eeh patch

Greg KH greg at
Wed Mar 17 08:39:08 EST 2004

On Tue, Mar 16, 2004 at 10:21:15AM -0800, Greg KH wrote:
> > In an attempt to ease the number of EEH events some people have
> > seen, it was decided to add the ability to hot-unplug certain
> > devices when they received an eeh event instead of calling panic().
> >
> > This is done via an user-space daemon that does a read on a /proc
> > file that blocks in the kernel until an eeh event occurs.  The
> > kernel then returns the device name to the daemon who handles
> > the actual unplugging of the device.
> >
> > The (2.6) patch and user-space daemon code have been tested, so I am
> > hoping to get some feedback before pushing this on to Ameslab.
> How about "what the hell are you doing?" as feedback.  :)

Ok, sorry if anyone got offended by this comment, I didn't mean to be so
snippy.  I apologize for this.

But my main point stands.  If you are going to try to invent a new way
to talk to the kernel from userspace, please take about 10 deep breaths
and back away from the keyboard.  Then go see how the kernel does things
today and please please please use that interface instead.

This is only about the 240th time this same topic has come up in the
past on any one of a zillion different kernel mailing lists
(linux-kernel, linux-hotplug-devel, etc.)  Feel free to read the
archives of those lists to see what Linus thinks of interfaces like you
just created, and what I think of /proc.  So again, if you are going to
be writing kernel code, please do your homework first before adding new

And if you think you do want to create a new interface, please, bring it
up in linux-kernel where it belongs.  There is no reason the ppc64 tree
should be doing things on their own for such an important interaction.
We should spread the goodness to all other arches too.

Again, sorry for any hurt feelings that my post might have caused, I was
only critiquing the code that was published.

greg k-h

** Sent via the linuxppc64-dev mail list. See

More information about the Linuxppc64-dev mailing list