[v1] Fix hibmc driver bugs

[PATCH drm-dp 10/10] drm/hisilicon/hibmc: fix no showing problem with loading hibmc manually

Posted by Yongbang Shi 6 months, 3 weeks ago

From: Baihan Li <libaihan@huawei.com>

When using command rmmod and insmod, there is no showing in second time
insmoding. Because DP controller won't send HPD signals, if connection
doesn't change or controller isn't reset. So add reset before unreset
in hibmc_dp_hw_init().

Fixes: 94ee73ee3020 ("drm/hisilicon/hibmc: add dp hw moduel in hibmc driver")
Signed-off-by: Baihan Li <libaihan@huawei.com>
---
 drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c b/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
index 4f93d60b932b..e1b9589ce639 100644
--- a/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
+++ b/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
@@ -172,13 +172,15 @@ int hibmc_dp_hw_init(struct hibmc_dp *dp)
 	dp_dev->link.cap.lanes = 0x2;
 	dp_dev->link.cap.link_rate = DP_LINK_BW_8_1;
 
-	/* hdcp data */
-	writel(HIBMC_DP_HDCP, dp_dev->base + HIBMC_DP_HDCP_CFG);
 	/* int init */
 	writel(0, dp_dev->base + HIBMC_DP_INTR_ENABLE);
 	writel(HIBMC_DP_INT_RST, dp_dev->base + HIBMC_DP_INTR_ORIGINAL_STATUS);
 	/* rst */
+	writel(0, dp_dev->base + HIBMC_DP_DPTX_RST_CTRL);
+	usleep_range(30, 50);
 	writel(HIBMC_DP_DPTX_RST, dp_dev->base + HIBMC_DP_DPTX_RST_CTRL);
+	/* hdcp data */
+	writel(HIBMC_DP_HDCP, dp_dev->base + HIBMC_DP_HDCP_CFG);
 	/* clock enable */
 	writel(HIBMC_DP_CLK_EN, dp_dev->base + HIBMC_DP_DPTX_CLK_CTRL);
 
-- 
2.33.0

Re: [PATCH drm-dp 10/10] drm/hisilicon/hibmc: fix no showing problem with loading hibmc manually

Posted by Dmitry Baryshkov 6 months, 1 week ago

On Fri, May 30, 2025 at 05:54:32PM +0800, Yongbang Shi wrote:
> From: Baihan Li <libaihan@huawei.com>
> 
> When using command rmmod and insmod, there is no showing in second time
> insmoding. Because DP controller won't send HPD signals, if connection
> doesn't change or controller isn't reset. So add reset before unreset
> in hibmc_dp_hw_init().
> 
> Fixes: 94ee73ee3020 ("drm/hisilicon/hibmc: add dp hw moduel in hibmc driver")

Technically... yes and no. The function was written this way in that
commit, however HPD signals were not handled until the latter commit.

> Signed-off-by: Baihan Li <libaihan@huawei.com>
> ---
>  drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c b/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
> index 4f93d60b932b..e1b9589ce639 100644
> --- a/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
> +++ b/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
> @@ -172,13 +172,15 @@ int hibmc_dp_hw_init(struct hibmc_dp *dp)
>  	dp_dev->link.cap.lanes = 0x2;
>  	dp_dev->link.cap.link_rate = DP_LINK_BW_8_1;
>  
> -	/* hdcp data */
> -	writel(HIBMC_DP_HDCP, dp_dev->base + HIBMC_DP_HDCP_CFG);
>  	/* int init */
>  	writel(0, dp_dev->base + HIBMC_DP_INTR_ENABLE);
>  	writel(HIBMC_DP_INT_RST, dp_dev->base + HIBMC_DP_INTR_ORIGINAL_STATUS);
>  	/* rst */
> +	writel(0, dp_dev->base + HIBMC_DP_DPTX_RST_CTRL);
> +	usleep_range(30, 50);
>  	writel(HIBMC_DP_DPTX_RST, dp_dev->base + HIBMC_DP_DPTX_RST_CTRL);
> +	/* hdcp data */
> +	writel(HIBMC_DP_HDCP, dp_dev->base + HIBMC_DP_HDCP_CFG);

There are two independent changes here. Split them into two commits.

>  	/* clock enable */
>  	writel(HIBMC_DP_CLK_EN, dp_dev->base + HIBMC_DP_DPTX_CLK_CTRL);
>  
> -- 
> 2.33.0
> 

-- 
With best wishes
Dmitry

Re: [PATCH drm-dp 10/10] drm/hisilicon/hibmc: fix no showing problem with loading hibmc manually

Posted by Yongbang Shi 6 months, 1 week ago

> On Fri, May 30, 2025 at 05:54:32PM +0800, Yongbang Shi wrote:
>> From: Baihan Li <libaihan@huawei.com>
>>
>> When using command rmmod and insmod, there is no showing in second time
>> insmoding. Because DP controller won't send HPD signals, if connection
>> doesn't change or controller isn't reset. So add reset before unreset
>> in hibmc_dp_hw_init().
>>
>> Fixes: 94ee73ee3020 ("drm/hisilicon/hibmc: add dp hw moduel in hibmc driver")
> Technically... yes and no. The function was written this way in that
> commit, however HPD signals were not handled until the latter commit.

Alright, can I put 2 commits here?


>> Signed-off-by: Baihan Li <libaihan@huawei.com>
>> ---
>>   drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c | 6 ++++--
>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c b/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
>> index 4f93d60b932b..e1b9589ce639 100644
>> --- a/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
>> +++ b/drivers/gpu/drm/hisilicon/hibmc/dp/dp_hw.c
>> @@ -172,13 +172,15 @@ int hibmc_dp_hw_init(struct hibmc_dp *dp)
>>   	dp_dev->link.cap.lanes = 0x2;
>>   	dp_dev->link.cap.link_rate = DP_LINK_BW_8_1;
>>   
>> -	/* hdcp data */
>> -	writel(HIBMC_DP_HDCP, dp_dev->base + HIBMC_DP_HDCP_CFG);
>>   	/* int init */
>>   	writel(0, dp_dev->base + HIBMC_DP_INTR_ENABLE);
>>   	writel(HIBMC_DP_INT_RST, dp_dev->base + HIBMC_DP_INTR_ORIGINAL_STATUS);
>>   	/* rst */
>> +	writel(0, dp_dev->base + HIBMC_DP_DPTX_RST_CTRL);
>> +	usleep_range(30, 50);
>>   	writel(HIBMC_DP_DPTX_RST, dp_dev->base + HIBMC_DP_DPTX_RST_CTRL);
>> +	/* hdcp data */
>> +	writel(HIBMC_DP_HDCP, dp_dev->base + HIBMC_DP_HDCP_CFG);
> There are two independent changes here. Split them into two commits.

Okay!


>>   	/* clock enable */
>>   	writel(HIBMC_DP_CLK_EN, dp_dev->base + HIBMC_DP_DPTX_CLK_CTRL);
>>   
>> -- 
>> 2.33.0
>>