For VLD3/VLD4 with double-spaced registers, add the implicit use of the
super register for both the instruction loading the even registers and the
instruction loading the odd registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113452 91177308-0d34-0410-b5e6-96231b3b80d8
Omission of memory form of PI2PD is intentional; this
does not use an MMX register and does not put the chip
into MMX mode (PI2PS, oddly enough, does).
Operands of PI2PS follow the gcc builtin, not Intel.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113388 91177308-0d34-0410-b5e6-96231b3b80d8
nodes to emit shuffles and don't do isel mask matching anymore.
- Add the selection of the remaining shuffle opcode (movddup)
- Introduce two new functions to "recognize" where we may get
potential folds and add several comments to them explaining why
they are not yet in the desidered shape.
- Add more patterns to fallback the case where we select
a specific shuffle opcode as if it could fold a load, but it
can't, so remap to a valid instruction.
- Add a couple of FIXMEs to address in the following days once
there's a good solution to the current folding problem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113369 91177308-0d34-0410-b5e6-96231b3b80d8
option to disable base pointer usage, pay attention to it when deciding
if we can realign (if no base pointer and VLAs, we can't).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113366 91177308-0d34-0410-b5e6-96231b3b80d8
register must be one of the destination registers for the load. Otherwise,
the tLDM instruction will write-back to the base register, which isn't what's
desired (otherwise, we'd have a t2LDM_UPD instead).
rdar://8394087
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113297 91177308-0d34-0410-b5e6-96231b3b80d8
always be disambiguated as sldtw. sldtw and sldtq with
a mem operands have the same effect, but sldtw is more
compact. Force it to sldtw, resolving rdar://8017530
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113186 91177308-0d34-0410-b5e6-96231b3b80d8
of a mneumonic, report operand errors with better location
info. For example, we now report:
t.s:6:14: error: invalid operand for instruction
cwtl $1
^
but we fail for common cases like:
t.s:11:4: error: invalid operand for instruction
addl $1, $1
^
because we don't know if this is supposed to be the reg/imm or imm/reg
form.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113178 91177308-0d34-0410-b5e6-96231b3b80d8
failed because a subtarget feature was not enabled. Use this to
remove a bunch of hacks from the X86AsmParser for rejecting things
like popfl in 64-bit mode. Previously these hacks weren't needed,
but were important to get a message better than "invalid instruction"
when used in the wrong mode.
This also fixes bugs where pushal would not be rejected correctly in
32-bit mode (just pusha).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113166 91177308-0d34-0410-b5e6-96231b3b80d8
into the middle of the class, and rework how the different sections of
the generated file are conditionally included for simplicity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113163 91177308-0d34-0410-b5e6-96231b3b80d8
Since mem2reg isn't run at -O0, we get a ton of reloads from the stack,
for example, before, this code:
int foo(int x, int y, int z) {
return x+y+z;
}
used to compile into:
_foo: ## @foo
subq $12, %rsp
movl %edi, 8(%rsp)
movl %esi, 4(%rsp)
movl %edx, (%rsp)
movl 8(%rsp), %edx
movl 4(%rsp), %esi
addl %edx, %esi
movl (%rsp), %edx
addl %esi, %edx
movl %edx, %eax
addq $12, %rsp
ret
Now we produce:
_foo: ## @foo
subq $12, %rsp
movl %edi, 8(%rsp)
movl %esi, 4(%rsp)
movl %edx, (%rsp)
movl 8(%rsp), %edx
addl 4(%rsp), %edx ## Folded load
addl (%rsp), %edx ## Folded load
movl %edx, %eax
addq $12, %rsp
ret
Fewer instructions and less register use = faster compiles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113102 91177308-0d34-0410-b5e6-96231b3b80d8
checking each standalone condition and decide whether emit target
specific nodes or remove the condition if it's already matched before.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113031 91177308-0d34-0410-b5e6-96231b3b80d8
"Use target specific nodes instead of relying in unpckl and
unpckh pattern fragments during isel time. Also place a
depth limit in getShuffleScalarElt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113020 91177308-0d34-0410-b5e6-96231b3b80d8
mask pattern fragment", which depends on r112934, which introduced some infinite
loop and select failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112998 91177308-0d34-0410-b5e6-96231b3b80d8
"For ARM stack frames that utilize variable sized objects and have either
large local stack areas or require dynamic stack realignment, allocate a
base register via which to access the local frame. This allows efficient
access to frame indices not accessible via the FP (either due to being out
of range or due to dynamic realignment) or the SP (due to variable sized
object allocation). In particular, this greatly improves efficiency of access
to spill slots in Thumb functions which contain VLAs."
r112986 fixed a latent bug exposed by the above.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112989 91177308-0d34-0410-b5e6-96231b3b80d8
alignment should be performed. Otherwise dynamic realignment may trigger
when the register allocator has already used the frame pointer as a general
purpose register. That is, we need to make sure that the list of reserved
registers doesn't change after register allocation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112986 91177308-0d34-0410-b5e6-96231b3b80d8
instructions prior to regalloc. Since it's getting a little close to
the 2.8 branch deadline, I'll have to leave the rest of the instructions
handled by the NEONPreAllocPass for now, but I didn't want to leave half
of the VLD instructions converted and the other half not.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112983 91177308-0d34-0410-b5e6-96231b3b80d8
vabd intrinsic and add and/or zext operations. In the case of vaba, this
also avoids the need for a DAG combine pattern to combine vabd with add.
Update tests. Auto-upgrade the old intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112941 91177308-0d34-0410-b5e6-96231b3b80d8
- Teach getShuffleScalarElt how to handle more target
specific nodes, so the DAGCombine can make use of it.
- Add another hack to avoid the node update problem
during legalization. More description on the comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112934 91177308-0d34-0410-b5e6-96231b3b80d8
large local stack areas or require dynamic stack realignment, allocate a
base register via which to access the local frame. This allows efficient
access to frame indices not accessible via the FP (either due to being out
of range or due to dynamic realignment) or the SP (due to variable sized
object allocation). In particular, this greatly improves efficiency of access
to spill slots in Thumb functions which contain VLAs.
rdar://7352504
rdar://8374540
rdar://8355680
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112883 91177308-0d34-0410-b5e6-96231b3b80d8
there are clearly no stores between the load and the store. This fixes
this miscompile reported as PR7833.
This breaks the test/CodeGen/X86/narrow_op-2.ll optimization, which is
safe, but awkward to prove safe. Move it to X86's README.txt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112861 91177308-0d34-0410-b5e6-96231b3b80d8
add, and subtract operations with zero-extended or sign-extended vectors.
Update tests. Add auto-upgrade support for the old intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112773 91177308-0d34-0410-b5e6-96231b3b80d8
check more strict, breaking some cases not checked in the
testsuite, but also exposes some foldings not done before,
as this example:
movaps (%rdi), %xmm0
movaps (%rax), %xmm1
movaps %xmm0, %xmm2
movss %xmm1, %xmm2
shufps $36, %xmm2, %xmm0
now is generated as:
movaps (%rdi), %xmm0
movaps %xmm0, %xmm1
movlps (%rax), %xmm1
shufps $36, %xmm1, %xmm0
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112753 91177308-0d34-0410-b5e6-96231b3b80d8
int x(int t) {
if (t & 256)
return -26;
return 0;
}
We generate this:
tst.w r0, #256
mvn r0, #25
it eq
moveq r0, #0
while gcc generates this:
ands r0, r0, #256
it ne
mvnne r0, #25
bx lr
Scandalous really!
During ISel time, we can look for this particular pattern. One where we have a
"MOVCC" that uses the flag off of a CMPZ that itself is comparing an AND
instruction to 0. Something like this (greatly simplified):
%r0 = ISD::AND ...
ARMISD::CMPZ %r0, 0 @ sets [CPSR]
%r0 = ARMISD::MOVCC 0, -26 @ reads [CPSR]
All we have to do is convert the "ISD::AND" into an "ARM::ANDS" that sets [CPSR]
when it's zero. The zero value will all ready be in the %r0 register and we only
need to change it if the AND wasn't zero. Easy!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112664 91177308-0d34-0410-b5e6-96231b3b80d8
kill flag.
This could cause duplicate kill flags when the same register was used twice in a
continuous sequence of STRs.
There is no small test case. <rdar://problem/8218046>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112534 91177308-0d34-0410-b5e6-96231b3b80d8
help relieve register pressure a bit. Recalculating the local address is
almost always going to be better than spilling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112503 91177308-0d34-0410-b5e6-96231b3b80d8
operand is killed, add it to the expanded instruction as an implicit kill
operand instead of marking the individual subregs with kill flags. This
should work better in general and also handles the case for VST3 where one
of the subregs was not referenced in the expanded instruction and so was
not marked killed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112494 91177308-0d34-0410-b5e6-96231b3b80d8
optional modified register (instead of reg0). Along with r112461 it will make
sure that the optional define of CPSR is marked as "def" and will thus mark the
instructions using these classes (t2ANDS*) as setting the 's' flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112462 91177308-0d34-0410-b5e6-96231b3b80d8
IR add/sub operations with one or both operands sign- or zero-extended.
Auto-upgrade the old intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112416 91177308-0d34-0410-b5e6-96231b3b80d8
times. This patch causes llc and llvm-mc (which both default to
verbose-asm) to print out comments after a few common shuffle
instructions which indicates the shuffle mask, e.g.:
insertps $113, %xmm3, %xmm0 ## xmm0 = zero,xmm0[1,2],xmm3[1]
unpcklps %xmm1, %xmm0 ## xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
pshufd $1, %xmm1, %xmm1 ## xmm1 = xmm1[1,0,0,0]
This is carefully factored to keep the information extraction (of the
shuffle mask) separate from the printing logic. I plan to move the
extraction part out somewhere else at some point for other parts of
the x86 backend that want to introspect on the behavior of shuffles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112387 91177308-0d34-0410-b5e6-96231b3b80d8
when the top elements of a vector are undefined. This happens all
the time for X86-64 ABI stuff because only the low 2 elements of
a 4 element vector are defined. For example, on:
_Complex float f32(_Complex float A, _Complex float B) {
return A+B;
}
We used to produce (with SSE2, SSE4.1+ uses insertps):
_f32: ## @f32
movdqa %xmm0, %xmm2
addss %xmm1, %xmm2
pshufd $16, %xmm2, %xmm2
pshufd $1, %xmm1, %xmm1
pshufd $1, %xmm0, %xmm0
addss %xmm1, %xmm0
pshufd $16, %xmm0, %xmm1
movdqa %xmm2, %xmm0
unpcklps %xmm1, %xmm0
ret
We now produce:
_f32: ## @f32
movdqa %xmm0, %xmm2
addss %xmm1, %xmm2
pshufd $1, %xmm1, %xmm1
pshufd $1, %xmm0, %xmm3
addss %xmm1, %xmm3
movaps %xmm2, %xmm0
unpcklps %xmm3, %xmm0
ret
This implements rdar://8368414
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112378 91177308-0d34-0410-b5e6-96231b3b80d8
a new EltStride variable instead of reusing NumElems variable
for a non-obvious purpose. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112377 91177308-0d34-0410-b5e6-96231b3b80d8
Also teach this logic how to handle target specific shuffles if
needed, this is necessary while searching recursively for zeroed
scalar elements in vector shuffle operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112348 91177308-0d34-0410-b5e6-96231b3b80d8
the special values that for ARM would be used with IB or DA modes. Fall
through and consider materializing a new base address is it would be
profitable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112329 91177308-0d34-0410-b5e6-96231b3b80d8
all the other LDM/STM instructions. This fixes asm printer crashes when
compiling with -O0. I've changed one of the NEON tests (vst3.ll) to run
with -O0 to check this in the future.
Prior to this change VLDM/VSTM used addressing mode #5, but not really.
The offset field was used to hold a count of the number of registers being
loaded or stored, and the AM5 opcode field was expanded to specify the IA
or DB mode, instead of the standard ADD/SUB specifier. Much of the backend
was not aware of these special cases. The crashes occured when rewriting
a frameindex caused the AM5 offset field to be changed so that it did not
have a valid submode. I don't know exactly what changed to expose this now.
Maybe we've never done much with -O0 and NEON. Regardless, there's no longer
any reason to keep a count of the VLDM/VSTM registers, so we can use
addressing mode #4 and clean things up in a lot of places.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112322 91177308-0d34-0410-b5e6-96231b3b80d8
still having a significant effect. It shouldn't be now that the pre-RA
virtual base reg stuff is in. Assuming that's valididated by the nightly
testers, we can simplify a lot of the PEI frame index code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112220 91177308-0d34-0410-b5e6-96231b3b80d8
fix: add a flag to MapValue and friends which indicates whether
any module-level mappings are being made. In the common case of
inlining, no module-level mappings are needed, so MapValue doesn't
need to examine non-function-local metadata, which can be very
expensive in the case of a large module with really deep metadata
(e.g. a large C++ program compiled with -g).
This flag is a little awkward; perhaps eventually it can be moved
into the ClonedCodeInfo class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112190 91177308-0d34-0410-b5e6-96231b3b80d8
comparison with 0. These two pieces of code should give identical results:
rsbs r1, r1, 0
cmp r0, r1
mov r0, #0
it ls
mov r0, #1
and:
cmn r0, r1
mov r0, #0
it ls
mov r0, #1
However, the CMN gives the *opposite* result when r1 is 0. This is because the
carry flag is set in the CMP case but not in the CMN case. In short, the CMP
instruction doesn't perform a truncate of the (logical) NOT of 0 plus the value
of r0 and the carry bit (because the "carry bit" parameter to AddWithCarry is
defined as 1 in this case, the carry flag will always be set when r0 >= 0). The
CMN instruction doesn't perform a NOT of 0 so there is never a "carry" when this
AddWithCarry is performed (because the "carry bit" parameter to AddWithCarry is
defined as 0).
The AddWithCarry in the CMP case seems to be relying upon the identity:
~x + 1 = -x
However when x is 0 and unsigned, this doesn't hold:
x = 0
~x = 0xFFFF FFFF
~x + 1 = 0x1 0000 0000
(-x = 0) != (0x1 0000 0000 = ~x + 1)
Therefore, we should disable *all* versions of CMN, especially when comparing
against zero, until we can limit when the CMN instruction is used (when we know
that the RHS is not 0) or when we have a hardware fix for this.
(See the ARM docs for the "AddWithCarry" pseudo-code.)
This is related to <rdar://problem/7569620>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112176 91177308-0d34-0410-b5e6-96231b3b80d8