If a global variable has a specified alignment that is less than the preferred

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

diff --git a/lib/Target/README.txt b/lib/Target/README.txt
index bda764deffe09bacae4f8ab3250fe7bdde485963..99616b47c841401f05ee90aa18dfa068a2794228 100644 (file)
--- a/lib/Target/README.txt
+++ b/lib/Target/README.txt
@@ -254,6 +254,20 @@ unsigned long reverse(unsigned v) {
 
 //===---------------------------------------------------------------------===//
 
+[LOOP DELETION]
+
+We don't delete this output free loop, because trip count analysis doesn't
+realize that it is finite (if it were infinite, it would be undefined).  Not
+having this blocks Loop Idiom from matching strlen and friends.  
+
+void foo(char *C) {
+  int x = 0;
+  while (*C)
+    ++x,++C;
+}
+
+//===---------------------------------------------------------------------===//
+
 [LOOP RECOGNITION]
 
 These idioms should be recognized as popcount (see PR1488):
@@ -378,34 +392,6 @@ PHI Slicing could be extended to do this.
 
 //===---------------------------------------------------------------------===//
 
-LSR should know what GPR types a target has from TargetData.  This code:
-
-volatile short X, Y; // globals
-
-void foo(int N) {
-  int i;
-  for (i = 0; i < N; i++) { X = i; Y = i*4; }
-}
-
-produces two near identical IV's (after promotion) on PPC/ARM:
-
-LBB1_2:
-       ldr r3, LCPI1_0
-       ldr r3, [r3]
-       strh r2, [r3]
-       ldr r3, LCPI1_1
-       ldr r3, [r3]
-       strh r1, [r3]
-       add r1, r1, #4
-       add r2, r2, #1   <- [0,+,1]
-       sub r0, r0, #1   <- [0,-,1]
-       cmp r0, #0
-       bne LBB1_2
-
-LSR should reuse the "+" IV for the exit test.
-
-//===---------------------------------------------------------------------===//
-
 Tail call elim should be more aggressive, checking to see if the call is
 followed by an uncond branch to an exit block.
 
@@ -1311,6 +1297,21 @@ codegen.
 
 //===---------------------------------------------------------------------===//
 
+simplifylibcalls should turn these snprintf idioms into memcpy (GCC PR47917)
+
+char buf1[6], buf2[6], buf3[4], buf4[4];
+int i;
+
+int foo (void) {
+  int ret = snprintf (buf1, sizeof buf1, "abcde");
+  ret += snprintf (buf2, sizeof buf2, "abcdef") * 16;
+  ret += snprintf (buf3, sizeof buf3, "%s", i++ < 6 ? "abc" : "def") * 256;
+  ret += snprintf (buf4, sizeof buf4, "%s", i++ > 10 ? "abcde" : "defgh")*4096;
+  return ret;
+}
+
+//===---------------------------------------------------------------------===//
+
 "gas" uses this idiom:
   else if (strchr ("+-/*%|&^:[]()~", *intel_parser.op_string))
 ..
@@ -1766,43 +1767,6 @@ case it choses instead to keep the max operation obvious.
 
 //===---------------------------------------------------------------------===//
 
-Take the following testcase on x86-64 (similar testcases exist for all targets
-with addc/adde):
-
-define void @a(i64* nocapture %s, i64* nocapture %t, i64 %a, i64 %b,
-i64 %c) nounwind {
-entry:
- %0 = zext i64 %a to i128                        ; <i128> [#uses=1]
- %1 = zext i64 %b to i128                        ; <i128> [#uses=1]
- %2 = add i128 %1, %0                            ; <i128> [#uses=2]
- %3 = zext i64 %c to i128                        ; <i128> [#uses=1]
- %4 = shl i128 %3, 64                            ; <i128> [#uses=1]
- %5 = add i128 %4, %2                            ; <i128> [#uses=1]
- %6 = lshr i128 %5, 64                           ; <i128> [#uses=1]
- %7 = trunc i128 %6 to i64                       ; <i64> [#uses=1]
- store i64 %7, i64* %s, align 8
- %8 = trunc i128 %2 to i64                       ; <i64> [#uses=1]
- store i64 %8, i64* %t, align 8
- ret void
-}
-
-Generated code:
-        addq   %rcx, %rdx
-        sbbq   %rax, %rax
-        subq   %rax, %r8
-        movq   %r8, (%rdi)
-        movq   %rdx, (%rsi)
-        ret
-
-Expected code:
-       addq    %rcx, %rdx
-       adcq    $0, %r8
-       movq    %r8, (%rdi)
-       movq    %rdx, (%rsi)
-       ret
-
-//===---------------------------------------------------------------------===//
-
 Switch lowering generates less than ideal code for the following switch:
 define void @a(i32 %x) nounwind {
 entry:
@@ -2110,11 +2074,12 @@ for.end:                                          ; preds = %entry
 }
 
 This shouldn't need the ((zext (%n - 1)) + 1) game, and it should ideally fold
-the two memset's together. The issue with %n seems to stem from poor handling
-of the original loop.
+the two memset's together.
 
-To simplify this, we need SCEV to know that "n != 0" because of the dominating
-conditional.  That would turn the second memset into a simple memset of 'n'.
+The issue with the addition only occurs in 64-bit mode, and appears to be at
+least partially caused by Scalar Evolution not keeping its cache updated: it
+returns the "wrong" result immediately after indvars runs, but figures out the
+expected result if it is run from scratch on IR resulting from running indvars.
 
 //===---------------------------------------------------------------------===//
 
@@ -2273,4 +2238,23 @@ missed cases:
 
 //===---------------------------------------------------------------------===//
 
+define i1 @test1(i32 %x) nounwind {
+  %and = and i32 %x, 3
+  %cmp = icmp ult i32 %and, 2
+  ret i1 %cmp
+}
+
+Can be folded to (x & 2) == 0.
+
+define i1 @test2(i32 %x) nounwind {
+  %and = and i32 %x, 3
+  %cmp = icmp ugt i32 %and, 1
+  ret i1 %cmp
+}
+
+Can be folded to (x & 2) != 0.
 
+SimplifyDemandedBits shrinks the "and" constant to 2 but instcombine misses the
+icmp transform.
+
+//===---------------------------------------------------------------------===//