Commit Graph

12308 Commits

Author SHA1 Message Date
Benjamin Kramer
68ed46ce9a Roll back the rest of r126557. It's a hack that will break in some obscure cases.
llvm-svn: 138130
2011-08-19 22:39:31 +00:00
Nick Lewycky
c1348074ec Eli points out that this is what report_fatal_error() is for.
llvm-svn: 138091
2011-08-19 21:45:19 +00:00
Nick Lewycky
3f73184d90 This is not actually unreachable, so don't use llvm_unreachable for it. Since
the intent seems to be to terminate even in Release builds, just use abort()
directly.

If program flow ever reaches a __builtin_unreachable (which llvm_unreachable is
#define'd to on newer GCCs) then the program is undefined.

llvm-svn: 138068
2011-08-19 20:14:27 +00:00
Jakob Stoklund Olesen
6949077f74 Add llc flags to disable machine DCE and CSE.
This is useful for unit tests.

llvm-svn: 138028
2011-08-19 02:05:35 +00:00
Benjamin Kramer
4938edb02c Make a bunch of symbols private.
llvm-svn: 138025
2011-08-19 01:42:18 +00:00
Jakob Stoklund Olesen
9eb77bf615 Don't treat a partial <def,undef> operand as a read.
Normally, a partial register def is treated as reading the
super-register unless it also defines the full register like this:

  %vreg110:sub_32bit<def> = COPY %vreg77:sub_32bit, %vreg110<imp-def>

This patch also uses the <undef> flag on partial defs to recognize
non-reading operands:

  %vreg110:sub_32bit<def,undef> = COPY %vreg77:sub_32bit

This fixes a subtle bug in RegisterCoalescer where LIS->shrinkToUses
would treat a coalesced copy as still reading the register, extending
the live range artificially.

My test case only works when I disable DCE so a dead copy is left for
RegisterCoalescer, so I am not including it.

<rdar://problem/9967101>

llvm-svn: 138018
2011-08-19 00:30:17 +00:00
Renato Golin
c8d4065781 add the comments of each declaration follow it, making it easier to read and compare to GCC's result.
llvm-svn: 138009
2011-08-18 23:43:14 +00:00
Devang Patel
0ecbcbd12c Eliminate unnecessary forwarding function.
llvm-svn: 138006
2011-08-18 23:17:55 +00:00
Devang Patel
a6576a146d Add new DIE into the map asap.
llvm-svn: 137998
2011-08-18 22:21:50 +00:00
Ivan Krasin
d7cbd4c518 FastISel: avoid function calls between the materialization of the constant and its use.
llvm-svn: 137993
2011-08-18 22:06:10 +00:00
Bill Wendling
247fd3bf59 Add the support in code-gen for the landingpad instruction lowering.
The landingpad instruction is lowered into the EXCEPTIONADDR and EHSELECTION
SDNodes. The information from the landingpad instruction is harvested by the
'AddLandingPadInfo' function. The new EH uses the current EH scheme in the
back-end. This will change once we switch over to the new scheme. (Reviewed by
Jakob!)

llvm-svn: 137880
2011-08-17 21:56:44 +00:00
Bill Wendling
a408e5bf31 Revert patch. Forgot a dependent commit.
llvm-svn: 137875
2011-08-17 21:28:05 +00:00
Bill Wendling
2a521948f0 Add the body of 'visitLandingPad'.
This generates the SDNodes for the new exception handling scheme. It takes the
two values coming from the landingpad instruction and assigns them to the
EXCEPTIONADDR and EHSELECTION nodes.

llvm-svn: 137873
2011-08-17 21:25:14 +00:00
Bill Wendling
1cdd7fdf54 Modify for the new EH scheme.
Things are much saner now. We no longer need to modify the laning pads, because
of the invariants we impose upon them. The only thing DwarfEHPrepare needs to do
is convert the 'resume' instruction into a call to '_Unwind_Resume'.

llvm-svn: 137855
2011-08-17 19:48:49 +00:00
Devang Patel
eb1bb4e419 Until now all debug info MDNodes referred to a root MDNode, a compile unit. This simplified handling of these needs in dwarf writer. However, one side effect of this is that during link time optimization all these MDNodes are _not_ uniqued. In other words there will be N number of MDNodes describing "int", "char" and all other types, which would suddenly grow when each object file starts using libraries like STL.
MDNodes graph structure such that compiler unit keeps track of important MDNodes and update dwarf writer to process mdnodes top-down instead of bottom up.

llvm-svn: 137778
2011-08-16 22:09:43 +00:00
Jim Grosbach
345768c9ff Remove unused Target argument from AsmParser construction methods.
The argument is unused, and is a layering violation in any case.

llvm-svn: 137735
2011-08-16 18:33:49 +00:00
Devang Patel
927840458e Remove unnecessary version check.
llvm-svn: 137728
2011-08-16 17:41:41 +00:00
Nadav Rotem
b66b866f46 Revert r137562 because it caused PR10674
llvm-svn: 137719
2011-08-16 14:34:29 +00:00
Devang Patel
07bb9eea33 Refactor.
llvm-svn: 137689
2011-08-15 23:47:24 +00:00
Devang Patel
1f4f98d664 Continue to hoist uses of getCompileUnit() up. The goal is to get rid of uses of getCompileUnit().
llvm-svn: 137683
2011-08-15 23:36:40 +00:00
Devang Patel
d2dfc5ec02 This is somewhat déjà-vu, but avoid using getCompileUnit() as much as possible.
llvm-svn: 137668
2011-08-15 22:24:32 +00:00
Devang Patel
3acc70e536 Refactor. Variables are part of compile unit so let CompileUnit create new variable.
llvm-svn: 137663
2011-08-15 22:04:40 +00:00
Devang Patel
d899444347 There is no need to maintain a set to keep track of variables that use location expressions. In such cases, AT_location attribute's value will be a label.
llvm-svn: 137659
2011-08-15 21:43:21 +00:00
Devang Patel
900d97719b Fix warning.
llvm-svn: 137658
2011-08-15 21:35:16 +00:00
Devang Patel
3e4a965519 Simplify. Let DbgVariable keep track of variable's DBG_VALUE machine instruction.
llvm-svn: 137656
2011-08-15 21:24:36 +00:00
Devang Patel
99819b527d Simplify mapping to variable from its abstract variable info.
When a variable is inlined multiple places, abstract variable keeps name, location, type etc.. info and all other concreate instances of the variable directly refers to abstract variable.

llvm-svn: 137637
2011-08-15 19:01:20 +00:00
Devang Patel
d7d80aadd1 Refactor.
llvm-svn: 137632
2011-08-15 18:40:16 +00:00
Devang Patel
6e4d2c9fb7 Refactor.
llvm-svn: 137631
2011-08-15 18:35:42 +00:00
Devang Patel
dfd6ec3ce1 Refactor. Global variables are part of compile unit so let CompileUnit create new global variable.
llvm-svn: 137621
2011-08-15 17:57:41 +00:00
Devang Patel
895437142a Refactor. A subprogram is part of compile unit so let CompileUnit construct new subprogram.
llvm-svn: 137618
2011-08-15 17:24:54 +00:00
Nadav Rotem
6858b344ed Fix PR 10635. When generating integer constants, the constant element type may
be illegal, even if the requested vector type is legal. Testcase is one of the
disabled ARM tests in the vector-select patch.

llvm-svn: 137562
2011-08-13 20:31:45 +00:00
Bill Wendling
fae1475823 Initial commit of the 'landingpad' instruction.
This implements the 'landingpad' instruction. It's used to indicate that a basic
block is a landing pad. There are several restrictions on its use (see
LangRef.html for more detail). These restrictions allow the exception handling
code to gather the information it needs in a much more sane way.

This patch has the definition, implementation, C interface, parsing, and bitcode
support in it.

llvm-svn: 137501
2011-08-12 20:24:12 +00:00
Devang Patel
444034783e Use ArrayRef.
llvm-svn: 137485
2011-08-12 18:10:19 +00:00
Chris Lattner
335d399a0e switch to use the new api for structtypes.
llvm-svn: 137480
2011-08-12 18:06:37 +00:00
Devang Patel
db4374a28a Provide fast path as Jakob suggested.
llvm-svn: 137478
2011-08-12 18:01:34 +00:00
Nadav Rotem
62da15a330 Revert r137310 because it does not optimize any code on ToT
llvm-svn: 137466
2011-08-12 17:15:04 +00:00
Duncan Sands
a41634e307 Silence a bunch (but not all) "variable written but not read" warnings
when building with assertions disabled.

llvm-svn: 137460
2011-08-12 14:54:45 +00:00
Jakob Stoklund Olesen
1f582ba609 Simplify the interference checking code a bit.
This is possible now that we now longer provide an interface to iterate
the interference overlaps.

llvm-svn: 137397
2011-08-12 00:22:04 +00:00
Jakob Stoklund Olesen
da0192d72b Remove the InterferenceResult class.
llvm-svn: 137381
2011-08-11 22:46:06 +00:00
Jakob Stoklund Olesen
cd14efaec2 Eliminate the last use of InterferenceResult.
The Query class now holds two iterators instead of an InterferenceResult
instance. The iterators are used as bookmarks for repeated
collectInterferingVRegs calls.

llvm-svn: 137380
2011-08-11 22:46:04 +00:00
Jakob Stoklund Olesen
da4f0eb12c Remove more dead code.
collectInterferingVRegs will be the primary function for interference
checks.

llvm-svn: 137354
2011-08-11 21:18:34 +00:00
Jakob Stoklund Olesen
7519336752 Privatize an unused part of the LiveIntervalUnion::Query interface.
No clients are iterating over interference overlaps.

llvm-svn: 137350
2011-08-11 21:00:42 +00:00
Jakob Stoklund Olesen
05ff9d1f6d Remove some dead code.
The InterferenceResult iterator turned out to be less important than we
thought it would be.  LiveIntervalUnion clients want higher level
information, like the list of interfering virtual registers.

llvm-svn: 137346
2011-08-11 20:41:41 +00:00
Benjamin Kramer
fa7e6a54b1 Plug a memory leak.
llvm-svn: 137321
2011-08-11 18:39:28 +00:00
Nadav Rotem
61140e1028 [AVX] When joining two XMM registers into a YMM register, make sure that the
lower XMM register gets in first. This will allow the SUBREG pattern to
elliminate the first vector insertion. 

llvm-svn: 137310
2011-08-11 16:49:36 +00:00
Chris Lattner
96710b4308 fix PR10605 / rdar://9930964 by adding a pretty scary missed check.
It's somewhat surprising anything works without this.  Before we would
compile the testcase into:

test:                                   # @test
	movl	$4, 8(%rdi)
	movl	8(%rdi), %eax
	orl	%esi, %eax
	cmpl	$32, %edx
	movl	%eax, -4(%rsp)          # 4-byte Spill
	je	.LBB0_2

now we produce:

test:                                   # @test
	movl	8(%rdi), %eax
	movl	$4, 8(%rdi)
	orl	%esi, %eax
	cmpl	$32, %edx
	movl	%eax, -4(%rsp)          # 4-byte Spill
	je	.LBB0_2
llvm-svn: 137303
2011-08-11 06:26:54 +00:00
Devang Patel
784077eb57 Stay within 80 columns.
llvm-svn: 137283
2011-08-10 23:58:09 +00:00
Devang Patel
bb23a4a9a5 Distinguish between two copies of one inlined variable. Take 2.
llvm-svn: 137253
2011-08-10 21:50:54 +00:00
Devang Patel
37a62058fe While extending definition range of a debug variable, consult lexical scopes also. There is no point extending debug variable out side its lexical block. This provides 6x compile time speedup in some cases.
llvm-svn: 137250
2011-08-10 21:25:34 +00:00
Devang Patel
e30746c844 Revert unintentional parts of previous check-in.
llvm-svn: 137249
2011-08-10 21:16:49 +00:00