Re: Updated Debian Ports installation images 2021-09-23
On 10/5/2021 9:16:40 AM, John Paul Adrian Glaubitz wrote: Hi Marc! On 10/5/21 18:10, Marc Zyngier wrote: > Many thanks for keeping Debian alive on these old architectures. You're welcome. > I just dusted off my XServe-G5 (ppc64) and gave this snapshot a go. > > The very good news is that it boots all the way to the installer. > The less good news is that it stops pretty quickly as none of the > required drivers are actually in the install media. > > To be able to install Debian, you would at least need: > CONFIG_SATA_SVW (sata) > CONFIG_TIGON3 (Ethernet) Alex Perez: While you're at it, if you could also add SATA_SIL24 to the default build, this is, by far, the most common legacy SATA PCI controller that's used in legacy PCI-based PowerPC Macs.
Re: X stopped working with 5.14 on iBook
Stan, FYI, There are nine powerpc-specific changes in 5.13. It may make sense to start rolling back those commits first: https://kernelnewbies.org/Linux_5.13#POWERPC On 10/6/2021 10:00:16 AM, Stan Johnson wrote: On 10/6/21 4:25 AM, John Paul Adrian Glaubitz wrote: > On 10/6/21 01:40, Riccardo Mottola wrote: >> Again, booting an old kernel leads to working X. > Very much sounds like a kernel regression that you should bisect. > > That isn't too difficult when cross-building with a fast x86_64 machine. > > Adrian > I can confirm that X is not working on a PowerPC G4 Cube with the stock Debian kernel "vmlinux-5.14.0-2-powerpc". X does work with the stock Debian kernel "vmlinux-5.10.0-8-powerpc". In both cases, I'm using the latest Debian SID with Xfce. In 5.14.0-2, wdm runs but the system console screen remains blank (the LCD backlight is on). (BTW, I have a PowerBook G4 12" 1.5GHz, but I didn't test the 5.14 kernel there. The PB G4 overheats and shuts off whenever it runs anything CPU-intensive, such as a Gentoo upgrade or even compressing a file. I think Apple had a way of throttling the CPU speed to prevent overheating -- it doesn't have any problems with Mac OS X Tiger or Leopard -- but they probably didn't share the details.) I'll attempt a bisect, starting with mainline v5.14 and going backwards if necessary (it looks like 5.12.9 worked; I'm not sure about v5.13). If there's a kernel regression, I'll confirm the regression exists on both the G4 Cube and the PB G4. I don't have an iBook to test. -Stan
Re: Poll: Is anyone using nouveau driver on ppc64 system?
Adrian, John Paul Adrian Glaubitz wrote on 10/8/2021 2:24 AM: Hi Thomas! On 10/7/21 23:41, Tom Grzybowski wrote: I have a related question: instead of us breaking what is working by updating the kernel regularly, why don't we find a good working kernel version and move on only when we can match that in terms of functionality? Is there some "requirement" to always feature the latest kernel? That's unfortunately not possible at the moment as Debian Ports is a pure unstable distribution which means we're rolling release. Can't you still be rolling-release, but with an LTS kernel? There are plans to add releases to Debian Ports but that would require setting up an instance of Debian's release management software called Britney [1] which is something someone from the release team needs to do. Perhaps there would be fewer barriers if someone were to prepare a "spin" of BE-ppc/ppc64 that is outside of the official Debian Ports. It seems like it would be possible to run our own instance of Britney2: https://release.debian.org/doc/britney/setting-up-britney.html ...but then the question would become "what else needs to be modified to make use of it".. Adrian [1] https://release.debian.org/doc/britney/
Installation report - Successful install via netinst images on Beige G3
I was successfully able to install Debian 11 to a 266MHz Beige G3 (Old World) using the 2020-10-06 NETINST image[1] Because this is an Old World machine, I copied the initrd and kernel from the CD to my System boot volume, and loaded them using BootX 1.2.2. Note that the initrd in the ISO below is much larger than the default setting for initrd size in BootX 1.2.2, which defaults to 8192 kilobytes. You must set it to 20,000,000 or thereabouts, or the first boot with BootX will fail. Since the old world macs, don't support GRUB, once you've performed a successful install, you still need to copy over the installed kernel/initrd to a place where you can use them to boot via BootX, from the Mac OS side. [1] https://cdimage.debian.org/cdimage/ports/snapshots/2021-10-06/debian-11.0.0-powerpc-NETINST-1.iso
Re: Installer doesn't find the mirrors
This does not appear to have happened (new images) Is it something you can automate, or quasi-automate? On 4/9/2023 11:16:01 PM, John Paul Adrian Glaubitz wrote: > On Apr 10, 2023, at 4:54 AM, Jeffrey Walton wrote: > > On Sun, Apr 9, 2023 at 6:38 PM Ken Cunningham > wrote: >> >> Your issue is in French, but it looks like the one I recently faced as well >> if I am reading it correctly. >> >> I fixed it using the information on this page: >> >> https://unix.stackexchange.com/questions/75807/no-public-key-available-on-apt-get-update > > Ugh, Debian has chronic problems in key management for ports. I've got > half a dozen to a dozen Chroot's for testing that have been broken for > months. There are no »chronic« problems. It’s intentional that keys expire. > I really wish they would fix them and stop this madness. I will create new images next week which addresses this problem. Adrian
Successfully installed Debian 10 on Lombard G3 PowerMac
Hi folks, I just wanted to share that I was able to succesfully boot and install Debian 10 using the "official" Debian 10 netinst CD from [1], with the exception of the bootloader, which failed, and I had to handle manually. I also tried booting from the latest snapshot CD image [2], which completely fails to boot, I assume due to the lack of yaboot on the latest image, which is unfortunate, as this appears to be the only way to boot Lombard-based systems, due to the version of OpenFirmware. Despite the hardware being 23 years old, it's usable with Debian 10, and the on-board 100 megabit Ethernet worked fine, out of the box. I'm booting from a 32GB SanDisk Ultra CompactFlash card, installed in the 2.5" drive bay. Subsequent to the install, I did an update to unstable, and am running kernel 5.9 without issues. Even Xorg starts, and works (ATI Rage LT), albeit with a bit of "wrong color" pixels here and there. I'm not sure why, but it's tolerable as-is. So, I'm wondering what the rationale for removal of yaboot was, given that it excludes a class of machines that were produced in large quantities, where Debian otherwise works fine, except for GRUB/yaboot support. [1] https://cdimage.debian.org/cdimage/ports/10.0/powerpc/iso-cd/debian-10.0-powerpc-NETINST-1.iso [2] https://cdimage.debian.org/cdimage/ports/snapshots/2020-12-03/debian-10.0.0-powerpc-NETINST-1.iso
Re: Updated installer images 2020-12-02
Adrian, John Paul Adrian Glaubitz wrote on 12/28/20 11:58 AM: >> Yet at the end, no apparent important errors, not even noticeable at any step >> in console, the request of reboot yields a flashing OS folder. > You should try to install one of the Yaboot images to make sure there is no > problem > with your machine. > Which yaboot images, and where?
Steps to boot a fresh PowerPC32 install of Debian from GRUB on installer CD
If you've successfully gotten 32-bit PowerPC Debian installed using one of the recent Debian Ports snapshot CDs (I used https://ftp.acc.umu.se/cdimage/ports/snapshots/2020-11-09/debian-10.0.0-powerpc-NETINST-1.iso) but have no installed bootloader (grub install fails), here's what you can do to temporarily boot in to this installed environment: 1) After the install completes, re-boot _with the installation CD still in the drive_ 2) Break in to GRUB, and type 'c' to get to a grub command line 3) Run the following commands from the GRUB CLI, substituting apple3/sda3 for the appropriate volume where /boot was installed to. insmod part_apple insmod ext2 'linux (ieee1275/hd,apple3)/boot/vmlinux root=/dev/sda3' initrd (ieee1275/hd,apple3)/boot/initrd.img boot ...and your machine will boot to the installed environment. I then installed the 'git' Debian package, and cloned the yaboot4 git repo: git clone --recurse-submodules https://github.com/rsaxvc/yaboot4 ...and then compiled it with 'make' ...and then did a 'sudo make install' ...and then manually edited /etc/yaboot.conf with correct arguments: image=/boot/vmlinux label=Linux root=/dev/sda3 initrd=/boot/initrd.img read-only ...then ran 'ybin' and rebooted. And it results in a working environment.
Re: Steps to boot a fresh PowerPC32 install of Debian from GRUB on installer CD
On 1/3/2021 4:54:18 PM, John Paul Adrian Glaubitz wrote: Hello Alex! On 1/4/21 1:36 AM, Alex Perez wrote: > If you've successfully gotten 32-bit PowerPC Debian installed using one of > the recent Debian Ports > snapshot CDs (I used > https://ftp.acc.umu.se/cdimage/ports/snapshots/2020-11-09/debian-10.0.0-powerpc-\ > NETINST-1.iso) but have no installed bootloader (grub install fails), here's > what you can do to temporarily > boot in to this installed environment: Are you sure you used this image? I'm asking because all newer images are known to have issues with partman on 32-bit PowerPC machines. Are you not seeing any issues with the partioner on 32-bit PowerPC? You know what...now that you mention it, I did a clean install _over the top of_ one I had partitioned with the 2018 powerpc32 netinst image. Therefore, I didn't exercise the disk partitioning code path in this image at all. I guess that's a reasonable way to work around the current breakagepartition with the old installer, then reboot with a newer one after the partitioning step is complete. Adrian
Re: More debug infos - was: Re: iMac G5 "windfarm"
John Paul Adrian Glaubitz wrote on 1/4/2021 1:56 AM: My guess is that this issue can be worked around by zeroing the partition table of the system before installing Debian. Can someone give it a try? I did so, zeroing out my drive first, with the latest 2020-01-03 netinst image, and it seems to maybe get a little bit further, but still hangs. In the TUI, it hangs at 52% instead of ~47% percent, as it did before. It seems to stall in when executing /lib/partman/init.d/35dump (which seems to never exit) All the above nine-line script does is cat the device strings (via /var/lib/partman/devices/*) and then the debian-installer stuffs this in a dialog box, presumably to select the drive to install to. Since there's only one drive on my system, and I've verified I get expected values from the one entry for sda, I'm not sure why it's hanging, but it's progress of a sort. Adrian
Re: Radeon HD 5770
Johannes, Johannes Brakensiek wrote on 1/26/2021 10:47 AM: Hi, On 26 Jan 2021, at 15:24, luigi burdo wrote: i had been test linux with G5 Quad . Quadro 7800Gtx in 16x slot, radeonhd on 8x slot. Result, it work with mesa 3D enabled on 4650, 5450, 6570, and R5 220. only 2D working with SI cards 7750, r7 250 because mesa crash for endianess someware. For have the PC bios card working you need 2 Monitors. one monitor for the apple firmaware gpu and one for the "pc firmware gpu". This is a video of my g5 with Debian Port (thanx Adrian Glubitz) and radeonhd 6570 https://www.youtube.com/watch?v=QsB5xlLyZwE&t Ciao Luigi Da: John Paul Adrian Glaubitz Inviato: martedì 26 gennaio 2021 15:14 A: Linux User #330250 Cc: Johannes Brakensiek ; debian-powerpc@lists.debian.org Oggetto: Re: Radeon HD 5770 Hi! On 1/26/21 3:06 PM, Linux User #330250 wrote: If the card doesn't have a Firmware with FCode for the Open Firmware of the Power Mac, it's won't be possible to see anything prior to the Linux boot, where Linux initializes the card using the PC-VideoBIOS. (...) The problem is that without an FCode graphics ROM you also don't get any error messages, neither from the Open Firmware (like when it doesn't find a bootable operating system: the folder symbol with the question mark in it) or from the early kernel. (early printk) I would probably just use a serial console for the boot sequence. Adrian thank you for all your replys. Sounds as if it works if I’ve already got a working system (which I have). And it seems to be even easier when using both graphics cards in parallel. I did not know I can use a serial console with a G5. Going to investigate that as well. It's possible, and basically involves re-purposing the modem interface, which is really just a serial port. See http://www.geethree.com/stealth/ for more information.
Re: OT: Huge Right to Repair Win for Consumers
Milan Kupcevic wrote on 6/10/21 6:10 AM: > On 6/10/21 12:53 AM, John Paul Adrian Glaubitz wrote: >> On 6/10/21 2:08 AM, Paul Wise wrote: The report and its recommendations may provide a means to pierce the veil of closed platforms, like closed-sourced firmware. >>> >>> It seems unlikely to me that we will ever see a "Right to Repair" for >>> software, firmware or gateware. >> >> So, why should laws protect the intellectual property of software >> companies >> but not the IP of hardware companies? >> >> What supporters euphemistically call a "right to repair" is in >> reality an >> initiative against the right of companies to protect their intellectual >> property. >> > Adrian, > > When you are bringing the question of property up, just ask yourself > what happens when you buy an item. Who is the owner of the item you've > just bought? Who decides from that point on how are you going to use > the item? Is it you, or somebody else? Has your property been > protected? Do you have any rights? At this point, you have effectively no protection. Here's a real-world example. I have two 1080p, Wi-Fi enabled HD security cameras that worked great for years, until FLIR Corporation decided they didn't want to be in that business sector anymore, sold the business unit to Lorex, and suddenly I have two 1080p doorstops instead of cameras. And incidentally, Lorex wants me to buy their new shiny cameras. These sorts of business practices are antithetical to good planetary stewardship. Any company who makes firmware for a device can abandon it, hope that a majority of the customers of the previous, abandoned (but not technically obsolete) product generation/line. IP Law is important, but it _should not trump all other considerations_, which, Adrian, seems to be what you're arguing for. If a manufacturer is going to abandon suport for a product mere years after they've introduced it, it should be their responsibility to ensure the majority of those devices do not end up in landfils, or scrapped for plastic and eWaste.
