mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2026-01-22 05:22:57 +00:00
specified in the same file that the library itself is created. This is more idiomatic for CMake builds, and also allows us to correctly specify dependencies that are missed due to bugs in the GenLibDeps perl script, or change from compiler to compiler. On Linux, this returns CMake to a place where it can relably rebuild several targets of LLVM. I have tried not to change the dependencies from the ones in the current auto-generated file. The only places I've really diverged are in places where I was seeing link failures, and added a dependency. The goal of this patch is not to start changing the dependencies, merely to move them into the correct location, and an explicit form that we can control and change when necessary. This also removes a serialization point in the build because we don't have to scan all the libraries before we begin building various tools. We no longer have a step of the build that regenerates a file inside the source tree. A few other associated cleanups fall out of this. This isn't really finished yet though. After talking to dgregor he urged switching to a single CMake macro to construct libraries with both sources and dependencies in the arguments. Migrating from the two macros to that style will be a follow-up patch. Also, llvm-config is still generated with GenLibDeps.pl, which means it still has slightly buggy dependencies. The internal CMake 'llvm-config-like' macro uses the correct explicitly specified dependencies however. A future patch will switch llvm-config generation (when using CMake) to be based on these deps as well. This may well break Windows. I'm getting a machine set up now to dig into any failures there. If anyone can chime in with problems they see or ideas of how to solve them for Windows, much appreciated. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@136433 91177308-0d34-0410-b5e6-96231b3b80d8
//===---------------------------------------------------------------------===//
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 lr, [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.