ardo has quit [Read error: Connection reset by peer]
ardo has joined #crux
_moth_ has quit [Quit: _moth_]
<jaeger>
Could use a smaller boot media to install, probably
<SiFuh>
jaeger: maybe, actually I was going to ask you something about the ISO
<SiFuh>
How come there iare two kernels identical on the CD? One under boot with an initramfs and another in the rootfs.tar.xz ?
<jaeger>
compiled kernel image? or just the config?
<SiFuh>
Compiled. I did a checksum
<jaeger>
probably just oversight... I don't think one would be needed in the rootfs
<SiFuh>
Me too, just curious.
<jaeger>
I'll give it a look
<SiFuh>
Yeah, I was poking around. I am woking on trying to get the install CD in the EFI partition of the hard disk
<SiFuh>
But I run out of RAM when I try to load it :-P
<SiFuh>
h and this is 3.6.1 iso I am talking about. It was the only one that coould load on a system with 1 GB of RAM
<SiFuh>
I remembered that the wife has a Beelink BT3-PRO with 4GB of RAM I will try using that machine for 3.7.rc3
<SiFuh>
jaeger: lsusb doesn't work on the live CD. It says libatomic.so.1 cannot open shared object file
<SiFuh>
crux-3.7.rc3.iso
<jaeger>
hrmm, ok
<SiFuh>
dmesg says "Unknown kernel command line parameters "BOOT_IMAGE=/boot/vmlinuz", will be passed to user space.
<SiFuh>
And all that work I did, figuring out how to get mmc to work in 3.6 and you modified the kernel for it. No longer is working again in 3.7.rc3. So I can't install without making an entirely new installation kernel
<SiFuh>
Shame there is no exfat module either.
ppetrov^ has joined #crux
<jaeger>
despite what you might think, it's NOT intentional to remove or break functionality with a kernel update... we can add it again
<SiFuh>
I know, I was mentionging that I am now going to have to find out what it is for the new kernel
<SiFuh>
Some of the modules are still there, just something extra is missing.
<SiFuh>
jaeger: Yeah something definately must have changed because all the stuff I asked you to enable are still there. I had to load mmc_block manually
<farkuhar>
SiFuh: I appreciate all the testing you're doing. (On bare metal too! virtual machines wouldn't provide as rigorous a test.) Here I am just reaping the fruits of your labors and turning them into sections of the handbook.
<SiFuh>
farkuhar: I havea 4GB EFI system partition on a dinosaur machine and everything seems to be fine. Just wish it had more RAM
<SiFuh>
I think I figured out what is wrong with MMC again
<farkuhar>
didn't you say that SD cards are notoriously unreliable as storage devices? or does CONFIG_MMC=y enable other devices too, and you're using one of those as your boot medium?
<SiFuh>
Some small portable computers use MMC as their main drive
<SiFuh>
Such as jaeger's Zotac
<SiFuh>
This Beelink has the MMC physically soldered to the board
<farkuhar>
true story about SD cards: one day I went for a run, and I found a 16GB SD card on the ground just outside the gate. I put it in my pocket and finished the run, then forgot about it and put the garment in my laundry basket. The SD card was still readable after going through the wash!
<SiFuh>
SPI is widely used by microcontrollers to talk with sensors, eeprom and flash memory, codecs and various other controller chips, analog to digital (and d-to-a) converters, and more. MMC and SD cards can be accessed using SPI protocol; and for DataFlash cards used in MMC sockets, SPI must always be used.
<jaeger>
yeah, I was just out running errands and thinking about testing it on my zotac, haven't used that thing in ages
<SiFuh>
I can make 3 kernel configs but I am pretty sure only the first config will be enough
<SiFuh>
Oh and I enabled exfat as a module on that one as well
SiFuh has quit [Ping timeout: 268 seconds]
SiFuh has joined #crux
groovy3shoes has joined #crux
groovy2shoes has quit [Read error: Connection reset by peer]
ty3r0x has quit [Ping timeout: 240 seconds]
<dlcusa>
Farkuhar, thunderous applause for turning Sifuh's research into Handbook sections!!!
<farkuhar>
Ah, shucks, dlcusa. I'm just following your example in FS#1201 and elaborating on the excellent structure that was already in place.
<jaeger>
That lsusb libatomic thing is annoying, don't want to add gcc to the ISO if not needed
<jaeger>
I wonder if usbutils can avoid linking it
<SiFuh>
jaeger: I want to do a rebuild of crux 3.7 rc. iso.git seems to only be 3.6
<jaeger>
There's a 3.7 branch
<SiFuh>
Oh okay I think I found it
ty3r0x has joined #crux
<jaeger>
I have minor changes locally that aren't pushed yet but it's pretty close
<SiFuh>
How to get the 3.7 ports though?
<jaeger>
git or rsync
<SiFuh>
Same as 3.6?
<jaeger>
yeah
<SiFuh>
So 3.7 ports are no different from the 3.6 ports I already have? Same versions?
<jaeger>
Not sure what you mean... 3.7 ports are 3.7, there are a lot of differences from 3.6
ty3r0x has quit [Ping timeout: 252 seconds]
<SiFuh>
Well, then I do not know how to get them then. The URL should be different the pub file should be different, the rsync file should be different
<jaeger>
change 3.6 to 3.7?
ardo has quit [Read error: Connection reset by peer]
aardo has joined #crux
<SiFuh>
Hahaha okay, that easy then
<jaeger>
:)
<SiFuh>
So many ports to buil in opt. That could be reduce by a lot if there was an option for a core only CD. This was the 'opt' ports I had selected for 3.6.1 https://dpaste.com/EV4RCB9WU.txt and no xorg at all.
<jaeger>
yeah, I've been considering building a couple of extra ISO styles for testing... core-only is one, another would be netinst, and that one might have some crossover with some kind of minimal one for smaller memory usage
<SiFuh>
Wouldn't be too hard to figure out. That list I had made was from what the CD needed and from what I needed on a core only system.
<SiFuh>
Hopefully and after this ISO may successfully build, it shouldn't be too hard to test the kernel and tweak it a little. I'd also end up creating a new opt list. I got rid of xorg though on this one. Hopefully nothing from opt tries to pull in anything xorg. I never checked
<jaeger>
You could save time by grabbing the built packages from an existing ISO
<SiFuh>
Actually I have them here
<jaeger>
ok
<jaeger>
I guess I could just add libatomic without the rest of gcc to make lsusb work
<SiFuh>
I was thinking that but then I was wondering, what about static build
<farkuhar>
SiFuh, these packages in opt have dependencies in xorg: at-spi2-atk at-spi2-core cairo firefox-bin gtk gtk3 harfbuzz openbox pango xterm
<SiFuh>
Cheers farkuhar
<farkuhar>
It's "grep -F" that deserves the credit here, and the file "setup.dependencies" on the install ISO.
<SiFuh>
Some of them were pretty obvious. Hah, I should have looked
<SiFuh>
jaeger: Doesn't look like static will work.
<jaeger>
Why wouldn't it?
<SiFuh>
I got a few hundred lines of errors
<jaeger>
doh
<SiFuh>
/usr/include/libusb-1.0/libusb.h:1365:18: note: declared here 1365 | void LIBUSB_CALL libusb_set_debug(libusb_context *ctx, int level);
<SiFuh>
jaeger: seems to be libusb that is relying on libatomic
ppetrov^ has quit [Quit: Leaving]
<SiFuh>
farkuhar: by the way ENCRYPTED is what ever you name it
<SiFuh>
I used my hostname machupo
<farkuhar>
SiFuh: yes, I should probably insert a comment about that
<SiFuh>
That was actuall an install I did and logged. Also you should note that install has a separate swap partition generated on boot through an rc script. So randomized swap each boot
<SiFuh>
farkuhar: CONFIG_CRYTPO_USER_API_SKCIPHER <-- should already be set. jaeger did that in 3.6.1
<SiFuh>
lz4 is what I used for kernel compression. I thought it was going to be added to core in 3.7 but I guess not.
<SiFuh>
You can shorten the dracut line to this and by pass 'mv' all together dracut --kver 5.15.55 /boot/efi/BOOT/initramfs-5.15.55.img
<SiFuh>
You can remove this efibootmgr
<SiFuh>
and bypass (erase) the 'mv' line*
<farkuhar>
Thanks SiFuh! I've revised the example FDE instructions accordingly.
<SiFuh>
farkuhar: there are so many options with encryption though. It might be best to introduce user install examples?
<SiFuh>
Arch has an extensive wiki on it. After a few lines you just go 'blurr' because they seem to be talking gibberish.
<SiFuh>
Unfortunately there is no security. Your password is accessible to anyone using your system and it is built into the kernel which is on a non enecrypted drive.