On Fri, Mar 13, 2009 at 08:52:01AM +0100, Hans de Goede wrote: > Almost there, see below for my comment. > > David Lehman wrote: > > <snip, all good so far> > >> @@ -1152,18 +1170,38 @@ class DeviceTree(object): >> rs.activate(mknod=True) >> # Create the DMRaidArray >> + if self.zeroMbr: >> + cb = lambda: True >> + else: >> + cb = lambda: questionInitializeDisk(self.intf, >> + rs.name) >> + >> + if not self.clearPartDisks or \ >> + rs.name in self.clearPartDisks: >> + # we'll need the array's sysfs path below >> + dm_node = dm.dm_node_from_name(rs.name) >> + _sysfs_path = "/class/block/%s" % dm_node >> + >> + # if the disk contains protected partitions >> + # we will not wipe the disklabel even if >> + # clearpart --initlabel was specified >> + initlabel = self.reinitializeDisks >> + for protected in self.protectedPartitions: >> + _p = "/sys/%s/%s" % (_sysfs_path, protected) >> + if os.path.exists(os.path.normpath(_p)): >> + initlabel = False >> + break > > This won't work with dm(raid) partitions as their block sysfs entries are not a subdir > of the containing "disk", they are a (partial) linear device mapping of their containing > disk, so they themselves have a sysfs path of /sys/class/block/dm-# too, just like > their containing disk (but with a higher #) > > All I can come up with here, is that you check if the protectedPartition name ends with p#, > and if so strip p#, and then see if what remains matches the name of the dmraid array. > > See udev.py: udev_device_get_dmraid_partition_disk(), but that needs udev info for > the partition which you do not have, although we could change its argument, as all it really > needs is the name. I wonder if its worth it to start handling sysfs paths in storage/devicelibs/dm.py or pyblock itself. I'd vote for dm.py. regards > > Regards, > > Hans > > _______________________________________________ > Anaconda-devel-list mailing list > Anaconda-devel-list@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/anaconda-devel-list -- Joel Andres Granados Brno, Czech Republic, Red Hat. _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
