Remove an incorrect overaggressive optimization

[oota-llvm.git] / lib / Target / README.txt

diff --git a/lib/Target/README.txt b/lib/Target/README.txt
index dce29b82755ac631c507f4421090b6ec354a4fd3..aad621f440ac7c3f424b371eacbbf39539ac01b5 100644 (file)
--- a/lib/Target/README.txt
+++ b/lib/Target/README.txt
@@ -339,6 +339,8 @@ we don't have whole-function selection dags.  On x86, this means we use one
 extra register for the function when effective_addr2 is declared as U64 than
 when it is declared U32.
 
+PHI Slicing could be extended to do this.
+
 //===---------------------------------------------------------------------===//
 
 LSR should know what GPR types a target has.  This code:
@@ -1612,12 +1614,6 @@ int int_char(char m) {if(m>7) return 0; return m;}
 
 //===---------------------------------------------------------------------===//
 
-IPSCCP is propagating elements of first class aggregates, but is not propagating
-the entire aggregate itself.  This leads it to miss opportunities, for example
-in test/Transforms/SCCP/ipsccp-basic.ll:test5b.
-
-//===---------------------------------------------------------------------===//
-
 int func(int a, int b) { if (a & 0x80) b |= 0x80; else b &= ~0x80; return b; }
 
 Generates this:
@@ -1686,3 +1682,55 @@ entry:
 }
 
 //===---------------------------------------------------------------------===//
+
+IPSCCP does not currently propagate argument dependent constants through
+functions where it does not not all of the callers.  This includes functions
+with normal external linkage as well as templates, C99 inline functions etc.
+Specifically, it does nothing to:
+
+define i32 @test(i32 %x, i32 %y, i32 %z) nounwind {
+entry:
+  %0 = add nsw i32 %y, %z                         
+  %1 = mul i32 %0, %x                             
+  %2 = mul i32 %y, %z                             
+  %3 = add nsw i32 %1, %2                         
+  ret i32 %3
+}
+
+define i32 @test2() nounwind {
+entry:
+  %0 = call i32 @test(i32 1, i32 2, i32 4) nounwind
+  ret i32 %0
+}
+
+It would be interesting extend IPSCCP to be able to handle simple cases like
+this, where all of the arguments to a call are constant.  Because IPSCCP runs
+before inlining, trivial templates and inline functions are not yet inlined.
+The results for a function + set of constant arguments should be memoized in a
+map.
+
+//===---------------------------------------------------------------------===//
+
+The libcall constant folding stuff should be moved out of SimplifyLibcalls into
+libanalysis' constantfolding logic.  This would allow IPSCCP to be able to
+handle simple things like this:
+
+static int foo(const char *X) { return strlen(X); }
+int bar() { return foo("abcd"); }
+
+//===---------------------------------------------------------------------===//
+
+InstCombine should use SimplifyDemandedBits to remove the or instruction:
+
+define i1 @test(i8 %x, i8 %y) {
+  %A = or i8 %x, 1
+  %B = icmp ugt i8 %A, 3
+  ret i1 %B
+}
+
+Currently instcombine calls SimplifyDemandedBits with either all bits or just
+the sign bit, if the comparison is obviously a sign test. In this case, we only
+need all but the bottom two bits from %A, and if we gave that mask to SDB it
+would delete the or instruction for us.
+
+//===---------------------------------------------------------------------===//