Fail to reproduce the safe relaxations testing of X86 architecture

[DehengYang]

Dear Authors,

Sorry for interrupting you here. I encountered an issue when I was trying to reproduce the safe relaxations testing of X86 architecture based on the "diy Seven tutorial" Section 7.2 (Page 45). I performed the following steps:

1. create a conf file named x86-rfi.conf and write the following configurations:

#rfi x86 conf file 
-arch X86 
-nprocs 4 
-size 6 
-name rfi 
-safe PosR* PodR* PodWW PosWW Rfe Wse Fre FencesWR FencedWR 
-relax Rfi

2. run diy7 -conf x86-rfi.conf

3. run litmus7 -mach saumur @all > saumur.rfi.00

4. run readRelax7 saumur.rfi.00

However, the output in the 4th step was:

** Relaxations **
** Non-Relaxations **
rfi000 Rfi With Safe Rfe Fre PodWW PodRR
rfi001 Rfi With Safe Fre MFencedWR
rfi002 Rfi With Safe Fre Coe PodWW MFencedWR
rfi003 Rfi With Safe Fre PodWW
rfi004 Rfi With Safe Rfe Fre PodWW PodRR
rfi005 Rfi With Safe Rfe Fre PosRW PodWW PodRR
rfi006 Rfi With Safe Fre PodWW MFencedWR
rfi007 Rfi With Safe Fre PosRR PodWW MFencedWR
rfi008 Rfi With Safe Fre Coe PodWW
rfi009 Rfi With Safe Fre Coe PosRR PodWW
rfi010 Rfi With Safe Rfe Fre Coe PodWW PodRW
rfi011 Rfi With Safe Rfe Coe PosRW PodWW PodRW
rfi012 Rfi With Safe Rfe Fre PodRW MFencedWR
rfi013 Rfi With Safe Rfe Fre Coe PodWW PodRW
rfi014 Rfi With Safe Rfe Fre PodWW PodRW
rfi015 Rfi With Safe Rfe Fre PosRR PodWW PodRW
rfi016 Rfi With Safe Rfe Fre PodRW
rfi017 Rfi With Safe Rfe PosRW PodRW
** Relaxation summary **

If I do not misunderstand, the above output indicates that Rfi is not allowed to be relaxed, which is as opposed to the result given in the "diy Seven tutorial" Section 7.2 (Page 45).

Thus, it would be sincerely appreciated if any help could be provided for this issue. Thank you!

Thank you in advance for your time and help.

Yours sincerely,
Deheng

[maranget]

Dear @DehengYang. This not normal. The default behaviour of diy7 has changed. As a consequence, the tool no longer generate the tests that would show Rfi to be relaxed. Moreover the new behaviour is not intended and contradicts the new documentation. In other words, this is a bug, which will be corrected soon --- see PR #552.

Thanks for your report.

Luc

[DehengYang]

Get it. Thank you very much for the kind help and great effort!