summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* images: Split dracut options out into a config file.Colin Guthrie2013-11-023-2/+7
| | | | | Also ensure that the compression method is xz -9 for much better compression (was previously defaulting to gzip)
* images/kernel: Ensure the previously hand-picked modules are included in the ↵Colin Guthrie2013-11-023-25/+11
| | | | | | | | | | | | | | new dracut-based initrd. This removes functions no longer called and adds a new one which will simply list the desired modules and pass them on the dracut command line. This should mean we have the same modules as before (at least) which should mean less regressions. This does increase the size of the initrd back up quite a lot (as expected) so we may want to look at optimising other areas to bring it back down a bit again.
* images: Use a generic, shared isolinux.cfg file.Colin Guthrie2013-10-302-2/+12
| | | | | | This then loads a x86_64 or i586 config depending on the detected arch. Anne asked for this kind of setup.
* images: Add a short circuit rule to make boot.iso for testingColin Guthrie2013-10-301-0/+7
|
* images: Clear out some stuff that no longer used.Colin Guthrie2013-10-303-103/+4
| | | | | Now that initrds are generated via dracut, there is no need to prepare kernel modules. This sadly removes support for passing in a kernel rpm file, but such is life.
* images: Fix isolinux.bin modules path and add needed modulesColin Guthrie2013-10-291-4/+16
|
* images: Fix kernel+ramdisk folder creationColin Guthrie2013-10-291-1/+1
|
* 1.94Anne Nicolas2013-10-292-1/+3
|
* 1.83Anne Nicolas2013-10-292-1/+2
|
* 15.6915.69ennael2013-10-292-1/+3
|
* stage1: Add debugging for net device name confusion.Colin Guthrie2013-10-291-1/+2
| | | | | | | | | Because dracut uses udev and because udev has likely already started the interface, we will see things out of order. Large sections of the probing code could probably be removed but for the time being we should probably consider removing this error message completely
* stage1: Switch to a new mode of knowing when to re-exec init.Colin Guthrie2013-10-292-31/+30
| | | | | | | | | | | | Now that dracut runs stage1 earlier, we need to store it's exit code for later reuse. In this, the separate dracut module will just write a flag file in /run/drakx/ to indicate that we want to re-exec /sbin/init (after doing some symlinks and /usr bind mount) rather than run the install. As a result there is no need to return 0x35 ('procced') so just return 0 and keep 0x35 return code for when we really do need to restart stage1 (e.g. on a fatal error). The dracut module should loop if 0x35 is ever returned.
* stage1: Remove the counter from init.Colin Guthrie2013-10-291-8/+2
| | | | | init is now solely responsibe for running stage2 (be it install or rescue) as stage1 is now run earlier by dracut.
* stage1: Don't bother mounting various API filesystems.Colin Guthrie2013-10-291-18/+0
| | | | Dracut mounts these already.
* install/rescue: Adapt to a dracut based stage1.Colin Guthrie2013-10-296-22/+14
| | | | | | | | | | | The main change here is not remounting a fresh /run. Dracut does this for us and we actually need to keep this one as the initial udev database (when udev is started by dracut) is stored here and thus mounting an empty fs could see us lose some metadata. We no longer touch the /bin, /sbin or /lib[64] symlinks as these are already relative symlinks into /usr and simply bind mounting /usr is enough.
* stage1: Move the symlink creation from stage1 to init.Colin Guthrie2013-10-297-108/+87
| | | | | | | | | | | | | | | | | | | | In order to run stage1 we now need to do things a little differently, namely that we move the symlinking fixups into 'init' itself which we will actually call as a kind of fake switch_root implementation to avoid hacking dracut too much[1]. As we run stage1 during dracut's 'mount' phase we should not do the symlink hacks so early otherwise it will mess up the rest of the dracut initqueue processing. Currently, the initrd's /usr is nuked and then symlinked, but to allow for easier debugging I now just bind mount the stage2 /usr over the top which allows us to unmount it again easily enough if we want to go back. We no longer touch the /bin, /sbin or /lib[64] symlinks as these are relative and simply bind mounting /usr is enough to fix all of them. 1. This scheme will have to change if we eventually swtich to a systemd based dracut invocation.
* stage1: Correct path to shellColin Guthrie2013-10-291-1/+1
|
* images: Convert image generation to dracut.Colin Guthrie2013-10-293-19/+44
| | | | | | | | | | | | | | | | | | | | | | | This converts the initrd generation from a home grown system to dracut. It uses the upstream modules where possible plus two specially written dracut modules: mgainstaller and mgakadeploy. The initrds are identical for all builds and there are lots of bits that are incomplete and unnecesary. e.g. copying the kernels to a folder and then not using them, and also the module-descriptions file is not included in the initrd itself which it should be. We can probably customise the initrd better for different use cases, e.g. perhaps network stuff is not needed for HD based installs? We may also need to install the mdraid and lvm modules for some HD support. This is the first attempt so I expect various bits need fixing
* stage1: Switch to a shared stage1.Colin Guthrie2013-10-293-14/+17
| | | | | | | | | | | Previously we would create a static stage1 for the installer, but as a dracut based initrd already includes glibc, there is very little advantage in this small, statically linked version. As a side note, it seems that if the initrd contains a /etc/modprobe.d/ folder (which can be the case with dracut) then dietlibc seems to return a ENOENT when calling readdir_r() on that folder (via the libkmod-config.c code).
* stage1: Do not hard code STAGE2 pathColin Guthrie2013-10-291-1/+1
|
* stage1: Add .gitignoreColin Guthrie2013-10-291-0/+13
|
* do not use arch dependant isolinux.bin anymoreennael2013-10-251-1/+1
|
* Merge branch 'master' of git://git.mageia.org/software/drakxennael2013-10-253-461/+471
|\
| * small fix in Slovenian translationfilip2013-10-241-4/+3
| |
| * small fix in Slovenian translationfilip2013-10-241-2/+3
| |
| * Update Russian translationYuri Chornoivan2013-10-241-420/+440
| |
| * fixed typo in Slovenian translationfilip2013-10-241-2/+2
| |
| * Slovenian translation updatedfilip2013-10-241-39/+29
| |
* | move isolinux in the root treeennael2013-10-221-5/+5
| |
* | copy images in $arch directory instead of alt0 for new syslinuxennael2013-10-221-2/+2
|/
* More fixes to Ukrainian translationYuri Chornoivan2013-10-161-3/+3
|
* Fix Ukrainian translationYuri Chornoivan2013-10-161-3/+3
|
* Do not add localhost to /etc/hosts as this is handled by nss-myhostnameOlav Vitters2013-10-151-3/+0
| | | | | The myhostname NSS module (which is default in nsswitch.conf) handles this happily.
* Remove "Clean /tmp at each boot" optionOlav Vitters2013-10-151-11/+3
| | | | | Not needed since systemd automatically already sets up /tmp using tmpfs. Proposed and accepted on dev mailing list.
* Update Ukrainian translationYuri Chornoivan2013-10-141-26/+16
|
* 1.45.1Thierry Vignaud2013-10-142-1/+3
|
* many btrfs tools no longer existThierry Vignaud2013-10-142-2/+4
|
* 1.45Thierry Vignaud2013-10-142-1/+3
|
* adapt to latest mdadmThierry Vignaud2013-10-142-2/+3
|
* 15.68.115.68.1Thierry Vignaud2013-10-142-1/+3
|
* adapt to latest mdadmThierry Vignaud2013-10-142-2/+4
|
* 15.6815.68Thierry Vignaud2013-10-143-1/+5
|
* list_modules.pm changes affect drakxtools & stage2 tooThierry Vignaud2013-10-142-0/+2
|
* 1.82Thierry Vignaud2013-10-142-1/+2
|
* 1.93Thierry Vignaud2013-10-142-1/+3
|
* sync with kernel-3.12Thierry Vignaud2013-10-143-6/+10
|
* Drop /etc/termcap as this is no longer included in our buildColin Guthrie2013-10-135-6/+2
|
* any: Fix path to urpmi.update (mga#11125)Colin Guthrie2013-10-132-1/+2
|
* finish-install: Do not check for /usr/sbin/urpmi.updateColin Guthrie2013-10-132-1/+3
| | | | | | | | | This file is now 'hidden' in /usr/libexec and only used via it's wrapper script in mgaonline. This was originally changed from /usr/bin/urpmi back in 2001 by Pixel, but the current path for urpmi is /usr/sbin/urpmi, which seems to be as reliable as urpmi.update was before.
* polkit: Use exec in wrapper scriptsColin Guthrie2013-10-121-1/+1
|