Date: Mon, 22 Oct 2012 17:47:37 +0000 (UTC) From: Dimitry Andric <dim@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r241895 - head/contrib/llvm/lib/Target/X86/MCTargetDesc Message-ID: <201210221747.q9MHlbdr060603@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: dim Date: Mon Oct 22 17:47:37 2012 New Revision: 241895 URL: http://svn.freebsd.org/changeset/base/241895 Log: Pull in r165878 from upstream llvm trunk: X86: Disable long nops for all cpus prior to pentiumpro/i686. This is the safest approach for now. If you think long nops matter a lot for performance, compile with -march=i686 or higher. :) MFC after: 3 days Modified: head/contrib/llvm/lib/Target/X86/MCTargetDesc/X86AsmBackend.cpp Modified: head/contrib/llvm/lib/Target/X86/MCTargetDesc/X86AsmBackend.cpp ============================================================================== --- head/contrib/llvm/lib/Target/X86/MCTargetDesc/X86AsmBackend.cpp Mon Oct 22 17:23:32 2012 (r241894) +++ head/contrib/llvm/lib/Target/X86/MCTargetDesc/X86AsmBackend.cpp Mon Oct 22 17:47:37 2012 (r241895) @@ -307,7 +307,9 @@ bool X86AsmBackend::writeNopData(uint64_ }; // This CPU doesnt support long nops. If needed add more. - if (CPU == "geode") { + // FIXME: Can we get this from the subtarget somehow? + if (CPU == "generic" || CPU == "i386" || CPU == "i486" || CPU == "i586" || + CPU == "pentium" || CPU == "pentium-mmx" || CPU == "geode") { for (uint64_t i = 0; i < Count; ++i) OW->Write8(0x90); return true;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201210221747.q9MHlbdr060603>