The recent problem with gcc-8 and RTAI kernel modules has led to some
interesting discussion on a Stack Overflow question:

https://stackoverflow.com/questions/60589636/what-can-cause-c-code-to-crash-when-an-array-is-initialised-at-declaration-but

Towards the bottom is a link to what compiler options are set if
compiling a kernel module with the recommended defaults:
https://pastebin.com/YB7spBCy

This is significantly more that are set in our Makefile. And includes
at least one ( -mno-sse ) which appears to cause a compile failure in
LinuxCNC:

cc1: warning: SSE instruction set disabled, using 387 arithmetics
In file included from hal/components/bldc.comp:394:
/home/andypugh/linuxcnc-dev/src/../include/rtapi_math.h: In function ‘atan’:
/home/andypugh/linuxcnc-dev/src/../include/rtapi_math.h:60:1: error:
SSE register return with SSE disabled
 extern __inline double atan (double __y) {

It seems that we might need to make some decisions about whether we
abandon kernel-mode realtime or put some effort in to future-proofing
the build system.

(And as we seem to be unable to make a decision to release 2.8 I am not hopeful)

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1912


_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to