summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen
Commit message (Collapse)AuthorAgeFilesLines
...
* A better fix for my previous patch, MOVZQI2PQIrr just requires SSE2.Chris Lattner2008-04-201-2/+2
| | | | llvm-svn: 49986
* Not all x86-64 machines have sse3 apparently.Chris Lattner2008-04-201-2/+2
| | | | llvm-svn: 49985
* rename *.llx -> *.llChris Lattner2008-04-1929-0/+0
| | | | llvm-svn: 49970
* 64-bit atomic operations.Evan Cheng2008-04-192-0/+48
| | | | llvm-svn: 49949
* Be more careful with insert_subreg and extract_subreg where either source or ↵Evan Cheng2008-04-171-0/+171
| | | | | | destination operand has already been coalesced with another register that's defined by a insert_subreg or extract_subreg. llvm-svn: 49843
* Fix a sub-register indice propagation bug.Evan Cheng2008-04-171-0/+33
| | | | llvm-svn: 49832
* Don't forget about sub-register indices when rematting instructions.Evan Cheng2008-04-161-0/+46
| | | | llvm-svn: 49830
* After reading memory that's already freed.Evan Cheng2008-04-161-0/+14
| | | | llvm-svn: 49810
* Really test what's intended.Evan Cheng2008-04-161-0/+1
| | | | llvm-svn: 49802
* Rewrite LiveVariable liveness computation. The new implementation is much ↵Evan Cheng2008-04-161-0/+48
| | | | | | | | simplified. It eliminated the nasty recursive routines and removed the partial def / use bookkeeping. There is also potential for performance improvement by replacing the conservative handling of partial physical register definitions. The code is currently disabled until live interval analysis is taught of the name scheme. This patch also fixed a couple of nasty corner cases. llvm-svn: 49784
* Add support for the form of the SSE41 extractps instruction thatDan Gohman2008-04-162-0/+31
| | | | | | puts its result in a 32-bit GPR. llvm-svn: 49762
* Recreate the size SDNode instead of reusing the old one in the x86Dan Gohman2008-04-161-0/+13
| | | | | | | | memcpy lowering code; this ensures that the size node has the desired result type. This fixes a regression from r49572 with @llvm.memcpy.i64 on x86-32. llvm-svn: 49761
* Add movd instructions to move from MMX registersDan Gohman2008-04-151-0/+26
| | | | | | to 64-bit GPR registers on x86-64. llvm-svn: 49757
* Treat EntryToken nodes as "passive" so that they aren't added to theDan Gohman2008-04-151-2/+2
| | | | | | | | | | | | | | | | | ScheduleDAG; they don't correspond to any actual instructions so they don't need to be scheduled. This fixes a bug where the EntryToken was being scheduled multiple times in some cases, though it ended up not causing any trouble because EntryToken doesn't expand into anything. With this fixed the schedulers reliably schedule the expected number of units, so we can check this with an assertion. This requires a tweak to test/CodeGen/X86/loop-hoist.ll because it ends up getting scheduled differently in a trivial way, though it was enough to fool the prcontext+grep that the test does. llvm-svn: 49701
* Upgrade these tests for the current intrinsic prototypes.Dan Gohman2008-04-141-6/+6
| | | | llvm-svn: 49669
* Remove -unwind-tables-optional everywhere, sinceDale Johannesen2008-04-1416-21/+21
| | | | | | this is now the default. llvm-svn: 49667
* This patch corrects the handling of byval arguments for tailcallArnold Schwaighofer2008-04-122-1/+34
| | | | | | | | | | | | | | | | | | | | | | optimized x86-64 (and x86) calls so that they work (... at least for my test cases). Should fix the following problems: Problem 1: When i introduced the optimized handling of arguments for tail called functions (using a sequence of copyto/copyfrom virtual registers instead of always lowering to top of the stack) i did not handle byval arguments correctly e.g they did not work at all :). Problem 2: On x86-64 after the arguments of the tail called function are moved to their registers (which include ESI/RSI etc), tail call optimization performs byval lowering which causes xSI,xDI, xCX registers to be overwritten. This is handled in this patch by moving the arguments to virtual registers first and after the byval lowering the arguments are moved from those virtual registers back to RSI/RDI/RCX. llvm-svn: 49584
* Drop ISD::MEMSET, ISD::MEMMOVE, and ISD::MEMCPY, which are not LegalDan Gohman2008-04-128-7/+67
| | | | | | | | | | | | | | | | | | | | | | | | | | | on any current target and aren't optimized in DAGCombiner. Instead of using intermediate nodes, expand the operations, choosing between simple loads/stores, target-specific code, and library calls, immediately. Previously, the code to emit optimized code for these operations was only used at initial SelectionDAG construction time; now it is used at all times. This fixes some cases where rep;movs was being used for small copies where simple loads/stores would be better. This also cleans up code that checks for alignments less than 4; let the targets make that decision instead of doing it in target-independent code. This allows x86 to use rep;movs in low-alignment cases. Also, this fixes a bug that resulted in the use of rep;stos for memsets of 0 with non-constant memory size when the alignment was at least 4. It's better to use the library in this case, which can be significantly faster when the size is large. This also preserves more SourceValue information when memory intrinsics are lowered into simple loads/stores. llvm-svn: 49572
* Fix a bug that prevented x86-64 from using rep.movsq forDan Gohman2008-04-121-1/+1
| | | | | | 8-byte-aligned data. llvm-svn: 49571
* If a PHI node has a single implicit_def source, replace it with an ↵Evan Cheng2008-04-111-0/+3544
| | | | | | implicit_def instead of a copy. llvm-svn: 49543
* New test.Evan Cheng2008-04-102-1/+259
| | | | llvm-svn: 49514
* A copy instruction may use a register multiple times on some targets. Change ↵Evan Cheng2008-04-101-0/+100
| | | | | | them all. llvm-svn: 49491
* Fix the x86-64 side of PR2108 by adding a v2f64 version ofChris Lattner2008-04-101-0/+13
| | | | | | | | MOVZQI2PQIrr. This would be better handled as a dag combine (with the goal of eliminating the bitconvert) but I don't know how to do that safely. Thoughts welcome. llvm-svn: 49463
* Teach branch folding pass about implicit_def instructions. Unfortunately we ↵Evan Cheng2008-04-102-3/+49
| | | | | | can't just eliminate them since register scavenger expects every register use to be defined. However, we can delete them when there are no intra-block uses. Carefully removing some implicit def's which enable more blocks to be optimized away. llvm-svn: 49461
* - More aggressively coalescing away copies whose source is defined by an ↵Evan Cheng2008-04-093-0/+124
| | | | | | | | implicit_def. - Added insert_subreg coalescing support. llvm-svn: 49448
* Missed a hasInterval check.Evan Cheng2008-04-091-0/+19
| | | | llvm-svn: 49415
* Rename -disable-required-unwind-tables to -unwind-tables-optional.Dale Johannesen2008-04-0816-21/+21
| | | | llvm-svn: 49391
* Missed one.Dale Johannesen2008-04-081-1/+1
| | | | llvm-svn: 49365
* Add -disable-required-unwind-tables to testsDale Johannesen2008-04-0816-21/+21
| | | | | | | that need it (usually, grepping for some string found in unwind info) llvm-svn: 49364
* Fix test.Evan Cheng2008-04-071-2/+2
| | | | llvm-svn: 49343
* fix this testcase to pass and remove a duplicate instance of itself.Chris Lattner2008-04-061-75/+3
| | | | llvm-svn: 49281
* Prefer to expand mask for xor to -1, so we have a chance to turn it into a not.Torok Edwin2008-04-061-0/+146
| | | | | | | If it cannot be expanded, it will keep the old behaviour and try to shrink the constant. Part of enhancement for PR2191. llvm-svn: 49280
* 1. IMPLICIT_DEF can *re-define* any register.Evan Cheng2008-04-051-0/+60
| | | | | | | 2. Coalescer can now create an interesting situation where a register def can reaches itself without being killed. llvm-svn: 49246
* Favors pshufd over shufps when shuffling elements from one vector. pshufd is ↵Evan Cheng2008-04-0510-19/+43
| | | | | | faster than shufps. llvm-svn: 49244
* New test case.Evan Cheng2008-04-031-0/+29
| | | | llvm-svn: 49190
* Testcase for EH with functions whose names are stripped.Dale Johannesen2008-04-021-0/+27
| | | | llvm-svn: 49111
* Speculatively micro-optimize memory-zeroing calls on Darwin 10.Dan Gohman2008-04-011-0/+8
| | | | llvm-svn: 49048
* More soft fp fixes.Evan Cheng2008-04-011-1/+3
| | | | llvm-svn: 49016
* Unbreak ARM / Thumb soft FP support.Evan Cheng2008-04-011-0/+10
| | | | llvm-svn: 49012
* Mark functions in some tests as 'nounwind'. GeneratingDale Johannesen2008-03-3114-31/+31
| | | | | | | | EH info for these functions causes the tests to fail for random reasons (e.g. looking for 'or' or counting lines with asm-printer; labels count as lines.) llvm-svn: 49003
* It's not safe to fold a load from GV stub or constantpool into a two-address ↵Evan Cheng2008-03-311-0/+40
| | | | | | use. llvm-svn: 49002
* Fix a DAGCombiner optimization to respect volatile qualification.Dan Gohman2008-03-311-0/+24
| | | | llvm-svn: 48994
* Fix a tokenfactor node to use the load chain rather than theDan Gohman2008-03-281-0/+35
| | | | | | load value. This fixes PR2177. llvm-svn: 48932
* Fix a memory bug: increment an iterator of a deleted machine instr.Evan Cheng2008-03-271-0/+24
| | | | llvm-svn: 48853
* One more coalescer fix wrt deadness propagation.Evan Cheng2008-03-261-0/+28
| | | | llvm-svn: 48837
* Avoid commuting a def MI in order to coalesce a copy instruction away if any ↵Evan Cheng2008-03-261-0/+49
| | | | | | use of the same val# is a copy instruction that has already been coalesced. llvm-svn: 48833
* Use ## for comment delimiter on darwin x86-32, soDale Johannesen2008-03-251-0/+49
| | | | | | | | llvm's output .s files will go through gcc -std=c99 without triggering preprocesser errors. Approach suggested by Daveed Vandevoorde. llvm-svn: 48808
* Handle a special case xor undef, undef -> 0. Technically this should be ↵Evan Cheng2008-03-251-0/+11
| | | | | | transformed to undef. But this is such a common idiom (misuse) we are going to handle it. llvm-svn: 48792
* Add CMP32mr and friends to the load-unfolding table. AmongDan Gohman2008-03-251-1/+1
| | | | | | | | other things, this allows the scheduler to unfold a load operand in the 2008-01-08-SchedulerCrash.ll testcase, so it now successfully clones the comparison to avoid a pushf+popf. llvm-svn: 48777
* Byebye llvm-upgrade!Tanya Lattner2008-03-2582-1240/+1166
| | | | llvm-svn: 48762
OpenPOWER on IntegriCloud