09:01 | suihkulokki> | agraf: lets say I've just compiled edk2 way too many times for mcbin :) |
09:01 | suihkulokki> | agraf: but my notes are slightly outdated, haven't tried 18.06 branch |
09:51 | agraf> | suihkulokki: got a link? ;) |
09:51 | agraf> | suihkulokki: last time i built the firmware is already more than half a year ago |
09:51 | agraf> | suihkulokki: and I'm quite sure things have changed in between |
10:59 | Ke> | agraf: the filename for SCP_BL2 has changed and now it's at least mandatory |
11:05 | suihkulokki> | agraf: let me try build and run 18.06, good chance to walk through and clean notes |
11:33 | agraf> | suihkulokki: thanks :) |
13:13 | suihkulokki> | what's the voltage on mcbin internal uart headers? 3.3v ? |
13:19 | Ke> | yes |
13:19 | Ke> | again they are not configured to be uarts by default |
13:38 | suihkulokki> | agraf: gotta skip the test part, so here's just updated build instructions https://people.debian.org/~riku/mcbin/edk2-build.txt |
13:39 | suihkulokki> | microusb UART header broken for good it seems |
13:42 | agraf> | suihkulokki: ouch, thanks a lot! |
17:54 | agraf> | suihkulokki: are you using upstream kernels with that? |
17:54 | agraf> | suihkulokki: seems to explode left and right for me :) |
17:55 | Ke> | agraf: debian buster kernels |
17:55 | Ke> | which is upstreamish |
17:56 | Ke> | for me boot frequently fails, but after successful boot, there is moderately good stability |
17:56 | agraf> | Ke: hmhm, with 4.17 DT based boot just stalls and ACPI based boot crashes in the ACPI interpreter |
17:56 | Ke> | always? |
17:57 | agraf> | I can try again I suppose ;) |
17:57 | agraf> | you're saying it's random memory corruption? |
17:57 | Ke> | or a race condition or any nondeterministic bug whatsoever |
17:58 | Ke> | but definitely nondeterministic for me |
17:58 | agraf> | the DT one occurs exactly the same way again |
17:59 | agraf> | looks like linux fails to properly make use of secondaries? |
18:00 | agraf> | no, that's not it - still broken ;) |
18:03 | agraf> | ACPI breakage is also deterministic |
18:03 | agraf> | so I guess that's progress - it's deterministically broken :) |
18:10 | Ke> | only, if you plan to fix it |
21:59 | suihkulokki> | agraf: something regressed then. Try with |
21:59 | suihkulokki> | whatever was the kernel option to keep clocks running |
22:00 | suihkulokki> | agraf: try appending clk_ignore_unused to kernel command line |
22:00 | suihkulokki> | I'm using DT but haven't updated from 4.16 kernel |