On 02/16/2012 03:16 PM, David Cantrell wrote:
On Thu, Feb 16, 2012 at 11:50:43AM +0100, Radek Vykydal wrote:
To sum up: to me it seems that readNetInfo shouldn't have been removed,
then the loaderData->netDev and loaderData->netDev_set flags would
be set properly and the condition in kickstartNetworkUp would prevent
device reactivation.
Yeah, that is one way to handle the problem and I did explore that angle,
but removal is far easier. If readNetInfo() was to stay, then linuxrc.s390
needed to learn how to read and parse dracut configuration data so it could
then write out /tmp/s390net. Unfortunately dracut does not currently
write out things like SUBCHANNELS and stuff like that, so if I wanted to
make it write /tmp/s390net as it had been, it would have had to add more
code to linuxrc.s390 to do that.
And really, all of that is a pointless exercise because dracut has already
taken care of configuring the interface. This is why I chose the route of
removal.
And that is why I put in all the #if's in loader's code to prevent certain
steps from running on s390x because the iface_t struct is not populated
correctly. I would rather not waste time teaching linuxrc.s390 how to
read and parse dracut options -or- teach loader how to read and parse
dracut options because both linuxrc.s390 and loader are going away.
The primary goal of this patchset is to just prop up the code on the
rhel7-alpha-branch enough to run and be moderately useful. All of this
is throw-away anyway.
Fair enough, ack then.
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/anaconda-devel-list
