X-Git-Url: http://git.annexia.org/?a=blobdiff_plain;f=helper%2Ffebootstrap-supermin-helper.pod;h=c6c551cee5828636005769f8351e0c0ba340ae38;hb=0f89ba0654de234429042ffcc91c8a0de94ec98b;hp=29f4b95e4443c1ef9c8bb32c39763b9bb6811a07;hpb=c0a9555e8559a0dbf432367f87c59c2a49520c99;p=febootstrap.git

diff --git a/helper/febootstrap-supermin-helper.pod b/helper/febootstrap-supermin-helper.pod
index 29f4b95..c6c551c 100644
--- a/helper/febootstrap-supermin-helper.pod
+++ b/helper/febootstrap-supermin-helper.pod
@@ -7,6 +7,10 @@ febootstrap-supermin-helper - Reconstruct initramfs from supermin appliance.
  febootstrap-supermin-helper supermin.img hostfiles.txt host_cpu kernel initrd
  febootstrap-supermin-helper input [...] host_cpu kernel initrd
 
+ febootstrap-supermin-helper -f ext2 input [...] host_cpu kernel initrd appliance
+
+ febootstrap-supermin-helper -f checksum input [...] host_cpu
+
 =head1 DESCRIPTION
 
 I<febootstrap-supermin-helper> reconstructs a bootable kernel and
@@ -16,8 +20,8 @@ L<febootstrap-to-supermin(8)>.
 
 =head1 PARAMETERS
 
-Of the four or five required parameters, the first few are I<input>
-files, and the last two are I<output> files.
+Of the required parameters, the first few are I<input> files, and the
+last two or three are I<output> files.
 
 C<supermin.img> and C<hostfiles.txt> are the input files which
 describe the supermin appliance.  (You can also use a directory name
@@ -25,14 +29,53 @@ here which is searched for files).
 
 C<host_cpu> should be the host CPU, eg. C<x86_64> or C<i686>.
 
-C<kernel> and C<initrd> are the temporary output files that this
-script produces.  These output files are meant to be used just for
-booting the appliance, and should be deleted straight afterwards.
+C<kernel>, C<initrd> and C<appliance> are the temporary output files
+that this script produces.  These output files are meant to be used
+just for booting the appliance, and should be deleted straight
+afterwards.  The extra C<appliance> parameter is only required when
+the format is C<ext2>.  None of these parameters are needed for
+the checksum output C<-f checksum>.
 
 =head1 OPTIONS
 
 =over 4
 
+=item B<-f fmt> | B<--format fmt>
+
+Select the output format for the appliance.  Possible formats are:
+
+=over 4
+
+=item cpio
+
+A Linux initramfs.  This is the default.
+
+In this case you have to supply names for the C<kernel>
+and C<initrd>, where the C<initrd> is the appliance.
+
+=item ext2
+
+An ext2 filesystem.
+
+In this case you have to supply names for the C<kernel>, a small
+C<initrd> which is used just to locate the appliance, and the
+C<appliance> (the ext2 filesystem).
+
+=item checksum
+
+Output a checksum.
+
+This prints a checksum which only changes when one of the input files
+changes.
+
+You can use this in order to cache the output of a previous run of
+this program: computing the checksum is much quicker than building an
+appliance, and you only need to invalidate the cache (and consequently
+rebuild the appliance) when the checksum changes.  Note that the
+host_cpu and the UID of the current user are included in the checksum.
+
+=back
+
 =item B<-k file> | B<--kmods file>
 
 If this option is specified, then C<file> should be a list of