x220 port not working as it is?

[tlaurion]

I've seperated x230 linux configs from x220, reverting the changes prior to updating x230 configs to be FBWhiptail compatible in this branch.

Basically, resulting in the board config, coreboot and linux configs being like this.

The screen stays black.

Someone has compiled and flashed x220 to their laptop recently?

[tlaurion]

@Akendo @jgrip @paulmenzel?

[snmcmillan]

@tlaurion
I have, it works. Wait 50 seconds.

[tlaurion]

Me delay you were talking about?

…

On March 27, 2019 5:20:21 PM EDT, Sebastian McMillan ***@***.***> wrote: I have, it works. Wait 50 seconds. -- You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: #543 (comment)

-- Sent from my Android device with K-9 Mail. Please excuse my brevity.

[snmcmillan]

@tlaurion

It's actually a TPM delay. And I think I've isolated it to the second measurement.

But yes, that's the issue, unless there's something else wrong with your config.

[tlaurion]

@SebastianMcMillan after around 50 seconds, from a clean build from your branch, with and without @flammit patch, there is a constant high pitch sound with lights flashing.

Will try to revive my beaglebone black and install screwdriver on it again.

Sent from my Galaxy S3 using FastHub-Libre

[snmcmillan]

@tlaurion

That's a RAM error. Try using a rubber eraser on the RAM contacts, and try every stick until it works. Trust me, I had this issue and it was a pain in the rear.

For reference, both this Repo's master and my Repo's master, as well as main branch coreboot are picky about this. However, all 3 of those work after trying almost every single stick of DDR3 RAM in my house. Try every combination of RAM in every slot.

[tlaurion]

@SebastianMcMillan flashing back original rom works.

Sent from my Galaxy S3 using FastHub-Libre

[snmcmillan]

@tlaurion

That's expected. Coreboot is picky. VERY VERY picky unlike stock in the case of the X220.

[techge]

I wanted to report this anyway, but did not find the time yet. I will try the waiting thing these days... I just wanted to confirm that I tried multiple times to build and flash on x220 and got a blank screen with heads while coreboot works fine for me.

[snmcmillan]

@techge wait 50 seconds.

Since Coreboot already works for you, you aren't affected by this issue, but instead you're affected by #541

[snmcmillan]

@tlaurion any update on your problems? I'm currently running a build from this repo's master on my X220, and although I haven't tested it on X220, my repo should, in theory, boot as well since the changes I made to X220 are the exact same changes I made to T420 and it boots fine. Both builds Only issue either of the builds have is the issue above, but they will work, given you get past the X220 being overly picky on RAM init with coreboot.

Currently setting up a new heads environment for my X220 to test my changes.

[tlaurion]

EDIT: I've swapped x220 memory from x230 to x220 and joy happened. Had 50 seconds delay before heads showing up. It was a memory issue with coreboot being pickier then stock kernel as Seb said.

Tried to debug with beaglebone black UHCI dongle, but can't get the serial output when booting heads and don't know why but ran out of time.
Instructions are not clear from beaglebone black screwdriver EHCI debug dongle, driver needed is not the one loaded by default and needs to be changed. rmmod -f g_multi, modprobe g_dbgp fixed it.
Got serial output from the x230 coreboot clean build default config with UHCI debug options added in (console output to usb dongle, letting the default UHCI dongle). Might have time to dig this deeper while removing NO_GFX into coreboot and bisect the changes between coreboot defconfig of x230 and heads x230 coreboot config until I get debug output on EHCI deebug dongle, but I can't guarantee..

Meanwhile, since x220 boots heads, cbmem seems to be a good way to debut things up. Apply this and add debug options as needed.

@Akendo @paulmenzel : Do you still have a x220? You have that 50 delay?

Notes:
The coreboot option to delay TPM init is supposed to dodge a race condition but had no effect.
@Akendo that ported this to the x220 initially never reported such bug. If I recall well it was for an older coreboot version but didn't verified. Maybe the patch was ported blindly.

[tlaurion]

My non booting and beeping was related to considered bad memory from coreboot. I now have the 50 seconds delay before Heads boot.

[tlaurion]

@SebastianMcMillan : Once snmcmillan#1 is merged in, you can close this issue since #541 is the real culprit here. I referred the cbmem PR there.

[snmcmillan]

It's been merged, you'll need to close as you're the owner of the issue.