This project could get us rid of the last fbdev drivers we're stil shipping: efifb and vesafb.
Platform drivers handle hardware that is not auto-detected, but somehow there. For graphics this would be VGA, VESA, or UEFI framebuffers. In SUSE Linux, we currently support VESA and UEFI with fbdev drivers. Those load early in the boot process and maintain graphics hardware until the actual driver takes over.
It would be nice to support VESA and UEFI with modern DRM drivers. Doing so would give us basic support for very old, very modern or very obscure graphics hardware. This in itself is not complicated. The problem is the handover to the actual HW driver. For the old fbdev, there's a mechanism for the HW driver to disable the platform driver. Not such thing exists for DRM. Implementing this mechanism would be the main goal of the project.
This project is part of:
Hack Week 19
Activity
Comments
-
almost 6 years ago by vliaskovitis | Reply
Interested though I only have limited kms experience (I have tested vkms/vgem, igt tools, and written a small vkms-like module in the past). Are the proposed drm drivers supposed to follow the fbdev->drm conversion from the fbconv helpers work https://gitlab.freedesktop.org/tzimmermann/linux/commits/fbconv , or are they meant to be from-scratch drm drivers? No idea about the handover mechanism though. If this is something that could be done and tested with VMs or a standard desktop without specific gfx cards (i assume so as vesafb is generic), and there are separable tasks that you don't plan on hacking, let me know.
-
almost 6 years ago by tdz | Reply
Hi!
I started working on this idea today.
The drivers are written from scratch. (1) The fbdev drivers are just trivial framebuffers that the screen image is being copied into. At least efifb cannot even do modesetting. It uses the display mode that has been set by the UEFI BIOS.
The hand-over mechanism is similar to how fbdev does it. The platform driver claims a piece of I/O memory, and once the actual driver wants to take the memory range, it kicks out the platform driver.
I was a bit optimistic about the number of fbdev drivers. There are 2 or 3 others besides efifb and vesafb. Yet, those would be a significant step forward.
I don't know about testing in VMs. If the current kernel's efifb works in the VM, the new driver should work as well.
Best regards Thomas
(1) It's nice that you recognized my work on fbconv. :)
-
-
-
almost 6 years ago by tdz | Reply
Day 2: Reviewing efifb and vesafb shows that these drivers don't actually do any modesetting or even call much of their respective interfaces. They only pick up a preconfigured framebuffer. Therefore I'm changing goal slightly. Instead of recreating these drivers for DRM, I'm porting over simplefb. It also picks up a preconfigured framebuffer, but it's entirely agnostic to the platform and works with firmware, bootloader or device tree. It even support EFI and VESA. AFAICT all the setup is done by the platform and the kernel's boot code. Just as simplefb, the new DRM driver would take these settings and make them available within the graphics framework.
-
almost 6 years ago by tdz | Reply
Day 3: After two days of tinkering with the platform setup code, I typed down the driver today. The kernel fetches the framebuffer from GRUB and provides it as framebuffer device. My driver picks up the device and provides it via DRM. So far, the fbdev console emulation works and I was able to get GDM partially working. Apparently, X11 wants a PCI ID, which the device doesn't have. But there's probably a way around this limitation. All in all, this looks really good so far.
Reply to tdz
And with a little weak, weston now works as well.
# A First Level Header ## A Second Level Header Use one asterisk to *emphasize* Use two asterisks for **strong emphasis** - Use hyphens - for unordereed - lists This is an [link to example.com](http://example.com/) This is an image  This is a user link @hans This is a project link hw#some-cool-title
-
-
almost 6 years ago by tdz | Reply
Day 4: I spent the day on cleaning up what I have and implementing the hand-over code. The hand-over is slightly tricky, as it requires to remove the DRM device while it's being used by the platform driver. And it's concurrent code that requires some locking. It's partially working ATM.
-
almost 6 years ago by tdz | Reply
Day 5: Today I finished the hand-over code. My test system is equipped with a Radeon graphics card. The bootloader, grub, configures a display mode before running the kernel. The built-in simplekms takes over early in the boot process and displays typical boot message. Once the radeon driver is ready, it kicks simplekms out and takes over the framebuffer. The hand-over is generic, so arbitrary drivers can kickout any platform driver.
If there's no native DRM driver for the hardware, simplekms remains active and provides a basic framebuffer for Wayland compositors or the framebuffer console. I did not get native X11 running, as X11 requested a PCI device. I did not look further into this problem. However, Gnome runs fine in Wayland mode and X applications can be started via Xwayland.
Yeah, I think this was a successful hackweek and the project is complete for now. After a bit more polishing, the driver could go upstream.
-
almost 6 years ago by tdz | Reply
I uploaded the current state of the project to the git branch at
https://gitlab.suse.de/tzimmermann/linux/tree/simplekms
-
-
-
Similar Projects
Backporting patches using LLM by jankara
Description
Backporting Linux kernel fixes (either for CVE issues or as part of general git-fixes workflow) is boring and mostly mechanical work (dealing with changes in context, renamed variables, new helper functions etc.). The idea of this project is to explore usage of LLM for backporting Linux kernel commits to SUSE kernels using LLM.
Goals
- Create safe environment allowing LLM to run and backport patches without exposing the whole filesystem to it (for privacy and security reasons).
- Write prompt that will guide LLM through the backporting process. Fine tune it based on experimental results.
- Explore success rate of LLMs when backporting various patches.
Resources
- Docker
- Gemini CLI
Repository
Current version of the container with some instructions for use are at: https://gitlab.suse.de/jankara/gemini-cli-backporter
Add Qualcomm Snapdragon 765G (SM7250) basic device tree to mainline linux kernel by pvorel
Qualcomm Snapdragon 765G (SM7250) (smartphone SoC) has no support in the linux kernel, nor in u-boot. Try to add basic device tree support. The hardest part will be to create boot.img which will be accepted by phone.
UART is available for smartphone :).
Improve UML page fault handler by ptesarik
Description
Improve UML handling of segmentation faults in kernel mode. Although such page faults are generally caused by a kernel bug, it is annoying if they cause an infinite loop, or panic the kernel. More importantly, a robust implementation allows to write KUnit tests for various guard pages, preventing potential kernel self-protection regressions.
Goals
Convert the UML page fault handler to use oops_* helpers, go through a few review rounds and finally get my patch series merged in 6.14.
Resources
Wrong initial attempt: https://lore.kernel.org/lkml/20231215121431.680-1-petrtesarik@huaweicloud.com/T/
early stage kdump support by mbrugger
Project Description
When we experience a early boot crash, we are not able to analyze the kernel dump, as user-space wasn't able to load the crash system. The idea is to make the crash system compiled into the host kernel (think of initramfs) so that we can create a kernel dump really early in the boot process.
Goal for the Hackweeks
- Investigate if this is possible and the implications it would have (done in HW21)
- Hack up a PoC (done in HW22 and HW23)
- Prepare RFC series (giving it's only one week, we are entering wishful thinking territory here).
update HW23
- I was able to include the crash kernel into the kernel Image.
- I'll need to find a way to load that from
init/main.c:start_kernel()probably afterkcsan_init() - I workaround for a smoke test was to hack
kexec_file_load()systemcall which has two problems:- My initramfs in the porduction kernel does not have a new enough kexec version, that's not a blocker but where the week ended
- As the crash kernel is part of init.data it will be already stale once I can call
kexec_file_load()from user-space.
The solution is probably to rewrite the POC so that the invocation can be done from init.text (that's my theory) but I'm not sure if I can reuse the kexec infrastructure in the kernel from there, which I rely on heavily.
update HW24
- Day1
- rebased on v6.12 with no problems others then me breaking the config
- setting up a new compilation and qemu/virtme env
- getting desperate as nothing works that used to work
- Day 2
- getting to call the invocation of loading the early kernel from
__initafterkcsan_init()
- getting to call the invocation of loading the early kernel from
Day 3
- fix problem of memdup not being able to alloc so much memory... use 64K page sizes for now
- code refactoring
- I'm now able to load the crash kernel
- When using virtme I can boot into the crash kernel, also it doesn't boot completely (major milestone!), crash in
elfcorehdr_read_notes()
Day 4
- crash systems crashes (no pun intended) in
copy_old_mempage()link; will need to understand elfcorehdr... - call path
vmcore_init() -> parse_crash_elf_headers() -> elfcorehdr_read() -> read_from_oldmem() -> copy_oldmem_page() -> copy_to_iter()
- crash systems crashes (no pun intended) in
Day 5
- hacking
arch/arm64/kernel/crash_dump.c:copy_old_mempage()to see if crash system really starts. It does. - fun fact: retested with more reserved memory and with UEFI FW, host kernel crashes in init but directly starts the crash kernel, so it works (somehow) \o/
- hacking
update HW25
- Day 1
- rebased crash-kernel on v6.12.59 (for now), still crashing
bpftrace contribution by mkoutny
Description
bpftrace is a great tool, no need to sing odes to it here. It can access any kernel data and process them in real time. It provides helpers for some common Linux kernel structures but not all.
Goals
- set up bpftrace toolchain
- learn about bpftrace implementation and internals
- implement support for
percpu_counters - look into some of the first issues
- send a refined PR (on Thu)
Resources