Tom Stellard [Wed, 23 Oct 2013 00:44:32 +0000 (00:44 +0000)]
R600: Fix handling of vector kernel arguments
The SelectionDAGBuilder was promoting vector kernel arguments to legal
types, but this won't work for R600 and SI since kernel arguments are
stored in memory and can't be promoted. In order to handle vector
arguments correctly we need to look at the original types from the LLVM IR
function.
Manman Ren [Tue, 22 Oct 2013 20:59:19 +0000 (20:59 +0000)]
Debug Info: code clean up.
Remove unnecessary creation of LexicalScope in collectDeadVariables.
The created LexicialScope was only used to get isAbstractScope, which
should be false from the creation:
"new LexicalScope(NULL, DIDescriptor(SP), NULL, false);".
We can also remove a DenseMap that holds the created LexicalScopes.
David Blaikie [Tue, 22 Oct 2013 20:28:55 +0000 (20:28 +0000)]
DIEHashing: Provide an assert for unreachable functionality regarding friends.
Since (as of r190716) Clang no longer emits debug info for C++ friend
declarations (and it seems GCC never has/does, which was the motivation
for the Clang change), there's no actual reachable case for implementing
the part of DWARF 4, Section 7.27 part 5 that pertains to friends.
Leave an assert here so that if/when we do have a client producing
friends and using type units, we can fill in the gap and add appropriate
(unit and feature) tests.
Tim Northover [Tue, 22 Oct 2013 19:00:39 +0000 (19:00 +0000)]
ARM: provide diagnostics on more writeback LDM/STM instructions
The set of circumstances where the writeback register is allowed to be in the
list of registers is rather baroque, but I think this implements them all on
the assembly parsing side.
For disassembly, we still warn about an ARM-mode LDM even if the architecture
revision is < v7 (the required architecture information isn't available). It's
a silly instruction anyway, so hopefully no-one will mind.
Tom Stellard [Tue, 22 Oct 2013 18:19:10 +0000 (18:19 +0000)]
R600: Simplify handling of private address space
The AMDGPUIndirectAddressing pass was previously responsible for
lowering private loads and stores to indirect addressing instructions.
However, this pass was buggy and way too complicated. The only
advantage it had over the new simplified code was that it saved one
instruction per direct write to private memory. This optimization
likely has a minimal impact on performance, and we may be able
to duplicate it using some other transformation.
For the private address space, we now:
1. Lower private loads/store to Register(Load|Store) instructions
2. Reserve part of the register file as 'private memory'
3. After regalloc lower the Register(Load|Store) instructions to
MOV instructions that use indirect addressing.
David Blaikie [Tue, 22 Oct 2013 18:14:41 +0000 (18:14 +0000)]
DWARF type hashing: pointers to members
Includes a test case/FIXME demonstrating a bug/limitation in pointer to
member hashing. To be honest I'm not sure why we don't just always use
summary hashing for referenced types... but perhaps I'm missing
something.
Bob Wilson [Tue, 22 Oct 2013 17:43:47 +0000 (17:43 +0000)]
Fix llvm-cov counts to be 64-bit integers to avoid overflows.
Line counts in llvm-cov are read in as 64-bit integers but were being truncated
to 32-bit in collectLineCounts(), which caused overflow for large counts.
This patch fixes all counts to be uint64_t.
Wan Xiaofei [Tue, 22 Oct 2013 08:02:02 +0000 (08:02 +0000)]
Using FoldingSet in SelectionDAG::getVTList.
VTList has a long life cycle through the module and getVTList is frequently called. In current getVTList, sequential search over a std::vector is used, this is inefficient in big module.
This patch use FoldingSet to implement hashing mechanism when searching.
Reviewer: Nadav Rotem
Test : Pass unit tests & LNT test suite
Bob Wilson [Tue, 22 Oct 2013 05:09:41 +0000 (05:09 +0000)]
Change llvm-cov output formatting to be more similar to gcov.
- Replaced tabs with proper padding
- print() takes two arguments, which are the GCNO and GCDA filenames
- Files are listed at the top of output, appended by line 0
- Stripped strings of trailing \0s
- Removed last two lines of whitespace in output
Bob Wilson [Tue, 22 Oct 2013 00:09:03 +0000 (00:09 +0000)]
Move the printing of llvm-cov information out from collectLineCounts().
collectLineCounts() should only organize the output data. This is done in
anticipation of subsequent changes which will pass in GCNO and GCDA filenames
into the print function where it is printed similar to the gcov output.
David Blaikie [Mon, 21 Oct 2013 18:59:40 +0000 (18:59 +0000)]
DWARF type hashing: Handle multiple (including recursive) references to the same type
This uses a map, keeping the type DIE numbering separate from the DIEs
themselves - alternatively we could do things the way GCC does if we
want to add an integer to the DIE type to record the numbering there.
Lang Hames [Mon, 21 Oct 2013 17:51:24 +0000 (17:51 +0000)]
X86 vector element shift-by-immediate instructions take i8 immediates. Make
the instruction defenitions and ISEL reflect this.
Prior to this patch these instructions took an i32i8imm, and the high bits were
dropped during encoding. This led to incorrect behavior for shifts by
immediates higher than 255. This patch fixes that issue by detecting large
immediate shifts and returning constant zero (for logical shifts) or capping
the shift amount at an encodable value (for arithmetic shifts).
Rafael Espindola [Mon, 21 Oct 2013 17:14:55 +0000 (17:14 +0000)]
Optimize more linkonce_odr values during LTO.
When a linkonce_odr value that is on the dso list is not unnamed_addr
we can still look to see if anything is actually using its address. If
not, it is safe to hide it.
This patch implements that by moving GlobalStatus to Transforms/Utils
and using it in Internalize.
David Blaikie [Mon, 21 Oct 2013 16:37:22 +0000 (16:37 +0000)]
DebugInfo: Hash DW_FORM_GNU_str_index as a string.
Found while adding type safety to the various DWARF enumerations (form,
attribute, tag, etc) that caused Clang to warn on an incompletely
covered switch. Converting the comment to a default/unreachable
uncovered this case of an unsupported form encoding. Seems we were
skipping fission strings entirely.
Matheus Almeida [Mon, 21 Oct 2013 12:26:50 +0000 (12:26 +0000)]
[mips][msa] Direct Object Emission support for CTCMSA and CFCMSA.
These instructions are logically related as they allow read/write of MSA control registers.
Currently MSA control registers are emitted by number but hopefully that will change as soon
as GAS starts accepting them by name as that would make the assembly easier to read.
Bill Wendling [Mon, 21 Oct 2013 04:09:17 +0000 (04:09 +0000)]
Don't eliminate a partially redundant load if it's in a landing pad.
A landing pad can be jumped to only by the unwind edge of an invoke
instruction. If we eliminate a partially redundant load in a landing pad, it
will create a basic block that violates this constraint. It then leads to other
problems down the line if it tries to merge that basic block with the landing
pad. Avoid this by not eliminating the load in a landing pad.
Teach simplify-cfg how to correctly create covered lookup tables for switches on iN with N >= 3.
One optimization simplify-cfg performs is the converting of switches to
lookup tables if the switch has > 4 cases. This is done by:
1. Finding the max/min case value and calculating the switch case range.
2. Create a lookup table basic block.
3. Perform a check in the switch's BB to see if the input value is in
the switch's case range. If the input value satisfies said predicate
branch to the lookup table BB, otherwise branch to the switch's default
destination BB using the default value as the result.
The conditional check consists of subtracting the min case value of the
table from any input iN value and then ensuring that said value is
unsigned less than the size of the lookup table represented as an iN
value.
If the lookup table is a covered lookup table, the size of the table will be N
which is 0 as an iN value. Thus the comparison will be an `icmp ult` of an iN
value against 0 which is always false yielding the incorrect result.
This patch fixes this problem by recognizing if we have a covered lookup table
and if we do, unconditionally jumps to the lookup table BB since the covering
property of the lookup table implies no input values could not be handled by
said BB.
This ensures that the prefix data is treated as part of the function for
the purpose of debug info. This provides a better debugging experience,
among other things by allowing a debug info client to correctly look up
a function in debug info given a function pointer.
Bill Wendling [Sat, 19 Oct 2013 11:27:12 +0000 (11:27 +0000)]
Perform an intelligent splice of the predecessor with the single successor.
If the predecessor's being spliced into a landing pad, then we need the PHIs to
come first and the rest of the predecessor's code to come *after* the landing
pad instruction.
Andrew Trick [Fri, 18 Oct 2013 23:43:53 +0000 (23:43 +0000)]
SCEV should use NSW to get trip count for positive nonunit stride loops.
SCEV currently fails to compute loop counts for nonunit stride
loops. This comes up frequently. It prevents loop optimization and
forces vectorization to insert extra loop checks.
For example:
void foo(int n, int *x) {
for (int i = 0; i < n; i += 3) {
x[i] = i;
x[i+1] = i+1;
x[i+2] = i+2;
}
}
We need to properly handle the case in which limit > INT_MAX-stride. In
the above case: n > INT_MAX-3. In this case the loop counter will step
beyond the limit and overflow at the same time. However, knowing that
signed integer overlow in undefined, we can assume the loop test
behavior is arbitrary after overflow. This obeys both C undefined
behavior rules, and the more strict LLVM poison value rules.
I'm finally fixing this in response to Hal Finkel's persistence.
The most probable reason that we never optimized this before is that
we were being careful to handle case where the developer expected a
side-effect free infinite loop relying on overflow:
for (int i = 0; i < n; i += s) {
++j;
}
return j;
If INT_MAX+1 is a multiple of s and n > INT_MAX-s, then we might
expect an infinite loop. However there are plenty of ways to achieve
this effect without relying on undefined behavior of signed overflow.
Manman Ren [Fri, 18 Oct 2013 21:14:19 +0000 (21:14 +0000)]
Debug Info: add a newly-created DIE to a parent in the same function.
With this commit, all DIEs created in CompileUnit will be added to parents
inside the same function. Also make getOrCreateTemplateType|Value functions
private.
Hans Wennborg [Fri, 18 Oct 2013 20:46:28 +0000 (20:46 +0000)]
MC asm parser: allow ?'s in symbol names, and handle @'s in names in MS asm
This is another (final?) stab at making us able to parse our own asm output
on Windows.
Symbols on Windows often contain @'s and ?'s in their names. Our asm parser
didn't like this. ?'s were not allowed, and @'s were intepreted as trying to
reference PLT/GOT/etc.
We can't just add quotes around the bad names, since e.g. for MinGW, we use gas
to assemble, and it doesn't like quotes in some places (notably in .def
directives).
This commit makes us allow ?'s in symbol names, and @'s in symbol names for MS
assembly.
projects / llvm / log