The 0x63 opcodes in 64-bit mode have a fixed source size of 32-bits, but the destination size is controlled by REX.W and the 0x66 opsize prefix. This instruction is normally used with a REX.W prefix which provides desired behavior. The other encodings are interpretted as valid by the processor, but aren't useful.
This patch makes us recognize them for the disassembler to match objdump.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343614 91177308-0d34-0410-b5e6-96231b3b80d8
-verify-machineinstrs inserts the MachineVerifier after every MachineInstr-based
pass. However, GlobalISel creates MachineInstr-based passes earlier than DAGISel
and the corresponding verifiers are not being added. This patch fixes that.
If GlobalISel triggers the fallback path then the MIR can be left in a bad
state that is going to be cleared by ResetMachineFunctions. In this situation
verifying between GlobalISel passes will prevent the fallback path from
recovering from this. As a result, we bail out of verifying a function if the
FailedISel attribute is present.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343613 91177308-0d34-0410-b5e6-96231b3b80d8
Add the .cv_fpo_stackalign directive so that we can define $T0, or the
VFRAME virtual register, with it. This was overlooked in the initial
implementation because unlike MSVC, we push CSRs before allocating stack
space, so this value is only needed to describe local variable
locations. Variables that the compiler now addresses via ESP are instead
described as being stored at offsets from VFRAME, which for us is ESP
after alignment in the prologue.
This adds tests that show that we use the VFRAME register properly in
our S_DEFRANGE records, and that we emit the correct FPO data to define
it.
Fixes PR38857
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343603 91177308-0d34-0410-b5e6-96231b3b80d8
These are candidates for the same fold that was implemented in
D52439, but FP types require bitcasting (and that changes the
extra uses profitability calculation).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343587 91177308-0d34-0410-b5e6-96231b3b80d8
This is an attempt to get out of a local-minimum that instcombine currently
gets stuck in. We essentially combine two optimisations at once, ~a - ~b = b-a
and min(~a, ~b) = ~max(a, b), only doing the transform if the result is at
least neutral. This involves using IsFreeToInvert, which has been expanded a
little to include selects that can be easily inverted.
This is trying to fix PR35875, using the ideas from Sanjay. It is a large
improvement to one of our rgb to cmy kernels.
Differential Revision: https://reviews.llvm.org/D52177
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343569 91177308-0d34-0410-b5e6-96231b3b80d8
This adds the memory tagging extension, which is an optional extension
introduced in v8.5A. The new instructions and registers will be added by
subsequent patches.
Patch by Pablo Barrio!
Differential revision: https://reviews.llvm.org/D52486
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343563 91177308-0d34-0410-b5e6-96231b3b80d8
Consistently try to use APFloat::toString for floating point constant comments to get rid of differences between Constant / ConstantDataSequential values - it should help stop some of the linux-windows buildbot failures matching NaN/INF etc. as well.
Differential Revision: https://reviews.llvm.org/D52702
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343562 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This is redundant, as FetchStage::getNextInstruction already checks this
and returns llvm::ErrorSuccess() as appropriate.
NFC.
Reviewers: andreadb
Subscribers: gbedwell, llvm-commits
Differential Revision: https://reviews.llvm.org/D52642
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343555 91177308-0d34-0410-b5e6-96231b3b80d8
getNumUses is linear in the number of uses. Since we're looking for a specific use count, we can use hasNUses which will stop as soon as it determines there are more than N uses instead of walking all of them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343550 91177308-0d34-0410-b5e6-96231b3b80d8
There is no variable in this function named CondBB, but there is one named ThenBB and I believe the comments are all refering to it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343548 91177308-0d34-0410-b5e6-96231b3b80d8
DIA SDK is returning several new sym tag types, so we update
the enumeration and printing code to support these.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343547 91177308-0d34-0410-b5e6-96231b3b80d8
There's a strange assertion on two of the Green Dragon bots that goes away when
this is reverted. The assertion is in RegBankAlloc and if it is this commit then
-verify-machine-instrs should have caught it earlier in the pipeline.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343546 91177308-0d34-0410-b5e6-96231b3b80d8
These assembler directives are still pretty unreadable and it would be
nice to clean them up at some point.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343544 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Before this change, LLVM would always describe locals on the stack as
being relative to some specific register, RSP, ESP, EBP, ESI, etc.
Variables in stack memory are pretty common, so there is a special
S_DEFRANGE_FRAMEPOINTER_REL symbol for them. This change uses it to
reduce the size of our debug info.
On top of the size savings, there are cases on 32-bit x86 where local
variables are addressed from ESP, but ESP changes across the function.
Unlike in DWARF, there is no FPO data to describe the stack adjustments
made to push arguments onto the stack and pop them off after the call,
which makes it hard for the debugger to find the local variables in
frames further up the stack.
To handle this, CodeView has a special VFRAME register, which
corresponds to the $T0 variable set by our FPO data in 32-bit. Offsets
to local variables are instead relative to this value.
This is part of PR38857.
Reviewers: hans, zturner, javed.absar
Subscribers: aprantl, hiraditya, JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D52217
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343543 91177308-0d34-0410-b5e6-96231b3b80d8
Clang-cl was complaining about some sort of constexpr narrowing bug:
C:\src\llvm-project\llvm\lib\CodeGen\GlobalISel\CombinerHelper.cpp(136,31): error: non-constant-expression cannot be narrowed from type 'llvm::TargetOpcode::(anonymous enum at C:\src\llvm-project\llvm\include\llvm/CodeGen/TargetOpcodes.h:22:1)' to 'unsigned int' in initializer list [-Wc++11-narrowing]
unsigned(MI.getOpcode()) == unsigned(TargetOpcode::G_LOAD)
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
C:\src\llvm-project\llvm\lib\CodeGen\GlobalISel\CombinerHelper.cpp(136,31): note: insert an explicit cast to silence this issue
unsigned(MI.getOpcode()) == unsigned(TargetOpcode::G_LOAD)
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
static_cast<unsigned int>(
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343541 91177308-0d34-0410-b5e6-96231b3b80d8
Original message:
This patch adds load folding support to the test shrinking code. This was noticed missing in the review for D52669
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343540 91177308-0d34-0410-b5e6-96231b3b80d8
This includes a fix to prevent i16 compares with i32/i64 ands from being shrunk if bit 15 of the and is set and the sign bit is used.
Original commit message:
Currently we skip looking through truncates if the sign flag is used. But that's overly restrictive.
It's safe to look through the truncate as long as we ensure one of the 3 things when we shrink. Either the MSB of the mask at the shrunken size isn't set. If the mask bit is set then either the shrunk size needs to be equal to the compare size or the sign
There are still missed opportunities to shrink a load and fold it in here. This will be fixed in a future patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343539 91177308-0d34-0410-b5e6-96231b3b80d8
Going from XForm Load to DSForm Load requires that the immediate be 4 byte
aligned.
If we are not aligned we must leave the load as LDX (XForm).
This bug is causing a compile-time failure in the benchmark h264ref.
Differential Revision: https://reviews.llvm.org/D51988
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343525 91177308-0d34-0410-b5e6-96231b3b80d8
This reverts commit r342387 as it's showing significant performance
regressions in a number of benchmarks. Followed up with the
committer and original thread with an example and will get performance
numbers before recommitting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343522 91177308-0d34-0410-b5e6-96231b3b80d8
Spill/reload instructions are artificially generated by the compiler and
have no relation to the original source code. So the best thing to do is
not attach any debug location to them (instead of just taking the next
debug location we find on following instructions).
Differential Revision: https://reviews.llvm.org/D52125
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343520 91177308-0d34-0410-b5e6-96231b3b80d8
There's a subtle bug in the handling of truncate from i32/i64 to i32 without minsize.
I'll be adding more test cases and trying to find a fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343516 91177308-0d34-0410-b5e6-96231b3b80d8
The pattern had a couple of problems:
- It was checking for loads of bytes in the reverse order to what it
should have been looking for.
- It would replace loads of bytes with a load of a word without making
sure that the alignment was correct.
Thanks to Eli Friedman for pointing it out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343514 91177308-0d34-0410-b5e6-96231b3b80d8
Currently it returns incorrect operand size for a target independet
node such as COPY if operand is a register with subreg. Instead of
correct subreg size it returns a size of the whole superreg.
Differential Revision: https://reviews.llvm.org/D52736
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343508 91177308-0d34-0410-b5e6-96231b3b80d8
These work a little differently because they are actually in
the globals stream and are treated as symbol records, even though
DIA presents them as types. So this also adds the necessary
infrastructure to cache records that live somewhere other than
the TPI stream as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343507 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
The AsmParser Lexer regards these as a seperate token.
Here we expand the instruction name with them if they are
adjacent (no whitespace).
Tested: the basic-assembly.s test case has one case with a / in it.
The currently are also instructions with : in them, which we intend
to rename rather than fix them here.
Reviewers: tlively, dschuff
Subscribers: sbc100, jgravelle-google, aheejin, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D52442
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343501 91177308-0d34-0410-b5e6-96231b3b80d8
Currently we skip looking through truncates if the sign flag is used. But that's overly restrictive.
It's safe to look through the truncate as long as we ensure one of the 3 things when we shrink. Either the MSB of the mask at the shrunken size isn't set. If the mask bit is set then either the shrunk size needs to be equal to the compare size or the sign flag needs to be unused.
There are still missed opportunities to shrink a load and fold it in here. This will be fixed in a future patch.
Differential Revision: https://reviews.llvm.org/D52669
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@343498 91177308-0d34-0410-b5e6-96231b3b80d8
Craig Topper