On Sat, Jul 06, 2019 at 08:17:29PM +1000, Stephen Rothwell wrote:
> Hi Greg,
>
> On Sat, 6 Jul 2019 11:46:47 +0200 Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> >
> > On Sat, Jul 06, 2019 at 07:44:12PM +1000, Stephen Rothwell wrote:
> > >
> > > On Sat, 6 Jul 2019 10:34:33 +0200 Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > > >
> > > > On Thu, Jul 04, 2019 at 10:24:50PM +1000, Stephen Rothwell wrote:
> > > > >
> > > > > This release produces a whole lot (over 200) of this message in my qemu
> > > > > boot tests:
> > > > >
> > > > > [ 1.698497] debugfs: File 'sched' already present!
> > > > >
> > > > > Introduced by commit
> > > > >
> > > > > 43e23b6c0b01 ("debugfs: log errors when something goes wrong")
> > > > >
> > > > > from the driver-core tree. I assume that the error(?) was already
> > > > > happening, but it is now being reported.
> > > >
> > > > What are you passing to qemu to get this? I just tried it myself and
> > > > see no error reports at all. Have a .config I can use to try to
> > > > reproduce this?
> > >
> > > It is a powerpc pseries_le_defconfig kernel and I run qemu like this:
> > >
> > > qemu-system-ppc64 -M pseries -m 2G -vga none -nographic -kernel vmlinux -initrd rootfs.cpio.gz
> >
> > Hm, I think my rootfs initrd might be quite simple compared to yours (it
> > drops me into a busybox shell). Any pointers to where you created yours
> > from?
>
> Michael Ellerman gave it to me. It is very simple. Its /init is just
>
> $ cat init
> #!/bin/sh
> # devtmpfs does not get automounted for initramfs
> /bin/mount -t devtmpfs devtmpfs /dev
> exec 0</dev/console
> exec 1>/dev/console
> exec 2>/dev/console
> exec /sbin/init $*
>
> and /sbin/init is a link to /bin/busybox
>
> It is all run by an expect script that just waits for the login:
> prompt, logs in a root and runs "halt".
>
> All the debugfs messages appear before the kernel finished booting.
Ah, found the bug. Turns out I caused it in the blk queue debugfs code
a short while ago. I'll post a patch for this in a few minutes...
thanks,
greg k-h
