[openchrome-users] openchrome drm on 5.14-next branch: lockup

[Kevin Brace]

Hi Jackson,

First of all, sorry for causing needless troubles on your end.
Unfortunately, when I converted OpenChrome DRM code to support atomic mode setting back in November 2020, I forgot to test the code on a laptop with the supported VIA chipset.
I developed the code on ASUS P5V800-VM mainboard with Intel Pentium D 2.8 GHz.
I used to develop OpenChrome DRM on HP 2133 mini-note for some time, but since it took about 30 minutes to build the kernel assuming that the Linux kernel is already precompiled, I transitioned to ASUS P5V800-VM mainboard since it can now do the same in about 8 minutes (as a developer, I do like fast compile times).
I know that VGA only configuration does work (otherwise, the code would not have been converted), but I only became aware of this issue only around May 2021 when I installed 64-bit Gentoo Linux to Averatec 2260 (K8N890 chipset).
I was hoping no one would have noticed it, but unfortunately, you got hit by it.
I could not really allocate adequate time to work on OpenChrome since late October 2020 due to sprinkler related flooding incident that happened near my apartment unit around the same time.
I got the atomic mode setting code prototype first working shortly before the sprinkler incident.
    You are correct about the buggy behavior OpenChrome DRM is exhibiting.
Even if FP is not properly detected, the system is not dead since the keyboard LED is still working, and you can still restart with Ctrl + Alt + Delete.
When I first noticed this issue on Averatec 2260, I could use the VGA due to the connector being bent (it probably happened with the previous owner), and was having issues with recording /var/log/kern.log.
I solved the first issue by inserting a flat screw driver between the chassis and the VGA connector.
The second issue was due to my lack of experience setting up a Gentoo Linux based system properly, and it was resolved by adding sysklogd to be started up automatocally by OpenRC each time the system is booted.
    Right now (at this moment, literally), I am working with a friend of mine trying to add support for NeoWare CA10 thin client's DVI (based on VT1632A DVI transmitter) to OpenChrome DDX UMS code.
I will likely have more time later this week to work on fixing FP (Flat Panel) issues with the OpenChrome DRM code.

Regards,

Kevin Brace
Brace Computer Laboratory blog
https://bracecomputerlab.com


> Date: Mon, 19 Jul 2021 17:35:39 -0400
> From: Jackson <dragonwarriorxtreme at gmail.com>
> To: openchrome-users at lists.freedesktop.org
> Subject: [openchrome-users] openchrome drm on 5.14-next branch: lockup
> Message-ID:
> 	<CAFenMTXAyx+KupH2zeJS2MfB_J7-bZJWRWZFxuOUZm-Pq7e7fg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> This happens regardless if I am switching from VESA/simpledrm or from VGA
> text console.
>
> When I set openchrome.modeset to 1 via modprobe, drm_ttm_helper (module)
> and openchrome (also a module) load, then nothing else happens; there is a
> lockup... but not even a black screen. Although, it's a soft lockup, since
> the shell is still functioning and I can type commands. However, trying to
> reboot or shutdown causes the system to go into a hard lockup, leaving no
> choice but to force the power off.
>
> Additionally, switching from the VGA text console also causes the VGA text
> console's characters to become heavily spaced out, with dot crawl happening
> rapidly on the screen... if that symptom is worth noticing.
>
> For what it's worth, this is a VIA C7 optimized kernel at -O2, and I am
> running it on an HP Mini 2133.