Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 19 May 2014 16:57:17 +0400
From:      Boris Samorodov <bsam@passap.ru>
To:        Ian Lepore <ian@FreeBSD.org>
Cc:        freebsd-current@FreeBSD.org
Subject:   Re: [arm cross-compiling, clang] Error: selected processor does not support `ldrexd r2,r3,[r1]'
Message-ID:  <5379FFAD.3060801@passap.ru>
In-Reply-To: <5379F3BF.9090000@passap.ru>
References:  <5378D64C.2000903@passap.ru> <5378DB4E.5030305@passap.ru>	 <1400431085.1152.24.camel@revolution.hippie.lan>	 <5379134B.9000502@passap.ru> <1400448314.1152.26.camel@revolution.hippie.lan> <5379F3BF.9090000@passap.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
19.05.2014 16:06, Boris Samorodov пишет:
> 19.05.2014 01:25, Ian Lepore пишет:
>> On Mon, 2014-05-19 at 00:08 +0400, Boris Samorodov wrote:
> 
>>> It's definitely not my day -- crochet build failed with:
>>> -----
>>> --- all_subdir_libllvmarmcodegen ---
>>> /usr/src/lib/clang/libllvmarmcodegen/../../../contrib/llvm/lib/Target/ARM/ARMBaseInstrInfo.cpp:3687:15:
>>> error: no member named
>>>  'VLD1d64TPseudoWB_fixed' in namespace 'llvm::ARM'; did you mean
>>> 'VST1d64TPseudoWB_fixed'?
>>>     case ARM::VLD1d64TPseudoWB_fixed:
>>>          ~~~~~^~~~~~~~~~~~~~~~~~~~~~
>>>               VST1d64TPseudoWB_fixed
>>> ./ARMGenInstrInfo.inc.h:1969:5: note: 'VST1d64TPseudoWB_fixed' declared here
>>>     VST1d64TPseudoWB_fixed      = 1953,
>>>     ^
>>> /usr/src/lib/clang/libllvmarmcodegen/../../../contrib/llvm/lib/Target/ARM/ARMBaseInstrInfo.cpp:3704:15:
>>> error: no member named
>>>  'VLD1d64QPseudoWB_fixed' in namespace 'llvm::ARM'; did you mean
>>> 'VST1d64QPseudoWB_fixed'?
>>>     case ARM::VLD1d64QPseudoWB_fixed:
>>>          ~~~~~^~~~~~~~~~~~~~~~~~~~~~
>>>               VST1d64QPseudoWB_fixed
>>> ./ARMGenInstrInfo.inc.h:1963:5: note: 'VST1d64QPseudoWB_fixed' declared here
>>>     VST1d64QPseudoWB_fixed      = 1947,
>>> -----
>>
>> I've seen others report this error recently, and it was caused by an
>> update to clang.  There's a dependency glitch so that some header files
>> don't get regenerated correctly; I think that has been fixed, but to get
>> the fix in place you have to clean out obj/arm.armv6 and build fresh.
> 
> Ian, thanks -- that helped!

Just a note: (crochet) buildworld finished successfully. However
buildkernel fails at the very beginning with "Malformed conditional
(${MK_ARM_EABI} != "no")" -- just as Michael Tuexen reported at arm@.

-- 
WBR, Boris Samorodov (bsam)
FreeBSD Committer, http://www.FreeBSD.org The Power To Serve



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5379FFAD.3060801>