Message-ID: <2c7ee056-546c-4891-abfc-c1d41e2c1632@amd.com>
Date: Tue, 28 Oct 2025 16:30:25 +0800
From: "Du, Bin" <bin.du@....com>
To: Krzysztof Kozlowski <krzk@...nel.org>, mchehab@...nel.org,
 hverkuil@...all.nl, laurent.pinchart+renesas@...asonboard.com,
 bryan.odonoghue@...aro.org, sakari.ailus@...ux.intel.com,
 prabhakar.mahadev-lad.rj@...renesas.com, linux-media@...r.kernel.org,
 linux-kernel@...r.kernel.org, sultan@...neltoast.com
Cc: pratap.nirujogi@....com, benjamin.chan@....com, king.li@....com,
 gjorgji.rosikopulos@....com, Phil.Jawich@....com, Dominic.Antony@....com,
 mario.limonciello@....com, richard.gong@....com, anson.tsao@....com,
 Svetoslav Stoilov <Svetoslav.Stoilov@....com>,
 Mario Limonciello <superm1@...nel.org>,
 Alexey Zagorodnikov <xglooom@...il.com>
Subject: Re: [PATCH v5 1/7] media: platform: amd: Introduce amd isp4 capture
 driver

Many thanks for the review, Krzysztof

On 10/24/2025 9:57 PM, Krzysztof Kozlowski wrote:
> On 24/10/2025 11:06, Bin Du wrote:
>> diff --git a/drivers/media/platform/amd/isp4/isp4.c b/drivers/media/platform/amd/isp4/isp4.c
>> new file mode 100644
>> index 000000000000..a3fc2462d70f
>> --- /dev/null
>> +++ b/drivers/media/platform/amd/isp4/isp4.c
>> @@ -0,0 +1,121 @@
>> +// SPDX-License-Identifier: GPL-2.0+
>> +/*
>> + * Copyright (C) 2025 Advanced Micro Devices, Inc.
>> + */
>> +
>> +#include <linux/pm_runtime.h>
>> +#include <linux/vmalloc.h>
>> +#include <media/v4l2-ioctl.h>
>> +
>> +#include "isp4.h"
>> +
>> +#define VIDEO_BUF_NUM 5
>> +
>> +#define ISP4_DRV_NAME "amd_isp_capture"
>> +
>> +const char *isp4_irq_name[] = {
> 
> Why isn't this static?
> 

Sure, will set it to static and check other similar variables.

>> +	"isp_irq_stream1",
>> +	"isp_irq_global"
>> +};
>> +
>> +/* interrupt num */
>> +static const u32 isp4_ringbuf_interrupt_num[] = {
>> +	0, /* ISP_4_1__SRCID__ISP_RINGBUFFER_WPT9 */
>> +	4, /* ISP_4_1__SRCID__ISP_RINGBUFFER_WPT12 */
>> +};
>> +
>> +static irqreturn_t isp4_irq_handler(int irq, void *arg)
>> +{
>> +	return IRQ_HANDLED;
>> +}
>> +
>> +static int isp4_capture_probe(struct platform_device *pdev)
>> +{
>> +	struct device *dev = &pdev->dev;
>> +	struct isp4_device *isp_dev;
>> +	int i, irq, ret;
>> +
>> +	isp_dev = devm_kzalloc(dev, sizeof(*isp_dev), GFP_KERNEL);
>> +	if (!isp_dev)
>> +		return -ENOMEM;
>> +
>> +	isp_dev->pdev = pdev;
>> +	dev->init_name = ISP4_DRV_NAME;
>> +
>> +	for (i = 0; i < ARRAY_SIZE(isp4_ringbuf_interrupt_num); i++) {
>> +		irq = platform_get_irq(pdev, isp4_ringbuf_interrupt_num[i]);
>> +		if (irq < 0)
>> +			return dev_err_probe(dev, irq,
>> +					     "fail to get irq %d\n",
>> +					     isp4_ringbuf_interrupt_num[i]);
>> +		ret = devm_request_irq(dev, irq, isp4_irq_handler, 0,
>> +				       isp4_irq_name[i], dev);
>> +		if (ret)
>> +			return dev_err_probe(dev, ret, "fail to req irq %d\n",
>> +					     irq);
>> +	}
>> +
>> +	/* Link the media device within the v4l2_device */
>> +	isp_dev->v4l2_dev.mdev = &isp_dev->mdev;
>> +
>> +	/* Initialize media device */
>> +	strscpy(isp_dev->mdev.model, "amd_isp41_mdev",
>> +		sizeof(isp_dev->mdev.model));
>> +	snprintf(isp_dev->mdev.bus_info, sizeof(isp_dev->mdev.bus_info),
>> +		 "platform:%s", ISP4_DRV_NAME);
>> +	isp_dev->mdev.dev = dev;
>> +	media_device_init(&isp_dev->mdev);
>> +
>> +	/* register v4l2 device */
>> +	snprintf(isp_dev->v4l2_dev.name, sizeof(isp_dev->v4l2_dev.name),
>> +		 "AMD-V4L2-ROOT");
>> +	ret = v4l2_device_register(dev, &isp_dev->v4l2_dev);
>> +	if (ret)
>> +		return dev_err_probe(dev, ret,
>> +				     "fail register v4l2 device\n");
>> +
>> +	ret = media_device_register(&isp_dev->mdev);
>> +	if (ret) {
>> +		dev_err(dev, "fail to register media device %d\n", ret);
>> +		goto err_unreg_v4l2;
>> +	}
>> +
>> +	platform_set_drvdata(pdev, isp_dev);
>> +
>> +	pm_runtime_set_suspended(dev);
>> +	pm_runtime_enable(dev);
>> +
>> +	return 0;
>> +
>> +err_unreg_v4l2:
>> +	v4l2_device_unregister(&isp_dev->v4l2_dev);
>> +
>> +	return dev_err_probe(dev, ret, "isp probe fail\n");
> 
> No, don't print generic error thus multiple errors. Drop this and keep
> informative dev_err_probe() in applicable places.
> 
> 

Very good suggestion, will implement it accordingly.

>> +}
>> +
>> +static void isp4_capture_remove(struct platform_device *pdev)
>> +{
>> +	struct isp4_device *isp_dev = platform_get_drvdata(pdev);
>> +
>> +	media_device_unregister(&isp_dev->mdev);
>> +	v4l2_device_unregister(&isp_dev->v4l2_dev);
>> +}
>> +
>> +static struct platform_driver isp4_capture_drv = {
>> +	.probe = isp4_capture_probe,
>> +	.remove = isp4_capture_remove,
>> +	.driver = {
>> +		.name = ISP4_DRV_NAME,
>> +		.owner = THIS_MODULE,
> 
> It is v5 but you still did not run standard tools. You try to upstream
> 12 year old code without cleaning it up.
> 

yes, no need to set .owner because the core will do it. Will clean it up.

> Please run standard kernel tools for static analysis, like coccinelle,
> smatch and sparse, and fix reported warnings. Also please check for
> warnings when building with W=1 for gcc and clang. Most of these
> commands (checks or W=1 build) can build specific targets, like some
> directory, to narrow the scope to only your code. The code here looks
> like it needs a fix. Feel free to get in touch if the warning is not clear.
> 

Really appreciate the instructions, will set up what you suggested and 
address the warnings. Suppose it can catch the issues you mentioned above.

>> +	}
>> +};
>> +
>> +module_platform_driver(isp4_capture_drv);
>> +
>> +MODULE_ALIAS("platform:" ISP4_DRV_NAME);
> 
> You should not need MODULE_ALIAS() in normal cases. If you need it,
> usually it means your device ID table is wrong (e.g. misses either
> entries or MODULE_DEVICE_TABLE()). MODULE_ALIAS() is not a substitute
> for incomplete ID table.
> 

Thanks for the clarification, yes, MODULE_ALIAS() does not apply in our 
case and should be removed.

> 
> Best regards,
> Krzysztof

-- 
Regards,
Bin

Powered by blists - more mailing lists