Summary:
Byval argument pair formation assumes that if a type is less than 8 bytes
it must be an integer and not a pointer, which is not true for x32 and NaCl.
Relax the assertion and add a test for a codegen case that triggered it.
Reviewers: jvoung
Subscribers: jfb, cfe-commits
Differential Revision: http://reviews.llvm.org/D10701
llvm-svn: 240600
As specified in the SysV AVX512 ABI drafts. It follows the same scheme
as AVX2:
Arguments of type __m512 are split into eight eightbyte chunks.
The least significant one belongs to class SSE and all the others
to class SSEUP.
This also means we change the OpenMP SIMD default alignment on AVX512.
Based on r240337.
Differential Revision: http://reviews.llvm.org/D9894
llvm-svn: 240338
The patch is generated using this command:
$ tools/extra/clang-tidy/tool/run-clang-tidy.py -fix \
-checks=-*,llvm-namespace-comment -header-filter='llvm/.*|clang/.*' \
work/llvm/tools/clang
To reduce churn, not touching namespaces spanning less than 10 lines.
llvm-svn: 240270
This patch fixes an assertion failure in method
'X86_64ABIInfo::GetByteVectorType'.
Method 'GetByteVectorType' (in TargetInfo.cpp) is responsible
for mapping a QualType 'Ty' (for an argument or return value) to an LLVM IR
type that, according to the ABI, must be passed in a XMM/YMM vector register.
When selecting the IR vector type, method 'GetByteVectorType' always tries to
choose the "best" IR vector type for the 'Ty' in input. In particular, if Ty
is a wrapper structure, it keeps unwrapping it until it finds a vector type VTy.
That VTy is the "preferred IR type".
However, function 'isSingleElementStructure' (used to unwrap structures) does
not know how to look through union types. So, before this patch, if Ty was in
a nest of wrapper structures with at least two union types, we would have
triggered an assertion failure (added at revision 230971).
With this patch, if method 'GetByteVectorType' fails to find the preferred
vector type, we just return a valid (although potentially 'less friendly')
vector type based on the type size. So, rather than asserting on an 'unexpected'
'Ty' in input, we conservatively return vector type <2 x double> if Ty is 16
bytes, or <4 x double> if Ty is 32 bytes.
Differential Revision: http://reviews.llvm.org/D10190
llvm-svn: 238861
If the type isn't trivially moveable emplace can skip a potentially
expensive move. It also saves a couple of characters.
Call sites were found with the ASTMatcher + some semi-automated cleanup.
memberCallExpr(
argumentCountIs(1), callee(methodDecl(hasName("push_back"))),
on(hasType(recordDecl(has(namedDecl(hasName("emplace_back")))))),
hasArgument(0, bindTemporaryExpr(
hasType(recordDecl(hasNonTrivialDestructor())),
has(constructExpr()))),
unless(isInTemplateInstantiation()))
No functional change intended.
llvm-svn: 238601
Re-land the change r238200, but with modifications in the tests that should
prevent new failures in some environments as reported with the original
change on the mailing list.
llvm-svn: 238253
On MIPS unsigned int type should not be zero extended but sign-extended.
Patch by Strahinja Petrovic.
Differential Revision: http://reviews.llvm.org/D9198
llvm-svn: 238200
We already have the ABI, we don't need a "HasAVX" flag.
This will also makes it easier to add an AVX512 ABI.
No functional change intended.
llvm-svn: 237989
Also add trivial handling of transparent unions.
PPC32, MSP430, and XCore apparently all rely on DefaultABIInfo. This
should worry you, because DefaultABIInfo is not implementing the rules
of any particular ABI.
Fixes PR23097, patch by Andy Gibbs.
llvm-svn: 237630
This patch adds support for the z13 architecture type. For compatibility
with GCC, a pair of options -mvx / -mno-vx can be used to selectively
enable/disable use of the vector facility.
When the vector facility is present, we default to the new vector ABI.
This is characterized by two major differences:
- Vector types are passed/returned in vector registers
(except for unnamed arguments of a variable-argument list function).
- Vector types are at most 8-byte aligned.
The reason for the choice of 8-byte vector alignment is that the hardware
is able to efficiently load vectors at 8-byte alignment, and the ABI only
guarantees 8-byte alignment of the stack pointer, so requiring any higher
alignment for vectors would require dynamic stack re-alignment code.
However, for compatibility with old code that may use vector types, when
*not* using the vector facility, the old alignment rules (vector types
are naturally aligned) remain in use.
These alignment rules are not only implemented at the C language level,
but also at the LLVM IR level. This is done by selecting a different
DataLayout string depending on whether the vector ABI is in effect or not.
Based on a patch by Richard Sandiford.
llvm-svn: 236531
- Changed CUDALaunchBounds arguments from integers to Expr* so they can
be saved in AST for instantiation.
- Added support for template instantiation of launch_bounds attrubute.
- Moved evaluation of launch_bounds arguments to NVPTXTargetCodeGenInfo::
SetTargetAttributes() where it can be done after template instantiation.
- Added a warning on negative launch_bounds arguments.
- Amended test cases.
Differential Revision: http://reviews.llvm.org/D8985
llvm-svn: 235452
Something like { void*, void * } would be passed to a function as a [2 x i64], but returned as an i128. This patch unifies the 2 behaviours so that we also return it as a [2 x i64].
This is better for the quality of the IR, and the size of the final LLVM binary as we tend to want to insert/extract values from these types and do so with the insert/extract instructions is less IR than shifting, truncating, and or'ing values.
Reviewed by Tim Northover.
llvm-svn: 235231
C structs.
This comes up when we have a function that takes a struct and is defined in a
C++ file and used in a C file.
Before this commit, we will generate byval for C++ and will expand the struct
for C, thus causing difference at IR level. We will use bitcast of function type
at the callsite, which causes the inliner to not inline the function.
This commit changes how we handle small C like structs at IR level, but at
backend, we should generate the same argument passing before and after the
commit.
Note that the condition for expanding is still over conservative. We should be
able to expand type that is spelled with “class” and types that are not C-like.
But this commit fixes the inconsistent argument passing between C/C++.
Reviewed by John.
rdar://20121030
llvm-svn: 234033
Running the GCC's inter-compiler ABI compatibility test suite uncovered
a couple of errors in clang's SystemZ ABI implementation. These all
affect only rare corner cases:
- Short vector types
GCC synthetic vector types defined with __attribute__ ((vector_size ...))
are always passed and returned by reference. (This is not documented in
the official ABI document, but is the de-facto ABI implemented by GCC.)
clang would do that only for vector sizes >= 16 bytes, but not for shorter
vector types.
- Float-like aggregates and empty bitfields
clang would consider any aggregate containing an empty bitfield as
first element to be a float-like aggregate. That's obviously wrong.
According to the ABI doc, the presence of an empty bitfield makes
an aggregate to be *not* float-like. However, due to a bug in GCC,
empty bitfields are ignored in C++; this patch changes clang to be
compatible with this "feature" of GCC.
- Float-like aggregates and va_arg
The va_arg implementation would mis-detect some aggregates as float-like
that aren't actually passed as such. This applies to aggregates that
have only a single element of type float or double, but using an aligned
attribute that increases the total struct size to more than 8 bytes.
This error occurred because the va_arg implement used to have an copy
of the float-like aggregate detection logic (i.e. it would call the
isFPArgumentType routine, but not perform the size check).
To simplify the logic, this patch removes the duplicated logic and
instead simply checks the (possibly coerced) LLVM argument type as
already determined by classifyArgumentType.
llvm-svn: 233543
Support for the QPX vector instruction set, used on the IBM BG/Q supercomputer,
has recently been added to the LLVM PowerPC backend. This vector instruction
set requires some ABI modifications because the ABI on the BG/Q expects
<4 x double> vectors to be provided with 32-byte stack alignment, and to be
handled as native vector types (similar to how Altivec vectors are handled on
mainline PPC systems). I've named this ABI variant elfv1-qpx, have made this
the default ABI when QPX is supported, and have updated the ABI handling code
to provide QPX vectors with the correct stack alignment and associated
register-assignment logic.
llvm-svn: 231960
When passing a type with large alignment byval, we were specifying the type's
alignment rather than the alignment that the backend is actually capable of
producing (ABIAlign).
This would be OK (if odd) assuming the backend dealt with it prooperly,
unfortunately it doesn't and trying to pass types with "byval align 16" can
cause it to set fp incorrectly and trash the stack during the prologue. I'll be
fixing that in a separate patch, but Clang should still be emitting IR that's
as close to its intent as possible.
rdar://20059039
llvm-svn: 231706
Opt in Win64 to supporting sjlj lowering. We have the backend lowering,
so I think this was just an oversight because WinX86_64TargetCodeGenInfo
doesn't inherit from X86_64TargetCodeGenInfo.
llvm-svn: 231280
isSingleElementStruct was a bit too tight in its definition of struct
so we got a mismatch between classify() and the actual code generation.
To make matters worse the code in GetByteVectorType still defaulted to
<2 x double> if it encountered a type it didn't know, making this a
silent miscompilation (PR22753).
Completely remove the "preferred type" stuff from GetByteVectorType and
make it fail an assertion if someone tries to use it with a type not
suitable for a vector register.
llvm-svn: 230971
The backend should now be able to handle all AAPCS rules based on argument
type, which means Clang no longer has to duplicate the register-counting logic
and the CodeGen can be significantly simplified.
llvm-svn: 230349
MSVC does not support C99 _Complex.
ICC, however, does support it on windows x86_64, and treats it, for purposes of parameter passing, as equivalent to a struct containing two fields (for the real and imaginary part).
Differential Revision: http://reviews.llvm.org/D7825
llvm-svn: 230315
llvm.eh.sjlj.setjmp / llvm.eh.sjlj.longjmp, if the backend is known to
support them outside the Exception Handling context. The default
handling in LLVM codegen doesn't work and will create incorrect code.
The ARM backend on the other hand will assert if the intrinsics are
used.
llvm-svn: 230255
This is a patch for PR22563 ( http://llvm.org/bugs/show_bug.cgi?id=22563 ).
We were not correctly unwrapping a single 256-bit AVX vector that was defined as an array of 1 inside a struct.
We would generate a <4 x float> param/return value instead of <8 x float> and lose half of the vector.
Differential Revision: http://reviews.llvm.org/D7614
llvm-svn: 229408
For #pragma comment(linker, ...) MSVC expects the comment string to be quoted, but for #pragma comment(lib, ...) the compiler itself quotes the library name.
Since this distinction disappears by the time the directive reaches the backend, move quoting for the "lib" version to the frontend.
Differential Revision: http://reviews.llvm.org/D7653
llvm-svn: 229376
Create a new TargetCodeGenInfo for Windows on ARM to permit annotating the
functions with stack-probe-size (for /Gs and -mstack-probe-support) for
generating the stack probe necessary for Windows targets. This will be used by
the backend when lowering the frame to generate the stack probe appropriately.
llvm-svn: 227641
Summary:
They just existed before to use NaCl's custom ABIInfos; now that those are gone,
the custom TargetCodeGenInfos are no longer needed either.
Test Plan: don't break the existing tests
Reviewers: jvoung
Subscribers: jfb, cfe-commits
Differential Revision: http://reviews.llvm.org/D7234
llvm-svn: 227406
Summary:
It was used for interoperability with PNaCl's calling conventions, but
it's no longer needed.
Also Remove NaCl*ABIInfo which just existed to delegate to either the portable
or native ABIInfo, and remove checkCallingConvention which was now a no-op
override.
Reviewers: jvoung
Subscribers: jfb, llvm-commits
Differential Revision: http://reviews.llvm.org/D7206
llvm-svn: 227362
Summary:
This fixes MultiSource/Applications/lemon on big-endian N32 by correcting the
handling of the argument to wait(). glibc defines it as a transparent union of
void* and int*. Such unions are passed according to the rules of the first
member so the argument must be passed as if it were a void* (sign extended from
i32 to i64) and not as a union (shifted to the upper bits of an i64).
wait() already behaves correctly on big-endian O32 and N64 since the union is
already the same size as an argument slot.
Reviewers: atanasyan
Reviewed By: atanasyan
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D6963
llvm-svn: 225981
Sorry for the noise, I managed to miss a bunch of recent regressions of
include orderings here. This should actually sort all the includes for
Clang. Again, no functionality changed, this is just a mechanical
cleanup that I try to run periodically to keep the #include lines as
regular as possible across the project.
llvm-svn: 225979
Summary:
The Mips ABI's treat pointers in the same way as integers. They are
sign-extended to 32-bit for O32, and 64-bit for N32/N64. This doesn't matter
for O32 and N64 where pointers are already the correct width but it does matter
for big-endian N32, where pointers are 32-bit and need promoting.
The caller side is already passing pointers correctly. This patch corrects the
callee.
Reviewers: vmedic, atanasyan
Reviewed By: atanasyan
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D6812
llvm-svn: 225782
The extension has the following syntax:
__builtin_call_with_static_chain(Call, Chain)
where Call must be a function call expression and Chain must be of pointer type
This extension performs a function call Call with a static chain pointer
Chain passed to the callee in a designated register. This is useful for
calling foreign language functions whose ABI uses static chain pointers
(e.g. to implement closures).
Differential Revision: http://reviews.llvm.org/D6332
llvm-svn: 224167
ARM ABI specifies that all the libcalls use soft FP ABI
(even hard FP binaries). These days clang emits _mulsc3 / _muldc3
calls with default (C) calling convention which would be translated
into AAPCS_VFP LLVM calling and thus the result of complex
multiplication will be bogus.
Introduce a way for a target to specify explicitly calling
convention for libcalls. Right now this is temporary correctness
fix. Ultimately, we'll end with intrinsic for complex
multiplication and all calling convention decisions for libcalls
will be put into backend.
llvm-svn: 223123
Richard rejected my Sema change to interpret an integer literal zero in
a varargs context as a null pointer, so -Wsentinel sees an integer
literal zero and fires off a warning. Only CodeGen currently knows that
it promotes integer literal zeroes in this context to pointer size on
Windows. I didn't want to teach -Wsentinel about that compatibility
hack. Therefore, I'm migrating to C++11 nullptr.
llvm-svn: 223079
Now that LLVM can count the registers needed to implement AAPCS rules, we don't
need to duplicate that logic here. This means we can drop the explicit padding
and also use more natural types in many cases (e.g. "struct { float arr[3]; }"
used to end up as "[2 x double]" to avoid holes on the stack.
The one wrinkle is that AAPCS va_arg was also using the register counting
machinery. But the local replacement isn't too bad.
llvm-svn: 222904
All supported environments on x86 Windows return structs in EAX:EDX.
This removes code added in r204978 that had to get updated in r222680.
We should now have the same behavior we had before r204978.
llvm-svn: 222697
Cygwin and MinGW fail to conform to the underlying system's structure passing
ABI. Make the check more precise to ensure that we correctly generate code for
the itanium environment.
llvm-svn: 222626
Summary:
With this patch, passing a va_list to another function and reading 10 int's from
it works correctly on a big-endian target.
Based on a pair of patches by David Chisnall, one of which I've reworked
for the current trunk.
Reviewers: theraven, atanasyan
Reviewed By: theraven, atanasyan
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D6248
llvm-svn: 222339
Homogeneous aggregates on AAPCS_VFP ARM need to be passed *without* being
flattened (e.g. [2 x float] rather than "float, float") for various weird ABI
reasons. However, this isn't the case for anything else; further, we know at
the ABIArgInfo::getDirect callsites whether this flattening is allowed.
So, we can get more unified ARM code, with a simpler Clang, by just using that
knowledge directly.
llvm-svn: 221559
The most complex aspect of the convention is the handling of homogeneous
vector and floating point aggregates. Reuse the homogeneous aggregate
classification code that we use on PPC64 and ARM for this.
This convention also has a C mangling, and we apparently implement that
in both Clang and LLVM.
Reviewed By: majnemer
Differential Revision: http://reviews.llvm.org/D6063
llvm-svn: 221006
Reuse the PPC64 HVA detection algorithm for ARM and AArch64. This is a
nice code deduplication, since they are roughly identical. A few virtual
method extension points are needed to understand how big an HVA can be
and what element types it can have for a given architecture.
Also make the record expansion code work in the presence of non-virtual
bases.
Reviewed By: uweigand, asl
Differential Revision: http://reviews.llvm.org/D6045
llvm-svn: 220972
Following the NVVM IR specifications, arguments of aggregate type should be
passed on the stack without splitting (byval).
http://reviews.llvm.org/D6020
Patch by Jacques Pienaar.
llvm-svn: 220854
As discussed in bug 21398, PowerPC ABI code needs to consider C++ base
classes when classifying a class as homogeneous aggregate (or not) for
ABI purposes.
llvm-svn: 220852
Summary:
This allows us to easily identify them in the backend which in turn allows us
to handle them correctly for big-endian targets (where they must be shifted
into the upper bits of the register).
Depends on D5961
Reviewers: atanasyan
Reviewed By: atanasyan
Subscribers: cfe-commits, theraven
Differential Revision: http://reviews.llvm.org/D5962
llvm-svn: 220566
Summary:
Ensure all integral/enumeration types are appropriately annotated with
signext/zeroext. In particular, i32 now has these attributes when using the
N32/N64 ABI. This paves the way for accurately representing the way the
N32/N64 ABI's promotes integer arguments to i64.
Reviewers: atanasyan
Reviewed By: atanasyan
Subscribers: cfe-commits, theraven
Differential Revision: http://reviews.llvm.org/D5961
llvm-svn: 220563
When the aligned clause of an OpenMP simd pragma is not provided with an
explicit alignment, a target-dependent default must be used. This adds such a
default of PPC targets.
This will become slightly more complicated when BG/Q support is added (because
then it will depend on the type). For now, 16 is a correct value for all
systems, and covers Altivec and VSX vectors.
llvm-svn: 218994
Summary:
Currently, with struct my_struct { int x; method_ptr y; };
a call to foo(my_struct s) may end up dropping the last 4 bytes
of the method pointer for x86_64 NaCl and x32.
When checking Has64BitPointers, also check if the method pointer
straddles an eightbyte boundary and classify Hi as well as Lo if needed.
Test Plan: test/CodeGenCXX/x86_64-arguments-nacl-x32.cpp
Reviewers: dschuff, pavel.v.chupin
Subscribers: jfb
Differential Revision: http://reviews.llvm.org/D5555
llvm-svn: 218889
On further investigation, COMDATs should work with .ctors, and the issue
I was hitting probably reproduces with .init_array.
This reverts commit r218287.
llvm-svn: 218313
In particular, pre-.init_array ELF uses the .ctors section mechanism.
MinGW COFF also uses .ctors, now that I think about it. Therefore,
restrict this optimization to the two platforms that are currently known
to work: ELF with .init_array and COFF with .CRT$XCU.
llvm-svn: 218287
Summary:
Vectors are normally 16-byte aligned, however the O32 ABI enforces a
maximum alignment of 8-bytes since the base of the stack is 8-byte aligned.
Previously, this was enforced on the caller side, but not on the callee side.
This fixes the output of OpenCL's printf when given vectors.
Reviewers: atanasyan
Reviewed By: atanasyan
Subscribers: llvm-commits, pekka.jaaskelainen
Differential Revision: http://reviews.llvm.org/D5433
llvm-svn: 218248
The field is defined as:
If the third field is present, non-null, and points to a global variable or function, the initializer function will only run if the associated data from the current module is not discarded.
And without COMDATs we can't implement that.
llvm-svn: 218097
Clang can already handle
-------------------------------------------
struct S {
static const int x;
};
template<typename T> struct U {
static const int k;
};
template<typename T> const int U<T>::k = T::x;
const int S::x = 42;
extern const int *f();
const int *g() { return &U<S>::k; }
int main() {
return *f() + U<S>::k;
}
const int *f() { return &U<S>::k; }
-------------------------------------------
since r217264 which puts the .inint_array section in the same COMDAT
as the variable.
This patch allows the linker to more easily delete some dead code and data by
putting the guard variable and init function in the same COMDAT.
llvm-svn: 218089
If control falls off the end of a function after an __asm block, MSVC
assumes that the inline assembly filled the EAX and possibly EDX
registers with an appropriate return value. This functionality is used
in inline functions returning 64-bit integers in system headers, so we
need some amount of compatibility.
This is implemented in Clang by adding extra output constraints to every
inline asm block, and storing the resulting output registers into the
return value slot. If we see an asm block somewhere in the function
body, we emit a normal epilogue instead of marking the end of the
function with a return type unreachable.
Normal returns in functions not using this functionality will overwrite
the return value slot, and in most cases LLVM should be able to
eliminate the dead stores.
Fixes PR17201.
Reviewed By: majnemer
Differential Revision: http://reviews.llvm.org/D5177
llvm-svn: 217187
Summary:
This allows us to easily find them in the backend after the aggregates have
been lowered to other types. This is important on big-endian targets using
the N32/N64 ABI's since these ABI's must shift small structures into the
upper bits of the register.
Reviewers: atanasyan
Reviewed By: atanasyan
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5005
llvm-svn: 217160
Summary:
They are returned indirectly which causes the other arguments to move to
the next argument slot.
With this, utils/ABITest does not discover any failing cases in the first
500 attempts on big/little endian for O32. Previously some of these failed.
Also tested N32/N64 little endian (big endian has other known issues) with
no issues.
Reviewers: atanasyan
Reviewed By: atanasyan
Subscribers: atanasyan, cfe-commits
Differential Revision: http://reviews.llvm.org/D4811
llvm-svn: 217147
ACLE 2.0 allows __fp16 to be used as a function argument or return
type. This enables this for AArch64.
This also fixes an existing bug that causes clang to not allow
homogeneous floating-point aggregates with a base type of __fp16. This
is valid for AAPCS64, but not for AAPCS-VFP.
llvm-svn: 216558
This tidies up some ARM-specific code added by r208417 to move it out
of the target-independent parts of clang into TargetInfo.cpp. This
also has the advantage that we can now flatten struct arguments to
variadic AAPCS functions.
llvm-svn: 216535
This time though, preserve the extension for bool types since that's compatible
with what MSVC expects.
See http://reviews.llvm.org/D4380
llvm-svn: 216507
Summary:
MSVC doesn't extend integer types smaller than 64bit, so to preserve
binary compatibility, clang shouldn't either.
For example, the following C code built with MSVC:
unsigned test(unsigned v);
unsigned foobar(unsigned short);
int main() { return test(0xffffffff) + foobar(28); }
Produces the following:
0000000000000004: B9 FF FF FF FF mov ecx,0FFFFFFFFh
0000000000000009: E8 00 00 00 00 call test
000000000000000E: 89 44 24 20 mov dword ptr [rsp+20h],eax
0000000000000012: 66 B9 1C 00 mov cx,1Ch
0000000000000016: E8 00 00 00 00 call foobar
And as you can see, when setting up the call to foobar, only cx is overwritten.
If foobar is compiled with clang, then the zero extension added by clang means
the rest of the register, which contains garbage, could be used.
For example if foobar is:
unsigned foobar(unsigned short v) {
return v;
}
Compiled with clang -fomit-frame-pointer -O3 gives the following assembly:
foobar:
0000000000000000: 89 C8 mov eax,ecx
0000000000000002: C3 ret
And that function would return garbage because the 16 most significant bits of
ecx still contain garbage from the first call.
With this change, the code for that function is now:
foobar:
0000000000000000: 0F B7 C1 movzx eax,cx
0000000000000003: C3 ret
Reviewers: chapuni, rnk
Reviewed By: rnk
Subscribers: majnemer, cfe-commits
Differential Revision: http://reviews.llvm.org/D4380
llvm-svn: 216491
It appears that the backend does not handle all cases that were handled by clang.
In particular, it does not handle structs as used in
SingleSource/UnitTests/2003-05-07-VarArgs.
llvm-svn: 214512
Summary:
This patch causes clang to emit va_arg instructions to the backend instead of
expanding them into an implementation itself. The backend already implements
va_arg since this is necessary for NaCl so this patch is removing redundant
code.
Together with the llvm patch (D4556) that accounts for the effect of endianness
on the expansion of va_arg, this fixes PR19612.
Depends on D4556
Reviewers: sstankovic, dsanders
Reviewed By: dsanders
Subscribers: rnk, cfe-commits
Differential Revision: http://reviews.llvm.org/D4742
llvm-svn: 214497
While Clang now supports both ELFv1 and ELFv2 ABIs, their use is currently
hard-coded via the target triple: powerpc64-linux is always ELFv1, while
powerpc64le-linux is always ELFv2.
These are of course the most common scenarios, but in principle it is
possible to support the ELFv2 ABI on big-endian or the ELFv1 ABI on
little-endian systems (and GCC does support that), and there are some
special use cases for that (e.g. certain Linux kernel versions could
only be built using ELFv1 on LE).
This patch implements the Clang side of supporting this, based on the
LLVM commit 214072. The command line options -mabi=elfv1 or -mabi=elfv2
select the desired ABI if present. (If not, Clang uses the same default
rules as now.)
Specifically, the patch implements the following changes based on the
presence of the -mabi= option:
In the driver:
- Pass the appropiate -target-abi flag to the back-end
- Select the correct dynamic loader version (/lib64/ld64.so.[12])
In the preprocessor:
- Define _CALL_ELF to the appropriate value (1 or 2)
In the compiler back-end:
- Select the correct ABI in TargetInfo.cpp
- Select the desired ABI for LLVM via feature (elfv1/elfv2)
llvm-svn: 214074
Previously we were building up the inalloca struct in the usual pattern
of return type followed by arguments. However, on Windows, 'this'
always precedes the 'sret' parameter, so we need to insert it into the
struct first as a special case.
llvm-svn: 213990
The main subtlety here is that the Darwin tools still need to be given "-arch
arm64" rather than "-arch aarch64". Fortunately this already goes via a custom
function to handle weird edge-cases in other architectures, and it tested.
I removed a few arm64_be tests because that really isn't an interesting thing
to worry about. No-one using big-endian is also referring to the target as
arm64 (at least as far as toolchains go). Mostly they date from when arm64 was
a separate target and we *did* need a parallel name simply to test it at all.
Now aarch64_be is sufficient.
llvm-svn: 213744
In addition to enabling ELFv2 homogeneous aggregate handling,
LLVM support to pass array types directly also enables a performance
enhancement. We can now pass (non-homogeneous) aggregates that fit
fully in registers as direct integer arrays, using an element type
to encode the alignment requirement (that would otherwise go to the
"byval align" field).
This is preferable since "byval" forces the back-end to write the
aggregate out to the stack, even if it could be passed fully in
registers. This is particularly annoying on ELFv2, if there is
no parameter save area available, since we then need to allocate
space on the callee's stack just to hold those aggregates.
Note that to implement this optimization, this patch does not attempt
to fully anticipate register allocation rules as (defined in the
ABI and) implemented in the back-end. Instead, the patch is simply
passing *any* aggregate passed by value using the array mechanism
if its size is up to 64 bytes. This means that some of those will
end up being passed in stack slots anyway, but the generated code
shouldn't be any worse either. (*Large* aggregates remain passed
using "byval" to enable optimized copying via memcpy etc.)
llvm-svn: 213495
This patch implements clang support for the PowerPC ELFv2 ABI.
Together with a series of companion patches in LLVM, this makes
clang/LLVM fully usable on powerpc64le-linux.
Most of the ELFv2 ABI changes are fully implemented on the LLVM side.
On the clang side, we only need to implement some changes in how
aggregate types are passed by value. Specifically, we need to:
- pass (and return) "homogeneous" floating-point or vector aggregates in
FPRs and VRs (this is similar to the ARM homogeneous aggregate ABI)
- return aggregates of up to 16 bytes in one or two GPRs
The second piece is trivial to implement in any case. To implement
the first piece, this patch makes use of infrastructure recently
enabled in the LLVM PowerPC back-end to support passing array types
directly, where the array element type encodes properties needed to
handle homogeneous aggregates correctly.
Specifically, the array element type encodes:
- whether the parameter should be passed in FPRs, VRs, or just
GPRs/stack slots (for float / vector / integer element types,
respectively)
- what the alignment requirements of the parameter are when passed in
GPRs/stack slots (8 for float / 16 for vector / the element type
size for integer element types) -- this corresponds to the
"byval align" field
With this support in place, the clang part simply needs to *detect*
whether an aggregate type implements a float / vector homogeneous
aggregate as defined by the ELFv2 ABI, and if so, pass/return it
as array type using the appropriate float / vector element type.
llvm-svn: 213494
r211898 introduced a regression where a large struct, which would
normally be passed ByVal, was causing padding to be inserted to
prevent the backend from using some GPRs, in order to follow the
AAPCS. However, the type of the argument was not being set correctly,
so the backend cannot align 8-byte aligned struct types on the stack.
The fix is to not insert the padding arguments when the argument is
being passed ByVal.
llvm-svn: 213359
This patch adds support for respecting the ABI and type alignment
of aggregates passed by value. Currently, all aggregates are aligned
at 8 bytes in the parameter save area. This is incorrect for two
reasons:
- Aggregates that need alignment of 16 bytes or more should be aligned
at 16 bytes in the parameter save area. This is implemented by
using an appropriate "byval align" attribute in the IR.
- Aggregates that need alignment beyond 16 bytes need to be dynamically
realigned by the caller. This is implemented by setting the Realign
flag of the ABIArgInfo::getIndirect call.
In addition, when expanding a va_arg call accessing a type that is
aligned at 16 bytes in the argument save area (either one of the
aggregate types as above, or a vector type which is already aligned
at 16 bytes), code needs to align the va_list pointer accordingly.
Reviewed by Hal Finkel.
llvm-svn: 212743
This patch adds support for passing arguments of non-Altivec vector type
(i.e. defined via attribute ((vector_size (...)))) on powerpc64-linux.
While such types are not mentioned in the formal ABI document, this
patch implements a calling convention compatible with GCC:
- Vectors of size < 16 bytes are passed in a GPR
- Vectors of size > 16 bytes are passed via reference
Note that vector types with a number of elements that is not a power
of 2 are not supported by GCC, so there is no pre-existing ABI to
follow. We choose to pass those (of size < 16) as if widened to the
next power of two, so they might end up in a vector register or
in a GPR. (Sizes > 16 are always passed via reference as well.)
Reviewed by Hal Finkel.
llvm-svn: 212734
The sret paramater consumes the register after the implicit 'this'
parameter, as with other calling conventions.
Fixes PR20278, which turned out to be very easy.
llvm-svn: 212669
Derek Schuff