Re: Regression in 4.6.0-git - bisected to commit dd254f5a382c

[Al Viro]

On Tue, May 24, 2016 at 09:06:03PM -0500, Matthew McClintock wrote:
> 
> > On May 24, 2016, at 8:28 PM, Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote:
> > 
> > The next obvious question is which binary it is and what's the return
> > address to userland; make that
> > 	if (!size)
> > 		printk(KERN_ERR "crap in %s[%x]",
> > 			current->comm,
> > 			current_pt_regs()->rip);
> > (in the same place)
> 
> This is an ARM board, no rip. Quick log without, can re-run again later tonight or maybe tomorrow:

probably ->ARM_pc, then...

> [    8.008054] crap in udevd

Indeed.  Which version of that Fine Piece Of Software is it?  Again, the
kernel-side bug is real and needs fixing, but "udev does something bogus"
is not surprising in the slightest...