On 1/24/20, Stefan Lehner <stefan-leh...@aon.at> wrote: > Hi! > I have a very unusal setup: a old Handheld PC with a StrongARM CPU on which > i like to get Debian 9 running. > > > Short summary of my Kernel: > 4.9.210 with BX emulation patch (to get things going on the old ARMv4) > CONFIG_AEABI=y > CONFIG_OABI_COMPAT=y > I can send a the full .config i fit necessary. > > The armel rootfs is copied onto a CF card with debootstrap Then I generated > the devices and set some things linke hostname, locale . > > > $ debootstrap --arch armel --foreign stretch /mnt/jornada > http://ftp.de.debian.org/debian > $ cp /usr/bin/qemu-arm-static /mnt/jornada/usr/bin > $ LANG=C.UTF-8 chroot /mnt/jornada qemu-arm-static /bin/bash > > $ /debootstrap/debootstrap --second-stage > > $ apt install makedev > $ mount none /proc -t proc > $ cd /dev > $ MAKEDEV generic > > And here is the problem: the kernel boots up fine, mounts the rootfs. After > a while it displays crng init done and thats it. Nothing more . > Can somebody help me? > > Here is the console output: > > --------linexec-------- > < SNIPPED! > > > SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1 > NR_IRQS:16 nr_irqs:61 61 > irq: Cannot allocate irq_descs @ IRQ1, assuming pre-allocated > irq: Cannot allocate irq_descs @ IRQ33, assuming pre-allocated
Upon further reflection, GRIN, what I shared was only a possible work-around *if* either even works. Doesn't *fix* whatever is happening. DISCLAIMER about this is: Yes, I see the following is specifically talking about cherryview-pinctrl. I'm looking at the error message and decided maybe what they're talking about can be tweaked and tested in this case. You said it's an unusual setup so there you go right there. Maybe no one's ever encountered this to be able to have fixed it before now. Here's hoping! :) Those two lines above came back to mind. The use of "assuming" doesn't mean that something useful was in fact invoked somehow. We're talking about Stretch so it seems ok to share a late 2017 kernel.org thread about those very lines. Rational for highlighting the following is that this might still be age appropriate for Stretch: https://lore.kernel.org/patchwork/patch/836748/ If it looks useful, there's a *clean* patch at the very bottom that maybe could be personalized per this setup *if* that fix is not already in this instance's kernel (?). One last peek found the words "still may fail" near the top so I'm feeling ok to go ahead and send this out. Next stop: Another romp around the Net to see if I can't find something to be able to finally actually play with this firsthand as you all chat up these topics. :) PS I saw the other *lore* thread after a search using "StrongARM". Am glad I didn't start with that search because I might have stopped right there instead of tripping over the above patch. Maybe the error they're wanting in their last post is those "assuming" lines because that was part of what was fixed in the above PATCH a couple years ago. Would be really cool if that helped speed things up somehow. Cindy :) -- Cindy-Sue Causey Talking Rock, Pickens County, Georgia, USA * runs with birdseed *
