diff mbox series

[v3,9/9] drm/tegra: Support context isolation

Message ID 20220218113952.3077606-10-mperttunen@nvidia.com (mailing list archive)
State New, archived
Headers show
Series Host1x context isolation support | expand

Commit Message

Mikko Perttunen Feb. 18, 2022, 11:39 a.m. UTC
For engines that support context isolation, allocate a context when
opening a channel, and set up stream ID offset and context fields
when submitting a job.

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
---
 drivers/gpu/drm/tegra/drm.h    |  2 ++
 drivers/gpu/drm/tegra/submit.c | 13 ++++++++++++
 drivers/gpu/drm/tegra/uapi.c   | 36 ++++++++++++++++++++++++++++++++--
 3 files changed, 49 insertions(+), 2 deletions(-)

Comments

Dmitry Osipenko Feb. 19, 2022, 6:35 p.m. UTC | #1
18.02.2022 14:39, Mikko Perttunen пишет:
> +	if (context->memory_context && context->client->ops->get_streamid_offset) {
            ^^^
> +		int offset = context->client->ops->get_streamid_offset(context->client);
> +
> +		if (offset >= 0) {
> +			job->context = context->memory_context;
> +			job->engine_streamid_offset = offset;
> +			host1x_context_get(job->context);
> +		}

You should bump refcount unconditionally or you'll get refcnt underflow
on put, when offset < 0.

> +	}
> +
>  	/*
>  	 * job_data is now part of job reference counting, so don't release
>  	 * it from here.
> diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c
> index 9ab9179d2026..be33da54d12c 100644
> --- a/drivers/gpu/drm/tegra/uapi.c
> +++ b/drivers/gpu/drm/tegra/uapi.c
> @@ -33,6 +33,9 @@ static void tegra_drm_channel_context_close(struct tegra_drm_context *context)
>  	struct tegra_drm_mapping *mapping;
>  	unsigned long id;
>  
> +	if (context->memory_context)
> +		host1x_context_put(context->memory_context);

The "if (context->memory_context &&
context->client->ops->get_streamid_offset)" above doesn't match the "if
(context->memory_context)". You'll get refcount underflow.
Mikko Perttunen Feb. 21, 2022, 12:06 p.m. UTC | #2
On 2/19/22 20:35, Dmitry Osipenko wrote:
> 18.02.2022 14:39, Mikko Perttunen пишет:
>> +	if (context->memory_context && context->client->ops->get_streamid_offset) {
>              ^^^
>> +		int offset = context->client->ops->get_streamid_offset(context->client);
>> +
>> +		if (offset >= 0) {
>> +			job->context = context->memory_context;
>> +			job->engine_streamid_offset = offset;
>> +			host1x_context_get(job->context);
>> +		}
> 
> You should bump refcount unconditionally or you'll get refcnt underflow
> on put, when offset < 0.

This refcount is intended to be dropped from 'release_job', where it's 
dropped if job->context is set, which it is from this path.

> 
>> +	}
>> +
>>   	/*
>>   	 * job_data is now part of job reference counting, so don't release
>>   	 * it from here.
>> diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c
>> index 9ab9179d2026..be33da54d12c 100644
>> --- a/drivers/gpu/drm/tegra/uapi.c
>> +++ b/drivers/gpu/drm/tegra/uapi.c
>> @@ -33,6 +33,9 @@ static void tegra_drm_channel_context_close(struct tegra_drm_context *context)
>>   	struct tegra_drm_mapping *mapping;
>>   	unsigned long id;
>>   
>> +	if (context->memory_context)
>> +		host1x_context_put(context->memory_context);
> 
> The "if (context->memory_context &&
> context->client->ops->get_streamid_offset)" above doesn't match the "if
> (context->memory_context)". You'll get refcount underflow.

And this drop is for the refcount implicitly added when allocating the 
memory context through host1x_context_alloc; so these two places should 
be independent.

Please elaborate if I missed something.

Thanks,
Mikko
Dmitry Osipenko Feb. 21, 2022, 8:02 p.m. UTC | #3
21.02.2022 15:06, Mikko Perttunen пишет:
> On 2/19/22 20:35, Dmitry Osipenko wrote:
>> 18.02.2022 14:39, Mikko Perttunen пишет:
>>> +    if (context->memory_context &&
>>> context->client->ops->get_streamid_offset) {
>>              ^^^
>>> +        int offset =
>>> context->client->ops->get_streamid_offset(context->client);
>>> +
>>> +        if (offset >= 0) {
>>> +            job->context = context->memory_context;
>>> +            job->engine_streamid_offset = offset;
>>> +            host1x_context_get(job->context);
>>> +        }
>>
>> You should bump refcount unconditionally or you'll get refcnt underflow
>> on put, when offset < 0.
> 
> This refcount is intended to be dropped from 'release_job', where it's
> dropped if job->context is set, which it is from this path.
> 
>>
>>> +    }
>>> +
>>>       /*
>>>        * job_data is now part of job reference counting, so don't
>>> release
>>>        * it from here.
>>> diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c
>>> index 9ab9179d2026..be33da54d12c 100644
>>> --- a/drivers/gpu/drm/tegra/uapi.c
>>> +++ b/drivers/gpu/drm/tegra/uapi.c
>>> @@ -33,6 +33,9 @@ static void tegra_drm_channel_context_close(struct
>>> tegra_drm_context *context)
>>>       struct tegra_drm_mapping *mapping;
>>>       unsigned long id;
>>>   +    if (context->memory_context)
>>> +        host1x_context_put(context->memory_context);
>>
>> The "if (context->memory_context &&
>> context->client->ops->get_streamid_offset)" above doesn't match the "if
>> (context->memory_context)". You'll get refcount underflow.
> 
> And this drop is for the refcount implicitly added when allocating the
> memory context through host1x_context_alloc; so these two places should
> be independent.
> 
> Please elaborate if I missed something.

You named context as memory_context and then have
context=context->memory_context. Please try to improve the variable
names, like drm_ctx->host1x_ctx for example.

I'm also not a big fan of the "if (ref) put(ref)" pattern. Won't be
better to move all the "if (!NULL)" checks inside of get()/put() and
make the invocations unconditional?
Mikko Perttunen Feb. 22, 2022, 8:37 a.m. UTC | #4
On 2/21/22 22:02, Dmitry Osipenko wrote:
> 21.02.2022 15:06, Mikko Perttunen пишет:
>> On 2/19/22 20:35, Dmitry Osipenko wrote:
>>> 18.02.2022 14:39, Mikko Perttunen пишет:
>>>> +    if (context->memory_context &&
>>>> context->client->ops->get_streamid_offset) {
>>>               ^^^
>>>> +        int offset =
>>>> context->client->ops->get_streamid_offset(context->client);
>>>> +
>>>> +        if (offset >= 0) {
>>>> +            job->context = context->memory_context;
>>>> +            job->engine_streamid_offset = offset;
>>>> +            host1x_context_get(job->context);
>>>> +        }
>>>
>>> You should bump refcount unconditionally or you'll get refcnt underflow
>>> on put, when offset < 0.
>>
>> This refcount is intended to be dropped from 'release_job', where it's
>> dropped if job->context is set, which it is from this path.
>>
>>>
>>>> +    }
>>>> +
>>>>        /*
>>>>         * job_data is now part of job reference counting, so don't
>>>> release
>>>>         * it from here.
>>>> diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c
>>>> index 9ab9179d2026..be33da54d12c 100644
>>>> --- a/drivers/gpu/drm/tegra/uapi.c
>>>> +++ b/drivers/gpu/drm/tegra/uapi.c
>>>> @@ -33,6 +33,9 @@ static void tegra_drm_channel_context_close(struct
>>>> tegra_drm_context *context)
>>>>        struct tegra_drm_mapping *mapping;
>>>>        unsigned long id;
>>>>    +    if (context->memory_context)
>>>> +        host1x_context_put(context->memory_context);
>>>
>>> The "if (context->memory_context &&
>>> context->client->ops->get_streamid_offset)" above doesn't match the "if
>>> (context->memory_context)". You'll get refcount underflow.
>>
>> And this drop is for the refcount implicitly added when allocating the
>> memory context through host1x_context_alloc; so these two places should
>> be independent.
>>
>> Please elaborate if I missed something.
> 
> You named context as memory_context and then have
> context=context->memory_context. Please try to improve the variable
> names, like drm_ctx->host1x_ctx for example.
> 
> I'm also not a big fan of the "if (ref) put(ref)" pattern. Won't be
> better to move all the "if (!NULL)" checks inside of get()/put() and
> make the invocations unconditional?

I agree that the naming here is confusing with different kinds of 
contexts flying around, though I would prefer not to change the original 
'struct tegra_drm_context *context' since it's used all around the code. 
But I'll try to make it clearer.

Regarding moving NULL checks inside get/put, I personally dislike that 
pattern (also with e.g. kfree) since when reading the code, it makes it 
more difficult to see that the pointer can be NULL.

Mikko
diff mbox series

Patch

diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index fc0a19554eac..717e9f81ee1f 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -80,6 +80,7 @@  struct tegra_drm_context {
 
 	/* Only used by new UAPI. */
 	struct xarray mappings;
+	struct host1x_context *memory_context;
 };
 
 struct tegra_drm_client_ops {
@@ -91,6 +92,7 @@  struct tegra_drm_client_ops {
 	int (*submit)(struct tegra_drm_context *context,
 		      struct drm_tegra_submit *args, struct drm_device *drm,
 		      struct drm_file *file);
+	int (*get_streamid_offset)(struct tegra_drm_client *client);
 };
 
 int tegra_drm_submit(struct tegra_drm_context *context,
diff --git a/drivers/gpu/drm/tegra/submit.c b/drivers/gpu/drm/tegra/submit.c
index 6d6dd8c35475..8d74b82b83a5 100644
--- a/drivers/gpu/drm/tegra/submit.c
+++ b/drivers/gpu/drm/tegra/submit.c
@@ -498,6 +498,9 @@  static void release_job(struct host1x_job *job)
 	struct tegra_drm_submit_data *job_data = job->user_data;
 	u32 i;
 
+	if (job->context)
+		host1x_context_put(job->context);
+
 	for (i = 0; i < job_data->num_used_mappings; i++)
 		tegra_drm_mapping_put(job_data->used_mappings[i].mapping);
 
@@ -599,6 +602,16 @@  int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data,
 	job->release = release_job;
 	job->timeout = 10000;
 
+	if (context->memory_context && context->client->ops->get_streamid_offset) {
+		int offset = context->client->ops->get_streamid_offset(context->client);
+
+		if (offset >= 0) {
+			job->context = context->memory_context;
+			job->engine_streamid_offset = offset;
+			host1x_context_get(job->context);
+		}
+	}
+
 	/*
 	 * job_data is now part of job reference counting, so don't release
 	 * it from here.
diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c
index 9ab9179d2026..be33da54d12c 100644
--- a/drivers/gpu/drm/tegra/uapi.c
+++ b/drivers/gpu/drm/tegra/uapi.c
@@ -33,6 +33,9 @@  static void tegra_drm_channel_context_close(struct tegra_drm_context *context)
 	struct tegra_drm_mapping *mapping;
 	unsigned long id;
 
+	if (context->memory_context)
+		host1x_context_put(context->memory_context);
+
 	xa_for_each(&context->mappings, id, mapping)
 		tegra_drm_mapping_put(mapping);
 
@@ -72,6 +75,7 @@  static struct tegra_drm_client *tegra_drm_find_client(struct tegra_drm *tegra, u
 
 int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data, struct drm_file *file)
 {
+	struct host1x *host = tegra_drm_to_host1x(drm->dev_private);
 	struct tegra_drm_file *fpriv = file->driver_priv;
 	struct tegra_drm *tegra = drm->dev_private;
 	struct drm_tegra_channel_open *args = data;
@@ -102,10 +106,29 @@  int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data, struct drm_
 		}
 	}
 
+	/* Only allocate context if the engine supports context isolation. */
+	if (client->ops->get_streamid_offset &&
+	    client->ops->get_streamid_offset(client) >= 0) {
+		context->memory_context =
+			host1x_context_alloc(host, get_task_pid(current, PIDTYPE_TGID));
+		if (IS_ERR(context->memory_context)) {
+			if (PTR_ERR(context->memory_context) != -EOPNOTSUPP) {
+				err = PTR_ERR(context->memory_context);
+				goto put_channel;
+			} else {
+				/*
+				 * OK, HW does not support contexts or contexts
+				 * are disabled.
+				 */
+				context->memory_context = NULL;
+			}
+		}
+	}
+
 	err = xa_alloc(&fpriv->contexts, &args->context, context, XA_LIMIT(1, U32_MAX),
 		       GFP_KERNEL);
 	if (err < 0)
-		goto put_channel;
+		goto put_memctx;
 
 	context->client = client;
 	xa_init_flags(&context->mappings, XA_FLAGS_ALLOC1);
@@ -118,6 +141,9 @@  int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data, struct drm_
 
 	return 0;
 
+put_memctx:
+	if (context->memory_context)
+		host1x_context_put(context->memory_context);
 put_channel:
 	host1x_channel_put(context->channel);
 free:
@@ -156,6 +182,7 @@  int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data, struct drm_f
 	struct tegra_drm_mapping *mapping;
 	struct tegra_drm_context *context;
 	enum dma_data_direction direction;
+	struct device *mapping_dev;
 	int err = 0;
 
 	if (args->flags & ~DRM_TEGRA_CHANNEL_MAP_READ_WRITE)
@@ -177,6 +204,11 @@  int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data, struct drm_f
 
 	kref_init(&mapping->ref);
 
+	if (context->memory_context)
+		mapping_dev = &context->memory_context->dev;
+	else
+		mapping_dev = context->client->base.dev;
+
 	mapping->bo = tegra_gem_lookup(file, args->handle);
 	if (!mapping->bo) {
 		err = -EINVAL;
@@ -201,7 +233,7 @@  int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data, struct drm_f
 		goto put_gem;
 	}
 
-	mapping->map = host1x_bo_pin(context->client->base.dev, mapping->bo, direction, NULL);
+	mapping->map = host1x_bo_pin(mapping_dev, mapping->bo, direction, NULL);
 	if (IS_ERR(mapping->map)) {
 		err = PTR_ERR(mapping->map);
 		goto put_gem;