On 10/29/2009 01:51 AM, David Cantrell wrote:
> CONFIG_NETIUCV has been disabled in the kernel, so remove support for
> these devices during installation.
> ---
> loader/linuxrc.s390 | 228 +++++++++++++--------------------------------------
> network.py | 4 +-
> 2 files changed, 59 insertions(+), 173 deletions(-)
>
> diff --git a/loader/linuxrc.s390 b/loader/linuxrc.s390
> index 3b2a8d4..d50dcbd 100644
> --- a/loader/linuxrc.s390
> +++ b/loader/linuxrc.s390
Regarding loader/linuxrc.s390, it would be nice to explicitly inform the user that NETTYPE=iucv in (e.g. old existing) parm/conf files is no longer allowed so they are aware and can adapt their files if desired. We do so for CHANDEV, NETWORK, and BROADCAST already. Otherwise the user would just see the following generic error message regarding the value of NETTYPE which applies to parm/conf file as well as interactive input:
"Incorrect format or value for network type (NETTYPE): iucv"
Maybe something like the following:
diff --git a/loader/linuxrc.s390 b/loader/linuxrc.s390
index 345eb5e..422e69e 100644
--- a/loader/linuxrc.s390
+++ b/loader/linuxrc.s390
@@ -2750,6 +2750,12 @@ fi
[ -n "$MMTU" ] && mmtu_was_set=$MMTU
[ -n "$VSWITCH" ] && vswitch_was_set=$VSWITCH
+if [ "$NETTYPE" = "iucv" ]; then
+ echo
+ echo $"The value iucv for the NETTYPE variable is no longer supported,"
+ echo $"please update your .parm or the .conf file to use another NETTYPE instead."
+ echo
+fi
[ -n "$CHANDEV" ] && do_chandev
[ -n "$NETWORK" ] && do_network
[ -n "$BROADCAST" ] && do_broadcast
> @@ -670,9 +668,7 @@ function syntax_check_domainname() {
> }
>
> function modprobe_alias() {
> - if [ ":$NETTYPE" = ":iucv" ]; then
> - echo "alias $DEVICE netiucv" >> /tmp/modprobe.conf
> - elif [ ":$NETTYPE" = ":ctc" ]; then
> + if [ ":$NETTYPE" = ":ctc" ]; then
> echo "alias $DEVICE ctcm" >> /tmp/modprobe.conf
> else
> echo "alias $DEVICE $NETTYPE" >> /tmp/modprobe.conf
I suppose we could even remove modprobe_alias entirely since it seems deprecated since a while and it should work without even on s390.
> @@ -752,28 +748,19 @@ function rollback_config() {
> [ -n "$DEVICE" ] && tv ip -6 route flush default dev $DEVICE
> # address flush seems to be effective for all address families
> [ -n "$DEVICE" ] && ip address flush dev $DEVICE
> - # iucv device needs to be down before removal
Removing the comment is fine.
> - [ -n "$DEVICE" ] && ip link set down $DEVICE &> /dev/null
I'd like to let the "link set down" stay here. It does not hurt and in fact makes it more robust because we need the device down anyway and probably shouldn't rely on it going down implicitly.
> diff --git a/network.py b/network.py
> index e355398..248255a 100644
> --- a/network.py
> +++ b/network.py
> @@ -125,7 +125,7 @@ def getDefaultHostname(anaconda):
>
> # return if the device is of a type that requires a ptpaddr to be specified
> def isPtpDev(devname):
> - if (devname.startswith("ctc") or devname.startswith("iucv")):
> + if devname.startswith("ctc"):
> return True
> return False
>
> @@ -261,8 +261,6 @@ class NetworkDevice(SimpleConfigFile):
> self.info = { "DEVICE" : dev }
> if dev.startswith('ctc'):
> self.info["TYPE"] = "CTC"
> - elif dev.startswith('iucv'):
> - self.info["TYPE"] = "IUCV"
>
> class Network:
> def __init__(self):
I refuse any statement on this last hunk. That code seems to be leftovers to me that are unused and I don't know what they were used for. I just hope partial removal here doesn't break anything.
Steffen
Linux on System z Development
IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Erich Baier
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/anaconda-devel-list
