[meta-arago,master/kirkstone,v2] gstreamer1.0-plugins-good: wave5-codec: Set max buffers same as min buffers

v4l2src plugin queues same external buffer with different v4l2 buffer
with different index. This workaround was implemented as a workaround
for this. Workaround needs all buffers to be allocated during the init
phase, if new buffers are allocated post stream start it fails. Avoid
such issues by enforcing max buffers same as min buffers

Signed-off-by: Rahul T R <r-ravikumar@ti.com>
Reviewed-by: Devarsh Thakkar <devarsht@ti.com>
Reviewed-by: Prasanth Babu Mantena <p-mantena@ti.com>
---

Notes:
    v2:
        Updated commit message as per the comments
        Added the reviewed bys

 .../0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch      | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

Unfortunately I had already applied, tested, and marked the v1 patch for 
inclusion in CICD.  And the patch has already been promoted onto the 
master branch.

You sent a message in webex that implied a level of urgency with getting 
this patch in for CICD.  In the future, if there is no urgency and you 
want feedback on a patch, then do not send urgency messages off mailing 
list.  I am watching this list and checking things on a regular basis.



On 10/18/2023 4:19 AM, Rahul T R wrote:
> v4l2src plugin queues same external buffer with different v4l2 buffer
> with different index. This workaround was implemented as a workaround
> for this. Workaround needs all buffers to be allocated during the init
> phase, if new buffers are allocated post stream start it fails. Avoid
> such issues by enforcing max buffers same as min buffers
> 
> Signed-off-by: Rahul T R <r-ravikumar@ti.com>
> Reviewed-by: Devarsh Thakkar <devarsht@ti.com>
> Reviewed-by: Prasanth Babu Mantena <p-mantena@ti.com>
> ---
> 
> Notes:
>      v2:
>          Updated commit message as per the comments
>          Added the reviewed bys
> 
>   .../0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch      | 7 ++++---
>   1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch b/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch
> index a1471661..bf73e257 100644
> --- a/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch
> +++ b/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch
> @@ -14,9 +14,9 @@ Signed-off-by: Prasanth Babu Mantena <p-mantena@ti.com>
>   ---
>    sys/v4l2/gstv4l2bufferpool.c | 17 ++++++++++++++---
>    sys/v4l2/gstv4l2bufferpool.h |  2 ++
> - sys/v4l2/gstv4l2object.c     | 15 ++++++++++++---
> + sys/v4l2/gstv4l2object.c     | 16 +++++++++++++---
>    sys/v4l2/gstv4l2videodec.c   | 20 ++++++++++++++------
> - 4 files changed, 42 insertions(+), 12 deletions(-)
> + 4 files changed, 43 insertions(+), 12 deletions(-)
>   
>   diff --git a/sys/v4l2/gstv4l2bufferpool.c b/sys/v4l2/gstv4l2bufferpool.c
>   index d85f036..e6a60dc 100644
> @@ -86,7 +86,7 @@ index ee60540..eff1cf2 100644
>        own_min = MAX (own_min, GST_V4L2_MIN_BUFFERS (obj));
>   
>        /* for the downstream pool, we keep what downstream wants, though ensure
> -@@ -5049,8 +5053,13 @@ gst_v4l2_object_decide_allocation (GstV4l2Object * obj, GstQuery * query)
> +@@ -5049,8 +5053,14 @@ gst_v4l2_object_decide_allocation (GstV4l2Object * obj, GstQuery * query)
>        min = MAX (min, GST_V4L2_MIN_BUFFERS (obj));
>   
>        /* To import we need the other pool to hold at least own_min */
> @@ -95,6 +95,7 @@ index ee60540..eff1cf2 100644
>   +    if (obj_pool == pool) {
>   +        if (0 == strcmp(obj->vcap.driver, "wave5-dec")) {
>   +            min = own_min;
> ++            max = min;
>   +        } else {
>   +            min += own_min;
>   +        }

On 18/10/23 20:05, Ryan Eatmon wrote:

>
> Unfortunately I had already applied, tested, and marked the v1 patch 
> for inclusion in CICD.  And the patch has already been promoted onto 
> the master branch.
>
> You sent a message in webex that implied a level of urgency with 
> getting this patch in for CICD.  In the future, if there is no urgency 
> and you want feedback on a patch, then do not send urgency messages 
> off mailing list.  I am watching this list and checking things on a 
> regular basis.
>
Hi Ryan,

Sorry about the confusion, It was indeed urgent
anyway the change was only in the commit message

Regards
Rahul T R

>
>
> On 10/18/2023 4:19 AM, Rahul T R wrote:
>> v4l2src plugin queues same external buffer with different v4l2 buffer
>> with different index. This workaround was implemented as a workaround
>> for this. Workaround needs all buffers to be allocated during the init
>> phase, if new buffers are allocated post stream start it fails. Avoid
>> such issues by enforcing max buffers same as min buffers
>>
>> Signed-off-by: Rahul T R <r-ravikumar@ti.com>
>> Reviewed-by: Devarsh Thakkar <devarsht@ti.com>
>> Reviewed-by: Prasanth Babu Mantena <p-mantena@ti.com>
>> ---
>>
>> Notes:
>>      v2:
>>          Updated commit message as per the comments
>>          Added the reviewed bys
>>
>>   .../0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch      | 7 ++++---
>>   1 file changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git 
>> a/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch 
>> b/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch 
>>
>> index a1471661..bf73e257 100644
>> --- 
>> a/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch
>> +++ 
>> b/meta-arago-extras/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good/0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch
>> @@ -14,9 +14,9 @@ Signed-off-by: Prasanth Babu Mantena 
>> <p-mantena@ti.com>
>>   ---
>>    sys/v4l2/gstv4l2bufferpool.c | 17 ++++++++++++++---
>>    sys/v4l2/gstv4l2bufferpool.h |  2 ++
>> - sys/v4l2/gstv4l2object.c     | 15 ++++++++++++---
>> + sys/v4l2/gstv4l2object.c     | 16 +++++++++++++---
>>    sys/v4l2/gstv4l2videodec.c   | 20 ++++++++++++++------
>> - 4 files changed, 42 insertions(+), 12 deletions(-)
>> + 4 files changed, 43 insertions(+), 12 deletions(-)
>>     diff --git a/sys/v4l2/gstv4l2bufferpool.c 
>> b/sys/v4l2/gstv4l2bufferpool.c
>>   index d85f036..e6a60dc 100644
>> @@ -86,7 +86,7 @@ index ee60540..eff1cf2 100644
>>        own_min = MAX (own_min, GST_V4L2_MIN_BUFFERS (obj));
>>          /* for the downstream pool, we keep what downstream wants, 
>> though ensure
>> -@@ -5049,8 +5053,13 @@ gst_v4l2_object_decide_allocation 
>> (GstV4l2Object * obj, GstQuery * query)
>> +@@ -5049,8 +5053,14 @@ gst_v4l2_object_decide_allocation 
>> (GstV4l2Object * obj, GstQuery * query)
>>        min = MAX (min, GST_V4L2_MIN_BUFFERS (obj));
>>          /* To import we need the other pool to hold at least own_min */
>> @@ -95,6 +95,7 @@ index ee60540..eff1cf2 100644
>>   +    if (obj_pool == pool) {
>>   +        if (0 == strcmp(obj->vcap.driver, "wave5-dec")) {
>>   +            min = own_min;
>> ++            max = min;
>>   +        } else {
>>   +            min += own_min;
>>   +        }
>