From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Mon, 3 May 2021 08:57:33 +0200 Subject: [Buildroot] [PATCH v2 3/6] package/riscv64-elf-toolchain: new package In-Reply-To: <87tunk2veq.fsf@dell.be.48ers.dk> References: <20210502212141.934384-1-thomas.petazzoni@bootlin.com> <20210502212141.934384-11-thomas.petazzoni@bootlin.com> <87tunk2veq.fsf@dell.be.48ers.dk> Message-ID: <20210503085733.59777dce@windsurf> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, On Mon, 03 May 2021 08:17:33 +0200 Peter Korsgaard wrote: > Out of interest, why is that? For the ARM32 case it was a question of > 32bit / 64bit, but that does not seem to be the issue here? > > Just crappy firmware or are there any deeper differences between a bare > metal and Linux toolchain on riscv64? It is not yet 100% sure what's going on. The firmware clearly doesn't build with a Linux toolchain. I've done a bit of work to try to get it to build, by adding -ffrestanding -nostdinc, but it does include a few standard headers (stdint.h for example). Once this was fixed, the build system of the firmware uses a specific gcc spec file, called nano.spec, which isn't available except in their bare-metal toolchain. So for now, I opened a bug report at https://github.com/starfive-tech/beagle_ddrlnit/issues/4, and decided to go on with building this with the recommended bare-metal toolchain. Best regards, Thomas -- Thomas Petazzoni, co-owner and CEO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com