Date: Sat, 28 Feb 2015 07:31:15 -0500 From: John Baldwin <jhb@freebsd.org> To: "David E. O'Brien" <obrien@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r279381 - head/sys/boot/amd64/efi Message-ID: <1817281.ovAgFISPXy@ralph.baldwin.cx> In-Reply-To: <201502272222.t1RMM6rX042414@svn.freebsd.org> References: <201502272222.t1RMM6rX042414@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday, February 27, 2015 10:22:06 PM David E. O'Brien wrote: > Author: obrien > Date: Fri Feb 27 22:22:05 2015 > New Revision: 279381 > URL: https://svnweb.freebsd.org/changeset/base/279381 > > Log: > Use sys/boot/userboot/libstand/ and not /usr/lib/libstand.a. > > Modified: > head/sys/boot/amd64/efi/Makefile Ugh, this is wrong, please revert. For one, I just ran into a bug where the EFI bits are prone to random corruption because the EFI ABI doesn't use a redzone like the System V amd64 ABI. userboot is a C program though and will use the System V ABI, so it can't share libstand with loader.efi. Locally I have changed /usr/lib/libstand on amd64 to use -mno-red-zone as loader.efi is the only thing on amd64 that uses it (so this avoids mixing the ABIs). Also, if anything this should be using the lib/libstand from OBJDIR as that is what other loader binaries do. -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1817281.ovAgFISPXy>