[Sound-open-firmware] [PATCH] apl: memory: expand number of 256 Bytes block to be 64

Keyon Jie yang.jie at linux.intel.com
Fri Jun 8 14:10:29 CEST 2018



On 2018年06月07日 22:55, Liam Girdwood wrote:
> On Thu, 2018-06-07 at 16:23 +0800, Keyon Jie wrote:
>>
>> On 2018年06月07日 16:06, Keyon Jie wrote:
>>>
>>>
>>> On 2018年05月31日 19:51, Liam Girdwood wrote:
>>>> On Thu, 2018-05-31 at 16:05 +0800, Keyon Jie wrote:
>>>>> The 256 Bytes block are used frequently for component new, 16 is
>>>>> not enough for GP-MRB multiple SSPs/pipelines support, here expand
>>>>> it to be 64.
>>>>>
>>>>> Signed-off-by: Keyon Jie <yang.jie at linux.intel.com>
>>>>> ---
>>>>> Test with:
>>>>> APL-MRB with tdf8532 codec
>>>>> Linux topic/sof-v4.14: 3ffd39334cb53aced88b8aba17cae3df9ee2736f
>>>>> SOF master: 48a61423f6c2d2bcf5fad854a9b3d37dfb43b546
>>>>> Tool master: 7ee114d2ed822a0dbfc0dc3a37fd5c4977e00be8
>>>>> ---
>>>>>    src/platform/apollolake/include/platform/memory.h | 2 +-
>>>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/src/platform/apollolake/include/platform/memory.h
>>>>> b/src/platform/apollolake/include/platform/memory.h
>>>>> index e7f3ee3..652db96 100644
>>>>> --- a/src/platform/apollolake/include/platform/memory.h
>>>>> +++ b/src/platform/apollolake/include/platform/memory.h
>>>>> @@ -159,7 +159,7 @@
>>>>>    #define HEAP_RT_COUNT32            128
>>>>>    #define HEAP_RT_COUNT64            64
>>>>>    #define HEAP_RT_COUNT128        32
>>>>> -#define HEAP_RT_COUNT256        16
>>>>> +#define HEAP_RT_COUNT256        64
>>>>>    #define HEAP_RT_COUNT512        8
>>>>>    #define HEAP_RT_COUNT1024        4
>>>>
>>>> Can you check this does not overflow the HEAP into another section. We
>>>> could
>>>> also look at deprecating 1024 & 512 and giving this space up to the
>>>> others.
>>>
>>> this change will increase HEAP_RUNTIME_SIZE with 4KB, and
>>> HEAP_BUFFER_BASE will be shifted 4KB upper also, actually
>>> HEAP_BUFFER_SIZE will be decreased 4KB at the end.
>>
>> Sorry, it should be (4 - 1) * 4KB = 12KB shift.
>>
> 
> Ok, can you do likewise for the other platforms and decrease 1024 to 0 (giving
> it's space to 128, 64 and 32

sure, I can do that, but we have only 4KB room increase with this change 
and that is not enough for what COUNT256 needed. and are you sure nobody 
is using 1024 ones?

Thanks,
~Keyon

> 
> Liam
> _______________________________________________
> Sound-open-firmware mailing list
> Sound-open-firmware at alsa-project.org
> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmware
> 


More information about the Sound-open-firmware mailing list