| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On Jaguar, XCHG has a latency of 1cy and decodes to 2 macro-opcodes. Maximum
throughput for XCHG is 1 IPC. The byte exchange has worse latency and decodes to
1 extra uOP; maximum observed throughput is 0.5 IPC.
```
xchgb %cl, %dl # Latency: 2cy - uOPs: 3 - 2 ALU
xchgw %cx, %dx # Latency: 1cy - uOPs: 2 - 2 ALU
xchgl %ecx, %edx # Latency: 1cy - uOPs: 2 - 2 ALU
xchgq %rcx, %rdx # Latency: 1cy - uOPs: 2 - 2 ALU
```
The reg-mem forms of XCHG are atomic operations with an observed latency of
16cy. The resource usage is similar to the XCHGrr variants. The biggest
difference is obviously the bus-locking, which prevents the LS to issue other
memory uOPs in parallel until the unlocking store uOP is executed.
```
xchgb %cl, (%rsp) # Latency: 16cy - uOPs: 3 - ECX latency: 11cy
xchgw %cx, (%rsp) # Latency: 16cy - uOPs: 3 - ECX latency: 11cy
xchgl %ecx, (%rsp) # Latency: 16cy - uOPs: 3 - ECX latency: 11cy
xchgq %rcx, (%rsp) # Latency: 16cy - uOPs: 3 - ECX latency: 11cy
```
The exchanged in/out register operand becomes available after 11cy from the
start of execution. Added test xchg.s to verify that we correctly see that
register write committed in 11cy (and not 16cy).
Reg-reg XADD instructions have the same latency/throughput than the byte
exchange (register-register variant).
```
xaddb %cl, %dl # latency: 2cy - uOPs: 3 - 3 ALU
xaddw %cx, %dx # latency: 2cy - uOPs: 3 - 3 ALU
xaddl %ecx, %edx # latency: 2cy - uOPs: 3 - 3 ALU
xaddq %rcx, %rdx # latency: 2cy - uOPs: 3 - 3 ALU
```
The non-atomic RM variants have a latency of 11cy, and decode to 4
macro-opcodes. They still consume 2 ALU pipes, and the exchange in/out register
operand becomes available in 3cy (it matches the 'load-to-use latency').
```
xaddb %cl, (%rsp) # latency: 11cy - uOPs: 4 - 3 ALU
xaddw %cx, (%rsp) # latency: 11cy - uOPs: 4 - 3 ALU
xaddl %ecx, (%rsp) # latency: 11cy - uOPs: 4 - 3 ALU
xaddq %rcx, (%rsp) # latency: 11cy - uOPs: 4 - 3 ALU
```
The atomic XADD variants execute in 16cy. The in/out register operand is
available after 11cy from the start of execution.
```
lock xaddb %cl, (%rsp) # latency: 16cy - uOPs: 4 - 3 ALU -- ECX latency: 11cy
lock xaddw %cx, (%rsp) # latency: 16cy - uOPs: 4 - 3 ALU -- ECX latency: 11cy
lock xaddl %ecx, (%rsp) # latency: 16cy - uOPs: 4 - 3 ALU -- ECX latency: 11cy
lock xaddq %rcx, (%rsp) # latency: 16cy - uOPs: 4 - 3 ALU -- ECX latency: 11cy
```
Added test xadd.s to verify those latencies as well as read-advance values.
Differential Revision: https://reviews.llvm.org/D66535
llvm-svn: 369642
|
| |
|
|
|
|
| |
Allows for some cleanup in a lot of SSE/AVX vector splitting code
llvm-svn: 369640
|
| |
|
|
|
|
|
|
|
| |
The CodeGen/Thumb2/mve-vaddv.ll test needed to be amended to reflect the
changes from the above patch.
This reverts commit cd53ff6, reapplying 7c6b229.
llvm-svn: 369638
|
| |
|
|
|
|
|
| |
We must update loop metedata before we moved to parent loop if
it is present.
llvm-svn: 369637
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It broke the bots, see e.g. http://lab.llvm.org:8011/builders/clang-cuda-build/builds/36275/
> This patch fixes shifts by a 128/256 bit shift amount. It also fixes
> codegen for shifts of 32 by delegating to LLVM's default optimisation
> instead of emitting a long shift.
>
> Tests that used to generate long shifts of 32 are updated to check for the
> more optimised codegen.
>
> Differential revision: https://reviews.llvm.org/D66519
>
> llvm-svn: 369626
llvm-svn: 369636
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
legalization.
I don't really understand the costs we're using for fp_to_sint,
but prior to widening legalization we used 20 as the cost for this
via the v2i64->v2f64 entry. That number seems better than the 40
we got with widening legalization. So now we need either a
v2i32->v2f64 entry or a v4i32->v2f64 entry depending on whether
AVX is enabled or not since we skip the first SSE2 table look up
under AVX.
llvm-svn: 369628
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
There was a subtle, but pretty important difference between the Slice
and regular versions of this function. The Slice function was
zero-initializing the rest of the buffer when the read syscall returned
less bytes than expected, while the regular function did not.
This patch removes the inconsistency by making both functions *not*
zero-initialize the buffer. The zeroing code is moved to the
MemoryBuffer class, which is currently the only user of this code. This
makes the API more consistent, and the code shorter.
While in there, I also refactor the functions to return the number of
bytes through the regular return value (via Expected<size_t>) instead of
a separate by-ref argument.
Reviewers: aganea, rnk
Subscribers: kristina, Bigcheese, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66471
llvm-svn: 369627
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This patch fixes shifts by a 128/256 bit shift amount. It also fixes
codegen for shifts of 32 by delegating to LLVM's default optimisation
instead of emitting a long shift.
Tests that used to generate long shifts of 32 are updated to check for the
more optimised codegen.
Differential revision: https://reviews.llvm.org/D66519
llvm-svn: 369626
|
| |
|
|
|
|
|
|
|
|
| |
The patch introduces MakeLibCallOptions struct as suggested by @efriedma on D65497.
The struct contain argument flags which will pass to makeLibCall function.
The patch should not has any functionality changes.
Differential Revision: https://reviews.llvm.org/D65795
llvm-svn: 369622
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Reviewers: wxiao3, LuoYuanke, andrew.w.kaylor, craig.topper, annita.zhang, liutianle, pengfei, xiangzhangllvm, RKSimon, spatel, andreadb
Reviewed By: RKSimon
Subscribers: andreadb, hiraditya, llvm-commits
Tags: #llvm
Patch by Gen Pei (gpei)
Differential Revision: https://reviews.llvm.org/D65933
llvm-svn: 369612
|
| |
|
|
|
|
|
|
|
|
|
|
| |
APIntToHexString returns wrong value ("0000000000000000ffffffffffffffff")
for integer larger than 64 bits, and thus
TargetLoweringObjectFileCOFF::getSectionForConstant returns same section name
for all numbers larger than 64 bits. This patch tries to fix it.
Differential Revision: https://reviews.llvm.org/D66458
Patch by Senran Zhang
llvm-svn: 369610
|
| |
|
|
|
|
|
| |
Buildbots that use GCC failed to compile because overwritten
namespace with variable name
llvm-svn: 369602
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
The intention for this is to allow reading and printing symbols out from
llvm-nm. Tapi file, and Tapi universal follow a similiar format to
their respective MachO Object format.
The tests are dependent on llvm-nm processing tbd files which is why its in D66160
Reviewers: ributzka, steven_wu, lhames
Reviewed By: ributzka, lhames
Subscribers: mgorny, hiraditya, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66159
llvm-svn: 369600
|
| |
|
|
|
|
|
|
|
|
|
|
| |
instructions.
We had an odd combination of WriteJump applied to some memory
instructions and WriteJumpLd applied to register and immediate
instructions.
Thsi should hopefully assign them all correctly.
llvm-svn: 369599
|
| |
|
|
|
|
| |
readability. NFC
llvm-svn: 369598
|
| |
|
|
| |
llvm-svn: 369590
|
| |
|
|
|
|
|
| |
Function can have users that are not instructions, e.g., bitcasts. For
now, we simply give up when we see them.
llvm-svn: 369588
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
AddressRanges
The full GSYM patch started with: https://reviews.llvm.org/D53379
This patch add the ability to encode data using the new llvm::gsym::FileWriter class.
FileWriter is a simplified binary data writer class that doesn't require targets, target definitions, architectures, or require any other optional compile time libraries to be enabled via the build process. This class needs the ability to seek to different spots in the binary data that it produces to fix up offsets and sizes in GSYM data. It currently uses std::ostream over llvm::raw_ostream because llvm::raw_ostream doesn't support seeking which is required when encoding and decoding GSYM data.
AddressRange objects are encoded and decoded to be relative to a base address. This will be the FunctionInfo's start address if the AddressRange is directly contained in a FunctionInfo, or a base address of the containing parent AddressRange or AddressRanges. This allows address ranges to be efficiently encoded using ULEB128 encodings as we encode the offset and size of each range instead of full addresses. This also makes encoded addresses easy to relocate as we just need to relocate one base address.
Differential Revision: https://reviews.llvm.org/D63828
llvm-svn: 369587
|
| |
|
|
| |
llvm-svn: 369580
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Tapi files are YAML files that start with the !tapi tag. The only execption are
TBD v1 files, which don't have a tag. In that case we have to scan a little
further and check if the first key "archs" exists.
This is the first patch in a series of patches to add libObject support for
text-based dynamic library (.tbd) files.
This patch is practically exactly the same as D37820, that was never pushed to master,
and is needed for future commits related to reading tbd files for llvm-nm
Reviewers: ributzka, steven_wu, bollu, espindola, jfb, shafik, jdoerfert
Reviewed By: steven_wu
Subscribers: dexonsmith, llvm-commits
Tags: #llvm, #clang, #sanitizers, #lldb, #libc, #openmp
Differential Revision: https://reviews.llvm.org/D66149
llvm-svn: 369579
|
| |
|
|
| |
llvm-svn: 369577
|
| |
|
|
| |
llvm-svn: 369576
|
| |
|
|
|
|
|
| |
So far we split the unreachable off and placed a new one, this is not
necessary.
llvm-svn: 369575
|
| |
|
|
| |
llvm-svn: 369573
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
unavailable pred.
Currently we do not properly translate addresses with PHIs if LoadBB !=
LI->getParent(), because PHITranslateAddr expects a direct predecessor as argument,
because it considers all instructions outside of the current block to
not requiring translation.
The amount of cases that trigger this should be very low, as most single
predecessor blocks should be folded into their predecessor by GVN before
we actually start with value numbering. It is still not guaranteed to
happen, so we should do PHI translation along all edges between the
loads' block and the predecessor where we have to place a load.
There are a few test cases showing current limits of the PHI translation, which
could be improved later.
Reviewers: spatel, reames, efriedma, john.brawn
Reviewed By: efriedma
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65020
llvm-svn: 369570
|
| |
|
|
|
|
| |
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-bootstrap/builds/13605/
llvm-svn: 369569
|
| |
|
|
| |
llvm-svn: 369567
|
| |
|
|
|
|
|
|
| |
Fixes: https://github.com/emscripten-core/emscripten/issues/8770
Differential Revision: https://reviews.llvm.org/D66508
llvm-svn: 369566
|
| |
|
|
|
|
|
|
|
| |
I might look at improving PR43065 which will require being
able to mark a 256 and 512 bit vector of f16 as Legal.
Differential Revision: https://reviews.llvm.org/D66515
llvm-svn: 369565
|
| |
|
|
|
|
|
|
| |
In case of expanding `lw/sw $reg, symbol($reg)` instruction for PIC it's
enough to call the `loadAndAddSymbolAddress` method. Additional work
performed by the `expandLoadAddress` is not required here.
llvm-svn: 369563
|
| |
|
|
| |
llvm-svn: 369562
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
These calls change the order in which some nodes are processed and so have an effect on codegen.
The change in fixup-bw-copy.ll is due to (and (load anyext)) gets transformed into (load zext) while previously the and was removed by SimplifyDemandedBits, so the (load anyext) remained.
Reviewers: craig.topper, efriedma, RKSimon, lebedev.ri
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66543
llvm-svn: 369561
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We cannot create null constants for certain types, e.g. VoidTy,
FunctionTy or LabelTy. getNullValue asserts if we pass in an
unsupported type. We should also check for opaque types, but I'm not
sure how.
This fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14795.
Reviewers: t.p.northover, jfb, vsk
Reviewed By: vsk
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65897
llvm-svn: 369557
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Since clang does not support comment style fallthrough annotations
these should be switched.
Reviewers: aaron.ballman, nickdesaulniers, xbolva00
Reviewed By: aaron.ballman, nickdesaulniers, xbolva00
Subscribers: xbolva00, nickdesaulniers, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66487
llvm-svn: 369549
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
preserve it.
Summary:
Add a flag to the FunctionToLoopAdaptor that allows enabling MemorySSA only for the loop pass managers that are known to preserve it.
If an LPM is known to have only loop transforms that *all* preserve MemorySSA, then use MemorySSA if `EnableMSSALoopDependency` is set.
If an LPM has loop passes that do not preserve MemorySSA, then the flag passed is `false`, regardless of the value of `EnableMSSALoopDependency`.
When using a custom loop pass pipeline via `passes=...`, use keyword `loop` vs `loop-mssa` to use MemorySSA in that LPM. If a loop that does not preserve MemorySSA is added while using the `loop-mssa` keyword, that's an error.
Add the new `loop-mssa` keyword to a few tests where a difference occurs when enabling MemorySSA.
Reviewers: chandlerc
Subscribers: mehdi_amini, Prazek, george.burgess.iv, sanjoy.google, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66376
llvm-svn: 369548
|
| |
|
|
|
|
|
|
| |
This is necessary for handling <3 x s16> on AMDGPU, assuming this
should be handled as 2 separate legalization actions. The alternative
would be for fewerElementsVector to handle 3->2.
llvm-svn: 369547
|
| |
|
|
|
|
|
| |
This is just some formatting cleanup, prior to the masked load and store patch
in D66534.
llvm-svn: 369545
|
| |
|
|
|
|
| |
Noticed while looking at pr43028.
llvm-svn: 369541
|
| |
|
|
| |
llvm-svn: 369533
|
| |
|
|
|
|
|
| |
Differential Revision: https://reviews.llvm.org/D63731
Reviewers: qcolombet, rampitec
llvm-svn: 369532
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
This is patch is part of a serie to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: jfb
Subscribers: hiraditya, dexonsmith, llvm-commits, courbet
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65800
llvm-svn: 369531
|
| |
|
|
|
|
|
|
|
| |
DWARFUnitHeader::getLength() returns uint64_t.
DWARFUnit::getLength() should do the same.
Differential Revision: https://reviews.llvm.org/D66472
llvm-svn: 369529
|
| |
|
|
|
|
|
|
|
| |
The hint instructions are enabled by default (if the standard C extension is
enabled). To disable them pass -mattr=-rvc-hints.
Differential Revision: https://reviews.llvm.org/D62592
llvm-svn: 369528
|
| |
|
|
| |
llvm-svn: 369520
|
| |
|
|
|
|
|
|
|
|
|
| |
An intermediate extend is used to widen the narrow operand to the width of
the other (wider) operand. At that point, we have the same logic as the
existing transform that was restricted to folds of equal width zext/sext.
This mostly solves PR42700:
https://bugs.llvm.org/show_bug.cgi?id=42700
llvm-svn: 369519
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
The code for serializing minidumps was living in MinidumpYAML.cpp
so that it would be accessible from unit tests. While this had its
advantages, it was also unfortunate because it broke symmetry with all
other yaml2obj serializers.
Fortunately, nowadays all of yaml2obj is a library, so we don't need to
do anything special. This patch improves the code consistency by moving
the serialization code to MinidumpEmitter.cpp to match the style used in
other backends. It also removes the writeAsBinary entry point in favor
of the more general convertYAML interface.
This patch is just massaging the code a bit. There shouldn't be any
functional change here.
Reviewers: jhenderson, abrachet
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66474
llvm-svn: 369517
|
| |
|
|
|
|
|
|
| |
NarrowScalar G_ZEXTLOAD and G_SEXTLOAD to s32 for MIPS32.
Differential Revision: https://reviews.llvm.org/D66205
llvm-svn: 369512
|
| |
|
|
|
|
|
|
| |
NarrowScalar G_ZEXT and G_SEXT to s32 for MIPS32.
Differential Revision: https://reviews.llvm.org/D66204
llvm-svn: 369511
|
| |
|
|
|
|
|
|
|
| |
r351882 allows different type for shift amount then result and value
being shifted. Fix MIPS Legalizer rules to take r351882 into account.
Differential Revision: https://reviews.llvm.org/D66203
llvm-svn: 369510
|
| |
|
|
|
|
|
|
|
| |
Add NarrowScalar for G_TRUNC when NarrowTy is half the size of source.
NarrowScalar G_TRUNC to s32 for MIPS32.
Differential Revision: https://reviews.llvm.org/D66202
llvm-svn: 369509
|