Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

This was the first implementation that downloads the complete ISO,
but now we have the posibility of download only needed *.sfs files.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

This was useful to me during development of dm-snapshot support
to create very quickly isos without wating for SquashFS compression.

Is time to remove this, I think the is no practical usage,
and make the code a bit more simple.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Simple way to enable self-PXE-server live-enviroment with HTTP.

dnsmasq --port=0 \
        --enable-tftp \
        --tftp-root=/run/archiso/bootmnt \
        --dhcp-range=192.168.0.2,192.168.0.254,86400 \
        --dhcp-boot=/arch/boot/syslinux/gpxelinux.0 \
        --dhcp-option-force=209,boot/syslinux/archiso.cfg \
        --dhcp-option-force=210,/arch/

darkhttpd /run/archiso/bootmnt

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

We are only using these files only for SYSLINUX bootloader that supports
".." as directories.

ISOLINUX jumps to archiso.cfg via isolinux.cfg
PXELINUX goes directly to archiso.cfg via DHCP configuration.

Suggested by Thomas.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* In that way we can have two differents syslinux menus depending if booting from
PXE or from non-PXE.
Some file renames are made to keep things ordered:

archiso_pxe* all things related to PXELINUX
archiso_sys* all things related to SYSLINUX
*_inc.cfg only contains INCLUDE directive

* By default NBD PXE option is selected for booting on timeout.
* As (NDB/NFS/HTTP/FTP) server is used the same IP from pxeserver.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* Now we have more than one PXE boot methods.
* archiso_ndb_srv= accepts special ${pxeserver} parameter like in others PXE hooks.
* This allows to setup network from initramfs without forcing using a special PXE hook.
* Update README

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* UI is more correct than DEFAULT for simple menu system.
* Remove TIMEOUT
* Remove PROMPT 0, this is the default value.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

On machines with common graphics chipsets (radeon, intel, nvidia), enable
KMS early. Downloading images from a remote server can take time, better
view the status with a nice resolution.

Otherwise does not work.

22100 Nov 29 09:29 ./lib/libnss_dns-2.14.1.so
   20 Nov 29 09:29 ./lib/libnss_dns.so.2 -> libnss_dns-2.14.1.so
46732 Nov 29 09:29 ./lib/libnss_files-2.14.1.so
   22 Nov 29 09:29 ./lib/libnss_files.so.2 -> libnss_files-2.14.1.so

Looks like my test was made before introducing symlinks
and using directly libnss_dns-$VERSION.so.2 as libnss_dns.so.2.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

We are currently checksuming all files inside $archisobasedir
(aitab, *sfs, kernel/initramfs and boot/syslinux related files)
this is stored in one file only. Its works good for now,
but I think we only need to checksum only aitab and *.sfs if they are needed.
Currently if you boot a dual-iso in i686, also checksum is done
for *.sfs of x86_64. Not a big issue, this just take a bit more time.
The real issue is when booting via PXE with HTTP/FTP methods,
since they download only aitab and needed *.sfs files,
can not use directly the only one file where checksums are stored.

This patch does:

(1) Do not checksum syslinux related files, anyway you are already reach initramfs stage.
(2) Do not checksum kernel/initramfs files, for the same as (1)
(3) Two checksum.${arch}.md5 for each i686 and x86_64 for only aitab and *.sfs.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* In that way, we have unified directory scheme among {pxe,sys,ext,iso}linux,
  sharing the same directory structure and config files.
* Also enable IPAPPEND (will show a warning message about unknown directive in non-pxelinux before menu)

NEW:
* Rename syslinux*.cfg -> archiso*.cfg
* Now syslinux.cfg is only a "loadconfig" trick like we are doing for isolinux
* All paths are now relative to %INSTALL_DIR% ($archisobasedir) (Default: /arch/)

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

[ramfs /]# cat /etc/resolv.conf
nameserver 192.168.0.77
[ramfs /]# curl -O http://releng.archlinux.org/pxeboot/README


  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  1129  100  1129    0     0   2191      0 --:--:-- --:--:-- --:--:--  4589
[ramfs /]#

Do not "add_binary /lib/libnss_dns.so.2", because will add libnss_dns-VERSION.so.2 as libnss_dns.so.2
Is not an issue will work (nobody links directly to VERSION), but is not nice.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

It add support for using *.sfs images via NFS.

It must be located after archiso_pxe_nbd in mkinitcpio.conf, in that way
NBD is the default.

New boot parameters:

    archiso_nfs_srv=IP:/path
         Set the NFS IP and path. Accepts ${pxeserver} as IP.

    archiso_nfs_opt= Sets NFS mount opts (comma separated)
                    Defaults to:
                      port            = as given by server portmap daemon
                      rsize           = 1024
                      wsize           = 1024
                      timeo           = 7
                      retrans         = 3
                      acregmin        = 3
                      acregmax        = 60
                      acdirmin        = 30
                      acdirmax        = 60
                      flags           = hard, nointr, noposix, cto, ac

:: Running Hook [archiso_pxe_common]
IP-Config: eth0 hardware address 52:54:00:12:34:57 mtu 1500
IP-Config: eth0 guessed broadcast address 192.168.0.255
IP-Config: eth0 guessed nameserver address 192.168.0.218
IP-Config: eth0 complete (from 192.168.0.218):
 address: 192.168.0.90     broadcast: 192.168.0.255    netmask: 255.255.255.0
 gateway: 192.168.0.218    dns0     : 192.168.0.218    dns1   : 0.0.0.0
 rootserver: 192.168.0.218 rootpath:
 filename  :
:: Running Hook [archiso_pxe_nbd]
:: Running Hook [archiso_pxe_curl]
:: Running Hook [archiso_pxe_nfs]
:: Mounting '192.168.0.77:/mnt/iso'

...
192.168.0.77:/mnt/iso/ on /run/archiso/bootmnt type nfs (rw,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,nolock,proto=tcp,port=65535,timeo=7,retrans=3,sec=sys,local_lock=all,addr=192.168.0.77)
...

My /etc/exports
/mnt/iso 192.168.0.90(ro,no_subtree_check,no_root_squash)

Its add support for downloading the ISO image via HTTP/FTP with curl or download *.sfs files.
The files are downloaded on /run/archiso/curlspace (tmpfs) and setup a loop dev from it
(like in archiso_loop_mnt) in the ISO case, or just do a bind mount to
/run/archiso/bootmnt in the other case.

This hook must be located after archiso_pxe_nbd in mkinitcpio.conf, in that way
NBD is the default.

New bootparm:
* archiso_curl_url=(first form)
    archiso_curl_url=http://192.168.0.7/archlinux/iso/archlinux.iso (absolute form)
        Use an absolute URL. Fetch the entire ISO
    archiso_curl_url=ftp://${pxeserver}/archlinux/iso/archlinux.iso (relative form)
        Use the as IP the same server where PXE reside.  Fetch the entire ISO

* archiso_curl_url=(second form)
    archiso_curl_url=http://192.168.0.7/archlinux/iso/unpacked

 (absolute form)
        Use an absolute URL. Where unpacked (an example name) is a directory where
        ${archisobasedir} ("arch" by default) is located with all *.sfs.
        This fetch only needed files.
    archiso_curl_url=ftp://${pxeserver}/archlinux/iso/unpacked (relative form)
        Use the as IP the same server where PXE reside. Where unpacked (an example name)
        is a directory where ${archisobasedir} ("arch" by default)
        is located with all *.sfs. This fetch only needed files.

* curlspace_size= Set a size for a tmpfs filesystem where files are downloaded.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

One hook only setup the network device (archiso_pxe_common) the other
setup the NBD client (archiso_pxe_nbd).

New bootparam:
    archiso_pxe_srv=IP
        Allow to set an IP different from the PXE server.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Avoid issues with -b BASEDIR (*).
We became independent of the mkinitcpio version of the build environment.

(*) For example of a recent issue:
http://mailman.archlinux.org/pipermail/arch-projects/2011-November/002202.html
http://mailman.archlinux.org/pipermail/arch-projects/2011-November/002205.html



Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

In this way allow to mount /run/archiso/bootmnt from another hook,
but still use the logic from the main mount_hook, skipping only this mount.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* Now bootmnt/img_dev is always a ro-mount,
  and cowspace is first ro-mount then rw-mount.

* Fix a "hidden" bug: at the time of test [[ -ef ]], if devices nodes
  are not ready, such test will fail then archisodevice will mounted ro and
  when cow_device is mounted to be rw it fails.
  (I recently suffered this on a machine with slow USB)

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Boot media is already available in /run/archiso/bootmnt
if no copytoram is used.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Is in {base} since 1.8-4.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* ipconfig cmd writes a file in /tmp that is ready for direct evaluation.
We can use this instead of parsing the output, to do this some variable
renames are needed.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* Remove IDE stuff.
* Remove modules that can be added via relevant hooks.
* Replace sata with virtio in baseline profile.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

STATUS: Working (std boot, loop_mnt, pxe) with copytoram=[y|n].

NEEDS: initscript > 2011.07.3, mkinitcpio > 0.7.2, mkinitcpio-busybox > 1.18.5-1

Purpose: we need this for propertly unmount $cow_device,
used for persistent dm-snapshot devices.

This hook is based on work from Tom Gundersen[#1],
but adapted for archiso things (specially the shutdown script)

[#1] http://mailman.archlinux.org/pipermail/arch-projects/2011-July/001549.html
[#2] http://projects.archlinux.org/initscripts.git/commit/?id=1fa7b4b453e96533ae1db3630031285e5fc302b3
[#3] http://mailman.archlinux.org/pipermail/arch-projects/2011-August/001749.html



Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

* Support img_label=
* Allow use img_dev= as cow_device=
* Use new function _mnt_dev() from archiso.
* Check for error while setting loopback device.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

This allow to take control again of these mountpoints outside initramfs.
(i.e: on deinitramfs stage at shutdown for unmount it)

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Add some options to control where all COW files will be located.
Until this moment all files are located in a tmpfs filesystem.
Now is posible to set a device via a filesystem label or device node
plus a directory, where all these files will be stored.
All dm-snapshot devices will be persistent by default,
but this can be changed if wanted.

Take care, a filesystem that does not support sparse files maybe
is not the best choice for COW files, because they are created with
the same size (is apparent) like the read-only device (the image.fs inside .sfs).
Of course sooner or later, depending on use, these files actually end up
being as big as the read-only device.

KNOW-ISSUE: On shutdown in step "Unmounting Filesystems" will [FAIL],
    all filesystem except $cow_device, will be unmounted/remounted-ro.
    For this reason this change needs an archiso_shutdown hook, thats
    unmount all block devices used by archiso. (Work in progress)

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Separate this code from main mount hook, and make it more generic.

_mnt_dev(device, mountpoint, flags) ->
    wait for device and mount, launch a shell if something goes wrong.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Image size in MiB:

old -- new -- iso
679 -> 693    core-dual
374 -> 381    core-i686
380 -> 387    core-x86_64
309 -> 323    netinstall-dual
182 -> 189    netinstall-i686
185 -> 192    netinstall-x86_64

Was requested: mailing-list and bugtracker.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

This is now supported since util-linux pkg enable it.

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>

Fix: FS#18959 - download_repos.sh does not report tcp timeouts and such

* Always use pacman cache. Previously pacman cache was used if package
  already exists on it, now is also used to download pkg on it.
  Finally copy all packages from cache with cp and make core.db.
* Using pacman, we can sure that package integrity is right.
* Always get a [core] package, in case of [testing] enabled.
* Now works when mirror is: Server = file:// ;)

Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>