Re: rootdelay

From: David N. Welton
Date: Mon Apr 25 2005 - 04:49:49 EST

Daniel Drake <dsd@xxxxxxxxxx> writes:

[ Please CC replies to me - thanks! ]

> Hi David,

> David N. Welton wrote:

> > [ Please CC replies to me, thanks! ]

> > Hi, I was looking at your patch:

> > Very small, which is nice. I

> > was wondering if there were any interest in my own efforts in that
> > direction:

> > which

> > is far more intrusive, and perhaps isn't good kernel programming
> > style, but, on the other hand, is the optimal solution in terms of
> > boot time because it wakes up the boot process right when the
> > device comes on line. Since I saw your patch included, it looks
> > like there is interest in this, and I'd toot my own horn once more
> > before just leaving my patch to the bit rot of the ages...
> > Thanks!

> As simple as it may be, it's a bit of a shame that we actually need
> rootdelay as its something that the kernel should do
> automatically. At the time when we last discussed it, we didn't come
> up with a better (and safe) way to handle it, but I don't think we
> considered anything like your implementation.

> I've CC'd a few people who were involved the last time around to see
> if they have any input for you.

Thanks! I don't wish to be a pest, but not having heard a "no", I'll
send another ping out. Perhaps a simple description is better than
the patch for busy people:

In init/do_mounts.c, mount_root does an interruptible_sleep_on a
wait queue, and goes on about its business after register_blkdev
in drivers/block/genhd.c does a wake_up_interruptible on it, so
that mounting the root device happens exactly when it needs to, no
sooner, no later, and doesn't depend on any fiddly timing issues.

David N. Welton

Apache, Linux, Tcl Consulting
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at