[blfs-support] Internal compiler error ?

Michael Shell list1 at michaelshell.org
Tue Feb 2 20:14:16 PST 2016


On Mon, 1 Feb 2016 20:05:11 +0000
Ken Moffat <zarniwhoop at ntlworld.com> wrote:

> The only hardware-specific thing I recall "recently" was gmp - since
> September 2015 I have been using the configfsf.{guess,sub} over the
> config. versions.  Discussed on the lfs-dev list back then.


Yeah, I sure do remember that ordeal as I participated in it:

https://www.mail-archive.com/blfs-support@lists.linuxfromscratch.org/msg02166.html

and as the thread went on we finally tracked down the specific
illegal instruction.

The offender turned out to be the mulx instruction:

https://www.mail-archive.com/blfs-support@lists.linuxfromscratch.org/msg02228.html

and therein is a patch for the gmp source to lockout the use of bmi2
instruction set (of which mulx is one). 

Also, note from my post a way to check to see if a CPU supports
the bmi2 instruction set:

: cat /proc/cpuinfo
:
: To support mulx, the CPU must list bmi2 in the cpuinfo flags section:
:
: http://unix.stackexchange.com/questions/43539/what-do-the-flags-in-proc-cpuinfo-mean
:
: (see under "Intel-defined CPU features, CPUID level 0x00000007:0 (ebx)")


In this way, Edgar can compare the CPUs of his two machines. I bet they
differ with regard to bmi2 capability.


  Cheers,

  Mike Shell




More information about the blfs-support mailing list