public inbox for linux-mediatek@lists.infradead.org
 help / color / mirror / Atom feed
From: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
To: "kyrie.wu" <kyrie.wu@mediatek.com>,
	Hans Verkuil <hverkuil-cisco@xs4all.nl>,
	Mauro Carvalho Chehab <mchehab@kernel.org>,
	Rob Herring <robh+dt@kernel.org>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	Tzung-Bi Shih <tzungbi@chromium.org>
Cc: Rob Herring <robh@kernel.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Bin Liu <bin.liu@mediatek.com>,
	kyrie wu <kyrie.wu@mediatek.corp-partner.google.com>,
	linux-media@vger.kernel.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-mediatek@lists.infradead.org
Subject: Re: [V1,04/12] media: mediatek: jpeg: add jpeg smmu sid setting
Date: Fri, 10 Jan 2025 10:39:38 +0100	[thread overview]
Message-ID: <73144187-7f7f-4f68-aa5a-5fc019e646ec@collabora.com> (raw)
In-Reply-To: <20250109133513.20151-5-kyrie.wu@mediatek.com>

Il 09/01/25 14:35, kyrie.wu ha scritto:
> Add a configuration to set jpeg dec & enc smmu sid
> 
> Signed-off-by: kyrie.wu <kyrie.wu@mediatek.com>
> ---
>   .../platform/mediatek/jpeg/mtk_jpeg_core.c    |  6 ++++
>   .../platform/mediatek/jpeg/mtk_jpeg_core.h    | 11 +++++++
>   .../platform/mediatek/jpeg/mtk_jpeg_dec_hw.c  | 30 +++++++++++++++++--
>   .../platform/mediatek/jpeg/mtk_jpeg_dec_hw.h  |  4 +++
>   .../platform/mediatek/jpeg/mtk_jpeg_enc_hw.c  | 26 ++++++++++++++++
>   .../platform/mediatek/jpeg/mtk_jpeg_enc_hw.h  |  4 +++
>   6 files changed, 78 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c
> index c3ccc525d9fd..77b3bd6c4d3f 100644
> --- a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c
> +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c
> @@ -1655,6 +1655,9 @@ static void mtk_jpegenc_worker(struct work_struct *work)
>   	jpeg_dst_buf->frame_num = ctx->total_frame_num;
>   	ctx->total_frame_num++;
>   	mtk_jpeg_enc_reset(comp_jpeg[hw_id]->reg_base);
> +#if IS_ENABLED(CONFIG_ARM_SMMU_V3)
> +	mtk_jpeg_enc_set_smmu_sid(hw_id);
> +#endif
>   	mtk_jpeg_set_enc_dst(ctx,
>   			     comp_jpeg[hw_id]->reg_base,
>   			     &dst_buf->vb2_buf);
> @@ -1771,6 +1774,9 @@ static void mtk_jpegdec_worker(struct work_struct *work)
>   	spin_lock_irqsave(&comp_jpeg[hw_id]->hw_lock, flags);
>   	ctx->total_frame_num++;
>   	mtk_jpeg_dec_reset(comp_jpeg[hw_id]->reg_base);
> +#if IS_ENABLED(CONFIG_ARM_SMMU_V3)

You can avoid using preprocessor conditionals - and then, this doesn't look like an
ARM SMMUv3 configuration, but rather a JPEG dec/enc HW config, so in that case the
enclosing of this in the proposed config option would even be wrong.

Use platform data for that.

> +	mtk_jpeg_dec_set_smmu_sid(hw_id);
> +#endif
>   	mtk_jpeg_dec_set_config(comp_jpeg[hw_id]->reg_base,
>   				jpeg->variant->support_34bit,
>   				&jpeg_src_buf->dec_param,
> diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h
> index 8fddc133c46c..d3aba1e6cae8 100644
> --- a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h
> +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h
> @@ -36,6 +36,17 @@
>   
>   #define MTK_JPEG_ADDR_MASK GENMASK(1, 0)
>   
> +#if IS_ENABLED(CONFIG_ARM_SMMU_V3)
> +#define JPG_REG_CORE0_GUSER_ID                          0x380d0000
> +#define JPG_REG_CORE1_GUSER_ID                          0x388d0000

At least these two definitions shall go in the devicetree reg node.

> +#define JPG_REG_GUSER_ID_MASK                           0x7
> +#define JPG_REG_GUSER_ID_DEC_SID                        0x4
> +#define JPG_REG_GUSER_ID_ENC_SID                        0x5
> +#define JPG_REG_DEC_GUSER_ID_SHIFT                      8
> +#define JPG_REG_ENC_GUSER_ID_SHIFT                      4

If this is setting an IOMMU SID, then you can just use the "iommus" property
to pass a handle to the IOMMU that this device is using plus the streamid,
which you can retrieve and write to the guser_id_{dec,enc}_sid register.

There's no reason to hardcode that in this header, and actually hardcoding
will give you a number of issues (example: unpowered/unclocked access to the
IOMMU, and many others).

> +#define GUSER_ID_MAPRANGE                               4
> +#endif
> +
>   /**
>    * enum mtk_jpeg_ctx_state - states of the context state machine
>    * @MTK_JPEG_INIT:		current state is initialized
> diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c
> index d868e46aaf37..fadfc4b5e366 100644
> --- a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c
> +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c
> @@ -274,6 +274,32 @@ void mtk_jpeg_dec_reset(void __iomem *base)
>   }
>   EXPORT_SYMBOL_GPL(mtk_jpeg_dec_reset);
>   
> +#if IS_ENABLED(CONFIG_ARM_SMMU_V3)
> +void mtk_jpeg_dec_set_smmu_sid(int hwid)
> +{
> +	void __iomem *dec_reg_base;
> +	u32 val, mask;
> +
> +	if (hwid)
> +		dec_reg_base = ioremap(JPG_REG_CORE1_GUSER_ID, GUSER_ID_MAPRANGE);
> +	else
> +		dec_reg_base = ioremap(JPG_REG_CORE0_GUSER_ID, GUSER_ID_MAPRANGE);
> +	if (!dec_reg_base) {
> +		dev_err(jpeg->dev, "Failed to map jpgdec JPG_REG_GUSER_ID\n");
> +		return;
> +	}
> +
> +	val = ioread32(dec_reg_base);
> +	mask = ~(JPG_REG_GUSER_ID_MASK << JPG_REG_DEC_GUSER_ID_SHIFT);
> +	val &= mask;
> +	val |= (JPG_REG_GUSER_ID_DEC_SID << JPG_REG_DEC_GUSER_ID_SHIFT);
> +
> +	iowrite32(val, dec_reg_base);
> +	iounmap(dec_reg_base);
> +}
> +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_set_smmu_sid);
> +#endif
> +
>   static void mtk_jpeg_dec_set_brz_factor(void __iomem *base, u8 yscale_w,
>   				u8 yscale_h, u8 uvscale_w, u8 uvscale_h)
>   {
> @@ -552,7 +578,6 @@ static irqreturn_t mtk_jpegdec_hw_irq_handler(int irq, void *priv)
>   	struct vb2_v4l2_buffer *src_buf, *dst_buf;
>   	struct mtk_jpeg_src_buf *jpeg_src_buf;
>   	enum vb2_buffer_state buf_state;
> -	struct mtk_jpeg_ctx *ctx;
>   	u32 dec_irq_ret;
>   	u32 irq_status;
>   	int i;
> @@ -562,7 +587,6 @@ static irqreturn_t mtk_jpegdec_hw_irq_handler(int irq, void *priv)
>   
>   	cancel_delayed_work(&jpeg->job_timeout_work);
>   
> -	ctx = jpeg->hw_param.curr_ctx;
>   	src_buf = jpeg->hw_param.src_buffer;
>   	dst_buf = jpeg->hw_param.dst_buffer;
>   	v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true);
> @@ -585,7 +609,7 @@ static irqreturn_t mtk_jpegdec_hw_irq_handler(int irq, void *priv)
>   	buf_state = VB2_BUF_STATE_DONE;
>   	v4l2_m2m_buf_done(src_buf, buf_state);
>   	mtk_jpegdec_put_buf(jpeg);
> -	pm_runtime_put(ctx->jpeg->dev);
> +	pm_runtime_put(jpeg->dev);

You're doing more than what you're describing in the commit title and description.

If this change was intentional, please move it to a diffrent commit.

Regards,
Angelo



  reply	other threads:[~2025-01-10  9:41 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-09 13:35 [V1,00/12] Enable jpeg enc & dec multi-hardwares for MT8196 kyrie.wu
2025-01-09 13:35 ` [V1,01/12] dt-bindings: mediatek: Add mediatek, mt8196-jpgdec compatible kyrie.wu
2025-01-09 13:54   ` Krzysztof Kozlowski
2025-04-07  7:13     ` Kyrie Wu (吴晗)
2025-04-07  7:19       ` Krzysztof Kozlowski
2025-04-07  9:39         ` Kyrie Wu (吴晗)
2025-01-09 13:35 ` [V1,02/12] dt-bindings: mediatek: Add mediatek, mt8196-jpgenc compatible kyrie.wu
2025-01-09 13:35 ` [V1,03/12] media: mediatek: jpeg: add jpeg compatible kyrie.wu
2025-01-10  9:46   ` AngeloGioacchino Del Regno
2025-04-07  7:15     ` Kyrie Wu (吴晗)
2025-01-09 13:35 ` [V1,04/12] media: mediatek: jpeg: add jpeg smmu sid setting kyrie.wu
2025-01-10  9:39   ` AngeloGioacchino Del Regno [this message]
2025-04-07  7:21     ` Kyrie Wu (吴晗)
2025-01-09 13:35 ` [V1,05/12] media: mediatek: jpeg: fix jpeg hw count setting kyrie.wu
2025-01-10  9:45   ` AngeloGioacchino Del Regno
2025-04-07  7:23     ` Kyrie Wu (吴晗)
2025-01-09 13:35 ` [V1,06/12] media: mediatek: jpeg: refactor jpeg buffer payload setting kyrie.wu
2025-01-09 13:35 ` [V1,07/12] media: mediatek: jpeg: refactor jpeg dst buffer layout kyrie.wu
2025-01-09 13:35 ` [V1,08/12] media: mediatek: jpeg: fix stop streaming flow for multi-core kyrie.wu
2025-01-09 13:35 ` [V1,09/12] media: mediatek: jpeg: refactor multi-core clk suspend and resume setting kyrie.wu
2025-01-09 13:35 ` [V1,10/12] media: mediatek: jpeg: fix decoding buffer number setting timing issue kyrie.wu
2025-01-09 13:35 ` [V1,11/12] media: mediatek: jpeg: refactor decoding resolution change operation kyrie.wu
2025-01-09 13:35 ` [V1,12/12] media: mediatek: jpeg: fix remove buffer operation for multi-core kyrie.wu
2025-01-09 14:06 ` [V1,00/12] Enable jpeg enc & dec multi-hardwares for MT8196 Nicolas Dufresne
2025-01-10  5:31   ` Kyrie Wu (吴晗)
2025-04-07  6:27   ` Kyrie Wu (吴晗)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=73144187-7f7f-4f68-aa5a-5fc019e646ec@collabora.com \
    --to=angelogioacchino.delregno@collabora.com \
    --cc=bin.liu@mediatek.com \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=hverkuil-cisco@xs4all.nl \
    --cc=krzk+dt@kernel.org \
    --cc=kyrie.wu@mediatek.com \
    --cc=kyrie.wu@mediatek.corp-partner.google.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=matthias.bgg@gmail.com \
    --cc=mchehab@kernel.org \
    --cc=robh+dt@kernel.org \
    --cc=robh@kernel.org \
    --cc=tzungbi@chromium.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox