lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <926021c1-d9d1-4449-f01b-f405b48f326a@quicinc.com>
Date:   Tue, 24 Jan 2023 15:52:46 -0800
From:   Kuogee Hsieh <quic_khsieh@...cinc.com>
To:     Marijn Suijten <marijn.suijten@...ainline.org>
CC:     <dri-devel@...ts.freedesktop.org>, <robdclark@...il.com>,
        <sean@...rly.run>, <swboyd@...omium.org>, <dianders@...omium.org>,
        <vkoul@...nel.org>, <daniel@...ll.ch>, <airlied@...il.com>,
        <agross@...nel.org>, <dmitry.baryshkov@...aro.org>,
        <andersson@...nel.org>, <quic_abhinavk@...cinc.com>,
        <quic_sbillaka@...cinc.com>, <freedreno@...ts.freedesktop.org>,
        <linux-arm-msm@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 10/14] drm/msm/disp/dpu: add supports of DSC encoder
 v1.2 engine


On 1/23/2023 12:11 PM, Marijn Suijten wrote:
> add support for*
>
> drm/msm/dpu*
>
> On 2023-01-23 10:24:30, Kuogee Hsieh wrote:
>> DSC V1.2 encoder engine is newly added hardware module. This patch
>> add support functions to configure and enable DSC V1.2 encoder engine.
>>
>> Signed-off-by: Kuogee Hsieh <quic_khsieh@...cinc.com>
>> ---
>>   drivers/gpu/drm/msm/Makefile                   |   1 +
>>   drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c    |   2 +-
>>   drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h |  60 +++-
>>   drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.c     |  23 +-
>>   drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.h     |  23 +-
>>   drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc_1_2.c | 371 +++++++++++++++++++++++++
>>   6 files changed, 463 insertions(+), 17 deletions(-)
>>   create mode 100644 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc_1_2.c
>>
>> diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
>> index 28cf52b..271c29a15 100644
>> --- a/drivers/gpu/drm/msm/Makefile
>> +++ b/drivers/gpu/drm/msm/Makefile
>> @@ -65,6 +65,7 @@ msm-$(CONFIG_DRM_MSM_DPU) += \
>>   	disp/dpu1/dpu_hw_catalog.o \
>>   	disp/dpu1/dpu_hw_ctl.o \
>>   	disp/dpu1/dpu_hw_dsc.o \
>> +	disp/dpu1/dpu_hw_dsc_1_2.o \
>>   	disp/dpu1/dpu_dsc_helper.o \
>>   	disp/dpu1/dpu_hw_interrupts.o \
>>   	disp/dpu1/dpu_hw_intf.o \
>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
>> index 7f4a439..901e317 100644
>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
>> @@ -1821,7 +1821,7 @@ static void dpu_encoder_dsc_pipe_cfg(struct dpu_hw_dsc *hw_dsc,
>>   				     u32 initial_lines)
>>   {
>>   	if (hw_dsc->ops.dsc_config)
>> -		hw_dsc->ops.dsc_config(hw_dsc, dsc, common_mode, initial_lines);
>> +		hw_dsc->ops.dsc_config(hw_dsc, dsc, common_mode, initial_lines, false);
> As usual, an enum is better: readers have no idea what a free-floating
> bool means.
>
>>   
>>   	if (hw_dsc->ops.dsc_config_thresh)
>>   		hw_dsc->ops.dsc_config_thresh(hw_dsc, dsc);
>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
>> index 978e3bd..7b0b092 100644
>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
>> @@ -1,6 +1,6 @@
>>   /* SPDX-License-Identifier: GPL-2.0-only */
>>   /*
>> - * Copyright (c) 2022. Qualcomm Innovation Center, Inc. All rights reserved.
>> + * Copyright (c) 2022-2023, Qualcomm Innovation Center, Inc. All rights reserved.
>>    * Copyright (c) 2015-2018, 2020 The Linux Foundation. All rights reserved.
>>    */
>>   
>> @@ -11,6 +11,7 @@
>>   #include <linux/bug.h>
>>   #include <linux/bitmap.h>
>>   #include <linux/err.h>
>> +#include "dpu_hw_mdss.h"
> Unused if you remove the unused DECLARE_BITMAP(dsc_pair_mask, DSC_MAX).
>
>>   
>>   /**
>>    * Max hardware block count: For ex: max 12 SSPP pipes or
>> @@ -182,6 +183,7 @@ enum {
>>    * @DPU_PINGPONG_TE2        Additional tear check block for split pipes
>>    * @DPU_PINGPONG_SPLIT      PP block supports split fifo
>>    * @DPU_PINGPONG_SLAVE      PP block is a suitable slave for split fifo
>> + * @DPU_PINGPONG_DSC,       Display stream compression blocks
>>    * @DPU_PINGPONG_DITHER,    Dither blocks
>>    * @DPU_PINGPONG_MAX
>>    */
>> @@ -190,10 +192,32 @@ enum {
>>   	DPU_PINGPONG_TE2,
>>   	DPU_PINGPONG_SPLIT,
>>   	DPU_PINGPONG_SLAVE,
>> +	DPU_PINGPONG_DSC,
> This is not used.
>
>>   	DPU_PINGPONG_DITHER,
>>   	DPU_PINGPONG_MAX
>>   };
>>   
>> +
>> +/** DSC sub-blocks/features
> Newline between /** and the text.
>
>> + * @DPU_DSC_OUTPUT_CTRL         Supports the control of the pp id which gets
>> + *                              the pixel output from this DSC.
> The original comment is much more concise, can we keep it?
>
>> + * @DPU_DSC_HW_REV_1_1          dsc block supports dsc 1.1 only
>> + * @DPU_DSC_HW_REV_1_2          dsc block supports dsc 1.1 and 1.2
> Capitalize DSC just like elsewhere.
>
>> + * @DPU_DSC_NATIVE_422_EN,      Supports native422 and native420 encoding
>> + * @DPU_DSC_ENC,                DSC encoder sub block
>> + * @DPU_DSC_CTL,                DSC ctl sub block
> No need for trailing commas in doc comments; if anything replace them
> with colons?
>
>> + * @DPU_DSC_MAX
>> + */
>> +enum {
>> +	DPU_DSC_OUTPUT_CTRL = 0x1,
>> +	DPU_DSC_HW_REV_1_1,
>> +	DPU_DSC_HW_REV_1_2,
>> +	DPU_DSC_NATIVE_422_EN,
>> +	DPU_DSC_ENC,
>> +	DPU_DSC_CTL,
> These two enum values only have a meaning within the dpu_hw_dsc_1_2.c
> file, and have nothing to do with the other feature flags/block
> description.  Please move them there (and give _dsc_subblk_offset a
> proper enum type).
>
>> +	DPU_DSC_MAX
>> +};
>> +
>>   /**
>>    * CTL sub-blocks
>>    * @DPU_CTL_SPLIT_DISPLAY:	CTL supports video mode split display
>> @@ -276,15 +300,6 @@ enum {
>>   };
>>   
>>   /**
>> - * DSC features
>> - * @DPU_DSC_OUTPUT_CTRL       Configure which PINGPONG block gets
>> - *                            the pixel output from this DSC.
>> - */
>> -enum {
>> -	DPU_DSC_OUTPUT_CTRL = 0x1,
> Did this have to move?
>
>> -};
>> -
>> -/**
>>    * MACRO DPU_HW_BLK_INFO - information of HW blocks inside DPU
>>    * @name:              string name for debug purposes
>>    * @id:                enum identifying this block
>> @@ -346,6 +361,14 @@ struct dpu_pp_blk {
>>   };
>>   
>>   /**
>> + * struct dpu_dsc_blk : DSC Encoder sub-blk information
> Use a hyphen here and everywhere else:
> https://docs.kernel.org/doc-guide/kernel-doc.html
>
>> + * @info:   HW register and features supported by this sub-blk
>> + */
>> +struct dpu_dsc_blk {
>> +	DPU_HW_SUBBLK_INFO;
>> +};
>> +
>> +/**
>>    * enum dpu_qos_lut_usage - define QoS LUT use cases
>>    */
>>   enum dpu_qos_lut_usage {
>> @@ -403,6 +426,7 @@ struct dpu_rotation_cfg {
>>    * @pixel_ram_size     size of latency hiding and de-tiling buffer in bytes
>>    * @max_hdeci_exp      max horizontal decimation supported (max is 2^value)
>>    * @max_vdeci_exp      max vertical decimation supported (max is 2^value)
>> + * @max_dsc_width      max dsc line width support.
> DSC*
>
>>    */
>>   struct dpu_caps {
>>   	u32 max_mixer_width;
>> @@ -419,6 +443,7 @@ struct dpu_caps {
>>   	u32 pixel_ram_size;
>>   	u32 max_hdeci_exp;
>>   	u32 max_vdeci_exp;
>> +	u32 max_dsc_width;
> This is never read.
>
>>   };
>>   
>>   /**
>> @@ -494,9 +519,20 @@ struct dpu_dspp_sub_blks {
>>   struct dpu_pingpong_sub_blks {
>>   	struct dpu_pp_blk te;
>>   	struct dpu_pp_blk te2;
>> +	struct dpu_pp_blk dsc;
> Unused.
>
>>   	struct dpu_pp_blk dither;
>>   };
>>   
>> +
>> +/**
>> + * struct dpu_dsc_sub_blks : DSC sub-blks
>> + *
> A sub-block of sub-blocks?  Use the documentation to explain what this
> is for, describe @enc and @ctl.
>
>> + */
>> +struct dpu_dsc_sub_blks {
>> +	struct dpu_dsc_blk enc;
>> +	struct dpu_dsc_blk ctl;
>> +};
>> +
>>   /**
>>    * dpu_clk_ctrl_type - Defines top level clock control signals
>>    */
>> @@ -641,10 +677,14 @@ struct dpu_merge_3d_cfg  {
>>    * struct dpu_dsc_cfg - information of DSC blocks
>>    * @id                 enum identifying this block
>>    * @base               register offset of this block
>> + * @len:               length of hardware block
>>    * @features           bit mask identifying sub-blocks/features
>> + * @dsc_pair_mask:     Bitmask of DSCs that can be controlled by same CTL
>>    */
>>   struct dpu_dsc_cfg {
>>   	DPU_HW_BLK_INFO;
>> +	DECLARE_BITMAP(dsc_pair_mask, DSC_MAX);
> This bitmask is unused.
>
>> +	const struct dpu_dsc_sub_blks *sblk;
>>   };
>>   
>>   /**
>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.c
>> index 619926d..51e8890 100644
>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.c
>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.c
>> @@ -1,6 +1,7 @@
>>   // SPDX-License-Identifier: GPL-2.0-only
>>   /*
>>    * Copyright (c) 2020-2022, Linaro Limited
>> + * Copyright (c) 2023 Qualcomm Innovation Center, Inc. All rights reserved
>>    */
>>   
>>   #include "dpu_kms.h"
>> @@ -41,10 +42,11 @@ static void dpu_hw_dsc_disable(struct dpu_hw_dsc *dsc)
>>   static void dpu_hw_dsc_config(struct dpu_hw_dsc *hw_dsc,
>>   			      struct drm_dsc_config *dsc,
>>   			      u32 mode,
>> -			      u32 initial_lines)
>> +			      u32 initial_lines,
>> +			      bool ich_reset_override)
>>   {
>>   	struct dpu_hw_blk_reg_map *c = &hw_dsc->hw;
>> -	u32 data;
>> +	u32 data, lsb, bpp;
>>   	u32 slice_last_group_size;
>>   	u32 det_thresh_flatness;
>>   	bool is_cmd_mode = !(mode & DSC_MODE_VIDEO);
>> @@ -58,7 +60,14 @@ static void dpu_hw_dsc_config(struct dpu_hw_dsc *hw_dsc,
>>   	data = (initial_lines << 20);
>>   	data |= ((slice_last_group_size - 1) << 18);
>>   	/* bpp is 6.4 format, 4 LSBs bits are for fractional part */
>> -	data |= (dsc->bits_per_pixel << 8);
>> +	data |= dsc->bits_per_pixel << 12;
>> +	lsb = dsc->bits_per_pixel % 4;
>> +	bpp = dsc->bits_per_pixel / 4;
>> +	bpp *= 4;
>> +	bpp <<= 4;
>> +	bpp |= lsb;
>> +
>> +	data |= bpp << 8;
> Why are you re-adding this nonsense?  It was removed in [1] _and_ does
> not account for bits_per_pixel _already being in x.4 format_. This will
> regress existing hardware.
>
> [1]: https://lore.kernel.org/linux-arm-msm/20221026182824.876933-10-marijn.suijten@somainline.org/
>
>>   	data |= (dsc->block_pred_enable << 7);
>>   	data |= (dsc->line_buf_depth << 3);
>>   	data |= (dsc->simple_422 << 2);
>> @@ -221,7 +230,13 @@ struct dpu_hw_dsc *dpu_hw_dsc_init(enum dpu_dsc idx, void __iomem *addr,
>>   
>>   	c->idx = idx;
>>   	c->caps = cfg;
>> -	_setup_dsc_ops(&c->ops, c->caps->features);
>> +
>> +	if (test_bit(DPU_DSC_HW_REV_1_1, &c->caps->features))
>> +		_setup_dsc_ops(&c->ops, c->caps->features);
>> +	else if (test_bit(DPU_DSC_HW_REV_1_2, &c->caps->features))
>> +		dpu_dsc_1_2_setup_ops(&c->ops, c->caps->features);
>> +	else
>> +		_setup_dsc_ops(&c->ops, c->caps->features);
>>   
>>   	return c;
>>   }
>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.h
>> index ae9b5db..a48f572 100644
>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.h
>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc.h
>> @@ -1,5 +1,8 @@
>>   /* SPDX-License-Identifier: GPL-2.0-only */
>> -/* Copyright (c) 2020-2022, Linaro Limited */
>> +/*
>> + * Copyright (c) 2020-2022, Linaro Limited
>> + * Copyright (c) 2023 Qualcomm Innovation Center, Inc. All rights reserved
>> + */
>>   
>>   #ifndef _DPU_HW_DSC_H
>>   #define _DPU_HW_DSC_H
>> @@ -33,7 +36,8 @@ struct dpu_hw_dsc_ops {
>>   	void (*dsc_config)(struct dpu_hw_dsc *hw_dsc,
>>   			   struct drm_dsc_config *dsc,
>>   			   u32 mode,
>> -			   u32 initial_lines);
>> +			   u32 initial_lines,
>> +			   bool ich_reset_override);
>>   
>>   	/**
>>   	 * dsc_config_thresh - programs panel thresholds
>> @@ -43,6 +47,12 @@ struct dpu_hw_dsc_ops {
>>   	void (*dsc_config_thresh)(struct dpu_hw_dsc *hw_dsc,
>>   				  struct drm_dsc_config *dsc);
>>   
>> +	/**
>> +	 * bind_pingpong_blk - enable/disable the connection with pp
> Inherit docs from the enum.
>
>> +	 * @hw_dsc: Pointer to dsc context
> DSC*
>
>> +	 * @enable: enable/disable connection
>> +	 * @pp: pingpong blk id
> It's documentation, write out block fully.
>
>> +	 */
>>   	void (*dsc_bind_pingpong_blk)(struct dpu_hw_dsc *hw_dsc,
>>   				  bool enable,
>>   				  enum dpu_pingpong pp);
>> @@ -51,6 +61,7 @@ struct dpu_hw_dsc_ops {
>>   struct dpu_hw_dsc {
>>   	struct dpu_hw_blk base;
>>   	struct dpu_hw_blk_reg_map hw;
>> +	struct dpu_hw_ctl *hw_ctl;
> Unused.
>
>>   
>>   	/* dsc */
>>   	enum dpu_dsc idx;
>> @@ -76,9 +87,17 @@ struct dpu_hw_dsc *dpu_hw_dsc_init(enum dpu_dsc idx, void __iomem *addr,
>>    */
>>   void dpu_hw_dsc_destroy(struct dpu_hw_dsc *dsc);
>>   
>> +/**
>> + * dpu_hw_dsc - convert base object dpu_hw_base to container
>> + * @hw: Pointer to base hardware block
>> + * return: Pointer to hardware block container
>> + */
>>   static inline struct dpu_hw_dsc *to_dpu_hw_dsc(struct dpu_hw_blk *hw)
>>   {
>>   	return container_of(hw, struct dpu_hw_dsc, base);
>>   }
>>   
>> +void dpu_dsc_1_2_setup_ops(struct dpu_hw_dsc_ops *ops,
>> +		const unsigned long features);
>> +
>>   #endif /* _DPU_HW_DSC_H */
>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc_1_2.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc_1_2.c
>> new file mode 100644
>> index 00000000..2be74ae
>> --- /dev/null
>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_dsc_1_2.c
>> @@ -0,0 +1,371 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (c) 2020-2021, The Linux Foundation. All rights reserved.
>> + * Copyright (c) 2023 Qualcomm Innovation Center, Inc. All rights reserved
>> + */
>> +
>> +#include "dpu_kms.h"
>> +#include "dpu_hw_catalog.h"
>> +#include "dpu_hwio.h"
>> +#include "dpu_hw_mdss.h"
>> +#include "dpu_hw_dsc.h"
>> +
>> +
>> +#define DSC_CMN_MAIN_CNF           0x00
>> +
>> +/* DPU_DSC_ENC register offsets */
>> +#define ENC_DF_CTRL                0x00
>> +#define ENC_GENERAL_STATUS         0x04
>> +#define ENC_HSLICE_STATUS          0x08
>> +#define ENC_OUT_STATUS             0x0C
>> +#define ENC_INT_STAT               0x10
>> +#define ENC_INT_CLR                0x14
>> +#define ENC_INT_MASK               0x18
>> +#define DSC_MAIN_CONF              0x30
>> +#define DSC_PICTURE_SIZE           0x34
>> +#define DSC_SLICE_SIZE             0x38
>> +#define DSC_MISC_SIZE              0x3C
>> +#define DSC_HRD_DELAYS             0x40
>> +#define DSC_RC_SCALE               0x44
>> +#define DSC_RC_SCALE_INC_DEC       0x48
>> +#define DSC_RC_OFFSETS_1           0x4C
>> +#define DSC_RC_OFFSETS_2           0x50
>> +#define DSC_RC_OFFSETS_3           0x54
>> +#define DSC_RC_OFFSETS_4           0x58
>> +#define DSC_FLATNESS_QP            0x5C
>> +#define DSC_RC_MODEL_SIZE          0x60
>> +#define DSC_RC_CONFIG              0x64
>> +#define DSC_RC_BUF_THRESH_0        0x68
>> +#define DSC_RC_BUF_THRESH_1        0x6C
>> +#define DSC_RC_BUF_THRESH_2        0x70
>> +#define DSC_RC_BUF_THRESH_3        0x74
>> +#define DSC_RC_MIN_QP_0            0x78
>> +#define DSC_RC_MIN_QP_1            0x7C
>> +#define DSC_RC_MIN_QP_2            0x80
>> +#define DSC_RC_MAX_QP_0            0x84
>> +#define DSC_RC_MAX_QP_1            0x88
>> +#define DSC_RC_MAX_QP_2             0x8C
>> +#define DSC_RC_RANGE_BPG_OFFSETS_0  0x90
>> +#define DSC_RC_RANGE_BPG_OFFSETS_1  0x94
>> +#define DSC_RC_RANGE_BPG_OFFSETS_2  0x98
> Reindent to line this back up.
>
>> +
>> +/* DPU_DSC_CTL register offsets */
>> +#define DSC_CTL                    0x00
>> +#define DSC_CFG                    0x04
>> +#define DSC_DATA_IN_SWAP           0x08
>> +#define DSC_CLK_CTRL               0x0C
>> +
>> +
>> +static int _dsc_calc_ob_max_addr(struct dpu_hw_dsc *hw_dsc, int num_ss)
>> +{
>> +	enum dpu_dsc idx;
>> +
>> +	idx = hw_dsc->idx;
>> +
>> +	if (!(hw_dsc->caps->features & BIT(DPU_DSC_NATIVE_422_EN))) {
> Why not swap the bodies instead of inverting this.
>
>> +		if (num_ss == 1)
>> +			return 2399;
>> +		else if (num_ss == 2)
>> +			return 1199;
>> +	} else {
>> +		if (num_ss == 1)
>> +			return 1199;
>> +		else if (num_ss == 2)
>> +			return 599;
>> +	}
>> +	return 0;
>> +}
>> +
>> +static inline int _dsc_subblk_offset(struct dpu_hw_dsc *hw_dsc, int s_id,
>> +		u32 *idx)
>> +{
>> +	const struct dpu_dsc_sub_blks *sblk;
>> +
>> +	if (!hw_dsc)
>> +		return -EINVAL;
>> +
>> +	*idx = 0;
>> +
>> +	sblk = hw_dsc->caps->sblk;
>> +
>> +	switch (s_id) {
>> +
>> +	case DPU_DSC_ENC:
>> +		*idx = sblk->enc.base;
>> +		break;
>> +	case DPU_DSC_CTL:
>> +		*idx = sblk->ctl.base;
>> +		break;
>> +	default:
>> +		return -EINVAL;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static void dpu_hw_dsc_disable_1_2(struct dpu_hw_dsc *hw_dsc)
>> +{
>> +	struct dpu_hw_blk_reg_map *hw;
>> +	u32 idx;
> Can we rename these to offset or subblk_offset or something more clear?
>
>> +
>> +	if (!hw_dsc)
>> +		return;
>> +
>> +	if (_dsc_subblk_offset(hw_dsc, DPU_DSC_CTL, &idx))
>> +		return;
> These error checks are excessive: you pass in a non-null hw_dsc and
> known enum constant - _dsc_subblk_offset should perhaps not return
> errors at all.
>
>> +
>> +	hw = &hw_dsc->hw;
>> +	DPU_REG_WRITE(hw, DSC_CFG + idx, 0);
> Swap the arguments to + so that it's clear that DSC_CFG is a register on
> the subblock offset denoted by "idx", not the other way around.
>
>> +
>> +	if (_dsc_subblk_offset(hw_dsc, DPU_DSC_ENC, &idx))
>> +		return;
>> +
>> +	DPU_REG_WRITE(hw, ENC_DF_CTRL + idx, 0);
>> +	DPU_REG_WRITE(hw, DSC_MAIN_CONF + idx, 0);
>> +}
>> +
>> +static void dpu_hw_dsc_config_1_2(struct dpu_hw_dsc *hw_dsc,
>> +		struct drm_dsc_config *dsc, u32 mode,
>> +		u32 initial_lines, bool ich_reset_override)
>> +{
>> +	struct dpu_hw_blk_reg_map *hw;
>> +	struct msm_display_dsc_info *dsc_info;
>> +	u32 idx;
>> +	u32 data = 0;
>> +	u32 bpp;
>> +	void __iomem *off;
>> +
>> +	if (!hw_dsc || !dsc)
>> +		return;
>> +
>> +	hw = &hw_dsc->hw;
>> +
>> +	dsc_info = to_msm_dsc_info(dsc);
>> +
>> +	if (_dsc_subblk_offset(hw_dsc, DPU_DSC_ENC, &idx))
>> +		return;
>> +
>> +	if (mode & DSC_MODE_SPLIT_PANEL)
>> +		data |= BIT(0);
>> +
>> +	if (mode & DSC_MODE_MULTIPLEX)
>> +		data |= BIT(1);
> These are well known bitwise definitions for a reason, data |= mode will
> do (or out DSC_MODE_VIDEO since you have to shift that one at BIT(9).
>
>> +
>> +	data |= (dsc_info->num_active_ss_per_enc & 0x3) << 7;
>> +
>> +	DPU_REG_WRITE(hw, DSC_CMN_MAIN_CNF, data);
>> +
>> +	data = (dsc_info->initial_lines & 0xff);
> You already get initial_lines passed as function argument, but ignore
> it?
>
>> +	data |= ((mode & DSC_MODE_VIDEO) ? 1 : 0) << 9;
> Yuck. if (mode & DSC_MODE_VIDEO) data |= BIT(9);.
>
>> +	if (ich_reset_override)
>> +		data |= 0xC00; // set bit 10 and 11
> Instead of a comment, make this self-describing BIT(10) | BIT(11) code.
>
>> +	data |= (_dsc_calc_ob_max_addr(hw_dsc, dsc_info->num_active_ss_per_enc) << 18);
>> +
>> +	DPU_REG_WRITE(hw, ENC_DF_CTRL + idx, data);
>> +
>> +	data = (dsc->dsc_version_minor & 0xf) << 28;
>> +	if (dsc->dsc_version_minor == 0x2) {
>> +		if (dsc->native_422)
>> +			data |= BIT(22);
>> +		if (dsc->native_420)
>> +			data |= BIT(21);
>> +	}
>> +
>> +	bpp = dsc->bits_per_pixel;
> As above, don't forget to read the documentation on this field:
>
>      Target bits per pixel with 4 fractional bits, bits_per_pixel << 4
>
>> +	/* as per hw requirement bpp should be programmed
>> +	 * twice the actual value in case of 420 or 422 encoding
>> +	 */
>> +	if (dsc->native_422 || dsc->native_420)
>> +		bpp = 2 * bpp;
>> +	data |= (dsc->block_pred_enable ? 1 : 0) << 20;
>> +	data |= (bpp << 10);
> Either wrap everything or nothing in ().
>
>> +	data |= (dsc->line_buf_depth & 0xf) << 6;
>> +	data |= dsc->convert_rgb << 4;
>> +	data |= dsc->bits_per_component & 0xf;
>> +
>> +	DPU_REG_WRITE(hw, DSC_MAIN_CONF + idx, data);
>> +
>> +	data = (dsc->pic_width & 0xffff) |
>> +		((dsc->pic_height & 0xffff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_PICTURE_SIZE + idx, data);
>> +
>> +	data = (dsc->slice_width & 0xffff) |
>> +		((dsc->slice_height & 0xffff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_SLICE_SIZE + idx, data);
>> +
>> +	DPU_REG_WRITE(hw, DSC_MISC_SIZE + idx,
>> +			(dsc->slice_chunk_size) & 0xffff);
>> +
>> +	data = (dsc->initial_xmit_delay & 0xffff) |
>> +		((dsc->initial_dec_delay & 0xffff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_HRD_DELAYS + idx, data);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_SCALE + idx,
>> +			dsc->initial_scale_value & 0x3f);
>> +
>> +	data = (dsc->scale_increment_interval & 0xffff) |
>> +		((dsc->scale_decrement_interval & 0x7ff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_SCALE_INC_DEC + idx, data);
>> +
>> +	data = (dsc->first_line_bpg_offset & 0x1f) |
>> +		((dsc->second_line_bpg_offset & 0x1f) << 5);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_OFFSETS_1 + idx, data);
>> +
>> +	data = (dsc->nfl_bpg_offset & 0xffff) |
>> +		((dsc->slice_bpg_offset & 0xffff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_OFFSETS_2 + idx, data);
>> +
>> +	data = (dsc->initial_offset & 0xffff) |
>> +		((dsc->final_offset & 0xffff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_OFFSETS_3 + idx, data);
>> +
>> +	data = (dsc->nsl_bpg_offset & 0xffff) |
>> +		((dsc->second_line_offset_adj & 0xffff) << 16);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_OFFSETS_4 + idx, data);
>> +
>> +	data = (dsc->flatness_min_qp & 0x1f);
>> +	data |= (dsc->flatness_max_qp & 0x1f) << 5;
>> +	data |= (dsc_info->det_thresh_flatness & 0xff) << 10;
> dpu_hw_dsc.c computes this on the fly.  After removing that, and
> using initial_lines from the function parameters, only
> dsc_info->num_active_ss_per_enc remains.  Do you really need that
> msm_display_dsc_info struct here, do you need it at all?

I ported these code from our down stream code base.

I make it work first, then clean it up will follow.

I submit it for review since it looks like you guy like to have code sooner.

yes, eliminate msm_display_dsc_info is my next target and hope it can be 
done.





>
>> +
>> +	DPU_REG_WRITE(hw, DSC_FLATNESS_QP + idx, data);
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_MODEL_SIZE + idx,
>> +			(dsc->rc_model_size) & 0xffff);
>> +
>> +	data = dsc->rc_edge_factor & 0xf;
>> +	data |= (dsc->rc_quant_incr_limit0 & 0x1f) << 8;
>> +	data |= (dsc->rc_quant_incr_limit1 & 0x1f) << 13;
>> +	data |= (dsc->rc_tgt_offset_high & 0xf) << 20;
>> +	data |= (dsc->rc_tgt_offset_low & 0xf) << 24;
>> +
>> +	DPU_REG_WRITE(hw, DSC_RC_CONFIG + idx, data);
>> +
>> +	/* program the dsc wrapper */
>> +	if (_dsc_subblk_offset(hw_dsc, DPU_DSC_CTL, &idx))
>> +		return;
>> +
>> +	off = hw->blk_addr + idx;
>> +
>> +	data = BIT(0); /* encoder enable */
>> +	if (dsc->native_422)
>> +		data |= BIT(8);
>> +	else if (dsc->native_420)
>> +		data |= BIT(9);
>> +	if (!dsc->convert_rgb)
>> +		data |= BIT(10);
>> +	if (dsc->bits_per_component == 8)
>> +		data |= BIT(11);
>> +	if (mode & DSC_MODE_SPLIT_PANEL)
>> +		data |= BIT(12);
>> +	if (mode & DSC_MODE_MULTIPLEX)
>> +		data |= BIT(13);
>> +	if (!(mode & DSC_MODE_VIDEO))
>> +		data |= BIT(17);
>> +
>> +	if (dsc_info->dsc_4hsmerge_en) {
>> +		data |= dsc_info->dsc_4hsmerge_padding << 18;
>> +		data |= dsc_info->dsc_4hsmerge_alignment << 22;
>> +		data |= BIT(16);
>> +	}
>> +
>> +	DPU_REG_WRITE(hw, DSC_CFG + idx, data);
>> +
>> +//	DPU_REG_WRITE(hw, DSC_DATA_IN_SWAP + idx, 0x14e5);
> No commented-out code please, especially not with //
>
>> +}
>> +
>> +static void dpu_hw_dsc_config_thresh_1_2(struct dpu_hw_dsc *hw_dsc,
>> +		struct drm_dsc_config *dsc)
>> +{
>> +	struct dpu_hw_blk_reg_map *hw;
>> +	struct msm_display_dsc_info *dsc_info;
>> +	u32 idx, off;
>> +	int i, j = 0;
>> +	struct drm_dsc_rc_range_parameters *rc;
>> +	u32 data = 0, min_qp = 0, max_qp = 0, bpg_off = 0;
>> +
>> +	if (!hw_dsc || !dsc)
>> +		return;
>> +
>> +	if (_dsc_subblk_offset(hw_dsc, DPU_DSC_ENC, &idx))
>> +		return;
>> +
>> +	hw = &hw_dsc->hw;
>> +
>> +	dsc_info = to_msm_dsc_info(dsc);
>> +
>> +	rc = dsc->rc_range_params;
>> +
>> +	off = 0;
>> +	for (i = 0; i < DSC_NUM_BUF_RANGES - 1; i++) {
>> +		data |= dsc->rc_buf_thresh[i] << (8*j);
> Lack of spaces does not make this multiplication any prettier to read.
>
> * has precedence over << but it's better to replicate the () below as
> well.
>
>> +		j++;
>> +		if ((j == 4) || (i == DSC_NUM_BUF_RANGES - 2)) {
>> +			DPU_REG_WRITE(hw, DSC_RC_BUF_THRESH_0 + idx + off,
>> +					data);
>> +			off += 4;
>> +			j = 0;
>> +			data = 0;
>> +		}
>> +	}
>> +
>> +	off = 0;
>> +	for (i = 0; i < DSC_NUM_BUF_RANGES; i++) {
>> +		min_qp |= (rc[i].range_min_qp & 0x1f) << 5*j;
>> +		max_qp |= (rc[i].range_max_qp & 0x1f) << 5*j;
>> +		bpg_off |= (rc[i].range_bpg_offset & 0x3f) << 6*j;
> These values _must_ already be masked to be useful in
> drm_dsc_compute_rc_parameters(), no need to mask them again just like
> the v1.1 block implementation.
>
>> +		j++;
>> +		if (j == 5) {
>> +			DPU_REG_WRITE(hw, DSC_RC_MIN_QP_0 + idx + off,
>> +					min_qp);
>> +			DPU_REG_WRITE(hw, DSC_RC_MAX_QP_0 + idx + off,
>> +					max_qp);
>> +			DPU_REG_WRITE(hw,
>> +					DSC_RC_RANGE_BPG_OFFSETS_0 + idx + off,
>> +					bpg_off);
>> +			off += 4;
>> +			j = 0;
>> +			min_qp = 0;
>> +			max_qp = 0;
>> +			bpg_off = 0;
>> +		}
>> +	}
>> +}
>> +
>> +static void dpu_hw_dsc_bind_pingpong_blk_1_2(
>> +		struct dpu_hw_dsc *hw_dsc,
>> +		bool enable,
>> +		const enum dpu_pingpong pp)
>> +{
>> +	struct dpu_hw_blk_reg_map *hw;
>> +	int idx;
>> +	int mux_cfg = 0xF; /* Disabled */
> Lowercase hex (and anywhere else if I skipped any).
>
>> +
>> +	if (!hw_dsc)
>> +		return;
> As with the v1.1 implementation, we don't check this, and your function
> below also checks it (but it does not need to).
>
>> +	if (_dsc_subblk_offset(hw_dsc, DPU_DSC_CTL, &idx))
>> +		return;
>> +
>> +	hw = &hw_dsc->hw;
>> +	if (enable)
>> +		mux_cfg = (pp - PINGPONG_0) & 0x7;
>> +
>> +	DPU_REG_WRITE(hw, DSC_CTL + idx, mux_cfg);
>> +}
>> +
>> +void dpu_dsc_1_2_setup_ops(struct dpu_hw_dsc_ops *ops,
>> +		const unsigned long features)
>> +{
>> +	ops->dsc_disable = dpu_hw_dsc_disable_1_2;
>> +	ops->dsc_config = dpu_hw_dsc_config_1_2;
>> +	ops->dsc_config_thresh = dpu_hw_dsc_config_thresh_1_2;
>> +	ops->dsc_bind_pingpong_blk = dpu_hw_dsc_bind_pingpong_blk_1_2;
>> +}
>> -- 
>> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
>> a Linux Foundation Collaborative Project
>>
> All in all you really need to revise and clean your patches before
> sending them to the lists; these are already far too many comments and
> nits, and massively take away from reviewing code behaviour which I have
> not even started with after looking at 1 out of 14 patches :(
>
> - Marijn

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ