[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250331171755.GC289482@nvidia.com>
Date: Mon, 31 Mar 2025 14:17:55 -0300
From: Jason Gunthorpe <jgg@...dia.com>
To: Dan Williams <dan.j.williams@...el.com>
Cc: Dave Jiang <dave.jiang@...el.com>,
"Paul E. McKenney" <paulmck@...nel.org>, linux-cxl@...r.kernel.org,
dave@...olabs.net, jonathan.cameron@...wei.com,
alison.schofield@...el.com, vishal.l.verma@...el.com,
ira.weiny@...el.com, gourry@...rry.net,
linux-kernel@...r.kernel.org, linux-next@...r.kernel.org,
sfr@...b.auug.org.au
Subject: Re: [BUG -next] ./usr/include/cxl/features.h:11:10: fatal error:
uuid/uuid.h: No such file or directory
On Mon, Mar 31, 2025 at 09:54:55AM -0700, Dan Williams wrote:
> Jason Gunthorpe wrote:
> > On Fri, Mar 28, 2025 at 05:26:42PM -0700, Dave Jiang wrote:
> > > > For now the following builds for me, but it is a quite a mess to undo
> > > > the assumption that that the hardware object definitions can not use
> > > > uuid_t:
> > >
> > > +Jason.
> >
> > Seems invasive?
>
> Yeah, it left a bad taste for me as well.
>
> > Maybe just like below?
>
> I like that this avoids converting to the kernel's uuid API, however,
> not quite happy that it forces userspace to contend with the
> type-conflict with uuid/uuid.h.
Oh I see
> So how about one more riff on your idea?
Sure, works for me, please post it..
Jason
Powered by blists - more mailing lists