aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen
AgeCommit message (Collapse)Author
2012-05-10misched: Introducing Top and Bottom register pressure trackers during ↵Andrew Trick
scheduling. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156571 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-10RegPressure: API for speculatively checking instruction pressure.Andrew Trick
Added getMaxExcessUpward/DownwardPressure. They somewhat abuse the tracker by speculatively handling an instruction out of order. But it is convenient for now. In the future, we will cache each instruction's pressure contribution to make this efficient. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156561 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-10RegPressure: fix array index iteration style.Andrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156560 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-10Revert: 156550 "ARM: peephole optimization to remove cmp instruction"Manman Ren
This commit broke an external linux bot and gave a compile-time warning. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156556 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-10ARM: peephole optimization to remove cmp instructionManman Ren
This patch will optimize the following cases: sub r1, r3 | sub r1, imm cmp r3, r1 or cmp r1, r3 | cmp r1, imm bge L1 TO subs r1, r3 bge L1 or ble L1 If the branch instruction can use flag from "sub", then we can replace "sub" with "subs" and eliminate the "cmp" instruction. rdar: 10734411 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156550 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-08Fix thinko in conditional.Eric Christopher
Part of rdar://11352000 and should bring the buildbots back. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156421 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-08DAGCombiner should not change the type of an extract_vector index.Jim Grosbach
When a combine twiddles an extract_vector, care should be take to preserve the type of the index operand. No luck extracting a reasonable testcase, unfortunately. rdar://11391009 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156419 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-08Formatting fixes.Akira Hatanaka
Patch by Jack Carter. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156409 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-08Handle OpDeref in case it comes in as a register operand.Eric Christopher
Part of rdar://11352000 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156405 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-08Extract methods for joining physregs.Jakob Stoklund Olesen
No functional change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156345 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07Naming convention and whitespace. No functional change.Jakob Stoklund Olesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156342 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07Coalesce subreg-subreg copies.Jakob Stoklund Olesen
At least some of them: %vreg1:sub_16bit = COPY %vreg2:sub_16bit; GR64:%vreg1, GR32: %vreg2 Previously, we couldn't figure out that the above copy could be eliminated by coalescing %vreg2 with %vreg1:sub_32bit. The new getCommonSuperRegClass() hook makes it possible. This is not very useful yet since the unmodified part of the destination register usually interferes with the source register. The coalescer needs to understand sub-register interference checking first. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156334 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07Add an MF argument to TRI::getPointerRegClass() and TII::getRegClass().Jakob Stoklund Olesen
The getPointerRegClass() hook can return register classes that depend on the calling convention of the current function (ptr_rc_tailcall). So far, we have been able to infer the calling convention from the subtarget alone, but as we add support for multiple calling conventions per target, that no longer works. Patch by Yiannis Tsiouris! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156328 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07Teach DAG combine to fold x-x to 0.0 when unsafe FP math is enabled.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156324 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-05Add a new target hook "predictableSelectIsExpensive".Benjamin Kramer
This will be used to determine whether it's profitable to turn a select into a branch when the branch is likely to be predicted. Currently enabled for everything but Atom on X86 and Cortex-A9 devices on ARM. I'm not entirely happy with the name of this flag, suggestions welcome ;) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156233 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04Make sure findRepresentativeClass picks the widest super-register.Jakob Stoklund Olesen
We want the representative register class to contain the largest super-registers available. This makes the function less sensitive to the register class numbering. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156220 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04Remove extra comma in debug output.Jakob Stoklund Olesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156219 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04Use SuperRegClassIterator for findRepresentativeClass().Jakob Stoklund Olesen
The masks returned by SuperRegClassIterator are computed automatically by TableGen. This is better than depending on the manually specified SuperRegClasses. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156147 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-03Fix two-address pass's aggressive instruction commuting heuristics. It's meantEvan Cheng
to catch cases like: %reg1024<def> = MOV r1 %reg1025<def> = MOV r0 %reg1026<def> = ADD %reg1024, %reg1025 r0 = MOV %reg1026 By commuting ADD, it let coalescer eliminate all of the copies. However, there was a bug in the heuristics where it ended up commuting the ADD in: %reg1024<def> = MOV r0 %reg1025<def> = MOV 0 %reg1026<def> = ADD %reg1024, %reg1025 r0 = MOV %reg1026 That did no benefit but rather ensure the last MOV would not be coalesced. rdar://11355268 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156048 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-03Added TargetRegisterInfo::getAllocatableClass.Andrew Trick
The ensures that virtual registers always belong to an allocatable class. If your target attempts to create a vreg for an operand that has no allocatable register subclass, you will crash quickly. This ensures that targets define register classes as intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156046 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-02Teach DAGCombine the same multiply-by-1.0 folding trick when doing FMAs, ↵Owen Anderson
just like it now knows for FMULs. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156029 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-02Teach DAG combine that multiplication by 1.0 can always be constant folded.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156023 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01Tidy up. Naming conventions.Jim Grosbach
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155960 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01Use dyn_cast instead of checking opcode and cast.Jakub Staszak
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155957 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01Strip the pointer casts off of allocas so that the selection DAG can find them.Bill Wendling
PR10799 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155954 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01Target independent Hexagon Packetizer fix.Sirish Pande
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155947 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01Change the PassManager from a reference to a pointer.Bill Wendling
The TargetPassManager's default constructor wants to initialize the PassManager to 'null'. But it's illegal to bind a null reference to a null l-value. Make the ivar a pointer instead. PR12468 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155902 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-30Add some constantness. No functionality change.Jakub Staszak
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155859 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-29RegisterPressure: ArrayRefize some functions for better readability. No ↵Benjamin Kramer
functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155795 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-28Don't update spill weights when joining intervals.Jakob Stoklund Olesen
We don't compute spill weights until after coalescing anyway. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155766 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-28Spring cleaning - Delete dead code.Jakob Stoklund Olesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155765 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-28Reapply 155668: Fix the SD scheduler to avoid gluing the same node twice.Andrew Trick
This time, also fix the caller of AddGlue to properly handle incomplete chains. AddGlue had failure modes, but shamefully hid them from its caller. It's luck ran out. Fixes rdar://11314175: BuildSchedUnits assert. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155749 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-27Temporarily revert r155668: Fix the SD scheduler to avoid gluing.Andrew Trick
This definitely caused regression with ARM -mno-thumb. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155743 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-26Fix the SD scheduler to avoid gluing the same node twice.Andrew Trick
DAGCombine strangeness may result in multiple loads from the same offset. They both may try to glue themselves to another load. We could insist that the redundant loads glue themselves to each other, but the beter fix is to bail out from bad gluing at the time we detect it. Fixes rdar://11314175: BuildSchedUnits assert. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155668 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-25Remove more dead code.Jakob Stoklund Olesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155566 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-25Remove the -disable-cross-class-join option.Jakob Stoklund Olesen
Cross-class joins have been normal and fully supported for a while now. With TableGen generating the getMatchingSuperRegClass() hook, they are unlikely to cause problems again. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155552 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-25Cross-class joining is winning.Jakob Stoklund Olesen
Remove the heuristic for disabling cross-class joins. The greedy register allocator can handle the narrow register classes, and when it splits a live range, it can pick a larger register class. Benchmarks were unaffected by this change. <rdar://problem/11302212> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155551 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24Fix a naughty header include that breaks "installed" builds.Andrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155486 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24MachineBasicBlock::SplitCriticalEdge() should follow LLVM IR variant and ↵Evan Cheng
refuse to break edge to EH landing pad. rdar://11300144 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155470 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24cmake: new fileAndrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155460 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24misched: DAG builder must special case earlyclobberAndrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155459 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24misched: try (not too hard) to place debug values where they belongAndrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155458 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24misched: ignore debug values during schedulingAndrew Trick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155457 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24misched: DAG builder support for tracking register pressure within the ↵Andrew Trick
current scheduling region. The DAG builder is a convenient place to do it. Hopefully this is more efficient than a separate traversal over the same region. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155456 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24RegisterPressure: A utility for computing register pressure within aAndrew Trick
MachineInstr sequence. This uses the new target interface for tracking register pressure using pressure sets to model overlapping register classes and subregisters. RegisterPressure results can be tracked incrementally or stored at region boundaries. Global register pressure can be deduced from local RegisterPressure results if desired. This is an early, somewhat untested implementation. I'm working on testing it within the context of a register pressure reducing MachineScheduler. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155454 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24Look for the 'Is Simulated' module flag. This indicates that the program is ↵Bill Wendling
compiled to run on a simulator. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155435 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-23This patch fixes a problem which arose when using the Post-RA schedulerPreston Gurd
on X86 Atom. Some of our tests failed because the tail merging part of the BranchFolding pass was creating new basic blocks which did not contain live-in information. When the anti-dependency code in the Post-RA scheduler ran, it would sometimes rename the register containing the function return value because the fact that the return value was live-in to the subsequent block had been lost. To fix this, it is necessary to run the RegisterScavenging code in the BranchFolding pass. This patch makes sure that the register scavenging code is invoked in the X86 subtarget only when post-RA scheduling is being done. Post RA scheduling in the X86 subtarget is only done for Atom. This patch adds a new function to the TargetRegisterClass to control whether or not live-ins should be preserved during branch folding. This is necessary in order for the anti-dependency optimizations done during the PostRASchedulerList pass to work properly when doing Post-RA scheduling for the X86 in general and for the Intel Atom in particular. The patch adds and invokes the new function trackLivenessAfterRegAlloc() instead of using the existing requiresRegisterScavenging(). It changes BranchFolding.cpp to call trackLivenessAfterRegAlloc() instead of requiresRegisterScavenging(). It changes the all the targets that implemented requiresRegisterScavenging() to also implement trackLivenessAfterRegAlloc(). It adds an assertion in the Post RA scheduler to make sure that post RA liveness information is available when it is needed. It changes the X86 break-anti-dependencies test to use –mcpu=atom, in order to avoid running into the added assertion. Finally, this patch restores the use of anti-dependency checking (which was turned off temporarily for the 3.1 release) for Intel Atom in the Post RA scheduler. Patch by Andy Zhang! Thanks to Jakob and Anton for their reviews. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155395 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-23Temporarily revert r155364 until the upstream review can complete, perChandler Carruth
the stated developer policy. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155373 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-23Hexagon Packetizer's target independent fix.Sirish Pande
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155364 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-22ZERO_EXTEND/SIGN_EXTEND/TRUNCATE optimization for AVX2Elena Demikhovsky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155309 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-09-23 09:45:45 +0000