Summary:
It is incorrect to compare TripCount (which is BECount + 1)
with extraiters (or Count) to check if we should enter unrolled
loop or not, because TripCount can potentially overflow
(when BECount is max unsigned integer).
While comparing BECount with (Count - 1) is overflow safe and
therefore correct.
Reviewer: hfinkel
Differential Revision: http://reviews.llvm.org/D19256
From: Evgeny Stupachenko <evstupac@gmail.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267662 91177308-0d34-0410-b5e6-96231b3b80d8
We run after PEI, so we need to AddPristinesAndCSRs.
In practice, that makes no difference here, because we only ask about
liveness of super-registers of defined GR8/GR16 registers, so they
can't be pristine. Still, it's the correct thing to do.
Thanks to Quentin for noticing!
Follow-up to r267495.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267658 91177308-0d34-0410-b5e6-96231b3b80d8
It's probably the case for all 3 MMX users out there, but with
hand-crafted IR, you can trigger selection failures. Fix that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267652 91177308-0d34-0410-b5e6-96231b3b80d8
This effectively adds back the extractelt combine removed by r262358:
the direct case can still occur (because x86_mmx is special, see
r262446), but it's the indirect case that's now superseded by the
generic combine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267651 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This tries to anchor down the concept of domains a bit better. I had
trouble initially relating this to anything. Also talking to David
Majnemer on IRC suggested that I wasn't the only one.
Reviewers: hfinkel
Subscribers: llvm-commits, majnemer
Differential Revision: http://reviews.llvm.org/D18799
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267647 91177308-0d34-0410-b5e6-96231b3b80d8
In the case where isLegalAddressingMode is used for cases
not related to addressing modes, such as pure adds and muls,
it should not be using address space 0. LSR already passes -1
as the address space in these cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267645 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
If the linker requested to preserve a linkonce function, we should
honor this even if we drop all uses.
Reviewers: dexonsmith
Subscribers: llvm-commits, joker.eph
Differential Revision: http://reviews.llvm.org/D19527
From: Mehdi Amini <mehdi.amini@apple.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267644 91177308-0d34-0410-b5e6-96231b3b80d8
This splits out the per-loop functionality from the Pass class.
With this the fact whether the loop is forced-distribute with the new
metadata/pragma can be cached in the per-loop class rather than passed
around.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267643 91177308-0d34-0410-b5e6-96231b3b80d8
When encountering a non-local pointer, LVI would eagerly scan the block for dereferences of the given object to prove the pointer to be non null. That's all well and good, but *then* we'd go recurse through our input blocks. As a result, we could end up scanning each and every block we traverse, even if the final definition was obviously non null or we found a constant value somewhere up the chain. The previous code papered over this by using the isKnownNonNull routine from value tracking. This made the duplication less painful in the common case.
Instead, we know do the block scan only *after* we've gotten the recursive results back. This lets us stop scanning individual blocks as soon as we've determined it to be non-null in any predecessor block and use our usual merge rules to propagate that information cheaply through successor blocks. For a pointer which can be found non-null, this does strictly less work and sometimes substaintially so.
Note that the case where we *can't* prove something non-null is still the really expensive case. We end up scanning each and every block looking for a dereference and never end up finding one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267642 91177308-0d34-0410-b5e6-96231b3b80d8
is defined!
The users were checking the proper thing (Defined + PartialDeadDef), but the
information may have been wrong for other use cases, so fix that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267641 91177308-0d34-0410-b5e6-96231b3b80d8
I missed read the comment when I commited r267621 and thought the
comment did not need update. Matthias kindly proved me wrong.
Fixing that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267638 91177308-0d34-0410-b5e6-96231b3b80d8
the prologue.
Do not use basic blocks that have EFLAGS live-in as prologue if we need
to realign the stack. Realigning the stack uses AND instruction and this
clobbers EFLAGS.
An other alternative would have been to save and restore EFLAGS around
the stack realignment code, but this is likely inefficient.
Fixes PR27531.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267634 91177308-0d34-0410-b5e6-96231b3b80d8
Previously we were recursing on our operands for unary and binary operators regardless of whether we knew how to reason about the operator in question. This has the effect of doing a potentially large amount of work, only to throw it away. By checking whether the operation is one LVI can handle, we can cut short the search and return the (overdefined) answer more quickly. The quality of the results produced should not change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267626 91177308-0d34-0410-b5e6-96231b3b80d8
When the simple analysis provided by MachineBasicBlock::computeRegisterLiveness
fails, fall back on the LivePhysReg utility.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267623 91177308-0d34-0410-b5e6-96231b3b80d8
Now, it is possible to know that partial definitions are dead definitions and
recognize that clobbered registers are also dead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267621 91177308-0d34-0410-b5e6-96231b3b80d8
As pointed out by John Regehr over in http://reviews.llvm.org/D19485, LVI was being incredibly stupid about applying its transfer rules. Rather than gathering local facts from the expression itself, it was simply giving up entirely if one of the inputs was overdefined. This greatly impacts the precision of the overall analysis and makes it far more fragile as well.
This patch builds on 267609 which did the same thing for unary casts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267620 91177308-0d34-0410-b5e6-96231b3b80d8
NVPTXLowerKernelArgs is required for correctness, so it should not be guarded
by CodeGenOpt::None.
NVPTXPeephole is optimization only, so it should be skipped when
CodeGenOpt::None.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267619 91177308-0d34-0410-b5e6-96231b3b80d8
Essentially, I was using the wrong size function. For types which were sized, but not primitive, I wasn't getting a useful size for the operand and failed an assert. I fixed this, and also added a guard that the input is a sized type. Test case is for the original mistake. I'm not sure how to actually exercise the sized type check.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267618 91177308-0d34-0410-b5e6-96231b3b80d8
I'll clean this up and add a test case shortly. I want to make sure this does actually fix the bots; if not, I'll revert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267617 91177308-0d34-0410-b5e6-96231b3b80d8
We need the default ratio to be sufficiently large that it triggers transforms
based on block frequency info (BFI) and plays well with the recently introduced
BranchProbability used by CGP.
Differential Revision: http://reviews.llvm.org/D19435
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267615 91177308-0d34-0410-b5e6-96231b3b80d8
As pointed out by John Regehr over in http://reviews.llvm.org/D19485, LVI was being incredibly stupid about applying its transfer rules. Rather than gathering local facts from the expression itself, it was simply giving up entirely if one of the inputs was overdefined. This greatly impacts the precision of the overall analysis and makes it far more fragile as well.
This patch implements only the unary operation case. Once this is in, I'll implement the same for the binary operations.
Differential Revision: http://reviews.llvm.org/D19492
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267609 91177308-0d34-0410-b5e6-96231b3b80d8
The destination buffer that sprintf uses is restrict qualified, we do
not need to worry about derived pointers referenced via format
specifiers.
This reverts commit r267580.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267605 91177308-0d34-0410-b5e6-96231b3b80d8
The DBI stream contains a lot of bookkeeping information for other
streams. In particular it contains information about section contributions
and linked modules. This patch is a first attempt at parsing some of the
information out of the DBI stream. It currently only parses and dumps the
headers of the DBI stream, so none of the module data or section
contribution data is pulled out.
This is just a proof of concept that we understand the basic properties of
the DBI stream's metadata, and followup patches will try to extract more
detailed information out.
Differential Revision: http://reviews.llvm.org/D19500
Reviewed By: majnemer, ruiu
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267585 91177308-0d34-0410-b5e6-96231b3b80d8
When a block is tail-duplicated, the PHI nodes from that block are
replaced with appropriate COPY instructions. When those PHI nodes
contained use operands with subregisters, the subregisters were
dropped from the COPY instructions, resulting in incorrect code.
Keep track of the subregister information and use this information
when remapping instructions from the duplicated block.
Differential Revision: http://reviews.llvm.org/D19337
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@267583 91177308-0d34-0410-b5e6-96231b3b80d8
