On Fri, 27 Nov 2020, Maciej W. Rozycki wrote:
> > That said, if it fixes the test suite errors you're seeing, it would
> > probably be OK to go with just this minimal change -- unless we can
> > just move to LRA as mentioned above.
>
> I've looked through the test results and indeed these suspicious ICEs
> remain:
>
> .../gcc/testsuite/gcc.dg/pr83623.c:13:1: internal compiler error: in
> change_address_1, at emit-rtl.c:2275
> .../gcc/testsuite/gcc.dg/torture/vshuf-main.inc:27:1: internal compiler
> error: in change_address_1, at emit-rtl.c:2275
I've double-checked these and this:
> corresponding to:
>
> FAIL: gcc.dg/pr83623.c (internal compiler error)
> FAIL: gcc.dg/pr83623.c (test for excess errors)
comes from this insn:
(insn 17 14 145 (set (reg:SI 1 %r1)
(zero_extract:SI (mem/c:SI (symbol_ref:SI ("x") <var_decl
0x7ffff7f80120 x>) [1 x+0 S4 A128])
(const_int 16 [0x10])
(const_int 16 [0x10]))) ".../gcc/testsuite/gcc.dg/pr83623.c":12:9
101 {*vax.md:805}
(nil))
and this:
> FAIL: gcc.dg/torture/vshuf-v16qi.c -O2 (internal compiler error)
> FAIL: gcc.dg/torture/vshuf-v16qi.c -O2 (test for excess errors)
likewise:
(insn 83 82 84 (set (reg:SI 5 %r5 [84])
(zero_extract:SI (mem/c:SI (symbol_ref:SI ("b") <var_decl
0x7ffff7f801b0 b>) [0 b+0 S4 A128])
(const_int 8 [0x8])
(const_int 16 [0x10])))
".../gcc/testsuite/gcc.dg/torture/vshuf-main.inc":28:1 101 {*vax.md:805}
(nil))
So these are not related (and addressed with 22/31 BTW).
I'll make the "|| strict_low" update then and push this change along with
the rest once all the verification has completed, presumably this coming
Monday.
Thanks for your review!
Maciej
