https://bugs.kde.org/show_bug.cgi?id=404406
--- Comment #2 from Julian Seward <jsew...@acm.org> ---
(In reply to Andreas Arnez from comment #1)
> Patch #1: Adds the new machine models z14 and z14 ZR1.
> Patch #2: Cleans up s390-check-opcodes.pl, to fix false positives when
> checking s390-opcodes.csv.
> Patch #4: Adds tests for the new instructions. Patch by Ilya Leoshkevich.
These 3 look fine; OK to land.
> Patch #3: Implements the requested instructions. Patch by Ilya Leoshkevich.
One minor and one maybe not so minor thing. First the small thing:
--- a/VEX/priv/guest_s390_toIR.c
+++ b/VEX/priv/guest_s390_toIR.c
+static const HChar *
+s390_irgen_BIC(UChar r1, IRTemp op2addr)
+{
+ IRTemp cond = newTemp(Ity_I32);
+
+ if (r1 == 0) {
+ } else if (r1 == 15) {
I assume the empty { } for the case r1 == 0 is intended, yes? If so
please put " /* nothing */" inside the { } so as to make this clear.
The maybe not so minor thing is:
In host_s390_isel.c, is ISelEnv::hwcaps checked before creating a MG or MGRK
instruction? If it isn't, and the host doesn't have that capability, but by
mistake a MG/MGRK instruction is created, then the system will fail with
SIGILL on the host (I assume). If the check is in place then at least we'll
get an assertion from somewhere inside host_s390_isel.c, saying "I can't
reduce this tree" (select insns for it).
--
You are receiving this mail because:
You are watching all bug changes.
