mbox series

[v10,00/11] Add support for MT8195 SCP 2nd core

Message ID 20230426091211.21557-1-tinghan.shen@mediatek.com
Headers show
Series Add support for MT8195 SCP 2nd core | expand

Message

Tinghan Shen April 26, 2023, 9:12 a.m. UTC
The mediatek remoteproc driver currently only allows bringing up a 
single core SCP, e.g. MT8183. It also only bringing up the 1st 
core in SoCs with a dual-core SCP, e.g. MT8195. This series support 
to bring-up the 2nd core of the dual-core SCP.

v9 -> v10:
1. move the global mtk_scp list into the platform device driver data structure
2. remove an unnecessary if() condition

v8 -> v9:
1. initialize l1tcm_size/l1tcm_phys at patchset 05/11 
2. rewrite patchset 06/11 to unify the flow and remove hacks

v7 -> v8:
1. update the node name of mt8192 asurada SCP rpmsg subnode
2. squash register definitions into driver patches
3. initialize local variables on the declaration at patch v8 06/11 

v6 -> v7:
1. merge the mtk_scp_cluster struct into the mtk_scp structure
   at the "Probe multi-core SCP" patch

v5 -> v6:
1. move the mtk_scp_of_regs structure from mtk_common.h to mtk_scp.c
2. rename the SCP core 0 label from 'scp' to 'scp_c0'

v4 -> v5:
1. move resource release actions to the platform driver remove operation 
2. fix dual-core watchdog handling

v3 -> v4:
1. change the representation of dual-core SCP in dts file and update SCP yaml
2. rewrite SCP driver to reflect the change of dts node
3. drop 'remove redundant call of rproc_boot for SCP' in v3 for further investigation

v2 -> v3:
1. change the representation of dual-core SCP in dts file and update SCP yaml
2. rewrite SCP driver to reflect the change of dts node
3. add SCP core 1 node to mt8195.dtsi
4. remove redundant call of rproc_boot for SCP
5. refine IPI error message

v1 -> v2:
1. update dt-binding property description
2. remove kconfig for scp dual driver
3. merge mtk_scp_dual.c and mtk_scp_subdev.c to mtk_scp.c

Tinghan Shen (11):
  dt-bindings: remoteproc: mediatek: Improve the rpmsg subnode
    definition
  arm64: dts: mediatek: Update the node name of SCP rpmsg subnode
  dt-bindings: remoteproc: mediatek: Support MT8195 dual-core SCP
  remoteproc: mediatek: Add MT8195 SCP core 1 operations
  remoteproc: mediatek: Extract remoteproc initialization flow
  remoteproc: mediatek: Probe multi-core SCP
  remoteproc: mediatek: Control SCP core 1 by rproc subdevice
  remoteproc: mediatek: Setup MT8195 SCP core 1 SRAM offset
  remoteproc: mediatek: Handle MT8195 SCP core 1 watchdog timeout
  remoteproc: mediatek: Refine ipi handler error message
  arm64: dts: mediatek: mt8195: Add SCP 2nd core

 .../bindings/remoteproc/mtk,scp.yaml          | 176 +++++++-
 .../arm64/boot/dts/mediatek/mt8183-kukui.dtsi |   2 +-
 .../boot/dts/mediatek/mt8192-asurada.dtsi     |   2 +-
 .../boot/dts/mediatek/mt8195-cherry.dtsi      |   6 +-
 arch/arm64/boot/dts/mediatek/mt8195.dtsi      |  32 +-
 drivers/remoteproc/mtk_common.h               |  33 ++
 drivers/remoteproc/mtk_scp.c                  | 415 ++++++++++++++++--
 7 files changed, 604 insertions(+), 62 deletions(-)

Comments

Mathieu Poirier May 1, 2023, 10:04 p.m. UTC | #1
Hi Tinghan,

On Wed, Apr 26, 2023 at 05:12:05PM +0800, Tinghan Shen wrote:
> This is the preparation for probing multi-core SCP. The remoteproc
> initialization flow is similar on cores and is reused to avoid
> redundant code.
> 
> The registers of config and l1tcm are shared for multi-core
> SCP. Reuse the mapped addresses for all cores.
> 
> Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> ---
>  drivers/remoteproc/mtk_scp.c | 69 ++++++++++++++++++++++++++----------
>  1 file changed, 50 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
> index 2bf66b1a8d80..5e4982f4d5dc 100644
> --- a/drivers/remoteproc/mtk_scp.c
> +++ b/drivers/remoteproc/mtk_scp.c
> @@ -23,6 +23,13 @@
>  #define MAX_CODE_SIZE 0x500000
>  #define SECTION_NAME_IPI_BUFFER ".ipi_buffer"
>  
> +struct mtk_scp_of_cluster {
> +	void __iomem *reg_base;
> +	void __iomem *l1tcm_base;
> +	size_t l1tcm_size;
> +	phys_addr_t l1tcm_phys;
> +};

This is a good start.

> +
>  /**
>   * scp_get() - get a reference to SCP.
>   *
> @@ -855,10 +862,11 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
>  	}
>  }
>  
> -static int scp_probe(struct platform_device *pdev)
> +static int scp_rproc_init(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct mtk_scp_of_cluster *of_cluster = platform_get_drvdata(pdev);

1) Because of the work done in the next patch, I think a "struct
mtk_scp_of_cluster *" should be given as a parameter to scp_rproc_init().

2) I would rename of_cluster to scp_cluster.

>  	struct mtk_scp *scp;
>  	struct rproc *rproc;
>  	struct resource *res;
> @@ -879,6 +887,11 @@ static int scp_probe(struct platform_device *pdev)
>  	scp->data = of_device_get_match_data(dev);
>  	platform_set_drvdata(pdev, scp);
>  
> +	scp->reg_base = of_cluster->reg_base;
> +	scp->l1tcm_base = of_cluster->l1tcm_base;
> +	scp->l1tcm_size = of_cluster->l1tcm_size;
> +	scp->l1tcm_phys = of_cluster->l1tcm_phys;
> +
>  	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram");
>  	scp->sram_base = devm_ioremap_resource(dev, res);
>  	if (IS_ERR(scp->sram_base))
> @@ -888,24 +901,6 @@ static int scp_probe(struct platform_device *pdev)
>  	scp->sram_size = resource_size(res);
>  	scp->sram_phys = res->start;
>  
> -	/* l1tcm is an optional memory region */
> -	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> -	scp->l1tcm_base = devm_ioremap_resource(dev, res);
> -	if (IS_ERR(scp->l1tcm_base)) {
> -		ret = PTR_ERR(scp->l1tcm_base);
> -		if (ret != -EINVAL) {
> -			return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> -		}
> -	} else {
> -		scp->l1tcm_size = resource_size(res);
> -		scp->l1tcm_phys = res->start;
> -	}
> -
> -	scp->reg_base = devm_platform_ioremap_resource_byname(pdev, "cfg");
> -	if (IS_ERR(scp->reg_base))
> -		return dev_err_probe(dev, PTR_ERR(scp->reg_base),
> -				     "Failed to parse and map cfg memory\n");
> -
>  	ret = scp->data->scp_clk_get(scp);
>  	if (ret)
>  		return ret;
> @@ -957,6 +952,42 @@ static int scp_probe(struct platform_device *pdev)
>  	return ret;
>  }
>  
> +static int scp_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct mtk_scp_of_cluster *of_cluster;
> +	struct resource *res;
> +	int ret;
> +
> +	of_cluster = devm_kzalloc(&pdev->dev, sizeof(*of_cluster), GFP_KERNEL);
> +	if (!of_cluster)
> +		return -ENOMEM;
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg");
> +	of_cluster->reg_base = devm_ioremap_resource(dev, res);
> +	if (IS_ERR(of_cluster->reg_base))
> +		return dev_err_probe(dev, PTR_ERR(of_cluster->reg_base),
> +				     "Failed to parse and map cfg memory\n");
> +
> +	/* l1tcm is an optional memory region */
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> +	of_cluster->l1tcm_base = devm_ioremap_resource(dev, res);
> +	if (IS_ERR(of_cluster->l1tcm_base)) {
> +		ret = PTR_ERR(of_cluster->l1tcm_base);
> +		if (ret != -EINVAL)
> +			return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> +
> +		of_cluster->l1tcm_base = NULL;
> +	} else {
> +		of_cluster->l1tcm_size = resource_size(res);
> +		of_cluster->l1tcm_phys = res->start;
> +	}
> +
> +	platform_set_drvdata(pdev, of_cluster);
> +

In scp_rproc_init() pdev->dev->driver_data is set to a *scp, which
defeats the purpose of setting here.  Once the driver data for @pdev is set, it
doesn't change after that.

> +	return scp_rproc_init(pdev);

Function scp_rproc_init() needs to return an *scp and that scp needs to be added
to the list of SCPs in scp_cluster.

> +}
> +
>  static int scp_remove(struct platform_device *pdev)
>  {
>  	struct mtk_scp *scp = platform_get_drvdata(pdev);

This should be:

        struct mtk_scp_of_cluster *scp_cluster = platform_get_drvdata(pdev);

And then iterate over the list of SCPs to remove each scp.  That way things work
the same way regardless of the amount of scp in the system.


> -- 
> 2.18.0
>