[meta-ti] meta-ti (a345e0c2) Linux kernel fails to boot on Beagleboard-xM (rev C)

[Kevin.E.Cribbs at L3Harris.com]

If I understand correctly, you are saying the beagleboard MACHINE should cause a working mainline LTS kernel (e.g. 4.19) to be built. It looks like git://git.ti.com/ti-linux-kernel/ti-linux-kernel.git pulls kernels (e.g. 4.19.73) from git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git so the git URL doesn't need changed.

It will be at least a day before I test a mainline kernel/build.


Regarding your " any issues using a mainline kernel for your platform " question ...

I did find a problem in the TI TPS65950 chip device driver code: "drivers/mfd/twl-core.c". The "twl_i2c_read_u8(TWL4030_MODULE_INTBR, &temp, REG_GPPUPDCTR1);" and two lines that immediately follow should only be executed for boards that use external I2C pull ups. Boards that rely on the TPS65950 internal I2C pull ups (that are on by default) can no longer use the TSP65950 I2C buses after the pull ups are disabled. Since the presence/absence of external pull ups cannot be detected and boards without external pull ups fail to operate after disabling the pull ups, they should only be disabled if a device tree property indicates they should be disabled.

Sorry for the rushed response. I'm using a remote connection that isn't working very well.


Kevin Cribbs
Principal Member Engineering Staff
INTEGRATED MISSION SYSTEMS / L3HARRIS TECHNOLOGIES
t +1 856 338 3763
L3Harris.com / kevin.e.cribbs at L3Harris.com
1 Federal Street / Camden, NJ 08103 / USA


-----Original Message-----
From: Denys Dmytriyenko <denys at ti.com> 
Sent: Monday, September 16, 2019 1:29 PM
To: Cribbs, Kevin E (US) @ CSG - CSE <Kevin.E.Cribbs at L3Harris.com>
Cc: meta-ti at yoctoproject.org
Subject: [EXT] Re: [meta-ti] meta-ti (a345e0c2) Linux kernel fails to boot on Beagleboard-xM (rev C)

On Thu, Sep 12, 2019 at 03:14:33PM +0000, Kevin.E.Cribbs at L3Harris.com wrote:
> I created a Yocto Project Bugzilla bug (13517) for this problem. Randy 
> MacLeod said "meta-ti is not tracked in this bugzilla. Please contact 
> the BSP owner".

> https://bugzilla.yoctoproject.org/show_bug.cgi?id=13517
> 
> What is the correct procedure for reporting problems with the meta-ti layer?

Kevin,

Thank you for the report.

Regarding the use of Yocto Project bugzilla for metat-ti bugs - I've discussed this with fellow Yocto Project TSC members and there's a confusion caused by limited resources. We'll try to get this sorted out...

Meanwhile, as of the booting issue itself. As you can see, meta-ti uses TI "staging" tree, where active development for current and latest platforms happens during the upstreaming process. OMAP3 SOC is 10-15 years old and what is called a legacy device. It has been fully supported in mainline Linux kernel for many years. Yes, TI is also responsible for maintaining existing support in the mainline kernel and ensuring no regressions. But that is not against the TI "staging" tree - i.e. there's no active testing being done of meta-ti working on OMAP3 platforms like Beagleboard-xM, unfortunately. This is left to corresponding communities, such as Beagleboard.org

I'll see what I can do to get this looked at in meta-ti. Otherwise, do you see any issues using a mainline kernel for your platform? Thanks.

--
Denys

  

CONFIDENTIALITY NOTICE: This email and any attachments are for the sole use of the intended recipient and may contain material that is proprietary, confidential, privileged or otherwise legally protected or restricted under applicable government laws. Any review, disclosure, distributing or other use without expressed permission of the sender is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies without reading, printing, or saving.