| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2016-04-29 | fix paths (in progress) | Andrew Cady | |
| 2016-04-29 | don't hard-code btrfs default subvolume name | Andrew Cady | |
| 2016-04-28 | copy initrd logs into rootfs | Andrew Cady | |
| 2016-04-28 | fixed: /outerfs not mounted on first boot | Andrew Cady | |
| 2016-04-28 | ramdisk is now allocated "all" of system memory | Andrew Cady | |
| 2016-04-28 | boot-ram was using erroneous math to calculate ramdisk size | Andrew Cady | |
| 2016-04-28 | share more code between boot-new and boot-ram -- this fixes a bug | Andrew Cady | |
| 2016-04-28 | samizdat-eject | Andrew Cady | |
| 2016-04-28 | more samizdat-eject fixes | Andrew Cady | |
| also, copy samizdat-eject from the initramfs to the new rootfs (into /sbin) | |||
| 2016-04-27 | Implement encrypted cdrom ejection | Andrew Cady | |
| This allows the cdrom to be copied onto the outer filesystem (mounted in /outerfs) without storing unencrypted gpg keys there. e.g.: samizdat-eject.sh /outerfs/samizdat.iso This was necessary because the other method probably causes btrfs deadlocks. We do end up copying data twice this way (or three times, probably -- if the ISO is saved), but not _from the cdrom_. And we get to eject immediately after the first copy. Future copies will be from the hard drive. Not too bad. | |||
| 2016-04-27 | put /root in separate btrfs snapshot | Andrew Cady | |
| 2016-04-27 | on failed install, remove /dev/mapper/samizdatcrypt | Andrew Cady | |
| 2016-04-27 | fix: only remove root's password on new install | Andrew Cady | |
| 2016-04-27 | get rid of unused isolinux files; change iso cdrom paths | Andrew Cady | |
| 2016-04-26 | fixup! Cdrom duplication | Andrew Cady | |
| 2016-04-26 | ejection script for cdrom duplication | Andrew Cady | |
| included some necessary dependencies | |||
| 2016-04-25 | fixup! Cdrom duplication | Andrew Cady | |
| 2016-04-25 | fixup! Cdrom duplication | Andrew Cady | |
| 2016-04-25 | Cdrom duplication | Andrew Cady | |
| The cdrom can now be copied into the btrfs filesystem (if there is enough space) and hot-removed. This isn't done automatically, though. In fact, the code to do it isn't available anymore after boot; it would have to be copied back onto the system to be run. | |||
| 2016-04-22 | mount seed layers when booting with RAM option | Andrew Cady | |
| 2016-04-22 | workaround for btrfs limitation (read-only seed fix) | Andrew Cady | |
| 2016-04-22 | Add multi-layered seed support | Andrew Cady | |
| The available filesystem with the largest number of seeds will be chosen for the root of the initial install. | |||
| 2016-04-22 | remove some cruft | Andrew Cady | |
| 2016-04-22 | move gnupghome onto installed fs root | Andrew Cady | |
| 2016-04-20 | btrfs-based encrypted persistent root | Andrew Cady | |
| This replaces the old LVM code with btrfs-based code. | |||
| 2016-04-19 | implement menu select option for booting with ramdisk root | Andrew Cady | |
| 2016-04-19 | make sure real openvt (not busybox) is available and used | Andrew Cady | |
| this fixes recovery after menu selection (menu is relaunched) | |||
| 2016-04-19 | offer to install even when disk space is below cdrom size | Andrew Cady | |
| 2016-04-19 | make the samizdat boot menu show up | Andrew Cady | |
| 2016-04-19 | cause the grok-block hooks to run | Andrew Cady | |
 |
index : sami.git | |
| summaryrefslogtreecommitdiff |
path: root/old-school