[PATCH v2] gpu: host1x: Avoid trying to use GART on Tegra20

[PATCH v2] gpu: host1x: Avoid trying to use GART on Tegra20

[Robin Murphy]

Since commit c7e3ca515e78 ("iommu/tegra: gart: Do not register with
bus") quite some time ago, the GART driver has effectively disabled
itself to avoid issues with the GPU driver expecting it to work in ways
that it doesn't. As of commit 57365a04c921 ("iommu: Move bus setup to
IOMMU device registration") that bodge no longer works, but really the
GPU driver should be responsible for its own behaviour anyway. Make the
workaround explicit.

Reported-by: Jon Hunter <jonathanh@nvidia.com>
Suggested-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---

v2: Cover DRM instance too, move into *_wants_iommu() for consistency

 drivers/gpu/drm/tegra/drm.c | 4 ++++
 drivers/gpu/host1x/dev.c    | 4 ++++
 2 files changed, 8 insertions(+)

diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index 6748ec1e0005..a1f909dac89a 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -1093,6 +1093,10 @@ static bool host1x_drm_wants_iommu(struct host1x_device *dev)
 	struct host1x *host1x = dev_get_drvdata(dev->dev.parent);
 	struct iommu_domain *domain;
 
+	/* Our IOMMU usage policy doesn't currently play well with GART */
+	if (of_machine_is_compatible("nvidia,tegra20"))
+		return false;
+
 	/*
 	 * If the Tegra DRM clients are backed by an IOMMU, push buffers are
 	 * likely to be allocated beyond the 32-bit boundary if sufficient
diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
index 0cd3f97e7e49..f60ea24db0ec 100644
--- a/drivers/gpu/host1x/dev.c
+++ b/drivers/gpu/host1x/dev.c
@@ -292,6 +292,10 @@ static void host1x_setup_virtualization_tables(struct host1x *host)
 
 static bool host1x_wants_iommu(struct host1x *host1x)
 {
+	/* Our IOMMU usage policy doesn't currently play well with GART */
+	if (of_machine_is_compatible("nvidia,tegra20"))
+		return false;
+
 	/*
 	 * If we support addressing a maximum of 32 bits of physical memory
 	 * and if the host1x firewall is enabled, there's no need to enable
-- 
2.36.1.dirty

Re: [PATCH v2] gpu: host1x: Avoid trying to use GART on Tegra20

[Jon Hunter]

On 20/10/2022 15:23, Robin Murphy wrote:
> Since commit c7e3ca515e78 ("iommu/tegra: gart: Do not register with
> bus") quite some time ago, the GART driver has effectively disabled
> itself to avoid issues with the GPU driver expecting it to work in ways
> that it doesn't. As of commit 57365a04c921 ("iommu: Move bus setup to
> IOMMU device registration") that bodge no longer works, but really the
> GPU driver should be responsible for its own behaviour anyway. Make the
> workaround explicit.
> 
> Reported-by: Jon Hunter <jonathanh@nvidia.com>
> Suggested-by: Dmitry Osipenko <digetx@gmail.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> 
> v2: Cover DRM instance too, move into *_wants_iommu() for consistency
> 
>   drivers/gpu/drm/tegra/drm.c | 4 ++++
>   drivers/gpu/host1x/dev.c    | 4 ++++
>   2 files changed, 8 insertions(+)
> 
> diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
> index 6748ec1e0005..a1f909dac89a 100644
> --- a/drivers/gpu/drm/tegra/drm.c
> +++ b/drivers/gpu/drm/tegra/drm.c
> @@ -1093,6 +1093,10 @@ static bool host1x_drm_wants_iommu(struct host1x_device *dev)
>   	struct host1x *host1x = dev_get_drvdata(dev->dev.parent);
>   	struct iommu_domain *domain;
>   
> +	/* Our IOMMU usage policy doesn't currently play well with GART */
> +	if (of_machine_is_compatible("nvidia,tegra20"))
> +		return false;
> +
>   	/*
>   	 * If the Tegra DRM clients are backed by an IOMMU, push buffers are
>   	 * likely to be allocated beyond the 32-bit boundary if sufficient
> diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
> index 0cd3f97e7e49..f60ea24db0ec 100644
> --- a/drivers/gpu/host1x/dev.c
> +++ b/drivers/gpu/host1x/dev.c
> @@ -292,6 +292,10 @@ static void host1x_setup_virtualization_tables(struct host1x *host)
>   
>   static bool host1x_wants_iommu(struct host1x *host1x)
>   {
> +	/* Our IOMMU usage policy doesn't currently play well with GART */
> +	if (of_machine_is_compatible("nvidia,tegra20"))
> +		return false;
> +
>   	/*
>   	 * If we support addressing a maximum of 32 bits of physical memory
>   	 * and if the host1x firewall is enabled, there's no need to enable


Thanks! Works for me.

Tested-by: Jon Hunter <jonathanh@nvidia.com>

Cheers
Jon

-- 
nvpublic

Re: [PATCH v2] gpu: host1x: Avoid trying to use GART on Tegra20

[Jon Hunter]

Thierry,

On 21/10/2022 08:41, Jon Hunter wrote:
> 
> On 20/10/2022 15:23, Robin Murphy wrote:
>> Since commit c7e3ca515e78 ("iommu/tegra: gart: Do not register with
>> bus") quite some time ago, the GART driver has effectively disabled
>> itself to avoid issues with the GPU driver expecting it to work in ways
>> that it doesn't. As of commit 57365a04c921 ("iommu: Move bus setup to
>> IOMMU device registration") that bodge no longer works, but really the
>> GPU driver should be responsible for its own behaviour anyway. Make the
>> workaround explicit.
>>
>> Reported-by: Jon Hunter <jonathanh@nvidia.com>
>> Suggested-by: Dmitry Osipenko <digetx@gmail.com>
>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>> ---
>>
>> v2: Cover DRM instance too, move into *_wants_iommu() for consistency
>>
>>   drivers/gpu/drm/tegra/drm.c | 4 ++++
>>   drivers/gpu/host1x/dev.c    | 4 ++++
>>   2 files changed, 8 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
>> index 6748ec1e0005..a1f909dac89a 100644
>> --- a/drivers/gpu/drm/tegra/drm.c
>> +++ b/drivers/gpu/drm/tegra/drm.c
>> @@ -1093,6 +1093,10 @@ static bool host1x_drm_wants_iommu(struct 
>> host1x_device *dev)
>>       struct host1x *host1x = dev_get_drvdata(dev->dev.parent);
>>       struct iommu_domain *domain;
>> +    /* Our IOMMU usage policy doesn't currently play well with GART */
>> +    if (of_machine_is_compatible("nvidia,tegra20"))
>> +        return false;
>> +
>>       /*
>>        * If the Tegra DRM clients are backed by an IOMMU, push buffers 
>> are
>>        * likely to be allocated beyond the 32-bit boundary if sufficient
>> diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
>> index 0cd3f97e7e49..f60ea24db0ec 100644
>> --- a/drivers/gpu/host1x/dev.c
>> +++ b/drivers/gpu/host1x/dev.c
>> @@ -292,6 +292,10 @@ static void 
>> host1x_setup_virtualization_tables(struct host1x *host)
>>   static bool host1x_wants_iommu(struct host1x *host1x)
>>   {
>> +    /* Our IOMMU usage policy doesn't currently play well with GART */
>> +    if (of_machine_is_compatible("nvidia,tegra20"))
>> +        return false;
>> +
>>       /*
>>        * If we support addressing a maximum of 32 bits of physical memory
>>        * and if the host1x firewall is enabled, there's no need to enable
> 
> 
> Thanks! Works for me.
> 
> Tested-by: Jon Hunter <jonathanh@nvidia.com>

We need to pick this fix up for v6.1.

Thanks
Jon

-- 
nvpublic

Re: [PATCH v2] gpu: host1x: Avoid trying to use GART on Tegra20

[Thierry Reding]

From: Thierry Reding <treding@nvidia.com>

On Thu, 20 Oct 2022 15:23:40 +0100, Robin Murphy wrote:
> Since commit c7e3ca515e78 ("iommu/tegra: gart: Do not register with
> bus") quite some time ago, the GART driver has effectively disabled
> itself to avoid issues with the GPU driver expecting it to work in ways
> that it doesn't. As of commit 57365a04c921 ("iommu: Move bus setup to
> IOMMU device registration") that bodge no longer works, but really the
> GPU driver should be responsible for its own behaviour anyway. Make the
> workaround explicit.
> 
> [...]

Applied, thanks!

[1/1] gpu: host1x: Avoid trying to use GART on Tegra20
      commit: c8899b7129927d96d2a948a0d969ab13b9f1dad1

Best regards,
-- 
Thierry Reding <treding@nvidia.com>