similar to: Rationale for hashed .so?

(Apologies to klibc@zytor.com for the re-send) This patch makes gen_init_cpio generate the initramfs_data.cpio from a file which contains a list of entries: file, dir, nod. I swapped the order of filename/location for the file arguments so that it would be more uniform with the dir and nod tyes. [thayne@torch linux-2.6.8]$ usr/gen_init_cpio --help ERROR: unable to open '--help': No such

Hi, Here's a try at writing an initramfs HOWTO. This is basically a write-up of a number of interesting emails I collected over time. It could probably use an editor, more fact-checking and a bunch of other good things, but it should be better than nothing ;-) Daniel -------------- next part -------------- INITRAMFS HOWTO 0) What are klibc and initramfs? Initramfs is a ramfs into which

In case people here don't follow the udev mailing list, udev seems to be quasi-dropping klibc support. Well, see the forwarded message below. As I don't agree with the "omg udev is complex! use glibc!" rationale, I want klibc and udev to remain working with each other.... Assuming some minor breakage in the near future, is klibc open to accepting patches for "basic C

Hi I am looking at inserting functions in the early user space and want to use initramfs for this. I understand this can be achieved using 'initrd' or by making changes in the 'usr/gen_init_cpio.c' file. Can someone kindly give me pointers as to how to go about this. I am looking at the 2.6.x kernel. Thanks Rachita.

I've googled to try and find discussion of this topic but didn't see any results particularly interesting... Is there any particular reason the kernel has gen_init_cpio instead of just using cpio itself? I'm working on some changes that would involve just cpio-ing the entire usr/root subtree as the initramfs image (instead of gen_init_cpio having to know ahead of time what files to

Hi Peter & others. Here are three patches that does the following: #1 - Update kbuild part of klibc so make clean works Adds gzip including a sample kbuild file #2 - Factor out definition of usr/ to two variables #3 - Move kbuild files to reflect location in the kernel As requested in earlier mail I need a bit of guidiance of what you expect from the kernel integrated parts of klibc.

This fix has been implemented in preparation for the upcoming klibc merge in -mm. But as it fixes a real (but rare I hope) bug in vanilla kernel it will be added to my pending 2.6.17 kbuild queue. I've done quite some changes to the gen_initramfs script and being shell programming newbie review by more experienced shell hackers would be appreciated. I did a git mv

hello hpa, nothing particular stands out, just syncing with latest Debian upload and subsequent patch emails. please review merge or nack. thanks :) maks git pull git://git.debian.org/~maks/klibc.git maks for the changes: Aaron Griffin (1): [klibc] kinit: skip md assembly if mdX exists Colin Watson (1): [klibc] mount/umount FUSE support Harald Jenny (1): [klibc] fstype:

hpa notified me in private mail that we have an issue that when building initramfs - if the content changes (say kinit) then the initramfs will not be rebuild. The solution I have come up with is to let make check all the files included in the initramfs and rebuild it if any files are newer than the initramfs and also rebuild if arguments changes (root_uid). In same process the usr/Kbuild file

Last patch for now. This gets gen_init_cpio into a state where it can pick up a "kinit" binary from usr and stuff it into the cpio archive as /sbin/init. This is obviously a crock, but it suffices for testing until something sane is done about integrating klibc into the kernel tree. The patch also adds a bit of instrumentation to initramfs.c, to try to help flush out bug 740.

Just curious if anyone has done anything similar, as I'd rather not reporduce work. I'm going to put alias handling (and only alias handling) into a modified version of insmod. The reason I'm doing this is to aid in boot time module loading via aliases exposed in sysfs. Sure, I could include modprobe directly in the initramfs, but it's a bit large, and I really only need alias

At present the configuration items for initramfs is located in: Device drivers | Block Drivers | xxx This is maybe not the most natural place to have it. So with the following patch it is moved below "General setup", and relevant config items are collected in a file with a new home in usr/. The original reason why I looked into this is the upcoming merge of klibc and I missed a good

Generate dependencies almost like done by fixdep in the kernel. This teaches make to rebuild cpio image when content has changed. Also restructured usr/Kbuild a little so we better utilise parallel makes (dash, gzip, utils and kinit are now build in parallel). Signed-off-by: Sam Ravnborg <sam@ravnborg.org> --- patch is on top of hpa's linux tree, so path differ if applied to the klibc

Harcoding default content of initramfs in the gen_initramfs_list.sh script seems ackward. Move it to a file named: usr/initramfs.default This also fixes a small bug when no arguments was passed to CONFIG_INITRAMFS_SOURCE (-d was passed twice to the script) Idea and bug report from: "H. Peter Anvin" <hpa@zytor.com> Signed-off-by: Sam Ravnborg <sam@ravnborg.org> ---

I'm not sure when exactly this started, but currently when I specify a ramfs image via the initrd line in grub, kinit picks it up as an initrd and tries (and fails) to unpack it and run linuxrc from it. I'm assuming this is not the intended behavior. Can anyone point me into the proper direction for unpacking an initramfs image ontop of the in-kernel image? Currently, I am using kernel

Fix usr/Kbuild so it utilise the new gen_initramfs.sh script which will rebuild initramfs when the content has changed. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> --- usr/Kbuild | 74 +++++++++++++++++++---------------------------------------------- 1 file changed, 22 insertions(+), 52 deletions(-) This patch shall not be applied until my kbuild.git tree is merged upstream. Or it

The patch makes gen_init_cpio generate the initramfs_data.cpio from a file which contains a list of entries: file, dir, nod. I swapped the order of filename/location for the file arguments so that it would be more uniform with the dir and nod tyes. [thayne@torch linux-2.6.8]$ usr/gen_init_cpio --help ERROR: unable to open '--help': No such file or directory Usage:

Harcoding default content of initramfs in the gen_initramfs_list.sh script seems ackward. Move it to a file named: usr/initramfs.default This also fixes a small bug when no arguments was passed to CONFIG_INITRAMFS_SOURCE (-d was passed twice to the script) Idea and bug report from: "H. Peter Anvin" <hpa at zytor.com> Signed-off-by: Sam Ravnborg <sam at ravnborg.org>

Hi Bryan and other klibc people. I have taken a stamp on the Makefiles for klibc, this is what I came up with. 1) No longer recompile on every invocation 2) Correct checking on dependencies 3) Simpler makefile syntax (almost all over the place) I compile-time tested it only. Two open issues: a) Do we realy use .a files for initramfs. I renemed that to the executable. b) I renamed

gen_initramfs_list.sh failed to generate a dependency list when it was supplied the -d option. Restructured gen_initramfs_list.sh to support dependency info with -d. This also fixes a bug where we did not properly process when more than one dir was specified. Finally update dependency info for the -d case (list_default_initramfs) In usr/Kbuild remove "" around CONFIG_RAMFS_SOURCE so the