Commit Graph

93 Commits

Author SHA1 Message Date
Dan Gohman
65fd6564b8 Reapply r143206, with fixes. Disallow physical register lifetimes
across calls, and only check for nested dependences on the special
call-sequence-resource register.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@143660 91177308-0d34-0410-b5e6-96231b3b80d8
2011-11-03 21:49:52 +00:00
Dan Gohman
6f3ddef7c5 Revert r143206, as there are still some failing tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@143262 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-29 00:41:52 +00:00
Dan Gohman
bf923b815d Reapply r143177 and r143179 (reverting r143188), with scheduler
fixes: Use a separate register, instead of SP, as the
calling-convention resource, to avoid spurious conflicts with
actual uses of SP. Also, fix unscheduling of calling sequences,
which can be triggered by pseudo-two-address dependencies.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@143206 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-28 17:55:38 +00:00
Duncan Sands
62c1d00dfd Speculatively disable Dan's commits 143177 and 143179 to see if
it fixes the dragonegg self-host (it looks like gcc is miscompiled).
Original commit messages:
Eliminate LegalizeOps' LegalizedNodes map and have it just call RAUW
on every node as it legalizes them. This makes it easier to use
hasOneUse() heuristics, since unneeded nodes can be removed from the
DAG earlier.

Make LegalizeOps visit the DAG in an operands-last order. It previously
used operands-first, because LegalizeTypes has to go operands-first, and
LegalizeTypes used to be part of LegalizeOps, but they're now split.
The operands-last order is more natural for several legalization tasks.
For example, it allows lowering code for nodes with floating-point or
vector constants to see those constants directly instead of seeing the
lowered form (often constant-pool loads). This makes some things
somewhat more complicated today, though it ought to allow things to be
simpler in the future. It also fixes some bugs exposed by Legalizing
using RAUW aggressively.

Remove the part of LegalizeOps that attempted to patch up invalid chain
operands on libcalls generated by LegalizeTypes, since it doesn't work
with the new LegalizeOps traversal order. Instead, define what
LegalizeTypes is doing to be correct, and transfer the responsibility
of keeping calls from having overlapping calling sequences into the
scheduler.

Teach the scheduler to model callseq_begin/end pairs as having a
physical register definition/use to prevent calls from having
overlapping calling sequences. This is also somewhat complicated, though
there are ways it might be simplified in the future.

This addresses rdar://9816668, rdar://10043614, rdar://8434668, and others.
Please direct high-level questions about this patch to management.

Delete #if 0 code accidentally left in.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@143188 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-28 09:55:57 +00:00
Dan Gohman
2ba60e5930 Eliminate LegalizeOps' LegalizedNodes map and have it just call RAUW
on every node as it legalizes them. This makes it easier to use
hasOneUse() heuristics, since unneeded nodes can be removed from the
DAG earlier.

Make LegalizeOps visit the DAG in an operands-last order. It previously
used operands-first, because LegalizeTypes has to go operands-first, and
LegalizeTypes used to be part of LegalizeOps, but they're now split.
The operands-last order is more natural for several legalization tasks.
For example, it allows lowering code for nodes with floating-point or
vector constants to see those constants directly instead of seeing the
lowered form (often constant-pool loads). This makes some things
somewhat more complicated today, though it ought to allow things to be
simpler in the future. It also fixes some bugs exposed by Legalizing
using RAUW aggressively.

Remove the part of LegalizeOps that attempted to patch up invalid chain
operands on libcalls generated by LegalizeTypes, since it doesn't work
with the new LegalizeOps traversal order. Instead, define what
LegalizeTypes is doing to be correct, and transfer the responsibility
of keeping calls from having overlapping calling sequences into the
scheduler.

Teach the scheduler to model callseq_begin/end pairs as having a
physical register definition/use to prevent calls from having
overlapping calling sequences. This is also somewhat complicated, though
there are ways it might be simplified in the future.

This addresses rdar://9816668, rdar://10043614, rdar://8434668, and others.
Please direct high-level questions about this patch to management.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@143177 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-28 01:29:32 +00:00
Bill Wendling
e575499d83 Revert r141529. This is causing failures in the test-suite, like bigstack and ReedSolomon. Boo...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@141716 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-11 21:40:47 +00:00
Bill Wendling
ef2c86f876 Reapply r141365 now that PR11107 is fixed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@141591 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-10 22:59:55 +00:00
Bill Wendling
eba564ceac Revert r141365. It was causing MultiSource/Benchmarks/MiBench/consumer-lame to
hang, and possibly SPEC/CINT2006/464_h264ref.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@141560 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-10 18:27:30 +00:00
Bill Wendling
8129d21396 When getting the number of bits necessary for addressing mode
ARMII::AddrModeT1_s, we need to take into account that if the frame register is
ARM::SP, then the number of bits is 8. If it's not ARM::SP, then the number of
bits is 5.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@141529 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-10 07:24:23 +00:00
Anton Korobeynikov
244455e6d6 Peephole optimization for ABS on ARM.
Patch by Ana Pazos!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@141365 91177308-0d34-0410-b5e6-96231b3b80d8
2011-10-07 16:15:08 +00:00
Eli Friedman
47d3ee559a Convert more tests to new atomic instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140567 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 21:36:10 +00:00
Nick Lewycky
756a0a1f39 Disable these tests harder. They're XFAIL'd, but that means they still run, and
these tests all infinitely recurse, bringing my system down into swapping hell.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139192 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-06 22:08:18 +00:00
Dan Gohman
6b13cd4154 Revert r131152, r129796, r129761. This code is currently considered
to be unreliable on platforms which require memcpy calls, and it is
complicating broader legalize cleanups. It is hoped that these cleanups
will make memcpy byval easier to implement in the future.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138977 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-01 23:07:08 +00:00
Evan Cheng
39dfb0ff84 Change some ARM subtarget features to be single bit yes/no in order to sink them down to MC layer. Also fix tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@134590 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-07 03:55:05 +00:00
Eric Christopher
73744df0c4 Add support for the 'h' constraint.
Part of rdar://9119939


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@134203 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-30 23:23:01 +00:00
Chris Lattner
b85e4eba85 rip out a ton of intrinsic modernization logic from AutoUpgrade.cpp, which is
for pre-2.9 bitcode files.  We keep x86 unaligned loads, movnt, crc32, and the
target indep prefetch change.

As usual, updating the testsuite is a PITA.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133337 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-18 06:05:24 +00:00
Chris Lattner
26b0000166 manually upgrade a bunch of tests to modern syntax, and remove some that
are either unreduced or only test old syntax.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133228 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-17 03:14:27 +00:00
Jakob Stoklund Olesen
c02a6fa7d8 Don't use register classes larger than TLI->getRegClassFor(VT).
In Thumb mode we cannot handle GPR virtual registers, even though some
instructions can. When isel is lowering a CopyFromReg, it should limit
itself to subclasses of getRegClassFor(VT).

<rdar://problem/9624323>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133210 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-16 22:50:38 +00:00
Eli Friedman
e8e1e4444b FileCheck-ize test, and make it work on EABI hosts, like clang-native-arm-cortex-a9.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133139 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-16 02:36:32 +00:00
Stuart Hastings
007cfce220 Move this test to CodeGen/Thumb. rdar://problem/9416774
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131196 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-11 19:41:28 +00:00
Jakob Stoklund Olesen
96169b189c Fix more register and coalescing dependencies.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130859 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 19:02:11 +00:00
Evan Cheng
554daa67bd Be careful about scheduling nodes above previous calls. It increase usages of
more callee-saved registers and introduce copies. Only allows it if scheduling
a node above calls would end up lessen register pressure.

Call operands also has added ABI restrictions for register allocation, so be
extra careful with hoisting them above calls.

rdar://9329627


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130245 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-26 21:31:35 +00:00
Jakob Stoklund Olesen
775c3e5824 Make tests register allocation independent again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129739 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-19 00:14:43 +00:00
Evan Cheng
06b2a60ef9 Follow up on r127913. Fix Thumb revsh isel. rdar://9286766
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129548 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-14 23:27:44 +00:00
Jakob Stoklund Olesen
c3178f85b5 Fix Thumb and Thumb2 tests to be register allocator independent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128690 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 23:31:50 +00:00
Jakob Stoklund Olesen
1db952d0c6 Provide a legal pointer register class when targeting thumb1.
The LocalStackSlotAllocation pass was creating illegal registers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128687 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-31 23:02:15 +00:00
Cameron Zwarich
899eaa3569 Roll r127459 back in:
Optimize trivial branches in CodeGenPrepare, which often get created from the
lowering of objectsize intrinsics. Unfortunately, a number of tests were relying
on llc not optimizing trivial branches, so I had to add an option to allow them
to continue to test what they originally tested.

This fixes <rdar://problem/8785296> and <rdar://problem/9112893>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@127498 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-11 21:52:04 +00:00
Daniel Dunbar
950d3db5f4 Revert r127459, "Optimize trivial branches in CodeGenPrepare, which often get
created from the", it broke some GCC test suite tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@127477 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-11 19:30:30 +00:00
Cameron Zwarich
592ca3fda9 Optimize trivial branches in CodeGenPrepare, which often get created from the
lowering of objectsize intrinsics. Unfortunately, a number of tests were relying
on llc not optimizing trivial branches, so I had to add an option to allow them
to continue to test what they originally tested.

This fixes <rdar://problem/8785296> and <rdar://problem/9112893>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@127459 91177308-0d34-0410-b5e6-96231b3b80d8
2011-03-11 04:54:27 +00:00
Evan Cheng
9fe2009956 Sorry, several patches in one.
TargetInstrInfo:
Change produceSameValue() to take MachineRegisterInfo as an optional argument.
When in SSA form, targets can use it to make more aggressive equality analysis.

Machine LICM:
1. Eliminate isLoadFromConstantMemory, use MI.isInvariantLoad instead.
2. Fix a bug which prevent CSE of instructions which are not re-materializable.
3. Use improved form of produceSameValue.

ARM:
1. Teach ARM produceSameValue to look pass some PIC labels.
2. Look for operands from different loads of different constant pool entries
   which have same values.
3. Re-implement PIC GA materialization using movw + movt. Combine the pair with
   a "add pc" or "ldr [pc]" to form pseudo instructions. This makes it possible
   to re-materialize the instruction, allow machine LICM to hoist the set of
   instructions out of the loop and make it possible to CSE them. It's a bit
   hacky, but it significantly improve code quality.
4. Some minor bug fixes as well.

With the fixes, using movw + movt to materialize GAs significantly outperform the
load from constantpool method. 186.crafty and 255.vortex improved > 20%, 254.gap
and 176.gcc ~10%.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123905 91177308-0d34-0410-b5e6-96231b3b80d8
2011-01-20 08:34:58 +00:00
Evan Cheng
df55fea807 Spill R4 if it's going to be used to restore SP from FP.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123567 91177308-0d34-0410-b5e6-96231b3b80d8
2011-01-16 05:14:33 +00:00
Jim Grosbach
3e333637f1 Thumb1 had two patterns for the same load-from-constant-pool instruction.
Canonicalize on tLDRpci and remove tLDRcp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@121920 91177308-0d34-0410-b5e6-96231b3b80d8
2010-12-15 23:52:36 +00:00
Devang Patel
afeaae7a94 If dbg_declare() or dbg_value() is not lowered by isel then emit DEBUG message instead of creating DBG_VALUE for undefined value in reg0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@121059 91177308-0d34-0410-b5e6-96231b3b80d8
2010-12-06 22:39:26 +00:00
Evan Cheng
ab5c703fdb Fix epilogue codegen to avoid leaving the stack pointer in an invalid
state. Previously Thumb2 would restore sp from fp like this:
mov sp, r7
sub, sp, #4
If an interrupt is taken after the 'mov' but before the 'sub', callee-saved
registers might be clobbered by the interrupt handler. Instead, try
restoring directly from sp:
add sp, #4
Or, if necessary (with VLA, etc.) use a scratch register to compute sp and
then restore it:
sub.w r4, r7, #8
mov sp, r7
rdar://8465407


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119977 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-22 18:12:04 +00:00
Dale Johannesen
8abe08d7f9 These tests are looking for library function names that
appear to differ on Linux.  Try to make them pass on Linux.
Would be good for a Linux person to review this.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119572 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 21:57:32 +00:00
Bob Wilson
54f9256380 Do not use MEMBARRIER_MCR for any Thumb code.
It is only supported for ARM code.  Normally Thumb2 code would use DMB instead,
but depending on how the compiler is invoked (e.g., -mattr=-db) that might be
disabled.  This prevents a "cannot select MEMBARRIER_MCR" error in that
situation.  Radar 8644195

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118642 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-09 22:50:44 +00:00
Bob Wilson
f74a429816 Overhaul memory barriers in the ARM backend. Radar 8601999.
There were a number of issues to fix up here:
* The "device" argument of the llvm.memory.barrier intrinsic should be
used to distinguish the "Full System" domain from the "Inner Shareable"
domain.  It has nothing to do with using DMB vs. DSB instructions.
* The compiler should never need to emit DSB instructions.  Remove the
ARMISD::SYNCBARRIER node and also remove the instruction patterns for DSB.
* Merge the separate DMB/DSB instructions for options only used for the
disassembler with the default DMB/DSB instructions.  Add the default
"full system" option ARM_MB::SY to the ARM_MB::MemBOpt enum.
* Add a separate ARMISD::MEMBARRIER_MCR node for subtargets that implement
a data memory barrier using the MCR instruction.
* Fix up encodings for these instructions (except MCR).
I also updated the tests and added a few new ones to check for DMB options
that were not currently being exercised.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117756 91177308-0d34-0410-b5e6-96231b3b80d8
2010-10-30 00:54:37 +00:00
Jakob Stoklund Olesen
7eb589d3f9 Try again to disable critical edge splitting in CodeGenPrepare.
The bug that broke i386 linux has been fixed in r115191.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115204 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-30 20:51:52 +00:00
Jakob Stoklund Olesen
08342f2d11 Revert "Disable codegen prepare critical edge splitting. Machine instruction passes now"
This reverts revision 114633. It was breaking llvm-gcc-i386-linux-selfhost.

It seems there is a downstream bug that is exposed by
-cgp-critical-edge-splitting=0. When that bug is fixed, this patch can go back
in.

Note that the changes to tailcallfp2.ll are not reverted. They were good are
required.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114859 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-27 18:43:48 +00:00
Evan Cheng
dd3a41a6b3 Disable codegen prepare critical edge splitting. Machine instruction passes now
break critical edges on demand.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114633 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-23 06:55:34 +00:00
Jim Grosbach
d0bd76b0fb Re-enable usage of the ARM base pointer. r113394 fixed the known failures.
Re-running some nightly testers w/ it enabled to verify.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113399 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-08 20:12:02 +00:00
Jim Grosbach
e1e6d18786 disable for the moment while tracking down a few Thumb2-O0 failure that look
related. (attempt deux, complete w/ test update this time)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113333 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-08 02:00:34 +00:00
Jim Grosbach
65482b1bb8 Re-apply r112883:
"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
2010-09-03 18:37:12 +00:00
Daniel Dunbar
6a8700301c Revert "For ARM stack frames that utilize variable sized objects and have either", it is breaking oggenc with Clang for ARMv6.
This reverts commit 8d6e29cfda270be483abf638850311670829ee65.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112962 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-03 15:26:42 +00:00
Jim Grosbach
1755b3964f 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.

rdar://7352504
rdar://8374540
rdar://8355680



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112883 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-02 22:29:01 +00:00
Jim Grosbach
85e75afc8e Update test for 112609
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112610 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31 17:58:47 +00:00
Duncan Sands
cdd4f8c7cb Correct bogus module triple specifications.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112469 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-30 10:48:29 +00:00
Jim Grosbach
ae47c6d69e Enable pre-RA virtual frame base register allocation. rdar://8277890
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112127 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-26 00:58:06 +00:00
Evan Cheng
e1f0859789 Fix test and re-enable it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110829 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-11 17:25:51 +00:00
Dan Gohman
f63fed141e Temporarily disable some failing tests, until they can be
properly investigated.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110808 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-11 15:09:00 +00:00