* Rusty Russell <ru...@rustcorp.com.au> [2010-06-20 20:31]: > On Sat, 19 Jun 2010 04:08:03 am Ryan Harper wrote: > > With the availablility of a sysfs device attribute for examining disk serial > > numbers the ioctl is no longer needed. The user-space changes for this > > aren't > > upstream yet so we don't have any users to worry about. > > If John Cooper acks this, I'll push it to Linus immediately. > > Unfortunately we offered this interface in 2.6.34, and we're now removing it. > That's unpleasant.
Yes; well. There's a story as there always is. John can tell it better than I, but it goes something like: John cooked up some patches, one of which was an example use of the serial string including a VBID ioctl. No one got around to doing a sysfs interface and somehow the ioctl side in virtio-blk got picked up. Working with what was available, I pushed some patches to linux-hotplug to get this whole virtio-blk serial and disk/by-id symlinks working and was met with: why does a new kernel driver have an ioctl interface and we don't want to collect additional single-use binaries in the udev tree. So now, the sysfs serial attribute patch and with it, no need and no users of ioctl. *whew* > > Thanks, > Rusty. > PS. John should have been cc'd on these patches! He's cc'ed on the others, just not on this removal one. I need to learn git-send-email better, I explicitly added him as --cc on when sending; next time I'll look closer at the headers during --dry-run. -- Ryan Harper Software Engineer; Linux Technology Center IBM Corp., Austin, Tx ry...@us.ibm.com
