Message ID | 20230207151300.3223-1-sinthu.raja@ti.com |
---|---|
State | Accepted |
Delegated to: | Ryan Eatmon |
Headers | show |
Series | [meta-ti,dunfell,V2] conf: machine: j721s2: Add ehrpwm overlay for am68-sk | expand |
On 2/7/2023 9:13, Sinthu Raja wrote: > From: Sinthu Raja <sinthu.raja@ti.com> > > The AM68 SK has PWM output pins connected to its RPi headers. Add > overlay file to enable PWM pins. > > Signed-off-by: Sinthu Raja <sinthu.raja@ti.com> > --- > > Changes in v2: > ============== > Address review comment: > - Remove extra space > > v1:https://lists.yoctoproject.org/g/meta-ti/message/15730 > > conf/machine/j721s2-evm.conf | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/conf/machine/j721s2-evm.conf b/conf/machine/j721s2-evm.conf > index 7c120dcc..f1127ed5 100644 > --- a/conf/machine/j721s2-evm.conf > +++ b/conf/machine/j721s2-evm.conf > @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ > ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ > ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ > ti/k3-am68-sk-fpdlink-fusion.dtbo \ > + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ > ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ > ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ > ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \ This file does not exists in the kernel (yet?). I'm assuming that you all have put in the patch for it and we are just playing a waiting game? If that is the case, then I would appreciate (in the future) not submitting these patches to the meta-ti mailing list until the kernel .dtb changes have been accepted into the kernel. This is the second patch in recent weeks that I'm having to sit on until the kernel catches up... It makes it really hard to test these patches if the kernel part is not there, and I do not have the bandwidth to keep checking daily if the change has made it into the public/cicd branches.
On 2/8/2023 10:37 AM, Ryan Eatmon via lists.yoctoproject.org wrote: > > > On 2/7/2023 9:13, Sinthu Raja wrote: >> From: Sinthu Raja <sinthu.raja@ti.com> >> >> The AM68 SK has PWM output pins connected to its RPi headers. Add >> overlay file to enable PWM pins. >> >> Signed-off-by: Sinthu Raja <sinthu.raja@ti.com> >> --- >> >> Changes in v2: >> ============== >> Address review comment: >> - Remove extra space >> >> v1:https://lists.yoctoproject.org/g/meta-ti/message/15730 >> conf/machine/j721s2-evm.conf | 1 + >> 1 file changed, 1 insertion(+) >> >> diff --git a/conf/machine/j721s2-evm.conf b/conf/machine/j721s2-evm.conf >> index 7c120dcc..f1127ed5 100644 >> --- a/conf/machine/j721s2-evm.conf >> +++ b/conf/machine/j721s2-evm.conf >> @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ >> ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ >> ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ >> ti/k3-am68-sk-fpdlink-fusion.dtbo \ >> + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ >> ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ >> ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ >> ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \ > > > This file does not exists in the kernel (yet?). I'm assuming that you > all have put in the patch for it and we are just playing a waiting game? > > If that is the case, then I would appreciate (in the future) not > submitting these patches to the meta-ti mailing list until the kernel > .dtb changes have been accepted into the kernel. This is the second > patch in recent weeks that I'm having to sit on until the kernel catches > up... > > It makes it really hard to test these patches if the kernel part is not > there, and I do not have the bandwidth to keep checking daily if the > change has made it into the public/cicd branches. > Agreed here. Sinthu, Please post the associated meta-ti change only once the kernel dts change is merged to cicd build. from what i see, the kernel patch is still under review. > > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#15750): https://lists.yoctoproject.org/g/meta-ti/message/15750 > Mute This Topic: https://lists.yoctoproject.org/mt/96808923/2167429 > Group Owner: meta-ti+owner@lists.yoctoproject.org > Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/leave/6695566/2167429/1935739249/xyzzy [praneeth@ti.com] > -=-=-=-=-=-=-=-=-=-=-=- >
On Wed, Feb 8, 2023 at 10:31 PM Bajjuri, Praneeth <praneeth@ti.com> wrote: > > > > On 2/8/2023 10:37 AM, Ryan Eatmon via lists.yoctoproject.org wrote: > > > > > > On 2/7/2023 9:13, Sinthu Raja wrote: > >> From: Sinthu Raja <sinthu.raja@ti.com> > >> > >> The AM68 SK has PWM output pins connected to its RPi headers. Add > >> overlay file to enable PWM pins. > >> > >> Signed-off-by: Sinthu Raja <sinthu.raja@ti.com> > >> --- > >> > >> Changes in v2: > >> ============== > >> Address review comment: > >> - Remove extra space > >> > >> v1:https://lists.yoctoproject.org/g/meta-ti/message/15730 > >> conf/machine/j721s2-evm.conf | 1 + > >> 1 file changed, 1 insertion(+) > >> > >> diff --git a/conf/machine/j721s2-evm.conf b/conf/machine/j721s2-evm.conf > >> index 7c120dcc..f1127ed5 100644 > >> --- a/conf/machine/j721s2-evm.conf > >> +++ b/conf/machine/j721s2-evm.conf > >> @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ > >> ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ > >> ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ > >> ti/k3-am68-sk-fpdlink-fusion.dtbo \ > >> + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ > >> ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ > >> ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ > >> ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \ > > > > > > This file does not exists in the kernel (yet?). I'm assuming that you > > all have put in the patch for it and we are just playing a waiting game? > > > > If that is the case, then I would appreciate (in the future) not > > submitting these patches to the meta-ti mailing list until the kernel > > .dtb changes have been accepted into the kernel. This is the second > > patch in recent weeks that I'm having to sit on until the kernel catches > > up... > > > > It makes it really hard to test these patches if the kernel part is not > > there, and I do not have the bandwidth to keep checking daily if the > > change has made it into the public/cicd branches. > > > > Agreed here. > Sinthu, Please post the associated meta-ti change only once the kernel > dts change is merged to cicd build. > Agreed. > from what i see, the kernel patch is still under review. > > > > > > > > > -=-=-=-=-=-=-=-=-=-=-=- > > Links: You receive all messages sent to this group. > > View/Reply Online (#15750): https://lists.yoctoproject.org/g/meta-ti/message/15750 > > Mute This Topic: https://lists.yoctoproject.org/mt/96808923/2167429 > > Group Owner: meta-ti+owner@lists.yoctoproject.org > > Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/leave/6695566/2167429/1935739249/xyzzy [praneeth@ti.com] > > -=-=-=-=-=-=-=-=-=-=-=- > >
On 2/8/23 11:08 AM, Sinthu Raja M via lists.yoctoproject.org wrote: > On Wed, Feb 8, 2023 at 10:31 PM Bajjuri, Praneeth <praneeth@ti.com> wrote: >> >> >> >> On 2/8/2023 10:37 AM, Ryan Eatmon via lists.yoctoproject.org wrote: >>> >>> >>> On 2/7/2023 9:13, Sinthu Raja wrote: >>>> From: Sinthu Raja <sinthu.raja@ti.com> >>>> >>>> The AM68 SK has PWM output pins connected to its RPi headers. Add >>>> overlay file to enable PWM pins. >>>> >>>> Signed-off-by: Sinthu Raja <sinthu.raja@ti.com> >>>> --- >>>> >>>> Changes in v2: >>>> ============== >>>> Address review comment: >>>> - Remove extra space >>>> >>>> v1:https://lists.yoctoproject.org/g/meta-ti/message/15730 >>>> conf/machine/j721s2-evm.conf | 1 + >>>> 1 file changed, 1 insertion(+) >>>> >>>> diff --git a/conf/machine/j721s2-evm.conf b/conf/machine/j721s2-evm.conf >>>> index 7c120dcc..f1127ed5 100644 >>>> --- a/conf/machine/j721s2-evm.conf >>>> +++ b/conf/machine/j721s2-evm.conf >>>> @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ >>>> ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ >>>> ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ >>>> ti/k3-am68-sk-fpdlink-fusion.dtbo \ >>>> + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ >>>> ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ >>>> ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ >>>> ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \ >>> >>> >>> This file does not exists in the kernel (yet?). I'm assuming that you >>> all have put in the patch for it and we are just playing a waiting game? >>> >>> If that is the case, then I would appreciate (in the future) not >>> submitting these patches to the meta-ti mailing list until the kernel >>> .dtb changes have been accepted into the kernel. This is the second >>> patch in recent weeks that I'm having to sit on until the kernel catches >>> up... >>> >>> It makes it really hard to test these patches if the kernel part is not >>> there, and I do not have the bandwidth to keep checking daily if the >>> change has made it into the public/cicd branches. >>> >> >> Agreed here. >> Sinthu, Please post the associated meta-ti change only once the kernel >> dts change is merged to cicd build. >> > Agreed. > I'm going to slightly disagree, I don't see an issue with getting some early reviews. Just as long as you only post it as an "RFC" and only post the real "PATCH" when ready. I'd also like to poke again the idea of pulling in DTB/DTBO files based on some pattern matching instead of manually adding each time. I know the idea got shot down, but I think this is going to be a growing problem as we start to increase our usage of DT Overlays in the near future.. Andrew >> from what i see, the kernel patch is still under review. >> >>> >>> >>> >>> >>> > > > > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#15752): https://lists.yoctoproject.org/g/meta-ti/message/15752 > Mute This Topic: https://lists.yoctoproject.org/mt/96808923/3619733 > Group Owner: meta-ti+owner@lists.yoctoproject.org > Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/unsub [afd@ti.com] > -=-=-=-=-=-=-=-=-=-=-=- >
On 2/8/2023 12:28, Andrew Davis wrote: > On 2/8/23 11:08 AM, Sinthu Raja M via lists.yoctoproject.org wrote: >> On Wed, Feb 8, 2023 at 10:31 PM Bajjuri, Praneeth <praneeth@ti.com> >> wrote: >>> >>> >>> >>> On 2/8/2023 10:37 AM, Ryan Eatmon via lists.yoctoproject.org wrote: >>>> >>>> >>>> On 2/7/2023 9:13, Sinthu Raja wrote: >>>>> From: Sinthu Raja <sinthu.raja@ti.com> >>>>> >>>>> The AM68 SK has PWM output pins connected to its RPi headers. Add >>>>> overlay file to enable PWM pins. >>>>> >>>>> Signed-off-by: Sinthu Raja <sinthu.raja@ti.com> >>>>> --- >>>>> >>>>> Changes in v2: >>>>> ============== >>>>> Address review comment: >>>>> - Remove extra space >>>>> >>>>> v1:https://lists.yoctoproject.org/g/meta-ti/message/15730 >>>>> conf/machine/j721s2-evm.conf | 1 + >>>>> 1 file changed, 1 insertion(+) >>>>> >>>>> diff --git a/conf/machine/j721s2-evm.conf >>>>> b/conf/machine/j721s2-evm.conf >>>>> index 7c120dcc..f1127ed5 100644 >>>>> --- a/conf/machine/j721s2-evm.conf >>>>> +++ b/conf/machine/j721s2-evm.conf >>>>> @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ >>>>> ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ >>>>> ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ >>>>> ti/k3-am68-sk-fpdlink-fusion.dtbo \ >>>>> + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ >>>>> ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ >>>>> ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ >>>>> ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \ >>>> >>>> >>>> This file does not exists in the kernel (yet?). I'm assuming that you >>>> all have put in the patch for it and we are just playing a waiting >>>> game? >>>> >>>> If that is the case, then I would appreciate (in the future) not >>>> submitting these patches to the meta-ti mailing list until the kernel >>>> .dtb changes have been accepted into the kernel. This is the second >>>> patch in recent weeks that I'm having to sit on until the kernel >>>> catches >>>> up... >>>> >>>> It makes it really hard to test these patches if the kernel part is not >>>> there, and I do not have the bandwidth to keep checking daily if the >>>> change has made it into the public/cicd branches. >>>> >>> >>> Agreed here. >>> Sinthu, Please post the associated meta-ti change only once the kernel >>> dts change is merged to cicd build. >>> >> Agreed. >> > > I'm going to slightly disagree, I don't see an issue with getting some > early reviews. Just as long as you only post it as an "RFC" and only > post the real "PATCH" when ready. If the patch is doing more than just adding a dtb, or changing a SRCREV, or anything that depends on an external repo to be updated then I would agree. Posting the change under the RFC banner (and clearly stating that it will not build yet), it a great thing. > I'd also like to poke again the idea of pulling in DTB/DTBO files based > on some pattern matching instead of manually adding each time. I know the > idea got shot down, but I think this is going to be a growing problem as > we start to increase our usage of DT Overlays in the near future.. The one annoying with making all of the DTBs being patterns is that we have to implement the logic to expand the wildcarding and set KERNEL_DEVICETREE. If anyone tries to point to a different kernel (say an upstream version) then nothing from meta-ti will work. And that may not be a show stopper issue, it's just the one that popped up in my head when you first brought up the idea for doing it for all platforms. Maybe this would be a good feature to promote to bitbake to support? And alternative to KERNEL_DEVICETREE? A different variable? > Andrew > >>> from what i see, the kernel patch is still under review. >>> >>>> >>>> >>>> >>>> >>>> >> >> >> >> >> >> -=-=-=-=-=-=-=-=-=-=-=- >> Links: You receive all messages sent to this group. >> View/Reply Online (#15752): >> https://lists.yoctoproject.org/g/meta-ti/message/15752 >> Mute This Topic: https://lists.yoctoproject.org/mt/96808923/3619733 >> Group Owner: meta-ti+owner@lists.yoctoproject.org >> Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/unsub [afd@ti.com] >> -=-=-=-=-=-=-=-=-=-=-=- >>
On 2/8/23 12:45 PM, Ryan Eatmon wrote: > > > On 2/8/2023 12:28, Andrew Davis wrote: >> On 2/8/23 11:08 AM, Sinthu Raja M via lists.yoctoproject.org wrote: >>> On Wed, Feb 8, 2023 at 10:31 PM Bajjuri, Praneeth <praneeth@ti.com> wrote: >>>> >>>> >>>> >>>> On 2/8/2023 10:37 AM, Ryan Eatmon via lists.yoctoproject.org wrote: >>>>> >>>>> >>>>> On 2/7/2023 9:13, Sinthu Raja wrote: >>>>>> From: Sinthu Raja <sinthu.raja@ti.com> >>>>>> >>>>>> The AM68 SK has PWM output pins connected to its RPi headers. Add >>>>>> overlay file to enable PWM pins. >>>>>> >>>>>> Signed-off-by: Sinthu Raja <sinthu.raja@ti.com> >>>>>> --- >>>>>> >>>>>> Changes in v2: >>>>>> ============== >>>>>> Address review comment: >>>>>> - Remove extra space >>>>>> >>>>>> v1:https://lists.yoctoproject.org/g/meta-ti/message/15730 >>>>>> conf/machine/j721s2-evm.conf | 1 + >>>>>> 1 file changed, 1 insertion(+) >>>>>> >>>>>> diff --git a/conf/machine/j721s2-evm.conf b/conf/machine/j721s2-evm.conf >>>>>> index 7c120dcc..f1127ed5 100644 >>>>>> --- a/conf/machine/j721s2-evm.conf >>>>>> +++ b/conf/machine/j721s2-evm.conf >>>>>> @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ >>>>>> ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ >>>>>> ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ >>>>>> ti/k3-am68-sk-fpdlink-fusion.dtbo \ >>>>>> + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ >>>>>> ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ >>>>>> ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ >>>>>> ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \ >>>>> >>>>> >>>>> This file does not exists in the kernel (yet?). I'm assuming that you >>>>> all have put in the patch for it and we are just playing a waiting game? >>>>> >>>>> If that is the case, then I would appreciate (in the future) not >>>>> submitting these patches to the meta-ti mailing list until the kernel >>>>> .dtb changes have been accepted into the kernel. This is the second >>>>> patch in recent weeks that I'm having to sit on until the kernel catches >>>>> up... >>>>> >>>>> It makes it really hard to test these patches if the kernel part is not >>>>> there, and I do not have the bandwidth to keep checking daily if the >>>>> change has made it into the public/cicd branches. >>>>> >>>> >>>> Agreed here. >>>> Sinthu, Please post the associated meta-ti change only once the kernel >>>> dts change is merged to cicd build. >>>> >>> Agreed. >>> >> >> I'm going to slightly disagree, I don't see an issue with getting some >> early reviews. Just as long as you only post it as an "RFC" and only >> post the real "PATCH" when ready. > > If the patch is doing more than just adding a dtb, or changing a SRCREV, or anything that depends on an external repo to be updated then I would agree. Posting the change under the RFC banner (and clearly stating that it will not build yet), it a great thing. > Fair, but in this case, even this simple change had an issue and needed a v2, what might be trivial to us might need review for folks new to all this. :) > >> I'd also like to poke again the idea of pulling in DTB/DTBO files based >> on some pattern matching instead of manually adding each time. I know the >> idea got shot down, but I think this is going to be a growing problem as >> we start to increase our usage of DT Overlays in the near future.. > > The one annoying with making all of the DTBs being patterns is that we have to implement the logic to expand the wildcarding and set KERNEL_DEVICETREE. If anyone tries to point to a different kernel (say an upstream version) then nothing from meta-ti will work. And that may not be a show stopper issue, it's just the one that popped up in my head when you first brought up the idea for doing it for all platforms. > > Maybe this would be a good feature to promote to bitbake to support? And alternative to KERNEL_DEVICETREE? A different variable? > That is an option, no need to make it TI specific, I'd imagine wildcard filtering would be useful for many other's bsp layers. Andrew >> Andrew >> >>>> from what i see, the kernel patch is still under review. >>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>> >>> >>> >>> >>> >>> -=-=-=-=-=-=-=-=-=-=-=- >>> Links: You receive all messages sent to this group. >>> View/Reply Online (#15752): https://lists.yoctoproject.org/g/meta-ti/message/15752 >>> Mute This Topic: https://lists.yoctoproject.org/mt/96808923/3619733 >>> Group Owner: meta-ti+owner@lists.yoctoproject.org >>> Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/unsub [afd@ti.com] >>> -=-=-=-=-=-=-=-=-=-=-=- >>> >
On Wed, Feb 08, 2023 at 02:10:23PM -0600, Andrew Davis via lists.yoctoproject.org wrote: > >>I'd also like to poke again the idea of pulling in DTB/DTBO files based > >>on some pattern matching instead of manually adding each time. I know the > >>idea got shot down, but I think this is going to be a growing problem as > >>we start to increase our usage of DT Overlays in the near future.. > > > >The one annoying with making all of the DTBs being patterns is that we have > >to implement the logic to expand the wildcarding and set > >KERNEL_DEVICETREE. If anyone tries to point to a different kernel (say an > >upstream version) then nothing from meta-ti will work. And that may not be > >a show stopper issue, it's just the one that popped up in my head when you > >first brought up the idea for doing it for all platforms. > > > >Maybe this would be a good feature to promote to bitbake to support? And > >alternative to KERNEL_DEVICETREE? A different variable? > > That is an option, no need to make it TI specific, I'd imagine wildcard > filtering would be useful for many other's bsp layers. To be fair - it wasn't "shut down" completely. I just pointed out that any dynamic processing of DTBs/DTBOs list would require more work. KERNEL_DEVICETREE is a top-level global variable used and referenced from other places besides the kernel recipe, specifically for image packaging, e.g. by WIC or FIT or anything else. If it only gets populated later on during the build process when the kernel gets built, none of those places will be able to see the list. There are some hurdles to overcome, but it may be a nice generic feature to add...
diff --git a/conf/machine/j721s2-evm.conf b/conf/machine/j721s2-evm.conf index 7c120dcc..f1127ed5 100644 --- a/conf/machine/j721s2-evm.conf +++ b/conf/machine/j721s2-evm.conf @@ -17,6 +17,7 @@ KERNEL_DEVICETREE = " \ ti/k3-am68-sk-bb-csi2-ov5640.dtbo \ ti/k3-am68-sk-bb-rpi-cam-imx219.dtbo \ ti/k3-am68-sk-fpdlink-fusion.dtbo \ + ti/k3-am68-sk-rpi-hdr-ehrpwm.dtbo \ ti/k3-am68-sk-som-ddr_mem_carveout.dtbo \ ti/k3-j721e-fpdlink-imx390-cm-0-0.dtbo \ ti/k3-j721e-fpdlink-imx390-cm-0-1.dtbo \