From mboxrd@z Thu Jan 1 00:00:00 1970 From: Maxime Ripard Subject: Re: [PATCH v3 00/12] cmd: fdt: Add device tree overlays support Date: Mon, 27 Jun 2016 08:29:12 +0200 Message-ID: <20160627062912.GM4000@lukather> References: <20160624142757.32735-1-maxime.ripard@free-electrons.com> <20160626153719.GP15625@voom.fritz.box> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="JvUS8mwutKMHKosv" Return-path: Content-Disposition: inline In-Reply-To: <20160626153719.GP15625-RXTfZT5YzpxwFLYp8hBm2A@public.gmane.org> Sender: devicetree-compiler-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: To: David Gibson Cc: Pantelis Antoniou , Simon Glass , Boris Brezillon , Alexander Kaplan , Thomas Petazzoni , devicetree-compiler-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Antoine =?iso-8859-1?Q?T=E9nart?= , Hans de Goede , Tom Rini , u-boot-0aAXYlwwYIKGBzrmiIFOJg@public.gmane.org, Stefan Agner --JvUS8mwutKMHKosv Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi David, On Mon, Jun 27, 2016 at 01:37:19AM +1000, David Gibson wrote: > On Fri, Jun 24, 2016 at 04:27:45PM +0200, Maxime Ripard wrote: > > Hi, > >=20 > > The device tree overlays are a great solution to the issue raised by > > the bunch expandable boards we find everywhere these days, like the > > Beaglebone, Raspberry Pi or CHIP. > >=20 > > However, most of the time, the overlays are applied through a > > mechanism involving the firmware request interface in Linux, that is > > only fully functional once the userspace has been mounted and is > > running. > >=20 > > Some expansion boards might need to be enabled before that, because > > they simply need to patch the DT early on, or need to be initialized > > early in order to be fully functional, or because they provide access > > to the root filesystem. > >=20 > > In these cases, having the bootloader applying the overlay before > > Linux starts seems like the easiest solution. > >=20 > > This implementation doesn't provide all the Linux fancyness though, > > there's no transactional application, which means that if the overlay > > cannot be applied for a reason while you're still halfway through the > > application, you're probably screwed. It also cannot remove an > > overlay, but I don't think that it is currently a use-case. > >=20 > > There's still a bunch of work to extend the libfdt unit tests to test > > the new functions introduced, but these patches will be submitted > > in a near future. >=20 > Are you planning to send a new version of (the relevant portions of) > these against upstream libfdt? Of course, I just only have to work on the test cases now. Simon wanted to merge this in U-Boot for the next merge window, and then sync with libfdt when it would be merged, hence why I posted it first. Maxime --=20 Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com --JvUS8mwutKMHKosv Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAEBAgAGBQJXcMe4AAoJEBx+YmzsjxAgTDEP/iv1QRMN6RQVfVJD4/kkYjBu ccX7DUuDVAktu3u0ohA+XW3I2Wgt3xkpskXxojfgZX8zYhXH0xu+/zcalpax0GT4 7PDD1WcYra7giheryh7WI5fw5Q26olAaod2EpmsJzDVjCIlxzE5LeDFJdq93V+wj gKxYW7hek1rU/35BB8f3vNeC7oDj1QTo8ftblO8jHOK/xes4WeQW8jnCFlfVEp4C sJdlGDE7pXqFBG3XwshFISaqC0Tyi5Lh9WBYPkjxNMovkYsO0TLZqOenFHbZWCrJ E3km0tefLGcSPNqg0N8/Welw3mk9KN7T5d/uehK0B074J9FRcM2fX1tZln7lHztx DvfT4r/7Dmtmpv/Wl5wpc2WFRiec/NCgSLnVFN3xsTz26wEQQPtHcc3ig+h8Z/E8 H+6CLh1U6MJTPgqTnL2NyVYhtCq86F7Ui9npuTvZU4hA/Q5Sqx/jESOtqJm2DsYv Jp1EToIcEXLbEONZd8sep2OkEdYt466z7owJ8xPgzzsRLkklKTNXCutA4EFWVyD8 MPig00rBL0ZhxVDRw8+SwviZHdzYdx/+O/i0KKBuK2rGYYD8LTIaXVolvn5QCx0z lBBW92YXILnpk1eEsr70VsT1/8dbLFuHvKuPKcMLzMoDV9jHCBSWKQcemSruDDEA P+DAlMDUZ7ufCYQ+kihF =ShYd -----END PGP SIGNATURE----- --JvUS8mwutKMHKosv--