Hi Liviu
On 3/25/2022 3:19 AM, Liviu Dudau wrote:
On Thu, Mar 24, 2022 at 09:36:50AM -0700, Abhinav Kumar wrote:
Hi Liviu
Hello,
Thanks for the response.
On 3/24/2022 3:12 AM, Liviu Dudau wrote:
On Wed, Mar 23, 2022 at 11:28:56AM -0700, Abhinav Kumar wrote:
Hi Liviu
Hello,
Thanks for the review.
On 3/23/2022 9:46 AM, Liviu Dudau wrote:
On Mon, Mar 21, 2022 at 04:56:43PM -0700, Abhinav Kumar wrote:
For vendors drivers which pass an already allocated and
initialized encoder especially for cases where the encoder
hardware is shared OR the writeback encoder shares the resources
with the rest of the display pipeline introduce a new API,
drm_writeback_connector_init_with_encoder() which expects
an initialized encoder as a parameter and only sets up the
writeback connector.
changes in v5:
- reorder this change to come before in the series
to avoid incorrect functionality in subsequent changes
- continue using struct drm_encoder instead of
struct drm_encoder * and switch it in next change
Signed-off-by: Abhinav Kumar <quic_abhin...@quicinc.com>
Hi Abhinav,
---
drivers/gpu/drm/drm_writeback.c | 143
++++++++++++++++++++++++++++------------
include/drm/drm_writeback.h | 5 ++
2 files changed, 106 insertions(+), 42 deletions(-)
diff --git a/drivers/gpu/drm/drm_writeback.c b/drivers/gpu/drm/drm_writeback.c
index dc2ef12..abe78b9 100644
--- a/drivers/gpu/drm/drm_writeback.c
+++ b/drivers/gpu/drm/drm_writeback.c
@@ -149,37 +149,15 @@ static const struct drm_encoder_funcs
drm_writeback_encoder_funcs = {
.destroy = drm_encoder_cleanup,
};
-/**
- * drm_writeback_connector_init - Initialize a writeback connector and its
properties
- * @dev: DRM device
- * @wb_connector: Writeback connector to initialize
- * @con_funcs: Connector funcs vtable
- * @enc_helper_funcs: Encoder helper funcs vtable to be used by the internal
encoder
- * @formats: Array of supported pixel formats for the writeback engine
- * @n_formats: Length of the formats array
- * @possible_crtcs: possible crtcs for the internal writeback encoder
- *
- * This function creates the writeback-connector-specific properties if they
- * have not been already created, initializes the connector as
- * type DRM_MODE_CONNECTOR_WRITEBACK, and correctly initializes the property
- * values. It will also create an internal encoder associated with the
- * drm_writeback_connector and set it to use the @enc_helper_funcs vtable for
- * the encoder helper.
- *
- * Drivers should always use this function instead of drm_connector_init() to
- * set up writeback connectors.
- *
- * Returns: 0 on success, or a negative error code
- */
-int drm_writeback_connector_init(struct drm_device *dev,
- struct drm_writeback_connector *wb_connector,
- const struct drm_connector_funcs *con_funcs,
- const struct drm_encoder_helper_funcs
*enc_helper_funcs,
- const u32 *formats, int n_formats, uint32_t
possible_crtcs)
+static int drm_writeback_connector_setup(struct drm_device *dev,
+ struct drm_writeback_connector *wb_connector,
+ const struct drm_connector_funcs *con_funcs, const u32 *formats,
+ int n_formats)
{
struct drm_property_blob *blob;
- struct drm_connector *connector = &wb_connector->base;
struct drm_mode_config *config = &dev->mode_config;
+ struct drm_connector *connector = &wb_connector->base;
+
Point of this reordering the statements is...?
This diff can be avoided. There was no reason to reorder this. I will remove
this re-order.
int ret = create_writeback_properties(dev);
if (ret != 0)
@@ -187,18 +165,10 @@ int drm_writeback_connector_init(struct drm_device *dev,
blob = drm_property_create_blob(dev, n_formats * sizeof(*formats),
formats);
- if (IS_ERR(blob))
- return PTR_ERR(blob);
-
- drm_encoder_helper_add(&wb_connector->encoder, enc_helper_funcs);
-
- wb_connector->encoder.possible_crtcs = possible_crtcs;
-
- ret = drm_encoder_init(dev, &wb_connector->encoder,
- &drm_writeback_encoder_funcs,
- DRM_MODE_ENCODER_VIRTUAL, NULL);
- if (ret)
- goto fail;
+ if (IS_ERR(blob)) {
+ ret = PTR_ERR(blob);
+ return ret;
+ }
I don't see why you have changed the earlier code to store the result of
PTR_ERR into
ret other than to help your debugging. I suggest that you keep the existing
code that
returns PTR_ERR(blob) directly and you will have a nicer diff stat as well.
Sure, i can fix this for a smaller diff stat.
connector->interlace_allowed = 0;
@@ -237,13 +207,102 @@ int drm_writeback_connector_init(struct drm_device *dev,
attach_fail:
drm_connector_cleanup(connector);
connector_fail:
- drm_encoder_cleanup(&wb_connector->encoder);
-fail:
drm_property_blob_put(blob);
return ret;
}
+
+/**
+ * drm_writeback_connector_init - Initialize a writeback connector and its
properties
+ * @dev: DRM device
+ * @wb_connector: Writeback connector to initialize
+ * @con_funcs: Connector funcs vtable
+ * @enc_helper_funcs: Encoder helper funcs vtable to be used by the internal
encoder
+ * @formats: Array of supported pixel formats for the writeback engine
+ * @n_formats: Length of the formats array
+ * @possible_crtcs: possible crtcs for the internal writeback encoder
+ *
+ * This function creates the writeback-connector-specific properties if they
+ * have not been already created, initializes the connector as
+ * type DRM_MODE_CONNECTOR_WRITEBACK, and correctly initializes the property
+ * values. It will also create an internal encoder associated with the
+ * drm_writeback_connector and set it to use the @enc_helper_funcs vtable for
+ * the encoder helper.
+ *
+ * Drivers should always use this function instead of drm_connector_init() to
+ * set up writeback connectors.
+ *
+ * Returns: 0 on success, or a negative error code
+ */
+int drm_writeback_connector_init(struct drm_device *dev,
+ struct drm_writeback_connector *wb_connector,
+ const struct drm_connector_funcs *con_funcs,
+ const struct drm_encoder_helper_funcs *enc_helper_funcs,
+ const u32 *formats, int n_formats, uint32_t possible_crtcs)
+{
+ int ret = 0;
+
+ drm_encoder_helper_add(&wb_connector->encoder, enc_helper_funcs);
+
+ wb_connector->encoder.possible_crtcs = possible_crtcs;
+
+ ret = drm_encoder_init(dev, &wb_connector->encoder,
+ &drm_writeback_encoder_funcs,
+ DRM_MODE_ENCODER_VIRTUAL, NULL);
+ if (ret)
+ return ret;
+
+ ret = drm_writeback_connector_setup(dev, wb_connector, con_funcs,
formats,
+ n_formats);
+
+ if (ret)
+ drm_encoder_cleanup(&wb_connector->encoder);
+
+ return ret;
+}
EXPORT_SYMBOL(drm_writeback_connector_init);
+/**
+ * drm_writeback_connector_init_with_encoder - Initialize a writeback
connector and its properties
+ * using the encoder which already assigned and initialized
+ *
+ * @dev: DRM device
+ * @wb_connector: Writeback connector to initialize
+ * @con_funcs: Connector funcs vtable
+ * @enc_helper_funcs: Encoder helper funcs vtable to be used by the internal
encoder
+ * @formats: Array of supported pixel formats for the writeback engine
+ * @n_formats: Length of the formats array
+ *
+ * This function creates the writeback-connector-specific properties if they
+ * have not been already created, initializes the connector as
+ * type DRM_MODE_CONNECTOR_WRITEBACK, and correctly initializes the property
+ * values.
+ *
+ * This function assumes that the drm_writebac_connector's encoder has already
been
spelling: writeback
Ack. will fix this.
+ * created and initialized before invoking this function.
+ *
+ * In addition, this function also assumes that callers of this API will manage
+ * assigning the encoder helper functions, possible_crtcs and any other encoder
+ * specific operation which is otherwise handled by
drm_writeback_connector_init().
+ *
+ * Drivers should always use this function instead of drm_connector_init() to
+ * set up writeback connectors.
.... if they want to manage themselves the lifetime of the associated encoder.
We're not trying to replace drm_writeback_connector_init() function here, only
to
provide an alternative function to call for special cases.
Yes, we are trying to provide an alternative function call for special case
where the encoder is a shared encoder and/or where the resources of the
writeback encoder are shared with other hardware blocks.
I can add that comment to this function's doc.
+ *
+ * Returns: 0 on success, or a negative error code
+ */
+int drm_writeback_connector_init_with_encoder(struct drm_device *dev,
+ struct drm_writeback_connector *wb_connector,
+ const struct drm_connector_funcs *con_funcs, const u32 *formats,
+ int n_formats)
Where is the encoder parameter? Isn't that the reason why the function is called
`_with_encoder`?
The encoder parameter is skipped here because this function assumes that
wb_connector->encoder has already been initialized, setup with functions and
its possible_crts have already been set prior to calling this function like
the vc4 example shows. So this function doesnt need an explicit encoder
parameter. Let me know if thats a concern.
I think there might have been too many version of these patchsets and things are
starting to be confusing. Please revisit the series without rushing and come up
with
a plan of action. My understanding of watching this series has been that you're
trying to come up with a function that does *connector* initialisation but
skips the
*encoder* initialisation because it might have been already done by the driver.
The
idea will be then to have a function
`drm_writeback_connector_init_with_encoder()`
that does *all* the work that `drm_writeback_connector_init()` does at the
moment minus
the encoder initialisation part. Then `drm_writeback_connector_init()` only
initialises the internal encoder and calls
`drm_writeback_connector_init_with_encoder()`.
There is no need to have the `drm_writeback_connector_setup()` function at all.
Best regards,
Liviu
I agree there have been a 4 revisions prior to this because of the way this
affects the existing writeback drivers. So initial revision was a bit
intrusive into other drivers (which was just mostly a take over from the
previous patchset posted by the Co-developer ) and since rev3 we have
decided to have a separate API drm_writeback_connector_init_with_encoder()
so that existing clients are unaffected and it works seamlessly under the
hood.
Only clients which already embed an encoder (vc4) and the new ones which
have special encoder requirements like the MSM driver for which I am
preparing these changes for will use the new API.
Apologies for the revisions, but thanks to some great feedback from you and
laurent its shaping up nicely and reaching its conclusion I feel.
I think it is only natural that there will be some iterations. If I remember
correctly, the initial writeback series has something like 13 revisions :)
So i think this revision is pretty close to being clean thanks to the
feedback you gave on rev4. Between rev4 and this one I didnt drastically
change the design other than re-ordering the changes to avoid the
intermediate patches having an incorrect state for the vc4 encoder. So all
your comments related to the encoder_cleanup() and vc4's encoder not getting
initialized would have gotten addressed but overall concept remained same.
You are right, we are trying to come up with a function which does connector
initialization but skips the encoder part because that has already been done
in the client side of this API ( hence _with_encoder() name ). The
"_with_encoder" indicates that the caller already has an encoder for the
writeback connector which is being passed so there is no need to pass the
encoder again.
I thought this addresses all the concerns posted in the previous series.
So are you suggesting something like below?
1) rename drm_writeback_connector_setup() to
drm_writeback_connector_init_with_encoder()
(essentially thats what will end up happening )
Correct. Without the pointless reordering of code it should be about 3 lines of
code
that get removed (the calls to drm_encoder_helper_add() and drm_encoder_init()).
But isnt thats how it already looks.
int drm_writeback_connector_init(struct drm_device *dev,
struct drm_writeback_connector *wb_connector,
const struct drm_connector_funcs *con_funcs,
const struct drm_encoder_helper_funcs *enc_helper_funcs,
const u32 *formats, int n_formats, uint32_t possible_crtcs)
{
int ret = 0;
wb_connector->encoder = &wb_connector->internal_encoder;
drm_encoder_helper_add(wb_connector->encoder, enc_helper_funcs);
wb_connector->encoder->possible_crtcs = possible_crtcs;
ret = drm_encoder_init(dev, wb_connector->encoder,
&drm_writeback_encoder_funcs,
DRM_MODE_ENCODER_VIRTUAL, NULL);
if (ret)
return ret;
ret = drm_writeback_connector_setup(dev, wb_connector, con_funcs,
formats,
n_formats);
So the only change you are requesting is that, instead of having a new
drm_writeback_connector_setup(), just call
drm_writeback_connector_init_with_encoder().
It will essentially look like
int drm_writeback_connector_init(struct drm_device *dev,
struct drm_writeback_connector *wb_connector,
const struct drm_connector_funcs *con_funcs,
const struct drm_encoder_helper_funcs *enc_helper_funcs,
const u32 *formats, int n_formats, uint32_t possible_crtcs)
{
int ret = 0;
wb_connector->encoder = &wb_connector->internal_encoder;
(1)
drm_encoder_helper_add(wb_connector->encoder, enc_helper_funcs);
wb_connector->encoder->possible_crtcs = possible_crtcs;
ret = drm_encoder_init(dev, wb_connector->encoder,
&drm_writeback_encoder_funcs,
DRM_MODE_ENCODER_VIRTUAL, NULL);
if (ret)
return ret;
ret = drm_writeback_connector_init_with_encoder(dev, wb_connector,
con_funcs, formats,
n_formats);
Yes, this is exactly what I had in mind.
Thank you for confirming.
drm_writeback_connector_init_with_encoder() will still not have an encoder
parameter because of what I wrote previously.
But in your patch drm_writeback_connector_init_with_encoder() still has an
encoder_funcs pointer which is useless, as the expectations are (AFAII) that the
whole encoder init dance has already happened. And while you have a point that
you
can set the encoder pointer in the connector before calling
drm_writeback_connector_init_with_encoder() I think it would be easier to read
if you
pass the encoder explicitly in the parameter list and skip the assignment in
(1) and
do it inside drm_writeback_connector_init_with_encoder(). Again, your code is
not
wrong, I just think we should be explicit so that code is easier to read.
Sorry, but I am missing something here.
Where is the encoder_funcs pointer in
drm_writeback_connector_init_with_encoder()? It only has
drm_connector_funcs pointer. Because, like I mentioned in the earlier
comment, the callers of this API would have already setup the encoder
with the required functions/possible_crtcs etc.
int drm_writeback_connector_init_with_encoder(struct drm_device *dev,
struct drm_writeback_connector *wb_connector,
const struct drm_connector_funcs *con_funcs, const u32 *formats,
int n_formats)
{
int ret = 0;
ret = drm_writeback_connector_setup(dev, wb_connector, con_funcs,
formats,
n_formats);
return ret;
}
So are you suggesting I pass the drm_encoder as a parameter to
drm_writeback_connector_init_with_encoder as well?
That has two potential issues:
1) If we move just the assignment of wb_connector->encoder to inside
drm_writeback_connector_init_with_encoder, we cannot do these before
calling drm_writeback_connector_init_with_encoder() as
wb_connector->encoder will be NULL for the cases which use internal_encoder
drm_encoder_helper_add(wb_connector->encoder, enc_helper_funcs);
wb_connector->encoder->possible_crtcs = possible_crtcs;
ret = drm_encoder_init(dev, wb_connector->encoder,
&drm_writeback_encoder_funcs,
DRM_MODE_ENCODER_VIRTUAL, NULL);
2) We can only do these changes once wb_connector->encoder has been
changed to a pointer which happens in patch 4 of this series
https://patchwork.freedesktop.org/patch/479084/?series=101260&rev=5
I was told to split the functionality of adding the new API in a
separate patch by laurent.
To keep changes independent and functionally correct I thought this is
the best split.
So the cleanup you are suggesting can be done in patch 4 of this series
but still I am not sure how to get around concern (1) because the
encoder has to be fully initialized before attaching to a wb_connector.
Hope this is what you had in mind as well.
2) Inside drm_writeback_connector_init() check:
drm_writeback_connector_init(.....)
{
if (!wb_conn->encoder)
initialize the encoder
No, the assumption of drm_writeback_connector_init() is that we will provide the
encoder, so no need to check that one is already provided. What you could do is:
WARN_ON(wb_conn->encoder);
Got it, i will add a warning inside drm_writeback_connector_init() to
emphasize that its only meant for cases where an encoder is not provided.
before we overwrite the encoder. That way we will get a nice warning in the
kernel
log if someone tries to call drm_writeback_connector_init() with a preset
encoder.
call drm_writeback_**_init_with_encoder()
}
This will also work but have the foll concerns/questions:
1) Will drm_writeback_connector_init_with_encoder() be exported if we change
as per your suggestion or will all clients continue to use
drm_writeback_connector_init() only? We wanted to have a separate function
for new clients.
Yes, we will need to export drm_writeback_connector_init_with_encoder() as well.
Alright, so vc4 and new vendors which provide their own encoder will use
this one. So no changes to the rest of the series.
2) How will we detect that encoder needs to be initialized without it being
a pointer which happens in the later change. So ordering becomes an issue.
I think the init problem is simple. You either call
drm_writeback_connector_init()
and the framework provides you with an encoder, or you call
drm_writeback_connector_init_with_encoder() and the framework will use yours.
The
problems will show up on the cleanup and exit codes, where we need to be able
to skip
the cleanup if the encoder pointer is not the internal one. I think a simple
if (connector->encoder == &connector->internal_encoder)
do_encoder_cleanup_work_here()
should work.
Sorry, I am missing something here.
Even in this current patch, the drm_encoder_cleanup() is done only inside
drm_writeback_connector_init() where internal_encoder is used.
For drm_writeback_connector_init_with_encoder(), we dont do the cleanup and
expect the caller to do it like vc4 does in the next patch.
So why do we need such a condition?
You're right. I was thinking that at cleanup time we also need to do work with
the
right encoder, but that should accomplished by passing the right .destroy hook
in the
drm_encoder_funcs pointer via drm_encoder_init. So if the special drivers to the
initialisation correctly it should all work fine, please disregard my comments.
Best regards,
Liviu
Thats why I thought this is the best way to address the comments and keep
the functionality intact with each change.
Let me know if I have misunderstood some comment or idea.
I hope that with these clarifications we are both on the same page.
Best regards,
Liviu
+{
+ int ret = 0;
+
+ ret = drm_writeback_connector_setup(dev, wb_connector, con_funcs,
formats,
+ n_formats);
+
+ return ret;
+}
+EXPORT_SYMBOL(drm_writeback_connector_init_with_encoder);
+
int drm_writeback_set_fb(struct drm_connector_state *conn_state,
struct drm_framebuffer *fb)
{
diff --git a/include/drm/drm_writeback.h b/include/drm/drm_writeback.h
index db6214f..0093bab 100644
--- a/include/drm/drm_writeback.h
+++ b/include/drm/drm_writeback.h
@@ -152,6 +152,11 @@ int drm_writeback_connector_init(struct drm_device *dev,
const struct drm_encoder_helper_funcs
*enc_helper_funcs,
const u32 *formats, int n_formats, uint32_t
possible_crtcs);
+int drm_writeback_connector_init_with_encoder(struct drm_device *dev,
+ struct drm_writeback_connector *wb_connector,
+ const struct drm_connector_funcs *con_funcs,
const u32 *formats,
+ int n_formats);
+
int drm_writeback_set_fb(struct drm_connector_state *conn_state,
struct drm_framebuffer *fb);
--
2.7.4
