03:00 <NishanthMenon> pivi: marex https://lore.kernel.org/all/20240501185331.1189647-1-alexander.sverdlin@siemens.com/

03:29 <marex> NishanthMenon: joy, thanks for digging into it

03:29 <marex> NishanthMenon: I'm sure pivi is even happier

10:08 <pivi> marex: I am always happy!

10:10 <pivi> marex: I even discovered a secret stash of beer I forgot about in my appartment this week, this makes me even happier!

11:23 <marex> pivi: hehehehe

11:26 <NishanthMenon> lol ;)

17:28 <yann-kaelig> Hello

17:41 <yann-kaelig> Well, still can not boot properly from my hdd-raid subboard but I found exactly what to do to get it work. In summary for those who haven't followed :D . To get the sata disk recognized by scsi scan, I have to boot a first time the board ( with or without the SSD connected the SATA cable it doesn't make any difference ). There is a LED on the sata line which lights up and then lights off. Next I poweroff the board wit hthe

17:41 <yann-kaelig> button and I plug the SSD at the SATA line. The LED lights ON but at this time doesn't light off and that make the big difference because when I turn ON the board again then the SSD drive is recognized by the scsi scan command line.

17:49 <yann-kaelig> Maybe the LED light ON and OFF is a sign of a cut in the SATA line which is preventing communication to the disk ?

17:54 <yann-kaelig> I'm playing with preboot, rebuilding uboot and maybe a regulator command before the scsi scan could help, maybe something else

19:15 <marex> yann-kaelig: it is probably some regulator/gpio/retained state in some register issue

19:29 <NishanthMenon> Tartarus: we are at it again https://lists.linaro.org/archives/list/boot-architecture@lists.linaro.org/thread/DZCZSOCRH5BN7YOXEL2OQKSDIY7DCW2M/

20:15 <pivi> NishanthMenon: I like this. System Ready IR or EBBR or ... is just not enough at the moment

20:16 <pivi> NishanthMenon: not sure if we can help on any way on this topic, you as TI should have way more resources than me on such a topic

20:19 <NishanthMenon> pivi: Jon(my boss) will be talking about exactly that https://www.kitefor.events/events/linaro-connect-24/submissions/127 .. I keep telling him to rename it as 'myth of just works' but he doesn't agree :(

20:21 <pivi> NishanthMenon: I have the same problem with my boss not listening to me, maybe they both need a shrink to help out

20:22 <pivi> anyway, done for today, if you feel like we can help on some lobbying just let me know

21:41 <yann-kaelig> well. The same issue occur with older uboot version, so I suppose that I'm the first on this problem

21:45 <yann-kaelig> So, what information should I provide for a bug report ? I suppose that I have to turn ON config logging in uboot

21:46 <marex> yann-kaelig: ideally a patch, scratch your own itch, people can provide hints, but they won't debug a bug they dont even have hardware to debug on

21:48 <yann-kaelig> I understand but I do not have the level to find the source of the issue. I don't know what to do now. I tried each of your ideas, none of them fix the issue, the issue seems to be somewhere else.

21:49 <yann-kaelig> But I don't think I wil be able to isolate it.

21:49 <marex> yann-kaelig: dump registers and compare, really

21:49 <marex> you have a test case, right ?

21:57 <yann-kaelig> I know when the issue occur and I know how to fix the issue by manually unplugging and reconnecting the SSD, but it is done when the board is off, so how can I know what is happening at this very moment when I plug the SSD ?

21:59 <yann-kaelig> I mean reconnecting the SSD is done when the board is off, and during this time there is a signal that it is send to the hardware and the LED turn ON.

22:00 <marex> yann-kaelig: didnt you also say that the problem does not occur if you boot Linux once and then warm-reset out of Linux ?

22:08 <yann-kaelig> marex: yes, that right tested again right now. I boot linux, reboot, get a uboot command prompt. scsi scan works, indeed.

22:09 <yann-kaelig> boot linux from the sdcard*

22:21 <marex> yann-kaelig: analyze everything which is somehow involved

22:21 <marex> yann-kaelig: the clock IP seems like a suitable first culprit

23:08 <yann-kaelig> well, thc for your help. Cu later

23:53 <crb> my board comes up with a random ethernet mac address every time which make DHCP hard, if I wanted to hard code a MAC address where would I look in the source?