Re: [PATCH v2 5/5] venus: register separate driver for firmware device

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Tomasz,

On 06/04/2018 04:18 PM, Tomasz Figa wrote:
> Hi Vikash,
> 
> On Sat, Jun 2, 2018 at 5:27 AM Vikash Garodia <vgarodia@xxxxxxxxxxxxxx> wrote:
>>
>> A separate child device is added for video firmware.
>> This is needed to
>> [1] configure the firmware context bank with the desired SID.
>> [2] ensure that the iova for firmware region is from 0x0.
>>
>> Signed-off-by: Vikash Garodia <vgarodia@xxxxxxxxxxxxxx>
>> ---
>>  .../devicetree/bindings/media/qcom,venus.txt       |  8 +++-
>>  drivers/media/platform/qcom/venus/core.c           | 48 +++++++++++++++++++---
>>  drivers/media/platform/qcom/venus/firmware.c       | 20 ++++++++-
>>  drivers/media/platform/qcom/venus/firmware.h       |  2 +
>>  4 files changed, 71 insertions(+), 7 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/media/qcom,venus.txt b/Documentation/devicetree/bindings/media/qcom,venus.txt
>> index 00d0d1b..701cbe8 100644
>> --- a/Documentation/devicetree/bindings/media/qcom,venus.txt
>> +++ b/Documentation/devicetree/bindings/media/qcom,venus.txt
>> @@ -53,7 +53,7 @@
>>
>>  * Subnodes
>>  The Venus video-codec node must contain two subnodes representing
>> -video-decoder and video-encoder.
>> +video-decoder and video-encoder, one optional firmware subnode.
>>
>>  Every of video-encoder or video-decoder subnode should have:
>>
>> @@ -79,6 +79,8 @@ Every of video-encoder or video-decoder subnode should have:
>>                     power domain which is responsible for collapsing
>>                     and restoring power to the subcore.
>>
>> +The firmware sub node must contain the iommus specifiers for ARM9.
> 
> Please document the compatible string here as well.
> 
>> +
>>  * An Example
>>         video-codec@1d00000 {
>>                 compatible = "qcom,msm8916-venus";
>> @@ -105,4 +107,8 @@ Every of video-encoder or video-decoder subnode should have:
>>                         clock-names = "core";
>>                         power-domains = <&mmcc VENUS_CORE1_GDSC>;
>>                 };
>> +               venus-firmware {
>> +                       compatible = "qcom,venus-firmware-no-tz";
> 
> I don't think "-no-tz" should be mentioned here in DT, since it's a
> firmware/software detail.

I have to agree with Tomasz, non-tz or tz is a software detail and it
shouldn't be reflected in compatible string.

Also I'm not sure but what will happen if this video-firmware subnode is
not added, do you expect that backward compatibility is satisfied for
older venus versions?

> 
>> +                       iommus = <&apps_smmu 0x10b2 0x0>;
>> +               }
>>         };
>> diff --git a/drivers/media/platform/qcom/venus/core.c b/drivers/media/platform/qcom/venus/core.c
>> index 101612b..5cfb3c2 100644
>> --- a/drivers/media/platform/qcom/venus/core.c
>> +++ b/drivers/media/platform/qcom/venus/core.c
>> @@ -179,6 +179,19 @@ static u32 to_v4l2_codec_type(u32 codec)
>>         }
>>  }
>>
>> +static int store_firmware_dev(struct device *dev, void *data)
>> +{
>> +       struct venus_core *core = data;
>> +
>> +       if (!core)
>> +               return -EINVAL;
>> +
> 
> No need to check this AFAICT.>
>> +       if (of_device_is_compatible(dev->of_node, "qcom,venus-firmware-no-tz"))
>> +               core->fw.dev = dev;
>> +
>> +       return 0;
>> +}
>> +
>>  static int venus_enumerate_codecs(struct venus_core *core, u32 type)
>>  {
>>         const struct hfi_inst_ops dummy_ops = {};
>> @@ -279,6 +292,13 @@ static int venus_probe(struct platform_device *pdev)
>>         if (ret < 0)
>>                 goto err_runtime_disable;
>>
>> +       ret = of_platform_populate(dev->of_node, NULL, NULL, dev);
>> +       if (ret)
>> +               goto err_runtime_disable;
>> +
>> +       /* Attempt to store firmware device */
>> +       device_for_each_child(dev, core, store_firmware_dev);
>> +
>>         ret = venus_boot(core);
>>         if (ret)
>>                 goto err_runtime_disable;
>> @@ -303,10 +323,6 @@ static int venus_probe(struct platform_device *pdev)
>>         if (ret)
>>                 goto err_core_deinit;
>>
>> -       ret = of_platform_populate(dev->of_node, NULL, NULL, dev);
>> -       if (ret)
>> -               goto err_dev_unregister;
>> -
>>         ret = pm_runtime_put_sync(dev);
>>         if (ret)
>>                 goto err_dev_unregister;
>> @@ -483,7 +499,29 @@ static __maybe_unused int venus_runtime_resume(struct device *dev)
>>                 .pm = &venus_pm_ops,
>>         },
>>  };
>> -module_platform_driver(qcom_venus_driver);
>> +
>> +static int __init venus_init(void)
>> +{
>> +       int ret;
>> +
>> +       ret = platform_driver_register(&qcom_video_firmware_driver);
>> +       if (ret)
>> +               return ret;
> 
> Do we really need this firmware driver? As far as I can see, the
> approach used here should work even without any driver bound to the
> firmware device.

We need device/driver bind because we need to call dma_configure() which
internally doing iommus sID parsing.

-- 
regards,
Stan
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux