5:55 a.m.
Add a directory to contain required augeas lenses which aren't yet upstream.
Include a new lens for grub's device.map.
---
augeas/README.txt | 4 ++++
augeas/device_map.aug | 28 ++++++++++++++++++++++++++++
2 files changed, 32 insertions(+), 0 deletions(-)
create mode 100644 augeas/README.txt
create mode 100644 augeas/device_map.aug
diff --git a/augeas/README.txt b/augeas/README.txt
new file mode 100644
index 0000000..709be7c
--- /dev/null
+++ b/augeas/README.txt
@@ -0,0 +1,4 @@
+This directory contains augeas lenses required by virt-v2v.
+
+device_map.aug posted upstream:
+http://www.redhat.com/archives/augeas-devel/2010-June/msg00027.html
diff --git a/augeas/device_map.aug b/augeas/device_map.aug
new file mode 100644
index 0000000..090edd5
--- /dev/null
+++ b/augeas/device_map.aug
@@ -0,0 +1,28 @@
+(* Parsing grub's device.map *)
+
+module Device_map =
+ autoload xfm
+
+ let sep_tab = Sep.tab
+ let eol = Util.eol
+ let fspath = Rx.fspath
+ let del_str = Util.del_str
+
+ let comment = Util.comment
+ let empty = Util.empty
+
+ let dev_name = /(h|f|c)d[0-9]+(,[0-9a-zA-Z]+){0,2}/
+ let dev_hex = /0x[0-9a-fA-F]+/
+ let dev_dec = /[0-9]+/
+
+ let device = del_str "(" . key ( dev_name | dev_hex | dev_dec ) . del_str
")"
+
+ let map = [ device . sep_tab . store fspath . eol ]
+
+ let lns = ( empty | comment | map ) *
+
+ let xfm = transform lns (incl "/boot/grub/device.map")
+
+(* Local Variables: *)
+(* mode: caml *)
+(* End: *)
--
1.7.1
5:55 a.m.
New subject: [PATCH 2/2] Update grub's device.map
This patch updates device paths in grub's device.map.
Fixes RHBZ#609448
---
lib/Sys/VirtV2V/GuestOS/RedHat.pm | 43 ++++++++++++++++++++++++++++++++----
1 files changed, 38 insertions(+), 5 deletions(-)
diff --git a/lib/Sys/VirtV2V/GuestOS/RedHat.pm b/lib/Sys/VirtV2V/GuestOS/RedHat.pm
index f881612..8898576 100644
--- a/lib/Sys/VirtV2V/GuestOS/RedHat.pm
+++ b/lib/Sys/VirtV2V/GuestOS/RedHat.pm
@@ -1540,13 +1540,29 @@ sub remap_block_devices
# that all Fedora distributions in use use libata.
if ($libata) {
- # Look for IDE and SCSI devices in fstab for the guest
+ # Look for IDE and SCSI devices in guest config files
my %guestif;
eval {
+ my @guestdevs;
+
foreach my $spec ($g->aug_match('/files/etc/fstab/*/spec')) {
my $device = $g->aug_get($spec);
- next unless($device =~ m{^/dev/(sd|hd)([a-z]+)});
+ push(@guestdevs, $device);
+ }
+
+ foreach my $key ($g->aug_match('/files/boot/grub/device.map/*'))
{
+ $key =~ m{/files/boot/grub/device.map/(.*)} or die;
+ my $gdev = $1;
+
+ next if ($gdev =~ /^#comment/);
+
+ my $odev = $g->aug_get($key);
+ push(@guestdevs, $odev);
+ }
+
+ foreach my $dev (@guestdevs) {
+ next unless($dev =~ m{^/dev/(sd|hd)([a-z]+)});
$guestif{$1} ||= {};
$guestif{$1}->{$1.$2} = 1;
}
@@ -1554,8 +1570,8 @@ sub remap_block_devices
$self->_augeas_error($@) if ($@);
- # If fstab contains references to sdX, these could refer to IDE or SCSI
- # devices. We may need to update them.
+ # If guest config contains references to sdX, these could refer to IDE
+ # or SCSI devices. We may need to update them.
if (exists($guestif{sd})) {
# Look for IDE and SCSI devices from the domain definition
my %domainif;
@@ -1613,7 +1629,7 @@ sub remap_block_devices
}
eval {
- # Go through fstab again, updating bare device references
+ # Update bare device references in fstab
foreach my $spec ($g->aug_match('/files/etc/fstab/*/spec')) {
my $device = $g->aug_get($spec);
@@ -1625,6 +1641,23 @@ sub remap_block_devices
$g->aug_set($spec, "/dev/".$map{$name}.$part);
}
+ # Update device.map
+ foreach my $key ($g->aug_match('/files/boot/grub/device.map/*')) {
+ $key =~ m{/files/boot/grub/device.map/(.*)} or die;
+ my $gdev = $1;
+
+ next if ($gdev =~ /^#comment/);
+
+ my $odev = $g->aug_get($key);
+
+ next unless($odev =~ m{^/dev/([a-z]+)(\d*)});
+ my $name = $1;
+ my $part = $2;
+
+ next unless(exists($map{$name}));
+
+ $g->aug_set($key, "/dev/".$map{$name}.$part);
+ }
$g->aug_save();
};
--
1.7.1
12:13 p.m.
New subject: [PATCH 2/2] Update grub's device.map
On Wed, Jun 30, 2010 at 11:55:51AM +0100, Matthew Booth wrote:
+ foreach my $key
($g->aug_match('/files/boot/grub/device.map/*')) {
+ $key =~ m{/files/boot/grub/device.map/(.*)} or die;
+ my $gdev = $1;
+
+ next if ($gdev =~ /^#comment/);
Shouldn't the Augeas lens take care of these?
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
New in Fedora 11: Fedora Windows cross-compiler. Compile Windows
programs, test, and build Windows installers. Over 70 libraries supprt'd
http://fedoraproject.org/wiki/MinGW http://www.annexia.org/fedora_mingw
4:31 p.m.
New subject: [PATCH 2/2] Update grub's device.map
On 02/07/10 18:13, Richard W.M. Jones wrote:
On Wed, Jun 30, 2010 at 11:55:51AM +0100, Matthew Booth wrote:
> + foreach my $key
($g->aug_match('/files/boot/grub/device.map/*')) {
> + $key =~ m{/files/boot/grub/device.map/(.*)} or die;
> + my $gdev = $1;
> +
> + next if ($gdev =~ /^#comment/);
Shouldn't the Augeas lens take care of these?
Augeas parses comments and returns them in the tree.
Matt
4:52 p.m.
New subject: [PATCH 2/2] Update grub's device.map
On Mon, Jul 12, 2010 at 10:31:05PM +0100, Matthew Booth wrote:
On 02/07/10 18:13, Richard W.M. Jones wrote:
> On Wed, Jun 30, 2010 at 11:55:51AM +0100, Matthew Booth wrote:
>> + foreach my $key
($g->aug_match('/files/boot/grub/device.map/*')) {
>> + $key =~ m{/files/boot/grub/device.map/(.*)} or die;
>> + my $gdev = $1;
>> +
>> + next if ($gdev =~ /^#comment/);
>
> Shouldn't the Augeas lens take care of these?
Augeas parses comments and returns them in the tree.
Oh I see, it's matching on the key name. In that case, ACK.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
virt-top is 'top' for virtual machines. Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://et.redhat.com/~rjones/virt-top
12:11 p.m.
New subject: [PATCH 1/2] Add new augeas directory with grub device.map lens
On Wed, Jun 30, 2010 at 11:55:50AM +0100, Matthew Booth wrote:
Add a directory to contain required augeas lenses which aren't
yet upstream.
Include a new lens for grub's device.map.
And these are not upstream why? I think we should just push these
upstream fast as possible, and push them into the Fedora augeas
package as patches as we need them.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
New in Fedora 11: Fedora Windows cross-compiler. Compile Windows
programs, test, and build Windows installers. Over 70 libraries supprt'd
http://fedoraproject.org/wiki/MinGW http://www.annexia.org/fedora_mingw
5061
days inactive
5073
days old
5 comments
2 participants
participants (2)
-
Matthew Booth -
Richard W.M. Jones