The post register allocator scheduler can generate poor schedules
because the scoreboard hazard recognizer is unable to identify
hazards for Hexagon precisely. Instead, Hexagon should use a DFA
based hazard recognizer.
Patch by Brendon Cahoon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@277143 91177308-0d34-0410-b5e6-96231b3b80d8
Normally, CFI instructions should be inserted after allocframe, but
if allocframe is in the same packet with a call, the CFI instructions
should be inserted before that packet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@277020 91177308-0d34-0410-b5e6-96231b3b80d8
Before adding a new preheader block, check if there is a candidate block
where the loop setup could be placed speculatively. This will be off by
default.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276919 91177308-0d34-0410-b5e6-96231b3b80d8
The callee-saved registers that are saved in a function are not pristine,
and so they can be defined and used. In case of shrink-wrapping though,
there are blocks that are outside of the save/restore range, and in those
blocks the saved registers must be treated as pristine. To avoid any uses
of these registers, add them as live-in in all those blocks.
This was already done for blocks reaching function exits after restore,
add code that does the same for blocks reached from the function entry
before save.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276886 91177308-0d34-0410-b5e6-96231b3b80d8
- Generate vector post-increment stores more aggressively.
- Predicate post-increment and vector stores in early if-conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276800 91177308-0d34-0410-b5e6-96231b3b80d8
Consider this case:
vreg1 = A2_zxth vreg0 (1)
...
vreg2 = A2_zxth vreg1 (2)
Redundant instruction elimination could delete the instruction (1)
because the user (2) only cares about the low 16 bits. Then it could
delete (2) because the input is already zero-extended. The problem
is that the properties allowing each individual instruction to be
deleted depend on the existence of the other instruction, so either
one can be deleted, but not both.
The existing check for this situation in RIE was insufficient. The
fix is to update all dependent cells when an instruction is removed
(replaced via COPY) in RIE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276792 91177308-0d34-0410-b5e6-96231b3b80d8
Change the bit simplifier to generate REG_SEQUENCE instructions in
addition to COPY, which will handle cases of word insert/extract.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276787 91177308-0d34-0410-b5e6-96231b3b80d8
When the packetizer wants to put a store to a stack slot in the same
packet with an allocframe, it updates the store offset to reflect the
value of SP before it is updated by allocframe. If the store cannot
be packetized with the allocframe after all, the offset needs to be
updated back to the previous value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276749 91177308-0d34-0410-b5e6-96231b3b80d8
Some targets, notably AArch64 for ILP32, have different relocation encodings
based upon the ABI. This is an enabling change, so a future patch can use the
ABIName from MCTargetOptions to chose which relocations to use. Tested using
check-llvm.
The corresponding change to clang is in: http://reviews.llvm.org/D16538
Patch by: Joel Jones
Differential Revision: https://reviews.llvm.org/D16213
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276654 91177308-0d34-0410-b5e6-96231b3b80d8
- FuncNode::findBlock traverses the function every time. Avoid using it,
and keep a cache of block addresses in DataFlowGraph instead.
- The operator[] in the map of definition stacks was very slow. Replace
the map with unordered_map.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@276429 91177308-0d34-0410-b5e6-96231b3b80d8
Schedule a load and its use in the same packet in MISched. Previously,
isResourceAvailable was returning false for dependences in the same
packet, which prevented MISched from packetizing a load and its use in
the same packet for v60.
Patch by Ikhlas Ajbar.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275804 91177308-0d34-0410-b5e6-96231b3b80d8
The machine scheduler needs to account for available resources
more accurately in order to avoid scheduling an instruction that
forces a new packet to be created.
This occurs in two ways: First, an instruction without an available
resource may have a large priority due to other metrics and be
scheduled when there are other instructions with available resources.
Second, an instruction with a non-zero latency may become available
prematurely. In both these cases, we attempt change the priority
in order to allow a better instruction to be scheduled.
Patch by Brendon Cahoon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275793 91177308-0d34-0410-b5e6-96231b3b80d8
An instruction may have multiple predecessors that are candidates
for using .cur. However, only one of them can use .cur in the
packet. When this case occurs, we need to make sure that only
one of the dependences gets a 0 latency value.
Patch by Brendon Cahoon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275790 91177308-0d34-0410-b5e6-96231b3b80d8
The Hexagon schedulers need to handle instructions with a latency
of 0 or 2 more accurately. The problem, in v60, is that a dependence
between two instructions with a 2 cycle latency can use a .cur version
of the source to achieve a 0 cycle latency when the use is in the
same packet. Any othe use, must be at least 2 packets later, or a
stall occurs. In other words, the compiler does not want to schedule
the dependent instructions 1 cycle later.
To achieve this, the latency adjustment code allows only a single
dependence to have a zero latency. All other instructions have the
other value, which is typically 2 cycles. We use a heuristic to
determine which instruction gets the 0 latency.
The Hexagon machine scheduler was also changed to increase the cost
associated with 0 latency dependences than can be scheduled in the
same packet.
Patch by Brendon Cahoon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275625 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Instead, we take a single flags arg (a bitset).
Also add a default 0 alignment, and change the order of arguments so the
alignment comes before the flags.
This greatly simplifies many callsites, and fixes a bug in
AMDGPUISelLowering, wherein the order of the args to getLoad was
inverted. It also greatly simplifies the process of adding another flag
to getLoad.
Reviewers: chandlerc, tstellarAMD
Subscribers: jholewinski, arsenm, jyknight, dsanders, nemanjai, llvm-commits
Differential Revision: http://reviews.llvm.org/D22249
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275592 91177308-0d34-0410-b5e6-96231b3b80d8
- Treat bitwise OR with a frame index as an ADD wherever possible, fold it
into addressing mode.
- Extend patterns for memops to allow memops with frame indexes as address
operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275569 91177308-0d34-0410-b5e6-96231b3b80d8
On Hexagon is it legal to packetize the instructions setting up call
arguments with the call instruction itself. This was already done,
except for tail calls. Make sure tail calls are handled as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275458 91177308-0d34-0410-b5e6-96231b3b80d8
Avoid implicit iterator conversions from MachineInstrBundleIterator to
MachineInstr* in the Hexagon backend, mostly by preferring MachineInstr&
over MachineInstr* and switching to range-based for loops.
There's a long tail of API cleanup here, but I'm planning to leave the
rest to the Hexagon maintainers. HexagonInstrInfo defines many of its
own predicates, and most of them still take MachineInstr*. Some of
those actually check for nullptr, so I didn't feel comfortable changing
them to MachineInstr& en masse.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@275142 91177308-0d34-0410-b5e6-96231b3b80d8
Stop using an implicit conversion from the return of
MachineBasicBlock::getFirstTerminator to MachineInstr*. In two cases,
directly dereference to a MachineInstr& since later code assumes it's
valid. In a third case, change to an iterator since later code checks
against MachineBasicBlock::end.
Although the fix for the third case avoids undefined behaviour, I expect
this doesn't cause a functionality change in practice (since the basic
block already has a terminator).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274898 91177308-0d34-0410-b5e6-96231b3b80d8
This could of course be a simple binary search with no global state
involved at all if someone cares enough. Just don't make everyone
linking the hexagon backend pay for it on process startup and shutdown.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274437 91177308-0d34-0410-b5e6-96231b3b80d8
This is a mechanical change to make TargetLowering API take MachineInstr&
(instead of MachineInstr*), since the argument is expected to be a valid
MachineInstr. In one case, changed a parameter from MachineInstr* to
MachineBasicBlock::iterator, since it was used as an insertion point.
As a side effect, this removes a bunch of MachineInstr* to
MachineBasicBlock::iterator implicit conversions, a necessary step
toward fixing PR26753.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274287 91177308-0d34-0410-b5e6-96231b3b80d8
