[blivet] small md cleanups (v2)

Brian C. Lane bcl at redhat.com
Wed Aug 6 00:34:01 UTC 2014


On Mon, Aug 04, 2014 at 10:14:56AM -0500, David Lehman wrote:
> New patch 1 teaches udev.device_get_md_device_uuid to ask udev first
> instead of relying solely on having run mdexamine.
> 
> New patch 2 takes advantage of md member and array UUIDs being in udev
> to delay running mdexamine until handleUdevMDMemberFormat. This also
> allows us to drop the md caching patch from v2 of the pyudev set.
> 
> Patch 3 is unchanged.
> 
> Patch 4 is generalized so the name is updated regardless of how we found
> the device and also to bypass the name setter, which disallows setting a
> name for existing devices.
> 
> Patch 5 has updated comment text.
> 
> New patch 6 removes obsolete code to avoid returning an md array UUID as
> the UUID of an md member.
> 
> 
> At some point it might be worthwhile to change things so that the udev
> functions handle the ID_FS_UUID_SUB situation transparently, but that will
> have to be a separate effort.
> 
> David Lehman (6):
>   Look in udev data for md member UUID.
>   Get md member and array UUIDs for format ctor from udev.
>   Remove an obsolete block related to unpredictable md device names.
>   Update md name when lookup relied on UUID.
>   Don't pass md array UUID as member format UUID.
>   Simplify udev.device_get_uuid.
> 
>  blivet/devicetree.py | 50 ++++++++++----------------------------------------
>  blivet/udev.py       | 11 ++++-------
>  2 files changed, 14 insertions(+), 47 deletions(-)

These all look ok to me.

-- 
Brian C. Lane | Anaconda Team | IRC: bcl #anaconda | Port Orchard, WA (PST8PDT)


More information about the anaconda-patches mailing list