mbox series

[00/17] ARM: dts: dra7/am57xx: remoteproc support

Message ID 20200424151244.3225-1-t-kristo@ti.com (mailing list archive)
Headers show
Series ARM: dts: dra7/am57xx: remoteproc support | expand

Message

Tero Kristo April 24, 2020, 3:12 p.m. UTC
Hi Tony,

This series adds the DT nodes necessary for remoteproc support, now that
the driver side changes are (mostly) in. Couple of things to note
though.

1) There is a new IOMMU issue, for which I posted a fix today [1]
2) The remoteproc core still has an issue for which there is ongoing
   discussion [2]

With these two issue taken care of, the omap remoteproc support is
functional. The question though is, whether we should just wait until
the above two issues are resolved and merge the DT patches post that, or
merge the DT patches with status = "disabled".

There aren't any boot failures without the mentioned two issues though,
as one needs to enable the RPMSG_VIRTIO support before the failures
really kick in (issue [2]), and this config is not enabled for OMAPs
yet. Also, multi-v7 config doesn't seem to enable omap remoteproc,
so that is safe also.

Another thing I was considering myself was to squash all the board
specific reserved-memory region patches into the
dra7-ipu-dsp-common.dtsi files. However Suman wants to have these
separate and as he is the actual author for these, I posted them in this
form. But anyway, just so you know it would be possible to merge them
together.

-Tero

[1] https://lore.kernel.org/linux-iommu/20200424145828.3159-1-t-kristo@ti.com/T/#u
[2] https://lkml.org/lkml/2020/4/20/1094



--
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Comments

Tony Lindgren April 24, 2020, 3:51 p.m. UTC | #1
* Tero Kristo <t-kristo@ti.com> [200424 08:13]:
> Hi Tony,
> 
> This series adds the DT nodes necessary for remoteproc support, now that
> the driver side changes are (mostly) in. Couple of things to note
> though.
> 
> 1) There is a new IOMMU issue, for which I posted a fix today [1]
> 2) The remoteproc core still has an issue for which there is ongoing
>    discussion [2]
> 
> With these two issue taken care of, the omap remoteproc support is
> functional. The question though is, whether we should just wait until
> the above two issues are resolved and merge the DT patches post that, or
> merge the DT patches with status = "disabled".

If there are no dependencies between the pending driver fixes and
the dts changes I see no reason to not merge the dts changes.

> There aren't any boot failures without the mentioned two issues though,
> as one needs to enable the RPMSG_VIRTIO support before the failures
> really kick in (issue [2]), and this config is not enabled for OMAPs
> yet. Also, multi-v7 config doesn't seem to enable omap remoteproc,
> so that is safe also.

OK thanks for checking that.

> Another thing I was considering myself was to squash all the board
> specific reserved-memory region patches into the
> dra7-ipu-dsp-common.dtsi files. However Suman wants to have these
> separate and as he is the actual author for these, I posted them in this
> form. But anyway, just so you know it would be possible to merge them
> together.

OK. The combining of common features can be done in later patches
too.

Regards,

Tony

> [1] https://lore.kernel.org/linux-iommu/20200424145828.3159-1-t-kristo@ti.com/T/#u
> [2] https://lkml.org/lkml/2020/4/20/1094
Tero Kristo April 24, 2020, 3:54 p.m. UTC | #2
On 24/04/2020 18:51, Tony Lindgren wrote:
> * Tero Kristo <t-kristo@ti.com> [200424 08:13]:
>> Hi Tony,
>>
>> This series adds the DT nodes necessary for remoteproc support, now that
>> the driver side changes are (mostly) in. Couple of things to note
>> though.
>>
>> 1) There is a new IOMMU issue, for which I posted a fix today [1]
>> 2) The remoteproc core still has an issue for which there is ongoing
>>     discussion [2]
>>
>> With these two issue taken care of, the omap remoteproc support is
>> functional. The question though is, whether we should just wait until
>> the above two issues are resolved and merge the DT patches post that, or
>> merge the DT patches with status = "disabled".
> 
> If there are no dependencies between the pending driver fixes and
> the dts changes I see no reason to not merge the dts changes.

Yeah, no hard dependencies as such, just that things won't work properly 
before they are in.

-Tero

> 
>> There aren't any boot failures without the mentioned two issues though,
>> as one needs to enable the RPMSG_VIRTIO support before the failures
>> really kick in (issue [2]), and this config is not enabled for OMAPs
>> yet. Also, multi-v7 config doesn't seem to enable omap remoteproc,
>> so that is safe also.
> 
> OK thanks for checking that.
> 
>> Another thing I was considering myself was to squash all the board
>> specific reserved-memory region patches into the
>> dra7-ipu-dsp-common.dtsi files. However Suman wants to have these
>> separate and as he is the actual author for these, I posted them in this
>> form. But anyway, just so you know it would be possible to merge them
>> together.
> 
> OK. The combining of common features can be done in later patches
> too.
> 
> Regards,
> 
> Tony
> 
>> [1] https://lore.kernel.org/linux-iommu/20200424145828.3159-1-t-kristo@ti.com/T/#u
>> [2] https://lkml.org/lkml/2020/4/20/1094

--
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
Tony Lindgren May 5, 2020, 6:14 p.m. UTC | #3
* Tero Kristo <t-kristo@ti.com> [200424 15:55]:
> On 24/04/2020 18:51, Tony Lindgren wrote:
> > * Tero Kristo <t-kristo@ti.com> [200424 08:13]:
> > > Hi Tony,
> > > 
> > > This series adds the DT nodes necessary for remoteproc support, now that
> > > the driver side changes are (mostly) in. Couple of things to note
> > > though.
> > > 
> > > 1) There is a new IOMMU issue, for which I posted a fix today [1]
> > > 2) The remoteproc core still has an issue for which there is ongoing
> > >     discussion [2]
> > > 
> > > With these two issue taken care of, the omap remoteproc support is
> > > functional. The question though is, whether we should just wait until
> > > the above two issues are resolved and merge the DT patches post that, or
> > > merge the DT patches with status = "disabled".
> > 
> > If there are no dependencies between the pending driver fixes and
> > the dts changes I see no reason to not merge the dts changes.
> 
> Yeah, no hard dependencies as such, just that things won't work properly
> before they are in.

Applying these all into omap-for-v5.8/dt thanks.

Tony