|
broken clock opsec posted:
Yawn. Call me when you get Mir working.
|
#
?
Apr 12, 2016 18:40
|
|
x server functional in windows 10 with ubuntu 
|
|
| # ? May 26, 2024 19:16 |
|
|
Lol why the gently caress doesn't CPIO support xattrs? What in the god drat gently caress?
|
|
#
?
Apr 12, 2016 21:25
|
|
|
It looks like it CPIO is too old, use tar.
|
|
#
?
Apr 12, 2016 21:45
|
|
|
ratbert90 posted:Lol why the gently caress doesn't CPIO support xattrs? What in the god drat gently caress? IBM paid red hat a lot of money to write a CPIO fork that could support files greater than 2GB.
|
|
#
?
Apr 12, 2016 22:12
|
|
|
MrMoo posted:It looks like it Initramfs built into the kernel doesn't support tar. Is it possible to use initramfs in tar format? Probably not.
|
|
#
?
Apr 12, 2016 22:59
|
|
|
Suspicious Dish posted:IBM paid red hat a lot of money to write a CPIO fork that could support files greater than 2GB. for aix users? mainframe? what on earth was this needed for?
|
|
#
?
Apr 13, 2016 00:25
|
|
|
ratbert90 posted:Initramfs built into the kernel doesn't support tar. Is it possible to use initramfs in tar format? Probably not. use initrd instead and then you don't have to worry about it
|
|
#
?
Apr 13, 2016 00:26
|
|
|
i'm installing centos on a vps that i'm switching over to, as is correct and proper how do I make tab completion for yum/dnf work with either zsh or bash? it doesn't seem to function out of the box, and since there's no aptitude equivalent and all the packages have their versions tacked onto the end, there's basically no way to install a package (that I've stumbled across yet)
|
|
#
?
Apr 13, 2016 01:19
|
|
|
install bash-completion; you can omit the versions
|
|
#
?
Apr 13, 2016 02:12
|
|
|
Notorious b.s.d. posted:for aix users? mainframe? what on earth was this needed for? they ship disk images as rpm files
|
|
#
?
Apr 13, 2016 06:44
|
|
|
Zom Aur posted:heads up ah so that's why firefox now looks complete batshit. thanks arch. tharch. you'd think someone would have tested this with a web browser, the one thing on linux where it actually makes sense to use a gui app
|
|
#
?
Apr 13, 2016 09:07
|
|
|
Wait, what the gently caress, RPM uses CPIO? I was actually just wondering the other day how RPMs/DEBs indicate how the tarball is compressed. I never imagined that the answer was that they don't use a tarball. (Well, DEB still might.)
|
|
#
?
Apr 13, 2016 14:31
|
|
|
Athas posted:Wait, what the gently caress, RPM uses CPIO? I was actually just wondering the other day how RPMs/DEBs indicate how the tarball is compressed. I never imagined that the answer was that they don't use a tarball. (Well, DEB still might.) deb uses ar seriously
|
|
#
?
Apr 13, 2016 14:59
|
|
|
Deb uses AR to combine the control and the data archives into one file. the archives themselves are tarballs that can be compressed with your favourite open sores fad compression.
|
|
#
?
Apr 13, 2016 16:44
|
|
|
broken clock opsec posted:
2038: civilization collapses
|
|
#
?
Apr 13, 2016 19:16
|
|
|
Gazpacho posted:2032: all restaurants are taco bell, all operating systems are linux im at the combination microsoft and linux shell
|
|
#
?
Apr 13, 2016 20:08
|
|
|
Notorious b.s.d. posted:use initrd instead and then you don't have to worry about it Not going to happen. SELinux Xatters aren't supported by initrd or initramfs as far as I can tell. Thanks Linux for making things lovely on the embedded side again!
|
|
#
?
Apr 13, 2016 20:15
|
|
|
an initrd is a block device, you can put whatever filesystem you want on it
|
|
#
?
Apr 13, 2016 20:17
|
|
|
Mr Dog posted:an initrd is a block device, you can put whatever filesystem you want on it But can it be the entire filing system?
|
|
#
?
Apr 13, 2016 20:18
|
|
|
if you really want you could probably write a statically linked initramfs containing just a handful of files that untar the "real" rootfs into ram pretty sure ramfs supports xattrs even if cpio doesn't. ramfs is just the linux inode and block cache without any backing store, it's a pretty sw8 design
|
|
#
?
Apr 13, 2016 20:19
|
|
|
Mr Dog posted:if you really want you could probably write a statically linked initramfs containing just a handful of files that untar the "real" rootfs into ram Yes, it does. However I want to preserve the xattrs into the ramfs image. CPIO fucks with this so initramfs is right out. Initrd would be the next best option I guess, just frustrating getting it going on the environment.
|
|
#
?
Apr 13, 2016 20:21
|
|
|
ratbert90 posted:But can it be the entire filing system? yes. it's just a block device. it's not a super efficient way of using ram because everything ends up in ram twice, and initramfs is a way better solution to the problem that initrds used to solve but if you really had to then you could use a ramdisk as the block device behind your rootfs. suggest using squashfs as the fs on your initrd, but it is a fundamentally read-only filesystem so you'll need to mount the rw bits of your filesystem elsewhere. mind giving a more detailed description of your use case?
|
|
#
?
Apr 13, 2016 20:23
|
|
|
Mr Dog posted:yes. it's just a block device. The use case is that I want a single file FS (That way updates to the product are easy in the future), and SELinux enabled (To lock the device down and prevent tampering, as telnet (Yes telnet, and no, I am not allowed to disable it) will be enabled. So my idea was originally to create a SELinux policy, set the file contexts before hand, create the initramfs, bundle it in with the kernel, and call it a day. But that's not going to happen. So perhaps I just need to really dig in and get initrd going.
|
|
#
?
Apr 13, 2016 20:54
|
|
|
LeftistMuslimObama posted:im at the combination microsoft and linux shell 
|
|
#
?
Apr 13, 2016 21:08
|
|
|
Yeah I'd say for your particular case go with a squashfs initrd and have it mount a ramfs on /var or whatever. Wikipedia sez squashfs supports xattrs and has done for some time. I have similar problems (minus the selinux-to-secure-telnet bit (lol btw)) Except my bootloader (U-Boot) is so loving bad that I ended up going with some awful three-stage setup because it loads the kernel and initrd at about 300KB/sec. You'd think this would be a solved problem. I investigated integrating OSTree with OpenEmbedded but holy poo poo it was a lot of work and my deadlines were way too tight to do an acceptable job of it.
|
|
#
?
Apr 13, 2016 21:11
|
|
|
disclaimer: Embedded Linux still is still slightly new to me so I might possibly not be the best person to ask.
|
|
#
?
Apr 13, 2016 21:12
|
|
|
Mr Dog posted:Yeah I'd say for your particular case go with a squashfs initrd and have it mount a ramfs on /var or whatever. Wikipedia sez squashfs supports xattrs and has done for some time. Uboot isn't bad, that's what embedded Linux uses for the most part to boot a kernel. And yes, the telnet thing is stupid, but this product is a stop gap between generations of product, and has to have a backwards compatibility mode to act and look like the old product, which has telnet.
|
|
#
?
Apr 13, 2016 21:18
|
|
|
Yeah don't get me wrong I like U-Boot for the most part but this particular build of U-Boot sux in that particular regard so unfortunately I have to live with it. Boot speed is not super important for this application but all the same I'd rather not spend literally two minutes in the bootloader, if for no other reason than it makes development frustrating.
|
|
#
?
Apr 13, 2016 21:23
|
|
|
broken clock opsec posted:
cool Cygwin could do the same thing also why would i want x server on windows
|
|
#
?
Apr 13, 2016 21:23
|
|
|
LeftistMuslimObama posted:im at the combination microsoft and linux shell 
|
|
#
?
Apr 13, 2016 22:00
|
|
|
ratbert90 posted:Yes, it does. However I want to preserve the xattrs into the ramfs image. CPIO fucks with this so initramfs is right out. Initrd would be the next best option I guess, just frustrating getting it going on the environment. that sucks xnu can use any supported filesystem on a disk image, including one entirely in RAM
|
|
#
?
Apr 14, 2016 01:36
|
|
|
LeftistMuslimObama posted:im at the combination microsoft and linux shell
|
|
#
?
Apr 14, 2016 01:44
|
|
|
eschaton posted:that sucks your derivative shaggar gimmick needs some work, try again
|
|
#
?
Apr 14, 2016 02:06
|
|
|
Mr Dog posted:your derivative shaggar gimmick needs some work, try again I don't think shaggars gimmick is licensed for that anyway
|
|
#
?
Apr 14, 2016 02:28
|
|
|
LeftistMuslimObama posted:im at the combination microsoft and linux shell new forum title please
|
|
#
?
Apr 14, 2016 02:30
|
|
|
Breakfast All Day posted:new forum title please 
|
|
#
?
Apr 14, 2016 03:44
|
|
|
ratbert90 posted:Yes, it does. However I want to preserve the xattrs into the ramfs image. CPIO fucks with this so initramfs is right out. Initrd would be the next best option I guess, just frustrating getting it going on the environment. initrd is literally just a compressed filesystem image expanded into memory. if you want xattrs or whatever, it's there. it's totally up to you. it sounds like you wanted initrd all along and built your poo poo around initramfs, a dead 1990s technology
|
|
#
?
Apr 14, 2016 05:17
|
|
|
eschaton posted:that sucks this has been the case for linux longer than "xnu" has been called "xnu" embedded developers just have terrible tunnel vision. (and i can hardly blame them for it: embedded dev promotes that viewpoint.)
|
|
#
?
Apr 14, 2016 05:18
|
|
|
Notorious b.s.d. posted:initrd is literally just a compressed filesystem image expanded into memory. if you want xattrs or whatever, it's there. it's totally up to you. Initramfs is newer than initrd.
|
|
#
?
Apr 14, 2016 06:12
|
|
|
|
| # ? May 26, 2024 19:16 |
|
|
initrd came first and was replaced by initramfs, hth
pseudorandom name fucked around with this message at 06:19 on Apr 14, 2016 |
|
#
?
Apr 14, 2016 06:13
|
|