summaryrefslogtreecommitdiffstats
path: root/scripts/mk-images
Commit message (Collapse)AuthorAgeFilesLines
...
* Let's try pulling libsqlite into the initrd one more time.Chris Lumens2008-09-101-1/+1
|
* Include NetworkManager and dbus libraries on 64-bit arches (#461632).Chris Lumens2008-09-091-8/+8
|
* We need libsqlite3.so in upd-instroot before it can be in the initrd.Chris Lumens2008-09-091-1/+1
|
* Add more kernel crypto modules (#443545).Chris Lumens2008-09-081-1/+1
|
* Add libsqlite to the initrd, which is needed by NSS libs.Chris Lumens2008-09-051-0/+1
|
* Add more dlopen()ed libraries to the initrd.Chris Lumens2008-09-051-0/+2
|
* Make sure /bin/sh is linked to /bin/bashDavid Cantrell2008-09-041-0/+1
| | | | | Using bash on the initrd.img now, but make sure we keep the /bin/sh link.
* Do not include /usr/lib/gconv in install.imgDavid Cantrell2008-09-041-4/+0
| | | | | | New NetworkManager appeared in rawhide that does not perform UTF-8 conversions on DHCP options, so we don't need to include /usr/lib/gconv in the install image.
* Support accessing preexisting LUKS devs using LRW or XTS ciphers. (#455063)David Lehman2008-08-261-1/+1
|
* Use NetworkManager instead of libdhcp. (#458183)David Cantrell2008-08-251-7/+123
| | | | | | | | | | | Finally, no more libdhcp. This is the first set of changes to take anaconda over to the wonderful world of NetworkManager. We are no longer linking with libdhcp to do interface configuration. NM is started early in the installation and opens the door to things like WPA installation support and things like that.
* Rename stage2.img to install.imgDavid Cantrell2008-08-131-3/+8
| | | | | | To make the installation image files more clear and since we have removed minstg2.img, rename stage2.img to install.img.
* More fixes to include udev rules in the initrd (#458570).Chris Lumens2008-08-121-0/+4
|
* udev rules have changed location (#458570).Chris Lumens2008-08-111-5/+5
|
* mk-images: replace kernel-xen with pv_ops kernelMark McLoughlin2008-08-071-5/+4
| | | | | | | | | | | | | | | | | | | | kernel-xen is dead, replaced now with pv_ops Xen guest support in the stock kernels. On x86_64 and ia64, we now just point the [images-xen] .treeinfo stanza at the stock kernels in images/pxeboot. On x86_32, things are a bit more complicated. The stock kernel is kernel.i586 but Xen guests support is only included in kernel-PAE.i686 because Xen only runs in PAE mode. In this case, we just copy the kernel-PAE vmlinuz to images/pxeboot/vmlinuz-PAE and build another initrd as images/pxeboot/initrd-PAE.img. It's these images we point the [images-xen] stanza at. Signed-off-by: Mark McLoughlin <markmc@redhat.com>
* More iSeries removal from scripts (#456878)David Cantrell2008-08-051-5/+1
|
* Leave anaconda-runtime around for mk-images run.David Cantrell2008-07-291-5/+8
| | | | | | | | Do not remove the anaconda-runtime or syslinux directories before calling doPostImages. We need files from there for the mkisofs step. In makemainimage, make sure we exclude the anaconda-runtime and syslinux directories when making the second stage image there.
* Clean up some mistakes in the minstg2 removal.David Cantrell2008-07-251-0/+1
| | | | | Fix up some errors in mk-images and upd-instroot that are preventing images from building correctly.
* Remove support for generating a minstg2.img image.David Cantrell2008-07-241-20/+17
| | | | | | The minstg2 image has outlived its usefulness. We are at a point where we have been maintaining two stage 2 images and they are slowly becoming identical.
* Support booting from FCP-attached CD/DVD drive on s390 (#184648)David Cantrell2008-07-171-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The s390 can boot El Torito CD or DVD images iff they are attached by zFCP, used as the IPL device, and contain a specially formatted boot image on the disc. IBM provided the tool to combobulate the boot image together and a description of the desired execution path. When you boot on s390, the linuxrc.s390 will look to see if you IPL'ed from a CD or DVD. If you did, it will ask if you also want to install from that device. If you answer yes (y, Y, or any case spelling of 'yes'), the script will bring the IPL device online so the kernel assigns it a device name. Then it skips over the network configuration and starts you in to loader. If you tell it no or did not IPL from a CD or DVD, it'll launch the missiles--wait, no, I mean you get the normal network installation questions before loader starts. I have no way to test this as it requires the following changes: (1) Rel-eng needs to build s390x media with -no-emul-boot and specify the new cdboot.img file on that platform. I have already contacted rel-eng about making this change. (2) I don't have a CD-ROM drive in my mainframe. IBM does and testing is all falling on them. IBM knows this...maybe. I explain all of this like anyone else on the team will ever get a chance to experience it. So there you have it. A letter opener.
* It's called crypto_blkcipher.ko these days.Chris Lumens2008-07-161-2/+2
|
* Deal with udev losing udevcontrol/udevtriggerJeremy Katz2008-06-171-2/+0
| | | | | udev is stopping supporting udevcontrol/udevtrigger as symlinks, so we should change what we call
* Link ld-linux.so.2 to ld-*.*.*.soDavid Cantrell2008-06-091-2/+16
| | | | | Avoid having two copies of the dynamic linker in the initrd.
* Stop neutering DRIBill Nottingham2008-05-121-1/+2
| | | | | | | | | | | | | | | | | | | | | The attached patch makes anaconda stop neutering DRI/DRM (at least it should.) It does it by including the DRI userspace modules (from mesa) and the kernel DRM drivers. Why do this? To be more like the 'normal' installed system. Heck, if DRI's going to fail, the system's just going to blow up on the first boot anyway. Furthermore, there are cards/chips that do 2D accel via the 3D pipeline. Plus, if we want to actually disable it in anaconda, we should do it explicitly, not by accident as a consequence of some files not being there. Caveats: - listing the kernel drivers explicitly is a hack - this doesn't pull libGL onto the second stage. Assuming nothing dlopens() it, that shouldn't be a problem
* Revert PS1 and PATH changes as they don't work with busybox as used in minstg2Jeremy Katz2008-04-291-3/+1
|
* Add missing backslashes for the .profile here document.David Cantrell2008-04-251-1/+1
|
* Use the correct ls(1) alias.David Cantrell2008-04-231-1/+1
| | | | The default ls alias does not use -F, so removed that.
* Set PS1 and ls(1) alias for tty2 shell.David Cantrell2008-04-231-1/+3
| | | | | Make the prompt and ls command more like the final system. Removed \u@\h from the prompt, but it's still more useful.
* mk-images: Module balls are no moreMark McLoughlin2008-04-101-5/+0
| | | | | | | | | | | | | | notting removed module balls some time ago: commit 7971fbec83064b096a6592055d66df573bdb8eea Author: Bill Nottingham <notting@redhat.com> Date: Wed Dec 19 14:53:45 2007 -0500 Don't ship module balls; just ship a 'normal' /lib/modules and /lib/firmware tree. but mk-images still has some remnants of them. Signed-off-by: Mark McLoughlin <markmc@redhat.com>
* ide-cd_mod, not ide-cd_rom (thanks to jwb)Jeremy Katz2008-04-091-1/+1
|
* More ia64 kernel finding fixage (#441708)Jeremy Katz2008-04-091-1/+5
|
* Merge branch 'master' of git+ssh://git.fedoraproject.org/git/hosted/anacondaPeter Jones2008-04-081-4/+9
|\
| * Handle ia64 kernel pathJeremy Katz2008-04-081-2/+7
| | | | | | | | | | The ia64 kernel is under /boot/efi not /boot. Fix up our finding of the vmlinuz (patch from Yi Zhan <yi.zhan@intel.com> #441470)
| * fix compression of modulesBill Nottingham2008-04-071-2/+2
| |
* | Build efiboot.img on x86_64 and i386 .Peter Jones2008-04-081-7/+2
| |
* | Add code to create efiboot.img on i386 and x86_64Peter Jones2008-04-071-0/+13
|/
* Make a little bit more future proof for kernel version changingJeremy Katz2008-03-261-2/+2
| | | | | uname -r as reported by the kernel should be figured out based on some files that get written, not by depending on "knowing" that it's version-release
* Put "ide-cd_mod" in the list of modules to pull in.Peter Jones2008-03-191-1/+1
|
* Use i586 kernel (#437641)Jeremy Katz2008-03-161-1/+1
|
* Convinced to change the name back to boot.isoJeremy Katz2008-03-101-1/+1
| | | | | | At least for in the tree, having the ISO as boot.iso helps tools such as virt-manager, which is a convincing reason to leave it. When pungi copies into the iso/ directory, it's probably reasonable to change the name
* Pass the --archlist option to yumdownloaderJesse Keating2008-03-061-1/+1
| | | | | | | This is necessary on ppc where we are trying to download the ppc64 kernel from a ppc environment. Requires an updated yum-utils that supports the --archlist option to yumdownloader, but I don't know the n-v-r of that build yet, so no change to anaconda.spec yet.
* Add --archlist to repoquery call.Jesse Keating2008-03-041-1/+1
| | | | | | This is necessary for composing ppc, where the compose is done on ppc, but we want to find a ppc64 kernel. By default, repoquery won't return results that don't match the running arch.
* Test for using efi rather than arch==ia64Peter Jones2008-03-041-1/+1
|
* If we don't find a kernel package, then give a better errorJeremy Katz2008-03-031-0/+5
|
* Add virtio_pci in hopes of getting virtio workingJeremy Katz2008-03-021-1/+1
|
* Change the ISOs we buildJeremy Katz2008-02-281-0/+5
| | | | | | Stop building boot.iso (and diskboot.img) in preference for building a netinst.iso with the boot bits as well as the second stage. This will replace the rescue image and is overall far more useful
* Take advantage of yum repos being availableJeremy Katz2008-02-281-24/+16
| | | | | | Instead of doing tricks with find, use repoquery and yumdownloader to download things like anaconda-runtime and kernels and firmware packages. This also has the nice side effect of letting us run buildinstall against an http repo
* Move doPostImages to be run after the second stage buildJeremy Katz2008-02-271-2/+2
|
* More crud to deal with the fact that rawhide trees are composed weirdJeremy Katz2008-02-261-2/+2
|
* Clean up invocation of mk-images from buildinstallJeremy Katz2008-02-171-31/+63
| | | | | | | Use named options when passing things to mk-images so that it's more clear as to what's going on. Also, make things like arch more explict. This also adds support for output from mk-images to go elsewhere, although that's currently not used by anything
* Fix up firmware inclusion. This didn't actually ever work.Jeremy Katz2008-02-141-35/+36
|
generated by cgit (git 2.34.1) at 2025-08-31 16:45:25 +0000