Jessica Paquette [Tue, 17 Oct 2017 19:03:23 +0000 (19:03 +0000)]
[MachineOutliner][NFC] Move decrement logic to OutlinedFunction
This commit moves the decrement logic for outlined functions into the class,
and makes OccurrenceCount private. It can now be accessed via
getOccurrenceCount().
This makes it more difficult to accidentally introduce bugs by incorrectly
decrementing the occurrence count on OutlinedFunctions.
Jessica Paquette [Tue, 17 Oct 2017 18:43:15 +0000 (18:43 +0000)]
[MachineOutliner][NFC] Move end index calculation into Candidate
Cleanup to Candidate that moves all end index calculations into
Candidate.endIdx(). For the sake of consistency, StartIdx and Len are now
private members, and can be accessed with length() and startIdx() respectively.
Gadi Haber [Tue, 17 Oct 2017 13:45:39 +0000 (13:45 +0000)]
[X86][Broadwell] Added the broadwell cpu to the scheduling regression tests.<NFC>
NFC.
Added the Broadwell cpu and the BROADWELL prefix to all the scheduling regression tests, as part of prepartion for a larger commit of adding all Broadwell scheduiling.
Nikolai Bozhenov [Tue, 17 Oct 2017 11:50:48 +0000 (11:50 +0000)]
Improve clamp recognition in ValueTracking.
Summary:
ValueTracking was recognizing not all variations of clamp. Swapping of
true value and false value of select was added to fix this problem. This
change breaks the canonical form of cmp inside the matchMinMax function,
that is why additional checks for compare predicates is needed. Added
corresponding test cases.
Yichao Yu [Tue, 17 Oct 2017 11:47:36 +0000 (11:47 +0000)]
Fix implicit null check with negative offset
Summary:
It seems that negative offset was accidentally allowed in D17967.
AFAICT small negative offset should be valid (always raise segfault) on all archs that I'm aware of (especially x86, which is the only one with this optimization enabled) and such case can be useful when loading hiden metadata from an object.
However, like the positive side, it should only be done within a certain limit.
For now, use the same limit on the positive side for the negative side.
A separate option can be added if needs appear.
Yichao Yu [Tue, 17 Oct 2017 11:44:34 +0000 (11:44 +0000)]
Fix `FaultMaps` crash when the out streamer is reused
Summary:
Make sure the map is cleared before processing a new module. Similar to what is done on `StackMaps`.
This issue is similar to D38588, though this time for FaultMaps (on x86) rather than ARM/AArch64. Other than possible mixing of information between modules, the crash is caused by the pointers values in the map that was allocated by the bump pointer allocator that is unwinded when emitting the next file. This issue has been around since 3.8.
This issue is likely much harder to write a test for since AFAICT it requires emitting something much more compilcated (and possibly real code) instead of just some random bytes.
Gadi Haber [Tue, 17 Oct 2017 06:47:04 +0000 (06:47 +0000)]
[X86][SKL] Updated scheduling information for the SkylakeClient target
Updated the scheduling information for the SkylakeClient target with the following changes:
1. regrouped the instructions after adding load and store latencies.
2. regrouped the instructions after adding identified missing ports in several groups.
The changes were made after revisiting the latencies impact of all the load and store uOps.
Philip Reames [Tue, 17 Oct 2017 06:21:07 +0000 (06:21 +0000)]
Revert 315440 on behalf of mkazantsev
This patch reverts rL315440 because of the bug described at
https://bugs.llvm.org/show_bug.cgi?id=34937
The fix for the bug is on review as D38944, but not yet ready. Given this is a regression reverting until a fix is ready is called for.
Max would have done the revert himself, but is having trouble doing a build of fresh LLVM for some reason. I did the build and test to ensure the revert worked as expected on his behalf.
Vedant Kumar [Tue, 17 Oct 2017 01:34:41 +0000 (01:34 +0000)]
[llvm-cov] Add one correction to r315960 (PR34962)
In r315960, I accidentally assumed that the first line segment is
guaranteed to be the non-gap region entry segment (given that one is
present). It can actually be any segment on the line, and the test I
checked in demonstrates that.
Vedant Kumar [Mon, 16 Oct 2017 23:47:10 +0000 (23:47 +0000)]
[llvm-cov] Remove workaround in line execution count calculation (PR34962)
Gap areas make it possible to correctly determine when to use counts
from deferred regions. Before gap areas were introduced, llvm-cov needed
to use a heuristic to do this: it ignored counts from segments that
start, but do not end, on a line. This heuristic breaks down on a simple
example (see PR34962).
This patch removes the heuristic and picks counts from any region entry
segment which isn't a gap area.
Mark Searles [Mon, 16 Oct 2017 23:38:53 +0000 (23:38 +0000)]
Use the return value of UpdateNodeOperands(); in some cases, UpdateNodeOperands() modifies the node in-place and using the return value isn’t strictly necessary. However, it does not necessarily modify the node, but may return a resultant node if it already exists in the DAG. See comments in UpdateNodeOperands(). In that case, the return value must be used to avoid such scenarios as an infinite loop (node is assumed to have been updated, so added back to the worklist, and re-processed; however, node hasn’t changed so it is once again passed to UpdateNodeOperands(), assumed modified, added back to worklist; cycle infinitely repeats).
Matthias Braun [Mon, 16 Oct 2017 23:03:21 +0000 (23:03 +0000)]
cmake: BSD: Mark /usr/local/include as system include directory
We add /usr/local/include to the include directory list for some BSD
systems. We should mark this as a system directory to avoid files from
/usr/local/include getting picked over files shipping with llvm.
This typically manifested as gtest headers installed with the system
getting picked over the ones shipping with llvm.
Quentin Colombet [Mon, 16 Oct 2017 22:28:40 +0000 (22:28 +0000)]
Re-apply [AArch64][RegisterBankInfo] Use the statically computed mappings for COPY
This reverts commit r315823, thus re-applying r315781.
Also make sure we don't use G_BITCAST mapping for non-generic registers.
Non-generic registers don't have a type but do have a reg bank.
Something the COPY mapping now how to deal with but the G_BITCAST
mapping don't.
-- Original Commit Message --
We use to resort on the generic implementation to get the mappings for
COPYs. The generic implementation resorts on table lookup and
dynamically allocated objects to get the valid mappings.
Given we already know how to map G_BITCAST and have the static mappings
for them, use that code path for COPY as well. This is much more
efficient.
Improve the compile time of RegBankSelect by up to 20%.
Note: When we eventually generate all the mappings via TableGen, we
wouldn't have to do that dance to shave compile time. The intent of this
change was to make sure that moving to static structure really pays off.
Tony Tye [Mon, 16 Oct 2017 20:44:29 +0000 (20:44 +0000)]
Add base relative relocation record that can be used for the following case (OpenCL example):
static __global int Var = 0;
__global int* Ptr[] = {&Var};
...
In this case Var is a non premptable symbol and so its address can be used as the value of Ptr, with a base relative relocation that will add the delta between the ELF address and the actual load address. Such relocations do not require a symbol.
This patch adds the ability to perform IPSCCP-like interprocedural analysis to
the generic sparse propagation solver. The patch gives clients the ability to
define their own custom LatticeKey types that the generic solver maps to custom
LatticeVal types. The custom lattice keys can be used, for example, to
distinguish among mappings for regular values, values returned from functions,
and values stored in global variables. Clients are responsible for defining how
to convert between LatticeKeys and LLVM Values by providing a specialization of
the LatticeKeyInfo template.
The added unit tests demonstrate how the generic solver can be used to perform
a simplified version of interprocedural constant propagation.
Akira Hatanaka [Mon, 16 Oct 2017 16:46:59 +0000 (16:46 +0000)]
[ObjCARC] Do not move a release that has the clang.imprecise_release tag
above PHI instructions.
ARC optimizer has an optimization that moves a call to an ObjC runtime
function above a phi instruction when the phi has a null operand and is
an argument passed to the function call. This optimization should not
kick in when the runtime function is an objc_release that releases an
object with precise lifetime semantics.
Sjoerd Meijer [Mon, 16 Oct 2017 14:07:30 +0000 (14:07 +0000)]
ISel type legalizer: debug messages. NFC.
Minor addition and follow up of r314773 and r311533: this adds more
debug messages to the type legalizer. For each node, it dumps
legalization info for results and operands nodes, rather than just the
final legalized node.
This patch is a result of D37262: The issues with X86 prefixes. It closes PR7709, PR17697, PR19251, PR32809 and PR21640. There could be other bugs closed by this patch.
Nikolai Bozhenov [Mon, 16 Oct 2017 09:19:21 +0000 (09:19 +0000)]
Move folding of icmp with zero after checking for min/max idioms.
Summary:
The following transformation for cmp instruction:
icmp smin(x, PositiveValue), 0 -> icmp x, 0
should only be done after checking for min/max to prevent infinite
looping caused by a reverse canonicalization. That is why this
transformation was moved to place after the mentioned check.
Yonghong Song [Mon, 16 Oct 2017 04:14:53 +0000 (04:14 +0000)]
bpf: fix bug on silently truncating 64-bit immediate
We came across an llvm bug when compiling some testcases that 64-bit
immediates are silently truncated into 32-bit and then packed into
BPF_JMP | BPF_K encoding. This caused comparison with wrong value.
This bug looks to be introduced by r308080. The Select_Ri pattern is
supposed to be lowered into J*_Ri while the latter only support 32-bit
immediate encoding, therefore Select_Ri should have similar immediate
predicate check as what J*_Ri are doing.
Reported-by: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: Jiong Wang <jiong.wang@netronome.com> Reviewed-by: Yonghong Song <yhs@fb.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@315889 91177308-0d34-0410-b5e6-96231b3b80d8
Hiroshi Inoue [Mon, 16 Oct 2017 04:12:57 +0000 (04:12 +0000)]
[PowerPC] Eliminate sign- and zero-extensions if already sign- or zero-extended
This patch enables redundant sign- and zero-extension elimination in PowerPC MI Peephole pass.
If the input value of a sign- or zero-extension is known to be already sign- or zero-extended, the operation is redundant and can be eliminated.
One common case is sign-extensions for a method parameter or for a method return value; they must be sign- or zero-extended as defined in PPC ELF ABI.
For example of the following simple code, two extsw instructions are generated before the invocation of int_func and before the return. With this patch, both extsw are eliminated.
void int_func(int);
void ii_test(int a) {
if (a & 1) return int_func(a);
}
Such redundant sign- or zero-extensions are quite common in many programs; e.g. I observed about 60,000 occurrences of the elimination while compiling the LLVM+CLANG.
Daniel Sanders [Mon, 16 Oct 2017 03:36:29 +0000 (03:36 +0000)]
Re-commit r315885: [globalisel][tblgen] Add support for iPTR and implement am_unscaled* and am_indexed*
Summary:
iPTR is a pointer of subtarget-specific size to any address space. Therefore
type checks on this size derive the SizeInBits from a subtarget hook.
At this point, we can import the simplests G_LOAD rules and select load
instructions using them. Further patches will support for the predicates to
enable additional loads as well as the stores.
The previous commit failed on MSVC due to a failure to convert an
initializer_list to a std::vector. Hopefully, MSVC will accept this version.
Daniel Sanders [Mon, 16 Oct 2017 01:16:35 +0000 (01:16 +0000)]
[globalisel][tblgen] Add support for iPTR and implement am_unscaled* and am_indexed*
Summary:
iPTR is a pointer of subtarget-specific size to any address space. Therefore
type checks on this size derive the SizeInBits from a subtarget hook.
At this point, we can import the simplests G_LOAD rules and select load
instructions using them. Further patches will support for the predicates to
enable additional loads as well as the stores.
Daniel Sanders [Mon, 16 Oct 2017 00:56:30 +0000 (00:56 +0000)]
[globalisel][tablegen] Implement unindexed load, non-extending load, and MemVT checks
Summary:
This includes some context-sensitivity in the MVT to LLT conversion so that
pointer types are tested correctly.
FIXME: I'm not happy with the way this is done since everything is a
special-case. I've yet to find a reasonable way to implement it.
select-load.mir fails because <1 x s64> loads in tablegen get priority over s64
loads. This is fixed in the next patch and as such they should be committed
together, I've posted them separately to help with the review.
Nico Weber [Sun, 15 Oct 2017 19:13:57 +0000 (19:13 +0000)]
Search for libxml2 on macOS too.
This allows lld-link to process /manifestinput: flags on macOS too.
Also makes the `REQUIRES: manifesttool` lld tests run on macOS.
Setting LLVM_ENABLE_LIBXML2 to off can suppress this behavior, like on Linux.
Daniel Sanders [Sun, 15 Oct 2017 18:22:54 +0000 (18:22 +0000)]
Re-commit r315863: [globalisel][tablegen] Import ComplexPattern when used as an operator
Summary:
It's possible for a ComplexPattern to be used as an operator in a match
pattern. This is used by the load/store patterns in AArch64 to name the
suboperands returned by ComplexPattern predicate so that they can be broken
apart and referenced independently in the result pattern.
This patch adds support for this in order to enable the import of load/store
patterns.
Depends on D37445
Hopefully fixed the ambiguous constructor that a large number of bots reported.
Daniel Sanders [Sun, 15 Oct 2017 17:03:36 +0000 (17:03 +0000)]
[globalisel][tablegen] Import ComplexPattern when used as an operator
Summary:
It's possible for a ComplexPattern to be used as an operator in a match
pattern. This is used by the load/store patterns in AArch64 to name the
suboperands returned by ComplexPattern predicate so that they can be broken
apart and referenced independently in the result pattern.
This patch adds support for this in order to enable the import of load/store
patterns.
Craig Topper [Sun, 15 Oct 2017 16:57:33 +0000 (16:57 +0000)]
[X86] Remove the SlowBTMem feature flag entirely
Turns out we have no patterns on the instructions that were using this feature flag for other reasons. These instructions are slow on all modern CPUs so it seems unlikely that we will spend any effort supporting these instructions going forward. So we might as well just kill of the feature flag and just fix up the comments.
Craig Topper [Sun, 15 Oct 2017 16:41:15 +0000 (16:41 +0000)]
[X86] Add FeatureSlowBTMem to Haswell, Broadwell, Skylake, Cannonlake, and Knights Landing CPUs.
Summary: I see nothing in Agner Fog's tables to indicate that this improved between Ivy Bridge and Haswell. It's also set for all Atom CPUs so I assume KNL should have it too.
[TableGen] Remove error checks incorrectly failing on non-error conditions
In type inference, an empty type set for a specific hw mode is not an
error. In earlier stages of the design it was, but having to use non-
parameterized types with target intrinsics necessarily led to type
contradictions: since the intrinsics used specific types, they were
only valid for a specific hw mode, and the resulting type set for other
modes ended up empty. To accommodate the existence of such intrinsics
individual type sets were allowed to be empty as long as not all sets
were empty.
whitequark [Sun, 15 Oct 2017 12:29:09 +0000 (12:29 +0000)]
[MergeFunctions] Merge small functions if possible without a thunk.
This can result in significant code size savings in some cases,
e.g. an interrupt table all filled with the same assembly stub
in a certain Cortex-M BSP results in code blowup by a factor of 2.5.
Craig Topper [Sun, 15 Oct 2017 06:05:50 +0000 (06:05 +0000)]
[X86] Don't use constant condition for select instruction when testing masking ops.
We should be able to fold constant conditions by converting to shuffles, but fixing that would break these tests in their current form. Since they are really trying to test masking ops, add a non-constant mask to the selects.
projects / llvm / log