SOC_FAMILY Rework

SOC_FAMILY Rework

[Daiane Angolini]

We have today the following SOC_FAMILY tree (for meta-fsl-arm)

i.MX -> mx3 -> mx31;
i.MX -> mx3 -> mx35;
i.MX -> mx5 -> mx51;
i.MX -> mx5 -> mx53;
i.MX -> mx6 -> mx6dl;
i.MX -> mx6 -> mx6q;
i.MX -> mx6 -> mx6sl;
i.MX -> mx6 -> mx6sx;
i.MX -> mxs -> mx23;
i.MX -> mxs -> mx28;
Layerscape -> ls102xa;
Vybrid -> vf -> vf50;
Vybrid -> vf -> vf60;

Where i.MX, Layerscape and Vybrid are only included for visual
purposes, not being a real OVERRIDE today (the imx inclusion is a
pending patch in ML right now)

Some of the points I think is important to consider when choosing a
OVERRIDE tree is the SoC divergence and convergence. In i.MX case, we
have some obvious examples such as GPU type, existence of VPU, and
machine features such as CAN (although imx6 generally supports CAN,
there are some board with connector, and other without it).

In CAN case, CAN is a SoC feature enabled by the board or not. In WIFI
case, it’s only a board feature.

So, the points to rule the OVERRIDE is SoC features and board features.

Another important thing to keep in mind is the deepness. How deep
should be the OVERRIDE tree? Two or tree levels?

OVERRIDE should be used to cluster BSP packages and configurations
accordingly with SoC and machine features. Today the meta-fsl-arm BSP
packages can be seen in the table [1]. Most of packages are the same
for all i.MX boards (including kernel and u-boot not included in the
table), but some are very specific to a group of SoCs only (i.e.
fsl-alsa-plugins), and there is the case where a package is for only
one SoC Family, but has big inner segmentation (Vivante)

[1] https://github.com/Freescale/Documentation/blob/master/release-notes/source/soc-pkg-optimization.inc

The classic example of a crazy OVERRIDE is ‘mxs’. It was included in
the past to cluster imx23 and imx28 because they are the only 2 board
using imx-bootlets. Does ‘mxs’ make sense nowaday? Would imx6UL be
more related with imx28 than with imx6Q? Would imx6SX be clustered
with vf60?

Maybe, we can concentrate the OVERRIDE only around GPU and VPU. We
have an extensive segmentation of OVERRIDE for imx6, and a cluster for
all the SoCs without GPU (vf,layerscape?,imx2,imx3,imx6UL, imx7)

I only thing this is the right timing to review and rework SOC_FAMILY.
I'm i.MX biased, so I can only say about i.MX, would be glad to hear
feedback from Vybrid and QoirQ.

Regards,
Daiane

Re: SOC_FAMILY Rework

[Otavio Salvador]

Hello Daiane,

First, thank you for preparing this e-mail.

On Wed, Jul 8, 2015 at 4:42 PM, Daiane Angolini <daiane.list@gmail.com> wrote:
> We have today the following SOC_FAMILY tree (for meta-fsl-arm)
>
> i.MX -> mx3 -> mx31;
> i.MX -> mx3 -> mx35;
> i.MX -> mx5 -> mx51;
> i.MX -> mx5 -> mx53;
> i.MX -> mx6 -> mx6dl;
> i.MX -> mx6 -> mx6q;
> i.MX -> mx6 -> mx6sl;
> i.MX -> mx6 -> mx6sx;
> i.MX -> mxs -> mx23;
> i.MX -> mxs -> mx28;

Those are fine, as far as I can see.

> Layerscape -> ls102xa;

I think this should be:

qoriq -> qoriq-arm -> ls102xa

> Vybrid -> vf -> vf50;
> Vybrid -> vf -> vf60;

I would change this for:

vybrid > vf50;
vybrid > vf60;

...
> Some of the points I think is important to consider when choosing a
> OVERRIDE tree is the SoC divergence and convergence. In i.MX case, we
> have some obvious examples such as GPU type, existence of VPU, and
> machine features such as CAN (although imx6 generally supports CAN,
> there are some board with connector, and other without it).
>
> In CAN case, CAN is a SoC feature enabled by the board or not. In WIFI
> case, it’s only a board feature.
>
> So, the points to rule the OVERRIDE is SoC features and board features.

Not really.

The SoC override is to offer a way to modify the source for ALL boards
and also to offer the possibility to cluster binary compatible
binaries.

This ends affecting the MACHINE_SOCARCH and the SoC subarch we
designed to avoid too much rebuilds. Features as CAN or WIFI are
machine features and shouldn't be mapped for SoC overrides.

> Another important thing to keep in mind is the deepness. How deep
> should be the OVERRIDE tree? Two or tree levels?

Not a big deal; it depends if it makes usage sense or not. Current set
seems good for me.

> OVERRIDE should be used to cluster BSP packages and configurations
> accordingly with SoC and machine features. Today the meta-fsl-arm BSP
> packages can be seen in the table [1]. Most of packages are the same
> for all i.MX boards (including kernel and u-boot not included in the
> table), but some are very specific to a group of SoCs only (i.e.
> fsl-alsa-plugins), and there is the case where a package is for only
> one SoC Family, but has big inner segmentation (Vivante)
>
> [1] https://github.com/Freescale/Documentation/blob/master/release-notes/source/soc-pkg-optimization.inc

Agreed.

> The classic example of a crazy OVERRIDE is ‘mxs’. It was included in
> the past to cluster imx23 and imx28 because they are the only 2 board
> using imx-bootlets. Does ‘mxs’ make sense nowaday? Would imx6UL be
> more related with imx28 than with imx6Q? Would imx6SX be clustered
> with vf60?

I think UL will end being a mx6ul and it will demand the check of all
mx6 uses to see if it is compatible or not. See Lauren's patch
regarding the GPU presence in Weston for an example.

> Maybe, we can concentrate the OVERRIDE only around GPU and VPU. We
> have an extensive segmentation of OVERRIDE for imx6, and a cluster for
> all the SoCs without GPU (vf,layerscape?,imx2,imx3,imx6UL, imx7)

It does not work. The point of OVERRIDE is to make changes in a set of
SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
Qualcomm...)

> I only thing this is the right timing to review and rework SOC_FAMILY.
> I'm i.MX biased, so I can only say about i.MX, would be glad to hear
> feedback from Vybrid and QoirQ.

As I said before, the QorIQ should have common and arch specific
overrides (I see the former going to be most used) and i.MX global
override will be rarely used.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Daiane Angolini]

On Thu, Jul 9, 2015 at 11:05 AM, Otavio Salvador
<otavio@ossystems.com.br> wrote:
> Hello Daiane,
>
> First, thank you for preparing this e-mail.
>
> On Wed, Jul 8, 2015 at 4:42 PM, Daiane Angolini <daiane.list@gmail.com> wrote:
>> We have today the following SOC_FAMILY tree (for meta-fsl-arm)
>>
>> i.MX -> mx3 -> mx31;
>> i.MX -> mx3 -> mx35;
>> i.MX -> mx5 -> mx51;
>> i.MX -> mx5 -> mx53;
>> i.MX -> mx6 -> mx6dl;
>> i.MX -> mx6 -> mx6q;
>> i.MX -> mx6 -> mx6sl;
>> i.MX -> mx6 -> mx6sx;
>> i.MX -> mxs -> mx23;
>> i.MX -> mxs -> mx28;
>
> Those are fine, as far as I can see.

I'm sorry, Otavio, but you are not re-thinking what we have. I don't
think this list is fine. I don't see why to separate imx5 from imxs.
And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
exactly like a imx3.

>
>> Layerscape -> ls102xa;
>
> I think this should be:
>
> qoriq -> qoriq-arm -> ls102xa
>
>> Vybrid -> vf -> vf50;
>> Vybrid -> vf -> vf60;
>
> I would change this for:
>
> vybrid > vf50;
> vybrid > vf60;
>
> ...
>> Some of the points I think is important to consider when choosing a
>> OVERRIDE tree is the SoC divergence and convergence. In i.MX case, we
>> have some obvious examples such as GPU type, existence of VPU, and
>> machine features such as CAN (although imx6 generally supports CAN,
>> there are some board with connector, and other without it).
>>
>> In CAN case, CAN is a SoC feature enabled by the board or not. In WIFI
>> case, it’s only a board feature.
>>
>> So, the points to rule the OVERRIDE is SoC features and board features.
>
> Not really.
>
> The SoC override is to offer a way to modify the source for ALL boards
> and also to offer the possibility to cluster binary compatible
> binaries.
>
> This ends affecting the MACHINE_SOCARCH and the SoC subarch we
> designed to avoid too much rebuilds. Features as CAN or WIFI are
> machine features and shouldn't be mapped for SoC overrides.

ok

>
>> Another important thing to keep in mind is the deepness. How deep
>> should be the OVERRIDE tree? Two or tree levels?
>
> Not a big deal; it depends if it makes usage sense or not. Current set
> seems good for me.

In fact I think it's a big deal.

Maybe there is not so much difference between 2 and 3, but if you
think about 30 levels it is indeed a big deal.

I understand the number of levels is arbitrary. But it only means we
must decide which are the prefect deepness.


>
>> OVERRIDE should be used to cluster BSP packages and configurations
>> accordingly with SoC and machine features. Today the meta-fsl-arm BSP
>> packages can be seen in the table [1]. Most of packages are the same
>> for all i.MX boards (including kernel and u-boot not included in the
>> table), but some are very specific to a group of SoCs only (i.e.
>> fsl-alsa-plugins), and there is the case where a package is for only
>> one SoC Family, but has big inner segmentation (Vivante)
>>
>> [1] https://github.com/Freescale/Documentation/blob/master/release-notes/source/soc-pkg-optimization.inc
>
> Agreed.
>
>> The classic example of a crazy OVERRIDE is ‘mxs’. It was included in
>> the past to cluster imx23 and imx28 because they are the only 2 board
>> using imx-bootlets. Does ‘mxs’ make sense nowaday? Would imx6UL be
>> more related with imx28 than with imx6Q? Would imx6SX be clustered
>> with vf60?
>
> I think UL will end being a mx6ul and it will demand the check of all
> mx6 uses to see if it is compatible or not. See Lauren's patch
> regarding the GPU presence in Weston for an example.
>
>> Maybe, we can concentrate the OVERRIDE only around GPU and VPU. We
>> have an extensive segmentation of OVERRIDE for imx6, and a cluster for
>> all the SoCs without GPU (vf,layerscape?,imx2,imx3,imx6UL, imx7)
>
> It does not work. The point of OVERRIDE is to make changes in a set of
> SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
> Qualcomm...)

GPU is a SoC characteristic, and it's obvious it's the most important
BSP piece on i.MX6 family. Instead of having a tree like:

imx -> imx6 -> everysingle soc

We could have something like:

imx -> imx6-light -> ligh socs
imx -> imx6-heavy -> heavy socs

and, instead of list "everysingle soc" when setting GPU BSP, we can
list only "light and heavy".

The downside is to list "light and heavy" in packages for imx6


I don't see how it will contaminate other BSP. I'm trying to stress
the imx OVERRIDE just included.


>
>> I only thing this is the right timing to review and rework SOC_FAMILY.
>> I'm i.MX biased, so I can only say about i.MX, would be glad to hear
>> feedback from Vybrid and QoirQ.
>
> As I said before, the QorIQ should have common and arch specific
> overrides (I see the former going to be most used) and i.MX global
> override will be rarely used.

One argument to keep things like it is today is learning curve
one argument to change it in meta-freescale is upgrade path


Daiane
>
> --
> Otavio Salvador                             O.S. Systems
> http://www.ossystems.com.br        http://code.ossystems.com.br
> Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Otavio Salvador]

On Thu, Jul 9, 2015 at 11:54 AM, Daiane Angolini <daiane.list@gmail.com> wrote:
> On Thu, Jul 9, 2015 at 11:05 AM, Otavio Salvador
> <otavio@ossystems.com.br> wrote:
>> Hello Daiane,
>>
>> First, thank you for preparing this e-mail.
>>
>> On Wed, Jul 8, 2015 at 4:42 PM, Daiane Angolini <daiane.list@gmail.com> wrote:
>>> We have today the following SOC_FAMILY tree (for meta-fsl-arm)
>>>
>>> i.MX -> mx3 -> mx31;
>>> i.MX -> mx3 -> mx35;
>>> i.MX -> mx5 -> mx51;
>>> i.MX -> mx5 -> mx53;
>>> i.MX -> mx6 -> mx6dl;
>>> i.MX -> mx6 -> mx6q;
>>> i.MX -> mx6 -> mx6sl;
>>> i.MX -> mx6 -> mx6sx;
>>> i.MX -> mxs -> mx23;
>>> i.MX -> mxs -> mx28;
>>
>> Those are fine, as far as I can see.
>
> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
> think this list is fine. I don't see why to separate imx5 from imxs.
> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
> exactly like a imx3.

Yes; when it goes than mx5 will be in mainline use as will be mx28
(mx23 is already using it).

However this is a new environment and when this happens we can rework
this. I am not sure we can kill the SoC families as it is needed for
U-Boot and other things internally as well (see mxs and imx-base
files).

...
>>> Another important thing to keep in mind is the deepness. How deep
>>> should be the OVERRIDE tree? Two or tree levels?
>>
>> Not a big deal; it depends if it makes usage sense or not. Current set
>> seems good for me.
>
> In fact I think it's a big deal.
>
> Maybe there is not so much difference between 2 and 3, but if you
> think about 30 levels it is indeed a big deal.
>
> I understand the number of levels is arbitrary. But it only means we
> must decide which are the prefect deepness.

Every modularization imposes complexity so I think usefulness is the
metric to be put in place here.

>>> OVERRIDE should be used to cluster BSP packages and configurations
>>> accordingly with SoC and machine features. Today the meta-fsl-arm BSP
>>> packages can be seen in the table [1]. Most of packages are the same
>>> for all i.MX boards (including kernel and u-boot not included in the
>>> table), but some are very specific to a group of SoCs only (i.e.
>>> fsl-alsa-plugins), and there is the case where a package is for only
>>> one SoC Family, but has big inner segmentation (Vivante)
>>>
>>> [1] https://github.com/Freescale/Documentation/blob/master/release-notes/source/soc-pkg-optimization.inc
>>
>> Agreed.
>>
>>> The classic example of a crazy OVERRIDE is ‘mxs’. It was included in
>>> the past to cluster imx23 and imx28 because they are the only 2 board
>>> using imx-bootlets. Does ‘mxs’ make sense nowaday? Would imx6UL be
>>> more related with imx28 than with imx6Q? Would imx6SX be clustered
>>> with vf60?
>>
>> I think UL will end being a mx6ul and it will demand the check of all
>> mx6 uses to see if it is compatible or not. See Lauren's patch
>> regarding the GPU presence in Weston for an example.
>>
>>> Maybe, we can concentrate the OVERRIDE only around GPU and VPU. We
>>> have an extensive segmentation of OVERRIDE for imx6, and a cluster for
>>> all the SoCs without GPU (vf,layerscape?,imx2,imx3,imx6UL, imx7)
>>
>> It does not work. The point of OVERRIDE is to make changes in a set of
>> SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
>> Qualcomm...)
>
> GPU is a SoC characteristic, and it's obvious it's the most important
> BSP piece on i.MX6 family. Instead of having a tree like:
>
> imx -> imx6 -> everysingle soc
>
> We could have something like:
>
> imx -> imx6-light -> ligh socs
> imx -> imx6-heavy -> heavy socs
>
> and, instead of list "everysingle soc" when setting GPU BSP, we can
> list only "light and heavy".
>
> The downside is to list "light and heavy" in packages for imx6

I don't agree; we don't know what i.MX8 will look like, what i.MX9
will look like so we should put it as close as possible for product
family.

The feature set is too fragmented and too SoC specific so grouping it
in super-groups is risky.

> I don't see how it will contaminate other BSP. I'm trying to stress
> the imx OVERRIDE just included.

IF we use the overrides wrong it does.

There is an arm override, we cannot use this for an i.MX patch as TI
would be affected, for example.

>>> I only thing this is the right timing to review and rework SOC_FAMILY.
>>> I'm i.MX biased, so I can only say about i.MX, would be glad to hear
>>> feedback from Vybrid and QoirQ.
>>
>> As I said before, the QorIQ should have common and arch specific
>> overrides (I see the former going to be most used) and i.MX global
>> override will be rarely used.
>
> One argument to keep things like it is today is learning curve
> one argument to change it in meta-freescale is upgrade path

I think Vybrid and Layerscape does need rework. mx5 and mxs may be
reduced once we move to mainline  but not killed and they are not same
thing as well.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Eric Bénard]

Hi Daiane,

Le Thu, 9 Jul 2015 11:54:14 -0300,
Daiane Angolini <daiane.list@gmail.com> a écrit :
> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
> think this list is fine. I don't see why to separate imx5 from imxs.
> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
> exactly like a imx3.
> 
Does that mean there is an updated GPU library and full IPU drivers for
more recent kernels than 2.6.35 for i.MX53 ?

Thanks,
Eric

Re: SOC_FAMILY Rework

[Nikolay Dimitrov]

Hi Daiane,

On 07/09/2015 05:54 PM, Daiane Angolini wrote:
>> It does not work. The point of OVERRIDE is to make changes in a set of
>> SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
>> Qualcomm...)
>
> GPU is a SoC characteristic, and it's obvious it's the most important
> BSP piece on i.MX6 family. Instead of having a tree like:
>
> imx -> imx6 -> everysingle soc
>
> We could have something like:
>
> imx -> imx6-light -> ligh socs
> imx -> imx6-heavy -> heavy socs
>
> and, instead of list "everysingle soc" when setting GPU BSP, we can
> list only "light and heavy".
>
> The downside is to list "light and heavy" in packages for imx6
>
>
> I don't see how it will contaminate other BSP. I'm trying to stress
> the imx OVERRIDE just included.

This is an excellent example. The OVERRIDEs should generalize, not
specialize (although you can always interpret a root-leaf traversal as
a specialization :D). Detailed specialization will lead to OVERRIDEs
tree explosion while trying to express every single SOC feature on each
level of the tree.

If we try to express SOC features via OVERRIDE mechanism, it's imho
doomed.

Regards,
Nikolay

Re: SOC_FAMILY Rework

[Otavio Salvador]

On Thu, Jul 9, 2015 at 12:18 PM, Eric Bénard <eric@eukrea.com> wrote:
> Hi Daiane,
>
> Le Thu, 9 Jul 2015 11:54:14 -0300,
> Daiane Angolini <daiane.list@gmail.com> a écrit :
>> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
>> think this list is fine. I don't see why to separate imx5 from imxs.
>> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
>> exactly like a imx3.
>>
> Does that mean there is an updated GPU library and full IPU drivers for
> more recent kernels than 2.6.35 for i.MX53 ?

This means mx5 will likely don't have GPU support in Yocto Project
1.9; VPU will likely work using mainline CODA.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Otavio Salvador]

On Thu, Jul 9, 2015 at 12:20 PM, Nikolay Dimitrov <picmaster@mail.bg> wrote:
> On 07/09/2015 05:54 PM, Daiane Angolini wrote:
>>>
>>> It does not work. The point of OVERRIDE is to make changes in a set of
>>> SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
>>> Qualcomm...)
>>
>>
>> GPU is a SoC characteristic, and it's obvious it's the most important
>> BSP piece on i.MX6 family. Instead of having a tree like:
>>
>> imx -> imx6 -> everysingle soc
>>
>> We could have something like:
>>
>> imx -> imx6-light -> ligh socs
>> imx -> imx6-heavy -> heavy socs
>>
>> and, instead of list "everysingle soc" when setting GPU BSP, we can
>> list only "light and heavy".
>>
>> The downside is to list "light and heavy" in packages for imx6
>>
>>
>> I don't see how it will contaminate other BSP. I'm trying to stress
>> the imx OVERRIDE just included.
>
>
> This is an excellent example. The OVERRIDEs should generalize, not
> specialize (although you can always interpret a root-leaf traversal as
> a specialization :D). Detailed specialization will lead to OVERRIDEs
> tree explosion while trying to express every single SOC feature on each
> level of the tree.
>
> If we try to express SOC features via OVERRIDE mechanism, it's imho
> doomed.

That's why we use the family. mx53; mx6q; ...

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Eric Bénard]

Hi Otavio,

Le Thu, 9 Jul 2015 12:25:27 -0300,
Otavio Salvador <otavio@ossystems.com.br> a écrit :

> On Thu, Jul 9, 2015 at 12:18 PM, Eric Bénard <eric@eukrea.com> wrote:
> > Hi Daiane,
> >
> > Le Thu, 9 Jul 2015 11:54:14 -0300,
> > Daiane Angolini <daiane.list@gmail.com> a écrit :
> >> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
> >> think this list is fine. I don't see why to separate imx5 from imxs.
> >> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
> >> exactly like a imx3.
> >>
> > Does that mean there is an updated GPU library and full IPU drivers for
> > more recent kernels than 2.6.35 for i.MX53 ?
> 
> This means mx5 will likely don't have GPU support in Yocto Project
> 1.9; VPU will likely work using mainline CODA.
> 
and also most of the tools mixing VPU & IPU (CSI for example) ...
Is that what you want in a "reference" BSP for reference platforms ?

Eric

Re: SOC_FAMILY Rework

[Otavio Salvador]

On Thu, Jul 9, 2015 at 12:47 PM, Eric Bénard <eric@eukrea.com> wrote:
> Le Thu, 9 Jul 2015 12:25:27 -0300,
> Otavio Salvador <otavio@ossystems.com.br> a écrit :
>
>> On Thu, Jul 9, 2015 at 12:18 PM, Eric Bénard <eric@eukrea.com> wrote:
>> > Hi Daiane,
>> >
>> > Le Thu, 9 Jul 2015 11:54:14 -0300,
>> > Daiane Angolini <daiane.list@gmail.com> a écrit :
>> >> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
>> >> think this list is fine. I don't see why to separate imx5 from imxs.
>> >> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
>> >> exactly like a imx3.
>> >>
>> > Does that mean there is an updated GPU library and full IPU drivers for
>> > more recent kernels than 2.6.35 for i.MX53 ?
>>
>> This means mx5 will likely don't have GPU support in Yocto Project
>> 1.9; VPU will likely work using mainline CODA.
>>
> and also most of the tools mixing VPU & IPU (CSI for example) ...
> Is that what you want in a "reference" BSP for reference platforms ?

If Freescale does not upgrade the AMD support components for newer
kernel someone needs to do it.

As a community member you are welcome to-do so, especially as Eukrea
has mx5-based SoMs. If no one does it, it will be removed.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Eric Bénard]

Hi Otavio,

Le Thu, 9 Jul 2015 13:45:10 -0300,
Otavio Salvador <otavio@ossystems.com.br> a écrit :

> On Thu, Jul 9, 2015 at 12:47 PM, Eric Bénard <eric@eukrea.com> wrote:
> > Le Thu, 9 Jul 2015 12:25:27 -0300,
> > Otavio Salvador <otavio@ossystems.com.br> a écrit :
> >
> >> On Thu, Jul 9, 2015 at 12:18 PM, Eric Bénard <eric@eukrea.com> wrote:
> >> > Hi Daiane,
> >> >
> >> > Le Thu, 9 Jul 2015 11:54:14 -0300,
> >> > Daiane Angolini <daiane.list@gmail.com> a écrit :
> >> >> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
> >> >> think this list is fine. I don't see why to separate imx5 from imxs.
> >> >> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
> >> >> exactly like a imx3.
> >> >>
> >> > Does that mean there is an updated GPU library and full IPU drivers for
> >> > more recent kernels than 2.6.35 for i.MX53 ?
> >>
> >> This means mx5 will likely don't have GPU support in Yocto Project
> >> 1.9; VPU will likely work using mainline CODA.
> >>
> > and also most of the tools mixing VPU & IPU (CSI for example) ...
> > Is that what you want in a "reference" BSP for reference platforms ?
> 
> If Freescale does not upgrade the AMD support components for newer
> kernel someone needs to do it.
> 
> As a community member you are welcome to-do so, especially as Eukrea
> has mx5-based SoMs. If no one does it, it will be removed.
> 
So just remove it, no interest to spend time on that old chipset support
in 2015 especially when AMD's libraries are proprietary and when
Freescale didn't even bother to provide a hard float version despite
several requests.

Best regards
Eric

Re: SOC_FAMILY Rework

[Luo Zhenhua]

The OVERRIDE scenario is more complex for QorIQ than i.MX, it is hard to have a fixed definition for the SOC_FAMILY/OVERRIDE and levels, the definition depends on the actual needs and might be changes along with new boards involvement. 

To be specific for the SOC_FAMILY and OVERRIDE definition of QorIQ, the following is the basic OVERRIDE tree currently, we also have special cases besides of the following tree. E.g. e5500  core includes p50x0, t102x, t104x, some  packages and features(l2switch, auto-response) are specific to t104x series, so we need to define a t104x SOC family(qoriq -> qoriq-ppc -> qoriq-ppce6500 -> t104x). 

Basic SOC family definition
* qoriq -> qoriq-arm -> ls1021a 
* qoriq -> qoriq-arm -> ls104xa (not upstreamed)
* qoriq -> qoriq-arm -> ls108xa (not upstreamed)
* qoriq -> qoriq-arm -> ls208xa (not upstreamed)
* qoriq -> qoriq-ppc -> qoriq-ppce500v2
* qoriq -> qoriq-ppc -> qoriq-ppce500mc
* qoriq -> qoriq-ppc -> qoriq-ppce5500
* qoriq -> qoriq-ppc -> qoriq-ppc64e5500
* qoriq -> qoriq-ppc -> qoriq-ppce6500
* qoriq -> qoriq-ppc -> qoriq-ppc64e6500


Best Regards,

Zhenhua

> -----Original Message-----
> From: meta-freescale-bounces@yoctoproject.org [mailto:meta-freescale-
> bounces@yoctoproject.org] On Behalf Of Daiane Angolini
> Sent: Thursday, July 09, 2015 3:43 AM
> To: meta-freescale@yoctoproject.org
> Subject: [meta-freescale] SOC_FAMILY Rework
> 
> We have today the following SOC_FAMILY tree (for meta-fsl-arm)
> 
> i.MX -> mx3 -> mx31;
> i.MX -> mx3 -> mx35;
> i.MX -> mx5 -> mx51;
> i.MX -> mx5 -> mx53;
> i.MX -> mx6 -> mx6dl;
> i.MX -> mx6 -> mx6q;
> i.MX -> mx6 -> mx6sl;
> i.MX -> mx6 -> mx6sx;
> i.MX -> mxs -> mx23;
> i.MX -> mxs -> mx28;
> Layerscape -> ls102xa;
> Vybrid -> vf -> vf50;
> Vybrid -> vf -> vf60;
> 
> Where i.MX, Layerscape and Vybrid are only included for visual purposes, not
> being a real OVERRIDE today (the imx inclusion is a pending patch in ML right
> now)
> 
> Some of the points I think is important to consider when choosing a OVERRIDE
> tree is the SoC divergence and convergence. In i.MX case, we have some
> obvious examples such as GPU type, existence of VPU, and machine features
> such as CAN (although imx6 generally supports CAN, there are some board
> with connector, and other without it).
> 
> In CAN case, CAN is a SoC feature enabled by the board or not. In WIFI case, it’s
> only a board feature.
> 
> So, the points to rule the OVERRIDE is SoC features and board features.
> 
> Another important thing to keep in mind is the deepness. How deep should be
> the OVERRIDE tree? Two or tree levels?
> 
> OVERRIDE should be used to cluster BSP packages and configurations
> accordingly with SoC and machine features. Today the meta-fsl-arm BSP
> packages can be seen in the table [1]. Most of packages are the same for all
> i.MX boards (including kernel and u-boot not included in the table), but some
> are very specific to a group of SoCs only (i.e.
> fsl-alsa-plugins), and there is the case where a package is for only one SoC
> Family, but has big inner segmentation (Vivante)
> 
> [1] https://github.com/Freescale/Documentation/blob/master/release-
> notes/source/soc-pkg-optimization.inc
> 
> The classic example of a crazy OVERRIDE is ‘mxs’. It was included in the past to
> cluster imx23 and imx28 because they are the only 2 board using imx-bootlets.
> Does ‘mxs’ make sense nowaday? Would imx6UL be more related with imx28
> than with imx6Q? Would imx6SX be clustered with vf60?
> 
> Maybe, we can concentrate the OVERRIDE only around GPU and VPU. We have
> an extensive segmentation of OVERRIDE for imx6, and a cluster for all the SoCs
> without GPU (vf,layerscape?,imx2,imx3,imx6UL, imx7)
> 
> I only thing this is the right timing to review and rework SOC_FAMILY.
> I'm i.MX biased, so I can only say about i.MX, would be glad to hear feedback
> from Vybrid and QoirQ.
> 
> Regards,
> Daiane
> --
> _______________________________________________
> meta-freescale mailing list
> meta-freescale@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-freescale

Re: SOC_FAMILY Rework

[Daiane Angolini]

On Thu, Jul 9, 2015 at 12:17 PM, Otavio Salvador
<otavio@ossystems.com.br> wrote:
> On Thu, Jul 9, 2015 at 11:54 AM, Daiane Angolini <daiane.list@gmail.com> wrote:
>> On Thu, Jul 9, 2015 at 11:05 AM, Otavio Salvador
>> <otavio@ossystems.com.br> wrote:
>>> Hello Daiane,
>>>
>>> First, thank you for preparing this e-mail.
>>>
>>> On Wed, Jul 8, 2015 at 4:42 PM, Daiane Angolini <daiane.list@gmail.com> wrote:
>>>> We have today the following SOC_FAMILY tree (for meta-fsl-arm)
>>>>
>>>> i.MX -> mx3 -> mx31;
>>>> i.MX -> mx3 -> mx35;
>>>> i.MX -> mx5 -> mx51;
>>>> i.MX -> mx5 -> mx53;
>>>> i.MX -> mx6 -> mx6dl;
>>>> i.MX -> mx6 -> mx6q;
>>>> i.MX -> mx6 -> mx6sl;
>>>> i.MX -> mx6 -> mx6sx;
>>>> i.MX -> mxs -> mx23;
>>>> i.MX -> mxs -> mx28;
>>>
>>> Those are fine, as far as I can see.
>>
>> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
>> think this list is fine. I don't see why to separate imx5 from imxs.
>> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
>> exactly like a imx3.
>
> Yes; when it goes than mx5 will be in mainline use as will be mx28
> (mx23 is already using it).
>
> However this is a new environment and when this happens we can rework
> this. I am not sure we can kill the SoC families as it is needed for
> U-Boot and other things internally as well (see mxs and imx-base
> files).

Only to be extra clear on my words. I'm talking about a future change,
mainly if meta-freescale takes place. I don't think we should rework
SOC_FAMILY in a normal meta-fsl-arm development cycle.


>
> ...
>>>> Another important thing to keep in mind is the deepness. How deep
>>>> should be the OVERRIDE tree? Two or tree levels?
>>>
>>> Not a big deal; it depends if it makes usage sense or not. Current set
>>> seems good for me.
>>
>> In fact I think it's a big deal.
>>
>> Maybe there is not so much difference between 2 and 3, but if you
>> think about 30 levels it is indeed a big deal.
>>
>> I understand the number of levels is arbitrary. But it only means we
>> must decide which are the prefect deepness.
>
> Every modularization imposes complexity so I think usefulness is the
> metric to be put in place here.
>
>>>> OVERRIDE should be used to cluster BSP packages and configurations
>>>> accordingly with SoC and machine features. Today the meta-fsl-arm BSP
>>>> packages can be seen in the table [1]. Most of packages are the same
>>>> for all i.MX boards (including kernel and u-boot not included in the
>>>> table), but some are very specific to a group of SoCs only (i.e.
>>>> fsl-alsa-plugins), and there is the case where a package is for only
>>>> one SoC Family, but has big inner segmentation (Vivante)
>>>>
>>>> [1] https://github.com/Freescale/Documentation/blob/master/release-notes/source/soc-pkg-optimization.inc
>>>
>>> Agreed.
>>>
>>>> The classic example of a crazy OVERRIDE is ‘mxs’. It was included in
>>>> the past to cluster imx23 and imx28 because they are the only 2 board
>>>> using imx-bootlets. Does ‘mxs’ make sense nowaday? Would imx6UL be
>>>> more related with imx28 than with imx6Q? Would imx6SX be clustered
>>>> with vf60?
>>>
>>> I think UL will end being a mx6ul and it will demand the check of all
>>> mx6 uses to see if it is compatible or not. See Lauren's patch
>>> regarding the GPU presence in Weston for an example.
>>>
>>>> Maybe, we can concentrate the OVERRIDE only around GPU and VPU. We
>>>> have an extensive segmentation of OVERRIDE for imx6, and a cluster for
>>>> all the SoCs without GPU (vf,layerscape?,imx2,imx3,imx6UL, imx7)
>>>
>>> It does not work. The point of OVERRIDE is to make changes in a set of
>>> SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
>>> Qualcomm...)
>>
>> GPU is a SoC characteristic, and it's obvious it's the most important
>> BSP piece on i.MX6 family. Instead of having a tree like:
>>
>> imx -> imx6 -> everysingle soc
>>
>> We could have something like:
>>
>> imx -> imx6-light -> ligh socs
>> imx -> imx6-heavy -> heavy socs
>>
>> and, instead of list "everysingle soc" when setting GPU BSP, we can
>> list only "light and heavy".
>>
>> The downside is to list "light and heavy" in packages for imx6
>
> I don't agree; we don't know what i.MX8 will look like, what i.MX9
> will look like so we should put it as close as possible for product
> family.

We never know the future. So the argument "we don't know the future"
is not good enough.

But I understand what you mean instead. You mean imx6 today is the
exception, not the rule. In case another exception (imx8 or imx9)
this can become the rule in the future.


Daiane
>
> The feature set is too fragmented and too SoC specific so grouping it
> in super-groups is risky.
>
>> I don't see how it will contaminate other BSP. I'm trying to stress
>> the imx OVERRIDE just included.
>
> IF we use the overrides wrong it does.
>
> There is an arm override, we cannot use this for an i.MX patch as TI
> would be affected, for example.And as an exception it should be cared
>
>>>> I only thing this is the right timing to review and rework SOC_FAMILY.
>>>> I'm i.MX biased, so I can only say about i.MX, would be glad to hear
>>>> feedback from Vybrid and QoirQ.
>>>
>>> As I said before, the QorIQ should have common and arch specific
>>> overrides (I see the former going to be most used) and i.MX global
>>> override will be rarely used.
>>
>> One argument to keep things like it is today is learning curve
>> one argument to change it in meta-freescale is upgrade path
>
> I think Vybrid and Layerscape does need rework. mx5 and mxs may be
> reduced once we move to mainline  but not killed and they are not same
> thing as well.
>
> --
> Otavio Salvador                             O.S. Systems
> http://www.ossystems.com.br        http://code.ossystems.com.br
> Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: SOC_FAMILY Rework

[Daiane Angolini]

On Thu, Jul 9, 2015 at 12:18 PM, Eric Bénard <eric@eukrea.com> wrote:
> Hi Daiane,
>
> Le Thu, 9 Jul 2015 11:54:14 -0300,
> Daiane Angolini <daiane.list@gmail.com> a écrit :
>> I'm sorry, Otavio, but you are not re-thinking what we have. I don't
>> think this list is fine. I don't see why to separate imx5 from imxs.
>> And, after GCC 5.X is up and linux-imx_2.6.35 is gone, imx5 will be
>> exactly like a imx3.
>>
> Does that mean there is an updated GPU library and full IPU drivers for
> more recent kernels than 2.6.35 for i.MX53 ?

Not that I know, Eric.

The topic about i.MX53 become mailine was discussed in last community
meeting, and we was not able to figure out a better option. At least I
personally cannot take this task.

If needed we can setup another meeting or open a new thread on this topic.

Daiane

>
> Thanks,
> Eric

Re: SOC_FAMILY Rework

[Daiane Angolini]

On Thu, Jul 9, 2015 at 12:20 PM, Nikolay Dimitrov <picmaster@mail.bg> wrote:
> Hi Daiane,
>
> On 07/09/2015 05:54 PM, Daiane Angolini wrote:
>>>
>>> It does not work. The point of OVERRIDE is to make changes in a set of
>>> SoCs and we cannot contaminate other BSP (Intel, TI, Samsung,
>>> Qualcomm...)
>>
>>
>> GPU is a SoC characteristic, and it's obvious it's the most important
>> BSP piece on i.MX6 family. Instead of having a tree like:
>>
>> imx -> imx6 -> everysingle soc
>>
>> We could have something like:
>>
>> imx -> imx6-light -> ligh socs
>> imx -> imx6-heavy -> heavy socs
>>
>> and, instead of list "everysingle soc" when setting GPU BSP, we can
>> list only "light and heavy".
>>
>> The downside is to list "light and heavy" in packages for imx6
>>
>>
>> I don't see how it will contaminate other BSP. I'm trying to stress
>> the imx OVERRIDE just included.
>
>
> This is an excellent example. The OVERRIDEs should generalize, not
> specialize (although you can always interpret a root-leaf traversal as
> a specialization :D). Detailed specialization will lead to OVERRIDEs
> tree explosion while trying to express every single SOC feature on each
> level of the tree.

The problem is that the i.MX6 shares only the name and the core. And
share the core is nothing (as i.MX can share the ARM core of any other
product in the world)

It's like i.MX6UL was the adopted brother. It shares the name but no
genetic features.

Of course we can use the marketing name on OVERRIDE, but I would
prefer if we focus on the SoC convergence instead.


>
> If we try to express SOC features via OVERRIDE mechanism, it's imho
> doomed.

I have a crush on philosophy, so I apologize in advance. But i.MX is
basicaly it's GPU. Mainly if you think about i.MX6. Represent GPU
characteristics on OVERRIDE is not represent a feature, but a core
feature.

However, I remember Otavio's argument i.MX6 is the exception now, not the rule.

Daiane
>
> Regards,
> Nikolay