6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-22 07:32:48 +00:00
Code Issues Projects Releases Wiki Activity
df40801dff
llvm-6502/test/CodeGen/X86/StackColoring.ll

432 lines
13 KiB
LLVM
Raw Normal View History

Evict local live ranges if they can be reassigned. The previous change to local live range allocation also suppressed eviction of local ranges. In rare cases, this could result in more expensive register choices. This commit actually revives a feature that I added long ago: check if live ranges can be reassigned before eviction. But now it only happens in rare cases of evicting a local live range because another local live range wants a cheaper register. The benefit is improved code size for some benchmarks on x86 and armv7. I measured no significant compile time increase and performance changes are noise. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187140 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-25 18:35:19 +00:00
; RUN: llc -mcpu=corei7 -no-stack-coloring=false < %s | FileCheck %s --check-prefix=YESCOLOR
; RUN: llc -mcpu=corei7 -no-stack-coloring=true < %s | FileCheck %s --check-prefix=NOCOLOR
Add a new optimization pass: Stack Coloring, that merges disjoint static allocations (allocas). Allocas are known to be disjoint if they are marked by disjoint lifetime markers (@llvm.lifetime.XXX intrinsics). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163299 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-06 09:17:37 +00:00
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.8.0"
Enable MI Sched for x86. This changes the SelectionDAG scheduling preference to source order. Soon, the SelectionDAG scheduler can be bypassed saving a nice chunk of compile time. Performance differences that result from this change are often a consequence of register coalescing. The register coalescer is far from perfect. Bugs can be filed for deficiencies. On x86 SandyBridge/Haswell, the source order schedule is often preserved, particularly for small blocks. Register pressure is generally improved over the SD scheduler's ILP mode. However, we are still able to handle large blocks that require latency hiding, unlike the SD scheduler's BURR mode. MI scheduler also attempts to discover the critical path in single-block loops and adjust heuristics accordingly. The MI scheduler relies on the new machine model. This is currently unimplemented for AVX, so we may not be generating the best code yet. Unit tests are updated so they don't depend on SD scheduling heuristics. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@192750 91177308-0d34-0410-b5e6-96231b3b80d8
2013-10-15 23:33:07 +00:00
;YESCOLOR: subq $144, %rsp
;NOCOLOR: subq $272, %rsp
Add a new optimization pass: Stack Coloring, that merges disjoint static allocations (allocas). Allocas are known to be disjoint if they are marked by disjoint lifetime markers (@llvm.lifetime.XXX intrinsics). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163299 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-06 09:17:37 +00:00
define i32 @myCall_w2(i32 %in) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b)
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
call void @llvm.lifetime.end(i64 -1, i8* %b)
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
call void @llvm.lifetime.end(i64 -1, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
}
;YESCOLOR: subq $272, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @myCall2_no_merge(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b)
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
br i1 %d, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
call void @llvm.lifetime.end(i64 -1, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
call void @llvm.lifetime.end(i64 -1, i8* %b)
ret i32 %t7
bb3:
call void @llvm.lifetime.end(i64 -1, i8* %b)
ret i32 0
}
;YESCOLOR: subq $144, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @myCall2_w2(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b)
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
call void @llvm.lifetime.end(i64 -1, i8* %b)
br i1 %d, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
call void @llvm.lifetime.end(i64 -1, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
bb3:
ret i32 0
}
Allocate local registers in order for optimal coloring. Also avoid locals evicting locals just because they want a cheaper register. Problem: MI Sched knows exactly how many registers we have and assumes they can be colored. In cases where we have large blocks, usually from unrolled loops, greedy coloring fails. This is a source of "regressions" from the MI Scheduler on x86. I noticed this issue on x86 where we have long chains of two-address defs in the same live range. It's easy to see this in matrix multiplication benchmarks like IRSmk and even the unit test misched-matmul.ll. A fundamental difference between the LLVM register allocator and conventional graph coloring is that in our model a live range can't discover its neighbors, it can only verify its neighbors. That's why we initially went for greedy coloring and added eviction to deal with the hard cases. However, for singly defined and two-address live ranges, we can optimally color without visiting neighbors simply by processing the live ranges in instruction order. Other beneficial side effects: It is much easier to understand and debug regalloc for large blocks when the live ranges are allocated in order. Yes, global allocation is still very confusing, but it's nice to be able to comprehend what happened locally. Heuristics could be added to bias register assignment based on instruction locality (think late register pairing, banks...). Intuituvely this will make some test cases that are on the threshold of register pressure more stable. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187139 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-25 18:35:14 +00:00
;YESCOLOR: subq $200, %rsp
;NOCOLOR: subq $408, %rsp
Add a new optimization pass: Stack Coloring, that merges disjoint static allocations (allocas). Allocas are known to be disjoint if they are marked by disjoint lifetime markers (@llvm.lifetime.XXX intrinsics). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163299 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-06 09:17:37 +00:00
define i32 @myCall_w4(i32 %in) {
entry:
%a1 = alloca [14 x i8*], align 8
%a2 = alloca [13 x i8*], align 8
%a3 = alloca [12 x i8*], align 8
%a4 = alloca [11 x i8*], align 8
%b1 = bitcast [14 x i8*]* %a1 to i8*
%b2 = bitcast [13 x i8*]* %a2 to i8*
%b3 = bitcast [12 x i8*]* %a3 to i8*
%b4 = bitcast [11 x i8*]* %a4 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b4)
call void @llvm.lifetime.start(i64 -1, i8* %b1)
%t1 = call i32 @foo(i32 %in, i8* %b1)
%t2 = call i32 @foo(i32 %in, i8* %b1)
call void @llvm.lifetime.end(i64 -1, i8* %b1)
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t9 = call i32 @foo(i32 %in, i8* %b2)
%t8 = call i32 @foo(i32 %in, i8* %b2)
call void @llvm.lifetime.end(i64 -1, i8* %b2)
call void @llvm.lifetime.start(i64 -1, i8* %b3)
%t3 = call i32 @foo(i32 %in, i8* %b3)
%t4 = call i32 @foo(i32 %in, i8* %b3)
call void @llvm.lifetime.end(i64 -1, i8* %b3)
%t11 = call i32 @foo(i32 %in, i8* %b4)
call void @llvm.lifetime.end(i64 -1, i8* %b4)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
}
;YESCOLOR: subq $112, %rsp
;NOCOLOR: subq $400, %rsp
define i32 @myCall2_w4(i32 %in) {
entry:
%a1 = alloca [14 x i8*], align 8
%a2 = alloca [13 x i8*], align 8
%a3 = alloca [12 x i8*], align 8
%a4 = alloca [11 x i8*], align 8
%b1 = bitcast [14 x i8*]* %a1 to i8*
%b2 = bitcast [13 x i8*]* %a2 to i8*
%b3 = bitcast [12 x i8*]* %a3 to i8*
%b4 = bitcast [11 x i8*]* %a4 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b1)
%t1 = call i32 @foo(i32 %in, i8* %b1)
%t2 = call i32 @foo(i32 %in, i8* %b1)
call void @llvm.lifetime.end(i64 -1, i8* %b1)
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t9 = call i32 @foo(i32 %in, i8* %b2)
%t8 = call i32 @foo(i32 %in, i8* %b2)
call void @llvm.lifetime.end(i64 -1, i8* %b2)
call void @llvm.lifetime.start(i64 -1, i8* %b3)
%t3 = call i32 @foo(i32 %in, i8* %b3)
%t4 = call i32 @foo(i32 %in, i8* %b3)
call void @llvm.lifetime.end(i64 -1, i8* %b3)
br i1 undef, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.start(i64 -1, i8* %b4)
%t11 = call i32 @foo(i32 %in, i8* %b4)
call void @llvm.lifetime.end(i64 -1, i8* %b4)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
bb3:
ret i32 0
}
;YESCOLOR: subq $144, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @myCall2_noend(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b)
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
call void @llvm.lifetime.end(i64 -1, i8* %b)
br i1 %d, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
bb3:
ret i32 0
}
;YESCOLOR: subq $144, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @myCall2_noend2(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %b)
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
br i1 %d, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.end(i64 -1, i8* %b)
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
bb3:
ret i32 0
}
;YESCOLOR: subq $144, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @myCall2_nostart(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
call void @llvm.lifetime.end(i64 -1, i8* %b)
br i1 %d, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
bb3:
ret i32 0
}
; Adopt the test from Transforms/Inline/array_merge.ll'
;YESCOLOR: subq $816, %rsp
;NOCOLOR: subq $1616, %rsp
define void @array_merge() nounwind ssp {
entry:
%A.i1 = alloca [100 x i32], align 4
%B.i2 = alloca [100 x i32], align 4
%A.i = alloca [100 x i32], align 4
%B.i = alloca [100 x i32], align 4
%0 = bitcast [100 x i32]* %A.i to i8*
call void @llvm.lifetime.start(i64 -1, i8* %0) nounwind
%1 = bitcast [100 x i32]* %B.i to i8*
call void @llvm.lifetime.start(i64 -1, i8* %1) nounwind
call void @bar([100 x i32]* %A.i, [100 x i32]* %B.i) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %0) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %1) nounwind
%2 = bitcast [100 x i32]* %A.i1 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %2) nounwind
%3 = bitcast [100 x i32]* %B.i2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %3) nounwind
call void @bar([100 x i32]* %A.i1, [100 x i32]* %B.i2) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %2) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %3) nounwind
ret void
}
Teach the DAGBuilder about lifetime markers which are generated from PHINodes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163494 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-10 08:43:23 +00:00
;YESCOLOR: subq $272, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @func_phi_lifetime(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
call void @llvm.lifetime.end(i64 -1, i8* %b)
br i1 %d, label %bb0, label %bb1
bb0:
%I1 = bitcast [17 x i8*]* %a to i8*
br label %bb2
bb1:
%I2 = bitcast [16 x i8*]* %a2 to i8*
br label %bb2
bb2:
%split = phi i8* [ %I1, %bb0 ], [ %I2, %bb1 ]
call void @llvm.lifetime.start(i64 -1, i8* %split)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
call void @llvm.lifetime.end(i64 -1, i8* %split)
ret i32 %t7
bb3:
ret i32 0
}
Stack Coloring: Add support for multiple regions of the same slot, within a single basic block. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163507 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-10 12:39:35 +00:00
Update to more CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change. All changes were made by the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" grep -q "^; *RUN: *llc.*debug" $NAME && continue grep -q "^; *RUN:.*llvm-objdump" $NAME && continue grep -q "^; *RUN: *opt.*" $NAME && continue TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\([A-Za-z0-9_-]*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC[:]* *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME done This script catches a superset of the cases caught by the script associated with commit r186280. It initially found some false positives due to unusual constructs in a minority of tests; all such cases were disambiguated first in commit r186621. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186624 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-18 22:47:09 +00:00
;YESCOLOR-LABEL: multi_region_bb:
;NOCOLOR-LABEL: multi_region_bb:
Stack Coloring: Add support for multiple regions of the same slot, within a single basic block. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163507 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-10 12:39:35 +00:00
define void @multi_region_bb() nounwind ssp {
entry:
%A.i1 = alloca [100 x i32], align 4
%B.i2 = alloca [100 x i32], align 4
%A.i = alloca [100 x i32], align 4
%B.i = alloca [100 x i32], align 4
%0 = bitcast [100 x i32]* %A.i to i8*
call void @llvm.lifetime.start(i64 -1, i8* %0) nounwind ; <---- start #1
%1 = bitcast [100 x i32]* %B.i to i8*
call void @llvm.lifetime.start(i64 -1, i8* %1) nounwind
call void @bar([100 x i32]* %A.i, [100 x i32]* %B.i) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %0) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %1) nounwind
%2 = bitcast [100 x i32]* %A.i1 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %2) nounwind
%3 = bitcast [100 x i32]* %B.i2 to i8*
call void @llvm.lifetime.start(i64 -1, i8* %3) nounwind
call void @llvm.lifetime.start(i64 -1, i8* %0) nounwind ; <---- start #2
call void @bar([100 x i32]* %A.i1, [100 x i32]* %B.i2) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %2) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %0) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %3) nounwind
ret void
}
Stack Coloring: Handle the case where END markers come before BEGIN markers properly. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163530 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-10 18:51:09 +00:00
;YESCOLOR: subq $272, %rsp
;NOCOLOR: subq $272, %rsp
define i32 @myCall_end_before_begin(i32 %in, i1 %d) {
entry:
%a = alloca [17 x i8*], align 8
%a2 = alloca [16 x i8*], align 8
%b = bitcast [17 x i8*]* %a to i8*
%b2 = bitcast [16 x i8*]* %a2 to i8*
%t1 = call i32 @foo(i32 %in, i8* %b)
%t2 = call i32 @foo(i32 %in, i8* %b)
call void @llvm.lifetime.end(i64 -1, i8* %b)
call void @llvm.lifetime.start(i64 -1, i8* %b)
br i1 %d, label %bb2, label %bb3
bb2:
call void @llvm.lifetime.start(i64 -1, i8* %b2)
%t3 = call i32 @foo(i32 %in, i8* %b2)
%t4 = call i32 @foo(i32 %in, i8* %b2)
%t5 = add i32 %t1, %t2
%t6 = add i32 %t3, %t4
%t7 = add i32 %t5, %t6
ret i32 %t7
bb3:
ret i32 0
}
Fix miscompile due to StackColoring incorrectly merging stack slots (PR15707) IR optimisation passes can result in a basic block that contains: llvm.lifetime.start(%buf) ... llvm.lifetime.end(%buf) ... llvm.lifetime.start(%buf) Before this change, calculateLiveIntervals() was ignoring the second lifetime.start() and was regarding %buf as being dead from the lifetime.end() through to the end of the basic block. This can cause StackColoring to incorrectly merge %buf with another stack slot. Fix by removing the incorrect Starts[pos].isValid() and Finishes[pos].isValid() checks. Just doing: Starts[pos] = Indexes->getMBBStartIdx(MBB); Finishes[pos] = Indexes->getMBBEndIdx(MBB); unconditionally would be enough to fix the bug, but it causes some test failures due to stack slots not being merged when they were before. So, in order to keep the existing tests passing, treat LiveIn and LiveOut separately rather than approximating the live ranges by merging LiveIn and LiveOut. This fixes PR15707. Patch by Mark Seaborn. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181922 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-15 21:15:09 +00:00
; Regression test for PR15707. %buf1 and %buf2 should not be merged
; in this test case.
Update to more CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change. All changes were made by the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" grep -q "^; *RUN: *llc.*debug" $NAME && continue grep -q "^; *RUN:.*llvm-objdump" $NAME && continue grep -q "^; *RUN: *opt.*" $NAME && continue TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\([A-Za-z0-9_-]*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC[:]* *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME done This script catches a superset of the cases caught by the script associated with commit r186280. It initially found some false positives due to unusual constructs in a minority of tests; all such cases were disambiguated first in commit r186621. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186624 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-18 22:47:09 +00:00
;YESCOLOR-LABEL: myCall_pr15707:
Fix miscompile due to StackColoring incorrectly merging stack slots (PR15707) IR optimisation passes can result in a basic block that contains: llvm.lifetime.start(%buf) ... llvm.lifetime.end(%buf) ... llvm.lifetime.start(%buf) Before this change, calculateLiveIntervals() was ignoring the second lifetime.start() and was regarding %buf as being dead from the lifetime.end() through to the end of the basic block. This can cause StackColoring to incorrectly merge %buf with another stack slot. Fix by removing the incorrect Starts[pos].isValid() and Finishes[pos].isValid() checks. Just doing: Starts[pos] = Indexes->getMBBStartIdx(MBB); Finishes[pos] = Indexes->getMBBEndIdx(MBB); unconditionally would be enough to fix the bug, but it causes some test failures due to stack slots not being merged when they were before. So, in order to keep the existing tests passing, treat LiveIn and LiveOut separately rather than approximating the live ranges by merging LiveIn and LiveOut. This fixes PR15707. Patch by Mark Seaborn. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181922 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-15 21:15:09 +00:00
;YESCOLOR: subq $200008, %rsp
Update to more CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change. All changes were made by the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" grep -q "^; *RUN: *llc.*debug" $NAME && continue grep -q "^; *RUN:.*llvm-objdump" $NAME && continue grep -q "^; *RUN: *opt.*" $NAME && continue TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\([A-Za-z0-9_-]*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC[:]* *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME done This script catches a superset of the cases caught by the script associated with commit r186280. It initially found some false positives due to unusual constructs in a minority of tests; all such cases were disambiguated first in commit r186621. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186624 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-18 22:47:09 +00:00
;NOCOLOR-LABEL: myCall_pr15707:
Fix miscompile due to StackColoring incorrectly merging stack slots (PR15707) IR optimisation passes can result in a basic block that contains: llvm.lifetime.start(%buf) ... llvm.lifetime.end(%buf) ... llvm.lifetime.start(%buf) Before this change, calculateLiveIntervals() was ignoring the second lifetime.start() and was regarding %buf as being dead from the lifetime.end() through to the end of the basic block. This can cause StackColoring to incorrectly merge %buf with another stack slot. Fix by removing the incorrect Starts[pos].isValid() and Finishes[pos].isValid() checks. Just doing: Starts[pos] = Indexes->getMBBStartIdx(MBB); Finishes[pos] = Indexes->getMBBEndIdx(MBB); unconditionally would be enough to fix the bug, but it causes some test failures due to stack slots not being merged when they were before. So, in order to keep the existing tests passing, treat LiveIn and LiveOut separately rather than approximating the live ranges by merging LiveIn and LiveOut. This fixes PR15707. Patch by Mark Seaborn. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181922 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-15 21:15:09 +00:00
;NOCOLOR: subq $200008, %rsp
define void @myCall_pr15707() {
%buf1 = alloca i8, i32 100000, align 16
%buf2 = alloca i8, i32 100000, align 16
call void @llvm.lifetime.start(i64 -1, i8* %buf1)
call void @llvm.lifetime.end(i64 -1, i8* %buf1)
call void @llvm.lifetime.start(i64 -1, i8* %buf1)
call void @llvm.lifetime.start(i64 -1, i8* %buf2)
%result1 = call i32 @foo(i32 0, i8* %buf1)
%result2 = call i32 @foo(i32 0, i8* %buf2)
ret void
}
Stack coloring: remove lifetime intervals which contain escaped allocas. The input program may contain intructions which are not inside lifetime markers. This can happen due to a bug in the compiler or due to a bug in user code (for example, returning a reference to a local variable). This commit adds checks that all of the instructions in the function and invalidates lifetime ranges which do not contain all of the instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163678 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-12 04:57:37 +00:00
; Check that we don't assert and crash even when there are allocas
; outside the declared lifetime regions.
Update to more CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change. All changes were made by the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" grep -q "^; *RUN: *llc.*debug" $NAME && continue grep -q "^; *RUN:.*llvm-objdump" $NAME && continue grep -q "^; *RUN: *opt.*" $NAME && continue TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\([A-Za-z0-9_-]*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC[:]* *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME done This script catches a superset of the cases caught by the script associated with commit r186280. It initially found some false positives due to unusual constructs in a minority of tests; all such cases were disambiguated first in commit r186621. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186624 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-18 22:47:09 +00:00
;YESCOLOR-LABEL: bad_range:
;NOCOLOR-LABEL: bad_range:
Stack coloring: remove lifetime intervals which contain escaped allocas. The input program may contain intructions which are not inside lifetime markers. This can happen due to a bug in the compiler or due to a bug in user code (for example, returning a reference to a local variable). This commit adds checks that all of the instructions in the function and invalidates lifetime ranges which do not contain all of the instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163678 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-12 04:57:37 +00:00
define void @bad_range() nounwind ssp {
entry:
%A.i1 = alloca [100 x i32], align 4
%B.i2 = alloca [100 x i32], align 4
%A.i = alloca [100 x i32], align 4
%B.i = alloca [100 x i32], align 4
%0 = bitcast [100 x i32]* %A.i to i8*
call void @llvm.lifetime.start(i64 -1, i8* %0) nounwind
%1 = bitcast [100 x i32]* %B.i to i8*
call void @llvm.lifetime.start(i64 -1, i8* %1) nounwind
call void @bar([100 x i32]* %A.i, [100 x i32]* %B.i) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %0) nounwind
call void @llvm.lifetime.end(i64 -1, i8* %1) nounwind
br label %block2
block2:
; I am used outside the marked lifetime.
call void @bar([100 x i32]* %A.i, [100 x i32]* %B.i) nounwind
ret void
}
Stack Coloring: We have code that checks that all of the uses of allocas are within the lifetime zone. Sometime legitimate usages of allocas are hoisted outside of the lifetime zone. For example, GEPS may calculate the address of a member of an allocated struct. This commit makes sure that we only check (abort regions or assert) for instructions that read and write memory using stack frames directly. Notice that by allowing legitimate usages outside the lifetime zone we also stop checking for instructions which use derivatives of allocas. We will catch less bugs in user code and in the compiler itself. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163791 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-13 12:38:37 +00:00
; Check that we don't assert and crash even when there are usages
; of allocas which do not read or write outside the declared lifetime regions.
Update to more CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change. All changes were made by the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" grep -q "^; *RUN: *llc.*debug" $NAME && continue grep -q "^; *RUN:.*llvm-objdump" $NAME && continue grep -q "^; *RUN: *opt.*" $NAME && continue TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\([A-Za-z0-9_-]*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC[:]* *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME done This script catches a superset of the cases caught by the script associated with commit r186280. It initially found some false positives due to unusual constructs in a minority of tests; all such cases were disambiguated first in commit r186621. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186624 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-18 22:47:09 +00:00
;YESCOLOR-LABEL: shady_range:
;NOCOLOR-LABEL: shady_range:
Stack Coloring: We have code that checks that all of the uses of allocas are within the lifetime zone. Sometime legitimate usages of allocas are hoisted outside of the lifetime zone. For example, GEPS may calculate the address of a member of an allocated struct. This commit makes sure that we only check (abort regions or assert) for instructions that read and write memory using stack frames directly. Notice that by allowing legitimate usages outside the lifetime zone we also stop checking for instructions which use derivatives of allocas. We will catch less bugs in user code and in the compiler itself. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163791 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-13 12:38:37 +00:00
%struct.Klass = type { i32, i32 }
define i32 @shady_range(i32 %argc, i8** nocapture %argv) uwtable {
%a.i = alloca [4 x %struct.Klass], align 16
%b.i = alloca [4 x %struct.Klass], align 16
%a8 = bitcast [4 x %struct.Klass]* %a.i to i8*
%b8 = bitcast [4 x %struct.Klass]* %b.i to i8*
; I am used outside the lifetime zone below:
%z2 = getelementptr inbounds [4 x %struct.Klass]* %a.i, i64 0, i64 0, i32 0
call void @llvm.lifetime.start(i64 -1, i8* %a8)
call void @llvm.lifetime.start(i64 -1, i8* %b8)
%z3 = load i32* %z2, align 16
%r = call i32 @foo(i32 %z3, i8* %a8)
%r2 = call i32 @foo(i32 %z3, i8* %b8)
call void @llvm.lifetime.end(i64 -1, i8* %a8)
call void @llvm.lifetime.end(i64 -1, i8* %b8)
ret i32 9
}
Add a new optimization pass: Stack Coloring, that merges disjoint static allocations (allocas). Allocas are known to be disjoint if they are marked by disjoint lifetime markers (@llvm.lifetime.XXX intrinsics). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163299 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-06 09:17:37 +00:00
declare void @bar([100 x i32]* , [100 x i32]*) nounwind
declare void @llvm.lifetime.start(i64, i8* nocapture) nounwind
declare void @llvm.lifetime.end(i64, i8* nocapture) nounwind
Stack coloring: remove lifetime intervals which contain escaped allocas. The input program may contain intructions which are not inside lifetime markers. This can happen due to a bug in the compiler or due to a bug in user code (for example, returning a reference to a local variable). This commit adds checks that all of the instructions in the function and invalidates lifetime ranges which do not contain all of the instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163678 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-12 04:57:37 +00:00
declare i32 @foo(i32, i8*)
Reference in New Issue Copy Permalink