Update linalg to affine lowering for convop to use affine load for input
whenever there is no padding. It had always been using std.loads because
max in index functions (needed for non-zero padding if not materializing
zeros) couldn't be represented in the non-zero padding cases.
In the future, the non-zero padding case could also be made to use
affine - either by materializing or using affine.execute_region. The
latter approach will not impact the scf/std output obtained after
lowering out affine.
Differential Revision: https://reviews.llvm.org/D81191
This patch addresses the comment in [D80972](https://reviews.llvm.org/D80972#inline-744217).
Before this patch, the initial length field of .debug_aranges section should be declared as:
```
## 32-bit DWARF
debug_aranges:
- Length:
TotalLength: 0x20
Version: 2
...
## 64-bit DWARF
debug_aranges:
- Length:
TotalLength: 0xffffffff
TotalLength64: 0x20
Version: 2
...
```
After this patch:
```
## 32-bit DWARF
debug_aranges:
- [[Format: DWARF32]] ## Optional
Length: 0x20
Version: 2
...
## 64-bit DWARF
debug_aranges:
- Format: DWARF64
Length: 0x20
Version: 2
```
Current implementation of generating DWARF64 .debug_aranges section is buggy. A follow-up patch will improve it and add test cases for DWARF64.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D81063
It should not be necessary to use weak linkage for these. Doing so
implies interposablity and thus PIC generates indirections and
dynamic relocations, which are unnecessary and suboptimal. Aside
from this, ASan instrumentation never introduces GOT indirection
relocations where there were none before--only new absolute relocs
in RELRO sections for metadata, which are less problematic for
special linkage situations that take pains to avoid GOT generation.
Patch By: mcgrathr
Differential Revision: https://reviews.llvm.org/D80605
trivial.
We previously took a shortcut by assuming that if a subobject had a
trivial copy assignment operator (with a few side-conditions), we would
always invoke it, and could avoid going through overload resolution.
That turns out to not be correct in the presenve of ref-qualifiers (and
also won't be the case for copy-assignments with requires-clauses
either). Use the same logic for lazy declaration of copy-assignments
that we use for all other special member functions.
Summary:
Cache the results from getMachineBasicBlocks in LexicalScopes to speed
up UserValueScopes::dominates queries. This replaces the caching done
in UserValueScopes. Compared to the old caching method, this reduces
memory traffic when a VarLoc is copied (e.g. when a VarLocMap grows),
and enables caching across basic blocks.
When compiling sqlite 3.5.7 (CTMark version), this patch reduces the
number of calls to getMachineBasicBlocks from 10,207 to 1,093. I also
measured a small compile-time reduction (~ 0.1% of total wall time, on
average, on my machine).
As a drive-by, I made the DebugLoc in UserValueScopes a const reference
to cut down on MetadataTracking traffic.
Reviewers: jmorse, Orlando, aprantl, nikic
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80957
This simplifies a lot of handling of BoolAttr/IntegerAttr. For example, a lot of places currently have to handle both IntegerAttr and BoolAttr. In other places, a decision is made to pick one which can lead to surprising results for users. For example, DenseElementsAttr currently uses BoolAttr for i1 even if the user initialized it with an Array of i1 IntegerAttrs.
Differential Revision: https://reviews.llvm.org/D81047
As a followup to D62922, add a sysroot command-line option to this test
to ensure that the output is independent of any default sysroot options,
and adjust the reactor test to be more consistent with the command test.
This revision adds a helper function to hoist alloc/dealloc pairs and
alloca op out of immediately enclosing scf::ForOp if both conditions are true:
1. all operands are defined outside the loop.
2. all uses are ViewLikeOp or DeallocOp.
This is now considered Linalg-specific and will be generalized on a per-need basis.
Differential Revision: https://reviews.llvm.org/D81152
Now that we have an operand based form for the GC arguments to a statepoint intrinsic, update RS4GC to use it and update tests to reflect. This is pretty straight forward. I nearly landed without review, but figured a second set of eyes didn't hurt.
Differential Revision: https://reviews.llvm.org/D81121
Follow the model used on Linux, where the clang driver passes the
linker a -u switch to force the profile runtime to be linked in,
rather than having every TU emit a dead function with a reference.
Differential Revision: https://reviews.llvm.org/D79835
- Change the reference to salvageDebugInfoOrUndef to salvageDebugInfo
(in accordance with https://reviews.llvm.org/D78369).
- Reorganize a few sections in preparation for an upcoming change that
attempts to specify rules for updating debug locations.
- Fix some intra-document links.
- Some spelling / wording fixes.
Global TableGen let override blocks are pretty dangerous and override
any local special cases. In this case, the broader HasFlatGlobalInsts
was overriding the more specific predicate for
FeatureAtomicFaddInsts. Make sure HasFlatGlobalInsts is implied by
FeatureAtomicFaddInsts, and make sure the right predicate is used.
One issue with independently setting the subtarget features on
incompatible targets is all of the encoding families do not define all
opcodes. This will hit an assert on gfx10 for example, since we set
the encoding independently based on the generation and not based on a
feature.
This may be missing a few overrides to set it off still in some
special cases. Since the flags set during selection should now be
reliably preserved, this should not change codegen for non-strictfp
functions.
As mentioned in the post-commit comments of D81013 -
the mask check API has to assume the shuffle is
not length-changing, but we have not ruled that out
in this code. Use the ShuffleVectorInst call instead.
Follow the model used on Linux, where the clang driver passes the
linker a -u switch to force the profile runtime to be linked in,
rather than having every TU emit a dead function with a reference.
Patch By: mcgrathr
Differential Revision: https://reviews.llvm.org/D79835
Summary:
In current implementation, D2D memcpy is first to copy data back to host and then
copy from host to device. This is very efficient if the device supports D2D
memcpy, like CUDA.
In this patch, D2D memcpy will first try to use native supported driver API. If
it fails, fall back to original way. It is worth noting that D2D memcpy in this
scenerio contains two ideas:
- Same devices: this is the D2D memcpy in the CUDA context.
- Different devices: this is the PeerToPeer memcpy in the CUDA context.
My implementation merges this two parts. It chooses the best API according to
the source device and destination device.
Reviewers: jdoerfert, AndreyChurbanov, grokos
Reviewed By: jdoerfert
Subscribers: yaxunl, guansong, sstefan1, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D80649
recommit e03394c6a6 with fix
When implicit HD function calls a function in device compilation,
if one candidate is an implicit HD function, current resolution rule is:
D wins over HD and H
HD and H are equal
this caused regression when there is an otherwise worse D candidate
This patch changes that to
D, HD and H are all equal
The rationale is that we already know for host compilation there is already
a valid candidate in HD and H candidates that will not cause error. Allowing
HD and H gives us a fall back candidate that will not cause error. If D wins,
that means D has to be a better match otherwise, therefore D should also
be a valid candidate that will not cause error. In this way, we can guarantee
no regression.
Differential Revision: https://reviews.llvm.org/D80450
There are a number of MIR tests using instructions on subtargets where
they don't really exist. These are some of the easy cases that don't
require splitting up test functions.
Summary:
Unlike normal traps, debug traps are allowed to return and can have
additional instructions in the same basic block. Without explicit
backend support for debug traps, they are lowered in ISel as normal
traps. Since normal traps are lowered in the WebAssembly backend to
the UNREACHABLE instruction, which is a terminator, using debug traps
could lead to invalid MBBs when there are additional instructions
after the trap. This patch fixes the issue by lowering debug traps to
a new version of the UNREACHABLE instruction, DEBUG_UNREACHABLE, that
is not a terminator.
An alternative approach would have been to make UNREACHABLE not a
terminator, but that breaks a large number of tests. In particular, it
would require removing the traps inserted after noreturn calls to
@llvm.wasm.throw because otherwise the terminator throw would be
followed by a non-terminator UNREACHABLE and we would be back to
having invalid MBBs. Overall the approach in this patch seems simpler.
Reviewers: aheejin, dschuff
Subscribers: sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D81055
Summary:
Using a forward reference to define a `len` parameter causes a crash.
The underlying cause was that a previously declared type had an
erroneous expression for its `LEN` param value. When this expression
was referenced to evaluate a subsequent expression, bad things happened.
I fixed this by putting in code to detect this case.
Reviewers: tskeith, klausler, DavidTruby
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80593
Part of lowering is to convert the front-end types to their FIR dialect representations. These conversions are done by here in the ConvertType module.
proactively update the code to conform better with LLVM coding conventions
Differential Revision: https://reviews.llvm.org/D81034
We can simplify the code a bit by using iterator_range instead of
plain iterators. Matrix type support here (added in 6f6e91d193)
already uses an iterator_range.
Reviewers: rjmccall, arphaman, jfb, Bigcheese
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D81138
Uday Bondhugula