It was just a less powerful and more confusing version of
MCCFIInstruction. A side effect is that, since MCCFIInstruction uses
dwarf register numbers, calls to getDwarfRegNum are pushed out, which
should allow further simplifications.
I left the MachineModuleInfo::addFrameMove interface unchanged since
this patch was already fairly big.
llvm-svn: 181680
This is only tested for global variables at the moment (& includes tests
for the unnamed parameter case, since apparently this entire function
was completely untested previously)
llvm-svn: 181632
for constructors and destructors since the original declaration given
by the AT_specification both won't and can't.
Patch by Yacine Belkadi, I've cleaned up the testcases.
llvm-svn: 181471
This provides basic functionality for imported declarations. For
subprograms and types some amount of lazy construction is supported (so
the definition of a function can proceed the using declaration), but it
still doesn't handle declared-but-not-defined functions (since we don't
generally emit function declarations).
Variable support is really rudimentary at the moment - simply looking up
the existing definition with no support for out of order (declaration,
imported_module, then definition).
llvm-svn: 181392
DIBuilder::createImportedDeclaration isn't fully plumbed through (note,
lacking in AsmPrinter/DwarfDebug support) but this seemed like a
sufficiently useful division of code to make the subsequent patch(es)
easier to follow.
llvm-svn: 181364
Apparently we didn't keep an association of Compile Unit metadata nodes
to DIEs so looking up that parental context failed & thus caused no
DW_TAG_imported_modules to be emitted at the CU scope. Fix this by
adding the mapping & sure up the test case to verify this.
llvm-svn: 181339
Now even the small structures could be passed within byval (small enough
to be stored in GPRs).
In regression tests next function prototypes are checked:
PR15293:
%artz = type { i32 }
define void @foo(%artz* byval %s)
define void @foo2(%artz* byval %s, i32 %p, %artz* byval %s2)
foo: "s" stored in R0
foo2: "s" stored in R0, "s2" stored in R2.
Next AAPCS rules are checked:
5.5 Parameters Passing, C.4 and C.5,
"ParamSize" is parameter size in 32bit words:
-- NSAA != 0, NCRN < R4 and NCRN+ParamSize > R4.
Parameter should be sent to the stack; NCRN := R4.
-- NSAA != 0, and NCRN < R4, NCRN+ParamSize < R4.
Parameter stored in GPRs; NCRN += ParamSize.
llvm-svn: 181148
at all of the operands. Previously it was skipping over implicit operands which
cause infinite looping when the two-address pass try to reschedule a
two-address instruction below the kill of tied operand.
I'm unable to come up with a reasonably sized test case.
rdar://13747577
llvm-svn: 180906
the things, and renames it to CBindingWrapping.h. I also moved
CBindingWrapping.h into Support/.
This new file just contains the macros for defining different wrap/unwrap
methods.
The calls to those macros, as well as any custom wrap/unwrap definitions
(like for array of Values for example), are put into corresponding C++
headers.
Doing this required some #include surgery, since some .cpp files relied
on the fact that including Wrap.h implicitly caused the inclusion of a
bunch of other things.
This also now means that the C++ headers will include their corresponding
C API headers; for example Value.h must include llvm-c/Core.h. I think
this is harmless, since the C API headers contain just external function
declarations and some C types, so I don't believe there should be any
nasty dependency issues here.
llvm-svn: 180881
report a fatal error. This allows us to continue processing the translation
unit. Test case to come on the clang side because we need an inline asm
diagnostics handler in place.
rdar://13446483
llvm-svn: 180873
register-indirect address with an offset of 0.
It used to be that a DBG_VALUE is a register-indirect value if the offset
(operand 1) is nonzero. The new convention is that a DBG_VALUE is
register-indirect if the first operand is a register and the second
operand is an immediate. For plain registers use the combination reg, reg.
rdar://problem/13658587
llvm-svn: 180816
First, taking advantage of the fact that the virtual base registers are allocated in order of the local frame offsets, remove the quadratic register-searching behavior. Because of the ordering, we only need to check the last virtual base register created.
Second, store the frame index in the FrameRef structure, and get the frame index and the local offset from this structure at the top of the loop iteration. This allows us to de-nest the loops in insertFrameReferenceRegisters (and I think makes the code cleaner). I also moved the needsFrameBaseReg check into the first loop over instructions so that we don't bother pushing FrameRefs for instructions that don't want a virtual base register anyway.
Lastly, and this is the only functionality change, avoid the creation of single-use virtual base registers. These are currently not useful because, in general, they end up replacing what would be one r+r instruction with an add and a r+i instruction. Committing this removes the XFAIL in CodeGen/PowerPC/2007-09-07-LoadStoreIdxForms.ll
Jim has okayed this off-list.
llvm-svn: 180799
The `llvm.tls_init_funcs' (created by the front-end) holds pointers to the TLS
initialization functions. These need to be placed into the correct section so
that they are run before `main()'.
<rdar://problem/13733006>
llvm-svn: 180737
to determine whether or not we're on a darwin platform for debug code
emitting.
Solves the problem of a module with no triple on the command line
and no triple in the module using non-gdb ok features on darwin. Fix
up the member-pointers test to check the correct things for cross
platform (DW_FORM_flag is a good prefix).
Unfortunately no testcase because I have no ideas how to test something
without a triple and without a triple in the module yet check
precisely on two platforms. Ideas welcome.
llvm-svn: 180660
Clarify documentation and API to make the difference between register and
register-indirect addressed locations more explicit. Put in a comment
to point out that with the current implementation we cannot specify
a register-indirect location with offset 0 (a breg 0 in DWARF).
No functionality change intended.
rdar://problem/13658587
llvm-svn: 180641
This already helps SSE2 x86 a lot because it lacks an efficient way to
represent a vector select. The long term goal is to enable the backend to match
a canonicalized pattern into a single instruction (e.g. vabs or pabs).
llvm-svn: 180597
Summary:
This is modelled on the Mach-O linker options implementation and should
support a Clang implementation of #pragma comment(lib/linker).
Reviewers: rafael
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D724
llvm-svn: 180569
Fixes PR15838. Need to check for blocks with nothing but dbg.value.
I'm not sure how to force this situation with a unit test. I tried to
reduce the test case in PR15838 (1k lines of metadata) but gave up.
llvm-svn: 180227
For now, we just reschedule instructions that use the copied vregs and
let regalloc elliminate it. I would really like to eliminate the
copies on-the-fly during scheduling, but we need a complete
implementation of repairIntervalsInRange() first.
The general strategy is for the register coalescer to eliminate as
many global copies as possible and shrink live ranges to be
extended-basic-block local. The coalescer should not have to worry
about resolving local copies (e.g. it shouldn't attemp to reorder
instructions). The scheduler is a much better place to deal with local
interference. The coalescer side of this equation needs work.
llvm-svn: 180193
When MachineScheduler is enabled, this functionality can be
removed. Until then, provide a way to disable it for test cases and
designing MachineScheduler heuristics.
llvm-svn: 180192
This exposed an issue with PowerPC AltiVec where it appears it was setting the wrong vector boolean contents. The included change
fixes the PowerPC tests, and was OK'd by Hal.
llvm-svn: 180129
1) Disallow 'returned' on parameter that is also 'sret' (no sensible semantics, as far as I can tell).
2) Conservatively disallow tail calls through 'returned' parameters that also are 'zext' or 'sext' (for consistency with treatment of other zero-extending and sign-extending operations in tail call position detection...can be revised later to handle situations that can be determined to be safe).
This is a new attribute that is not yet used, so there is no impact.
llvm-svn: 180118
This reverts commit r179840 with a fix to test/DebugInfo/two-cus-from-same-file.ll
I'm not sure why that test only failed on ARM & MIPS and not X86 Linux, even
though the debug info was clearly invalid on all of them, but this ought to fix
it.
llvm-svn: 179996
Rather than just splitting the input type and hoping for the best, apply
a bit more cleverness. Just splitting the types until the source is
legal often leads to an illegal result time, which is then widened and a
scalarization step is introduced which leads to truly horrible code
generation. With the loop vectorizer, these sorts of operations are much
more common, and so it's worth extra effort to do them well.
Add a legalization hook for the operands of a TRUNCATE node, which will
be encountered after the result type has been legalized, but if the
operand type is still illegal. If simple splitting of both types
ends up with the result type of each half still being legal, just
do that (v16i16 -> v16i8 on ARM, for example). If, however, that would
result in an illegal result type (v8i32 -> v8i8 on ARM, for example),
we can get more clever with power-two vectors. Specifically,
split the input type, but also widen the result element size, then
concatenate the halves and truncate again. For example on ARM,
To perform a "%res = v8i8 trunc v8i32 %in" we transform to:
%inlo = v4i32 extract_subvector %in, 0
%inhi = v4i32 extract_subvector %in, 4
%lo16 = v4i16 trunc v4i32 %inlo
%hi16 = v4i16 trunc v4i32 %inhi
%in16 = v8i16 concat_vectors v4i16 %lo16, v4i16 %hi16
%res = v8i8 trunc v8i16 %in16
This allows instruction selection to generate three VMOVN instructions
instead of a sequences of moves, stores and loads.
Update the ARMTargetTransformInfo to take this improved legalization
into account.
Consider the simplified IR:
define <16 x i8> @test1(<16 x i32>* %ap) {
%a = load <16 x i32>* %ap
%tmp = trunc <16 x i32> %a to <16 x i8>
ret <16 x i8> %tmp
}
define <8 x i8> @test2(<8 x i32>* %ap) {
%a = load <8 x i32>* %ap
%tmp = trunc <8 x i32> %a to <8 x i8>
ret <8 x i8> %tmp
}
Previously, we would generate the truly hideous:
.syntax unified
.section __TEXT,__text,regular,pure_instructions
.globl _test1
.align 2
_test1: @ @test1
@ BB#0:
push {r7}
mov r7, sp
sub sp, sp, #20
bic sp, sp, #7
add r1, r0, #48
add r2, r0, #32
vld1.64 {d24, d25}, [r0:128]
vld1.64 {d16, d17}, [r1:128]
vld1.64 {d18, d19}, [r2:128]
add r1, r0, #16
vmovn.i32 d22, q8
vld1.64 {d16, d17}, [r1:128]
vmovn.i32 d20, q9
vmovn.i32 d18, q12
vmov.u16 r0, d22[3]
strb r0, [sp, #15]
vmov.u16 r0, d22[2]
strb r0, [sp, #14]
vmov.u16 r0, d22[1]
strb r0, [sp, #13]
vmov.u16 r0, d22[0]
vmovn.i32 d16, q8
strb r0, [sp, #12]
vmov.u16 r0, d20[3]
strb r0, [sp, #11]
vmov.u16 r0, d20[2]
strb r0, [sp, #10]
vmov.u16 r0, d20[1]
strb r0, [sp, #9]
vmov.u16 r0, d20[0]
strb r0, [sp, #8]
vmov.u16 r0, d18[3]
strb r0, [sp, #3]
vmov.u16 r0, d18[2]
strb r0, [sp, #2]
vmov.u16 r0, d18[1]
strb r0, [sp, #1]
vmov.u16 r0, d18[0]
strb r0, [sp]
vmov.u16 r0, d16[3]
strb r0, [sp, #7]
vmov.u16 r0, d16[2]
strb r0, [sp, #6]
vmov.u16 r0, d16[1]
strb r0, [sp, #5]
vmov.u16 r0, d16[0]
strb r0, [sp, #4]
vldmia sp, {d16, d17}
vmov r0, r1, d16
vmov r2, r3, d17
mov sp, r7
pop {r7}
bx lr
.globl _test2
.align 2
_test2: @ @test2
@ BB#0:
push {r7}
mov r7, sp
sub sp, sp, #12
bic sp, sp, #7
vld1.64 {d16, d17}, [r0:128]
add r0, r0, #16
vld1.64 {d20, d21}, [r0:128]
vmovn.i32 d18, q8
vmov.u16 r0, d18[3]
vmovn.i32 d16, q10
strb r0, [sp, #3]
vmov.u16 r0, d18[2]
strb r0, [sp, #2]
vmov.u16 r0, d18[1]
strb r0, [sp, #1]
vmov.u16 r0, d18[0]
strb r0, [sp]
vmov.u16 r0, d16[3]
strb r0, [sp, #7]
vmov.u16 r0, d16[2]
strb r0, [sp, #6]
vmov.u16 r0, d16[1]
strb r0, [sp, #5]
vmov.u16 r0, d16[0]
strb r0, [sp, #4]
ldm sp, {r0, r1}
mov sp, r7
pop {r7}
bx lr
Now, however, we generate the much more straightforward:
.syntax unified
.section __TEXT,__text,regular,pure_instructions
.globl _test1
.align 2
_test1: @ @test1
@ BB#0:
add r1, r0, #48
add r2, r0, #32
vld1.64 {d20, d21}, [r0:128]
vld1.64 {d16, d17}, [r1:128]
add r1, r0, #16
vld1.64 {d18, d19}, [r2:128]
vld1.64 {d22, d23}, [r1:128]
vmovn.i32 d17, q8
vmovn.i32 d16, q9
vmovn.i32 d18, q10
vmovn.i32 d19, q11
vmovn.i16 d17, q8
vmovn.i16 d16, q9
vmov r0, r1, d16
vmov r2, r3, d17
bx lr
.globl _test2
.align 2
_test2: @ @test2
@ BB#0:
vld1.64 {d16, d17}, [r0:128]
add r0, r0, #16
vld1.64 {d18, d19}, [r0:128]
vmovn.i32 d16, q8
vmovn.i32 d17, q9
vmovn.i16 d16, q8
vmov r0, r1, d16
bx lr
llvm-svn: 179989
I think it's almost impossible to fold atomic fences profitably under
LLVM/C++11 semantics. As a result, this is now unused and just
cluttering up the target interface.
llvm-svn: 179940
Adding another CU-wide list, in this case of imported_modules (since they
should be relatively rare, it seemed better to add a list where each element
had a "context" value, rather than add a (usually empty) list to every scope).
This takes care of DW_TAG_imported_module, but to fully address PR14606 we'll
need to expand this to cover DW_TAG_imported_declaration too.
llvm-svn: 179836
This is a rework of the broken parts in r179373 which were subsequently reverted in r179374 due to incompatibility with C++98 compilers. This version should be ok under C++98.
llvm-svn: 179520
The register allocator expects minimal physreg live ranges. Schedule
physreg copies accordingly. This is slightly tricky when they occur in
the middle of the scheduling region. For now, this is handled by
rescheduling the copy when its associated instruction is
scheduled. Eventually we may instead bundle them, but only if we can
preserve the bundles as parallel copies during regalloc.
llvm-svn: 179449
When debugging performance regressions we often ask ourselves if the regression
that we see is due to poor isel/sched/ra or due to some micro-architetural
problem. When comparing two code sequences one good way to rule out front-end
bottlenecks (and other the issues) is to force code alignment. This pass adds
a flag that forces the alignment of all of the basic blocks in the program.
llvm-svn: 179353
In the simple and triangle if-conversion cases, when CopyAndPredicateBlock is
used because the to-be-predicated block has other predecessors, we need to
explicitly remove the old copied block from the successors list. Normally if
conversion relies on TII->AnalyzeBranch combined with BB->CorrectExtraCFGEdges
to cleanup the successors list, but if the predicated block contained an
un-analyzable branch (such as a now-predicated return), then this will fail.
These extra successors were causing a problem on PPC because it was causing
later passes (such as PPCEarlyReturm) to leave dead return-only basic blocks in
the code.
llvm-svn: 179227
The target hooks are getting out of hand. What does it mean to run
before or after regalloc anyway? Allowing either Pass* or AnalysisID
pass identification should make it much easier for targets to use the
substitutePass and insertPass APIs, and create less need for badly
named target hooks.
llvm-svn: 179140
therefore not at all) of the pc or statement list. We also don't
need to emit the compilation dir so save so space and time
and don't bother.
Fix up the testcase accordingly and verify that we don't emit
the attributes or the items that they use.
llvm-svn: 179114
This pattern occurs in SROA output due to the way vector arguments are lowered
on ARM.
The testcase from PR15525 now compiles into this, which is better than the code
we got with the old scalarrepl:
_Store:
ldr.w r9, [sp]
vmov d17, r3, r9
vmov d16, r1, r2
vst1.8 {d16, d17}, [r0]
bx lr
Differential Revision: http://llvm-reviews.chandlerc.com/D647
llvm-svn: 179106
a relocation across sections. Do this for DW_AT_stmt list in the
skeleton CU and check the relocations in the debug_info section.
Add a FIXME for multiple CUs.
llvm-svn: 178969
This fixes PEI as previously described, but correctly handles the case where
the instruction defining the virtual register to be scavenged is the first in
the block. Arnold provided me with a bugpoint-reduced test case, but even that
seems too large to use as a regression test. If I'm successful in cleaning it
up then I'll commit that as well.
Original commit message:
This change fixes a bug that I introduced in r178058. After a register is
scavenged using one of the available spills slots the instruction defining the
virtual register needs to be moved to after the spill code. The scavenger has
already processed the defining instruction so that registers killed by that
instruction are available for definition in that same instruction. Unfortunately,
after this, the scavenger needs to iterate through the spill code and then
visit, again, the instruction that defines the now-scavenged register. In order
to avoid confusion, the register scavenger needs the ability to 'back up'
through the spill code so that it can again process the instructions in the
appropriate order. Prior to this fix, once the scavenger reached the
just-moved instruction, it would assert if it killed any registers because,
having already processed the instruction, it believed they were undefined.
Unfortunately, I don't yet have a small test case. Thanks to Pranav Bhandarkar
for diagnosing the problem and testing this fix.
llvm-svn: 178919
During LTO, the target options on functions within the same Module may
change. This would necessitate resetting some of the back-end. Do this for X86,
because it's a Friday afternoon.
llvm-svn: 178917
Reverting because this breaks one of the LTO builders. Original commit message:
This change fixes a bug that I introduced in r178058. After a register is
scavenged using one of the available spills slots the instruction defining the
virtual register needs to be moved to after the spill code. The scavenger has
already processed the defining instruction so that registers killed by that
instruction are available for definition in that same instruction. Unfortunately,
after this, the scavenger needs to iterate through the spill code and then
visit, again, the instruction that defines the now-scavenged register. In order
to avoid confusion, the register scavenger needs the ability to 'back up'
through the spill code so that it can again process the instructions in the
appropriate order. Prior to this fix, once the scavenger reached the
just-moved instruction, it would assert if it killed any registers because,
having already processed the instruction, it believed they were undefined.
Unfortunately, I don't yet have a small test case. Thanks to Pranav Bhandarkar
for diagnosing the problem and testing this fix.
llvm-svn: 178916
This change fixes a bug that I introduced in r178058. After a register is
scavenged using one of the available spills slots the instruction defining the
virtual register needs to be moved to after the spill code. The scavenger has
already processed the defining instruction so that registers killed by that
instruction are available for definition in that same instruction. Unfortunately,
after this, the scavenger needs to iterate through the spill code and then
visit, again, the instruction that defines the now-scavenged register. In order
to avoid confusion, the register scavenger needs the ability to 'back up'
through the spill code so that it can again process the instructions in the
appropriate order. Prior to this fix, once the scavenger reached the
just-moved instruction, it would assert if it killed any registers because,
having already processed the instruction, it believed they were undefined.
Unfortunately, I don't yet have a small test case. Thanks to Pranav Bhandarkar
for diagnosing the problem and testing this fix.
llvm-svn: 178845
For now, just save the compile time since the ConvergingScheduler
heuristics don't use this analysis. We'll probably enable it later
after compile-time investigation.
llvm-svn: 178822
On certain architectures we can support efficient vectorized version of
instructions if the operand value is uniform (splat) or a constant scalar.
An example of this is a vector shift on x86.
We can efficiently support
for (i = 0 ; i < ; i += 4)
w[0:3] = v[0:3] << <2, 2, 2, 2>
but not
for (i = 0; i < ; i += 4)
w[0:3] = v[0:3] << x[0:3]
This patch adds a parameter to getArithmeticInstrCost to further qualify operand
values as uniform or uniform constant.
Targets can then choose to return a different cost for instructions with such
operand values.
A follow-up commit will test this feature on x86.
radar://13576547
llvm-svn: 178807
There is a difference for FORM_ref_addr between DWARF 2 and DWARF 3+.
Since Eric is against guarding DWARF 2 ref_addr with DarwinGDBCompat, we are
still in discussion on how to handle this.
The correct solution is to update our header to say version 4 instead of version
2 and update tool chains as well.
rdar://problem/13559431
llvm-svn: 178806
the target system.
It was hard-coded to 4 bytes before. I can't get llvm to generate a
ref_addr on a reasonably sized testing case.
rdar://problem/13559431
llvm-svn: 178722
For this we need to use a libcall. Previously LLVM didn't implement
libcall support for frem, so I've added it in the usual
straightforward manner. A test case from the bug report is included.
llvm-svn: 178639
The new instruction scheduling models provide information about the
number of cycles consumed on each processor resource. This makes it
possible to estimate ILP more accurately than simply counting
instructions / issue width.
The functions getResourceDepth() and getResourceLength() now identify
the limiting processor resource, and return a cycle count based on that.
This gives more precise resource information, particularly in traces
that use one resource a lot more than others.
llvm-svn: 178553
This is helps on architectures where i8,i16 are not legal but we have byte, and
short loads/stores. Allowing us to merge copies like the one below on ARM.
copy(char *a, char *b, int n) {
do {
int t0 = a[0];
int t1 = a[1];
b[0] = t0;
b[1] = t1;
radar://13536387
llvm-svn: 178546
We would also like to merge sequences that involve a variable index like in the
example below.
int index = *idx++
int i0 = c[index+0];
int i1 = c[index+1];
b[0] = i0;
b[1] = i1;
By extending the parsing of the base pointer to handle dags that contain a
base, index, and offset we can handle examples like the one above.
The dag for the code above will look something like:
(load (i64 add (i64 copyfromreg %c)
(i64 signextend (i8 load %index))))
(load (i64 add (i64 copyfromreg %c)
(i64 signextend (i32 add (i32 signextend (i8 load %index))
(i32 1)))))
The code that parses the tree ignores the intermediate sign extensions. However,
if there is a sign extension it needs to be on all indexes.
(load (i64 add (i64 copyfromreg %c)
(i64 signextend (add (i8 load %index)
(i8 1))))
vs
(load (i64 add (i64 copyfromreg %c)
(i64 signextend (i32 add (i32 signextend (i8 load %index))
(i32 1)))))
radar://13536387
llvm-svn: 178483
immediate in a register. I don't believe this should ever fail, but I see no
harm in trying to make this code bullet proof.
I've added an assert to ensure my assumtion is correct. If the assertion fires
something is wrong and we should fix it, rather then just silently fall back to
SelectionDAG isel.
llvm-svn: 178305
This is a follow-up to r178073 (which should actually make target-customized
spilling work again).
I still don't have a regression test for this (but it would be good to have
one; Thumb 1 and Mips16 use this callback as well).
Patch by Richard Sandiford.
llvm-svn: 178137
As pointed out by Richard Sandiford, my recent updates to the register
scavenger broke targets that use custom spilling (because the new code assumed
that if there were no valid spill slots, than spilling would be impossible).
I don't have a test case, but it should be possible to create one for Thumb 1,
Mips 16, etc.
llvm-svn: 178073
The previous algorithm could not deal properly with scavenging multiple virtual
registers because it kept only one live virtual -> physical mapping (and
iterated through operands in order). Now we don't maintain a current mapping,
but rather use replaceRegWith to completely remove the virtual register as
soon as the mapping is established.
In order to allow the register scavenger to return a physical register killed
by an instruction for definition by that same instruction, we now call
RS->forward(I) prior to eliminating virtual registers defined in I. This
requires a minor update to forward to ignore virtual registers.
These new features will be tested in forthcoming commits.
llvm-svn: 178058
- Handle the case where the result of 'insert_subvect' is bitcasted
before 'extract_subvec'. This removes the redundant insertf128/extractf128
pair on unaligned 256-bit vector load/store on vectors of non 64-bit integer.
llvm-svn: 177945
For instance, following transformation will be disabled:
x + x + x => 3.0f * x;
The problem of these transformations is that it introduces a FP constant, which
following Instruction-Selection pass cannot handle.
Reviewed by Nadav, thanks a lot!
rdar://13445387
llvm-svn: 177933
Performing this check unilaterally prevented us from generating FMAs when the incoming IR contained illegal vector types which would eventually be legalized to underlying types that *did* support FMA.
For example, an @llvm.fmuladd on an OpenCL float16 should become a sequence of float4 FMAs, not float4 fmul+fadd's.
NOTE: Because we still call the target-specific profitability hook, individual targets can reinstate the old behavior, if desired, by simply performing the legality check inside their callback hook. They can also perform more sophisticated legality checks, if, for example, some illegal vector types can be productively implemented as FMAs, but not others.
llvm-svn: 177820
177774 broke the lld-x86_64-darwin11 builder; error:
error: comparison of integers of different signs: 'int' and 'size_type' (aka 'unsigned long')
for (SI = 0; SI < Scavenged.size(); ++SI)
~~ ^ ~~~~~~~~~~~~~~~~
Fix this by making SI also unsigned.
llvm-svn: 177780
This patch lets the register scavenger make use of multiple spill slots in
order to guarantee that it will be able to provide multiple registers
simultaneously.
To support this, the RS's API has changed slightly: setScavengingFrameIndex /
getScavengingFrameIndex have been replaced by addScavengingFrameIndex /
isScavengingFrameIndex / getScavengingFrameIndices.
In forthcoming commits, the PowerPC backend will use this capability in order
to implement the spilling of condition registers, and some special-purpose
registers, without relying on r0 being reserved. In some cases, spilling these
registers requires two GPRs: one for addressing and one to hold the value being
transferred.
llvm-svn: 177774
This reverts commit 06091513c283c863296f01cc7c2e86b56bb50d02.
The code is obviously wrong, but the trivial fix causes
inefficient code generation on X86. Somebody with more
knowledge of the code needs to take a look here.
Signed-off-by: Christian König <christian.koenig@amd.com>
llvm-svn: 177529
A node's ordering is only propagated during legalization if (a) the new node does
not have an ordering (is not a CSE'd node), or (b) the new node has an ordering
that is higher than the node being legalized.
llvm-svn: 177465
The linker sorts the .tls$<xyz> sections by name, and we need
to make sure any extra sections we produce (e.g. for weak globals)
always end up between .tls$AAA and .tls$ZZZ, even if the name
starts with e.g. an underscore.
Patch by David Nadlinger!
llvm-svn: 177256
Implicit defs are not currently positional and not modeled by the
per-operand machine model. Unfortunately, we treat defs that are part
of the architectural instruction description, like flags, the same as
other implicit defs. Really, they should have a fixed MachineInstr
layout and probably shouldn't be "implicit" at all.
For now, we'll change the default latency to be the max operand
latency. That will give flag setting operands full latency for x86
folded loads. Other kinds of "fake" implicit defs don't occur prior to
regalloc anyway, and we would like them to go away postRegAlloc as
well.
llvm-svn: 177227
This is a generic function (derived from PEI); moving it into
MachineFrameInfo eliminates a current redundancy between the ARM and AArch64
backends, and will allow it to be used by the PowerPC target code.
No functionality change intended.
llvm-svn: 177111
Add the current PEI register scavenger as a parameter to the
processFunctionBeforeFrameFinalized callback.
This change is necessary in order to allow the PowerPC target code to
set the register scavenger frame index after the save-area offset
adjustments performed by processFunctionBeforeFrameFinalized. Only
after these adjustments have been made is it possible to estimate
the size of the stack frame.
llvm-svn: 177108
This doesn't reset all of the target options within the TargetOptions
object. This is because some of those are ABI-specific and must be determined if
it's okay to change those on the fly.
llvm-svn: 176986
belongs to a different compile unit.
DW_FORM_ref_addr should be used for cross compile-unit reference.
When compiling a large application, we got a dwarfdump verification error where
abstract_origin points to nowhere.
This error can't be reproduced on any testing case in MultiSource.
We may have other cases where we use DW_FORM_ref4 unconditionally.
rdar://problem/13370501
llvm-svn: 176882
Versioned debug info support has been a burden to maintain & also compromised
current debug info verification by causing test cases testing old debug info to
remain rather than being updated to the latest. It also makes it hard to add or
change the metadata schema by requiring various backwards-compatibility in the
DI* hierarchy.
So it's being removed in preparation for new changes to the schema to tidy up
old/unnecessary fields and add new fields needed for new debug info (well, new
to LLVM at least).
The more surprising part of this is the changes to DI*::Verify - this became
necessary due to the changes to AsmWriter. AsmWriter was relying on the version
test to decide which bits of metadata were actually debug info when printing
the comment annotations. Without the version information the tag numbers were
too common & it would print debug info on random metadata that happened to
start with an integer that matched a tag number. Instead this change makes the
Verify functions more precise (just adding "number of operands" checks - not
type checking those operands yet) & relies on that to decide which metadata is
debug info metadata.
llvm-svn: 176838
Summary:
Statistics are still available in Release+Asserts (any +Asserts builds),
and stats can also be turned on with LLVM_ENABLE_STATS.
Move some of the FastISel stats that were moved under DEBUG()
back out of DEBUG(), since stats are disabled across the board now.
Many tests depend on grepping "-stats" output. Move those into
a orig_dir/Stats/. so that they can be marked as unsupported
when building without statistics.
Differential Revision: http://llvm-reviews.chandlerc.com/D486
llvm-svn: 176733
LegalizeDAG.cpp uses the value of the comparison operands when checking
the legality of BR_CC, so DAGCombiner should do the same.
v2:
- Expand more BR_CC value types for NVPTX
v3:
- Expand correct BR_CC value types for Hexagon, Mips, and XCore.
llvm-svn: 176694
This verifies live intervals both before and after scheduling. It's
useful for anyone hacking on live interval update.
Note that we don't yet pass verification all the time. We don't yet
handle updating nonallocatable live intervals perfectly.
llvm-svn: 176685
Code generation makes some basic assumptions about the IR it's been given. In
particular, if there is only one 'invoke' in the function, then that invoke
won't be going away. However, with the advent of the `llvm.donothing' intrinsic,
those invokes may go away. If all of them go away, the landing pad no longer has
any users. This confuses the back-end, which asserts.
This happens with SjLj exceptions, because that's the model that modifies the IR
based on there being invokes, etc. in the function.
Remove any invokes of `llvm.donothing' during SjLj EH preparation. This will
give us a CFG that the back-end won't be confused about. If all of the invokes
in a function are removed, then the SjLj EH prepare pass won't insert the bogus
code the relies upon the invokes being there.
<rdar://problem/13228754&13316637>
llvm-svn: 176677
In very rare cases caused by irreducible control flow, the dominating
block can have the same trace head without actually being part of the
trace.
As long as such a dominator still has valid instruction depths, it is OK
to use it for computing instruction depths.
Rename the function to avoid lying, and add a check that instruction
depths are computed for the dominator.
llvm-svn: 176668
rdar:13370002 [pre-RA-sched] assertion: released too many times
I tracked this down to an earlier hack that is no longer applicable
and interfered with normal scheduler logic. With the changes in
r176037, it was causing an instruction to be scheduled multiple times.
I have an external test case that I tried hard to reduce and
failed. I can't even reproduce with llc.
llvm-svn: 176636
We now emit a line table for each compile unit. To reduce the prologue size
of each line table, the files and directories used by each compile unit are
stored in std::map<unsigned, std::vector< > > instead of std::vector< >.
The prologue for a lto'ed image can be as big as 93K. Duplicating 93K for each
compile unit causes a huge increase of debug info. With this patch, each
prologue will only emit the files required by the compile unit.
rdar://problem/13342023
llvm-svn: 176605
- ISD::SHL/SRL/SRA must have either both scalar or both vector operands
but TLI.getShiftAmountTy() so far only return scalar type. As a
result, backend logic assuming that breaks.
- Rename the original TLI.getShiftAmountTy() to
TLI.getScalarShiftAmountTy() and re-define TLI.getShiftAmountTy() to
return target-specificed scalar type or the same vector type as the
1st operand.
- Fix most TICG logic assuming TLI.getShiftAmountTy() a simple scalar
type.
llvm-svn: 176364
We avoided computing DAG height/depth during Node printing because it
shouldn't depend on an otherwise valid DAG. But this has become far
too annoying for the common case of a valid DAG where we want to see
valid values. If doing the computation on-the-fly turns out to be a
problem in practice, then I'll add a mode to the diagnostics to only
force it when we're likely to have a valid DAG, otherwise explicitly
print INVALID instead of bogus numbers. For now, just go for it all
the time.
llvm-svn: 176314
SelectionDAGIsel::LowerArguments needs a function, not a basic block. So it
makes sense to pass it the function instead of extracting a basic-block from
the function and then tossing it. This is also more self-documenting (functions
have arguments, BBs don't).
In addition, added comments to a couple of Select* methods.
llvm-svn: 176305
We make the cost for calling libm functions extremely high as emitting the
calls is expensive and causes spills (on x86) so performance suffers. We still
vectorize important calls like ceilf and friends on SSE4.1. and fabs.
Differential Revision: http://llvm-reviews.chandlerc.com/D466
llvm-svn: 176287
definition DIE (TAG_variable), and put AT_MIPS_linkage_name to TAG_member when
DarwinGDBCompat is true.
Darwin GDB needs AT_MIPS_linkage_name at both places to work.
Follow-up patch to r176143.
rdar://problem/13291234
llvm-svn: 176220
definition DIE, to make old GDB happy.
We have a regression for old GDB when Clang uses DW_TAG_member to declare
static members inside a class, instead of DW_TAG_variable. This patch will fix
this regression.
rdar://problem/13291234
llvm-svn: 176143
TAG_member inside a class to the specification DIE.
Having AT_MIPS_linkage_name on TAG_member caused old gdb (GNU 6.3.50) to
error out. Also gcc 4.7 has AT_MIPS_linkage_name on the specification DIE.
rdar://problem/13291234
llvm-svn: 176120
fewer scalar integer (i32 or i64) arguments. It completely eliminates the need
for SDISel for trivial functions.
Also, add the new llc -fast-isel-abort-args option, which is similar to
-fast-isel-abort option, but for formal argument lowering.
llvm-svn: 176052
memory intrinsics in the SDAG builder.
When alignment is zero, the lang ref says that *no* alignment
assumptions can be made. This is the exact opposite of the internal API
contracts of the DAG where alignment 0 indicates that the alignment can
be made to be anything desired.
There is another, more explicit alignment that is better suited for the
role of "no alignment at all": an alignment of 1. Map the intrinsic
alignment to this early so that we don't end up generating aligned DAGs.
It is really terrifying that we've never seen this before, but we
suddenly started generating a large number of alignment 0 memcpys due to
the new code to do memcpy-based copying of POD class members. That patch
contains a bug that rounds bitfield alignments down when they are the
first field. This can in turn produce zero alignments.
This fixes weird crashes I've seen in library users of LLVM on 32-bit
hosts, etc.
llvm-svn: 176022
itself recursively with a new instruction that has not been finalized, in order
to determine whether to keep the instruction. On 'make check' and test-suite the
only cases where the recursive invocation made any transformations were simple
instruction commutations, so I am restricting the recursive invocation to do
only this.
The other cases wouldn't work correctly when updating LiveIntervals, since the
new instructions don't have slot indices and LiveIntervals hasn't yet been
updated. If the other transformations were actually triggering in any test case
it would be possible to support it with a lot of effort, but since they don't
it's not worth it.
llvm-svn: 175979
unless it was requested to with an optional parameter that defaults to false, so
we don't need to handle that case in TwoAddressInstructionPass.
llvm-svn: 175974
TwoAddressInstructionPass. The code in rescheduleMIBelowKill() is a bit tricky,
since multiple instructions need to be moved down, one-at-a-time, in reverse
order.
llvm-svn: 175955
One of the phases of SelectionDAG is LegalizeVectors. We don't need to sort the DAG and copy nodes around if there are no vector ops.
Speeds up the compilation time of SelectionDAG on a big scalar workload by ~8%.
llvm-svn: 175929
It was incorrectly checking a Function* being an IntrinsicInst* which
isn't possible. It should always have been checking the CallInst* instead.
Added test case for x86 which ensures we only get one constant load.
It was 2 before this change.
rdar://problem/13267920
llvm-svn: 175853
pass. One of the callers of isKilled() can cope with overapproximation of kills
and the other can't, so I added a flag to indicate this.
In theory this could pessimize code slightly, but in practice most physical
register uses are kills, and most important kills of physical registers are the
only uses of that register prior to register allocation, so we can recognize
them as kills even without kill flags.
This is relevant because LiveIntervals gets rid of all kill flags.
llvm-svn: 175821
to TargetFrameLowering, where it belongs. Incidentally, this allows us
to delete some duplicated (and slightly different!) code in TRI.
There are potentially other layering problems that can be cleaned up
as a result, or in a similar manner.
The refactoring was OK'd by Anton Korobeynikov on llvmdev.
Note: this touches the target interfaces, so out-of-tree targets may
be affected.
llvm-svn: 175788
This fixes some problems with too conservative checking where we were
marking all aliases of a register as used, and then also checking all
aliases when allocating a register.
<rdar://problem/13249625>
llvm-svn: 175782
A legal BUILD_VECTOR goes in and gets constant folded into another legal
BUILD_VECTOR so we don't lose any legality here. The problematic PPC
optimization that made this check necessary was fixed recently.
llvm-svn: 175759
available.
With this commit there are no longer any assertion or verifier failures when
running 'make check' without LiveVariables. There are still 56 failing tests
with codegen differences and 1 unexpectedly passing test.
llvm-svn: 175719
Rewrite value numbers directly in the 'Other' LiveInterval which is
moribund anyway. This avoids allocating the OtherAssignments vector.
llvm-svn: 175690
When findReachingDefs() finds that only one value can reach the basic
block, just copy the work list of visited blocks directly into the live
interval.
Sort the block list and use a LiveRangeUpdater to make the bulk add
fast.
When multiple reaching defs are found, transfer the work list to the
updateSSA() work list as before. Also use LiveRangeUpdater in
updateLiveIns() following updateSSA().
This makes live interval analysis more than 3x faster on one huge test
case.
llvm-svn: 175685
related failures when running 'make check' without LiveVariables with the
verifier enabled. Some of the remaining failures elsewhere may still be fallout
from incorrect updating of LiveIntervals or the few missing cases left in the
two-address pass.
llvm-svn: 175672
(2xi32) (truncate ((2xi64) bitcast (buildvector i32 a, i32 x, i32 b, i32 y)))
can be folded into a (2xi32) (buildvector i32 a, i32 b).
Such a DAG would cause uneccessary vdup instructions followed by vmovn
instructions.
We generate this code on ARM NEON for a setcc olt, 2xf64, 2xf64. For example, in
the vectorized version of the code below.
double A[N];
double B[N];
void test_double_compare_to_double() {
int i;
for(i=0;i<N;i++)
A[i] = (double)(A[i] < B[i]);
}
radar://13191881
Fixes bug 15283.
llvm-svn: 175670
Adding new segments to large LiveIntervals can be expensive because the
LiveRange objects after the insertion point may need to be moved left or
right. This can cause quadratic behavior when adding a large number of
segments to a live range.
The LiveRangeUpdater class allows the LIveInterval to be in a temporary
invalid state while segments are being added. It maintains an internal
gap in the LiveInterval when it is shrinking, and it has a spill area
for new segments when the LiveInterval is growing.
The behavior is similar to the existing mergeIntervalRanges() function,
except it allocates less memory for the spill area, and the algorithm is
turned inside out so the loop is driven by the clients.
llvm-svn: 175644
- When extloading from a vector with non-byte-addressable element, e.g.
<4 x i1>, the current logic breaks. Extend the current logic to
fix the case where the element type is not byte-addressable by loading
all bytes, bit-extracting/packing each element.
llvm-svn: 175642
and removing instructions. The implementation seems more complicated than it
needs to be, but I couldn't find something simpler that dealt with all of the
corner cases.
Also add a call to repairIndexesInRange() from repairIntervalsInRange().
llvm-svn: 175601
Target implementations of getRegAllocationHints() should use the
provided allocation order, and they can never return hints outside the
order. This is already documented in TargetRegisterInfo.h.
<rdar://problem/13240556>
llvm-svn: 175540
Due to the execution order of doFinalization functions, the GC information were
deleted before AsmPrinter::doFinalization was executed. Thus, the
GCMetadataPrinter::finishAssembly was never called.
The patch fixes that by moving the code of the GCInfoDeleter::doFinalization to
Printer::doFinalization.
llvm-svn: 175528
arguably better than forward iterators for this use case, they are confusing and
there are some implementation problems with reverse iterators and MI bundles.
llvm-svn: 175393
MachineBasicBlock::SplitCriticalEdge. Since this is an iterator rather than
an instr_iterator, the isBundled() check only passes if getFirstTerminator()
returned end() and the garbage memory happens to lean that way.
Multiple successors can be present without any terminator instructions in the
case of exception handling with a fallthrough.
llvm-svn: 175383
terminators that actually have register uses when splitting critical edges.
This commit also introduces a method repairIntervalsInRange() on LiveIntervals,
which allows for repairing LiveIntervals in a small range after an arbitrary
target hook modifies, inserts, and removes instructions. It's pretty limited
right now, but I hope to extend it to support all of the things that are done
by the convertToThreeAddress() target hooks.
llvm-svn: 175382
If the frame pointer is omitted, and any stack changes occur in the inline
assembly, e.g.: "pusha", then any C local variable or C argument references
will be incorrect.
I pass no judgement on anyone who would do such a thing. ;)
rdar://13218191
llvm-svn: 175334
If two functions require different features (e.g., `-mno-sse' vs. `-msse') then
we want to honor that, especially during LTO. We can do that by resetting the
subtarget's features depending upon the 'target-feature' attribute.
llvm-svn: 175314
- add sincos to runtime library if target triple environment is GNU
- added canCombineSinCosLibcall() which checks that sincos is in the RTL and
if the environment is GNU then unsafe fpmath is enabled (required to
preserve errno)
- extended sincos-opt lit test
Reviewed by: Hal Finkel
llvm-svn: 175283
of the copy is a subregister def. The current code assumes that it can do a full
def of the destination register, but it is not checking that the def operand is
read-undef. It also doesn't clear the subregister index of the destination in
the new instruction to reflect the full subregister def.
These issues were found running 'make check' with my next commit that enables
rematerialization in more cases.
llvm-svn: 175122
RegisterCoalescer used to depend on LiveDebugVariable. LDV removes DBG_VALUEs
without emitting them at the end.
We fix this by removing LDV from RegisterCoalescer. Also add an assertion to
make sure we call emitDebugValues if DBG_VALUEs are removed at
runOnMachineFunction.
rdar://problem/13183203
Reviewed by Andy & Jakob
llvm-svn: 175023
DAGCombiner::ReduceLoadWidth was converting (trunc i32 (shl i64 v, 32))
into (shl i32 v, 32) into undef. To prevent this, check the shift count
against the final result size.
Patch by: Kevin Schoedel
Reviewed by: Nadav Rotem
llvm-svn: 174972
live ranges should always be extended, and the only successor that should be
considered for extension of other ranges is the target of the split edge.
llvm-svn: 174935
Sorry for the lack of a test case. I tried writing one for i386 as i know selects are illegal on this target, but they are actually considered legal by isel and expanded later.
I can't see any targets to trigger this, but checking for the legality of a node before forming it is general goodness.
llvm-svn: 174934
This is currently a bit hairier than it needs to be, since depending on where the
split block resides the end ListEntry of the split block may be the end ListEntry
of the original block or a new entry. Some changes to the SlotIndexes updating
should make it possible to eliminate the two cases here.
This also isn't as optimized as it could be. In the future Liveinterval should
probably get a flag that indicates whether the LiveInterval is within a single
basic block. We could ignore all such intervals when splitting an edge.
llvm-svn: 174870
This reverts my commit 171047. Now that I've removed my misguided attempt to
support backend warnings, these diagnostics are only about inline assembly.
It would take quite a bit more work to generalize them properly, so I'm
just reverting this.
llvm-svn: 174860
function is successfully handled by fast-isel. That's because function
arguments are *always* handled by SDISel. Introduce FastLowerArguments to
allow each target to provide hook to handle formal argument lowering.
As a proof-of-concept, add ARMFastIsel::FastLowerArguments to handle
functions with 4 or fewer scalar integer (i8, i16, or i32) arguments. It
completely eliminates the need for SDISel for trivial functions.
rdar://13163905
llvm-svn: 174855
support for updating SlotIndexes to MachineBasicBlock::SplitCriticalEdge(). This
calls renumberIndexes() every time; it should be improved to only renumber
locally.
llvm-svn: 174851
present, it currently verifies them with the MachineVerifier, and this passed
all of the test cases in 'make check' (when accounting for existing verifier
errors). There were some assertion failures in the two-address pass, but they
also happened on code without phis and look like they are caused by different
kill flags from LiveIntervals.
The only part that doesn't work is the critical edge splitting heuristic,
because there isn't currently an efficient way to update LiveIntervals after
splitting an edge. I'll probably start by implementing the slow fallback and
test that it works before tackling the fast path for single-block ranges. The
existing code that updates LiveVariables is fairly slow as it is.
There isn't a command-line option for enabling this; instead, just edit
PHIElimination.cpp to require LiveIntervals.
llvm-svn: 174831
This uses a liveness algorithm that does not depend on data from the
LiveVariables analysis, it is the first step towards removing
LiveVariables completely.
llvm-svn: 174774
This reverts r171041. This was a nice idea that didn't work out well.
Clang warnings need to be associated with warning groups so that they can
be selectively disabled, promoted to errors, etc. This simplistic patch didn't
allow for that. Enhancing it to provide some way for the backend to specify
a front-end warning type seems like overkill for the few uses of this, at
least for now.
llvm-svn: 174748
Previously, even when a pre-increment load or store was generated,
we often needed to keep a copy of the original base register for use
with other offsets. If all of these offsets are constants (including
the offset which was combined into the addressing mode), then this is
clearly unnecessary. This change adjusts these other offsets to use the
new incremented address.
llvm-svn: 174746
Aside from the question of whether we report a warning or an error when we
can't satisfy a requested stack object alignment, the current implementation
of this is not good. We're not providing any source location in the diagnostics
and the current warning is not connected to any warning group so you can't
control it. We could improve the source location somewhat, but we can do a
much better job if this check is implemented in the front-end, so let's do that
instead. <rdar://problem/13127907>
llvm-svn: 174741
Adds a function to target transform info to query for the cost of address
computation. The cost model analysis pass now also queries this interface.
The code in LoopVectorize adds the cost of address computation as part of the
memory instruction cost calculation. Only there, we know whether the instruction
will be scalarized or not.
Increase the penality for inserting in to D registers on swift. This becomes
necessary because we now always assume that address computation has a cost and
three is a closer value to the architecture.
radar://13097204
llvm-svn: 174713
Failure: undefined symbol 'Lline_table_start0'.
Root-cause: we use a symbol subtraction to calculate at_stmt_list, but
the line table entries are not dumped in the assembly.
Fix: use zero instead of a symbol subtraction for Compile Unit 0.
llvm-svn: 174479
base point of a load, and the overall alignment of the load. This caused infinite loops in DAG combine with the
original application of this patch.
ORIGINAL COMMIT LOG:
When the target-independent DAGCombiner inferred a higher alignment for a load,
it would replace the load with one with the higher alignment. However, it did
not place the new load in the worklist, which prevented later DAG combines in
the same phase (for example, target-specific combines) from ever seeing it.
This patch corrects that oversight, and updates some tests whose output changed
due to slightly different DAGCombine outputs.
llvm-svn: 174431
All targets are now adding return value registers as implicit uses on
return instructions, and there is no longer a need for the live out
lists.
llvm-svn: 174417
it would replace the load with one with the higher alignment. However, it did
not place the new load in the worklist, which prevented later DAG combines in
the same phase (for example, target-specific combines) from ever seeing it.
This patch corrects that oversight, and updates some tests whose output changed
due to slightly different DAGCombine outputs.
llvm-svn: 174343
Per discussion in rdar://13127907, we should emit a hard error only if
people write code where the requested alignment is larger than achievable
and assumes the low bits are zeros. A warning should be good enough when
we are not sure if the source code assumes the low bits are zeros.
rdar://13127907
llvm-svn: 174336
This required disabling a PowerPC optimization that did the following:
input:
x = BUILD_VECTOR <i32 16, i32 16, i32 16, i32 16>
lowered to:
tmp = BUILD_VECTOR <i32 8, i32 8, i32 8, i32 8>
x = ADD tmp, tmp
The add now gets folded immediately and we're back at the BUILD_VECTOR we
started from. I don't see a way to fix this currently so I left it disabled
for now.
Fix some trivially foldable X86 tests too.
llvm-svn: 174325
We used to create children DIEs for a scope, then check whether ScopeDIE is
null. If ScopeDIE is null, the children DIEs will be dangling. Other DIEs can
link to those dangling DIEs, which are not emitted at all, causing dwarf error.
The current testing case is 4k lines, from MultiSource/BenchMark/McCat/09-vor.
rdar://problem/13071959
llvm-svn: 174084
conditions are met:
1. They share the same operand and are in the same BB.
2. Both outputs are used.
3. The target has a native instruction that maps to ISD::FSINCOS node or
the target provides a sincos library call.
Implemented the generic optimization in sdisel and enabled it for
Mac OSX. Also added an additional optimization for x86_64 Mac OSX by
using an alternative entry point __sincos_stret which returns the two
results in xmm0 / xmm1.
rdar://13087969
PR13204
llvm-svn: 173755
The common code in the post-RA scheduler to break anti-dependencies on the
critical path contained a flaw. In the reported case, an anti-dependency
between the overlapping registers %X4 and %R4 exists:
%X29<def> = OR8 %X4, %X4
%R4<def>, %X3<def,dead,tied3> = LBZU 1, %X3<kill,tied1>
The unpatched code breaks the dependency by replacing %R4 and its uses
with %R3, the first register on the available list. However, %R3 and
%X3 overlap, so this creates two overlapping definitions on the same
instruction.
The fix is straightforward, preventing selection of a register that
overlaps any other defined register on the same instruction.
The test case is reduced from the bug report, and verifies that we no
longer produce "lbzu 3, 1(3)" when breaking this anti-dependency.
llvm-svn: 173706
Fix that by adding a cast to the shift expander. This came up with vector shifts
on sse-less X86 CPUs.
<2 x i64> = shl <2 x i64> <2 x i64>
-> i64,i64 = shl i64 i64; shl i64 i64
-> i32,i32,i32,i32 = shl_parts i32 i32 i64; shl_parts i32 i32 i64
Now we cast the last two i64s to the right type. Fixes the crash in PR14668.
llvm-svn: 173615
with an initial number of elements, instead of DenseMap, which has
zero initial elements, in order to avoid the copying of elements
when the size changes and to avoid allocating space every time
LegalizeTypes is run. This patch will not affect the memory footprint,
because DenseMap will increase the element size to 64
when the first element is added.
Patch by Wan Xiaofei.
llvm-svn: 173448
Maintain separate per-node and per-tree book-keeping.
Track all instructions above a DAG node including nested subtrees.
Seperately track instructions within a subtree.
Record subtree parents.
llvm-svn: 173426
For sanity, create a root when NumDataSuccs >= 4. Splitting large
subtrees will no longer be detrimental after my next checkin to handle
nested tree. A magic number of 4 is fine because single subtrees
seldom rejoin more than this. It makes subtrees easier to visualize
and heuristics more sane.
llvm-svn: 173399
The requirements of the strong heuristic are:
* A Protector is required for functions which contain an array, regardless of
type or length.
* A Protector is required for functions which contain a structure/union which
contains an array, regardless of type or length. Note, there is no limit to
the depth of nesting.
* A protector is required when the address of a local variable (i.e., stack
based variable) is exposed. (E.g., such as through a local whose address is
taken as part of the RHS of an assignment or a local whose address is taken as
part of a function argument.)
llvm-svn: 173231
SSPStrong applies a heuristic to insert stack protectors in these situations:
* A Protector is required for functions which contain an array, regardless of
type or length.
* A Protector is required for functions which contain a structure/union which
contains an array, regardless of type or length. Note, there is no limit to
the depth of nesting.
* A protector is required when the address of a local variable (i.e., stack
based variable) is exposed. (E.g., such as through a local whose address is
taken as part of the RHS of an assignment or a local whose address is taken as
part of a function argument.)
This patch implements the SSPString attribute to be equivalent to
SSPRequired. This will change in a subsequent patch.
llvm-svn: 173230
Previously we tried to infer it from the bit width size, with an added
IsIEEE argument for the PPC/IEEE 128-bit case, which had a default
value. This default value allowed bugs to creep in, where it was
inappropriate.
llvm-svn: 173138
A SparseMultiSet adds multiset behavior to SparseSet, while retaining SparseSet's desirable properties. Essentially, SparseMultiSet provides multiset behavior by storing its dense data in doubly linked lists that are inlined into the dense vector. This allows it to provide good data locality as well as vector-like constant-time clear() and fast constant time find(), insert(), and erase(). It also allows SparseMultiSet to have a builtin recycler rather than keeping SparseSet's behavior of always swapping upon removal, which allows it to preserve more iterators. It's often a better alternative to a SparseSet of a growable container or vector-of-vector.
llvm-svn: 173064
The optimization handles esoteric cases but adds a lot of complexity both to the X86 backend and to other backends.
This optimization disables an important canonicalization of chains of SEXT nodes and makes SEXT and ZEXT asymmetrical.
Disabling the canonicalization of consecutive SEXT nodes into a single node disables other DAG optimizations that assume
that there is only one SEXT node. The AVX mask optimizations is one example. Additionally this optimization does not update the cost model.
llvm-svn: 172968
Move the early if-conversion pass into this group.
ILP optimizations usually need to find the right balance between
register pressure and ILP using the MachineTraceMetrics analysis to
identify critical paths and estimate other costs. Such passes should run
together so they can share dominator tree and loop info analyses.
Besides if-conversion, future passes to run here here could include
expression height reduction and ARM's MLxExpansion pass.
llvm-svn: 172687
using the DW_FORM_GNU_addr_index and a separate .debug_addr section which
stays in the executable and is fully linked.
Sneak in two other small changes:
a) Print out the debug_str_offsets.dwo section.
b) Change form we're expecting the entries in the debug_str_offsets.dwo
section to take from ULEB128 to U32.
Add tests for all of this in the fission-cu.ll test.
llvm-svn: 172578
The included test case is derived from one of the GCC compatibility tests.
The problem arises after the selection DAG has been converted to type-legalized
form. The combiner first sees a 64-bit load that can be converted into a
pre-increment form. The original load feeds into a SRL that isolates the
upper 32 bits of the loaded doubleword. This looks like an opportunity for
DAGCombiner::ReduceLoadWidth() to replace the 64-bit load with a 32-bit load.
However, this transformation is not valid, as the replacement load is not
a pre-increment load. The pre-increment load produces an extra result,
which feeds a subsequent add instruction. The replacement load only has
one result value, and this value is propagated to all uses of the pre-
increment load, including the add. Because the add is looking for the
second result value as its operand, it ends up attempting to add a constant
to a token chain, resulting in a crash.
So the patch simply disables this transformation for any load with more than
two result values.
llvm-svn: 172480
Remember the minimum cost of the registers in an allocation order and
the number of registers at the end of the allocation order that have the
same cost per use.
This information can be used to limit the search space for
RAGreedy::tryEvict() when looking for a cheaper register.
llvm-svn: 172280
This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still
a complete mess but as long as the edges consist of virtual call it doesn't
cause breakage. BasicTTI did static calls and thus broke some build
configurations.
llvm-svn: 172246
the target if it supports the different CAST types. We didn't do this
on X86 because of the different register sizes and types, but on ARM
this makes sense.
llvm-svn: 172245
- recognize string "{memory}" in the MI generation
- mark as mayload/maystore when there's a memory clobber constraint.
PR14859.
Patch by Krzysztof Parzyszek
llvm-svn: 172228
When calling hasProperty() on an instruction inside a bundle, it should
always behave as if IgnoreBundle was passed, and just return properties
for the current instruction.
Only attempt to aggregate bundle properties whan asked about the bundle
header.
The assertion fires on existing ARM test cases without this fix.
llvm-svn: 172082
requirement when creating stack objects in MachineFrameInfo.
Add CreateStackObjectWithMinAlign to throw error when the minimal alignment
can't be achieved and to clamp the alignment when the preferred alignment
can't be achieved. Same is true for CreateVariableSizedObject.
Will not emit error in CreateSpillStackObject or CreateStackObject.
As long as callers of CreateStackObject do not assume the object will be
aligned at the requested alignment, we should not have miscompile since
later optimizations which look at the object's alignment will have the correct
information.
rdar://12713765
llvm-svn: 172027
It cahced XOR's operands before calling visitXOR() but failed to update the
operands when visitXOR changed the XOR node.
rdar://12968664
llvm-svn: 171999
It is possible to build MI bundles that don't begin with a BUNDLE
header. Add support for such bundles, counting all instructions inside
the bundle.
llvm-svn: 171985
fp128 is almost but not quite completely illegal as a type on AArch64. As a
result it needs to have a register class (for argument passing mainly), but all
operations need to be lowered to runtime calls. Currently there's no way for
targets to do this (without duplicating code), as the relevant functions are
hidden in SelectionDAG. This patch changes that.
llvm-svn: 171971
This was an experimental option, but needs to be defined
per-target. e.g. PPC A2 needs to aggressively hide latency.
I converted some in-order scheduling tests to A2. Hal is working on
more test cases.
llvm-svn: 171946
one file where it is called as a static function. Nuke the declaration
and the definition in lib/CodeGen, along with the include of
SelectionDAG.h from this file.
There is no dependency edge from lib/CodeGen to
lib/CodeGen/SelectionDAG, so it isn't valid for a routine in lib/CodeGen
to reference the DAG. There is a dependency from
lib/CodeGen/SelectionDAG on lib/CodeGen. This breaks one violation of
this layering.
llvm-svn: 171842
proposal. This leaves the strings in the skeleton die as strp,
but in all dwo files they're accessed now via DW_FORM_GNU_str_index.
Add support for dumping these sections and modify the fission-cu.ll
testcase to have the correct strings and form. Fix a small bug
in the fixed form sizes routine that involved out of array accesses
for the table and add a FIXME in the extractFast routine to fix
this up.
llvm-svn: 171779
peculiar headers under include/llvm.
This struct still doesn't make a lot of sense, but it makes more sense
down in TargetLowering than it did before.
llvm-svn: 171739
TargetTransformInfo rather than TargetLowering, removing one of the
primary instances of the layering violation of Transforms depending
directly on Target.
This is a really big deal because LSR used to be a "special" pass that
could only be tested fully using llc and by looking at the full output
of it. It also couldn't run with any other loop passes because it had to
be created by the backend. No longer is this true. LSR is now just
a normal pass and we should probably lift the creation of LSR out of
lib/CodeGen/Passes.cpp and into the PassManagerBuilder. =] I've not done
this, or updated all of the tests to use opt and a triple, because
I suspect someone more familiar with LSR would do a better job. This
change should be essentially without functional impact for normal
compilations, and only change behvaior of targetless compilations.
The conversion required changing all of the LSR code to refer to the TTI
interfaces, which fortunately are very similar to TargetLowering's
interfaces. However, it also allowed us to *always* expect to have some
implementation around. I've pushed that simplification through the pass,
and leveraged it to simplify code somewhat. It required some test
updates for one of two things: either we used to skip some checks
altogether but now we get the default "no" answer for them, or we used
to have no information about the target and now we do have some.
I've also started the process of removing AddrMode, as the TTI interface
doesn't use it any longer. In some cases this simplifies code, and in
others it adds some complexity, but I think it's not a bad tradeoff even
there. Subsequent patches will try to clean this up even further and use
other (more appropriate) abstractions.
Yet again, almost all of the formatting changes brought to you by
clang-format. =]
llvm-svn: 171735
This works fine with GDB for member variable pointers, but GDB's support for
member function pointers seems to be quite unrelated to
DW_TAG_ptr_to_member_type. (see GDB bug 14998 for details)
llvm-svn: 171698
a TargetMachine to construct (and thus isn't always available), to an
analysis group that supports layered implementations much like
AliasAnalysis does. This is a pretty massive change, with a few parts
that I was unable to easily separate (sorry), so I'll walk through it.
The first step of this conversion was to make TargetTransformInfo an
analysis group, and to sink the nonce implementations in
ScalarTargetTransformInfo and VectorTargetTranformInfo into
a NoTargetTransformInfo pass. This allows other passes to add a hard
requirement on TTI, and assume they will always get at least on
implementation.
The TargetTransformInfo analysis group leverages the delegation chaining
trick that AliasAnalysis uses, where the base class for the analysis
group delegates to the previous analysis *pass*, allowing all but tho
NoFoo analysis passes to only implement the parts of the interfaces they
support. It also introduces a new trick where each pass in the group
retains a pointer to the top-most pass that has been initialized. This
allows passes to implement one API in terms of another API and benefit
when some other pass above them in the stack has more precise results
for the second API.
The second step of this conversion is to create a pass that implements
the TargetTransformInfo analysis using the target-independent
abstractions in the code generator. This replaces the
ScalarTargetTransformImpl and VectorTargetTransformImpl classes in
lib/Target with a single pass in lib/CodeGen called
BasicTargetTransformInfo. This class actually provides most of the TTI
functionality, basing it upon the TargetLowering abstraction and other
information in the target independent code generator.
The third step of the conversion adds support to all TargetMachines to
register custom analysis passes. This allows building those passes with
access to TargetLowering or other target-specific classes, and it also
allows each target to customize the set of analysis passes desired in
the pass manager. The baseline LLVMTargetMachine implements this
interface to add the BasicTTI pass to the pass manager, and all of the
tools that want to support target-aware TTI passes call this routine on
whatever target machine they end up with to add the appropriate passes.
The fourth step of the conversion created target-specific TTI analysis
passes for the X86 and ARM backends. These passes contain the custom
logic that was previously in their extensions of the
ScalarTargetTransformInfo and VectorTargetTransformInfo interfaces.
I separated them into their own file, as now all of the interface bits
are private and they just expose a function to create the pass itself.
Then I extended these target machines to set up a custom set of analysis
passes, first adding BasicTTI as a fallback, and then adding their
customized TTI implementations.
The fourth step required logic that was shared between the target
independent layer and the specific targets to move to a different
interface, as they no longer derive from each other. As a consequence,
a helper functions were added to TargetLowering representing the common
logic needed both in the target implementation and the codegen
implementation of the TTI pass. While technically this is the only
change that could have been committed separately, it would have been
a nightmare to extract.
The final step of the conversion was just to delete all the old
boilerplate. This got rid of the ScalarTargetTransformInfo and
VectorTargetTransformInfo classes, all of the support in all of the
targets for producing instances of them, and all of the support in the
tools for manually constructing a pass based around them.
Now that TTI is a relatively normal analysis group, two things become
straightforward. First, we can sink it into lib/Analysis which is a more
natural layer for it to live. Second, clients of this interface can
depend on it *always* being available which will simplify their code and
behavior. These (and other) simplifications will follow in subsequent
commits, this one is clearly big enough.
Finally, I'm very aware that much of the comments and documentation
needs to be updated. As soon as I had this working, and plausibly well
commented, I wanted to get it committed and in front of the build bots.
I'll be doing a few passes over documentation later if it sticks.
Commits to update DragonEgg and Clang will be made presently.
llvm-svn: 171681
pass into the SelectionDAG itself rather than snooping on the
implementation of that pass as exposed by the TargetMachine. This
removes the last direct client of the ScalarTargetTransformInfo class
outside of the TTI pass implementation.
llvm-svn: 171625
This change essentially reverts r87069 which came without a test case. It
causes no regressions in the GDB 7.5 test suite & fixes 25 xfails (commit
to the test suite to follow). If anyone can present a test case that
demonstrates why this check is necessary I'd be happy to account for it in one
way or another.
llvm-svn: 171609
The series of patches leading up to this one makes llc -O0 run 8% faster.
When deallocating a MachineFunction, there is no need to visit all
MachineInstr and MachineOperand objects to deallocate them. All their
memory come from a BumpPtrAllocator that is about to be purged, and they
have empty destructors anyway.
This only applies when deallocating the MachineFunction.
DeleteMachineInstr() should still be used to recycle MI memory during
the codegen passes.
Remove the LeakDetector support for MachineInstr. I've never seen it
used before, and now it definitely doesn't work. With this patch, leaked
MachineInstrs would be much less of a problem since all of their memory
will be reclaimed by ~MachineFunction().
llvm-svn: 171599
Instead of an std::vector<MachineOperand>, use MachineOperand arrays
from an ArrayRecycler living in MachineFunction.
This has several advantages:
- MachineInstr now has a trivial destructor, making it possible to
delete them in batches when destroying MachineFunction. This will be
enabled in a later patch.
- Bypassing malloc() and free() can be faster, depending on the system
library.
- MachineInstr objects and their operands are allocated from the same
BumpPtrAllocator, so they will usually be next to each other in
memory, providing better locality of reference.
- Reduce MachineInstr footprint. A std::vector is 24 bytes, the new
operand array representation only uses 8+4+1 bytes in MachineInstr.
- Better control over operand array reallocations. In the old
representation, the use-def chains would be reordered whenever a
std::vector reached its capacity. The new implementation never changes
the use-def chain order.
Note that some decisions in the code generator depend on the use-def
chain orders, so this patch may cause different assembly to be produced
in a few cases.
llvm-svn: 171598
This function works like memmove() for MachineOperands, except it also
updates any use-def chains containing the moved operands.
The use-def chains are updated without affecting the order of operands
in the list. That isn't possible when using the
removeRegOperandFromUseList() and addRegOperandToUseList() functions.
Callers to follow soon.
llvm-svn: 171597
Most IMPLICIT_DEF instructions are removed by the ProcessImplicitDefs
pass, and a few are reinserted by PHIElimination when a PHI argument is
<undef>.
RegisterCoalescer was assuming that all IMPLICIT_DEF live ranges look
like those created by PHIElimination, and that their live range never
leaves the basic block.
The PR14732 test case does tricks with PHI nodes that causes a longer
IMPLICIT_DEF live range to appear. This happens very rarely, but
RegisterCoalescer should be able to handle it.
llvm-svn: 171435
DAGCombiner::reduceBuildVecConvertToConvertBuildVec() was making two
mistakes:
1. It was checking the legality of scalar INT_TO_FP nodes and then generating
vector nodes.
2. It was passing the result value type to
TargetLoweringInfo::getOperationAction() when it should have been
passing the value type of the first operand.
llvm-svn: 171420
into their new header subdirectory: include/llvm/IR. This matches the
directory structure of lib, and begins to correct a long standing point
of file layout clutter in LLVM.
There are still more header files to move here, but I wanted to handle
them in separate commits to make tracking what files make sense at each
layer easier.
The only really questionable files here are the target intrinsic
tablegen files. But that's a battle I'd rather not fight today.
I've updated both CMake and Makefile build systems (I think, and my
tests think, but I may have missed something).
I've also re-sorted the includes throughout the project. I'll be
committing updates to Clang, DragonEgg, and Polly momentarily.
llvm-svn: 171366
utils/sort_includes.py script.
Most of these are updating the new R600 target and fixing up a few
regressions that have creeped in since the last time I sorted the
includes.
llvm-svn: 171362
The later API is nicer than the former, and is correct regarding wrap-around offsets (if anyone cares).
There are a few more places left with duplicated code, which I'll remove soon.
llvm-svn: 171259
directly.
This is in preparation for removing the use of the 'Attribute' class as a
collection of attributes. That will shift to the AttributeSet class instead.
llvm-svn: 171253
These are now generally used for all diagnostics from the backend, not just
for inline assembly, so this drops the "InlineAsm" from the names. No
functional change. (I've left aliases for the old names but only for long
enough to let me switch over clang to use the new ones.)
llvm-svn: 171047
When the backend is used from clang, it should produce proper diagnostics
instead of just printing messages to errs(). Other clients may also want to
register their own error handlers with the LLVMContext, and the same handler
should work for warnings in the same way as the existing emitError methods.
llvm-svn: 171041
Instructions that are inserted in a basic block can still be decorated
with addOperand(MO).
Make the two-argument addOperand() function contain the actual
implementation. This function will now always have a valid MF reference
that it can use for memory allocation.
llvm-svn: 170798
This function is often used to decorate dangling instructions, so a
context reference is required to allocate memory for the operands.
Also add a corresponding MachineInstrBuilder method.
llvm-svn: 170797
Akira Hatanaka