Re: [PATCH] 2.5.22: ibm partition support.

[Pete Zaitcev <zaitcev@redhat.com>]

> another resend of the partition patch for ibm.c. Nobody sent a veto so far
> so please add it.

I suspect Linus wants it fixed "right", as we promised.
I made the kludge with genhd_dasd_ioctl because Red Hat
need to ship _something_ working while passing Viro approval.
I never intended it to be applied to 2.5.

Remember that what Al did first _almost_ worked. The recursion
happened because the SCSI code was so insanely convoluted,
and only for SCSI removables. Perhaps it may be doable to
fix that instead. Also I am thinking about special-casing
the minor 0 and make it open without partition probing;
then the probing happens on top of pre-opened minor 0
and spawns other minors. That should be good and proper.

The bad part is that the real fix for old broken ibm.c partitioning
is now a hostage of resolution of ioctl-during-partitioning
problem. I see no easy way to resolve it, it has to wait.
Good thing my Hercules uses FBAs :)

-- Pete