Re: [BUG 2.6.17-git] kmem_cache_create: duplicate cache scsi_cmd_cache

[Greg KH <greg@kroah.com>]

On Fri, May 12, 2006 at 11:28:16PM +0100, Al Viro wrote:
> On Fri, May 12, 2006 at 03:22:42PM -0700, Linus Torvalds wrote:
> > 
> > 
> > On Fri, 12 May 2006, Al Viro wrote:
> > > 
> > > Secondary question: who had resurrected that crap?  I distinctly remember
> > > killing it off...
> > 
> > If you did, I don't think it ever got into the kernel.
> > 
> > It was added by Kay Sievers on Nov 3, 2004, according to the old history 
> > (back then it was in drivers/block/genhd.c, and the function was called 
> > "block_hotplug()", but apart from renaming the function and moving the 
> > file, it's recognizably the same.
> > 
> > Of course, you may have killed off an even earlier incarnation..
> 
> Ah, right - there was another uevent mess in fs/super.c.  Sorry, got them
> confused... and that FPOS _is_ back.
> 
> gregkh, may I ask why had bdev_uevent() been resurrected?

One user, running an old version of HAL, on a distro that had an updated
version available (Gentoo), refused to upgrade it and was upset that
when they plugged a usb-storage device in, it did not show up as an icon
on the desktop.

Andrew pointed out that we broke the "don't break the userspace API"
kernel rule, so I put the patch back, and noted when it would finally be
removed.

I would gladly remove it again if everyone can agree that they will not
get upset at me again when someone running obsolete/broken userspace
programs complains...

thanks,

greg k-h