Summary:
In this change we introduce the notion of a "flight data recorder" mode
for XRay logging, where XRay logs in-memory first, and write out data
on-demand as required (as opposed to the naive implementation that keeps
logging while tracing is "on"). This depends on D26232 where we
implement the core data structure for holding the buffers that threads
will be using to write out records of operation.
This implementation only currently works on x86_64 and depends heavily
on the TSC math to write out smaller records to the inmemory buffers.
Also, this implementation defines two different kinds of records with
different sizes (compared to the current naive implementation): a
MetadataRecord (16 bytes) and a FunctionRecord (8 bytes). MetadataRecord
entries are meant to write out information like the thread ID for which
the metadata record is defined for, whether the execution of a thread
moved to a different CPU, etc. while a FunctionRecord represents the
different kinds of function call entry/exit records we might encounter
in the course of a thread's execution along with a delta from the last
time the logging handler was called.
While this implementation is not exactly what is described in the
original XRay whitepaper, this one gives us an initial implementation
that we can iterate and build upon.
Reviewers: echristo, rSerge, majnemer
Subscribers: mehdi_amini, llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D27038
llvm-svn: 293015
Sometime clang would be supplied -fobjc-arc -f(no)objc-arc-exceptions
and then later disable ARC with -fno-objc-arc, which only negate first
option, but not the latter, resulting usused argument warning. Silence
this warning only when -fno-objc-arc option is present.
Patch by Onha Choe!
llvm-svn: 293014
Looks like our cmake goop for handling .inc->td dependencies doesn't
track the .td files.
This manifests as cmake complaining about missing files since r293009.
Force a rerun to avoid that.
llvm-svn: 293012
loops.
We do this by reconstructing the newly added loops after the unroll
completes to avoid threading pass manager details through all the mess
of the unrolling infrastructure.
I've enabled some extra assertions in the LPM to try and catch issues
here and enabled a bunch of unroller tests to try and make sure this is
sane.
Currently, I'm manually running loop-simplify when needed. That should
go away once it is folded into the LPM infrastructure.
Differential Revision: https://reviews.llvm.org/D28848
llvm-svn: 293011
This surprisingly isn't NFC because there are patterns to select GPR
sub to SUBSWrr (rather than SUBWrr/rs); SUBS is later optimized to
SUB if NZCV is dead. From ISel's perspective, both are fine.
llvm-svn: 293010
Summary:
When we decide that the result of the invoke instruction need to be spilled, we need to insert the spill into a block that is on the normal edge coming out of the invoke instruction. (Prior to this change the code would insert the spill immediately after the invoke instruction, which breaks the IR, since invoke is a terminator instruction).
In the following example, we will split the edge going into %cont and insert the spill there.
```
%r = invoke double @print(double 0.0) to label %cont unwind label %pad
cont:
%0 = call i8 @llvm.coro.suspend(token none, i1 false)
switch i8 %0, label %suspend [i8 0, label %resume
i8 1, label %cleanup]
resume:
call double @print(double %r)
```
Reviewers: majnemer
Reviewed By: majnemer
Subscribers: mehdi_amini, llvm-commits, EricWF
Differential Revision: https://reviews.llvm.org/D29102
llvm-svn: 293006
This patch adds support for codegen of 'target teams' on the host.
This combined directive has two captured statements, one for the
'teams' region, and the other for the 'parallel'.
This target teams region is offloaded using the __tgt_target_teams()
call. The patch sets the number of teams as an argument to
this call.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29084
llvm-svn: 293005
Summary:
Clang appears to always use name as specified on the command
line, whereas gcc uses the name as specified in the linemarker at the
first line when compiling a preprocessed source. This results mismatch
between two compilers in FILE symbol table entry. This patch makes clang
to resemble gcc's behavior in finding the original source file name and
use it as an input file name.
Even with this patch, values of FILE symbol table entry may still be
different because clang uses dirname+basename for the entry whlie gcc
uses basename only. I'll write a patch for that once this patch is
committed.
Reviewers: dblaikie, inglorion
Reviewed By: inglorion
Subscribers: inglorion, aprantl, bruno
Differential Revision: https://reviews.llvm.org/D28796
llvm-svn: 293004
This patch adds support for codegen of 'target teams' on the host.
This combined directive has two captured statements, one for the
'teams' region, and the other for the 'parallel'.
This target teams region is offloaded using the __tgt_target_teams()
call. The patch sets the number of teams as an argument to
this call.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29084
llvm-svn: 293001
Summary:
This lets you select which sort of spilling you want, either s[0:1] or 64-bit loads from s[0:1].
Patch By: Dave Airlie
Reviewers: nhaehnle, arsenm, tstellarAMD
Reviewed By: arsenm
Subscribers: mareko, llvm-commits, kzhuravl, wdng, yaxunl, tony-tye
Differential Revision: https://reviews.llvm.org/D25428
llvm-svn: 293000
This patch adds support for the Spmd construct 'target parallel' on the
NVPTX device. This involves ignoring the num_threads clause on the device
since the number of threads in this combined construct is already set on
the host through the call to __tgt_target_teams().
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29083
llvm-svn: 292999
The num_threads-clause on the combined directive applies to the
'parallel' region of this construct. We modify the NumThreadsClause
class to capture the clause expression within the 'target' region.
The offload runtime call for 'target parallel' is changed to
__tgt_target_teams() with 1 team and the number of threads set by
this clause or a default if none.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29082
llvm-svn: 292997
There was a bug here where we were using p0 instead of s32 for the
selector type in the landingpad. Instead of hardcoding these types we
should get the types from the landingpad instruction directly.
Note that we replicate an assert from SDAG here to only support
two-valued landingpads.
llvm-svn: 292995
for CPU_SUBTYPE_ARM_V7S and CPU_SUBTYPE_ARM_V7K.
For these two cpusubtypes they should default to a cortex-a7 CPU
to give proper disassembly without a -mcpu= flag.
rdar://27431703
llvm-svn: 292993
Summary: This commit adds an option to set PC to the entry point of the file loaded using "target module load" command. In D28804, Greg asked me to separate this part under a different option.
Reviewers: clayborg
Reviewed By: clayborg
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D28944
llvm-svn: 292989
Summary: LLDB was using packet size advertised by the target as the max memory size to write in one go. It is wrong because packets have other overhead apart from memory payload. Also memory transferred through 'm' and 'M' packets needs 2 bytes in packet to transfer 1 of memory.
Reviewers: clayborg
Reviewed By: clayborg
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D28808
llvm-svn: 292987
The sequence like this:
v_cmpx_le_f32_e32 vcc, 0, v0
s_branch BB0_30
s_cbranch_execnz BB0_30
; BB#29:
exp null off, off, off, off done vm
s_endpgm
BB0_30:
; %endif110
is likely wrong. The s_branch instruction will unconditionally jump
to BB0_30 and the skip block (exp done + endpgm) inserted for
performing the kill instruction will never be executed. This results
in a GPU hang with Star Ruler 2.
The s_branch instruction is added during the "Control Flow Optimizer"
pass which seems to re-organize the basic blocks, and we assume
that SI_KILL_TERMINATOR is always the last instruction inside a
basic block. Thus, after inserting a skip block we just go to the
next BB without looking at the subsequent instructions after the
kill, and the s_branch op is never removed.
Instead, we should remove the unconditional out branches and let
skip the two instructions if the exec mask is non-zero.
This patch fixes the GPU hang and doesn't introduce any regressions
with "make check".
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99019
Patch by Samuel Pitoiset <samuel.pitoiset@gmail.com>
llvm-svn: 292985
This switches to the workaround that HSA defaults to
for the mesa path.
This should be applied to the 4.0 branch.
Patch by Vedran Miletić <vedran@miletic.net>
llvm-svn: 292982
TSan recently got the "ignore_noninstrumented_modules" flag, which disables tracking of read and writes that come from noninstrumented modules (via interceptors). This is a way of suppressing false positives coming from system libraries and other noninstrumented code. This patch turns this on by default on Darwin, where it's supposed to replace the previous solution, "ignore_interceptors_accesses", which disables tracking in *all* interceptors. The new approach should re-enable TSan's ability to find races via interceptors on Darwin.
Differential Revision: https://reviews.llvm.org/D29041
llvm-svn: 292981
Summary: In iterative sample pgo where profile is collected from PGOed binary, we may see indirect call targets promoted and inlined in the profile. Before profile annotation, we need to make this happen in order to annotate correctly on IR. This patch explicitly promotes these indirect calls and inlines them before profile annotation.
Reviewers: xur, davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29040
llvm-svn: 292979
When demangling a CV-qualified function type with a final reference type
parameter, we would treat the reference type parameter as a r-value ref
accidentally. This would result in the improper decoration of the
function type itself.
Resolves PR31741!
llvm-svn: 292976
Reason: broke ASAN bots with a global buffer overflow.
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/2291
Each test contains 20-30K test cases but takes only several (from 4 to 10)
seconds to complete on average machine. The tests cover the majority of
AMDGPU Gfx7/Gfx8 instructions, including many dark corners, and intended
to quickly find out if something is broken.
llvm-svn: 292974
When checking if the type is a r-value ref, we would not do a complete
check. This would result in us treating a trailing parameter reference
`&)` as a r-value ref, and improperly inject the cv qualifier on the
type. We now correctly demangle the type `KFvRmE` as a constant
function rather than a constant reference.
Fixes PR31741!
llvm-svn: 292973
Summary:
GVNHoist performs all the optimizations that MLSM does to loads, in a
more general way, and in a faster time bound (MLSM is N^3 in most
cases, N^4 in a few edge cases).
This disables the load portion.
Note that the way ld_hoist_st_sink.ll is written makes one think that
the loads should be moved to the while.preheader block, but
1. Neither MLSM nor GVNHoist do it (they both move them to identical places).
2. MLSM couldn't possibly do it anyway, as the while.preheader block
is not the head of the diamond, while.body is. (GVNHoist could do it
if it was legal).
3. At a glance, it's not legal anyway because the in-loop load
conflict with the in-loop store, so the loads must stay in-loop.
I am happy to update the test to use update_test_checks so that
checking is tighter, just was going to do it as a followup.
Note that i can find no particular benefit to the store portion on any
real testcase/benchmark i have (even size-wise). If we really still
want it, i am happy to commit to writing a targeted store sinker, just
taking the code from the MemorySSA port of MergedLoadStoreMotion
(which is N^2 worst case, and N most of the time).
We can do what it does in a much better time bound.
We also should be both hoisting and sinking stores, not just sinking
them, anyway, since whether we should hoist or sink to merge depends
basically on luck of the draw of where the blockers are placed.
Nonetheless, i have left it alone for now.
Reviewers: chandlerc, davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29079
llvm-svn: 292971
When demangling a CV-qualified function type with a final parameter with
a reference type, we would insert the CV qualification on the parameter
rather than the function, and in the process adjust the insertion point
by one extra, splitting the type name. This avoids doing so, even
though the attribution is still incorrect.
llvm-svn: 292965
Dean Michael Berris