| Commit message (Collapse) | Author | Age | Files | Lines |
| ... | |
| |
|
|
|
|
|
| |
This usually happens when TempScopInfo calls isMaxRegionInScop(..),
so don't fail there.
llvm-svn: 215624
|
| |
|
|
| |
llvm-svn: 215623
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
input node after manually adding it to the worklist and using CombineTo.
Once we use CombineTo the input node may have been deleted. Despite this
being *completely confusing* and somewhat broken, the only way to
"correctly" return from a DAG combine after potentially deleting the
input node is to return *that exact node*....
But really, this code should just never have used CombineTo. It won't do
what it wants (returning the node as mentioned above just causes the
combine to infloop). The correct way to combine away a casted load to
a load of the correct type is to RAUW the chain directly and then return
the loaded value to replace the actual value node.
I managed to find this with the vector shuffle fuzzer even though it
clearly has nothing at all to do with vector shuffles and rather those
happen to trigger a load of a constant pool that hits this combine *just
right*. I've included the test as it is small and a nice stress test
that the infrastructure isn't asserting.
llvm-svn: 215622
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Proof using CVC3 follows:
$ cat t.cvc
A, B : BITVECTOR(32);
QUERY BVXOR((A | ~B),(~A |B)) = BVXOR(A,B);
$ cvc3 t.cvc
Valid.
Patch by Mayur Pandey!
Differential Revision: http://reviews.llvm.org/D4883
llvm-svn: 215621
|
| |
|
|
|
|
| |
GCC was emitting a signed vs unsigned comparison warning.
llvm-svn: 215620
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Transform ((B | C) & A) | B --> B | (A & C)
Z3 Link: http://rise4fun.com/Z3/hP6p
Patch by Sonam Kumari!
Differential Revision: http://reviews.llvm.org/D4865
llvm-svn: 215619
|
| |
|
|
|
|
|
|
|
|
| |
It doesn't really make sense to try and do stuff with #pragma init_seg
when targeting non-Microsoft platforms; notions like library vs user
initializers don't exist for other targets.
This fixes PR20639.
llvm-svn: 215618
|
| |
|
|
|
|
|
| |
avoids users of AllDiagnostics.h from needing to pregenerate *all* generated
headers. Hopefully this will make my modules buildbot happier...
llvm-svn: 215617
|
| |
|
|
|
|
|
| |
declaration has its definition instantiated in two sibling modules and they use
a partial specialization.
llvm-svn: 215616
|
| |
|
|
|
|
|
|
|
|
|
|
| |
As X86MCAsmInfoDarwin uses '##' as CommentString although a single '#' starts a
comment a workaround for this special case is added.
Fixes divisions in constant expressions for the AArch64 assembler and other
targets which use '//' as CommentString.
Patch by Janne Grunau!
llvm-svn: 215615
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently when laying out bitfields that don't need any padding, we
represent them as a wide enough int to contain all of the bits. This
can be hard on the backend since we'll do things like represent stores
to a few bits as loading an i144, masking it with a large constant,
and storing it back.
This turns up in less pathological cases where we load and mask 64 bit
word on a 32 bit platform when we actually only need to access 32 bits.
This leads to bad code being generated in most of our 32 bit backends.
In practice, there are often natural breaks in bitfields, and it's a
fairly simple and effective heuristic to split these fields into legal
integer sized chunks when it will be equivalent (ie, it won't force us
to add any extra padding).
llvm-svn: 215614
|
| |
|
|
|
|
| |
Patch by Anthony Pesch. Thanks Anthony!
llvm-svn: 215613
|
| |
|
|
|
|
|
|
|
|
| |
definitions (because some other declaration declares a special member that
isn't present in the canonical definition), we need to search *all* of them; we
can't just stop when we find the requested name in any of the definitions,
because that can fail to find things (and in particular, it can fail to find
the member of the canonical declaration and return a bogus ODR failure).
llvm-svn: 215612
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
combining by replacing it with something else but not re-process the
node afterward to remove it.
In a truly remarkable stroke of bad luck, this would (in the test case
attached) end up getting some other node combined into it without ever
getting re-processed. By adding it back on to the worklist, in addition
to deleting the dead nodes more quickly we also ensure that if it
*stops* being dead for any reason it makes it back through the
legalizer. Without this, the test case will end up failing during
instruction selection due to an and node with a type we don't have an
instruction pattern for.
It took many million runs of the shuffle fuzz tester to find this.
llvm-svn: 215611
|
| |
|
|
|
|
| |
users.
llvm-svn: 215610
|
| |
|
|
|
|
| |
This fixes a regression I caused back in r211766.
llvm-svn: 215609
|
| |
|
|
|
|
|
|
| |
embedded python interpreter.
<rdar://problem/17949057>
llvm-svn: 215608
|
| |
|
|
| |
llvm-svn: 215607
|
| |
|
|
| |
llvm-svn: 215606
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In general two-level namespace means each program records exactly which dylib
each undefined (imported) symbol comes from. But, sometimes the implementor
wants to hide the implementation dylib. For instance libSytem.dylib is the base
dylib all Darwin programs must link with. A few years ago it was split up
into two dozen dylibs by all are hidden behind libSystem.dylib which re-exports
each sub-dylib. All clients still think libSystem.dylib is the implementor.
To support this, the linker must load "indirect" dylibs and not just the
"direct" dylibs specified on the command line. This is done in the
createImplicitFiles() method after all command line specified files are
loaded. Since an indirect dylib may have already been loaded as a direct dylib
(or indirectly via a previous direct dylib), the MachOLinkingContext keeps
a list of all loaded dylibs.
With this change hello world can now be linked against the real OS or SDK.
llvm-svn: 215605
|
| |
|
|
|
|
| |
Found by code inspection.
llvm-svn: 215604
|
| |
|
|
|
|
| |
found in global pool as well. rdar://16808765
llvm-svn: 215603
|
| |
|
|
|
|
|
|
|
|
| |
Split up the CRuntimeFile into one part for output types that need an entry
point and another part for output types that use stubs.
Add file 'test/mach-o/Inputs/libSystem.yaml' for use by test cases that
use -dylib and therefore may now need the helper symbol in libSystem.dylib.
llvm-svn: 215602
|
| |
|
|
|
|
| |
Sorry about the noise.
llvm-svn: 215601
|
| |
|
|
|
|
|
|
|
|
| |
Certain functions such as objc_autoreleaseReturnValue have to be called as
tail-calls even at -O0. Since normal fast-isel doesn't emit calls as tail calls,
we have to fall back to SelectionDAG to select calls that are marked as tail.
<rdar://problem/17991614>
llvm-svn: 215600
|
| |
|
|
|
|
|
| |
Implement __pld, __pldx, __pli and __plix builtin intrinsics as specified in
ARM ACLE 2.0.
llvm-svn: 215599
|
| |
|
|
|
|
|
|
|
|
|
| |
Mach-o uses "two-level namespace" where each undefined symbols is associated
with a specific dylib. This means at runtime the loader (dyld) does need to
search all loaded dylibs for that symbol but rather just the one specified.
Now that llvm-nm -m prints out that info, properly set that info, and test
it in the hello world test cases.
llvm-svn: 215598
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
FastISel didn't take much advantage of the different addressing modes available
to it on AArch64. This commit allows the ComputeAddress method to recognize more
addressing modes that allows shifts and sign-/zero-extensions to be folded into
the memory operation itself.
For Example:
lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3]
ldr x0, [x0, x1]
sxtw x1, w1
lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3]
ldr x0, [x0, x1]
llvm-svn: 215597
|
| |
|
|
|
|
| |
Use the right mechanism to put the XPC services now that Xcode supports the workflow.
llvm-svn: 215596
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
floating-point constants.
In the large code model for X86 floating-point constants are placed in the
constant pool and materialized by loading from it. Since the constant pool
could be far away, a PC relative load might not work. Therefore we first
materialize the address of the constant pool with a movabsq and then load
from there the floating-point value.
Fixes <rdar://problem/17674628>.
llvm-svn: 215595
|
| |
|
|
| |
llvm-svn: 215594
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
materialization.
This mostly affects the i64 value type, which always resulted in an 15byte
mobavsq instruction to materialize any constant. The custom code checks the
value of the immediate and tries to use a different and smaller mov
instruction when possible.
This fixes <rdar://problem/17420988>.
llvm-svn: 215593
|
| |
|
|
| |
llvm-svn: 215592
|
| |
|
|
|
|
|
|
|
|
| |
This change materializes now the value "0" from the zero register.
The zero register can be folded by several instruction, so no
materialization is need at all.
Fixes <rdar://problem/17924413>.
llvm-svn: 215591
|
| |
|
|
|
| |
FIXME: It seems this might be incompatible to dos path. Investigating.
llvm-svn: 215590
|
| |
|
|
| |
llvm-svn: 215589
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This changes the order in which FastISel tries to materialize a constant.
Originally it would try to use a simple target-independent approach, which
can lead to the generation of inefficient code.
On X86 this would result in the use of movabsq to materialize any 64bit
integer constant - even for simple and small values such as 0 and 1. Also
some very funny floating-point materialization could be observed too.
On AArch64 it would materialize the constant 0 in a register even the
architecture has an actual "zero" register.
On ARM it would generate unnecessary mov instructions or not use mvn.
This change simply changes the order and always asks the target first if it
likes to materialize the constant. This doesn't fix all the issues
mentioned above, but it enables the targets to implement such
optimizations.
Related to <rdar://problem/17420988>.
llvm-svn: 215588
|
| |
|
|
|
|
|
|
| |
This is a cleaner solution to the problem described in r215431.
When instructions are combined a dangling DBG_VALUE is removed.
This resolves bug 20598.
llvm-svn: 215587
|
| |
|
|
|
|
|
| |
Split the constant materialization code into three separate helper functions for
Integer-, Floating-Point-, and GlobalValue-Constants.
llvm-svn: 215586
|
| |
|
|
|
|
|
| |
The tests in r215568 hard code a value as %0 in their checks. This
isn't correct in asserts builds.
llvm-svn: 215585
|
| |
|
|
|
|
|
|
| |
This change is also in preparation for a future change to make sure that
the constant materialization uses MOVT/MOVW when available and not a load
from the constant pool.
llvm-svn: 215584
|
| |
|
|
|
|
|
|
|
|
|
|
| |
getRegClassFor returns the incorrect register class when in Thumb2 mode.
This fix simply manually selects the register class as in the code just a few
lines above.
There is no test case for this code, because the code is currently
unreachable. This will be changed in a future commit and existing test
cases will exercise this code.
llvm-svn: 215583
|
| |
|
|
|
|
| |
Cleanup and prepare constant materialization code for future commits.
llvm-svn: 215582
|
| |
|
|
|
|
| |
// rdar://16808765
llvm-svn: 215581
|
| |
|
|
|
|
|
|
|
|
| |
New function to erase a machine instruction and mark DBG_VALUE
for removal. A DBG_VALUE is marked for removal when it references
an operand defined in the instruction.
Use the new function to cleanup code in dead machine instruction
removal pass.
llvm-svn: 215580
|
| |
|
|
|
|
|
|
| |
recursively within the emission of another inline function. This ultimately
led to us emitting the same inline function definition twice, which we then
rejected because we believed we had a mangled name conflict.
llvm-svn: 215579
|
| |
|
|
|
|
| |
MSVC doesn't define __func__.
llvm-svn: 215578
|
| |
|
|
|
|
|
| |
expression to the best method found in global method pools.
This is wip. // rdar://16808765
llvm-svn: 215577
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
critical edge has been split. The MachineDominatorTree will when lazy update the
underlying dominance properties when require.
** Context **
This is a follow-up of r215410.
Each time a critical edge is split this invalidates the dominator tree
information. Thus, subsequent queries of that interface will be slow until the
underlying information is actually recomputed (costly).
** Problem **
Prior to this patch, splitting a critical edge needed to query the dominator
tree to update the dominator information.
Therefore, splitting a bunch of critical edges will likely produce poor
performance as each query to the dominator tree will use the slow query path.
This happens a lot in passes like MachineSink and PHIElimination.
** Proposed Solution **
Splitting a critical edge is a local modification of the CFG. Moreover, as soon
as a critical edge is split, it is not critical anymore and thus cannot be a
candidate for critical edge splitting anymore. In other words, the predecessor
and successor of a basic block inserted on a critical edge cannot be inserted by
critical edge splitting.
Using these observations, we can pile up the splitting of critical edge and
apply then at once before updating the DT information.
The core of this patch moves the update of the MachineDominatorTree information
from MachineBasicBlock::SplitCriticalEdge to a lazy MachineDominatorTree.
** Performance **
Thanks to this patch, the motivating example compiles in 4- minutes instead of
6+ minutes. No test case added as the motivating example as nothing special but
being huge!
The binaries are strictly identical for all the llvm test-suite + SPECs with and
without this patch for both Os and O3.
Regarding compile time, I observed only noise, although on average I saw a
small improvement.
<rdar://problem/17894619>
llvm-svn: 215576
|
| |
|
|
|
|
|
| |
This has been hiding really well. Hopefully brings the builders suffering from
outdated lit.site.cfg files back to life.
llvm-svn: 215575
|