[blfs-support] Internal compiler error ? SOLVED

Dr.-Ing. Edgar Alwers edgaralwers at gmx.de
Thu Mar 3 08:20:26 PST 2016

> On Tuesday 02 February 2016 23:14:16 Michael Shell wrote:
> > : cat /proc/cpuinfo
> > 
> > In this way, Edgar can compare the CPUs of his two machines. I bet they
> > differ with regard to bmi2 capability

Following Michael Shell's E-Mail :
The offender turned out to be the mulx instruction:
I changed in the gmp-6.1.0 source "config.guess" around line 815 to
modelstr = " coreisbr";

and in
"mpn/x86_64/fat/fat.c"  around line 293 I deleted the statement 
"if ((dummy_string[0 + 8 / 8] & (1 << (8 % 8))) != 0)
                CPUVEC_SETUP_coreihwl; "

With this changes, the image of the PC could be moved to the laptop, and the laptop compiler 

It is remarkable that the problem with the freezing ./configure ( my E-Mail 1.2.2016 ) was also 
related to gmp and also solved. Furthermore I am working with gmp-6.1.0 and not with 6.0.0a. 
This means, that the problem has not been solved yet in a general form.

The problem is important: I need the posibility of changing the installed system on the laptop, 
e.g to introduce wireless connections which I do not need on my PC, which is connected by 

I intend to create a short E-Mail with a different subject  " porting LFS/BLFS to another machine", 
as the actual subject does not reflect the real problem of porting systems.

Again, thank you very much  Bruce, Ken and Michael for the help


Dr.-Ing. Edgar Alwers <edgaralwers at gmx.de>

[1] https://www.mail-archive.com/blfs-support@lists.linuxfromscratch.org/msg02228.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfromscratch.org/pipermail/blfs-support/attachments/20160303/a738b391/attachment.html>

More information about the blfs-support mailing list