Re: [RFC PATCH] add wait_event_*_lock() functions

[Nishanth Aravamudan <nacc@us.ibm.com>]

On Thu, Feb 10, 2005 at 10:21:58AM -0800, David Brownell wrote:
> On Thursday 10 February 2005 9:39 am, Nishanth Aravamudan wrote:
> > Hi David, LKML,
> > 
> > It came up on IRC that the wait_cond*() functions from
> > usb/serial/gadget.c could be useful in other parts of the kernel. Does
> > the following patch make sense towards this? 
> 
> I know that Al Borchers -- who wrote those -- did so with that
> specific notion.  And it certainly makes sense to me, in
> principle, that such primitives exist in the kernel ... maybe
> with some tweaks first.  (And docs for all the wait_* calls?)

I would be happy to document all the wait_* callers, especially when
which should be used, their correspondence to the other sleep-functions,
etc.

> But nobody's pressed the issue before, to the relevant audience:
> namely, LKML.  I'd be interested to hear what other folk think.
> Clearly these particular primitives don't understand how to cope
> with nested spinlocks, but those are worth avoiding anyway.

Yes, I was considering that issue, but I figured let's go for the simple
case now and that should be good enough for *most* cases.

Thanks for the feedback!

-Nish