1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-30 15:32:52 +01:00
Commit Graph

97374 Commits

Author SHA1 Message Date
Benjamin Kramer
2d870f327a X86: Add a description for AMD bdver3 aka Steamroller.
This is just bdver2 + FSGSBase.

llvm-svn: 193984
2013-11-04 10:29:20 +00:00
Filip Pizlo
0080a77e7a Make the pretty stack trace be an opt-in, rather than opt-out, facility. Enable pretty
stack traces by default if you use PrettyStackTraceProgram, so that existing LLVM-based 
tools will continue to get it without any changes.

llvm-svn: 193971
2013-11-04 02:22:25 +00:00
Peter Zotov
4b13ee5f62 [OCaml] implement Llvm_passmgr_builder, bindings for PassManagerBuilder
llvm-svn: 193968
2013-11-04 01:39:42 +00:00
Peter Zotov
23e94a56d9 [OCaml] Documentation
llvm-svn: 193967
2013-11-04 01:39:31 +00:00
Peter Zotov
22f34b9373 [OCaml] Implement missing LLVMCore APIs
llvm-svn: 193966
2013-11-04 01:39:26 +00:00
Peter Zotov
d01f123fa7 [OCaml] Formatting
llvm-svn: 193965
2013-11-04 01:39:18 +00:00
Elena Demikhovsky
841cd7d09e AVX-512: added VPCONFLICT instruction and intrinsics,
added EVEX_KZ to tablegen

llvm-svn: 193959
2013-11-03 13:46:31 +00:00
Benjamin Kramer
15ebc47438 SLPVectorizer: Add a missing pair of parens. No functionality change.
llvm-svn: 193958
2013-11-03 12:54:32 +00:00
Venkatraman Govindaraju
8ce290e5fa [SparcV9] Handle i64 <-> float conversions in sparcv9 mode.
llvm-svn: 193957
2013-11-03 12:28:40 +00:00
Benjamin Kramer
f45bcf5480 SLPVectorizer: When CSEing generated gathers only scan blocks containing them.
Instead of doing a RPO traversal of the whole function remember the blocks
containing gathers (typically <= 2) and scan them in dominator-first order.

The actual CSE is still quadratic, but I'm not confident that adding a
scoped hash table here is worth it as we're only looking at the generated
instructions and not arbitrary code.

llvm-svn: 193956
2013-11-03 12:27:52 +00:00
David Majnemer
2bbbbaaeee Revert "Inliner: Handle readonly attribute per argument when adding memcpy"
This reverts commit r193356, it caused PR17781.

A reduced test case covering this regression has been added to the test suite.

llvm-svn: 193955
2013-11-03 12:22:13 +00:00
David Majnemer
fbb2338856 Spell "Actual" correctly
llvm-svn: 193954
2013-11-03 11:09:39 +00:00
Peter Zotov
47beae7a65 [OCaml] Implement Llvm.MemoryBuffer.{of_string,as_string}
llvm-svn: 193953
2013-11-03 08:27:45 +00:00
Peter Zotov
57d8f2c2e9 [OCaml] Fix ABI incompatibility
OCaml's type unit is not compatible with C's type void.

llvm-svn: 193952
2013-11-03 08:27:37 +00:00
Peter Zotov
aa2a604cf0 [OCaml] Implement Llvm_linker, bindings for the IR linker
llvm-svn: 193951
2013-11-03 08:27:32 +00:00
Peter Zotov
2aba543196 [OCaml] Implement Llvm_vectorize bindings
llvm-svn: 193950
2013-11-03 08:27:22 +00:00
Peter Zotov
f087cf7094 [OCaml] Refactor Llvm_target tests
Llvm_target tests did not check for return values. This actually
caused them to miss a bug.

llvm-svn: 193949
2013-11-03 08:27:13 +00:00
Peter Zotov
500cb8d4ee [OCaml] Llvm_target: fix typo (Int_val instead of Int64_val)
llvm-svn: 193948
2013-11-03 08:27:08 +00:00
Venkatraman Govindaraju
1b0f7d035d [Sparc] Expand FP_TO_UINT, UINT_TO_FP for fp128.
llvm-svn: 193947
2013-11-03 08:00:19 +00:00
Peter Zotov
9e58e5ad83 [OCaml] Llvm_scalar_opts: add missing transforms
llvm-svn: 193946
2013-11-03 07:54:17 +00:00
Peter Zotov
d9e82e73c8 [OCaml] Llvm_ipo: add missing transforms
llvm-svn: 193945
2013-11-03 07:54:08 +00:00
Bob Wilson
eb7943100b Convert calls to __sinpi and __cospi into __sincospi_stret
This adds an SimplifyLibCalls case which converts the special __sinpi and
__cospi (float & double variants) into a __sincospi_stret where appropriate to
remove duplicated work.

Patch by Tim Northover

llvm-svn: 193943
2013-11-03 06:48:38 +00:00
Bob Wilson
f7bb300deb Enable optimization of sin / cos pair into call to __sincos_stret for iOS7+.
rdar://12856873
Patch by Evan Cheng, with a fix for rdar://13209539 by Tilmann Scheller

llvm-svn: 193942
2013-11-03 06:14:38 +00:00
Venkatraman Govindaraju
c60684ae6f [SparcV9] Add ctpop instruction for i64. Also, expand ctlz, cttz and bswap.
llvm-svn: 193941
2013-11-03 05:59:07 +00:00
Bob Wilson
a6dc8812c7 Configure with --enable-terminfo=no for Apple's llvmCore builds.
Patch by Pete Cooper.

llvm-svn: 193940
2013-11-03 05:49:36 +00:00
Filip Pizlo
aba5b5499b Add a comment to note that LLVMDisablePrettyStackTrace() is likely not a good long-term solution.
llvm-svn: 193939
2013-11-03 04:38:31 +00:00
Filip Pizlo
8196f96988 When LLVM is embedded in a larger application, it's not OK for LLVM to intercept crashes. LLVM already has
the ability to disable this functionality.  This patch exposes it via the C API.

llvm-svn: 193937
2013-11-03 00:29:47 +00:00
Rafael Espindola
3c7f6cf61c move getSymbolNMTypeChar to the one program that needs it: nm.
llvm-svn: 193933
2013-11-02 21:16:09 +00:00
Rafael Espindola
844544bbd9 Convert another use of getSymbolNMTypeChar.
llvm-svn: 193932
2013-11-02 20:10:07 +00:00
Rafael Espindola
de649da692 Avoid some getSymbolNMTypeChar uses in COFFObjectFile.cpp itself.
This is a fixed version of 193928 which keeps these uses in sync.

llvm-svn: 193931
2013-11-02 18:07:48 +00:00
Rafael Espindola
0380ceeb65 Revert "Don't use getSymbolNMTypeChar for implementing COFFObjectFile::getSymbolFileOffset."
Investigating a bot failure.

This reverts commit r193928.

llvm-svn: 193929
2013-11-02 17:12:49 +00:00
Rafael Espindola
fc3d69f14a Don't use getSymbolNMTypeChar for implementing COFFObjectFile::getSymbolFileOffset.
llvm-svn: 193928
2013-11-02 16:55:21 +00:00
Benjamin Kramer
ae919396b6 SLPVectorizer: Remove duplicated function.
llvm-svn: 193927
2013-11-02 14:46:27 +00:00
Benjamin Kramer
abc7baa1dc LoopVectorize: Remove quadratic behavior the local CSE.
Doing this with a hash map doesn't change behavior and avoids calling
isIdenticalTo O(n^2) times. This should probably eventually move into a utility
class shared with EarlyCSE and the limited CSE in the SLPVectorizer.

llvm-svn: 193926
2013-11-02 13:39:00 +00:00
Rafael Espindola
29823465a8 A better fix that also works on ppc: add a target tripple.
llvm-svn: 193915
2013-11-02 06:00:09 +00:00
Rafael Espindola
2978e6cc09 Fix this test to pass on darwin now that llvm-nm is working.
llvm-svn: 193914
2013-11-02 05:29:22 +00:00
Rafael Espindola
1506f61a81 Fix llvm-nm to mach OS X's nm on some tests.
There is still a long way to go for llvm-nm, but at least we now match
nm's letter output in the cases we test for.

llvm-svn: 193912
2013-11-02 05:03:24 +00:00
Michael Liao
ae9a5c1116 Fix PR17764
- When selecting BLEND from vselect, the operands need swapping as due to the
  difference between vselect and SSE/AVX's BLEND insn

llvm-svn: 193900
2013-11-02 00:10:02 +00:00
Yuchen Wu
c01bdb27f5 Added command-line option to output llvm-cov to file.
Added -o option to llvm-cov. If no output file is specified, it defaults
to STDOUT.

llvm-svn: 193899
2013-11-02 00:09:17 +00:00
Arnold Schwaighofer
fed0c4f8e8 LoopVectorizer: Move cse code into its own function
llvm-svn: 193895
2013-11-01 23:28:54 +00:00
Eric Christopher
f77453e16f Comment some and reformat for clarity beginFunction.
llvm-svn: 193894
2013-11-01 23:14:17 +00:00
David Blaikie
aa11319eaa DebugInfo: regenerate test case from Clang to adjust for fixes/improvements
I hit some problems with future work due to the member subprogram of
'a_b's type having a subprogram (an implicit default ctor, !52 in the
pre-commit source) with no name. Clang now generates a name for such a
function but in this case doesn't even emit debug info for it as it is
unused (Clang never emits the body of the ctor, instead just emitting
memset if needed).

llvm-svn: 193892
2013-11-01 22:29:28 +00:00
Arnold Schwaighofer
fba1c74b67 LoopVectorizer: Perform redundancy elimination on induction variables
When the loop vectorizer was part of the SCC inliner pass manager gvn would
run after the loop vectorizer followed by instcombine. This way redundancy
(multiple uses) were removed and instcombine could perform scalarization on the
induction variables. Having moved the loop vectorizer to later we no longer run
any form of redundancy elimination before we perform instcombine. This caused
vectorized induction variables to survive that did not before.

On a recent iMac this helps linpack back from 6000Mflops to 7000Mflops.

This should also help lpbench and paq8p.

I ran a Release (without Asserts) build over the test-suite and did not see any
negative impact on compile time.

radar://15339680

llvm-svn: 193891
2013-11-01 22:18:19 +00:00
David Blaikie
2cee6afb96 DebugInfo: Improve readability of test case added in r193878
The point is to ensure that the attribute in question
(DW_AT_data_member_location) is associated with the prior tag, so ensure
that we don't see another tag starting between the intended tag and the
desired attribute.

llvm-svn: 193884
2013-11-01 20:59:53 +00:00
David Blaikie
a9b5d341da DebugInfo: add a test case for data member locations (coverage for r193835)
llvm-svn: 193878
2013-11-01 18:25:55 +00:00
David Blaikie
a3394eede2 Fix a test case broken by r193872
llvm-svn: 193876
2013-11-01 18:18:16 +00:00
Manman Ren
ca35ef0140 Add comments.
llvm-svn: 193874
2013-11-01 18:06:25 +00:00
David Blaikie
263e05531f DebugInfo: Make pubnames header printing similar to unit header printing
In a failed attempt to allow the gnu-public-names.ll test case to not
hardcode the size of the unit that the pubnames section referred to I've
at least managed to have unit headers and pubnames headers print out in
a similar style.

This failed to achieve the desired goal because the header in a unit
specifies the length of the unit without the length element of the
header whereas the length in the pubnames includes this element, so the
numbers are off by 4 bytes. I don't know of any arithmetic powers in
FileCheck so the test case can't simply say "CU_LENGTH + 4".

llvm-svn: 193872
2013-11-01 17:53:30 +00:00
Juergen Ributzka
1d4449f32d [Stackmap] Remove erroneous assert.
llvm-svn: 193871
2013-11-01 17:53:27 +00:00
Matt Arsenault
6936bdad5d Use isa<> instead of dyn_cast<> with unused value
llvm-svn: 193869
2013-11-01 17:39:26 +00:00