Random Linux fix: machine would not obey changes to efibootmgr... kept on repopulating (old) grub entry. Fix was to delete the folder in /boot/efi manually. Also UEFI sucks. #Linux#efibootmgr#uefi
This has been my only significant concern with Framework laptop ownership thus far. (There hasn't yet been a stable firmware release for Linux users in the 13+ months I've had my laptop.)
Zagłębiam się w kontrolę obrazu (wyświetlany na ekranie) na poziomie UEFI a potem systemu operacyjnego. Ale mam masę pytań.
Istnieje GOP (UEFI Graphics Output Protocol).
Wydaje się, że on kontroluje obraz, na etapie od naciśnięcia przycisku POWER na komputerze.
W pewnym momencie kontrolę przejmuje sterownik karty graficznej.
Czy tak jest? Czy dzieje się tam coś jeszcze pomiędzy?
Czyli karta graficzna jest w trybie GOP, a potem w trybie w którym obraz kontroluje sterownik karty graficznej?
Interesuje mnie kwestia rozdzielczości i częstotliwości odświeżania obrazu nadawanego przez kartę graficzną, do ekranu.
W jaki sposób GOP to kontroluje? Rozdzielczość i częstotliwość odświeżania obrazu?
Gdy włączę OSD monitora będąc w UEFI, to widać tam 2560x1440 60 Hz.
W pewnym momencie, przy chyba przełączaniu na system operacyjny, na chwilę gaśnie ekran, a potem wraca i już ma ustawioną właściwą rozdzielczość i częstotliwość odświeżania (2560x1440 165 Hz).
Dlaczego musi dojść do tego zgaśnięcia ekranu? Czemu to nie może być płynne?
Czemu na etapie UEFI nie może być 2560x1440 165 Hz?
Zresztą, gdy ustawię w systemie 2560x1440 60 Hz (czyli taki sam jak mówi OSD, że jest przy UEFI), to ekran i tak, tak samo gaśnie.
Więc z czego to wynika? Jaki dokładnie mechanizm za tym stoi, na poziomie karty graficznej?
Po co mi ta wiedza? Z ciekawości...
Bo u mnie to gaśnięcie ekranu trwa 6 sekund, niby tak długi czas, to wina monitora, że on tak po prostu ma, i jest wiele monitorów które tak po prostu mają. Ale jest to minimalnie problematyczne, i przez to się tak zagłębiam w temat...
I'm delving into controlling the image (displayed on the screen) at the UEFI level and then the operating system level. But I have lots of questions....
I'm delving into controlling the image (displayed on the screen) at the UEFI level and then the operating system level. But I have lots of questions.
There is GOP (UEFI Graphics Output Protocol).
He appears to be controlling the image right from pressing the POWER button on the computer.
At some point, the graphics card driver takes over.
Is that so? Is there something else going on in between?
So the graphics card is in GOP mode, and then in the mode in which the image is controlled by the graphics card driver?
I am interested in the issue of resolution and refresh rate of the image transmitted by the graphics card to the screen.
How does the GOP control this? Image resolution and refresh rate?
When I turn on the monitor's OSD while in UEFI, it shows 2560x1440 60 Hz.
At some point, probably when switching to the operating system, the screen goes blank for a moment and then comes back with the correct resolution and refresh rate set (2560x1440 165 Hz).
Why does this screen blackout have to happen? Why can't it be smooth?
Why can't it be 2560x1440 165 Hz at the UEFI stage?
Anyway, when I set the system to 2560x1440 60 Hz (which is the same as the OSD says for UEFI), the screen goes out the same way.
So what does this mean? What exactly is the mechanism behind this, at the graphics card level?
Why do I need this knowledge? Out of curiosity...
Because in my case the screen goes off for 6 seconds, it seems like such a long time, it's the monitor's fault, it's just like that, and there are many monitors that just have it like that. But it's minimally problematic, and that's why I'm delving deeper into the topic...
Just in case anyone is interested, and for the archives/searches, I recently asked if anyone had managed to use #LLM models to access #UEFI interfaces, or other interfaces without #A11Y, as a #blind user. The idea was to use a capture card to bring in the video information from the inaccessible machine, send pictures from that video stream to the LLM, and get descriptions/ask questions. This is how I did it. It's not pretty, but it's another helpful tool for the toolbox. It requires a video capture card, HDMI or display port to USB, the OpenAI #NVDA add-on, and a method of displaying the video from the capture card on screen. I tried four HDMI capture cards and all of them worked, I think the point is that the capture device should show up to Windows as a webcam. I haven't found a cheap capture device which didn't, the only reason I had to try four was that I was using audio input from the HDMI for another project and it's surprising how many devices will not receive the sound even in simple stereo. Anyhow, just searching for HDMI capture on google/amazon will probably get something to use. The Open AI NVDA Add-on is at https://github.com/aaclause/nvda-OpenAI/ The method I used to display the received video is at https://superuser.com/questions/1744688/how-can-i-view-the-video-coming-in-from-a-capture-card-on-windows-in-full-screen The steps are basically to put the puzzle pieces together. Set up the add-on with its instructions, copy and paste the HTML in the superuser link to a new HTML file, and open that HTML file in the browser. Having the file run from file explorer works fine, and firefox, at least, will ask for permission so make sure to allow it. Now, move the NVDA navigator cursor/focus to the video. Here, the object is called "document", the point is to avoid sending the entire screen, or even the firefox window. Having pressed the add-on command to capture the object, you will be placed in the prompt field and can ask any questions you like or rely on the default "describe this image" prompt. Generally, I will use the describe the image first and then ask follow-up questions or modify the image as best I can. Just a few tips. Maximizing the window and pressing the "full screen" button in firefox on the video appears to be helpful. The GPT 4 vision model does confabulate/hallucinate, and what it makes up is plausible. This is just another tool, not something to rely on exclusively. It is in addition to, rather than instead of, OCR, one's own knowledge, etc. The image is sometimes cut off, I'm not sure why this is but suspect at least some of it comes from its being displayed on the screen in the browser. I would welcome better ways to do this, as I said, it's not pretty and just what I could come up with in a few minutes of searching and with some trial/error. Having said that, it is a small step forward. Note that, as one would expect, the method also works to bring in pictures from a standard webcam. #nvdasr#ScreenReader
I know this is quite odd, but it's really just a shot in the dark so I don't reinvent the wheel. I'm sure someone has already done this and undoubtedly done it better than the way I want to. I am #blind and want to bring the HDMI output of one computer, running its UEFI configuration interface, to another computer, and then send the image of that output to OpenAI's #LLM so it can tell me what is selected, what is on the screen, and so on. This is so I can get access to the UEFI on machines, both to install systems and during those times when the machine doesn't boot and there's no sighted person around. I know of no #a11y method for blind users with #UEFI. My thinking is that a capture card would allow this. Has anyone managed this sort of thing on a windows machine? I don't mean to limit to the UEFI input, any sort of visual input from which static images are routed to a LLM from a capture card/visual input would be good to hear about. If so, I would be very grateful for any ideas on both the card and software to use, particularly so that the image is clear to the model. Thanks for having a look at what I'm sure is a very strange request.
I cant seem to get sound running under #linux on my machine...
Windows recognizes the onboard soundcard, but linux doesn´t, no matter which distro i am trying. Every singel one says dummy output. I tried restarting pipewire, alsamixer and pulseaudio. I tried another kernel, I reinstalled everything ...nothing. and windows is like: yup, there is the soundcard! Ahhh so frustrating...any recommendations what i can do to get soud work? #linux#support#fedihelp
@krevedkokun shared a cool thread about uki-bootloader implementation for guix, which brings alternative to grub and better encrypted root support. The solution has some drawbacks, but still very nice to see this work done!
> Multiple vulnerabilities were discovered in the TCP/IP stack (NetworkPkg) of Tianocore EDKII, an open source implementation of Unified Extensible Firmware Interface (UEFI). Researchers at Quarkslab have identified a total of 9 vulnerabilities that if exploited via network can lead to remote code execution, DoS attacks, DNS cache poisoning, and/or potential leakage of sensitive information. #Quarkslab have labeled these set of related vulnerabilities as PixieFail.
If you're at #37C3, go check out the self-organized sessions. I have submitted an intro to #oreboot and a #LinuxBoot/u-root/cpu workshop.
And at the Open Source #Firmware Foundation (OSFF) assembly, I'll run a little a #UEFI exploitation session.
The recent #UEFI#vulnerability - the decoder for the #boot#logo image is buggy, so you can fashion an image that will get full pre-secure-boot control of the machine at the #firmware level - seems to be getting way too much credit. Am I wrong here?
To be able to drop your specially-constructed image in the EFI system partition on a Linux machine, you need root privileges. So you need a chained remote-root vulnerability first.
Image control at the UEFI level and then the operating system level - I'm exploring the topic...
I'm delving into controlling the image (displayed on the screen) at the UEFI level and then the operating system level. But I have lots of questions....