All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday: - As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=rele...
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress: - With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work: - Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver - Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce. Is the release/db820c/qcomlt-5.15 basically the same as this branch? I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Thanks, Kim
From: Paul Neuhardt paul.neuhardt@linaro.org Sent: Friday, July 15, 2022 2:12 AM To: sightlineapplications@lists.linaro.org Subject: [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday: - As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=rele...
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress: - With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work: - Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver - Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul -- [Image removed by sender. Linaro]http://www.linaro.org/ Paul Neuhardt | Sr. Program Manager T: +44 0771 377 8664 paul.neuhardt@linaro.orgmailto:glen.valante@linaro.org IRC: pneuhardt
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=rele... https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboard820c/...
Maybe I only need the boot-XXX.img?
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=rele... https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboard820c/...
Maybe I only need the boot-XXX.img?
If you build all necessary drivers into the kernel, you can use just the boot.img. Moreover, you even don't have to flash it for the test. You can use `fastboot boot boot-xxx.img` to boot the image.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=rele... https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboard820c/...
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=rele... https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
One other question Dmitry. Do you have build instructions? I will try building how we usually build and I will try.
Not flashing the boot image would work for us initially just for bootability. But for testing in the cold, I think we would need to still press S1 button on the Dragonboard820 to do a fastboot to flash the board and wouldn't be accessible.
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboar d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/ ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboard820c/...
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboar d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/ ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
Hi,
Please excuse me, this took a bit longer due to issues with undervolting (see below).
On 15/07/2022 21:18, Kim Steiner wrote:
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboard820c/...
These instructions apply to the usual Linaro kernels. For you kernel I've dropped the patches adding distro.config, as we agreed to base on the latest stable kernel (rather than 5.15 Linaro kernel). So you can just drop 'distro.config' from the command line and continue as described.
Today I've pushed the commit enabling all necessary drivers and options into the builtin (instead of using the distro.config), so you can try the build after pulling the tree.
Note: At least on my board if I boot from the initramfs, the image will work. If I boot from the UFS, for now the board will undervolt with the default setup. Apply the following changes to the .config after running make ... defconfig:
--- .config.old 2022-07-19 12:10:37.909125521 +0300 +++ .config 2022-07-19 12:10:30.121075868 +0300 @@ -551,11 +551,11 @@ CONFIG_CPU_FREQ_GOV_COMMON=y CONFIG_CPU_FREQ_STAT=y # CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE is not set -# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y # CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set -CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL is not set CONFIG_CPU_FREQ_GOV_PERFORMANCE=y CONFIG_CPU_FREQ_GOV_POWERSAVE=y CONFIG_CPU_FREQ_GOV_USERSPACE=y @@ -1660,8 +1660,7 @@ CONFIG_FW_LOADER=y CONFIG_FW_LOADER_PAGED_BUF=y CONFIG_EXTRA_FIRMWARE="" -CONFIG_FW_LOADER_USER_HELPER=y -CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y +# CONFIG_FW_LOADER_USER_HELPER is not set # CONFIG_FW_LOADER_COMPRESS is not set CONFIG_FW_CACHE=y # end of Firmware loader
Then proceed with building as usual.
For your reference I have uploaded two images to our filserver: - boot-db820c-2cpu.img https://fileserver.linaro.org/s/nnpJdHZ3b3tW7HC This image will boot and use /dev/sda1 as the rootfs
- boot-initramfs-db820c-2cpu.img https://fileserver.linaro.org/s/H4L9pYSyCD8Yzdm
This image will boot into the debugging initramfs. Unlike the previous image it doesn't use the minimum possible CPU frequency, so it will work faster, but with the possible undervolts.
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboar d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log/ ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
Hi Dmitry,
I tried both of the images on the Dragonboard820c. The one named boot-db820c-2cpu.img booted, but after a few seconds it crashed (I didn't type anything or do anything). I had 4.14 rootfs on the board. See the attached log.
When we test the gpu temperature issue, we need an HDMI monitor attached. We run the application "glxgears". The other image, boot-initramfs-db820c-2cpu.img, doesn't seem to have HDMI working and it didn't know about glxgears.
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Tuesday, July 19, 2022 2:15 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi,
Please excuse me, this took a bit longer due to issues with undervolting (see below).
On 15/07/2022 21:18, Kim Steiner wrote:
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboar d820c/build/linux-kernel.md.html
These instructions apply to the usual Linaro kernels. For you kernel I've dropped the patches adding distro.config, as we agreed to base on the latest stable kernel (rather than 5.15 Linaro kernel). So you can just drop 'distro.config' from the command line and continue as described.
Today I've pushed the commit enabling all necessary drivers and options into the builtin (instead of using the distro.config), so you can try the build after pulling the tree.
Note: At least on my board if I boot from the initramfs, the image will work. If I boot from the UFS, for now the board will undervolt with the default setup. Apply the following changes to the .config after running make ... defconfig:
--- .config.old 2022-07-19 12:10:37.909125521 +0300 +++ .config 2022-07-19 12:10:30.121075868 +0300 @@ -551,11 +551,11 @@ CONFIG_CPU_FREQ_GOV_COMMON=y CONFIG_CPU_FREQ_STAT=y # CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE is not set -# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y # CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set -CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL is not set CONFIG_CPU_FREQ_GOV_PERFORMANCE=y CONFIG_CPU_FREQ_GOV_POWERSAVE=y CONFIG_CPU_FREQ_GOV_USERSPACE=y @@ -1660,8 +1660,7 @@ CONFIG_FW_LOADER=y CONFIG_FW_LOADER_PAGED_BUF=y CONFIG_EXTRA_FIRMWARE="" -CONFIG_FW_LOADER_USER_HELPER=y -CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y +# CONFIG_FW_LOADER_USER_HELPER is not set # CONFIG_FW_LOADER_COMPRESS is not set CONFIG_FW_CACHE=y # end of Firmware loader
Then proceed with building as usual.
For your reference I have uploaded two images to our filserver: - boot-db820c-2cpu.img https://fileserver.linaro.org/s/nnpJdHZ3b3tW7HC This image will boot and use /dev/sda1 as the rootfs
- boot-initramfs-db820c-2cpu.img https://fileserver.linaro.org/s/H4L9pYSyCD8Yzdm
This image will boot into the debugging initramfs. Unlike the previous image it doesn't use the minimum possible CPU frequency, so it will work faster, but with the possible undervolts.
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboa r d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log / ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/lo g /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
-- With best wishes Dmitry
Hi Dmitry,
I tried building the kernel for 5.15 myself and run into a similar reboot issue. When I built the kernel myself, it was confused by this: CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y
So it selected the "is not set" item. I can try building with the "=y" for this config, but seems to have same reboot issue as the image "boot-db820c-2cpu.img" given. I also disabled CONFIG_BT in the .config because it looked like it was rebooting doing some Bluetooth stuff, but that didn't help.
Could I be loading the image wrong? I just do "fastboot flash boot boot-db820c-2cpu.img". Do I need the filesystem modules? The only ones listed under \lib\modules are the 4.14 modules.
Thanks, Kim
-----Original Message----- From: Kim Steiner Sent: Tuesday, July 19, 2022 8:23 AM To: Dmitry Baryshkov dmitry.baryshkov@linaro.org; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: RE: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Dmitry,
I tried both of the images on the Dragonboard820c. The one named boot-db820c-2cpu.img booted, but after a few seconds it crashed (I didn't type anything or do anything). I had 4.14 rootfs on the board. See the attached log.
When we test the gpu temperature issue, we need an HDMI monitor attached. We run the application "glxgears". The other image, boot-initramfs-db820c-2cpu.img, doesn't seem to have HDMI working and it didn't know about glxgears.
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Tuesday, July 19, 2022 2:15 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi,
Please excuse me, this took a bit longer due to issues with undervolting (see below).
On 15/07/2022 21:18, Kim Steiner wrote:
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboar d820c/build/linux-kernel.md.html
These instructions apply to the usual Linaro kernels. For you kernel I've dropped the patches adding distro.config, as we agreed to base on the latest stable kernel (rather than 5.15 Linaro kernel). So you can just drop 'distro.config' from the command line and continue as described.
Today I've pushed the commit enabling all necessary drivers and options into the builtin (instead of using the distro.config), so you can try the build after pulling the tree.
Note: At least on my board if I boot from the initramfs, the image will work. If I boot from the UFS, for now the board will undervolt with the default setup. Apply the following changes to the .config after running make ... defconfig:
--- .config.old 2022-07-19 12:10:37.909125521 +0300 +++ .config 2022-07-19 12:10:30.121075868 +0300 @@ -551,11 +551,11 @@ CONFIG_CPU_FREQ_GOV_COMMON=y CONFIG_CPU_FREQ_STAT=y # CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE is not set -# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y # CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set -CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL is not set CONFIG_CPU_FREQ_GOV_PERFORMANCE=y CONFIG_CPU_FREQ_GOV_POWERSAVE=y CONFIG_CPU_FREQ_GOV_USERSPACE=y @@ -1660,8 +1660,7 @@ CONFIG_FW_LOADER=y CONFIG_FW_LOADER_PAGED_BUF=y CONFIG_EXTRA_FIRMWARE="" -CONFIG_FW_LOADER_USER_HELPER=y -CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y +# CONFIG_FW_LOADER_USER_HELPER is not set # CONFIG_FW_LOADER_COMPRESS is not set CONFIG_FW_CACHE=y # end of Firmware loader
Then proceed with building as usual.
For your reference I have uploaded two images to our filserver: - boot-db820c-2cpu.img https://fileserver.linaro.org/s/nnpJdHZ3b3tW7HC This image will boot and use /dev/sda1 as the rootfs
- boot-initramfs-db820c-2cpu.img https://fileserver.linaro.org/s/H4L9pYSyCD8Yzdm
This image will boot into the debugging initramfs. Unlike the previous image it doesn't use the minimum possible CPU frequency, so it will work faster, but with the possible undervolts.
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboa r d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing a preliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log / ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/lo g /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernel parameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to the CBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver
- Support for Sightline as required.
Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
-- With best wishes Dmitry
Hi Kim,
On 20/07/2022 02:35, Kim Steiner wrote:
Hi Dmitry,
I tried building the kernel for 5.15 myself and run into a similar reboot issue. When I built the kernel myself, it was confused by this: CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y
So it selected the "is not set" item. I can try building with the "=y" for this config, but seems to have same reboot issue as the image "boot-db820c-2cpu.img" given. I also disabled CONFIG_BT in the .config because it looked like it was rebooting doing some Bluetooth stuff, but that didn't help.
Could I be loading the image wrong? I just do "fastboot flash boot boot-db820c-2cpu.img". Do I need the filesystem modules? The only ones listed under \lib\modules are the 4.14 modules.
Well, default config comes with ext2, ext3 and ext4 built in. I suppose the issue is with the cpufreq. Could you please try disabling the CONFIG_ARM_QCOM_CPUFREQ_NVMEM and repeating the test?
Note, I haven't checked the GFX, so there might be voltage issues too. I'll check this today or tomorrow.
Thanks, Kim
-----Original Message----- From: Kim Steiner Sent: Tuesday, July 19, 2022 8:23 AM To: Dmitry Baryshkov dmitry.baryshkov@linaro.org; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: RE: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Dmitry,
I tried both of the images on the Dragonboard820c. The one named boot-db820c-2cpu.img booted, but after a few seconds it crashed (I didn't type anything or do anything). I had 4.14 rootfs on the board. See the attached log.
When we test the gpu temperature issue, we need an HDMI monitor attached. We run the application "glxgears". The other image, boot-initramfs-db820c-2cpu.img, doesn't seem to have HDMI working and it didn't know about glxgears.
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Tuesday, July 19, 2022 2:15 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi,
Please excuse me, this took a bit longer due to issues with undervolting (see below).
On 15/07/2022 21:18, Kim Steiner wrote:
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboar d820c/build/linux-kernel.md.html
These instructions apply to the usual Linaro kernels. For you kernel I've dropped the patches adding distro.config, as we agreed to base on the latest stable kernel (rather than 5.15 Linaro kernel). So you can just drop 'distro.config' from the command line and continue as described.
Today I've pushed the commit enabling all necessary drivers and options into the builtin (instead of using the distro.config), so you can try the build after pulling the tree.
Note: At least on my board if I boot from the initramfs, the image will work. If I boot from the UFS, for now the board will undervolt with the default setup. Apply the following changes to the .config after running make ... defconfig:
--- .config.old 2022-07-19 12:10:37.909125521 +0300 +++ .config 2022-07-19 12:10:30.121075868 +0300 @@ -551,11 +551,11 @@ CONFIG_CPU_FREQ_GOV_COMMON=y CONFIG_CPU_FREQ_STAT=y # CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE is not set -# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y # CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set -CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL is not set CONFIG_CPU_FREQ_GOV_PERFORMANCE=y CONFIG_CPU_FREQ_GOV_POWERSAVE=y CONFIG_CPU_FREQ_GOV_USERSPACE=y @@ -1660,8 +1660,7 @@ CONFIG_FW_LOADER=y CONFIG_FW_LOADER_PAGED_BUF=y CONFIG_EXTRA_FIRMWARE="" -CONFIG_FW_LOADER_USER_HELPER=y -CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y +# CONFIG_FW_LOADER_USER_HELPER is not set # CONFIG_FW_LOADER_COMPRESS is not set CONFIG_FW_CACHE=y # end of Firmware loader
Then proceed with building as usual.
For your reference I have uploaded two images to our filserver:
boot-db820c-2cpu.img https://fileserver.linaro.org/s/nnpJdHZ3b3tW7HC This image will boot and use /dev/sda1 as the rootfs
boot-initramfs-db820c-2cpu.img https://fileserver.linaro.org/s/H4L9pYSyCD8Yzdm
This image will boot into the debugging initramfs. Unlike the previous image it doesn't use the minimum possible CPU frequency, so it will work faster, but with the possible undervolts.
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonboa r d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing apreliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/log / ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/lo g /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernelparameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to theCBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver - Support for Sightline as required.Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
-- With best wishes Dmitry
Hi Dmitry,
I tried # CONFIG_ARM_QCOM_CPUFREQ_NVMEM is not set
I reboot 3 times. The 3rd time, the dragonboard820c did the same thing as before: [ 22.384255] atl1c 0002:01:00.0: atl1c: enP2p1s0 NIC Link is Up<100 Mbps Full Duplex> [ 28.191038] Bluetooth: hci0: setting up ROME/QCA6390 [ 28.675945] Bluetooth: hci0: QCA Product ID :0x00000008 [ 28.675998] Bluetooth: hci0: QCA SOC Version :0x00000044 [ 28.680413] Bluetooth: hci0: QCA ROM Version :0x00000302 [ 28.685769] Bluetooth: hci0: QCA Patch Version:0x00000111 [ 28.692543] Bluetooth: hci0: QCA controller version 0x00440302 [ 28.696558] Bluetooth: hci0: QCA Downloading qca/rampatch_00440302.bin [ 28.702982] bluetooth hci0: Direct firmware load for qca/rampatch_00440302.bin failed with error -2 [ 28.708807] Bluetooth: hci0: QCA Failed to request file: qca/rampatch_00440302.bin (-2) [ 28.717738] Bluetooth: hci0: QCA Failed to download patch (-2)
Debian GNU/Linux 10 linaro-alip ttyMSM0
linaro-alip login: root (automatic login)
Last login: Thu Jul 21 12:40:19 UTC 2022 on ttyMSM0 Linux linaro-alip 5.15.0-qcomlt-arm64 #5 SMP PREEMPT Thu Jul 21 05:21:57 PDT 2022 aarch64
The programs included with the Debian GNU/Linux system are free software; the exact distribution terms for each program are described in the individual files in /usr/share/doc/*/copyright.
Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent permitted by applicable law. root@linaro-alip:~# Format: Log Type - Time(microsec) - Message - Optional Info Log Type: B - Since Boot(Power On Reset), D - Delta, S - Statistic S - QC_IMAGE_VERSION_STRING=BOOT.XF.1.0-00301 S - IMAGE_VARIANT_STRING=M8996LAB
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Thursday, July 21, 2022 3:59 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 20/07/2022 02:35, Kim Steiner wrote:
Hi Dmitry,
I tried building the kernel for 5.15 myself and run into a similar reboot issue. When I built the kernel myself, it was confused by this: CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y
So it selected the "is not set" item. I can try building with the "=y" for this config, but seems to have same reboot issue as the image "boot-db820c-2cpu.img" given. I also disabled CONFIG_BT in the .config because it looked like it was rebooting doing some Bluetooth stuff, but that didn't help.
Could I be loading the image wrong? I just do "fastboot flash boot boot-db820c-2cpu.img". Do I need the filesystem modules? The only ones listed under \lib\modules are the 4.14 modules.
Well, default config comes with ext2, ext3 and ext4 built in. I suppose the issue is with the cpufreq. Could you please try disabling the CONFIG_ARM_QCOM_CPUFREQ_NVMEM and repeating the test?
Note, I haven't checked the GFX, so there might be voltage issues too. I'll check this today or tomorrow.
Thanks, Kim
-----Original Message----- From: Kim Steiner Sent: Tuesday, July 19, 2022 8:23 AM To: Dmitry Baryshkov dmitry.baryshkov@linaro.org; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: RE: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Dmitry,
I tried both of the images on the Dragonboard820c. The one named boot-db820c-2cpu.img booted, but after a few seconds it crashed (I didn't type anything or do anything). I had 4.14 rootfs on the board. See the attached log.
When we test the gpu temperature issue, we need an HDMI monitor attached. We run the application "glxgears". The other image, boot-initramfs-db820c-2cpu.img, doesn't seem to have HDMI working and it didn't know about glxgears.
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Tuesday, July 19, 2022 2:15 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi,
Please excuse me, this took a bit longer due to issues with undervolting (see below).
On 15/07/2022 21:18, Kim Steiner wrote:
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboa r d820c/build/linux-kernel.md.html
These instructions apply to the usual Linaro kernels. For you kernel I've dropped the patches adding distro.config, as we agreed to base on the latest stable kernel (rather than 5.15 Linaro kernel). So you can just drop 'distro.config' from the command line and continue as described.
Today I've pushed the commit enabling all necessary drivers and options into the builtin (instead of using the distro.config), so you can try the build after pulling the tree.
Note: At least on my board if I boot from the initramfs, the image will work. If I boot from the UFS, for now the board will undervolt with the default setup. Apply the following changes to the .config after running make ... defconfig:
--- .config.old 2022-07-19 12:10:37.909125521 +0300 +++ .config 2022-07-19 12:10:30.121075868 +0300 @@ -551,11 +551,11 @@ CONFIG_CPU_FREQ_GOV_COMMON=y CONFIG_CPU_FREQ_STAT=y # CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE is not set -# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y # CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set -CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL is not set CONFIG_CPU_FREQ_GOV_PERFORMANCE=y CONFIG_CPU_FREQ_GOV_POWERSAVE=y CONFIG_CPU_FREQ_GOV_USERSPACE=y @@ -1660,8 +1660,7 @@ CONFIG_FW_LOADER=y CONFIG_FW_LOADER_PAGED_BUF=y CONFIG_EXTRA_FIRMWARE="" -CONFIG_FW_LOADER_USER_HELPER=y -CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y +# CONFIG_FW_LOADER_USER_HELPER is not set # CONFIG_FW_LOADER_COMPRESS is not set CONFIG_FW_CACHE=y # end of Firmware loader
Then proceed with building as usual.
For your reference I have uploaded two images to our filserver:
boot-db820c-2cpu.img https://fileserver.linaro.org/s/nnpJdHZ3b3tW7HC This image will boot and use /dev/sda1 as the rootfs
boot-initramfs-db820c-2cpu.img https://fileserver.linaro.org/s/H4L9pYSyCD8Yzdm
This image will boot into the debugging initramfs. Unlike the previous image it doesn't use the minimum possible CPU frequency, so it will work faster, but with the possible undervolts.
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonbo a r d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing apreliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/lo g / ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/l o g /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernelparameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to theCBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver - Support for Sightline as required.Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
-- With best wishes Dmitry
-- With best wishes Dmitry
Hi Dmitry,
Just a random thought, but if this boot up issue that I'm seeing has anything to do with cpu frequency changes, could just setting the cpu frequency to some value help? On our custom board we set the cpu frequency to a single frequency. I don't know the internals of the system so maybe the cpu freq changes anyway.
Thanks, Kim
-----Original Message----- From: Kim Steiner Sent: Thursday, July 21, 2022 5:50 AM To: Dmitry Baryshkov dmitry.baryshkov@linaro.org; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: RE: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Dmitry,
I tried # CONFIG_ARM_QCOM_CPUFREQ_NVMEM is not set
I reboot 3 times. The 3rd time, the dragonboard820c did the same thing as before: [ 22.384255] atl1c 0002:01:00.0: atl1c: enP2p1s0 NIC Link is Up<100 Mbps Full Duplex> [ 28.191038] Bluetooth: hci0: setting up ROME/QCA6390 [ 28.675945] Bluetooth: hci0: QCA Product ID :0x00000008 [ 28.675998] Bluetooth: hci0: QCA SOC Version :0x00000044 [ 28.680413] Bluetooth: hci0: QCA ROM Version :0x00000302 [ 28.685769] Bluetooth: hci0: QCA Patch Version:0x00000111 [ 28.692543] Bluetooth: hci0: QCA controller version 0x00440302 [ 28.696558] Bluetooth: hci0: QCA Downloading qca/rampatch_00440302.bin [ 28.702982] bluetooth hci0: Direct firmware load for qca/rampatch_00440302.bin failed with error -2 [ 28.708807] Bluetooth: hci0: QCA Failed to request file: qca/rampatch_00440302.bin (-2) [ 28.717738] Bluetooth: hci0: QCA Failed to download patch (-2)
Debian GNU/Linux 10 linaro-alip ttyMSM0
linaro-alip login: root (automatic login)
Last login: Thu Jul 21 12:40:19 UTC 2022 on ttyMSM0 Linux linaro-alip 5.15.0-qcomlt-arm64 #5 SMP PREEMPT Thu Jul 21 05:21:57 PDT 2022 aarch64
The programs included with the Debian GNU/Linux system are free software; the exact distribution terms for each program are described in the individual files in /usr/share/doc/*/copyright.
Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent permitted by applicable law. root@linaro-alip:~# Format: Log Type - Time(microsec) - Message - Optional Info Log Type: B - Since Boot(Power On Reset), D - Delta, S - Statistic S - QC_IMAGE_VERSION_STRING=BOOT.XF.1.0-00301 S - IMAGE_VARIANT_STRING=M8996LAB
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Thursday, July 21, 2022 3:59 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 20/07/2022 02:35, Kim Steiner wrote:
Hi Dmitry,
I tried building the kernel for 5.15 myself and run into a similar reboot issue. When I built the kernel myself, it was confused by this: CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y
So it selected the "is not set" item. I can try building with the "=y" for this config, but seems to have same reboot issue as the image "boot-db820c-2cpu.img" given. I also disabled CONFIG_BT in the .config because it looked like it was rebooting doing some Bluetooth stuff, but that didn't help.
Could I be loading the image wrong? I just do "fastboot flash boot boot-db820c-2cpu.img". Do I need the filesystem modules? The only ones listed under \lib\modules are the 4.14 modules.
Well, default config comes with ext2, ext3 and ext4 built in. I suppose the issue is with the cpufreq. Could you please try disabling the CONFIG_ARM_QCOM_CPUFREQ_NVMEM and repeating the test?
Note, I haven't checked the GFX, so there might be voltage issues too. I'll check this today or tomorrow.
Thanks, Kim
-----Original Message----- From: Kim Steiner Sent: Tuesday, July 19, 2022 8:23 AM To: Dmitry Baryshkov dmitry.baryshkov@linaro.org; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: RE: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Dmitry,
I tried both of the images on the Dragonboard820c. The one named boot-db820c-2cpu.img booted, but after a few seconds it crashed (I didn't type anything or do anything). I had 4.14 rootfs on the board. See the attached log.
When we test the gpu temperature issue, we need an HDMI monitor attached. We run the application "glxgears". The other image, boot-initramfs-db820c-2cpu.img, doesn't seem to have HDMI working and it didn't know about glxgears.
Thanks, Kim
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Tuesday, July 19, 2022 2:15 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi,
Please excuse me, this took a bit longer due to issues with undervolting (see below).
On 15/07/2022 21:18, Kim Steiner wrote:
Hi Dmitry,
I think I need build instructions also. I tried the build instructions on the 96board forum, but I get stuck:
host> make defconfig distro.config *** Default configuration is based on 'defconfig' # # configuration written to .config # scripts/kconfig/Makefile:99: *** No configuration exists for this target on this architecture. Stop.
Maybe these instructions are not correct? https://www.96boards.org/documentation/consumer/dragonboard/dragonboa r d820c/build/linux-kernel.md.html
These instructions apply to the usual Linaro kernels. For you kernel I've dropped the patches adding distro.config, as we agreed to base on the latest stable kernel (rather than 5.15 Linaro kernel). So you can just drop 'distro.config' from the command line and continue as described.
Today I've pushed the commit enabling all necessary drivers and options into the builtin (instead of using the distro.config), so you can try the build after pulling the tree.
Note: At least on my board if I boot from the initramfs, the image will work. If I boot from the UFS, for now the board will undervolt with the default setup. Apply the following changes to the .config after running make ... defconfig:
--- .config.old 2022-07-19 12:10:37.909125521 +0300 +++ .config 2022-07-19 12:10:30.121075868 +0300 @@ -551,11 +551,11 @@ CONFIG_CPU_FREQ_GOV_COMMON=y CONFIG_CPU_FREQ_STAT=y # CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE is not set -# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE=y # CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set # CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set -CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL is not set CONFIG_CPU_FREQ_GOV_PERFORMANCE=y CONFIG_CPU_FREQ_GOV_POWERSAVE=y CONFIG_CPU_FREQ_GOV_USERSPACE=y @@ -1660,8 +1660,7 @@ CONFIG_FW_LOADER=y CONFIG_FW_LOADER_PAGED_BUF=y CONFIG_EXTRA_FIRMWARE="" -CONFIG_FW_LOADER_USER_HELPER=y -CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y +# CONFIG_FW_LOADER_USER_HELPER is not set # CONFIG_FW_LOADER_COMPRESS is not set CONFIG_FW_CACHE=y # end of Firmware loader
Then proceed with building as usual.
For your reference I have uploaded two images to our filserver:
boot-db820c-2cpu.img https://fileserver.linaro.org/s/nnpJdHZ3b3tW7HC This image will boot and use /dev/sda1 as the rootfs
boot-initramfs-db820c-2cpu.img https://fileserver.linaro.org/s/H4L9pYSyCD8Yzdm
This image will boot into the debugging initramfs. Unlike the previous image it doesn't use the minimum possible CPU frequency, so it will work faster, but with the possible undervolts.
-----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 9:12 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
On 15/07/2022 17:43, Kim Steiner wrote:
Hi Dmitry,
Thank you for the explanation!
We use fastboot to flash the dragonboard820 board with commands like this:
sudo fastboot flash boot boot-linaro-buster-dragonboard-820c-BUILD.img sudo fastboot flash rootfs linaro-buster-alip-dragonboard-820c-BUILD.img
this is described here: https://www.96boards.org/documentation/consumer/dragonboard/dragonbo a r d820c/installation/linux-fastboot.md.html
Maybe I only need the boot-XXX.img?
I will provide boot.img for you to test on Monday.
Thanks, Kim -----Original Message----- From: Dmitry Baryshkov dmitry.baryshkov@linaro.org Sent: Friday, July 15, 2022 7:35 AM To: Kim Steiner kim.steiner@sightlineapplications.com; Paul Neuhardt paul.neuhardt@linaro.org; sightlineapplications@lists.linaro.org Subject: Re: [Sightlineapplications] Re: Project Status EOD Thursday, 14 July 2022
Hi Kim,
On 15/07/2022 15:34, Kim Steiner wrote:
Hi Paul,
This all sounds like good news! We were building the /release/qcomlt-5.15 before and applying our patches from Inforce.
The release/qcomlt-5.15 is a generic branch targeting all the supported Qualcomm hardware. However I'm not sure if it supported 820 well. As you can see, the branch was not updated for quite some time.
Is the release/db820c/qcomlt-5.15 basically the same as this branch?
The release/db820c/qcomlt-5.15 is cleaned up branch, targeting only 820/APQ8096. Because it is based on 5.15.54, it received several patches backported from later kernel by the linux-stable team. On top of that it contains patches from the release/qcomlt-5.15 tree and additional patches targeting 820c. Later (next week) I'm going to update this branch with the patches that are being developed in the mean time (for the CBF clocks, regulators, etc).
I will try building from the branch you mentioned and applying our patches from Inforce to see if our customer board will boot.
I would suggest trying this kernel on the Dragonboard 820c first. I do not have access to IFC6601, so I can not test whether it will on your SoM and which modifications it might require. Later on I will test the kernel on the IFC6640 SBC, which hopefully is close enough to the IFC6601 SoM.
One question I have is would it be possible to get the binaries from this branch for the DragonBoard820c to do some quick temperature testing?
Could you please clarify, which binaries would you like to receive? For my own experiments I use the monolithic kernel with the initramfs containing a set of test utilities. Does that sound like what you would like to use for the tests?
Thanks,
Kim
*From:* Paul Neuhardt paul.neuhardt@linaro.org *Sent:* Friday, July 15, 2022 2:12 AM *To:* sightlineapplications@lists.linaro.org *Subject:* [Sightlineapplications] Project Status EOD Thursday, 14 July 2022
All,
Here is where we stand on the project at the end of the day on Thursday:
Progress since yesterday:
- As promised, Dmitry turned his attention to preparing apreliminary 5.15 kernel with his clock driver patches for Sightline testing. That kernel has been published on the following public branch and is ready for Sightline's initial testing:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/lo g / ?h=release/db820c/qcomlt-5.15 https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/l o g /?h=release/db820c/qcomlt-5.15
Dmitry notes that this kernel should be passed the kernelparameter 'maxcpus=2'.
Work in Progress:
- With the initial kernel published, work will return to theCBF/CPR patches, which when complete will be published on a separate branch.
Planned Work:
- Finish the CBF special handling in the CPR driver - Utilize CBF for CPUfreq scaling - Consider utilizing CPR driver for the GPU regulator powerup - Finalize extra corner cases in the CPR driver - Support for Sightline as required.Budget burn: Yesterday, Dmitry spent 1.5 hours on the project, equating to 0.2 person-days. This leaves 5.6 person-days (45.0 hours) remaining in the engineering budget. I spent .25 hours (for a running total of 3 hours including the kickoff meeting and project set-up), meaning 37% of the PM time has been consumed.
We feel we are still on-track to to deliver the expected outcomes within current budget. We are especially pleased that we have been able to deliver an initial, testable 5.15 kernel with over half of the engineering budget remaining, and we hope Sightline will be pleased with the results.
As always, if you have any questions or concerns please reach out at any time.
Kind regards, Paul
--
Image removed by sender. Linaro http://www.linaro.org/
*Paul Neuhardt*|/Sr. Program Manager/
T: +44 0771 377 8664 paul.neuhardt@linaro.org mailto:glen.valante@linaro.org IRC: pneuhardt
Sightlineapplications mailing list -- sightlineapplications@lists.linaro.org To unsubscribe send an email to sightlineapplications-leave@lists.linaro.org
-- With best wishes Dmitry
-- With best wishes Dmitry
-- With best wishes Dmitry
sightlineapplications@lists.linaro.org
-
Dmitry Baryshkov -
Kim Steiner -
Paul Neuhardt