public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r11-5785] VAX: Fix predicates and constraints for EXTV/EXTZV/INSV insns
@ 2020-12-05 18:29 Maciej W. Rozycki
0 siblings, 0 replies; only message in thread
From: Maciej W. Rozycki @ 2020-12-05 18:29 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:4c293413ca53ffa890e369af33db94264493d140
commit r11-5785-g4c293413ca53ffa890e369af33db94264493d140
Author: Maciej W. Rozycki <macro@linux-mips.org>
Date: Sat Dec 5 18:26:26 2020 +0000
VAX: Fix predicates and constraints for EXTV/EXTZV/INSV insns
It makes no sense for insn operand predicates, as long as they accept a
register operand, to be more restrictive than the set of the associated
constraints, because expand will choose the insn based on the relevant
operand being a pseudo register then and reload keep it happily as a
memory reference if a constraint permits it. So the restriction posed
by such a predicate will be happily ignored, and moreover if a splitter
is added, such as required for MODE_CC support, the new instructions
will reject the original operands supplied, causing an ICE. An actual
example will be given with a subsequent change.
Remove such inconsistencies we have with the EXTV/EXTZV/INSV insns then,
observing that a bit-field located in memory is byte-addressed by the
respective machine instructions and therefore SImode may only be used
with a register or an offsettable memory operand (i.e. not an indexed,
pre-decremented, or post-incremented one), which has already been taken
into account with the constraints currently used, except for `*insv_2'.
The QI machine mode may be used for the bit-field location with any kind
of memory operand, but we got the constraint wrong, although harmlessly
in reality, with `*insv'. Fix that for consistency though.
Also give the insns names, for easier reference here and elsewhere.
gcc/
* config/vax/vax.md (*insv_aligned, *extzv_aligned)
(*extv_aligned, *extv_non_const, *extzv_non_const): Name insns.
Fix location predicate.
(*extzv): Name insn.
(*insv): Likewise. Fix location constraint.
(*insv_2): Likewise, and the predicate.
Diff:
---
gcc/config/vax/vax.md | 32 ++++++++++++++++----------------
1 file changed, 16 insertions(+), 16 deletions(-)
diff --git a/gcc/config/vax/vax.md b/gcc/config/vax/vax.md
index f8e1c2eb02b..de90848a600 100644
--- a/gcc/config/vax/vax.md
+++ b/gcc/config/vax/vax.md
@@ -757,8 +757,8 @@
;; These handle aligned 8-bit and 16-bit fields,
;; which can usually be done with move instructions.
-(define_insn ""
- [(set (zero_extract:SI (match_operand:SI 0 "register_operand" "+ro")
+(define_insn "*insv_aligned"
+ [(set (zero_extract:SI (match_operand:SI 0 "nonimmediate_operand" "+ro")
(match_operand:QI 1 "const_int_operand" "n")
(match_operand:SI 2 "const_int_operand" "n"))
(match_operand:SI 3 "general_operand" "g"))]
@@ -786,9 +786,9 @@
return \"movw %3,%0\";
}")
-(define_insn ""
+(define_insn "*extzv_aligned"
[(set (match_operand:SI 0 "nonimmediate_operand" "=&g")
- (zero_extract:SI (match_operand:SI 1 "register_operand" "ro")
+ (zero_extract:SI (match_operand:SI 1 "nonimmediate_operand" "ro")
(match_operand:QI 2 "const_int_operand" "n")
(match_operand:SI 3 "const_int_operand" "n")))]
"(INTVAL (operands[2]) == 8 || INTVAL (operands[2]) == 16)
@@ -814,9 +814,9 @@
return \"movzwl %1,%0\";
}")
-(define_insn ""
+(define_insn "*extv_aligned"
[(set (match_operand:SI 0 "nonimmediate_operand" "=g")
- (sign_extract:SI (match_operand:SI 1 "register_operand" "ro")
+ (sign_extract:SI (match_operand:SI 1 "nonimmediate_operand" "ro")
(match_operand:QI 2 "const_int_operand" "n")
(match_operand:SI 3 "const_int_operand" "n")))]
"(INTVAL (operands[2]) == 8 || INTVAL (operands[2]) == 16)
@@ -842,7 +842,7 @@
return \"cvtwl %1,%0\";
}")
\f
-;; Register-only SImode cases of bit-field insns.
+;; Register and non-offsettable-memory SImode cases of bit-field insns.
(define_insn ""
[(set (cc0)
@@ -869,9 +869,9 @@
;; by a bicl or sign extension. Because we might end up choosing ext[z]v
;; anyway, we can't allow immediate values for the primary source operand.
-(define_insn ""
+(define_insn "*extv_non_const"
[(set (match_operand:SI 0 "nonimmediate_operand" "=g")
- (sign_extract:SI (match_operand:SI 1 "register_operand" "ro")
+ (sign_extract:SI (match_operand:SI 1 "nonimmediate_operand" "ro")
(match_operand:QI 2 "general_operand" "g")
(match_operand:SI 3 "general_operand" "nrmT")))]
""
@@ -886,9 +886,9 @@
return \"rotl %R3,%1,%0\;cvtwl %0,%0\";
}")
-(define_insn ""
+(define_insn "*extzv_non_const"
[(set (match_operand:SI 0 "nonimmediate_operand" "=g")
- (zero_extract:SI (match_operand:SI 1 "register_operand" "ro")
+ (zero_extract:SI (match_operand:SI 1 "nonimmediate_operand" "ro")
(match_operand:QI 2 "general_operand" "g")
(match_operand:SI 3 "general_operand" "nrmT")))]
""
@@ -962,7 +962,7 @@
""
"")
-(define_insn ""
+(define_insn "*extzv"
[(set (match_operand:SI 0 "nonimmediate_operand" "=g")
(zero_extract:SI (match_operand:QI 1 "memory_operand" "m")
(match_operand:QI 2 "general_operand" "g")
@@ -1015,8 +1015,8 @@
""
"")
-(define_insn ""
- [(set (zero_extract:SI (match_operand:QI 0 "memory_operand" "+g")
+(define_insn "*insv"
+ [(set (zero_extract:SI (match_operand:QI 0 "memory_operand" "+m")
(match_operand:QI 1 "general_operand" "g")
(match_operand:SI 2 "general_operand" "nrmT"))
(match_operand:SI 3 "general_operand" "nrmT"))]
@@ -1046,8 +1046,8 @@
return \"insv %3,%2,%1,%0\";
}")
-(define_insn ""
- [(set (zero_extract:SI (match_operand:SI 0 "register_operand" "+r")
+(define_insn "*insv_2"
+ [(set (zero_extract:SI (match_operand:SI 0 "nonimmediate_operand" "+ro")
(match_operand:QI 1 "general_operand" "g")
(match_operand:SI 2 "general_operand" "nrmT"))
(match_operand:SI 3 "general_operand" "nrmT"))]
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2020-12-05 18:29 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-12-05 18:29 [gcc r11-5785] VAX: Fix predicates and constraints for EXTV/EXTZV/INSV insns Maciej W. Rozycki
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).