mbox series

[0/3] Drop PRUSS and OMAP4 IPU/DSP hwmod data

Message ID 20200224232152.25562-1-s-anna@ti.com (mailing list archive)
Headers show
Series Drop PRUSS and OMAP4 IPU/DSP hwmod data | expand

Message

Suman Anna Feb. 24, 2020, 11:21 p.m. UTC 
Hi Tony,

The following patches drop the hwmod data for PRUSS on AM33xx and AM437x
SoCs, and for the IPU and DSP processors on OMAP4 SoC. Patches are based
on 5.6-rc1. Please consider these for the 5.7 merge window.

I will be submitting another series tomorrow to add the ti-sysc support
for PRUSS. Nevertheless, the PRUSS hwmods will not be used going forward
and can be dropped independently.

The IPU and DSP hwmods were never added for OMAP5 and DRA7xx/AM57xx SoCs,
and the MMU data was already dropped for 5.6-rc1, as it has all been
converted to ti-sysc and omap-prm. The DT nodes for these will follow
for the next merge window once the current OMAP remoteproc DT support
is accepted [1].

regards
Suman

[1] https://patchwork.kernel.org/cover/11396049/


Suman Anna (3):
  ARM: OMAP2+: Drop hwmod data for am3 and am4 PRUSS
  ARM: OMAP2+: Drop legacy platform data for OMAP4 DSP
  ARM: OMAP4: hwmod_data: Remove OMAP4 IPU hwmod data

 arch/arm/boot/dts/omap4.dtsi                  |   1 -
 .../omap_hwmod_33xx_43xx_common_data.h        |   2 -
 .../omap_hwmod_33xx_43xx_interconnect_data.c  |   8 --
 .../omap_hwmod_33xx_43xx_ipblock_data.c       |  33 -----
 arch/arm/mach-omap2/omap_hwmod_33xx_data.c    |  10 --
 arch/arm/mach-omap2/omap_hwmod_43xx_data.c    |   9 --
 arch/arm/mach-omap2/omap_hwmod_44xx_data.c    | 127 ------------------
 7 files changed, 190 deletions(-)

Comments

Tony Lindgren Feb. 26, 2020, 6:26 p.m. UTC | #1 
* Suman Anna <s-anna@ti.com> [200224 15:22]:
> Hi Tony,
> 
> The following patches drop the hwmod data for PRUSS on AM33xx and AM437x
> SoCs, and for the IPU and DSP processors on OMAP4 SoC. Patches are based
> on 5.6-rc1. Please consider these for the 5.7 merge window.
> 
> I will be submitting another series tomorrow to add the ti-sysc support
> for PRUSS. Nevertheless, the PRUSS hwmods will not be used going forward
> and can be dropped independently.
> 
> The IPU and DSP hwmods were never added for OMAP5 and DRA7xx/AM57xx SoCs,
> and the MMU data was already dropped for 5.6-rc1, as it has all been
> converted to ti-sysc and omap-prm. The DT nodes for these will follow
> for the next merge window once the current OMAP remoteproc DT support
> is accepted [1].

Well we still need the hwmod data until dts data is in place to reset
and idle these modules. So I'll wait with this set until we have the
dts changes in place.

Regards,

Tony
Suman Anna Feb. 26, 2020, 8:30 p.m. UTC | #2 
On 2/26/20 12:26 PM, Tony Lindgren wrote:
> * Suman Anna <s-anna@ti.com> [200224 15:22]:
>> Hi Tony,
>>
>> The following patches drop the hwmod data for PRUSS on AM33xx and AM437x
>> SoCs, and for the IPU and DSP processors on OMAP4 SoC. Patches are based
>> on 5.6-rc1. Please consider these for the 5.7 merge window.
>>
>> I will be submitting another series tomorrow to add the ti-sysc support
>> for PRUSS. Nevertheless, the PRUSS hwmods will not be used going forward
>> and can be dropped independently.
>>
>> The IPU and DSP hwmods were never added for OMAP5 and DRA7xx/AM57xx SoCs,
>> and the MMU data was already dropped for 5.6-rc1, as it has all been
>> converted to ti-sysc and omap-prm. The DT nodes for these will follow
>> for the next merge window once the current OMAP remoteproc DT support
>> is accepted [1].
> 
> Well we still need the hwmod data until dts data is in place to reset
> and idle these modules. So I'll wait with this set until we have the
> dts changes in place.

Not really, you cannot just idle these processors devices by themselves
without the cores running. The hwmod code actually keeps these in the
default reset asserted state itself, so no harm done in dropping the
data. The remoteproc bindings are acked, so for next merge window, we
can post the dts nodes once those get merged.

regards
Suman
Tony Lindgren Feb. 27, 2020, 1:27 a.m. UTC | #3 
* Suman Anna <s-anna@ti.com> [200226 20:31]:
> On 2/26/20 12:26 PM, Tony Lindgren wrote:
> > * Suman Anna <s-anna@ti.com> [200224 15:22]:
> >> Hi Tony,
> >>
> >> The following patches drop the hwmod data for PRUSS on AM33xx and AM437x
> >> SoCs, and for the IPU and DSP processors on OMAP4 SoC. Patches are based
> >> on 5.6-rc1. Please consider these for the 5.7 merge window.
> >>
> >> I will be submitting another series tomorrow to add the ti-sysc support
> >> for PRUSS. Nevertheless, the PRUSS hwmods will not be used going forward
> >> and can be dropped independently.
> >>
> >> The IPU and DSP hwmods were never added for OMAP5 and DRA7xx/AM57xx SoCs,
> >> and the MMU data was already dropped for 5.6-rc1, as it has all been
> >> converted to ti-sysc and omap-prm. The DT nodes for these will follow
> >> for the next merge window once the current OMAP remoteproc DT support
> >> is accepted [1].
> > 
> > Well we still need the hwmod data until dts data is in place to reset
> > and idle these modules. So I'll wait with this set until we have the
> > dts changes in place.
> 
> Not really, you cannot just idle these processors devices by themselves
> without the cores running. The hwmod code actually keeps these in the
> default reset asserted state itself, so no harm done in dropping the
> data. The remoteproc bindings are acked, so for next merge window, we
> can post the dts nodes once those get merged.

OK if they are in reset from the bootloader.. I'll try to check.

Regards,

Tony
Suman Anna Feb. 27, 2020, 5:40 p.m. UTC | #4 
On 2/26/20 7:27 PM, Tony Lindgren wrote:
> * Suman Anna <s-anna@ti.com> [200226 20:31]:
>> On 2/26/20 12:26 PM, Tony Lindgren wrote:
>>> * Suman Anna <s-anna@ti.com> [200224 15:22]:
>>>> Hi Tony,
>>>>
>>>> The following patches drop the hwmod data for PRUSS on AM33xx and AM437x
>>>> SoCs, and for the IPU and DSP processors on OMAP4 SoC. Patches are based
>>>> on 5.6-rc1. Please consider these for the 5.7 merge window.
>>>>
>>>> I will be submitting another series tomorrow to add the ti-sysc support
>>>> for PRUSS. Nevertheless, the PRUSS hwmods will not be used going forward
>>>> and can be dropped independently.
>>>>
>>>> The IPU and DSP hwmods were never added for OMAP5 and DRA7xx/AM57xx SoCs,
>>>> and the MMU data was already dropped for 5.6-rc1, as it has all been
>>>> converted to ti-sysc and omap-prm. The DT nodes for these will follow
>>>> for the next merge window once the current OMAP remoteproc DT support
>>>> is accepted [1].
>>>
>>> Well we still need the hwmod data until dts data is in place to reset
>>> and idle these modules. So I'll wait with this set until we have the
>>> dts changes in place.
>>
>> Not really, you cannot just idle these processors devices by themselves
>> without the cores running. The hwmod code actually keeps these in the
>> default reset asserted state itself, so no harm done in dropping the
>> data. The remoteproc bindings are acked, so for next merge window, we
>> can post the dts nodes once those get merged.
> 
> OK if they are in reset from the bootloader.. I'll try to check.

OK thanks. FWIW, we have never added the hwmods for IPUs and DSPs on
OMAP5 and DRA7xx/AM57xx, and most of them are pseudo hwmods anyway for
reset purposes, and do not get exercised until the remoteproc driver
would have specifically requested for them.

regards
Suman
Tony Lindgren Feb. 27, 2020, 6:03 p.m. UTC | #5 
* Suman Anna <s-anna@ti.com> [200227 17:40]:
> On 2/26/20 7:27 PM, Tony Lindgren wrote:
> > * Suman Anna <s-anna@ti.com> [200226 20:31]:
> >> On 2/26/20 12:26 PM, Tony Lindgren wrote:
> >>> * Suman Anna <s-anna@ti.com> [200224 15:22]:
> >>>> Hi Tony,
> >>>>
> >>>> The following patches drop the hwmod data for PRUSS on AM33xx and AM437x
> >>>> SoCs, and for the IPU and DSP processors on OMAP4 SoC. Patches are based
> >>>> on 5.6-rc1. Please consider these for the 5.7 merge window.
> >>>>
> >>>> I will be submitting another series tomorrow to add the ti-sysc support
> >>>> for PRUSS. Nevertheless, the PRUSS hwmods will not be used going forward
> >>>> and can be dropped independently.
> >>>>
> >>>> The IPU and DSP hwmods were never added for OMAP5 and DRA7xx/AM57xx SoCs,
> >>>> and the MMU data was already dropped for 5.6-rc1, as it has all been
> >>>> converted to ti-sysc and omap-prm. The DT nodes for these will follow
> >>>> for the next merge window once the current OMAP remoteproc DT support
> >>>> is accepted [1].
> >>>
> >>> Well we still need the hwmod data until dts data is in place to reset
> >>> and idle these modules. So I'll wait with this set until we have the
> >>> dts changes in place.
> >>
> >> Not really, you cannot just idle these processors devices by themselves
> >> without the cores running. The hwmod code actually keeps these in the
> >> default reset asserted state itself, so no harm done in dropping the
> >> data. The remoteproc bindings are acked, so for next merge window, we
> >> can post the dts nodes once those get merged.
> > 
> > OK if they are in reset from the bootloader.. I'll try to check.
> 
> OK thanks. FWIW, we have never added the hwmods for IPUs and DSPs on
> OMAP5 and DRA7xx/AM57xx, and most of them are pseudo hwmods anyway for
> reset purposes, and do not get exercised until the remoteproc driver
> would have specifically requested for them.

OK yeah and for omap4 they seem to be in reset like you said.
Applying these into omap-for-v5.7/accelerators thanks.

Tony