From: Ilias Apalodimas Date: Wed, 30 Dec 2020 23:17:44 +0200 Subject: [PATCH 0/8 v2] Change logic of EFI LoadFile2 protocol for initrd loading In-Reply-To: <2a0b4a86-0c36-f46f-1ce9-3f73693eca51@gmx.de> References: <20201230150722.154663-1-ilias.apalodimas@linaro.org> <2a0b4a86-0c36-f46f-1ce9-3f73693eca51@gmx.de> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Hi Heinrich, On Wed, Dec 30, 2020 at 09:44:39PM +0100, Heinrich Schuchardt wrote: > On 12/30/20 4:07 PM, Ilias Apalodimas wrote: > > Hi! > > > > This is v2 of [1]. > > > > There's been a couple of changes regarding where we install the protocol. > > The initial patchset was completely disregarding BootNext, so that's taken > > into account now and we can use the new feature. > > This brought a few changes on the selftests as well, since we now use the > > loaded image handle to install the protocol, as a consequence the custom > > handle in the tests is now uninstalled during the test .teardown(), but > > the overall approach remains identical. > > Boot#### contains a device path. > > Why should Initrd#### contain something U-Boot specific and not a device > path? Because there was no standard on it, since we used the path as a config option up to now I just kept it as is. > > Think of a Linux distribution. When updating the kernel it will have to > write Boot#### and Initrd####. It can determine the current boot device > via BootCurrent and Boot####. Next it can add the new file paths for > Boot#### and Initrd#### relative to the same device. This isn't a bad idea. Let me update the patch and send a v3. Cheers /Ilias