[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [rhel6-branch 4/4] mpath: filter out the slave devices and their partitions.



On 01/07/2011 06:36 PM, Brian C. Lane wrote:
On Fri, Jan 07, 2011 at 02:04:15PM +0100, Ales Kozumplik wrote:
On 01/06/2011 05:24 PM, David Lehman wrote:
Please do use udev_device_is_dm_mpath instead of adding another function
to do the same thing. I forgot that function had made it to
rhel6-branch.

Dave

I said yesterday this is good, but now I've found one more thing we
need to address first. I can not use udev_device_is_dm_mpath to test
for a partition of an mpath device, the function uses
udev_device_dm_subsystem_match() that looks at the first field in
DM_UUID which is like "mpath-2003013842bcb000c" for multipath
devices but "part1-mpath-2003013842bcb000c" for multipath
partitions.

I could workaround this on rhel6 branch where there still is the
disgraceful udev_device_is_multipath_partition(), but then I'd hit
the problem on master where the function was removed.

I am thinking about solving this by adding a parameter to
ude_device_dm_subsystem_match() that will make it skip the first
DM_UUID field if it was 'part\d+'.

FYI, this is related to rhbz#634771, so a special check for ^part\d+ is
valid for RHEL but not Fedora.


I'm not sure if this has been fixed on master yet, or do you think so? In that case we can simplify udev_dm_subystem_match() again.

Ales


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]