Commit Graph

11797 Commits

Author SHA1 Message Date
Eli Friedman
5e926ac651 Re-revert r130877; it's apparently causing a regression on 197.parser,
possibly related to cbnz formation.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130977 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-06 05:23:07 +00:00
Rafael Espindola
97a1216581 Remove DwarfTableException.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130964 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 23:19:54 +00:00
Rafael Espindola
254784f9e0 Remove the DwarfTable enum.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130959 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 22:14:31 +00:00
Devang Patel
51a04adc5e Remove little used statistical counter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130955 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 22:00:08 +00:00
Rafael Espindola
ecf58b9e3c Implement a really simple DwarfSjLjException.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130947 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 20:48:31 +00:00
Rafael Espindola
2d57a64701 List all exception types in a switch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130944 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 19:48:34 +00:00
Andrew Trick
ee10915892 ARM post RA scheduler compile time fix.
BuildSchedGraph was quadratic in the number of calls in the basic
block. After this fix, it keeps only a single call at the top of the
DefList so compile time doesn't blow up on large blocks. This reduces
postRA sched time on an external test case from 81s to 0.3s.  Although
r130800 (reduced ARM register alias defs) also partially fixes the
issue by reducing the constant overhead of checking call interference
by an order of magnitude.

Fixes <rdar://problem/7662664> very poor compile time with post RA scheduling.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130943 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 19:32:21 +00:00
Andrew Trick
f405b1ad19 whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130942 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 19:24:06 +00:00
Owen Anderson
d71867a8f4 Allow FastISel of three-register-operand instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130934 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 17:59:04 +00:00
Devang Patel
d96efb83c1 If debug info for inlined function is missing then handle it gracefully.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130933 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 17:54:26 +00:00
Jakob Stoklund Olesen
e9bd4ea5fd Add some statistics to the splitting and spilling frameworks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130931 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 17:22:53 +00:00
Eli Friedman
23d3243e62 Avoid extra vreg copies for arguments passed in registers. Specifically, this can make MachineCSE more effective in some cases (especially in small functions). PR8361 / part of rdar://problem/8259436 .
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130928 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 16:53:34 +00:00
Eli Friedman
6005b875b0 Small syntax cleanup; we don't need to #define constants in C++. No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130926 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 16:25:23 +00:00
Eli Friedman
5f6bf5d441 Minor correction to r130877; fixes PR9846 and hopefully the buildbot failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130925 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 16:18:11 +00:00
Bill Wendling
087aad44cb Remove a flag that would set the ".eh" symbol as .globl. MachO was the only one
who used this flag, and it now emits CFI and doesn't emit this anymore. All
other targets left this flag "false".
<rdar://problem/8486371>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130918 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05 06:49:15 +00:00
Jakob Stoklund Olesen
7a2718a3c1 Disable physical register coalescing by default.
Joining physregs is inherently dangerous because it uses a heuristic to avoid
creating invalid code. Linear scan had an emergency spilling mechanism to deal
with those rare cases. The new greedy allocator does not.

The greedy register allocator is much better at taking hints, so this has almost
no impact on code size and quality. The few cases where it matters show up as
unit tests that now have -join-physregs enabled explicitly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130896 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 23:59:00 +00:00
Bill Wendling
d29052bf26 SjLj EH could produce a machine basic block that legitimately has more than one
landing pad as its successor.

SjLj exception handling jumps to the correct landing pad via a switch statement
that's generated right before code-gen. Loosen the constraint in the machine
instruction verifier to allow for this. Note, this isn't the most rigorous check
since we cannot determine where that switch statement came from. But it's
marginally better than turning this check off when SjLj exceptions are used.
<rdar://problem/9187612>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130881 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 22:54:05 +00:00
Eli Friedman
baf717a08a Re-commit r130862 with a minor change to avoid an iterator running off the edge in some cases.
Original message:

Teach MachineCSE how to do simple cross-block CSE involving physregs.  This allows, for example, eliminating duplicate cmpl's on x86. Part of rdar://problem/8259436 .



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130877 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 22:10:36 +00:00
Eli Friedman
24d4c9911e Back out r130862; it appears to be breaking bootstrap.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130867 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 20:48:42 +00:00
Eli Friedman
49cec1d818 Teach MachineCSE how to do simple cross-block CSE involving physregs. This allows, for example, eliminating duplicate cmpl's on x86. Part of rdar://problem/8259436 .
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130862 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 19:54:24 +00:00
Rafael Espindola
597a7664e1 Producing a DW_FORM_addr for DW_AT_stmt_list is probably correct, but
it is both inefficient and unexpected by dwarfdump. Change to
a DW_FORM_data4.

While in here, change the predicate name to reflect that the position
is not really absolute (it is an offset), just that the linker needs a
relocation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130846 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 17:44:06 +00:00
Jakob Stoklund Olesen
9126601770 Rename -disable-physical-join to -join-physregs and invert it.
Physreg joining is still on by default, but I will turn it off shortly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130844 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 16:45:05 +00:00
Devang Patel
a938dd0dc9 Tighten up check for empty (i.e. no meaningful debug info) module. This fixes dwarf-die2.c test case from gcc test suite.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130842 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-04 16:34:02 +00:00
Devang Patel
ee4a3abede Even if the subprogram is going to use AT_specification, emit DW_AT_MIPS_linkage_name. This helps gdb and fixes var-path-expr.exp regression reported by gdb testsuite.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130794 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-03 21:50:34 +00:00
Jakob Stoklund Olesen
7d6b6a05b5 Gracefully handle invalid live ranges. Fix PR9831.
Register coalescing can sometimes create live ranges that end in the middle of a
basic block without any killing instruction. When SplitKit detects this, it will
repair the live range by shrinking it to its uses.

Live range splitting also needs to know about this. When the range shrinks so
much that it becomes allocatable, live range splitting fails because it can't
find a good split point. It is paranoid about making progress, so an allocatable
range is considered an error.

The coalescer should really not be creating these bad live ranges. They appear
when coalescing dead copies.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130787 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-03 20:42:13 +00:00
Devang Patel
30692abac7 If the front end has emitted llvm.dbg.cu and other debug info anchors (clang does it now) then use them directly. This saves one scan of entire module, to collect debug info, which in turns saves few machine cycles at compile time.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130759 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-03 16:45:22 +00:00
Owen Anderson
f7710af4ba Other parts of the SelectionDAG framework assume that targets use their pointer type for vector indices. Make the vector unrolling code respect that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130733 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 22:25:45 +00:00
Jakob Stoklund Olesen
d61c40360c Handle <def,undef> in the second loop as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130718 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 20:36:53 +00:00
Jakob Stoklund Olesen
e2f669fb3f Use the PrintReg adaptor to correctly print live-in registers in debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130715 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 20:06:30 +00:00
Jakob Stoklund Olesen
2048e4ab7f Only ignore <undef> use operands, keep the <def,undef> ops.
Def operands may also have an <undef> flag, but that just means that a
sub-register redef doesn't actually read the super-register. For physical
registers, it has no meaning.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130714 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 20:06:28 +00:00
Devang Patel
c455bb7dc7 Emit debug info for global variables first.
This works around a limitation in gdb which is reported by following inherit.exp test failures from gdb testsuite.

gdb.cp/inherit.exp: print g_vB.vB::vb
gdb.cp/inherit.exp: print g_vB.vB::vx
gdb.cp/inherit.exp: print g_vC.vC::vc
gdb.cp/inherit.exp: print g_vC.vC::vx
gdb.cp/inherit.exp: print g_vD.vB::vb
...



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 18:19:17 +00:00
Rafael Espindola
e17e6c01c8 Only produce the eh_frame section if we have at least one personality function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130692 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 15:49:52 +00:00
Jakob Stoklund Olesen
bb30dd40ed Minimize the slot indexes spanned by register ranges created when splitting.
When an interfering live range ends at a dead slot index between two
instructions, make sure that the inserted copy instruction gets a slot index
after the dead ones. This makes it possible to avoid the interference.

Ideally, there shouldn't be interference ending at a deleted instruction, but
physical register coalescing can sometimes do that to sub-registers.

This fixes PR9823.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130687 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02 05:29:58 +00:00
Rafael Espindola
5426a9ee37 GCC uses a different encoding of pointers in the FDE when using
-fno-dwarf2-cfi-asm. Implement the same behavior.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130637 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-01 04:49:54 +00:00
Jakob Stoklund Olesen
ec7e4fff96 When a physreg is live-in and live through a basic block, make sure its live
range covers the entire block.

The live range can't be terminated at a random instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130619 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 19:12:33 +00:00
Jakob Stoklund Olesen
6ee56e658a Avoid using stale entries form the sibling value map.
This could happen when trying to use a value that had been eliminated after dead
code elimination and folding loads.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130597 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 06:42:21 +00:00
Jakob Stoklund Olesen
66446c803a Use hysteresis for local live range splitting as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130596 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 05:07:46 +00:00
Rafael Espindola
f1a5c7ec04 Add all the plumbing needed for MC to expand cfi to the old tables in
the final assembly. It is the same technique used when targeting
assemblers that don't support .loc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130587 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 03:44:37 +00:00
Jakob Stoklund Olesen
141f529baf Update comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130582 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 03:13:08 +00:00
Jakob Stoklund Olesen
5aa3211e2d Use a greedy algorithm for allocating registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130568 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 01:37:54 +00:00
Bill Wendling
d65ba721cc Print out the 'nontemporal' info on a store.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130562 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29 23:45:22 +00:00
Eli Friedman
b2b03fc2fa Make FastEmit_ri_ try a bit harder to succeed for supported operations; FastEmit_i can fail for non-Thumb2 ARM. Makes ARMSimplifyAddress work correctly, and reduces the number of fast-isel bailouts on non-Thumb ARM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130560 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29 23:34:52 +00:00
Devang Patel
49a3ff9d17 Hoist MCLineEntry construction AsmPrinter so that anyone who derives from AsmPrinter can have line number entries.
PR 9810



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130518 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29 18:00:54 +00:00
Rafael Espindola
ac3ba1b27e The last hack for producing bit identical output with cfi on OS X.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130504 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29 15:09:53 +00:00
Rafael Espindola
7b11a4cbdc Change DwarfCFIException's member variables to track what it actually
emmits: .cfi_personality, .cfi_lsda and the moves.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130503 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29 14:48:51 +00:00
Rafael Espindola
a432997745 Factor some code to needsCFIMoves. Avoid printing moves when we don't have to.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130501 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29 14:14:06 +00:00
Devang Patel
c26f5447e3 Teach dwarf writer to handle complex address expression for .debug_loc entries.
This fixes clang generated blocks' variables' debug info.
Radar 9279956.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130373 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-28 02:22:40 +00:00
Eli Friedman
bd1253809b Fix a silly mistake in r130338.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130360 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-28 00:42:03 +00:00
Rafael Espindola
60246a9622 Remove unnecessary argument.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130343 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-27 23:17:57 +00:00
Rafael Espindola
7afec9cc0f Rename getPersonalityPICSymbol to getCFIPersonalitySymbol, document it, and
give it a bit more responsibility. Also implement it for MachO.

If hacked to use cfi, 32 bit MachO will produce

.cfi_personality 155, L___gxx_personality_v0$non_lazy_ptr

and 64 bit will produce

.cfi_presonality ___gxx_personality_v0

The general idea is that .cfi_personality gets passed the final symbol. It is
up to codegen to produce it if using indirect representation (like 32 bit
MachO), but it is up to MC to decide which relocations to create.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130341 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-27 23:08:15 +00:00