Mikael Holmen 954fd5590e [IfConversion] Maintain the CFG when predicating/merging blocks in IfConvert*
Summary:
This fixes PR32721 in IfConvertTriangle and possible similar problems in
IfConvertSimple, IfConvertDiamond and IfConvertForkedDiamond.

In PR32721 we had a triangle

   EBB
   | \
   |  |
   | TBB
   |  /
   FBB

where FBB didn't have any successors at all since it ended with an
unconditional return. Then TBB and FBB were be merged into EBB, but EBB
would still keep its successors, and the use of analyzeBranch and
CorrectExtraCFGEdges wouldn't help to remove them since the return
instruction is not analyzable (at least not on ARM).

The edge updating code and branch probability updating code is now pushed
into MergeBlocks() which allows us to share the same update logic between
more callsites. This lets us remove several dependencies on analyzeBranch
and completely eliminate RemoveExtraEdges.

One thing that showed up with this patch was that IfConversion sometimes
left a successor with 0% probability even if there was no branch or
fallthrough to the successor.

One such example from the test case ifcvt_bad_zero_prob_succ.mir. The
indirect branch tBRIND can only jump to bb.1, but without the patch we
got:

  bb.0:
    successors: %bb.1(0x80000000)

  bb.1:
    successors: %bb.1(0x80000000), %bb.2(0x00000000)
    tBRIND %r1, 1, %cpsr
    B %bb.1

  bb.2:

There is no way to jump from bb.1 to bb2, but still there is a 0% edge
from bb.1 to bb.2.

With the patch applied we instead get the expected:

  bb.0:
    successors: %bb.1(0x80000000)

  bb.1:
    successors: %bb.1(0x80000000)
    tBRIND %r1, 1, %cpsr
    B %bb.1

Since bb.2 had no predecessor at all, it was removed.

Several testcases had to be updated due to this since the removed
successor made the "Branch Probability Basic Block Placement" pass
sometimes place blocks in a different order.

Finally added a couple of new test cases:

* PR32721_ifcvt_triangle_unanalyzable.mir:
  Regression test for the original problem dexcribed in PR 32721.

* ifcvt_triangleWoCvtToNextEdge.mir:
  Regression test for problem that caused a revert of my first attempt
  to solve PR 32721.

* ifcvt_simple_bad_zero_prob_succ.mir:
  Test case showing the problem where a wrong successor with 0% probability
  was previously left.

* ifcvt_[diamond|forked_diamond|simple]_unanalyzable.mir
  Very simple test cases for the simple and (forked) diamond cases
  involving unanalyzable branches that can be nice to have as a base if
  wanting to write more complicated tests.

Reviewers: iteratee, MatzeB, grosser, kparzysz

Reviewed By: kparzysz

Subscribers: kbarton, davide, aemerson, nemanjai, javed.absar, kristof.beyls, llvm-commits

Differential Revision: https://reviews.llvm.org/D34099

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@310697 91177308-0d34-0410-b5e6-96231b3b80d8
2017-08-11 06:57:08 +00:00
..
2017-06-19 21:54:25 +00:00
2017-08-03 22:41:12 +00:00
2017-07-21 03:51:01 +00:00

//===---------------------------------------------------------------------===//

Common register allocation / spilling problem:

        mul lr, r4, lr
        str lr, [sp, #+52]
        ldr lr, [r1, #+32]
        sxth r3, r3
        ldr r4, [sp, #+52]
        mla r4, r3, lr, r4

can be:

        mul lr, r4, lr
        mov r4, lr
        str lr, [sp, #+52]
        ldr lr, [r1, #+32]
        sxth r3, r3
        mla r4, r3, lr, r4

and then "merge" mul and mov:

        mul r4, r4, lr
        str r4, [sp, #+52]
        ldr lr, [r1, #+32]
        sxth r3, r3
        mla r4, r3, lr, r4

It also increase the likelihood the store may become dead.

//===---------------------------------------------------------------------===//

bb27 ...
        ...
        %reg1037 = ADDri %reg1039, 1
        %reg1038 = ADDrs %reg1032, %reg1039, %NOREG, 10
    Successors according to CFG: 0x8b03bf0 (#5)

bb76 (0x8b03bf0, LLVM BB @0x8b032d0, ID#5):
    Predecessors according to CFG: 0x8b0c5f0 (#3) 0x8b0a7c0 (#4)
        %reg1039 = PHI %reg1070, mbb<bb76.outer,0x8b0c5f0>, %reg1037, mbb<bb27,0x8b0a7c0>

Note ADDri is not a two-address instruction. However, its result %reg1037 is an
operand of the PHI node in bb76 and its operand %reg1039 is the result of the
PHI node. We should treat it as a two-address code and make sure the ADDri is
scheduled after any node that reads %reg1039.

//===---------------------------------------------------------------------===//

Use local info (i.e. register scavenger) to assign it a free register to allow
reuse:
        ldr r3, [sp, #+4]
        add r3, r3, #3
        ldr r2, [sp, #+8]
        add r2, r2, #2
        ldr r1, [sp, #+4]  <==
        add r1, r1, #1
        ldr r0, [sp, #+4]
        add r0, r0, #2

//===---------------------------------------------------------------------===//

LLVM aggressively lift CSE out of loop. Sometimes this can be negative side-
effects:

R1 = X + 4
R2 = X + 7
R3 = X + 15

loop:
load [i + R1]
...
load [i + R2]
...
load [i + R3]

Suppose there is high register pressure, R1, R2, R3, can be spilled. We need
to implement proper re-materialization to handle this:

R1 = X + 4
R2 = X + 7
R3 = X + 15

loop:
R1 = X + 4  @ re-materialized
load [i + R1]
...
R2 = X + 7 @ re-materialized
load [i + R2]
...
R3 = X + 15 @ re-materialized
load [i + R3]

Furthermore, with re-association, we can enable sharing:

R1 = X + 4
R2 = X + 7
R3 = X + 15

loop:
T = i + X
load [T + 4]
...
load [T + 7]
...
load [T + 15]
//===---------------------------------------------------------------------===//

It's not always a good idea to choose rematerialization over spilling. If all
the load / store instructions would be folded then spilling is cheaper because
it won't require new live intervals / registers. See 2003-05-31-LongShifts for
an example.

//===---------------------------------------------------------------------===//

With a copying garbage collector, derived pointers must not be retained across
collector safe points; the collector could move the objects and invalidate the
derived pointer. This is bad enough in the first place, but safe points can
crop up unpredictably. Consider:

        %array = load { i32, [0 x %obj] }** %array_addr
        %nth_el = getelementptr { i32, [0 x %obj] }* %array, i32 0, i32 %n
        %old = load %obj** %nth_el
        %z = div i64 %x, %y
        store %obj* %new, %obj** %nth_el

If the i64 division is lowered to a libcall, then a safe point will (must)
appear for the call site. If a collection occurs, %array and %nth_el no longer
point into the correct object.

The fix for this is to copy address calculations so that dependent pointers
are never live across safe point boundaries. But the loads cannot be copied
like this if there was an intervening store, so may be hard to get right.

Only a concurrent mutator can trigger a collection at the libcall safe point.
So single-threaded programs do not have this requirement, even with a copying
collector. Still, LLVM optimizations would probably undo a front-end's careful
work.

//===---------------------------------------------------------------------===//

The ocaml frametable structure supports liveness information. It would be good
to support it.

//===---------------------------------------------------------------------===//

The FIXME in ComputeCommonTailLength in BranchFolding.cpp needs to be
revisited. The check is there to work around a misuse of directives in inline
assembly.

//===---------------------------------------------------------------------===//

It would be good to detect collector/target compatibility instead of silently
doing the wrong thing.

//===---------------------------------------------------------------------===//

It would be really nice to be able to write patterns in .td files for copies,
which would eliminate a bunch of explicit predicates on them (e.g. no side 
effects).  Once this is in place, it would be even better to have tblgen 
synthesize the various copy insertion/inspection methods in TargetInstrInfo.

//===---------------------------------------------------------------------===//

Stack coloring improvements:

1. Do proper LiveStackAnalysis on all stack objects including those which are
   not spill slots.
2. Reorder objects to fill in gaps between objects.
   e.g. 4, 1, <gap>, 4, 1, 1, 1, <gap>, 4 => 4, 1, 1, 1, 1, 4, 4

//===---------------------------------------------------------------------===//

The scheduler should be able to sort nearby instructions by their address. For
example, in an expanded memset sequence it's not uncommon to see code like this:

  movl $0, 4(%rdi)
  movl $0, 8(%rdi)
  movl $0, 12(%rdi)
  movl $0, 0(%rdi)

Each of the stores is independent, and the scheduler is currently making an
arbitrary decision about the order.

//===---------------------------------------------------------------------===//

Another opportunitiy in this code is that the $0 could be moved to a register:

  movl $0, 4(%rdi)
  movl $0, 8(%rdi)
  movl $0, 12(%rdi)
  movl $0, 0(%rdi)

This would save substantial code size, especially for longer sequences like
this. It would be easy to have a rule telling isel to avoid matching MOV32mi
if the immediate has more than some fixed number of uses. It's more involved
to teach the register allocator how to do late folding to recover from
excessive register pressure.