_div = 1, > > + .fifo.ext_preutra_en = 1, > > + .fifo.ultra_en = 0, > > + .fifo.pre_ultra_en = 1, > > + .fifo.ext_ultra_en = 1, > > + .fifo.extrd_arb_max_0 = 3, > > + .fifo.buf_resv_size_0 = 0, > > + .fifo.issue_req_th_0 = 0, > > + .fifo.ultra_h_con_0 = 156, > > + .fifo.ultra_l_con_0 = 104, > > +}; > > Now only support one SoC, so remove driver data and change these to > definition instead of variable. > OK, I will remove it. > > + > > +static const struct of_device_id pseudo_ovl_driver_dt_match[] = { > > + { .compatible = "mediatek,mt8195-disp-ethdr", > > + .data = &mt8195_pseudo_ovl_driver_data}, > > + {}, > > +}; > > +MODULE_DEVICE_TABLE(of, mtk_disp_pseudo_ovl_driver_dt_match); > > + > > +static int mtk_disp_pseudo_ovl_probe(struct platform_device *pdev) > > +{ > > + struct mtk_disp_pseudo_ovl *priv; > > + struct device *dev = &pdev->dev; > > + struct device_node *phandle = dev->parent->of_node; > > + const struct of_device_id *of_id; > > + int ret; > > + int i; > > + > > + dev_info(dev, "%s+\n", __func__); > > + > > + of_id = of_match_node(pseudo_ovl_driver_dt_match, phandle); > > + if (!of_id) > > + return -ENODEV; > > + > > + priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL); > > + if (!priv) > > + return -ENOMEM; > > + > > + priv->mmsys_dev = pdev->dev.platform_data; > > + > > + priv->data = of_id->data; > > + platform_set_drvdata(pdev, priv); > > + > > + ret = pseudo_ovl_comp_init(dev); > > + if (ret) { > > + dev_notice(dev, "pseudo_ovl comp init fail\n"); > > + return ret; > > + } > > + > > + dev_info(dev, "%s-\n", __func__); > > + return ret; > > +} > > + > > +static int mtk_disp_pseudo_ovl_remove(struct platform_device > > *pdev) > > +{ > > + struct mtk_disp_pseudo_ovl *pseudo_ovl = > > dev_get_drvdata(&pdev->dev); > > + int i; > > + > > + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_MERGE_BASE; > > i++) > > + pm_runtime_disable(pseudo_ovl- > > >pseudo_ovl_comp[i].dev); > > + > > + return 0; > > +} > > + > > +struct platform_driver mtk_disp_pseudo_ovl_driver = { > > + .probe = mtk_disp_pseudo_ovl_probe, > > + .remove = mtk_disp_pseudo_ovl_remove, > > + .driver = { > > + .name = "mediatek-disp-pseudo-ovl", > > + .owner = THIS_MODULE, > > + }, > > +}; > > +module_platform_driver(mtk_disp_pseudo_ovl_driver); > > diff --git a/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > > b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > > new file mode 100644 > > index 000000000000..b3fe1e1702b8 > > --- /dev/null > > +++ b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h > > @@ -0,0 +1,23 @@ > > +/* SPDX-License-Identifier: GPL-2.0-only */ > > +/* > > + * Copyright (c) 2021 MediaTek Inc. > > + */ > > + > > +#ifndef __MTK_DISP_PSEUDO_OVL_H__ > > +#define __MTK_DISP_PSEUDO_OVL_H__ > > + > > +#include <drm/mediatek_drm.h> > > + > > +void mtk_pseudo_ovl_start(struct device *dev); > > +void mtk_pseudo_ovl_stop(struct device *dev); > > +int mtk_pseudo_ovl_clk_enable(struct device *dev); > > +void mtk_pseudo_ovl_clk_disable(struct device *dev); > > +void mtk_pseudo_ovl_config(struct device *dev, unsigned int w, > > unsigned int h, > > + unsigned int vrefresh, unsigned int bpc, > > + struct cmdq_pkt *cmdq_pkt); > > +void mtk_pseudo_ovl_layer_config(struct device *dev, unsigned int > > idx, > > + struct mtk_plane_state *state, > > + struct cmdq_pkt *cmdq_pkt); > > + > > +#endif > > + > > diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > > b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > > new file mode 100644 > > index 000000000000..81d3cc4872eb > > --- /dev/null > > +++ b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c > > @@ -0,0 +1,456 @@ > > +// SPDX-License-Identifier: GPL-2.0-only > > +/* > > + * Copyright (c) 2021 MediaTek Inc. > > + */ > > + > > +#include <drm/drm_fourcc.h> > > +#include "mtk_drm_drv.h" > > +#include "mtk_mdp_reg_rdma.h" > > +#include "mtk_mdp_rdma.h" > > + > > +#define RDMA_INPUT_SWAP BIT(14) > > +#define RDMA_INPUT_10BIT BIT(18) > > +#define IRQ_INT_EN_ALL \ > > + (REG_FLD_MASK(FLD_UNDERRUN_INT_EN) |\ > > + REG_FLD_MASK(FLD_REG_UPDATE_INT_EN) |\ > > + REG_FLD_MASK(FLD_FRAME_COMPLETE_INT_EN)) > > + > > +static unsigned int rdma_get_y_pitch(unsigned int fmt, unsigned > > int width) > > +{ > > + switch (fmt) { > > + default: > > + case DRM_FORMAT_RGB565: > > + case DRM_FORMAT_BGR565: > > + return 2 * width; > > + case DRM_FORMAT_RGB888: > > + case DRM_FORMAT_BGR888: > > + return 3 * width; > > + case DRM_FORMAT_RGBX8888: > > + case DRM_FORMAT_RGBA8888: > > + case DRM_FORMAT_BGRX8888: > > + case DRM_FORMAT_BGRA8888: > > + case DRM_FORMAT_XRGB8888: > > + case DRM_FORMAT_ARGB8888: > > + case DRM_FORMAT_XBGR8888: > > + case DRM_FORMAT_ABGR8888: > > + case DRM_FORMAT_ARGB2101010: > > + case DRM_FORMAT_ABGR2101010: > > + case DRM_FORMAT_RGBA1010102: > > + case DRM_FORMAT_BGRA1010102: > > + return 4 * width; > > + case DRM_FORMAT_UYVY: > > + case DRM_FORMAT_YUYV: > > + return 2 * width; > > + case DRM_FORMAT_NV12: > > + case DRM_FORMAT_NV21: > > + return 1 * width; > > + } > > +} > > + > > +static unsigned int rdma_get_uv_pitch(unsigned int fmt, unsigned > > int width) > > +{ > > + switch (fmt) { > > + default: > > + return 0; > > + case DRM_FORMAT_NV12: > > + case DRM_FORMAT_NV21: > > + return 4 * width; > > + } > > +} > > + > > +static unsigned int rdma_get_block_h(unsigned int mode) > > +{ > > + switch (mode) { > > + default: > > + return 0; > > + case RDMA_BLOCK_8x8: > > + case RDMA_BLOCK_16x8: > > + case RDMA_BLOCK_32x8: > > + return 8; > > + case RDMA_BLOCK_8x16: > > + case RDMA_BLOCK_16x16: > > + case RDMA_BLOCK_32x16: > > + return 16; > > + case RDMA_BLOCK_8x32: > > + case RDMA_BLOCK_16x32: > > + case RDMA_BLOCK_32x32: > > + return 32; > > + } > > +} > > + > > +static unsigned int rdma_get_horizontal_shift_uv(unsigned int fmt) > > +{ > > + switch (fmt) { > > + default: > > + return 0; > > + case DRM_FORMAT_NV12: > > + case DRM_FORMAT_NV21: > > + return 1; > > + } > > +} > > + > > +static unsigned int rdma_get_vertical_shift_uv(unsigned int fmt) > > +{ > > + switch (fmt) { > > + default: > > + return 0; > > + case DRM_FORMAT_NV12: > > + case DRM_FORMAT_NV21: > > + return 1; > > + } > > +} > > + > > +static unsigned int rdma_get_bits_per_pixel_y(unsigned int fmt) > > +{ > > + switch (fmt) { > > + default: > > + case DRM_FORMAT_RGB565: > > + case DRM_FORMAT_BGR565: > > + return 16; > > + case DRM_FORMAT_RGB888: > > + case DRM_FORMAT_BGR888: > > + return 24; > > + case DRM_FORMAT_RGBX8888: > > + case DRM_FORMAT_RGBA8888: > > + case DRM_FORMAT_BGRX8888: > > + case DRM_FORMAT_BGRA8888: > > + case DRM_FORMAT_XRGB8888: > > + case DRM_FORMAT_ARGB8888: > > + case DRM_FORMAT_XBGR8888: > > + case DRM_FORMAT_ABGR8888: > > + return 32; > > + case DRM_FORMAT_UYVY: > > + case DRM_FORMAT_YUYV: > > + return 16; > > + case DRM_FORMAT_NV12: > > + case DRM_FORMAT_NV21: > > + return 8; > > + } > > +} > > + > > +static unsigned int rdma_get_bits_per_pixel_uv(unsigned int fmt) > > +{ > > + switch (fmt) { > > + default: > > + return 0; > > + case DRM_FORMAT_NV12: > > + case DRM_FORMAT_NV21: > > + return 16; > > + } > > +} > > + > > +static bool with_alpha(uint32_t format) > > +{ > > + const struct drm_format_info *fmt_info = > > drm_format_info(format); > > + > > + return fmt_info->has_alpha; > > +} > > + > > +static unsigned int rdma_fmt_convert(unsigned int fmt) > > +{ > > + switch (fmt) { > > + default: > > + case DRM_FORMAT_RGB565: > > + return RDMA_INPUT_FORMAT_RGB565; > > + case DRM_FORMAT_BGR565: > > + return RDMA_INPUT_FORMAT_RGB565 | RDMA_INPUT_SWAP; > > + case DRM_FORMAT_RGB888: > > + return RDMA_INPUT_FORMAT_RGB888; > > + case DRM_FORMAT_BGR888: > > + return RDMA_INPUT_FORMAT_RGB888 | RDMA_INPUT_SWAP; > > + case DRM_FORMAT_RGBX8888: > > + case DRM_FORMAT_RGBA8888: > > + return RDMA_INPUT_FORMAT_ARGB8888; > > + case DRM_FORMAT_BGRX8888: > > + case DRM_FORMAT_BGRA8888: > > + return RDMA_INPUT_FORMAT_ARGB8888 | > > RDMA_INPUT_SWAP; > > + case DRM_FORMAT_XRGB8888: > > + case DRM_FORMAT_ARGB8888: > > + return RDMA_INPUT_FORMAT_RGBA8888; > > + case DRM_FORMAT_XBGR8888: > > + case DRM_FORMAT_ABGR8888: > > + return RDMA_INPUT_FORMAT_RGBA8888 | > > RDMA_INPUT_SWAP; > > + case DRM_FORMAT_ABGR2101010: > > + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_SWAP > > | > > + RDMA_INPUT_10BIT; > > + case DRM_FORMAT_ARGB2101010: > > + return RDMA_INPUT_FORMAT_RGBA8888 | > > RDMA_INPUT_10BIT; > > + case DRM_FORMAT_RGBA1010102: > > + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_SWAP > > | > > + RDMA_INPUT_10BIT; > > + case DRM_FORMAT_BGRA1010102: > > + return RDMA_INPUT_FORMAT_ARGB8888 | > > RDMA_INPUT_10BIT; > > + case DRM_FORMAT_UYVY: > > + return RDMA_INPUT_FORMAT_UYVY; > > + case DRM_FORMAT_YUYV: > > + return RDMA_INPUT_FORMAT_YUY2; > > + } > > +} > > + > > +void mtk_mdp_rdma_start(void __iomem *base, struct cmdq_pkt > > *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base) > > +{ > > + unsigned int inten = IRQ_INT_EN_ALL; > > + > > + mtk_ddp_write_mask(cmdq_pkt, inten, cmdq_base, base, > > + MDP_RDMA_INTERRUPT_ENABLE, > > IRQ_INT_EN_ALL); > > Why do you enable rdma interrupt? > I will remove it. > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_ROT_ENABLE, > > 1), cmdq_base, > > + base, MDP_RDMA_EN, > > REG_FLD_MASK(FLD_ROT_ENABLE)); > > +} > > + > > +void mtk_mdp_rdma_stop(void __iomem *base, struct cmdq_pkt > > *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base) > > +{ > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_ROT_ENABLE, > > 0), cmdq_base, > > + base, MDP_RDMA_EN, > > REG_FLD_MASK(FLD_ROT_ENABLE)); > > + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, > > + MDP_RDMA_INTERRUPT_ENABLE, > > IRQ_INT_EN_ALL); > > + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, > > + MDP_RDMA_INTERRUPT_STATUS, > > IRQ_INT_EN_ALL); > > + mtk_ddp_write_mask(cmdq_pkt, 1, cmdq_base, base, > > MDP_RDMA_RESET, ~0); > > + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, > > MDP_RDMA_RESET, ~0); > > +} > > + > > +void mtk_mdp_rdma_fifo_config(void __iomem *base, struct cmdq_pkt > > *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base, > > + const struct mtk_mdp_rdma_fifo *fifo) > > Why do you pass fifo config from ovl to rdma? > I think just rdma know how to set these fifo parameter. > I will move fifo setting inside the rdma driver. > > +{ > > + int reg; > > + int reg_val; > > + int reg_mask; > > + > > + reg = MDP_RDMA_GMCIF_CON; > > + reg_val = REG_FLD_VAL(FLD_RD_REQ_TYPE, fifo- > > >read_request_type) | > > + REG_FLD_VAL(FLD_COMMAND_DIV, fifo->command_div) | > > + REG_FLD_VAL(FLD_EXT_PREULTRA_EN, fifo- > > >ext_preutra_en) | > > + REG_FLD_VAL(FLD_ULTRA_EN, fifo->ultra_en) | > > + REG_FLD_VAL(PRE_ULTRA_EN, fifo->pre_ultra_en) | > > + REG_FLD_VAL(FLD_EXT_ULTRA_EN, fifo- > > >ext_ultra_en); > > + reg_mask = REG_FLD_MASK(FLD_RD_REQ_TYPE) | > > + REG_FLD_MASK(FLD_COMMAND_DIV) | > > + REG_FLD_MASK(FLD_EXT_PREULTRA_EN) | > > + REG_FLD_MASK(FLD_ULTRA_EN) | > > + REG_FLD_MASK(PRE_ULTRA_EN) | > > + REG_FLD_MASK(FLD_EXT_ULTRA_EN); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_DMA_CON_0; > > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX, fifo- > > >extrd_arb_max_0) | > > + REG_FLD_VAL(FLD_BUF_RESV_SIZE, fifo- > > >buf_resv_size_0) | > > + REG_FLD_VAL(FLD_ISSUE_REQ_TH, fifo- > > >issue_req_th_0); > > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX) | > > + REG_FLD_MASK(FLD_BUF_RESV_SIZE) | > > + REG_FLD_MASK(FLD_ISSUE_REQ_TH); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_UTRA_H_CON_0; > > + reg_val = REG_FLD_VAL(FLD_PREUTRA_H_OFS_0, fifo- > > >ultra_h_con_0); > > + reg_mask = REG_FLD_MASK(FLD_PREUTRA_H_OFS_0); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_UTRA_L_CON_0; > > + reg_val = REG_FLD_VAL(FLD_PREUTRA_L_OFS_0, fifo- > > >ultra_l_con_0); > > + reg_mask = REG_FLD_MASK(FLD_PREUTRA_L_OFS_0); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_DMABUF_CON_1; > > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_1, 0) | > > + REG_FLD_VAL(FLD_BUF_RESV_SIZE_1, 0) | > > + REG_FLD_VAL(FLD_ISSUE_REQ_TH_1, 0); > > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_1) | > > + REG_FLD_MASK(FLD_BUF_RESV_SIZE_1) | > > + REG_FLD_MASK(FLD_ISSUE_REQ_TH_1); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_ULTRA_TH_HIGH_CON_1; > > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_1, 0); > > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_1); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_ULTRA_TH_LOW_CON_1; > > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_1, 0); > > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_1); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_DMABUF_CON_2; > > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_2, 0) | > > + REG_FLD_VAL(FLD_BUF_RESV_SIZE_2, 0) | > > + REG_FLD_VAL(FLD_ISSUE_REQ_TH_2, 0); > > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_2) | > > + REG_FLD_MASK(FLD_BUF_RESV_SIZE_2) | > > + REG_FLD_MASK(FLD_ISSUE_REQ_TH_2); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_UTRA_H_CON_2; > > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_2, 0); > > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_2); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_ULTRA_TH_LOW_CON_2; > > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_2, 0); > > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_2); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_DMABUF_CON_3; > > + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_3, 0) | > > + REG_FLD_VAL(FLD_BUF_RESV_SIZE_3, 0) | > > + REG_FLD_VAL(FLD_ISSUE_REQ_TH_3, 0); > > + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_3) | > > + REG_FLD_MASK(FLD_BUF_RESV_SIZE_3) | > > + REG_FLD_MASK(FLD_ISSUE_REQ_TH_3); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_UTRA_H_CON_3; > > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_3, 0); > > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_3); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > + > > + reg = MDP_RDMA_ULTRA_TH_LOW_CON_3; > > + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_3, 0); > > + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_3); > > + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, > > reg_mask); > > +} > > + > > +void mtk_mdp_rdma_config(void __iomem *base, struct > > mtk_mdp_rdma_cfg *cfg, > > + struct cmdq_pkt *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base) > > +{ > > + unsigned int src_pitch_uv = rdma_get_uv_pitch(cfg->fmt, > > cfg->source_width); > > + unsigned int src_pitch_y = rdma_get_y_pitch(cfg->fmt, cfg- > > >source_width); > > This is got from mtk_plane_atomic_update(). > OK,I will modify it. > > + unsigned int h_shift_uv = rdma_get_horizontal_shift_uv(cfg- > > >fmt); > > + unsigned int v_shift_uv = rdma_get_vertical_shift_uv(cfg- > > >fmt); > > + unsigned int bpp_uv = rdma_get_bits_per_pixel_uv(cfg->fmt); > > + unsigned int block_h = rdma_get_block_h(cfg->block_size); > > block_h = 0; OK, I will modify it. > > > + unsigned int bpp_y = rdma_get_bits_per_pixel_y(cfg->fmt); > > drm_format_info(). OK, I will modify it. > > > + unsigned int y_start_line = 0; > > + unsigned int offset_y = 0; > > + unsigned int offset_u = 0; > > + unsigned int offset_v = 0; > > + > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_UNIFORM_CONFIG, 1), > > + cmdq_base, base, MDP_RDMA_SRC_CON, > > + REG_FLD_MASK(FLD_UNIFORM_CONFIG)); > > + mtk_ddp_write_mask(cmdq_pkt, rdma_fmt_convert(cfg->fmt), > > cmdq_base, > > + base, MDP_RDMA_SRC_CON, > > REG_FLD_MASK(FLD_SWAP) | > > + REG_FLD_MASK(FLD_SRC_FORMAT) | > > + REG_FLD_MASK(FLD_BIT_NUMBER)); > > + > > + if (!cfg->csc_enable && with_alpha(cfg->fmt)) > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_OUTPUT_ARGB, 1), > > + cmdq_base, base, > > MDP_RDMA_SRC_CON, > > + REG_FLD_MASK(FLD_OUTPUT_ARGB)); > > + else > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_OUTPUT_ARGB, 0), > > + cmdq_base, base, > > MDP_RDMA_SRC_CON, > > + REG_FLD_MASK(FLD_OUTPUT_ARGB)); > > + > > + mtk_ddp_write_mask(cmdq_pkt, cfg->addr0, cmdq_base, base, > > + MDP_RDMA_SRC_BASE_0, > > REG_FLD_MASK(FLD_SRC_BASE_0)); > > + mtk_ddp_write_mask(cmdq_pkt, cfg->addr1, cmdq_base, base, > > + MDP_RDMA_SRC_BASE_1, > > REG_FLD_MASK(FLD_SRC_BASE_1)); > > + mtk_ddp_write_mask(cmdq_pkt, cfg->addr2, cmdq_base, base, > > + MDP_RDMA_SRC_BASE_2, > > REG_FLD_MASK(FLD_SRC_BASE_2)); > > + > > + mtk_ddp_write_mask(cmdq_pkt, src_pitch_y, cmdq_base, base, > > + MDP_RDMA_MF_BKGD_SIZE_IN_BYTE, > > + REG_FLD_MASK(FLD_MF_BKGD_WB)); > > + mtk_ddp_write_mask(cmdq_pkt, src_pitch_uv, cmdq_base, base, > > + MDP_RDMA_SF_BKGD_SIZE_IN_BYTE, > > + REG_FLD_MASK(FLD_SF_BKGD_WB)); > > + > > + if (cfg->encode_type == RDMA_ENCODE_AFBC) { > > Never be here, so remove. > OK, I will remove it. > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_MF_BKGD_WP, cfg->source_width), > > + cmdq_base, base, > > MDP_RDMA_MF_BKGD_SIZE_IN_PIXEL, > > + REG_FLD_MASK(FLD_MF_BKGD_WP)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_BKGD_HP, cfg->height), > > + cmdq_base, base, > > MDP_RDMA_MF_BKGD_H_SIZE_IN_PIXEL, > > + REG_FLD_MASK(FLD_BKGD_HP)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_AFBC_YUV_TRANSFORM, 1), > > + cmdq_base, base, > > MDP_RDMA_COMP_CON, > > + REG_FLD_MASK(FLD_AFBC_YUV_TRANSF > > ORM)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_UFBDC_EN, 1), cmdq_base, > > + base, MDP_RDMA_COMP_CON, > > REG_FLD_MASK(FLD_UFBDC_EN)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_AFBC_EN, 1), cmdq_base, > > + base, MDP_RDMA_COMP_CON, > > REG_FLD_MASK(FLD_AFBC_EN)); > > + } else { > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_AFBC_YUV_TRANSFORM, 0), > > + cmdq_base, base, > > MDP_RDMA_COMP_CON, > > + REG_FLD_MASK(FLD_AFBC_YUV_TRANSF > > ORM)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_UFBDC_EN, 0), cmdq_base, > > + base, MDP_RDMA_COMP_CON, > > REG_FLD_MASK(FLD_UFBDC_EN)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_AFBC_EN, 0), cmdq_base, > > + base, MDP_RDMA_COMP_CON, > > REG_FLD_MASK(FLD_AFBC_EN)); > > + } > > + > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_10B, > > 1), cmdq_base, > > + base, MDP_RDMA_CON, > > REG_FLD_MASK(FLD_OUTPUT_10B)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SIMPLE_MODE, > > 1), cmdq_base, > > + base, MDP_RDMA_CON, > > REG_FLD_MASK(FLD_SIMPLE_MODE)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_TRANS_EN, cfg- > > >csc_enable), > > + cmdq_base, base, MDP_RDMA_TRANSFORM_0, > > + REG_FLD_MASK(FLD_TRANS_EN)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_INT_MATRIX_SEL, cfg->profile), > > + cmdq_base, base, MDP_RDMA_TRANSFORM_0, > > + REG_FLD_MASK(FLD_INT_MATRIX_SEL)); > > + > > + if (cfg->block_size == RDMA_BLOCK_NONE) { > > + y_start_line = cfg->y_top; > > + > > + offset_y = (cfg->x_left * bpp_y >> 3) + > > y_start_line * src_pitch_y; > > + offset_u = ((cfg->x_left >> h_shift_uv) * bpp_uv > > >> 3) + > > + (y_start_line >> v_shift_uv) * > > src_pitch_uv; > > + offset_v = ((cfg->x_left >> h_shift_uv) * bpp_uv > > >> 3) + > > + (y_start_line >> v_shift_uv) * > > src_pitch_uv; > > + } else { > > Never be here, so remove. OK, I will remove it. > > > + offset_y = (cfg->x_left * block_h * bpp_y >> 3) + > > + (cfg->y_top) * src_pitch_y; > > + offset_u = ((cfg->x_left >> h_shift_uv) * (block_h > > >> v_shift_uv) * > > + bpp_uv >> 3) + (cfg->y_top) * > > src_pitch_uv; > > + offset_v = ((cfg->x_left >> h_shift_uv) * (block_h > > >> v_shift_uv) * > > + bpp_uv >> 3) + (cfg->y_top) * > > src_pitch_uv; > > + } > > + > > + if (cfg->encode_type == RDMA_ENCODE_AFBC) { > > Never be here, so remove. OK, I will remove it. > > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_SRC_OFFSET_WP, cfg->x_left), > > + cmdq_base, base, > > MDP_RDMA_SRC_OFFSET_WP, > > + REG_FLD_MASK(FLD_SRC_OFFSET_WP)) > > ; > > + > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_SRC_OFFSET_HP, cfg->y_top), > > + cmdq_base, base, > > MDP_RDMA_SRC_OFFSET_HP, > > + REG_FLD_MASK(FLD_SRC_OFFSET_HP)) > > ; > > + } > > + > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_0, > > offset_y), > > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_0, > > + REG_FLD_MASK(FLD_SRC_OFFSET_0)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_1, > > offset_u), > > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_1, > > + REG_FLD_MASK(FLD_SRC_OFFSET_1)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_2, > > offset_v), > > + cmdq_base, base, MDP_RDMA_SRC_OFFSET_2, > > + REG_FLD_MASK(FLD_SRC_OFFSET_2)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_SRC_W, cfg- > > >width), > > + cmdq_base, base, MDP_RDMA_MF_SRC_SIZE, > > + REG_FLD_MASK(FLD_MF_SRC_W)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_SRC_H, cfg- > > >height), > > + cmdq_base, base, MDP_RDMA_MF_SRC_SIZE, > > + REG_FLD_MASK(FLD_MF_SRC_H)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_CLIP_W, > > cfg->width), > > + cmdq_base, base, MDP_RDMA_MF_CLIP_SIZE, > > + REG_FLD_MASK(FLD_MF_CLIP_W)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_CLIP_H, > > cfg->height), > > + cmdq_base, base, MDP_RDMA_MF_CLIP_SIZE, > > + REG_FLD_MASK(FLD_MF_CLIP_H)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_OFFSET_W_1, > > 0), > > + cmdq_base, base, MDP_RDMA_MF_OFFSET_1, > > + REG_FLD_MASK(FLD_MF_OFFSET_W_1)); > > + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_OFFSET_H_1, > > 0), > > + cmdq_base, base, MDP_RDMA_MF_OFFSET_1, > > + REG_FLD_MASK(FLD_MF_OFFSET_H_1)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_LINE_THRESHOLD, cfg->height), > > + cmdq_base, base, MDP_RDMA_TARGET_LINE, > > + REG_FLD_MASK(FLD_LINE_THRESHOLD)); > > + mtk_ddp_write_mask(cmdq_pkt, > > REG_FLD_VAL(FLD_TARGET_LINE_EN, 1), > > + cmdq_base, base, MDP_RDMA_TARGET_LINE, > > + REG_FLD_MASK(FLD_TARGET_LINE_EN)); > > +} > > diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > > b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > > new file mode 100644 > > index 000000000000..c16bfb716610 > > --- /dev/null > > +++ b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h > > @@ -0,0 +1,109 @@ > > +/* SPDX-License-Identifier: GPL-2.0-only */ > > +/* > > + * Copyright (c) 2021 MediaTek Inc. > > + */ > > + > > +#ifndef __MTK_MDP_RDMA_H__ > > +#define __MTK_MDP_RDMA_H__ > > + > > +enum rdma_format { > > + RDMA_INPUT_FORMAT_RGB565 = 0, > > + RDMA_INPUT_FORMAT_RGB888 = 1, > > + RDMA_INPUT_FORMAT_RGBA8888 = 2, > > + RDMA_INPUT_FORMAT_ARGB8888 = 3, > > + RDMA_INPUT_FORMAT_UYVY = 4, > > + RDMA_INPUT_FORMAT_YUY2 = 5, > > + RDMA_INPUT_FORMAT_Y8 = 7, > > + RDMA_INPUT_FORMAT_YV12 = 8, > > + RDMA_INPUT_FORMAT_UYVY_3PL = 9, > > + RDMA_INPUT_FORMAT_NV12 = 12, > > + RDMA_INPUT_FORMAT_UYVY_2PL = 13, > > + RDMA_INPUT_FORMAT_Y410 = 14 > > +}; > > + > > +enum rdma_profile { > > + RDMA_CSC_RGB_TO_JPEG = 0, > > + RDMA_CSC_RGB_TO_FULL709 = 1, > > + RDMA_CSC_RGB_TO_BT601 = 2, > > + RDMA_CSC_RGB_TO_BT709 = 3, > > + RDMA_CSC_JPEG_TO_RGB = 4, > > + RDMA_CSC_FULL709_TO_RGB = 5, > > + RDMA_CSC_BT601_TO_RGB = 6, > > + RDMA_CSC_BT709_TO_RGB = 7, > > + RDMA_CSC_JPEG_TO_BT601 = 8, > > + RDMA_CSC_JPEG_TO_BT709 = 9, > > + RDMA_CSC_BT601_TO_JPEG = 10, > > + RDMA_CSC_BT709_TO_BT601 = 11, > > + RDMA_CSC_BT601_TO_BT709 = 12 > > +}; > > + > > +enum rdma_encode { > > + RDMA_ENCODE_NONE = 0, > > + RDMA_ENCODE_AFBC = 1, > > + RDMA_ENCODE_HYFBC = 2, > > + RDMA_ENCODE_UFO_DCP = 3 > > +}; > > + > > +enum rdma_block { > > + RDMA_BLOCK_NONE = 0, > > + RDMA_BLOCK_8x8 = 1, > > + RDMA_BLOCK_8x16 = 2, > > + RDMA_BLOCK_8x32 = 3, > > + RDMA_BLOCK_16x8 = 4, > > + RDMA_BLOCK_16x16 = 5, > > + RDMA_BLOCK_16x32 = 6, > > + RDMA_BLOCK_32x8 = 7, > > + RDMA_BLOCK_32x16 = 8, > > + RDMA_BLOCK_32x32 = 9 > > +}; > > + > > +struct mtk_mdp_rdma_cfg { > > + enum rdma_encode encode_type; > > + enum rdma_block block_size; > > + enum rdma_profile profile; > > + unsigned int source_width; > > + unsigned int addr0; > > + unsigned int addr1; > > + unsigned int addr2; > > + unsigned int width; > > + unsigned int height; > > + unsigned int x_left; > > + unsigned int y_top; > > + bool csc_enable; > > + int fmt; > > +}; > > + > > +struct mtk_mdp_rdma_fifo { > > + int read_request_type; > > + int command_div; > > + int ext_preutra_en; > > + int ultra_en; > > + int pre_ultra_en; > > + int ext_ultra_en; > > + int extrd_arb_max_0; > > + int buf_resv_size_0; > > + int issue_req_th_0; > > + int ultra_h_con_0; > > + int ultra_l_con_0; > > +}; > > Move rdma internal definition in .c file. > OK, I will move it to internal .c file. > > + > > +void mtk_mdp_rdma_start(void __iomem *base, > > + struct cmdq_pkt *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base); > > + > > +void mtk_mdp_rdma_stop(void __iomem *base, > > + struct cmdq_pkt *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base); > > + > > +void mtk_mdp_rdma_fifo_config(void __iomem *base, > > + struct cmdq_pkt *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base, > > + const struct mtk_mdp_rdma_fifo > > *fifo); > > + > > +void mtk_mdp_rdma_config(void __iomem *base, > > + struct mtk_mdp_rdma_cfg *cfg, > > + struct cmdq_pkt *cmdq_pkt, > > + struct cmdq_client_reg *cmdq_base); > > Move these prototype to mtk_disp_drv.h > OK. > > + > > +#endif // __MTK_MDP_RDMA_H__ > > + > > diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > > b/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > > new file mode 100644 > > index 000000000000..08abd9f39bd8 > > --- /dev/null > > +++ b/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h > > @@ -0,0 +1,160 @@ > > +/* SPDX-License-Identifier: GPL-2.0-only */ > > +/* > > + * Copyright (c) 2021 MediaTek Inc. > > + */ > > + > > +#ifndef __MDP_RDMA_REGS_H__ > > +#define __MDP_RDMA_REGS_H__ > > + > > +#define REG_FLD(width, shift) \ > > + ((unsigned int)((((width) & 0xff) << 16) | ((shift) & > > 0xff))) > > + > > +#define REG_FLD_WIDTH(field) ((unsigned int)(((field) >> 16) & > > 0xff)) > > + > > +#define REG_FLD_SHIFT(field) ((unsigned int)((field) & 0xff)) > > + > > +#define REG_FLD_MASK(field) \ > > + ((unsigned int)((1ULL << REG_FLD_WIDTH(field)) - 1) \ > > + << REG_FLD_SHIFT(field)) > > + > > +#define REG_FLD_VAL(field, val) \ > > + (((val) << REG_FLD_SHIFT(field)) & REG_FLD_MASK(field)) > > + > > +#define > > MDP_RDMA_EN 0x000 > > +#define FLD_ROT_ENABLE REG_FLD(1, 0) > > + > > +#define > > MDP_RDMA_RESET 0x008 > > + > > +#define > > MDP_RDMA_INTERRUPT_ENABLE 0x010 > > +#define FLD_UNDERRUN_INT_EN REG_FLD(1, 2) > > +#define FLD_REG_UPDATE_INT_EN REG_FLD(1, 1) > > +#define FLD_FRAME_COMPLETE_INT_EN REG_FLD(1, 0) > > + > > +#define > > MDP_RDMA_INTERRUPT_STATUS 0x018 > > + > > +#define > > MDP_RDMA_CON 0x020 > > +#define FLD_OUTPUT_10B REG_FLD(1, 5) > > +#define FLD_SIMPLE_MODE REG_FLD(1, 4) > > + > > +#define > > MDP_RDMA_GMCIF_CON 0x028 > > +#define FLD_EXT_ULTRA_EN REG_FLD(1, 18) > > +#define PRE_ULTRA_EN REG_FLD(2, 16) > > +#define FLD_ULTRA_EN REG_FLD(2, 12) > > +#define FLD_RD_REQ_TYPE REG_FLD(4, 4) > > +#define FLD_EXT_PREULTRA_EN REG_FLD(1, 3) > > +#define FLD_COMMAND_DIV REG_FLD(1, 0) > > + > > +#define > > MDP_RDMA_SRC_CON 0x030 > > +#define FLD_OUTPUT_ARGB REG_FLD(1, 25) > > +#define FLD_BIT_NUMBER REG_FLD(2, 18) > > +#define FLD_UNIFORM_CONFIG REG_FLD(1, 17) > > +#define FLD_SWAP REG_FLD(1, 14) > > +#define FLD_SRC_FORMAT REG_FLD(4, 0) > > + > > +#define > > MDP_RDMA_COMP_CON 0x038 > > +#define FLD_AFBC_EN REG_FLD(1, 22) > > +#define FLD_AFBC_YUV_TRANSFORM REG_FLD(1, 21) > > +#define FLD_UFBDC_EN REG_FLD(1, 12) > > + > > +#define > > MDP_RDMA_MF_BKGD_SIZE_IN_BYTE 0x060 > > +#define FLD_MF_BKGD_WB REG_FLD(23, 0) > > + > > +#define > > MDP_RDMA_MF_BKGD_SIZE_IN_PIXEL 0x068 > > +#define FLD_MF_BKGD_WP REG_FLD(23, 0) > > + > > +#define > > MDP_RDMA_MF_SRC_SIZE 0x070 > > +#define FLD_MF_SRC_H REG_FLD(15, 16) > > +#define FLD_MF_SRC_W REG_FLD(15, 0) > > + > > +#define > > MDP_RDMA_MF_CLIP_SIZE 0x078 > > +#define FLD_MF_CLIP_H REG_FLD(15, 16) > > +#define FLD_MF_CLIP_W REG_FLD(15, 0) > > + > > +#define > > MDP_RDMA_MF_OFFSET_1 0x080 > > +#define FLD_MF_OFFSET_H_1 REG_FLD(6, 16) > > +#define FLD_MF_OFFSET_W_1 REG_FLD(5, 0) > > + > > +#define > > MDP_RDMA_SF_BKGD_SIZE_IN_BYTE 0x090 > > +#define FLD_SF_BKGD_WB REG_FLD(23, 0) > > + > > +#define > > MDP_RDMA_MF_BKGD_H_SIZE_IN_PIXEL 0x098 > > +#define FLD_BKGD_HP REG_FLD(23, 0) > > + > > +#define > > MDP_RDMA_TARGET_LINE 0x0a0 > > +#define FLD_LINE_THRESHOLD REG_FLD(15, 17) > > +#define FLD_TARGET_LINE_EN REG_FLD(1, 16) > > + > > +#define > > MDP_RDMA_SRC_OFFSET_0 0x118 > > +#define FLD_SRC_OFFSET_0 REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_SRC_OFFSET_1 0x120 > > +#define FLD_SRC_OFFSET_1 REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_SRC_OFFSET_2 0x128 > > +#define FLD_SRC_OFFSET_2 REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_SRC_OFFSET_WP 0x148 > > +#define FLD_SRC_OFFSET_WP REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_SRC_OFFSET_HP 0x150 > > +#define FLD_SRC_OFFSET_HP REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_TRANSFORM_0 0x200 > > +#define FLD_INT_MATRIX_SEL REG_FLD(5, 23) > > +#define FLD_TRANS_EN REG_FLD(1, 16) > > + > > +#define > > MDP_RDMA_DMA_CON_0 0x240 > > +#define FLD_EXTRD_ARB_MAX REG_FLD(4, 24) > > +#define FLD_BUF_RESV_SIZE REG_FLD(8, 16) > > +#define FLD_ISSUE_REQ_TH REG_FLD(8, 0) > > + > > +#define > > MDP_RDMA_UTRA_H_CON_0 0x248 > > +#define FLD_PREUTRA_H_OFS_0 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_UTRA_L_CON_0 0x250 > > +#define FLD_PREUTRA_L_OFS_0 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_DMABUF_CON_1 0x258 > > +#define FLD_EXTRD_ARB_MAX_1 REG_FLD(4, 24) > > +#define FLD_BUF_RESV_SIZE_1 REG_FLD(7, 16) > > +#define FLD_ISSUE_REQ_TH_1 REG_FLD(7, 0) > > + > > +#define > > MDP_RDMA_ULTRA_TH_HIGH_CON_1 0x260 > > +#define FLD_PRE_ULTRA_TH_HIGH_OFS_1 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_ULTRA_TH_LOW_CON_1 0x268 > > +#define FLD_PRE_ULTRA_TH_LOW_OFS_1 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_DMABUF_CON_2 0x270 > > +#define FLD_EXTRD_ARB_MAX_2 REG_FLD(4, 24) > > +#define FLD_BUF_RESV_SIZE_2 REG_FLD(6, 16) > > +#define FLD_ISSUE_REQ_TH_2 REG_FLD(6, 0) > > + > > +#define > > MDP_RDMA_UTRA_H_CON_2 0x278 > > +#define FLD_PRE_ULTRA_TH_HIGH_OFS_2 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_ULTRA_TH_LOW_CON_2 0x280 > > +#define FLD_PRE_ULTRA_TH_LOW_OFS_2 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_DMABUF_CON_3 0x288 > > +#define FLD_EXTRD_ARB_MAX_3 REG_FLD(4, 24) > > +#define FLD_BUF_RESV_SIZE_3 REG_FLD(6, 16) > > +#define FLD_ISSUE_REQ_TH_3 REG_FLD(6, 0) > > + > > +#define > > MDP_RDMA_UTRA_H_CON_3 0x290 > > +#define FLD_PRE_ULTRA_TH_HIGH_OFS_3 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_ULTRA_TH_LOW_CON_3 0x298 > > +#define FLD_PRE_ULTRA_TH_LOW_OFS_3 REG_FLD(10, 10) > > + > > +#define > > MDP_RDMA_SRC_BASE_0 0xf00 > > +#define FLD_SRC_BASE_0 REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_SRC_BASE_1 0xf08 > > +#define FLD_SRC_BASE_1 REG_FLD(32, 0) > > + > > +#define > > MDP_RDMA_SRC_BASE_2 0xf10 > > +#define FLD_SRC_BASE_2 REG_FLD(32, 0) > > + > > +#endif /* __MDP_RDMA_REGS_H__ */ > > Only one file need these, so move register definiton in .c file. > OK. > Regards, > Chun-Kuang. > > > + > > -- > > 2.18.0 > >
diff --git a/drivers/gpu/drm/mediatek/Makefile b/drivers/gpu/drm/mediatek/Makefile index 27c89847d43b..31613564f499 100644 --- a/drivers/gpu/drm/mediatek/Makefile +++ b/drivers/gpu/drm/mediatek/Makefile @@ -13,7 +13,9 @@ mediatek-drm-y := mtk_disp_ccorr.o \ mtk_drm_gem.o \ mtk_drm_plane.o \ mtk_dsi.o \ - mtk_dpi.o + mtk_dpi.o \ + mtk_disp_pseudo_ovl.o \ + mtk_mdp_rdma.o obj-$(CONFIG_DRM_MEDIATEK) += mediatek-drm.o diff --git a/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c new file mode 100644 index 000000000000..0446fa99dd0a --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c @@ -0,0 +1,593 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2021 MediaTek Inc. + */ + +#include <drm/drm_fourcc.h> +#include <linux/clk.h> +#include <linux/reset.h> +#include <linux/component.h> +#include <linux/of_device.h> +#include <linux/of_address.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/soc/mediatek/mtk-mmsys.h> +#include <linux/soc/mediatek/mtk-cmdq.h> + +#include "mtk_drm_drv.h" +#include "mtk_drm_crtc.h" +#include "mtk_drm_ddp_comp.h" +#include "mtk_mdp_rdma.h" +#include "mtk_disp_pseudo_ovl.h" + +#define DISP_MERGE_ENABLE 0x0 + #define MERGE_ENABLE BIT(0) +#define DISP_MERGE_CFG_0 0x10 +#define DISP_MERGE_CFG_1 0x14 +#define DISP_MERGE_CFG_4 0x20 +#define DISP_MERGE_CFG_5 0x24 +#define DISP_MERGE_CFG_10 0x38 + #define CFG_10_NO_SWAP 0 +#define DISP_MERGE_CFG_12 0x40 + #define CFG12_10_10_1PI_2PO_BUF_MODE 6 + #define CFG12_11_10_1PI_2PO_MERGE 18 +#define DISP_MERGE_CFG_24 0x70 +#define DISP_MERGE_CFG_25 0x74 +#define DISP_MERGE_CFG_26 0x78 +#define DISP_MERGE_CFG_27 0x7c +#define DISP_MERGE_MUTE_0 0xf00 + +#define MTK_PSEUDO_OVL_SINGLE_PIPE_MAX_WIDTH 1920 + +enum mtk_pseudo_ovl_comp_type { + PSEUDO_OVL_TYPE_RDMA = 0, + PSEUDO_OVL_TYPE_MERGE, + PSEUDO_OVL_TYPE_NUM, +}; + +enum mtk_pseudo_ovl_comp_id { + PSEUDO_OVL_RDMA_BASE = 0, + PSEUDO_OVL_MDP_RDMA0 = PSEUDO_OVL_RDMA_BASE, + PSEUDO_OVL_MDP_RDMA1, + PSEUDO_OVL_MDP_RDMA2, + PSEUDO_OVL_MDP_RDMA3, + PSEUDO_OVL_MDP_RDMA4, + PSEUDO_OVL_MDP_RDMA5, + PSEUDO_OVL_MDP_RDMA6, + PSEUDO_OVL_MDP_RDMA7, + PSEUDO_OVL_MERGE_BASE, + PSEUDO_OVL_MERGE0 = PSEUDO_OVL_MERGE_BASE, + PSEUDO_OVL_MERGE1, + PSEUDO_OVL_MERGE2, + PSEUDO_OVL_MERGE3, + PSEUDO_OVL_ID_MAX +}; + +struct pseudo_ovl_data { + unsigned int layer_nr; + struct mtk_mdp_rdma_fifo fifo; +}; + +struct pseudo_ovl_comp_match { + enum mtk_pseudo_ovl_comp_type type; + int alias_id; +}; + +struct pseudo_ovl_merge_config { + unsigned int fmt; + unsigned int merge_mode; + unsigned int in_w[2]; + unsigned int out_w[2]; + unsigned int in_h; +}; + +struct mtk_pseudo_ovl_comp { + struct device *dev; + struct clk *clks[2]; + struct cmdq_client_reg cmdq_base; + void __iomem *regs; +}; + +struct mtk_disp_pseudo_ovl { + struct mtk_pseudo_ovl_comp pseudo_ovl_comp[PSEUDO_OVL_ID_MAX]; + const struct pseudo_ovl_data *data; + struct device *mmsys_dev; +}; + +static const char * const pseudo_ovl_comp_str[] = { + "PSEUDO_OVL_MDP_RDMA0", + "PSEUDO_OVL_MDP_RDMA1", + "PSEUDO_OVL_MDP_RDMA2", + "PSEUDO_OVL_MDP_RDMA3", + "PSEUDO_OVL_MDP_RDMA4", + "PSEUDO_OVL_MDP_RDMA5", + "PSEUDO_OVL_MDP_RDMA6", + "PSEUDO_OVL_MDP_RDMA7", + "PSEUDO_OVL_MERGE0", + "PSEUDO_OVL_MERGE1", + "PSEUDO_OVL_MERGE2", + "PSEUDO_OVL_MERGE3", + "PSEUDO_OVL_ID_MAX" +}; + +static const char * const private_comp_stem[PSEUDO_OVL_TYPE_NUM] = { + [PSEUDO_OVL_TYPE_RDMA] = "vdo1_rdma", + [PSEUDO_OVL_TYPE_MERGE] = "merge", +}; + +static const struct pseudo_ovl_comp_match comp_matches[PSEUDO_OVL_ID_MAX] = { + [PSEUDO_OVL_MDP_RDMA0] = { PSEUDO_OVL_TYPE_RDMA, 0 }, + [PSEUDO_OVL_MDP_RDMA1] = { PSEUDO_OVL_TYPE_RDMA, 1 }, + [PSEUDO_OVL_MDP_RDMA2] = { PSEUDO_OVL_TYPE_RDMA, 2 }, + [PSEUDO_OVL_MDP_RDMA3] = { PSEUDO_OVL_TYPE_RDMA, 3 }, + [PSEUDO_OVL_MDP_RDMA4] = { PSEUDO_OVL_TYPE_RDMA, 4 }, + [PSEUDO_OVL_MDP_RDMA5] = { PSEUDO_OVL_TYPE_RDMA, 5 }, + [PSEUDO_OVL_MDP_RDMA6] = { PSEUDO_OVL_TYPE_RDMA, 6 }, + [PSEUDO_OVL_MDP_RDMA7] = { PSEUDO_OVL_TYPE_RDMA, 7 }, + [PSEUDO_OVL_MERGE0] = { PSEUDO_OVL_TYPE_MERGE, 1 }, + [PSEUDO_OVL_MERGE1] = { PSEUDO_OVL_TYPE_MERGE, 2 }, + [PSEUDO_OVL_MERGE2] = { PSEUDO_OVL_TYPE_MERGE, 3 }, + [PSEUDO_OVL_MERGE3] = { PSEUDO_OVL_TYPE_MERGE, 4 }, +}; + +static int mtk_pseudo_ovl_fifo_setting(struct mtk_disp_pseudo_ovl *pseudo_ovl, + struct cmdq_pkt *handle) +{ + struct mtk_pseudo_ovl_comp *rdma = NULL; + const struct pseudo_ovl_data *data = pseudo_ovl->data; + const struct mtk_mdp_rdma_fifo *fifo = &data->fifo; + int i; + + for (i = PSEUDO_OVL_MDP_RDMA0; i <= PSEUDO_OVL_MDP_RDMA7; i++) { + rdma = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + i]; + mtk_mdp_rdma_fifo_config(rdma->regs, handle, &rdma->cmdq_base, fifo); + } + + return 0; +} + +static void mtk_pseudo_ovl_merge_config(struct mtk_pseudo_ovl_comp *comp, + struct pseudo_ovl_merge_config *merge_cfg, + struct cmdq_pkt *cmdq_pkt) +{ + switch (merge_cfg->merge_mode) { + case 6: + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_0); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_4); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_24); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_25); + break; + case 18: + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_0); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_1); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_4); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_24); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_25); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_26); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_27); + break; + default: + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_0); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->in_w[1]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_1); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[0]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_4); + mtk_ddp_write(cmdq_pkt, (merge_cfg->in_h << 16 | merge_cfg->out_w[1]), + &comp->cmdq_base, comp->regs, DISP_MERGE_CFG_5); + break; + } + + mtk_ddp_write(cmdq_pkt, merge_cfg->merge_mode, &comp->cmdq_base, + comp->regs, DISP_MERGE_CFG_12); + mtk_ddp_write(cmdq_pkt, CFG_10_NO_SWAP, &comp->cmdq_base, + comp->regs, DISP_MERGE_CFG_10); + mtk_ddp_write_mask(cmdq_pkt, 1, &comp->cmdq_base, comp->regs, + DISP_MERGE_ENABLE, MERGE_ENABLE); +} + +void mtk_pseudo_ovl_layer_config(struct device *dev, unsigned int idx, + struct mtk_plane_state *state, + struct cmdq_pkt *cmdq_pkt) +{ + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); + struct mtk_plane_pending_state *pending = &state->pending; + struct pseudo_ovl_merge_config merge_cfg = {0}; + struct mtk_mdp_rdma_cfg rdma_config = {0}; + struct mtk_pseudo_ovl_comp *rdma_l; + struct mtk_pseudo_ovl_comp *rdma_r; + struct mtk_pseudo_ovl_comp *merge; + const struct drm_format_info *fmt_info = drm_format_info(pending->format); + bool use_dual_pipe = false; + + dev_dbg(dev, "%s+ idx:%d, enable:%d, fmt:0x%x\n", __func__, idx, + pending->enable, pending->format); + dev_dbg(dev, "addr 0x%lx, fb w:%d, {%d,%d,%d,%d}\n", + pending->addr, (pending->pitch / fmt_info->cpp[0]), + pending->x, pending->y, pending->width, pending->height); + + rdma_l = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * idx]; + rdma_r = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * idx + 1]; + merge = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_MERGE_BASE + idx]; + + if (!pending->enable) { + mtk_ddp_write_mask(cmdq_pkt, 0x0, &merge->cmdq_base, merge->regs, + DISP_MERGE_ENABLE, MERGE_ENABLE); + mtk_mdp_rdma_stop(rdma_l->regs, cmdq_pkt, &rdma_l->cmdq_base); + mtk_mdp_rdma_stop(rdma_r->regs, cmdq_pkt, &rdma_r->cmdq_base); + return; + } + + if (pending->width > MTK_PSEUDO_OVL_SINGLE_PIPE_MAX_WIDTH) + use_dual_pipe = true; + + merge_cfg.out_w[0] = pending->width; + merge_cfg.in_h = pending->height; + merge_cfg.fmt = pending->format; + if (use_dual_pipe) { + merge_cfg.merge_mode = CFG12_11_10_1PI_2PO_MERGE; + merge_cfg.in_w[0] = (pending->width / 2) + ((pending->width / 2) % 2); + merge_cfg.in_w[1] = (pending->width / 2) - ((pending->width / 2) % 2); + } else { + merge_cfg.merge_mode = CFG12_10_10_1PI_2PO_BUF_MODE; + merge_cfg.in_w[0] = pending->width; + } + + mtk_pseudo_ovl_merge_config(merge, &merge_cfg, cmdq_pkt); + + mtk_mmsys_ddp_config(pseudo_ovl->mmsys_dev, MMSYS_CONFIG_MERGE_ASYNC_WIDTH, + idx, pending->width / 2, cmdq_pkt); + mtk_mmsys_ddp_config(pseudo_ovl->mmsys_dev, MMSYS_CONFIG_MERGE_ASYNC_HEIGHT, + idx, pending->height, cmdq_pkt); + + rdma_config.source_width = pending->pitch / fmt_info->cpp[0]; + rdma_config.csc_enable = fmt_info->is_yuv ? true : false; + rdma_config.profile = RDMA_CSC_FULL709_TO_RGB; + rdma_config.encode_type = RDMA_ENCODE_NONE; + rdma_config.block_size = RDMA_BLOCK_NONE; + rdma_config.width = merge_cfg.in_w[0]; + rdma_config.height = pending->height; + rdma_config.addr0 = pending->addr; + rdma_config.fmt = pending->format; + mtk_mdp_rdma_config(rdma_l->regs, &rdma_config, cmdq_pkt, &rdma_l->cmdq_base); + + rdma_config.x_left = merge_cfg.in_w[0]; + rdma_config.width = merge_cfg.in_w[1]; + mtk_mdp_rdma_config(rdma_r->regs, &rdma_config, cmdq_pkt, &rdma_r->cmdq_base); + + mtk_ddp_write_mask(cmdq_pkt, 0x1, &merge->cmdq_base, merge->regs, + DISP_MERGE_ENABLE, MERGE_ENABLE); + mtk_ddp_write_mask(cmdq_pkt, 0x0, &merge->cmdq_base, merge->regs, + DISP_MERGE_MUTE_0, 0x1); + + mtk_mdp_rdma_start(rdma_l->regs, cmdq_pkt, &rdma_l->cmdq_base); + if (use_dual_pipe) + mtk_mdp_rdma_start(rdma_r->regs, cmdq_pkt, &rdma_r->cmdq_base); + else + mtk_mdp_rdma_stop(rdma_r->regs, cmdq_pkt, &rdma_r->cmdq_base); +} + +void mtk_pseudo_ovl_config(struct device *dev, unsigned int w, unsigned int h, + unsigned int vrefresh, unsigned int bpc, + struct cmdq_pkt *cmdq_pkt) +{ + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); + + dev_info(dev, "%s w:%d, h:%d\n", __func__, w, h); + + mtk_pseudo_ovl_fifo_setting(pseudo_ovl, cmdq_pkt); +} + +void mtk_pseudo_ovl_start(struct device *dev) +{ +} + +void mtk_pseudo_ovl_stop(struct device *dev) +{ + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); + struct mtk_pseudo_ovl_comp *rdma_l; + struct mtk_pseudo_ovl_comp *rdma_r; + struct mtk_pseudo_ovl_comp *merge; + unsigned int reg; + u32 i; + + for (i = 0; i < pseudo_ovl->data->layer_nr; i++) { + rdma_l = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * i]; + rdma_r = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_RDMA_BASE + 2 * i + 1]; + merge = &pseudo_ovl->pseudo_ovl_comp[PSEUDO_OVL_MERGE_BASE + i]; + + mtk_mdp_rdma_stop(rdma_l->regs, NULL, &rdma_l->cmdq_base); + mtk_mdp_rdma_stop(rdma_r->regs, NULL, &rdma_r->cmdq_base); + + reg = readl(merge->regs + DISP_MERGE_ENABLE); + reg = reg & ~MERGE_ENABLE; + writel_relaxed(reg, merge->regs + DISP_MERGE_ENABLE); + + device_reset_optional(merge->dev); + } +} + +int mtk_pseudo_ovl_clk_enable(struct device *dev) +{ + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); + struct mtk_pseudo_ovl_comp *comp; + int ret; + int i; + int j; + + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_ID_MAX; i++) { + comp = &pseudo_ovl->pseudo_ovl_comp[i]; + if (!comp->dev) + continue; + + /* Need to power on for private rdma devices */ + if (i < PSEUDO_OVL_MERGE_BASE) { + ret = pm_runtime_get_sync(comp->dev); + if (ret < 0) + dev_err(dev, + "Failed to power on, err %d-%s\n", + ret, pseudo_ovl_comp_str[i]); + } + + for (j = 0; j < ARRAY_SIZE(comp->clks); j++) { + if (IS_ERR(comp->clks[j])) + break; + + ret = clk_prepare_enable(comp->clks[j]); + if (ret) + dev_err(dev, + "Failed to enable clock %d, err %d-%s\n", + i, ret, pseudo_ovl_comp_str[i]); + } + } + + return ret; +} + +void mtk_pseudo_ovl_clk_disable(struct device *dev) +{ + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(dev); + struct mtk_pseudo_ovl_comp *comp; + int ret; + int i; + int j; + + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_ID_MAX; i++) { + comp = &pseudo_ovl->pseudo_ovl_comp[i]; + if (!comp->dev) + continue; + + for (j = 0; i < ARRAY_SIZE(comp->clks); j++) { + if (IS_ERR(comp->clks[j])) + break; + clk_disable_unprepare(comp->clks[j]); + } + + /* Need to power off for private rdma devices */ + if (i < PSEUDO_OVL_MERGE_BASE) { + ret = pm_runtime_put(comp->dev); + if (ret < 0) + dev_err(dev, + "Failed to power off, err-%s\n", + ret, pseudo_ovl_comp_str[i]); + } + } +} + +static int pseudo_ovl_comp_get_id(struct device *dev, struct device_node *node, + enum mtk_pseudo_ovl_comp_type type) +{ + int alias_id = of_alias_get_id(node, private_comp_stem[type]); + int ret; + int i; + + for (i = 0; i < ARRAY_SIZE(comp_matches); i++) + if (comp_matches[i].type == type && + comp_matches[i].alias_id == alias_id) + return i; + + dev_err(dev, "Failed to get id. type: %d, alias: %d\n", type, alias_id); + return -EINVAL; +} + +static int private_comp_init(struct device *dev, struct device_node *node, + struct mtk_pseudo_ovl_comp *comp, + enum mtk_pseudo_ovl_comp_id id) +{ + struct platform_device *comp_pdev; + int ret; + int i; + + if (id < 0 || id >= PSEUDO_OVL_ID_MAX) { + dev_err(dev, "Invalid component id %d\n", id); + return -EINVAL; + } + + comp_pdev = of_find_device_by_node(node); + if (!comp_pdev) { + dev_warn(dev, "can't find platform device of node:%s\n", + node->name); + return -ENODEV; + } + comp->dev = &comp_pdev->dev; + comp->regs = of_iomap(node, 0); + + for (i = 0; i < ARRAY_SIZE(comp->clks); i++) { + comp->clks[i] = of_clk_get(node, i); + if (IS_ERR(comp->clks[i])) + break; + } + +#if IS_REACHABLE(CONFIG_MTK_CMDQ) + ret = cmdq_dev_get_client_reg(comp->dev, &comp->cmdq_base, 0); + if (ret) + dev_info(dev, "get mediatek,gce-client-reg fail!\n"); +#endif + + if (id < PSEUDO_OVL_MERGE_BASE) + pm_runtime_enable(comp->dev); + + dev_info(dev, "[DRM]regs:0x%p, node:%s\n", comp->regs, pseudo_ovl_comp_str[id]); + + return 0; +} + +static int mtk_disp_pseudo_ovl_comp_probe(struct platform_device *pdev) +{ + return 0; +} + +static int mtk_disp_pseudo_ovl_comp_remove(struct platform_device *pdev) +{ + return 0; +} + +static const struct of_device_id mtk_pseudo_ovl_comp_dt_ids[] = { + { + .compatible = "mediatek,mt8195-vdo1-rdma", + .data = (void *)PSEUDO_OVL_TYPE_RDMA, + }, { + .compatible = "mediatek,mt8195-vdo1-merge", + .data = (void *)PSEUDO_OVL_TYPE_MERGE, + }, + {}, +}; + +static struct platform_driver mtk_disp_pseudo_ovl_comp_driver = { + .probe = mtk_disp_pseudo_ovl_comp_probe, + .remove = mtk_disp_pseudo_ovl_comp_remove, + .driver = { + .name = "mediatek-disp-pseudo-ovl-comp", + .owner = THIS_MODULE, + .of_match_table = mtk_pseudo_ovl_comp_dt_ids, + }, +}; +module_platform_driver(mtk_disp_pseudo_ovl_comp_driver); + +static int pseudo_ovl_comp_init(struct device *dev) +{ + struct mtk_disp_pseudo_ovl *priv = dev_get_drvdata(dev); + struct device_node *node, *parent; + int i, ret; + + parent = dev->parent->of_node->parent; + + for_each_child_of_node(parent, node) { + const struct of_device_id *of_id; + enum mtk_pseudo_ovl_comp_type type; + struct mtk_pseudo_ovl_comp *comp; + int id; + + of_id = of_match_node(mtk_pseudo_ovl_comp_dt_ids, node); + if (!of_id) + continue; + + if (!of_device_is_available(node)) { + dev_info(dev, "Skipping disabled component %pOF\n", + node); + continue; + } + + type = (enum mtk_pseudo_ovl_comp_type)of_id->data; + id = pseudo_ovl_comp_get_id(dev, node, type); + if (id < 0) { + dev_warn(dev, "Skipping unknown component %pOF\n", + node); + continue; + } + + ret = private_comp_init(dev, node, &priv->pseudo_ovl_comp[id], id); + if (ret) + return ret; + } + + return 0; +} + +static const struct pseudo_ovl_data mt8195_pseudo_ovl_driver_data = { + .layer_nr = 4, + .fifo.read_request_type = 7, + .fifo.command_div = 1, + .fifo.ext_preutra_en = 1, + .fifo.ultra_en = 0, + .fifo.pre_ultra_en = 1, + .fifo.ext_ultra_en = 1, + .fifo.extrd_arb_max_0 = 3, + .fifo.buf_resv_size_0 = 0, + .fifo.issue_req_th_0 = 0, + .fifo.ultra_h_con_0 = 156, + .fifo.ultra_l_con_0 = 104, +}; + +static const struct of_device_id pseudo_ovl_driver_dt_match[] = { + { .compatible = "mediatek,mt8195-disp-ethdr", + .data = &mt8195_pseudo_ovl_driver_data}, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_disp_pseudo_ovl_driver_dt_match); + +static int mtk_disp_pseudo_ovl_probe(struct platform_device *pdev) +{ + struct mtk_disp_pseudo_ovl *priv; + struct device *dev = &pdev->dev; + struct device_node *phandle = dev->parent->of_node; + const struct of_device_id *of_id; + int ret; + int i; + + dev_info(dev, "%s+\n", __func__); + + of_id = of_match_node(pseudo_ovl_driver_dt_match, phandle); + if (!of_id) + return -ENODEV; + + priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL); + if (!priv) + return -ENOMEM; + + priv->mmsys_dev = pdev->dev.platform_data; + + priv->data = of_id->data; + platform_set_drvdata(pdev, priv); + + ret = pseudo_ovl_comp_init(dev); + if (ret) { + dev_notice(dev, "pseudo_ovl comp init fail\n"); + return ret; + } + + dev_info(dev, "%s-\n", __func__); + return ret; +} + +static int mtk_disp_pseudo_ovl_remove(struct platform_device *pdev) +{ + struct mtk_disp_pseudo_ovl *pseudo_ovl = dev_get_drvdata(&pdev->dev); + int i; + + for (i = PSEUDO_OVL_MDP_RDMA0; i < PSEUDO_OVL_MERGE_BASE; i++) + pm_runtime_disable(pseudo_ovl->pseudo_ovl_comp[i].dev); + + return 0; +} + +struct platform_driver mtk_disp_pseudo_ovl_driver = { + .probe = mtk_disp_pseudo_ovl_probe, + .remove = mtk_disp_pseudo_ovl_remove, + .driver = { + .name = "mediatek-disp-pseudo-ovl", + .owner = THIS_MODULE, + }, +}; +module_platform_driver(mtk_disp_pseudo_ovl_driver); diff --git a/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h new file mode 100644 index 000000000000..b3fe1e1702b8 --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2021 MediaTek Inc. + */ + +#ifndef __MTK_DISP_PSEUDO_OVL_H__ +#define __MTK_DISP_PSEUDO_OVL_H__ + +#include <drm/mediatek_drm.h> + +void mtk_pseudo_ovl_start(struct device *dev); +void mtk_pseudo_ovl_stop(struct device *dev); +int mtk_pseudo_ovl_clk_enable(struct device *dev); +void mtk_pseudo_ovl_clk_disable(struct device *dev); +void mtk_pseudo_ovl_config(struct device *dev, unsigned int w, unsigned int h, + unsigned int vrefresh, unsigned int bpc, + struct cmdq_pkt *cmdq_pkt); +void mtk_pseudo_ovl_layer_config(struct device *dev, unsigned int idx, + struct mtk_plane_state *state, + struct cmdq_pkt *cmdq_pkt); + +#endif + diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c new file mode 100644 index 000000000000..81d3cc4872eb --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.c @@ -0,0 +1,456 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2021 MediaTek Inc. + */ + +#include <drm/drm_fourcc.h> +#include "mtk_drm_drv.h" +#include "mtk_mdp_reg_rdma.h" +#include "mtk_mdp_rdma.h" + +#define RDMA_INPUT_SWAP BIT(14) +#define RDMA_INPUT_10BIT BIT(18) +#define IRQ_INT_EN_ALL \ + (REG_FLD_MASK(FLD_UNDERRUN_INT_EN) |\ + REG_FLD_MASK(FLD_REG_UPDATE_INT_EN) |\ + REG_FLD_MASK(FLD_FRAME_COMPLETE_INT_EN)) + +static unsigned int rdma_get_y_pitch(unsigned int fmt, unsigned int width) +{ + switch (fmt) { + default: + case DRM_FORMAT_RGB565: + case DRM_FORMAT_BGR565: + return 2 * width; + case DRM_FORMAT_RGB888: + case DRM_FORMAT_BGR888: + return 3 * width; + case DRM_FORMAT_RGBX8888: + case DRM_FORMAT_RGBA8888: + case DRM_FORMAT_BGRX8888: + case DRM_FORMAT_BGRA8888: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_ABGR8888: + case DRM_FORMAT_ARGB2101010: + case DRM_FORMAT_ABGR2101010: + case DRM_FORMAT_RGBA1010102: + case DRM_FORMAT_BGRA1010102: + return 4 * width; + case DRM_FORMAT_UYVY: + case DRM_FORMAT_YUYV: + return 2 * width; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + return 1 * width; + } +} + +static unsigned int rdma_get_uv_pitch(unsigned int fmt, unsigned int width) +{ + switch (fmt) { + default: + return 0; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + return 4 * width; + } +} + +static unsigned int rdma_get_block_h(unsigned int mode) +{ + switch (mode) { + default: + return 0; + case RDMA_BLOCK_8x8: + case RDMA_BLOCK_16x8: + case RDMA_BLOCK_32x8: + return 8; + case RDMA_BLOCK_8x16: + case RDMA_BLOCK_16x16: + case RDMA_BLOCK_32x16: + return 16; + case RDMA_BLOCK_8x32: + case RDMA_BLOCK_16x32: + case RDMA_BLOCK_32x32: + return 32; + } +} + +static unsigned int rdma_get_horizontal_shift_uv(unsigned int fmt) +{ + switch (fmt) { + default: + return 0; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + return 1; + } +} + +static unsigned int rdma_get_vertical_shift_uv(unsigned int fmt) +{ + switch (fmt) { + default: + return 0; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + return 1; + } +} + +static unsigned int rdma_get_bits_per_pixel_y(unsigned int fmt) +{ + switch (fmt) { + default: + case DRM_FORMAT_RGB565: + case DRM_FORMAT_BGR565: + return 16; + case DRM_FORMAT_RGB888: + case DRM_FORMAT_BGR888: + return 24; + case DRM_FORMAT_RGBX8888: + case DRM_FORMAT_RGBA8888: + case DRM_FORMAT_BGRX8888: + case DRM_FORMAT_BGRA8888: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_ABGR8888: + return 32; + case DRM_FORMAT_UYVY: + case DRM_FORMAT_YUYV: + return 16; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + return 8; + } +} + +static unsigned int rdma_get_bits_per_pixel_uv(unsigned int fmt) +{ + switch (fmt) { + default: + return 0; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + return 16; + } +} + +static bool with_alpha(uint32_t format) +{ + const struct drm_format_info *fmt_info = drm_format_info(format); + + return fmt_info->has_alpha; +} + +static unsigned int rdma_fmt_convert(unsigned int fmt) +{ + switch (fmt) { + default: + case DRM_FORMAT_RGB565: + return RDMA_INPUT_FORMAT_RGB565; + case DRM_FORMAT_BGR565: + return RDMA_INPUT_FORMAT_RGB565 | RDMA_INPUT_SWAP; + case DRM_FORMAT_RGB888: + return RDMA_INPUT_FORMAT_RGB888; + case DRM_FORMAT_BGR888: + return RDMA_INPUT_FORMAT_RGB888 | RDMA_INPUT_SWAP; + case DRM_FORMAT_RGBX8888: + case DRM_FORMAT_RGBA8888: + return RDMA_INPUT_FORMAT_ARGB8888; + case DRM_FORMAT_BGRX8888: + case DRM_FORMAT_BGRA8888: + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_SWAP; + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_ARGB8888: + return RDMA_INPUT_FORMAT_RGBA8888; + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_ABGR8888: + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_SWAP; + case DRM_FORMAT_ABGR2101010: + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_SWAP | + RDMA_INPUT_10BIT; + case DRM_FORMAT_ARGB2101010: + return RDMA_INPUT_FORMAT_RGBA8888 | RDMA_INPUT_10BIT; + case DRM_FORMAT_RGBA1010102: + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_SWAP | + RDMA_INPUT_10BIT; + case DRM_FORMAT_BGRA1010102: + return RDMA_INPUT_FORMAT_ARGB8888 | RDMA_INPUT_10BIT; + case DRM_FORMAT_UYVY: + return RDMA_INPUT_FORMAT_UYVY; + case DRM_FORMAT_YUYV: + return RDMA_INPUT_FORMAT_YUY2; + } +} + +void mtk_mdp_rdma_start(void __iomem *base, struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base) +{ + unsigned int inten = IRQ_INT_EN_ALL; + + mtk_ddp_write_mask(cmdq_pkt, inten, cmdq_base, base, + MDP_RDMA_INTERRUPT_ENABLE, IRQ_INT_EN_ALL); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_ROT_ENABLE, 1), cmdq_base, + base, MDP_RDMA_EN, REG_FLD_MASK(FLD_ROT_ENABLE)); +} + +void mtk_mdp_rdma_stop(void __iomem *base, struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base) +{ + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_ROT_ENABLE, 0), cmdq_base, + base, MDP_RDMA_EN, REG_FLD_MASK(FLD_ROT_ENABLE)); + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, + MDP_RDMA_INTERRUPT_ENABLE, IRQ_INT_EN_ALL); + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, + MDP_RDMA_INTERRUPT_STATUS, IRQ_INT_EN_ALL); + mtk_ddp_write_mask(cmdq_pkt, 1, cmdq_base, base, MDP_RDMA_RESET, ~0); + mtk_ddp_write_mask(cmdq_pkt, 0, cmdq_base, base, MDP_RDMA_RESET, ~0); +} + +void mtk_mdp_rdma_fifo_config(void __iomem *base, struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base, + const struct mtk_mdp_rdma_fifo *fifo) +{ + int reg; + int reg_val; + int reg_mask; + + reg = MDP_RDMA_GMCIF_CON; + reg_val = REG_FLD_VAL(FLD_RD_REQ_TYPE, fifo->read_request_type) | + REG_FLD_VAL(FLD_COMMAND_DIV, fifo->command_div) | + REG_FLD_VAL(FLD_EXT_PREULTRA_EN, fifo->ext_preutra_en) | + REG_FLD_VAL(FLD_ULTRA_EN, fifo->ultra_en) | + REG_FLD_VAL(PRE_ULTRA_EN, fifo->pre_ultra_en) | + REG_FLD_VAL(FLD_EXT_ULTRA_EN, fifo->ext_ultra_en); + reg_mask = REG_FLD_MASK(FLD_RD_REQ_TYPE) | + REG_FLD_MASK(FLD_COMMAND_DIV) | + REG_FLD_MASK(FLD_EXT_PREULTRA_EN) | + REG_FLD_MASK(FLD_ULTRA_EN) | + REG_FLD_MASK(PRE_ULTRA_EN) | + REG_FLD_MASK(FLD_EXT_ULTRA_EN); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_DMA_CON_0; + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX, fifo->extrd_arb_max_0) | + REG_FLD_VAL(FLD_BUF_RESV_SIZE, fifo->buf_resv_size_0) | + REG_FLD_VAL(FLD_ISSUE_REQ_TH, fifo->issue_req_th_0); + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX) | + REG_FLD_MASK(FLD_BUF_RESV_SIZE) | + REG_FLD_MASK(FLD_ISSUE_REQ_TH); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_UTRA_H_CON_0; + reg_val = REG_FLD_VAL(FLD_PREUTRA_H_OFS_0, fifo->ultra_h_con_0); + reg_mask = REG_FLD_MASK(FLD_PREUTRA_H_OFS_0); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_UTRA_L_CON_0; + reg_val = REG_FLD_VAL(FLD_PREUTRA_L_OFS_0, fifo->ultra_l_con_0); + reg_mask = REG_FLD_MASK(FLD_PREUTRA_L_OFS_0); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_DMABUF_CON_1; + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_1, 0) | + REG_FLD_VAL(FLD_BUF_RESV_SIZE_1, 0) | + REG_FLD_VAL(FLD_ISSUE_REQ_TH_1, 0); + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_1) | + REG_FLD_MASK(FLD_BUF_RESV_SIZE_1) | + REG_FLD_MASK(FLD_ISSUE_REQ_TH_1); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_ULTRA_TH_HIGH_CON_1; + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_1, 0); + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_1); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_ULTRA_TH_LOW_CON_1; + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_1, 0); + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_1); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_DMABUF_CON_2; + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_2, 0) | + REG_FLD_VAL(FLD_BUF_RESV_SIZE_2, 0) | + REG_FLD_VAL(FLD_ISSUE_REQ_TH_2, 0); + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_2) | + REG_FLD_MASK(FLD_BUF_RESV_SIZE_2) | + REG_FLD_MASK(FLD_ISSUE_REQ_TH_2); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_UTRA_H_CON_2; + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_2, 0); + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_2); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_ULTRA_TH_LOW_CON_2; + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_2, 0); + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_2); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_DMABUF_CON_3; + reg_val = REG_FLD_VAL(FLD_EXTRD_ARB_MAX_3, 0) | + REG_FLD_VAL(FLD_BUF_RESV_SIZE_3, 0) | + REG_FLD_VAL(FLD_ISSUE_REQ_TH_3, 0); + reg_mask = REG_FLD_MASK(FLD_EXTRD_ARB_MAX_3) | + REG_FLD_MASK(FLD_BUF_RESV_SIZE_3) | + REG_FLD_MASK(FLD_ISSUE_REQ_TH_3); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_UTRA_H_CON_3; + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_HIGH_OFS_3, 0); + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_HIGH_OFS_3); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); + + reg = MDP_RDMA_ULTRA_TH_LOW_CON_3; + reg_val = REG_FLD_VAL(FLD_PRE_ULTRA_TH_LOW_OFS_3, 0); + reg_mask = REG_FLD_MASK(FLD_PRE_ULTRA_TH_LOW_OFS_3); + mtk_ddp_write_mask(cmdq_pkt, reg_val, cmdq_base, base, reg, reg_mask); +} + +void mtk_mdp_rdma_config(void __iomem *base, struct mtk_mdp_rdma_cfg *cfg, + struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base) +{ + unsigned int src_pitch_uv = rdma_get_uv_pitch(cfg->fmt, cfg->source_width); + unsigned int src_pitch_y = rdma_get_y_pitch(cfg->fmt, cfg->source_width); + unsigned int h_shift_uv = rdma_get_horizontal_shift_uv(cfg->fmt); + unsigned int v_shift_uv = rdma_get_vertical_shift_uv(cfg->fmt); + unsigned int bpp_uv = rdma_get_bits_per_pixel_uv(cfg->fmt); + unsigned int block_h = rdma_get_block_h(cfg->block_size); + unsigned int bpp_y = rdma_get_bits_per_pixel_y(cfg->fmt); + unsigned int y_start_line = 0; + unsigned int offset_y = 0; + unsigned int offset_u = 0; + unsigned int offset_v = 0; + + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_UNIFORM_CONFIG, 1), + cmdq_base, base, MDP_RDMA_SRC_CON, + REG_FLD_MASK(FLD_UNIFORM_CONFIG)); + mtk_ddp_write_mask(cmdq_pkt, rdma_fmt_convert(cfg->fmt), cmdq_base, + base, MDP_RDMA_SRC_CON, REG_FLD_MASK(FLD_SWAP) | + REG_FLD_MASK(FLD_SRC_FORMAT) | + REG_FLD_MASK(FLD_BIT_NUMBER)); + + if (!cfg->csc_enable && with_alpha(cfg->fmt)) + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_ARGB, 1), + cmdq_base, base, MDP_RDMA_SRC_CON, + REG_FLD_MASK(FLD_OUTPUT_ARGB)); + else + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_ARGB, 0), + cmdq_base, base, MDP_RDMA_SRC_CON, + REG_FLD_MASK(FLD_OUTPUT_ARGB)); + + mtk_ddp_write_mask(cmdq_pkt, cfg->addr0, cmdq_base, base, + MDP_RDMA_SRC_BASE_0, REG_FLD_MASK(FLD_SRC_BASE_0)); + mtk_ddp_write_mask(cmdq_pkt, cfg->addr1, cmdq_base, base, + MDP_RDMA_SRC_BASE_1, REG_FLD_MASK(FLD_SRC_BASE_1)); + mtk_ddp_write_mask(cmdq_pkt, cfg->addr2, cmdq_base, base, + MDP_RDMA_SRC_BASE_2, REG_FLD_MASK(FLD_SRC_BASE_2)); + + mtk_ddp_write_mask(cmdq_pkt, src_pitch_y, cmdq_base, base, + MDP_RDMA_MF_BKGD_SIZE_IN_BYTE, + REG_FLD_MASK(FLD_MF_BKGD_WB)); + mtk_ddp_write_mask(cmdq_pkt, src_pitch_uv, cmdq_base, base, + MDP_RDMA_SF_BKGD_SIZE_IN_BYTE, + REG_FLD_MASK(FLD_SF_BKGD_WB)); + + if (cfg->encode_type == RDMA_ENCODE_AFBC) { + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_BKGD_WP, cfg->source_width), + cmdq_base, base, MDP_RDMA_MF_BKGD_SIZE_IN_PIXEL, + REG_FLD_MASK(FLD_MF_BKGD_WP)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_BKGD_HP, cfg->height), + cmdq_base, base, MDP_RDMA_MF_BKGD_H_SIZE_IN_PIXEL, + REG_FLD_MASK(FLD_BKGD_HP)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_YUV_TRANSFORM, 1), + cmdq_base, base, MDP_RDMA_COMP_CON, + REG_FLD_MASK(FLD_AFBC_YUV_TRANSFORM)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_UFBDC_EN, 1), cmdq_base, + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_UFBDC_EN)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_EN, 1), cmdq_base, + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_AFBC_EN)); + } else { + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_YUV_TRANSFORM, 0), + cmdq_base, base, MDP_RDMA_COMP_CON, + REG_FLD_MASK(FLD_AFBC_YUV_TRANSFORM)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_UFBDC_EN, 0), cmdq_base, + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_UFBDC_EN)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_AFBC_EN, 0), cmdq_base, + base, MDP_RDMA_COMP_CON, REG_FLD_MASK(FLD_AFBC_EN)); + } + + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_OUTPUT_10B, 1), cmdq_base, + base, MDP_RDMA_CON, REG_FLD_MASK(FLD_OUTPUT_10B)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SIMPLE_MODE, 1), cmdq_base, + base, MDP_RDMA_CON, REG_FLD_MASK(FLD_SIMPLE_MODE)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_TRANS_EN, cfg->csc_enable), + cmdq_base, base, MDP_RDMA_TRANSFORM_0, + REG_FLD_MASK(FLD_TRANS_EN)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_INT_MATRIX_SEL, cfg->profile), + cmdq_base, base, MDP_RDMA_TRANSFORM_0, + REG_FLD_MASK(FLD_INT_MATRIX_SEL)); + + if (cfg->block_size == RDMA_BLOCK_NONE) { + y_start_line = cfg->y_top; + + offset_y = (cfg->x_left * bpp_y >> 3) + y_start_line * src_pitch_y; + offset_u = ((cfg->x_left >> h_shift_uv) * bpp_uv >> 3) + + (y_start_line >> v_shift_uv) * src_pitch_uv; + offset_v = ((cfg->x_left >> h_shift_uv) * bpp_uv >> 3) + + (y_start_line >> v_shift_uv) * src_pitch_uv; + } else { + offset_y = (cfg->x_left * block_h * bpp_y >> 3) + + (cfg->y_top) * src_pitch_y; + offset_u = ((cfg->x_left >> h_shift_uv) * (block_h >> v_shift_uv) * + bpp_uv >> 3) + (cfg->y_top) * src_pitch_uv; + offset_v = ((cfg->x_left >> h_shift_uv) * (block_h >> v_shift_uv) * + bpp_uv >> 3) + (cfg->y_top) * src_pitch_uv; + } + + if (cfg->encode_type == RDMA_ENCODE_AFBC) { + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_WP, cfg->x_left), + cmdq_base, base, MDP_RDMA_SRC_OFFSET_WP, + REG_FLD_MASK(FLD_SRC_OFFSET_WP)); + + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_HP, cfg->y_top), + cmdq_base, base, MDP_RDMA_SRC_OFFSET_HP, + REG_FLD_MASK(FLD_SRC_OFFSET_HP)); + } + + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_0, offset_y), + cmdq_base, base, MDP_RDMA_SRC_OFFSET_0, + REG_FLD_MASK(FLD_SRC_OFFSET_0)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_1, offset_u), + cmdq_base, base, MDP_RDMA_SRC_OFFSET_1, + REG_FLD_MASK(FLD_SRC_OFFSET_1)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_SRC_OFFSET_2, offset_v), + cmdq_base, base, MDP_RDMA_SRC_OFFSET_2, + REG_FLD_MASK(FLD_SRC_OFFSET_2)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_SRC_W, cfg->width), + cmdq_base, base, MDP_RDMA_MF_SRC_SIZE, + REG_FLD_MASK(FLD_MF_SRC_W)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_SRC_H, cfg->height), + cmdq_base, base, MDP_RDMA_MF_SRC_SIZE, + REG_FLD_MASK(FLD_MF_SRC_H)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_CLIP_W, cfg->width), + cmdq_base, base, MDP_RDMA_MF_CLIP_SIZE, + REG_FLD_MASK(FLD_MF_CLIP_W)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_CLIP_H, cfg->height), + cmdq_base, base, MDP_RDMA_MF_CLIP_SIZE, + REG_FLD_MASK(FLD_MF_CLIP_H)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_OFFSET_W_1, 0), + cmdq_base, base, MDP_RDMA_MF_OFFSET_1, + REG_FLD_MASK(FLD_MF_OFFSET_W_1)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_MF_OFFSET_H_1, 0), + cmdq_base, base, MDP_RDMA_MF_OFFSET_1, + REG_FLD_MASK(FLD_MF_OFFSET_H_1)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_LINE_THRESHOLD, cfg->height), + cmdq_base, base, MDP_RDMA_TARGET_LINE, + REG_FLD_MASK(FLD_LINE_THRESHOLD)); + mtk_ddp_write_mask(cmdq_pkt, REG_FLD_VAL(FLD_TARGET_LINE_EN, 1), + cmdq_base, base, MDP_RDMA_TARGET_LINE, + REG_FLD_MASK(FLD_TARGET_LINE_EN)); +} diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h new file mode 100644 index 000000000000..c16bfb716610 --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_mdp_rdma.h @@ -0,0 +1,109 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2021 MediaTek Inc. + */ + +#ifndef __MTK_MDP_RDMA_H__ +#define __MTK_MDP_RDMA_H__ + +enum rdma_format { + RDMA_INPUT_FORMAT_RGB565 = 0, + RDMA_INPUT_FORMAT_RGB888 = 1, + RDMA_INPUT_FORMAT_RGBA8888 = 2, + RDMA_INPUT_FORMAT_ARGB8888 = 3, + RDMA_INPUT_FORMAT_UYVY = 4, + RDMA_INPUT_FORMAT_YUY2 = 5, + RDMA_INPUT_FORMAT_Y8 = 7, + RDMA_INPUT_FORMAT_YV12 = 8, + RDMA_INPUT_FORMAT_UYVY_3PL = 9, + RDMA_INPUT_FORMAT_NV12 = 12, + RDMA_INPUT_FORMAT_UYVY_2PL = 13, + RDMA_INPUT_FORMAT_Y410 = 14 +}; + +enum rdma_profile { + RDMA_CSC_RGB_TO_JPEG = 0, + RDMA_CSC_RGB_TO_FULL709 = 1, + RDMA_CSC_RGB_TO_BT601 = 2, + RDMA_CSC_RGB_TO_BT709 = 3, + RDMA_CSC_JPEG_TO_RGB = 4, + RDMA_CSC_FULL709_TO_RGB = 5, + RDMA_CSC_BT601_TO_RGB = 6, + RDMA_CSC_BT709_TO_RGB = 7, + RDMA_CSC_JPEG_TO_BT601 = 8, + RDMA_CSC_JPEG_TO_BT709 = 9, + RDMA_CSC_BT601_TO_JPEG = 10, + RDMA_CSC_BT709_TO_BT601 = 11, + RDMA_CSC_BT601_TO_BT709 = 12 +}; + +enum rdma_encode { + RDMA_ENCODE_NONE = 0, + RDMA_ENCODE_AFBC = 1, + RDMA_ENCODE_HYFBC = 2, + RDMA_ENCODE_UFO_DCP = 3 +}; + +enum rdma_block { + RDMA_BLOCK_NONE = 0, + RDMA_BLOCK_8x8 = 1, + RDMA_BLOCK_8x16 = 2, + RDMA_BLOCK_8x32 = 3, + RDMA_BLOCK_16x8 = 4, + RDMA_BLOCK_16x16 = 5, + RDMA_BLOCK_16x32 = 6, + RDMA_BLOCK_32x8 = 7, + RDMA_BLOCK_32x16 = 8, + RDMA_BLOCK_32x32 = 9 +}; + +struct mtk_mdp_rdma_cfg { + enum rdma_encode encode_type; + enum rdma_block block_size; + enum rdma_profile profile; + unsigned int source_width; + unsigned int addr0; + unsigned int addr1; + unsigned int addr2; + unsigned int width; + unsigned int height; + unsigned int x_left; + unsigned int y_top; + bool csc_enable; + int fmt; +}; + +struct mtk_mdp_rdma_fifo { + int read_request_type; + int command_div; + int ext_preutra_en; + int ultra_en; + int pre_ultra_en; + int ext_ultra_en; + int extrd_arb_max_0; + int buf_resv_size_0; + int issue_req_th_0; + int ultra_h_con_0; + int ultra_l_con_0; +}; + +void mtk_mdp_rdma_start(void __iomem *base, + struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base); + +void mtk_mdp_rdma_stop(void __iomem *base, + struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base); + +void mtk_mdp_rdma_fifo_config(void __iomem *base, + struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base, + const struct mtk_mdp_rdma_fifo *fifo); + +void mtk_mdp_rdma_config(void __iomem *base, + struct mtk_mdp_rdma_cfg *cfg, + struct cmdq_pkt *cmdq_pkt, + struct cmdq_client_reg *cmdq_base); + +#endif // __MTK_MDP_RDMA_H__ + diff --git a/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h b/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h new file mode 100644 index 000000000000..08abd9f39bd8 --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h @@ -0,0 +1,160 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2021 MediaTek Inc. + */ + +#ifndef __MDP_RDMA_REGS_H__ +#define __MDP_RDMA_REGS_H__ + +#define REG_FLD(width, shift) \ + ((unsigned int)((((width) & 0xff) << 16) | ((shift) & 0xff))) + +#define REG_FLD_WIDTH(field) ((unsigned int)(((field) >> 16) & 0xff)) + +#define REG_FLD_SHIFT(field) ((unsigned int)((field) & 0xff)) + +#define REG_FLD_MASK(field) \ + ((unsigned int)((1ULL << REG_FLD_WIDTH(field)) - 1) \ + << REG_FLD_SHIFT(field)) + +#define REG_FLD_VAL(field, val) \ + (((val) << REG_FLD_SHIFT(field)) & REG_FLD_MASK(field)) + +#define MDP_RDMA_EN 0x000 +#define FLD_ROT_ENABLE REG_FLD(1, 0) + +#define MDP_RDMA_RESET 0x008 + +#define MDP_RDMA_INTERRUPT_ENABLE 0x010 +#define FLD_UNDERRUN_INT_EN REG_FLD(1, 2) +#define FLD_REG_UPDATE_INT_EN REG_FLD(1, 1) +#define FLD_FRAME_COMPLETE_INT_EN REG_FLD(1, 0) + +#define MDP_RDMA_INTERRUPT_STATUS 0x018 + +#define MDP_RDMA_CON 0x020 +#define FLD_OUTPUT_10B REG_FLD(1, 5) +#define FLD_SIMPLE_MODE REG_FLD(1, 4) + +#define MDP_RDMA_GMCIF_CON 0x028 +#define FLD_EXT_ULTRA_EN REG_FLD(1, 18) +#define PRE_ULTRA_EN REG_FLD(2, 16) +#define FLD_ULTRA_EN REG_FLD(2, 12) +#define FLD_RD_REQ_TYPE REG_FLD(4, 4) +#define FLD_EXT_PREULTRA_EN REG_FLD(1, 3) +#define FLD_COMMAND_DIV REG_FLD(1, 0) + +#define MDP_RDMA_SRC_CON 0x030 +#define FLD_OUTPUT_ARGB REG_FLD(1, 25) +#define FLD_BIT_NUMBER REG_FLD(2, 18) +#define FLD_UNIFORM_CONFIG REG_FLD(1, 17) +#define FLD_SWAP REG_FLD(1, 14) +#define FLD_SRC_FORMAT REG_FLD(4, 0) + +#define MDP_RDMA_COMP_CON 0x038 +#define FLD_AFBC_EN REG_FLD(1, 22) +#define FLD_AFBC_YUV_TRANSFORM REG_FLD(1, 21) +#define FLD_UFBDC_EN REG_FLD(1, 12) + +#define MDP_RDMA_MF_BKGD_SIZE_IN_BYTE 0x060 +#define FLD_MF_BKGD_WB REG_FLD(23, 0) + +#define MDP_RDMA_MF_BKGD_SIZE_IN_PIXEL 0x068 +#define FLD_MF_BKGD_WP REG_FLD(23, 0) + +#define MDP_RDMA_MF_SRC_SIZE 0x070 +#define FLD_MF_SRC_H REG_FLD(15, 16) +#define FLD_MF_SRC_W REG_FLD(15, 0) + +#define MDP_RDMA_MF_CLIP_SIZE 0x078 +#define FLD_MF_CLIP_H REG_FLD(15, 16) +#define FLD_MF_CLIP_W REG_FLD(15, 0) + +#define MDP_RDMA_MF_OFFSET_1 0x080 +#define FLD_MF_OFFSET_H_1 REG_FLD(6, 16) +#define FLD_MF_OFFSET_W_1 REG_FLD(5, 0) + +#define MDP_RDMA_SF_BKGD_SIZE_IN_BYTE 0x090 +#define FLD_SF_BKGD_WB REG_FLD(23, 0) + +#define MDP_RDMA_MF_BKGD_H_SIZE_IN_PIXEL 0x098 +#define FLD_BKGD_HP REG_FLD(23, 0) + +#define MDP_RDMA_TARGET_LINE 0x0a0 +#define FLD_LINE_THRESHOLD REG_FLD(15, 17) +#define FLD_TARGET_LINE_EN REG_FLD(1, 16) + +#define MDP_RDMA_SRC_OFFSET_0 0x118 +#define FLD_SRC_OFFSET_0 REG_FLD(32, 0) + +#define MDP_RDMA_SRC_OFFSET_1 0x120 +#define FLD_SRC_OFFSET_1 REG_FLD(32, 0) + +#define MDP_RDMA_SRC_OFFSET_2 0x128 +#define FLD_SRC_OFFSET_2 REG_FLD(32, 0) + +#define MDP_RDMA_SRC_OFFSET_WP 0x148 +#define FLD_SRC_OFFSET_WP REG_FLD(32, 0) + +#define MDP_RDMA_SRC_OFFSET_HP 0x150 +#define FLD_SRC_OFFSET_HP REG_FLD(32, 0) + +#define MDP_RDMA_TRANSFORM_0 0x200 +#define FLD_INT_MATRIX_SEL REG_FLD(5, 23) +#define FLD_TRANS_EN REG_FLD(1, 16) + +#define MDP_RDMA_DMA_CON_0 0x240 +#define FLD_EXTRD_ARB_MAX REG_FLD(4, 24) +#define FLD_BUF_RESV_SIZE REG_FLD(8, 16) +#define FLD_ISSUE_REQ_TH REG_FLD(8, 0) + +#define MDP_RDMA_UTRA_H_CON_0 0x248 +#define FLD_PREUTRA_H_OFS_0 REG_FLD(10, 10) + +#define MDP_RDMA_UTRA_L_CON_0 0x250 +#define FLD_PREUTRA_L_OFS_0 REG_FLD(10, 10) + +#define MDP_RDMA_DMABUF_CON_1 0x258 +#define FLD_EXTRD_ARB_MAX_1 REG_FLD(4, 24) +#define FLD_BUF_RESV_SIZE_1 REG_FLD(7, 16) +#define FLD_ISSUE_REQ_TH_1 REG_FLD(7, 0) + +#define MDP_RDMA_ULTRA_TH_HIGH_CON_1 0x260 +#define FLD_PRE_ULTRA_TH_HIGH_OFS_1 REG_FLD(10, 10) + +#define MDP_RDMA_ULTRA_TH_LOW_CON_1 0x268 +#define FLD_PRE_ULTRA_TH_LOW_OFS_1 REG_FLD(10, 10) + +#define MDP_RDMA_DMABUF_CON_2 0x270 +#define FLD_EXTRD_ARB_MAX_2 REG_FLD(4, 24) +#define FLD_BUF_RESV_SIZE_2 REG_FLD(6, 16) +#define FLD_ISSUE_REQ_TH_2 REG_FLD(6, 0) + +#define MDP_RDMA_UTRA_H_CON_2 0x278 +#define FLD_PRE_ULTRA_TH_HIGH_OFS_2 REG_FLD(10, 10) + +#define MDP_RDMA_ULTRA_TH_LOW_CON_2 0x280 +#define FLD_PRE_ULTRA_TH_LOW_OFS_2 REG_FLD(10, 10) + +#define MDP_RDMA_DMABUF_CON_3 0x288 +#define FLD_EXTRD_ARB_MAX_3 REG_FLD(4, 24) +#define FLD_BUF_RESV_SIZE_3 REG_FLD(6, 16) +#define FLD_ISSUE_REQ_TH_3 REG_FLD(6, 0) + +#define MDP_RDMA_UTRA_H_CON_3 0x290 +#define FLD_PRE_ULTRA_TH_HIGH_OFS_3 REG_FLD(10, 10) + +#define MDP_RDMA_ULTRA_TH_LOW_CON_3 0x298 +#define FLD_PRE_ULTRA_TH_LOW_OFS_3 REG_FLD(10, 10) + +#define MDP_RDMA_SRC_BASE_0 0xf00 +#define FLD_SRC_BASE_0 REG_FLD(32, 0) + +#define MDP_RDMA_SRC_BASE_1 0xf08 +#define FLD_SRC_BASE_1 REG_FLD(32, 0) + +#define MDP_RDMA_SRC_BASE_2 0xf10 +#define FLD_SRC_BASE_2 REG_FLD(32, 0) + +#endif /* __MDP_RDMA_REGS_H__ */ +
Add pseudo ovl module files: Pseudo ovl is an encapsulated module and designed for simplified DRM control flow. This module is composed of 8 RDMAs, 4 MERGEs and an ETHDR. Two RDMAs merge into one layer, so this module support 4 layers. Signed-off-by: Nancy.Lin <nancy.lin@mediatek.com> --- drivers/gpu/drm/mediatek/Makefile | 4 +- .../gpu/drm/mediatek/mtk_disp_pseudo_ovl.c | 593 ++++++++++++++++++ .../gpu/drm/mediatek/mtk_disp_pseudo_ovl.h | 23 + drivers/gpu/drm/mediatek/mtk_mdp_rdma.c | 456 ++++++++++++++ drivers/gpu/drm/mediatek/mtk_mdp_rdma.h | 109 ++++ drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h | 160 +++++ 6 files changed, 1344 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.c create mode 100644 drivers/gpu/drm/mediatek/mtk_disp_pseudo_ovl.h create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_rdma.c create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_rdma.h create mode 100644 drivers/gpu/drm/mediatek/mtk_mdp_reg_rdma.h