summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Victor Lowther [Fri, 13 Feb 2009 12:41:40 +0000 (04:41 -0800)]
[PATCH 04/50] Modify root cmdline parsing to use a case statement.
This code takes less space and is easier to read at a glance.
Victor Lowther [Fri, 13 Feb 2009 12:41:38 +0000 (04:41 -0800)]
[PATCH 03/50] Use getarg to get the root kernel commandline parameter.
Victor Lowther [Fri, 13 Feb 2009 12:41:32 +0000 (04:41 -0800)]
[PATCH 02/50] Add a simple getarg function.
If we ever need to get more than one kernel command line argument, this
function will pay for itself. It relies on a feature of the way bash
handles string comparisons in [[ ]] statements -- the RHS is matched
according to globbing rules, and is not a straight string match.
Victor Lowther [Fri, 13 Feb 2009 12:41:29 +0000 (04:41 -0800)]
[PATCH 01/50] We have a path. Do not specify full paths to commands that are in it.
Jeremy Katz [Mon, 19 Jan 2009 21:00:24 +0000 (16:00 -0500)]
Slightly more reasonable timeout
Jeremy Katz [Mon, 19 Jan 2009 20:57:34 +0000 (15:57 -0500)]
Also rescan after each new lvm2 member appears so we can find all devs
Jeremy Katz [Mon, 19 Jan 2009 20:57:18 +0000 (15:57 -0500)]
Don't use deprecated git-* commands
Jeremy Katz [Mon, 19 Jan 2009 18:57:09 +0000 (13:57 -0500)]
Import volume info here too (noticed by wtogami with root-on-crypt-lv)
Bill Nottingham [Sat, 10 Jan 2009 19:59:13 +0000 (14:59 -0500)]
Go to an emergency shell if mounting the root filesystem fails.
Bill Nottingham [Sat, 10 Jan 2009 19:59:11 +0000 (14:59 -0500)]
Fix 'root=LABEL=/'
As you can't have /dev/disk/by-label//, we need to escape the '/'
character in the same way that udev does.
Bill Nottingham [Sat, 10 Jan 2009 19:59:12 +0000 (14:59 -0500)]
Fix tense.
Jeremy Katz [Mon, 5 Jan 2009 20:50:00 +0000 (15:50 -0500)]
Now that module dependencies work, we don't have to list as much
Jeremy Katz [Mon, 5 Jan 2009 20:45:39 +0000 (15:45 -0500)]
Fix module dependency handling
Module dependency handling wasn't being followed correctly; fix it
Jeremy Katz [Mon, 5 Jan 2009 20:42:56 +0000 (15:42 -0500)]
Clean up noise from eu-readelf
Jeremy Katz [Mon, 5 Jan 2009 19:40:34 +0000 (14:40 -0500)]
Add first pass of keyboard/console font setup
First pass of keyboard/console font setup. This ends up being Fedora/RHEL
specific as everyone specifies these things in a different place
Jeremy Katz [Mon, 5 Jan 2009 19:39:35 +0000 (14:39 -0500)]
Fix typo with local mode installing switch_root
Jeremy Katz [Mon, 5 Jan 2009 18:24:46 +0000 (13:24 -0500)]
Update the documentation a bit for current status
Jeremy Katz [Mon, 5 Jan 2009 18:16:39 +0000 (13:16 -0500)]
Basic support for loading SELinux from the initramfs
Jeremy Katz [Mon, 5 Jan 2009 17:46:39 +0000 (12:46 -0500)]
Stop depending on nash to find the program interpreter; use eu-readelf instead
Jeremy Katz [Fri, 19 Dec 2008 19:44:38 +0000 (14:44 -0500)]
Add a spec file for rpm building
Jeremy Katz [Fri, 19 Dec 2008 19:43:46 +0000 (14:43 -0500)]
Handle rules.d being from the installed location
Jeremy Katz [Fri, 19 Dec 2008 19:43:30 +0000 (14:43 -0500)]
Fedora calls --allow-missing as an arg to mkinitrd; noop it
Jeremy Katz [Fri, 19 Dec 2008 19:43:06 +0000 (14:43 -0500)]
plymouth calls set_verbose(), no noop it
Jeremy Katz [Thu, 18 Dec 2008 20:59:55 +0000 (15:59 -0500)]
Fix shebang at the top of the file
Jeremy Katz [Thu, 18 Dec 2008 20:58:15 +0000 (15:58 -0500)]
Add a makefile
Jeremy Katz [Thu, 18 Dec 2008 20:57:47 +0000 (15:57 -0500)]
Make using the version from the current directory explicit
Rather than looking at the contents of the current directory always, require
explicitly specifying --local to use those files
Jeremy Katz [Thu, 18 Dec 2008 20:55:28 +0000 (15:55 -0500)]
Handle locations where we'll be installing files
Handle that files will be in different places once installed
Jeremy Katz [Wed, 17 Dec 2008 22:51:15 +0000 (17:51 -0500)]
Make the generator start to be functional on a "real" system
Start to pull in modules from the system rather than the crude hack of
everything in a tree and explicitly list some classes of modules
including what's needed for dm-crypt
With this, I am now running a dracut initramfs on my laptop
Jeremy Katz [Wed, 17 Dec 2008 21:57:43 +0000 (16:57 -0500)]
Add some basic functions used for the initramfs
Add some functions used by the initramfs generator from Fedora's mkinitrd with
bits and pieces stripped out. This needs a closer look to ensure that we're
not bringing along any other old baggage
Jeremy Katz [Wed, 17 Dec 2008 21:56:51 +0000 (16:56 -0500)]
rstrode built a new plymouth with the changes, so we can drop that make install
Jeremy Katz [Wed, 17 Dec 2008 21:56:28 +0000 (16:56 -0500)]
Move generator to be called 'dracut'
Jeremy Katz [Wed, 17 Dec 2008 19:55:20 +0000 (14:55 -0500)]
let mount auto-determine the fs type
Jeremy Katz [Wed, 17 Dec 2008 18:16:12 +0000 (13:16 -0500)]
Now called dracut; add info about the mailing list
Jeremy Katz [Wed, 17 Dec 2008 18:15:55 +0000 (13:15 -0500)]
Add pointer to Peter's switchroot
Jeremy Katz [Tue, 16 Dec 2008 22:24:59 +0000 (17:24 -0500)]
Update the README and add some docs for TODO and HACKING
Jeremy Katz [Tue, 16 Dec 2008 21:56:28 +0000 (16:56 -0500)]
Sleep forever instead of exiting so you can scroll back
Jeremy Katz [Tue, 16 Dec 2008 19:29:43 +0000 (14:29 -0500)]
Don't run udevsettle; instead, we keep looking for the rootfs
Jeremy Katz [Tue, 16 Dec 2008 19:29:07 +0000 (14:29 -0500)]
Cut down to only creating the nodes that plymouth actually needs
Jeremy Katz [Tue, 16 Dec 2008 19:25:06 +0000 (14:25 -0500)]
Bump timeout to five minutes since people could be using encrypted root
Jeremy Katz [Tue, 16 Dec 2008 19:21:08 +0000 (14:21 -0500)]
Support root on LUKS (dm-crypt)
Support having root on LUKS with the password prompting handled by plymouth.
This requires ensuring our input is from /dev/console and also requires that
we import vol_id info about all block devices rather than ignoring dm devs
(which is what the persistent storage rules do by default)
Jeremy Katz [Tue, 16 Dec 2008 19:19:52 +0000 (14:19 -0500)]
Add some modules
dm-crypt support requires that we have some modules, so add some basic
support for copying over a module tree for testing
Jeremy Katz [Mon, 15 Dec 2008 21:10:51 +0000 (16:10 -0500)]
Add plymouth support
Add basic support for the use of plymouth in the new initrd world. Also
includes a hack at the moment to install from a local plymouth tree to make
testing easier
Jeremy Katz [Mon, 15 Dec 2008 19:40:36 +0000 (14:40 -0500)]
Some debugging helpers
Install terminfo files so that less is useful. Also, timeout mounting
the rootfs so that we can debug better
Jeremy Katz [Mon, 15 Dec 2008 19:07:55 +0000 (14:07 -0500)]
Add the basics for LVM support
For LVM, we need to do activation of the volume group within the initrd.
Stick lvm tools into the initrd as well as a (basic and stupid) udev rule
to run vgchange -ay. Ultimately we want to get a better rule for this into
udev proper
Jeremy Katz [Mon, 15 Dec 2008 18:33:54 +0000 (13:33 -0500)]
Various fixes to make things work; also listen to root=
Apparently what I had committed was broken. This fixes some pieces up
and also adds (basic) support for using root=
Jeremy Katz [Wed, 10 Dec 2008 22:17:09 +0000 (17:17 -0500)]
Kill udevd
Jeremy Katz [Wed, 10 Dec 2008 21:36:31 +0000 (16:36 -0500)]
And licensing...
Jeremy Katz [Wed, 10 Dec 2008 21:35:29 +0000 (16:35 -0500)]
Having a README is always nice
Jeremy Katz [Wed, 10 Dec 2008 21:32:21 +0000 (16:32 -0500)]
Initial commit
First pass at a new initrd implementation. Right now, it's a very simple
/init and we use nash's switchroot command via switch_root. The idea is to
get a switchroot binary into util-linux that can be used instead
generate.sh will let you generate an initrd image. Note that the generator
is intentionally super simple right now and is an area that will need a lot
of work once we're happier with how the /init process runs