Commit Graph

127 Commits

Author SHA1 Message Date
Duncan Sands
d102593b42 Use getPreferredAlignmentLog or getPreferredAlignment
to get the alignment of global variables, rather than
using hand-made versions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@46495 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-29 06:23:44 +00:00
Chris Lattner
9c4428b217 Fix PR1845 and rdar://5676945. Generic vectors smaller
than hardware supported type will be scalarized, so we
can infer their alignment from that info.

We now codegen pr1845 into:

_boolVectorSelect:
	lbz r2, 0(r3)
	stb r2, -16(r1)
	blr 



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45796 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-10 00:30:57 +00:00
Chris Lattner
4ee451de36 Remove attribution from file headers, per discussion on llvmdev.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45418 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-29 20:36:04 +00:00
Duncan Sands
5819799fa7 Fix a brain fart by our beloved leader (the content
of this patch is the last line).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45289 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-21 20:18:41 +00:00
Duncan Sands
67f1c493d1 Remove host endianness info from TargetData and
put it in a new header System/Host.h instead.
Instead of getting the endianness from configure,
calculate it directly.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44959 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-12 23:03:45 +00:00
Chris Lattner
57d7d3f2d7 Move TargetData::hostIsLittleEndian out of line, which means we
don't have to #include config.h in it.  #including config.h breaks
other projects that have their own autoconf stuff and try to #include
the llvm headers.  One obscure example is llvm-gcc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44825 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-11 00:28:59 +00:00
Anton Korobeynikov
b623ce9128 Silence a warning
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43954 91177308-0d34-0410-b5e6-96231b3b80d8
2007-11-09 19:06:14 +00:00
Duncan Sands
514ab348fd Executive summary: getTypeSize -> getTypeStoreSize / getABITypeSize.
The meaning of getTypeSize was not clear - clarifying it is important
now that we have x86 long double and arbitrary precision integers.
The issue with long double is that it requires 80 bits, and this is
not a multiple of its alignment.  This gives a primitive type for
which getTypeSize differed from getABITypeSize.  For arbitrary precision
integers it is even worse: there is the minimum number of bits needed to
hold the type (eg: 36 for an i36), the maximum number of bits that will
be overwriten when storing the type (40 bits for i36) and the ABI size
(i.e. the storage size rounded up to a multiple of the alignment; 64 bits
for i36).

This patch removes getTypeSize (not really - it is still there but
deprecated to allow for a gradual transition).  Instead there is:

(1) getTypeSizeInBits - a number of bits that suffices to hold all
values of the type.  For a primitive type, this is the minimum number
of bits.  For an i36 this is 36 bits.  For x86 long double it is 80.
This corresponds to gcc's TYPE_PRECISION.

(2) getTypeStoreSizeInBits - the maximum number of bits that is
written when storing the type (or read when reading it).  For an
i36 this is 40 bits, for an x86 long double it is 80 bits.  This
is the size alias analysis is interested in (getTypeStoreSize
returns the number of bytes).  There doesn't seem to be anything
corresponding to this in gcc.

(3) getABITypeSizeInBits - this is getTypeStoreSizeInBits rounded
up to a multiple of the alignment.  For an i36 this is 64, for an
x86 long double this is 96 or 128 depending on the OS.  This is the
spacing between consecutive elements when you form an array out of
this type (getABITypeSize returns the number of bytes).  This is
TYPE_SIZE in gcc.

Since successive elements in a SequentialType (arrays, pointers
and vectors) need to be aligned, the spacing between them will be
given by getABITypeSize.  This means that the size of an array
is the length times the getABITypeSize.  It also means that GEP
computations need to use getABITypeSize when computing offsets.
Furthermore, if an alloca allocates several elements at once then
these too need to be aligned, so the size of the alloca has to be
the number of elements multiplied by getABITypeSize.  Logically
speaking this doesn't have to be the case when allocating just
one element, but it is simpler to also use getABITypeSize in this
case.  So alloca's and mallocs should use getABITypeSize.  Finally,
since gcc's only notion of size is that given by getABITypeSize, if
you want to output assembler etc the same as gcc then getABITypeSize
is the size you want.

Since a store will overwrite no more than getTypeStoreSize bytes,
and a read will read no more than that many bytes, this is the
notion of size appropriate for alias analysis calculations.

In this patch I have corrected all type size uses except some of
those in ScalarReplAggregates, lib/Codegen, lib/Target (the hard
cases).  I will get around to auditing these too at some point,
but I could do with some help.

Finally, I made one change which I think wise but others might
consider pointless and suboptimal: in an unpacked struct the
amount of space allocated for a field is now given by the ABI
size rather than getTypeStoreSize.  I did this because every
other place that reserves memory for a type (eg: alloca) now
uses getABITypeSize, and I didn't want to make an exception
for unpacked structs, i.e. I did it to make things more uniform.
This only effects structs containing long doubles and arbitrary
precision integers.  If someone wants to pack these types more
tightly they can always use a packed struct.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43620 91177308-0d34-0410-b5e6-96231b3b80d8
2007-11-01 20:53:16 +00:00
Chris Lattner
b361ec3197 Fix PR1749 and InstCombine/2007-10-28-EmptyField.ll by handling
zero-length fields better.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43427 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-29 02:40:02 +00:00
Dan Gohman
8e41ed7c4e Simplify getIntPtrType, allowing it to work for arbitrary pointer sizes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42751 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-08 15:16:25 +00:00
Dale Johannesen
3b5b4cd1a5 Add getABITypeSize, getABITypeSizeInBits
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42488 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-01 16:03:14 +00:00
Rafael Espindola
a37ac9ff53 Don't add a default STACK_ALIGN (use the generic ABI alignment)
Implement calls to functions with byval arguments on X86


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42192 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-21 15:50:22 +00:00
Chris Lattner
76c1b97e40 Merge DenseMapKeyInfo & DenseMapValueInfo into DenseMapInfo
Add a new DenseMapInfo::isEqual method to allow clients to redefine
the equality predicate used when probing the hash table.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42042 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-17 18:34:04 +00:00
Rafael Espindola
588af2fb99 Add support for having different alignment for objects on call frames.
The x86-64 ABI states that objects passed on the stack have
8 byte alignment. Implement that.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@41768 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-07 14:52:14 +00:00
Reid Spencer
181b6c9cb5 Fix minor doxygen nits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40854 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-05 20:06:04 +00:00
Dale Johannesen
8c1e6a119a long double patch 2 of N. Handle it in TargetData.
(I've tried to get the info right for all targets,
but I'm not expert on all of them - check yours.)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40792 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-03 20:20:50 +00:00
Devang Patel
1997473cf7 Drop 'const'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36662 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-03 01:11:54 +00:00
Devang Patel
3e15bf33e0 Use 'static const char' instead of 'static const int'.
Due to darwin gcc bug, one version of darwin linker coalesces
static const int, which defauts PassID based pass identification.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36652 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-02 21:39:20 +00:00
Lauro Ramos Venancio
c718288f49 Fix build error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36648 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-02 20:37:47 +00:00
Devang Patel
794fd75c67 Do not use typeinfo to identify pass in pass manager.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36632 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-01 21:15:47 +00:00
Christopher Lamb
a7f0d3ce22 Support alignment queries for degenerate (length 1) vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36352 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-22 21:54:13 +00:00
Jeff Cohen
44f5fb46b0 When the number of elements is zero, don't malloc 32GB on 64-bit systems.
Fixes unexpected failures on FreeBSD/amd64 of:
  CFrontend/2005-09-24-BitFieldCrash.c:
  CFrontend/2007-02-04-EmptyStruct.c:
  CFrontend/2007-03-26-ZeroWidthBitfield.c:
  CodeGen/Generic/2005-10-18-ZeroSizeStackObject.ll:


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35828 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-09 19:26:30 +00:00
Jeff Cohen
ca5183d445 Unbreak VC++ build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34917 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-05 00:00:42 +00:00
Reid Spencer
1ede29c951 Wrap a long line.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34799 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-01 19:48:16 +00:00
Reid Spencer
4ecd9ee061 Simplify some code by moving variable declarations into the only block that
uses them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34432 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-19 23:30:10 +00:00
Reid Spencer
f734ea21a3 Implement support for non-standard integer bit widths of any size. The
rules alignment is to pick the alignment that corresponds to the smallest
specified alignment that is larger than the bit width of the type or the
largest specified integer alignment if none are larger than the bitwidth
of the type. For the byte size, the size returned is the next larger
multiple of the alignment for that type (using the above rule). This patch
also changes bit widths from "short" to "uint32_t" to ensure there are
enough bits to specify any bit width that LLVM can handle (currently 2^23);
16-bits isn't enough.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34431 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-19 22:35:00 +00:00
Chris Lattner
affeb56480 Do not dereference invalid ranges. Generalize targetdata alignment model.
This fixes the UnitTests/Vector/sumarray-dbl regressions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34358 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-17 00:41:42 +00:00
Chris Lattner
3ebb2e0034 Fix CodeGen/PowerPC/2007-02-16-AlignPacked.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34356 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-16 23:11:51 +00:00
Reid Spencer
52dc96842e Remove an unnecessary predicate.
Patch by Scott Michel.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34354 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-16 22:42:40 +00:00
Chris Lattner
ae6f1fa36b simplify some code, ensure that packed structures get abi alignment of 1.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34352 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-16 22:25:34 +00:00
Reid Spencer
1f43787bf0 For PR1195:
PACKED_ALIGN -> VECTOR_ALIGN


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34330 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-15 22:07:05 +00:00
Reid Spencer
16db3caba8 For PR1202:
Make sure we found an existing Alignment before overwriting it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34308 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-15 18:34:36 +00:00
Reid Spencer
9d6565a5b1 For PR1195:
Rename PackedType -> VectorType, ConstantPacked -> ConstantVector, and
PackedTyID -> VectorTyID. No functional changes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34293 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-15 02:26:10 +00:00
Reid Spencer
b7d61101b1 Fixed packed structure breakage from earlier TargetData patch; applied
Chris Lattner's code style suggestions.

Patch by Scott Michel!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34292 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-15 02:11:06 +00:00
Chris Lattner
d2b7cec527 Generalize TargetData strings, to support more interesting forms of data.
Patch by Scott Michel.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34266 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-14 05:52:17 +00:00
Chris Lattner
f6ca09a22b Switch LayoutInfo to be a DenseMap instead of an std::map. This speeds up
-load-vn -gcse by 2.3%.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34160 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 20:26:17 +00:00
Chris Lattner
a12bd03f07 add a typedef
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34159 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 20:18:06 +00:00
Chris Lattner
9182e3f205 eliminate the std::vector from StructLayout, allocating the elements immediately
after the StructLayout object in memory.  This marginally improves locality,
speeding up -load-vn -gcse by ~0.8%.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34158 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 20:15:41 +00:00
Chris Lattner
b0c39a3b4d encapsulate the rest of the StructLayout members.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34157 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 19:59:22 +00:00
Chris Lattner
b1919e2f08 Privatize StructLayout::MemberOffsets, adding an accessor
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34156 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 19:55:17 +00:00
Chris Lattner
ec6478b8f3 Use ManagedStatic to manage LayoutInfo, instead of rolling our own.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34154 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 19:43:18 +00:00
Chris Lattner
ddce8d21ea Change TargetData::getIndexedOffset interface to not require indices
in a vector.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34153 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-10 19:33:15 +00:00
Reid Spencer
d2a988caad Although targets are not required to support integers > 64bits, TargetData
must in order for backends that do want to support large integer types to be
able to function. Consequently, don't assert if the bitwidth > 64 bits
when computing the size and alignment. Instead, compute the size by rounding
up to the next even number of bytes for the size. Compute the alignment
as the same as the LongABIAlignment. These provide reasonable defaults
that the target can override.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33943 91177308-0d34-0410-b5e6-96231b3b80d8
2007-02-05 23:51:43 +00:00
Evan Cheng
dff1ab2001 Dead comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33719 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-31 21:31:25 +00:00
Reid Spencer
26f238589f For PR761:
The Module::setEndianness and Module::setPointerSize methods have been
removed. Instead you can get/set the DataLayout. Adjust thise accordingly.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33530 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-26 08:11:39 +00:00
Evan Cheng
de268f7dcf Renamed getTypeAlignmentShift() to getPreferredTypeAlignmentShift().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33482 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-24 07:03:39 +00:00
Evan Cheng
778900a5d1 - getTypeAlignmentShift() should be returning preferred alignment, not ABI
alignment.
- getPreferredAlignmentLog(): remove Double special case.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33445 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-22 23:08:19 +00:00
Reid Spencer
7c29243670 Implement a getTypeSizeInBits method. This helps in transforms that want
to ensure the bit size of a type is identical before proceeding.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33413 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-20 23:32:04 +00:00
Owen Anderson
1027a533d4 TargetData assumes (and some regression tests depend on it) that the size of
an unspecified datatype in the datalayout is capped by the size of a pointer.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33411 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-20 23:07:13 +00:00
Chris Lattner
4a8c32debf trivial cleanup
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@33410 91177308-0d34-0410-b5e6-96231b3b80d8
2007-01-20 22:39:15 +00:00