Chris Bieneman [Tue, 28 Jun 2016 21:10:26 +0000 (21:10 +0000)]
[YAML] Fix YAML tags appearing before the start of sequence elements
Our existing yaml::Output code writes tags immediately when mapTag is called, without any state handling. This results in tags on sequence elements being written before the element itself. For example, we see this:
Our reader handles reading properly, so this bug only impacts writing yaml sequences with tagged elements.
As a test for this I've modified the Mach-O yaml encoding to allways apply the !mach-o tag when encoding MachOYAML::Object entries. This results in the !mach-o tag appearing as expected in dumped fat files.
Zhan Jun Liau [Tue, 28 Jun 2016 21:03:19 +0000 (21:03 +0000)]
[SystemZ] Use NILL instruction instead of NILF where possible
Summary: SystemZ shift instructions only use the last 6 bits of the shift
amount. When the result of an AND operation is used as a shift amount, this
means that we can use the NILL instruction (which operates on the last 16 bits)
rather than NILF (which operates on the last 32 bits) for a 16-bit savings in
instruction size.
Rafael Espindola [Tue, 28 Jun 2016 20:13:36 +0000 (20:13 +0000)]
Use isPositionIndependent in a few more places.
I think this converts all the simple cases that really just care about
the generated code being position independent or not. The remaining
uses are a bit more complicated and are checking things like "is this
a library or executable" or "can this symbol be preempted".
Where command1 and command2 contain completely different sets of
valid options. This is backwards compatible with previous uses
of llvm cl which did not support subcommands, as any option
which specifies no optional subcommand (e.g. all existing
code) goes into a special "top level" subcommand that expects
dashed options to appear immediately after the program name.
For example, code which is subcommand unaware would generate
a command line such as the following, where no subcommand
is specified:
llvm-foo.exe -q1 -q2
The top level subcommand can co-exist with actual subcommands,
as it is implemented as an actual subcommand which is searched
if no explicit subcommand is specified. So llvm-foo.exe as
specified above could be written so as to support all three
aforementioned command lines simultaneously.
There is one additional "special" subcommand called AllSubCommands,
which can be used to inject an option into every subcommand.
This is useful to support things like help, so that commands
such as:
All work and display the help for the selected subcommand
without having to explicitly go and write code to handle each
one separately.
This patch is submitted without an example of anything actually
using subcommands, but a followup patch will convert the
llvm-pdbdump tool to use subcommands.
Artur Pilipenko [Tue, 28 Jun 2016 18:27:25 +0000 (18:27 +0000)]
Support arbitrary addrspace pointers in masked load/store intrinsics
This is a resubmittion of 263158 change after fixing the existing problem with intrinsics mangling (see LTO and intrinsics mangling llvm-dev thread for details).
This patch fixes the problem which occurs when loop-vectorize tries to use @llvm.masked.load/store intrinsic for a non-default addrspace pointer. It fails with "Calling a function with a bad signature!" assertion in CallInst constructor because it tries to pass a non-default addrspace pointer to the pointer argument which has default addrspace.
The fix is to add pointer type as another overloaded type to @llvm.masked.load/store intrinsics.
David Majnemer [Tue, 28 Jun 2016 16:04:46 +0000 (16:04 +0000)]
[X86] Make WRPKRU/RDPKRU pass -verify-machineinstrs
The original implementation attempted to zero registers using
XOR %foo, %foo. This is problematic because it constitutes a
read-modify-write of a register which might not be defined.
Instead, use MOV32r0 to avoid these problems; expandPostRAPseudo does
the right thing here.
Simon Pilgrim [Tue, 28 Jun 2016 13:24:05 +0000 (13:24 +0000)]
[X86][AVX] Peek through bitcasts to find the source of broadcasts (reapplied)
AVX1 can only broadcast vectors as floats/doubles, so for 256-bit vectors we insert bitcasts if we are shuffling v8i32/v4i64 types. Unfortunately the presence of these bitcasts prevents the current broadcast lowering code from peeking through cases where we have concatenated / extracted vectors to create the 256-bit vectors.
This patch allows us to peek through bitcasts as long as the number of elements doesn't change (i.e. element bitwidth is the same) so the broadcast index is not affected.
Note this bitcast peek is different from the stage later on which doesn't care about the type and is just trying to find a load node.
As we're being more aggressive with bitcasts, we also need to ensure that the broadcast type is correctly bitcasted
Vassil Vassilev [Tue, 28 Jun 2016 12:27:09 +0000 (12:27 +0000)]
[modules] Separate intrinsic_gen dependent headers from module LLVM_IR.
Some headers in IR depend on tablegen generated code. Modules builds triggered
generation of the LLVM_IR module (including headers dependant on intrinsic_gen),
imposing a unnecessary build dependency.
Simon Pilgrim [Tue, 28 Jun 2016 08:08:15 +0000 (08:08 +0000)]
[X86][SSE] Added support for combining target shuffles to (V)PSHUFD/VPERMILPD/VPERMILPS immediate permutes
This patch allows target shuffles to be combined to single input immediate permute instructions - (V)PSHUFD/VPERMILPD/VPERMILPS - allowing more general pattern matching than what we current do and improves the likelihood of memory folding compared to existing patterns which tend to reuse the input in multiple arguments.
Further permute instructions (V)PSHUFLW/(V)PSHUFHW/(V)PERMQ/(V)PERMPD may be added in the future but its proven tricky to create tests cases for them so far. (V)PSHUFLW/(V)PSHUFHW is already handled quite well in combineTargetShuffle so it may be that removing some of that code may allow us to perform more of the combining in one place without duplication.
This patch enhances dot graph viewer to show hot regions
with hot bbs/edges displayed in red. The ratio of the bb
freq to the max freq of the function needs to be no less
than the value specified by view-hot-freq-percent option.
The default value is 10 (i.e. 10%).
MBFI supports profile count dumping and function
name based filtering. Add these two feature to
BFI as well. The filtering option is shared between
BFI and MBFI: -view-bfi-func-name=..
Adam Nemet [Tue, 28 Jun 2016 04:02:47 +0000 (04:02 +0000)]
[LLE] Don't hoist conditionally executed loads
If the load is conditional we can't hoist its 0-iteration instance to
the preheader because that would make it unconditional. Thus we would
access a memory location that the original loop did not access.
BFI and MBFI's dot traits class share most of the
code and all future enhancement. This patch extracts
common implementation into base class BFIDOTGraphTraitsBase.
This patch also enables BFI graph to show branch probability
on edges as MBFI does before.
Vedant Kumar [Tue, 28 Jun 2016 02:09:39 +0000 (02:09 +0000)]
Reapply "[llvm-cov] Add an -output-dir option for the show sub-command""
Passing -output-dir path/to/dir to llvm-cov show creates path/to/dir if
it doesn't already exist, and prints reports into that directory.
In function view mode, all views are written into
path/to/dir/functions.$EXTENSION. In file view mode, all views are
written into path/to/dir/coverage/$PATH.$EXTENSION.
Vedant Kumar [Tue, 28 Jun 2016 00:18:57 +0000 (00:18 +0000)]
[llvm-cov] Add an -output-dir option for the show sub-command
Passing -output-dir path/to/dir to llvm-cov show creates path/to/dir if
it doesn't already exist, and prints reports into that directory.
In function view mode, all views are written into
path/to/dir/functions.$EXTENSION. In file view mode, all views are
written into path/to/dir/coverage/$PATH.$EXTENSION.
Chandler Carruth [Mon, 27 Jun 2016 23:26:08 +0000 (23:26 +0000)]
[PM] Improve the debugging and logging facilities of the CGSCC bits of
the new pass manager.
This adds operator<< overloads for the various bits of the
LazyCallGraph, dump methods for use from the debugger, and debug logging
using them to the CGSCC pass manager.
Having this was essential for debugging the call graph update patch, and
I've extracted what I could from that patch here to minimize the delta.
Kevin Enderby [Mon, 27 Jun 2016 21:39:39 +0000 (21:39 +0000)]
Change all but the last ErrorOr<...> use for MachOUniversalBinary to Expected<...> to
allow a good error message to be produced.
I added the one test case that the object file tools could produce an error
message. The other two errors can’t be triggered if the input file is passed
through sys::fs::identify_magic(). But the malformedError("bad magic number")
does get triggered by the logic in llvm-dsymutil when dealing with a normal
Mach-O file. The other "File too small ..." error would take a logic error
currently to produce and is not tested for.
Davide Italiano [Mon, 27 Jun 2016 20:38:39 +0000 (20:38 +0000)]
[llvm-ar] Ignore -plugin option.
binutils ar uses -plugin to specify the LTO plugin, but LLVM doesn't
need this as it doesn't use a plugin for LTO. Accepting (and ignoring)
the option allows interoperability with existing build systems and
make downstream consumers life much easier.
Chris Bieneman [Mon, 27 Jun 2016 19:53:53 +0000 (19:53 +0000)]
[yaml2obj] Remove --format option in favor of YAML tags
Summary:
Our YAML library's handling of tags isn't perfect, but it is good enough to get rid of the need for the --format argument to yaml2obj. This patch does exactly that.
Instead of requiring --format, it infers the format based on the tags found in the object file. The supported tags are:
!ELF
!COFF
!mach-o
!fat-mach-o
I have a corresponding patch that is quite large that fixes up all the in-tree test cases.
Artur Pilipenko [Mon, 27 Jun 2016 16:54:33 +0000 (16:54 +0000)]
Revert -r273892 "Support arbitrary addrspace pointers in masked load/store intrinsics" since some of the clang tests don't expect to see the updated signatures.
Artur Pilipenko [Mon, 27 Jun 2016 16:29:26 +0000 (16:29 +0000)]
Support arbitrary addrspace pointers in masked load/store intrinsics
This is a resubmittion of 263158 change after fixing the existing problem with intrinsics mangling (see LTO and intrinsics mangling llvm-dev thread for details).
This patch fixes the problem which occurs when loop-vectorize tries to use @llvm.masked.load/store intrinsic for a non-default addrspace pointer. It fails with "Calling a function with a bad signature!" assertion in CallInst constructor because it tries to pass a non-default addrspace pointer to the pointer argument which has default addrspace.
The fix is to add pointer type as another overloaded type to @llvm.masked.load/store intrinsics.
Kuba Brecka [Mon, 27 Jun 2016 15:57:08 +0000 (15:57 +0000)]
[asan] fix false dynamic-stack-buffer-overflow report with constantly-sized dynamic allocas, LLVM part
See the bug report at https://github.com/google/sanitizers/issues/691. When a dynamic alloca has a constant size, ASan instrumentation will treat it as a regular dynamic alloca (insert calls to poison and unpoison), but the backend will turn it into a regular stack variable. The poisoning/unpoisoning is then broken. This patch will treat such allocas as static.
Renato Golin [Mon, 27 Jun 2016 14:42:20 +0000 (14:42 +0000)]
[ARM] Fix Thumb text sections' flags under COFF/Windows
The main issue here is that the "thumb" flag wasn't set for some of these
sections, making MSVC's link.exe fails to correctly relocate code
against the symbols inside these sections. link.exe could fail for
instance with the "fixup is not aligned for target 'XX'" error. If
linking doesn't fail, the relocation process goes wrong in the end and
invalid code is generated by the linker.
This patch adds Thumb/ARM information so that the right flags are set
on COFF/Windows.
projects / llvm / log