[PATCH] Add infrastructure for support of multiple memory constraints.
hfinkel at anl.gov
hfinkel at anl.gov
Mon Mar 9 14:13:37 PDT 2015
Thanks for working on this!
================
Comment at: include/llvm/Target/TargetLowering.h:2597
@@ -2596,1 +2596,3 @@
+ virtual unsigned getInlineAsmMemConstraint(const std::string ConstraintCode) const {
+ // FIXME: Actually look at the constraint code.
----------------
Line is too long.
================
Comment at: include/llvm/Target/TargetLowering.h:2599
@@ +2598,3 @@
+ // FIXME: Actually look at the constraint code.
+ // This currently matches the existing behaviour to keep this change
+ // non-functional.
----------------
Once this lands, references to "this change" don't make sense. I'd say,
// Once all targets have been updated, we can use the constraint code.
================
Comment at: lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp:6598
@@ +6597,3 @@
+ if (ConstraintID == InlineAsm::Constraint_Unknown)
+ llvm_unreachable("Failed to convert memory constraint code to constraint id.");
+
----------------
Line too long. Also, why not write?
assert(ConstraintID != InlineAsm::Constraint_Unknown &&
"Failed to convert memory constraint code to constraint id.");
(same below too)
================
Comment at: lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp:6749
@@ +6748,3 @@
+ if (ConstraintID == InlineAsm::Constraint_Unknown)
+ llvm_unreachable("Failed to convert memory constraint code to constraint id.");
+
----------------
Line is too long.
================
Comment at: lib/Target/AArch64/AArch64ISelDAGToDAG.cpp:215
@@ -216,1 +214,3 @@
+ const SDValue &Op, unsigned ConstraintID, std::vector<SDValue> &OutOps) {
+ assert(ConstraintID == InlineAsm::Constraint_m && "unexpected asm memory constraint");
// Require the address to be in a register. That is safe for all AArch64
----------------
Line too long. (same with this assert in some of the other files too).
http://reviews.llvm.org/D8171
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
More information about the llvm-commits
mailing list