Date: Sun, 12 Nov 2017 08:40:25 -0700 From: Warner Losh <imp@bsdimp.com> To: Ed Maste <emaste@freebsd.org> Cc: Keith White <kwhite@eecs.uottawa.ca>, freebsd-arm@freebsd.org, Warner Losh <imp@freebsd.org> Subject: Re: RPI3 fails to boot: boot1.efi as bootaa64.efi on RPI3 [patch] Message-ID: <CANCZdfqvhbb_ZFHOC7qejr_dNP%2BAOn45vgpvPS5mP85KGdKWcw@mail.gmail.com> In-Reply-To: <CAPyFy2AkSBaMA1RV2usbkxhdO0BrefvFhNTx0ftd3=PKoR81%2BA@mail.gmail.com> References: <alpine.BSF.2.21.1711120758120.59718@e6220> <CAPyFy2AkSBaMA1RV2usbkxhdO0BrefvFhNTx0ftd3=PKoR81%2BA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Nov 12, 2017 7:56 AM, "Ed Maste" <emaste@freebsd.org> wrote: On 12 November 2017 at 08:11, Keith White <kwhite@eecs.uottawa.ca> wrote: > I've tried FreeBSD-12.0-CURRENT-arm64-aarch64-RPI3-20171030-r325156.img > and it fails to boot with an error similar to one reported by Mark Millard > in > https://lists.freebsd.org/pipermail/freebsd-arm/2017-September/016643.html > (Synchronous abort after displaying Load Path). Yes, unfortunately boot1.efi has been broken on PRi3 for months. A fix was committed in r325681 and will be in the next round of snapshots https://svnweb.freebsd.org/changeset/base/325681 jhb@ suggested checking the device for NULL as well, as in your patch, and I'll make that change shortly. In practice if the first call works the second will as well, but your version is an improvement in correctness. Makes sense to me. My fix for that just skipped it all if the first path was null. Of course, boot1.edi will be booted from the tree in the near term, so long term correctness doesn't matter. Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfqvhbb_ZFHOC7qejr_dNP%2BAOn45vgpvPS5mP85KGdKWcw>