b4f7b83dce
* Merge get_GVInitializer and getCharArrayLength into a single function
...
named getConstantStringLength. This is the common part of StrCpy and
StrLen optimizations and probably several others, yet to be written. It
performs all the validity checks for looking at constant arrays that are
supposed to be null-terminated strings and then computes the actual
length of the string.
* Implement the MemCpyOptimization class. This just turns memcpy of 1, 2, 4
and 8 byte data blocks that are properly aligned on those boundaries into
a load and a store. Much more could be done here but alignment
restrictions and lack of knowledge of the target instruction set prevent
use from doing significantly more. That will have to be delegated to the
code generators as they lower llvm.memcpy calls.
llvm-svn: 21562
2005-04-26 07:45:18 +00:00
89011b6f3c
Add a new test case for the llvm.memcpy simplification case.
...
llvm-svn: 21561
2005-04-26 07:40:40 +00:00
81ebb57771
* Add code to reduce multiplies by constant integers to shifts, adds and
...
subtracts. This is a very rough and nasty implementation of Lefevre's
"pattern finding" algorithm. With a few small changes though, it should
end up beating most other methods in common use, regardless of the size
of the constant (currently, it's often one or two shifts worse)
TODO: rewrite it so it's not hideously ugly (this is a translation from
perl, which doesn't help ;)
bypass most of it for multiplies by 2^n+1
(eventually) teach it that some combinations of shift+add are
cheaper than others (e.g. shladd on ia64, scaled adds on alpha)
get it to try multiple booth encodings in search of the cheapest
routine
make it work for negative constants
This is hacked up as a DAG->DAG transform, so once I clean it up I hope
it'll be pulled out of here and put somewhere else. The only thing backends
should really have to worry about for now is where to draw the line
between using this code vs. going ahead and doing an integer multiply
anyway.
llvm-svn: 21560
2005-04-26 07:23:02 +00:00
76dab9a523
* Implement StrLenOptimization
...
* Factor out commonalities between StrLenOptimization and StrCatOptimization
* Make sure that signatures return sbyte* not void*
llvm-svn: 21559
2005-04-26 05:24:00 +00:00
3f2710e64e
* Add a test case for StrLenOptimization
...
* Rename ExitInMain and StrCat tests so they don't have the date the
regression was entered since they are feature tests, not regressions.
llvm-svn: 21558
2005-04-26 05:22:38 +00:00
8ee5aacc38
Incorporate feedback from Chris:
...
* Change signatures of OptimizeCall and ValidateCalledFunction so they are
non-const, allowing the optimization object to be modified. This is in
support of caching things used across multiple calls.
* Provide two functions for constructing and caching function types
* Modify the StrCatOptimization to cache Function objects for strlen and
llvm.memcpy so it doesn't regenerate them on each call site. Make sure
these are invalidated each time we start the pass.
* Handle both a GEP Instruction and a GEP ConstantExpr
* Add additional checks to make sure we really are dealing with an arary of
sbyte and that all the element initializers are ConstantInt or
ConstantExpr that reduce to ConstantInt.
* Make sure the GlobalVariable is constant!
* Don't use ConstantArray::getString as it can fail and it doesn't give us
the right thing. We must check for null bytes in the middle of the array.
* Use llvm.memcpy instead of memcpy so we can factor alignment into it.
* Don't use void* types in signatures, replace with sbyte* instead.
llvm-svn: 21555
2005-04-26 03:26:15 +00:00
d45bdb45c1
Add SimplyLibCalls.cpp to VC++ build
...
llvm-svn: 21554
2005-04-26 02:57:49 +00:00
342fe9a833
Fix RUN: line to not always pass.
...
llvm-svn: 21553
2005-04-26 02:33:25 +00:00
cfa7ddd6e2
Fold (X > -1) | (Y > -1) --> (X&Y > -1)
...
llvm-svn: 21552
2005-04-26 01:18:33 +00:00
fe91dfec91
Changes due to code review and new implementation:
...
* Don't use std::string for the function names, const char* will suffice
* Allow each CallOptimizer to validate the function signature before
doing anything
* Repeatedly loop over the functions until an iteration produces
no more optimizations. This allows one optimization to insert a
call that is optimized by another optimization.
* Implement the ConstantArray portion of the StrCatOptimization
* Provide a template for the MemCpyOptimization
* Make ExitInMainOptimization split the block, not delete everything
after the return instruction.
(This covers revision 1.3 and 1.4, as the 1.3 comments were botched)
llvm-svn: 21548
2005-04-25 21:20:38 +00:00
f806459d90
implement some more logical compares with constants, so that:
...
int foo1(int x, int y) {
int t1 = x >= 0;
int t2 = y >= 0;
return t1 & t2;
}
int foo2(int x, int y) {
int t1 = x == -1;
int t2 = y == -1;
return t1 & t2;
}
produces:
_foo1:
or r2, r4, r3
srwi r2, r2, 31
xori r3, r2, 1
blr
_foo2:
and r2, r4, r3
addic r2, r2, 1
li r2, 0
addze r3, r2
blr
instead of:
_foo1:
srwi r2, r4, 31
xori r2, r2, 1
srwi r3, r3, 31
xori r3, r3, 1
and r3, r2, r3
blr
_foo2:
addic r2, r4, 1
li r2, 0
addze r2, r2
addic r3, r3, 1
li r3, 0
addze r3, r3
and r3, r2, r3
blr
llvm-svn: 21547
2005-04-25 21:20:28 +00:00
f2534c7291
Lots of changes based on review and new functionality:
...
* Use a
llvm-svn: 21546
2005-04-25 21:11:48 +00:00
172e9155c1
Update the test case to handle a few more (degenerate) cases and remove
...
the restriction that it is an XFAIL because it now passes.
llvm-svn: 21545
2005-04-25 21:08:34 +00:00
d373ff64aa
Codegen x < 0 | y < 0 as (x|y) < 0. This allows us to compile this to:
...
_foo:
or r2, r4, r3
srwi r3, r2, 31
blr
instead of:
_foo:
srwi r2, r4, 31
srwi r3, r3, 31
or r3, r2, r3
blr
llvm-svn: 21544
2005-04-25 21:03:25 +00:00
e093c6f565
Make dominates(A,B) work with post dominators. Patch contributed by
...
Naveen Neelakantam, thanks!
llvm-svn: 21543
2005-04-25 20:50:33 +00:00
af0ac2744e
Added question about turning off all optimizations. I think this has been asked once or twice.
...
llvm-svn: 21542
2005-04-25 20:36:56 +00:00
a21bf8d1be
implement getelementptr.ll:test10
...
llvm-svn: 21541
2005-04-25 20:17:30 +00:00
20621b1e94
rename fn
...
llvm-svn: 21540
2005-04-25 20:17:16 +00:00
e680ee2a48
new testcase
...
llvm-svn: 21539
2005-04-25 20:17:00 +00:00
ecac782786
Correctly handle global-argument aliases induced in main
...
llvm-svn: 21537
2005-04-25 19:16:31 +00:00
5965359d8f
Don't mess up SCC traversal when a node has null edges out of it.
...
llvm-svn: 21536
2005-04-25 19:16:17 +00:00
37b6b097ff
document 'opaque' types
...
llvm-svn: 21535
2005-04-25 17:34:15 +00:00
d1039cc581
Add feedback from Vikram
...
llvm-svn: 21534
2005-04-25 15:47:57 +00:00
478d3b930a
Make sure the target buffer is null terminated so we don't blow up
...
strcat when its called.
llvm-svn: 21533
2005-04-25 15:40:35 +00:00
20b0e43e1f
A test case for testing the StrCatOptimizer, currently XFAILed everywhere.
...
llvm-svn: 21532
2005-04-25 07:29:30 +00:00
9bbaa2ab7f
Post-Review Cleanup:
...
* Fix comments at top of file
* Change algorithm for running the call optimizations from n*n to something
closer to n.
* Use a hash_map to store and lookup the optimizations since there will
eventually (or potentially) be a large number of them. This gets lookup
based on the name of the function to O(1). Each CallOptimizer now has a
std::string member named func_name that tracks the name of the function
that it applies to. It is this string that is entered into the hash_map
for fast comparison against the function names encountered in the module.
* Cleanup some style issues pertaining to iterator invalidation
* Don't pass the Function pointer to the OptimizeCall function because if
the optimization needs it, it can get it from the CallInst passed in.
* Add the skeleton for a new CallOptimizer, StrCatOptimizer which will
eventually replace strcat's of constant strings with direct copies.
llvm-svn: 21526
2005-04-25 03:59:26 +00:00
23423346f8
Use the %name rather than the "name" format so those familiar with the
...
llvm-dis output don't go blind.
llvm-svn: 21525
2005-04-25 03:18:19 +00:00
ff7b16c1d6
Shut GCC 4.0 up about classes that have virtual functions but a non-virtual
...
destructor. Just add the do-nothing virtual destructor.
llvm-svn: 21524
2005-04-25 02:55:55 +00:00
0c2d046aa4
Declare a function to create the SimplifyLibCalls pass.
...
llvm-svn: 21523
2005-04-25 02:54:00 +00:00
39a762d149
A new pass to provide specific optimizations for certain well-known library
...
calls. The pass visits all external functions in the module and determines
if such function calls can be optimized. The optimizations are specific to
the library calls involved. This initial version only optimizes calls to
exit(3) when they occur in main(): it changes them to ret instructions.
llvm-svn: 21522
2005-04-25 02:53:12 +00:00
985f484263
A test case for the the ExitInMain libcall simplification.
...
llvm-svn: 21521
2005-04-25 02:50:08 +00:00
8edc8beacf
Older compilers won't like the inline virtual destructor in the header file
...
so we put the destructor in Pass.cpp and make it non-inline.
llvm-svn: 21520
2005-04-25 01:01:35 +00:00
fc104de06d
fix some bugs
...
llvm-svn: 21515
2005-04-25 00:38:52 +00:00
756d049c06
Fix a thinko in the documentation of the splitBasicBlock method. The branch
...
instruction is added to the original block, not the new block.
llvm-svn: 21513
2005-04-25 00:31:53 +00:00
9c47b25868
Shut GCC 4.0 up about classes with virtual functions but no virtual
...
destructor.
llvm-svn: 21510
2005-04-24 22:27:20 +00:00
ad750a80eb
Shut GCC 4.0 up when it complains about classes with virtual functions that
...
don't have virtual destructors.
llvm-svn: 21507
2005-04-24 22:20:32 +00:00
4da978466d
Make this readable for newbies and those who can only understand one set of
...
grammar rules for the English language.
llvm-svn: 21503
2005-04-24 20:56:18 +00:00
6818b33096
extract has been renamed to llvm-extract to avoid conflicting with another tool
...
llvm-svn: 21501
2005-04-24 17:46:58 +00:00
2f1457fd83
Eliminate cases where we could << by 64, which is undefined in C.
...
llvm-svn: 21500
2005-04-24 17:46:05 +00:00
ca1e0c6ae0
There are still uses for spaces in Makefiles -- to make text line up together,
...
regardless of the tab size/stop settings on the developer side
llvm-svn: 21499
2005-04-24 17:43:41 +00:00
e22594f3d7
extract has been renamed to llvm-extract to avoid conflicting with another tool
...
llvm-svn: 21498
2005-04-24 17:36:05 +00:00
3e9634eaa0
elisp code to help with LLVM code standards compliance
...
llvm-svn: 21497
2005-04-24 17:09:19 +00:00
831ad84eea
.vimrc file to aid in LLVM coding standards conformance
...
llvm-svn: 21496
2005-04-24 17:05:04 +00:00
d6f636a340
Implement xor.ll:test21: select (not C), A, B -> select C, B, A
...
llvm-svn: 21495
2005-04-24 07:30:14 +00:00
b57d040464
Test that xor/select are folded into a select with inverted operands.
...
llvm-svn: 21494
2005-04-24 07:28:53 +00:00
2c7d177d25
Allow these methods to take a generic Value* to simplify clients. Use
...
const_cast instead of c casts.
llvm-svn: 21493
2005-04-24 07:28:37 +00:00
b220952aca
allow these to take a generic Value*
...
llvm-svn: 21492
2005-04-24 07:28:04 +00:00
d1f46d3bf9
Use getPrimitiveSizeInBits() instead of getPrimitiveSize()*8
...
Completely rework the 'setcc (cast x to larger), y' code. This code has
the advantage of implementing setcc.ll:test19 (being more general than
the previous code) and being correct in all cases.
This allows us to unxfail 2004-11-27-SetCCForCastLargerAndConstant.ll,
and close PR454.
llvm-svn: 21491
2005-04-24 06:59:08 +00:00
0de0638522
unxfail this.
...
llvm-svn: 21490
2005-04-24 06:55:40 +00:00
624d5cb006
add a new testcase which occurs in 181.mcf
...
llvm-svn: 21489
2005-04-24 06:55:33 +00:00
1c8d93b08f
* The aesthetic police is on patrol!!...
...
* ... but it wasn't so busy as to not smell the roses and doxygenify comments
llvm-svn: 21487
2005-04-23 22:35:26 +00:00
396890d3ac
Add a helper method
...
llvm-svn: 21486
2005-04-23 22:20:22 +00:00
8a878ccb25
Fix a bug in my previous checkin
...
llvm-svn: 21485
2005-04-23 22:01:39 +00:00
2908bc1e7a
This file is never referenced
...
llvm-svn: 21484
2005-04-23 22:00:26 +00:00
6263766133
Add a method, remove last use of Type.def
...
llvm-svn: 21483
2005-04-23 22:00:09 +00:00
816394d7ad
add a method, remove a dead #include
...
llvm-svn: 21482
2005-04-23 21:59:42 +00:00
6226f941b0
Types.def is going away
...
llvm-svn: 21481
2005-04-23 21:59:11 +00:00
82639853c0
Eliminate tabs and trailing spaces
...
llvm-svn: 21480
2005-04-23 21:38:35 +00:00
8047f13a88
Eliminate tabs and trailing spaces
...
llvm-svn: 21479
2005-04-23 21:26:11 +00:00
4bbf66b7a2
Propagate eq sets through the bu graphs to the cbu and eq graphs, fixing
...
a crash of the sfv on 188.ammp
llvm-svn: 21478
2005-04-23 21:11:05 +00:00
fdfb25a9e1
add a bunch of documentation about the LLVM type resolution machinery
...
llvm-svn: 21475
2005-04-23 17:27:36 +00:00
7fa89cfc9b
Make the CBackend actually get included in llc by using USEDLIBS instead of
...
USEDLIB as the variable to which "CBackend" is appended. The surrounding
if clause is safe because currently the configure script ensures that the
CBackend target is always added to TARGETS_TO_BUILD. By using a non-hard
coded construct in the makefile, we gain uniformity and the ability to
change the default set of targets by only changing the configure script.
llvm-svn: 21474
2005-04-23 17:24:33 +00:00
c6ce71d434
make an advanced topics section, move symtab to it
...
llvm-svn: 21473
2005-04-23 16:10:52 +00:00
7eae28b9d9
This is not needed
...
llvm-svn: 21472
2005-04-23 16:00:27 +00:00
c5ad1ca942
add a search box for the documentation
...
llvm-svn: 21471
2005-04-23 15:55:18 +00:00
77c32c34d7
Generalize the setcc -> PHI and Select folding optimizations to work with
...
any constant RHS, not just a constant integer RHS. This implements
select.ll:test17
llvm-svn: 21470
2005-04-23 15:31:55 +00:00
64a8a7f9b7
add a new testcase
...
llvm-svn: 21469
2005-04-23 15:31:03 +00:00
01c81ab706
Always enable the C backend. This fixes a *vast* number of failures on the
...
testers last night, as llc was not getting the cbe linked in.
llvm-svn: 21468
2005-04-23 14:36:22 +00:00
0166db6529
Add method to return the type this type will be promoted to if it is
...
passed through a variable argument function.
llvm-svn: 21462
2005-04-23 00:11:21 +00:00
51efcefa49
Mention the --enable-targets= option to configure.
...
llvm-svn: 21459
2005-04-22 20:27:33 +00:00
7937b07940
* Order #includes as per style guide
...
* Combine multiple ``std::cerr <<'' statements into one for simplicity
llvm-svn: 21458
2005-04-22 19:13:22 +00:00
fd8c2265fa
Convert tabs to spaces
...
llvm-svn: 21457
2005-04-22 18:06:01 +00:00
36c6bc1bf4
Hyphenate the adjective `n-bit' when used to describe values and processors
...
llvm-svn: 21456
2005-04-22 18:02:52 +00:00
d5da38c9aa
Build the `Skeleton' target when building "all" targets
...
llvm-svn: 21455
2005-04-22 18:00:34 +00:00
1e195bdd41
Document the --enable-targets= feature of the configure script.
...
llvm-svn: 21454
2005-04-22 17:58:03 +00:00
7afc9558e0
Silence gcc-4.0.0 warnings.
...
llvm-svn: 21453
2005-04-22 17:56:01 +00:00
e73e76dc42
Convert tabs to spaces
...
llvm-svn: 21452
2005-04-22 17:54:37 +00:00
948f453254
Don't always build CBackend and Skeleton. Make use of the TARGETS_TO_BUILD
...
parameter instead which will correctly list the set of targets to be built.
llvm-svn: 21451
2005-04-22 17:32:05 +00:00
e4481c7c82
Implement the --enable-targets= feature of the configure script. The make
...
variable TARGETS_TO_BUILD is used to determine which targets in lib/Target
are built and which libraries are linked into llc. This effectively
implements the feature. One item remains: disabling targets in the dejagnu
test suite.
llvm-svn: 21450
2005-04-22 17:20:11 +00:00
ce9612b23c
Two changes:
...
1. Get rid of TOOLLINKOPTS as it is a hold over from llvm-test and only
used to communicate additional libraries to the linker. The *standard*
way to do that is with the LIBS variable which this change supports.
2. Allow the TARGETS_TO_BUILD variable to be set from the configuration
substitution. This is the result of the --enable-target= parameter to
the configure script.
llvm-svn: 21449
2005-04-22 17:14:14 +00:00
90dd65e4a7
Make sure the CBackend is always a target that is built.
...
llvm-svn: 21448
2005-04-22 17:06:18 +00:00
e482ca8bfa
Consolidate the target selection options into a single option,
...
--enable-target which can take values "all", "host-only" or a comma
separated list of target names (alpha,ia64,powerpc,skeleton,sparc,x86)
llvm-svn: 21447
2005-04-22 17:02:18 +00:00
8f4081a8c9
keep track of max depth stats
...
llvm-svn: 21446
2005-04-22 13:35:18 +00:00
4742804864
First step in avoiding compilation/usage of non-relevant targets. New
...
options have been added to the configure script that control which targets
will be used. The options are:
--enable-target-this (default=disabled)
This will specify that the target corresponding to the build host is
the target that will be compiled/used. You can't use this with any of
the other options (they'll be ignored). This is what most people want.
--disable-target-x86 (default=enabled)
This will prevent the X86 target(s) from being compiled/used.
--disable-target-sparc (default=enabled)
This will prevent both SparcV8 and SparcV9 from being compiled/used.
--disable-target-powerpc (default=enabled)
This will prevent the PowerPC target from being compiled/used.
--disable-target-alpha (default=enabled)
This will prevent the Alpha target from being compiled/used.
--disable-target-ia64 (default=enabled)
This will prevent the IA64 target from being compiled/used.
Note that without any of these options, the default behavior is to build
all targets, as is the current practice.
All these options do is set up the substititution variable TARGETS_TO_BUILD
which contains the targets that should be compiled/used. The variable is
intended to be used in the makefiles. Those changes will come later.
llvm-svn: 21445
2005-04-22 07:27:28 +00:00
42ed14860b
Updated dependence analyzer. Fixed numerous bugs. Same stage scheduling, etc.
...
llvm-svn: 21444
2005-04-22 06:32:48 +00:00
e25cfaee55
Malloc/Free have mod/ref effects. Do not allow CSE of function calls that
...
call malloc/free. This fixes PR555.
llvm-svn: 21443
2005-04-22 05:36:59 +00:00
16431f7b3c
remove 'another common example', which doesn't work with VC++, and indent
...
another example properly
llvm-svn: 21442
2005-04-22 04:49:59 +00:00
88e7b7288a
Eliminate tabs and trailing spaces
...
llvm-svn: 21441
2005-04-22 04:13:13 +00:00
5191b4b2d0
Convert tabs to spaces
...
llvm-svn: 21440
2005-04-22 04:08:30 +00:00
774511633d
Convert tabs to spaces
...
llvm-svn: 21439
2005-04-22 04:01:18 +00:00
1fabffbf37
Convert tabs to spaces
...
llvm-svn: 21438
2005-04-22 03:46:24 +00:00
4b45ec0be0
Convert tabs to spaces
...
llvm-svn: 21437
2005-04-22 03:35:49 +00:00
204572af9d
Convert tabs to spaces
...
llvm-svn: 21436
2005-04-22 03:27:20 +00:00
8c0ca76517
Convert tabs to spaces
...
llvm-svn: 21433
2005-04-22 03:18:56 +00:00
a3d4e04da8
Add the RUN: prefix to the start of the run line so this test doesn't fail.
...
BTW .. isn't the date on this funky? We haven't reachec 2005-05-09 yet???
llvm-svn: 21432
2005-04-22 02:49:25 +00:00
7275e3fb31
Change some old-style casts to C++ style casts to avoid warnings in XPS
...
compilation. This change has been waiting in the wings for a long time but
since Misha just did a global change, I figured now was the time to commit
it.
llvm-svn: 21431
2005-04-22 02:31:56 +00:00
d6312210d1
There is no =head3. Convert to =head2.
...
llvm-svn: 21430
2005-04-22 02:16:19 +00:00
84f3911b09
Correct the documentation to match the program.
...
llvm-svn: 21429
2005-04-22 02:12:41 +00:00
650ba8eb56
Remove trailing whitespace
...
llvm-svn: 21428
2005-04-22 00:00:37 +00:00
b1c9317bb4
Remove trailing whitespace
...
llvm-svn: 21427
2005-04-21 23:48:37 +00:00
c88330ad13
* Remove trailing whitespace
...
* Convert tabs to spaces
llvm-svn: 21426
2005-04-21 23:38:14 +00:00
b440243e94
Remove trailing whitespace
...
llvm-svn: 21425
2005-04-21 23:30:14 +00:00
89b8c8d215
Remove trailing whitespace
...
llvm-svn: 21424
2005-04-21 23:13:11 +00:00
10468d8a3c
Remove trailing whitespace
...
llvm-svn: 21422
2005-04-21 22:55:34 +00:00
91fb9ab60b
* Remove trailing whitespace
...
* Convert tabs to spaces
llvm-svn: 21421
2005-04-21 22:43:08 +00:00
835702a094
Remove trailing whitespace
...
llvm-svn: 21420
2005-04-21 22:36:52 +00:00
b47d28bfe8
* Remove trailing whitespace
...
* Convert tabs to spaces
llvm-svn: 21418
2005-04-21 21:48:46 +00:00
2ecc34dd3c
Remove trailing whitespace
...
llvm-svn: 21417
2005-04-21 21:44:41 +00:00
01808caded
Remove trailing whitespace
...
llvm-svn: 21416
2005-04-21 21:13:18 +00:00
13f332cd3f
* Remove trailing whitespace
...
* Convert tabs to spaces
llvm-svn: 21415
2005-04-21 21:10:11 +00:00
95e43ae162
Check for eqv matching
...
llvm-svn: 21414
2005-04-21 21:10:07 +00:00
3590ef1164
Match another form of eqv
...
llvm-svn: 21413
2005-04-21 21:09:11 +00:00
df586aa20e
Remove trailing whitespace
...
llvm-svn: 21412
2005-04-21 20:59:05 +00:00
95aab4ef6d
Remove trailing whitespace
...
llvm-svn: 21411
2005-04-21 20:48:15 +00:00
5c9ef1ba2d
this passes on darwin
...
llvm-svn: 21410
2005-04-21 20:48:08 +00:00
01230b24ae
Remove trailing whitespace
...
llvm-svn: 21409
2005-04-21 20:39:54 +00:00
01cada8797
Remove trailing whitespace
...
llvm-svn: 21408
2005-04-21 20:19:05 +00:00
486a1ec909
Handle stores of global address as stores of immediates. Instead of:
...
test1:
movl $N, %eax
movl %eax, G
ret
emit:
test1:
movl $N, G
ret
llvm-svn: 21407
2005-04-21 19:11:03 +00:00
adcfc1748b
Handle (store &GV -> mem) as a store immediate. This often occurs for
...
printf format strings and other stuff. Instead of generating this:
movl $l1__2E_str_1, %eax
movl %eax, (%esp)
we now emit:
movl $l1__2E_str_1, (%esp)
llvm-svn: 21406
2005-04-21 19:03:24 +00:00
c8b5d60937
Use the actual uid/gid for defaulting the fields in the archive.
...
llvm-svn: 21405
2005-04-21 17:49:57 +00:00
a3159af703
Fix a bug where we would not promote calls to invokes if they occured in
...
the same block as the setjmp. Thanks to Greg Pettyjohn for noticing this!
llvm-svn: 21403
2005-04-21 16:46:46 +00:00
ce7ce1628f
Eliminate calls to system dependent function getuid by using
...
the newly implemented sys::Process::GetCurrentUserId function. Replace
similarly for getgid.
llvm-svn: 21402
2005-04-21 16:15:19 +00:00
2d45e25125
Provide an implementation of the GetCurrentUserId and GetCurrentGroupId
...
methods that were recently added to the interface.
llvm-svn: 21401
2005-04-21 16:12:57 +00:00
5e92ead30d
Add two new methods for getting the User Id and Group Id values for the
...
current process.
llvm-svn: 21400
2005-04-21 16:12:04 +00:00
c346bf59ee
Add doxygen comments, patch contributed by Evan Jones.
...
llvm-svn: 21397
2005-04-21 16:10:03 +00:00
4c5c96f35f
add support for taking and resolving the address of free.
...
llvm-svn: 21396
2005-04-21 16:09:43 +00:00
631ee94316
add support for taking the address of free.
...
llvm-svn: 21395
2005-04-21 16:08:59 +00:00
954c64d532
Improve doxygen, from part of Evan's patch that didn't apply.
...
llvm-svn: 21394
2005-04-21 16:06:03 +00:00
7ceb081f3f
Improve doxygen documentation, patch contributed by Evan Jones!
...
llvm-svn: 21393
2005-04-21 16:04:49 +00:00
f6302441f0
Improve and elimination. On PPC, for:
...
bool %test(int %X) {
%Y = and int %X, 8
%Z = setne int %Y, 0
ret bool %Z
}
we now generate this:
rlwinm r2, r3, 0, 28, 28
srwi r3, r2, 3
instead of this:
rlwinm r2, r3, 0, 28, 28
srwi r2, r2, 3
rlwinm r3, r2, 0, 31, 31
I'll leave it to Nate to get it down to one instruction. :)
---------------------------------------------------------------------
llvm-svn: 21391
2005-04-21 06:28:15 +00:00
ab1ed77570
Fold (x & 8) != 0 and (x & 8) == 8 into (x & 8) >> 3.
...
This turns this PPC code:
rlwinm r2, r3, 0, 28, 28
cmpwi cr7, r2, 8
mfcr r2
rlwinm r3, r2, 31, 31, 31
into this:
rlwinm r2, r3, 0, 28, 28
srwi r2, r2, 3
rlwinm r3, r2, 0, 31, 31
Next up, nuking the extra and.
llvm-svn: 21390
2005-04-21 06:12:41 +00:00
374e659466
Instcombine this:
...
%shortcirc_val = select bool %tmp.1, bool true, bool %tmp.4 ; <bool> [#uses=1]
%tmp.6 = cast bool %shortcirc_val to int ; <int> [#uses=1]
into this:
%shortcirc_val = or bool %tmp.1, %tmp.4 ; <bool> [#uses=1]
%tmp.6 = cast bool %shortcirc_val to int ; <int> [#uses=1]
not this:
%tmp.4.cast = cast bool %tmp.4 to int ; <int> [#uses=1]
%tmp.6 = select bool %tmp.1, int 1, int %tmp.4.cast ; <int> [#uses=1]
llvm-svn: 21389
2005-04-21 05:43:13 +00:00
b38b443b15
Teach simplifycfg that setcc is cheap and non-trapping, so that it can
...
convert this:
%tmp.1 = seteq int %i, 0 ; <bool> [#uses=1]
br bool %tmp.1, label %shortcirc_done, label %shortcirc_next
shortcirc_next: ; preds = %entry
%tmp.4 = seteq int %j, 0 ; <bool> [#uses=1]
br label %shortcirc_done
shortcirc_done: ; preds = %shortcirc_next, %entry
%shortcirc_val = phi bool [ %tmp.4, %shortcirc_next ], [ true, %entry ] ; <bool> [#uses=1]
to this:
%tmp.1 = seteq int %i, 0 ; <bool> [#uses=1]
%tmp.4 = seteq int %j, 0 ; <bool> [#uses=1]
%shortcirc_val = select bool %tmp.1, bool true, bool %tmp.4 ; <bool> [#uses=1]
... which is later simplified by instcombine into an or.
llvm-svn: 21388
2005-04-21 05:31:13 +00:00
b7848974cd
Fix some broken links, taking care of PR554
...
llvm-svn: 21387
2005-04-21 04:53:58 +00:00
0727518f81
update to match build changes.
...
llvm-svn: 21386
2005-04-21 04:52:37 +00:00
8424ba373e
For Bug 543:
...
Standardize the error messages to be in "path: what failed: why" format.
Also attempt to use the correct errno to ThrowErrno in situations where
the errno value is erased by subsequent system calls.
llvm-svn: 21385
2005-04-21 02:50:10 +00:00
b02566dc1c
For Bug 543:
...
Allow the ThrowErrno function to optionally accept an error number
parameter so that callers can specify the error number to be used.
llvm-svn: 21384
2005-04-21 02:30:32 +00:00
2f72bafb60
Remove trailing whitespace at the end of lines
...
llvm-svn: 21380
2005-04-20 16:42:34 +00:00
ea548c0cb1
Remove trailing whitespace, patch by Markus Oberhumer.
...
llvm-svn: 21379
2005-04-20 16:05:03 +00:00
5e7bac21d5
Add FIXME by Markus Oberhumer from bug 545: not checking for "." in $PATH may
...
result in returning executable files that won't be runnable.
llvm-svn: 21378
2005-04-20 15:42:11 +00:00
a9a8c1b65b
Do not mark directories as `executable', we only want program files
...
Patch by Markus Oberhumer.
llvm-svn: 21377
2005-04-20 15:33:22 +00:00
ee6770f1a3
#include system headers after all LLVM headers
...
llvm-svn: 21374
2005-04-20 04:51:29 +00:00
17092bd0a4
Eliminate trailing spaces at end-of-line
...
llvm-svn: 21372
2005-04-20 04:08:35 +00:00
3f4fe46796
Consistently eschew space between `*' or `&' and function argument name
...
llvm-svn: 21371
2005-04-20 04:07:47 +00:00
1001aea4d3
Ignore dangling symlinks in getDirectoryContents()
...
Thanks to Markus Oberhumer for the patch!
llvm-svn: 21370
2005-04-20 04:04:07 +00:00
8cd50bf269
Initialize fields mode, uid, and gid.
...
Patch by Markus Oberhumer. Thanks!
llvm-svn: 21369
2005-04-20 03:55:35 +00:00
bebcaeb816
Align comments together for consistency
...
llvm-svn: 21368
2005-04-20 03:52:59 +00:00
64a8c3ed91
* Print commands as we execute them with `-v'
...
* Add option `-save-temps'
Patch contributed by Markus Oberhumer.
llvm-svn: 21367
2005-04-20 03:22:18 +00:00
8cb10a1775
Wrap some long lines.
...
Make IPSCCP strip off dead constant exprs that are using functions, making
them appear as though their address is taken. This allows us to propagate
some more pool descriptors, lowering the overhead of pool alloc.
llvm-svn: 21363
2005-04-19 19:16:19 +00:00
14db4a2895
ignore generated files
...
llvm-svn: 21362
2005-04-19 15:34:36 +00:00
b17483028f
fix bogus warning
...
llvm-svn: 21361
2005-04-19 15:32:30 +00:00
07ad98dbe2
fix PR549
...
llvm-svn: 21360
2005-04-19 15:27:29 +00:00
181c818373
Bug fixed
...
llvm-svn: 21355
2005-04-19 06:08:04 +00:00
5c219469a0
Eliminate a broken transformation, fixing PR548
...
llvm-svn: 21354
2005-04-19 06:04:18 +00:00
53c40624f9
Add completely untested support for mtcrf/mfcrf encoding
...
llvm-svn: 21353
2005-04-19 05:41:52 +00:00
baa9be572b
switch over the rest of the formats that use RC to use isDOT
...
llvm-svn: 21352
2005-04-19 05:21:30 +00:00
f9172e14c9
Convert the XForm instrs and XSForm instruction over to use isDOT
...
llvm-svn: 21351
2005-04-19 05:15:18 +00:00
5b78da4571
Now that the ppc64 and vmx operands of I are always 0, forward substitute
...
them away.
llvm-svn: 21350
2005-04-19 05:05:22 +00:00
116a9e5a34
convert over bform and iform instructions
...
llvm-svn: 21349
2005-04-19 05:00:59 +00:00
b2367e398e
Convert over DForm and DSForm instructions
...
llvm-svn: 21348
2005-04-19 04:59:28 +00:00
15709c2c33
Convert XLForm and XForm instructions over to use PPC64 when appropriate.
...
llvm-svn: 21347
2005-04-19 04:51:30 +00:00
d790d22f22
Convert XO XS and XFX forms to use isPPC64
...
llvm-svn: 21346
2005-04-19 04:40:07 +00:00
c7cb8c77fb
Turn PPC64 and VMX into classes that can be added to instructions instead of
...
bits that must be passed up the inheritance hierarchy. Convert MForm and AForm
instructions over
llvm-svn: 21345
2005-04-19 04:32:54 +00:00
ef94374a1c
Major change to tblgen: instead of resolving values every time a class is
...
finished up, only resolve fully when the def is defined. This allows things
to be changed and all uses to be propagated through. This implements
TableGen/LazyChange.td and fixes TemplateArgRename.td in the process.
None of the .td files used in LLVM backends are changed at all by this
patch.
llvm-svn: 21344
2005-04-19 03:36:21 +00:00
ab77740715
New testcase for a changing values late and allowing them to propagate
...
llvm-svn: 21343
2005-04-19 03:34:58 +00:00
d90d566dee
Make this significantly harder
...
llvm-svn: 21342
2005-04-19 02:58:57 +00:00
17bd7c5a88
Add a real run line
...
llvm-svn: 21341
2005-04-19 02:53:26 +00:00
c00f712fc0
fix this testcase
...
llvm-svn: 21340
2005-04-19 02:52:04 +00:00
2b76da7d1f
new testcase
...
llvm-svn: 21339
2005-04-19 02:26:23 +00:00
f2fab50974
add a run line
...
llvm-svn: 21338
2005-04-19 01:37:24 +00:00
ed8335e35b
Tell dj to run the tests in this directory
...
llvm-svn: 21337
2005-04-19 01:36:41 +00:00
3ff0e11294
implementing shifting of literal integers
...
llvm-svn: 21336
2005-04-19 01:17:35 +00:00
101fc501d0
Add initial lexer and parser support for shifting values. Every use of this
...
will lead to it being rejected though.
llvm-svn: 21335
2005-04-19 01:11:03 +00:00
2331c061ee
Next round of PPC CR optimizations. For the following code:
...
int %bar(float %a, float %b, float %c, float %d) {
entry:
%tmp.1 = setlt float %a, %d
%tmp.2 = setlt float %b, %d
%or = or bool %tmp.1, %tmp.2
%tmp.3 = setgt float %c, %d
%tmp.4 = or bool %or, %tmp.3
%tmp.5 = and bool %tmp.4, true
%retval = cast bool %tmp.5 to int
ret int %retval
}
We now emit:
_bar:
.LBB_bar_0: ; entry
fcmpu cr0, f1, f4
fcmpu cr1, f2, f4
cror 0, 0, 4
fcmpu cr1, f3, f4
cror 28, 0, 5
mfcr r2
rlwinm r3, r2, 29, 31, 31
blr
Instead of:
_bar:
.LBB_bar_0: ; entry
fcmpu cr7, f1, f4
mfcr r2
rlwinm r2, r2, 29, 31, 31
fcmpu cr7, f2, f4
mfcr r3
rlwinm r3, r3, 29, 31, 31
or r2, r2, r3
fcmpu cr7, f3, f4
mfcr r3
rlwinm r3, r3, 30, 31, 31
or r3, r2, r3
blr
llvm-svn: 21321
2005-04-18 07:48:09 +00:00
ee84413730
silence a bogus warning
...
llvm-svn: 21320
2005-04-18 05:26:21 +00:00
b61ecb5875
Fold setcc of MVT::i1 operands into logical operations
...
llvm-svn: 21319
2005-04-18 04:48:12 +00:00
6d40fd01fe
Another minor simplification: handle setcc (zero_extend x), c -> setcc(x, c')
...
llvm-svn: 21318
2005-04-18 04:30:45 +00:00
868d473009
Another simple xform
...
llvm-svn: 21317
2005-04-18 04:11:19 +00:00
bd22d83d15
Fold:
...
// (X != 0) | (Y != 0) -> (X|Y != 0)
// (X == 0) & (Y == 0) -> (X|Y == 0)
Compiling this:
int %bar(int %a, int %b) {
entry:
%tmp.1 = setne int %a, 0
%tmp.2 = setne int %b, 0
%tmp.3 = or bool %tmp.1, %tmp.2
%retval = cast bool %tmp.3 to int
ret int %retval
}
to this:
_bar:
or r2, r3, r4
addic r3, r2, -1
subfe r3, r3, r2
blr
instead of:
_bar:
addic r2, r3, -1
subfe r2, r2, r3
addic r3, r4, -1
subfe r3, r3, r4
or r3, r2, r3
blr
llvm-svn: 21316
2005-04-18 03:59:53 +00:00
d929f8bcd3
Make the AND elimination operation recursive and significantly more powerful,
...
eliminating an and for Nate's testcase:
int %bar(int %a, int %b) {
entry:
%tmp.1 = setne int %a, 0
%tmp.2 = setne int %b, 0
%tmp.3 = or bool %tmp.1, %tmp.2
%retval = cast bool %tmp.3 to int
ret int %retval
}
generating:
_bar:
addic r2, r3, -1
subfe r2, r2, r3
addic r3, r4, -1
subfe r3, r3, r4
or r3, r2, r3
blr
instead of:
_bar:
addic r2, r3, -1
subfe r2, r2, r3
addic r3, r4, -1
subfe r3, r3, r4
or r2, r2, r3
rlwinm r3, r2, 0, 31, 31
blr
llvm-svn: 21315
2005-04-18 03:48:41 +00:00
602a45f415
Change codegen for setcc to read the bit directly out of the condition
...
register. Added support in the .td file for the g5-specific variant
of cr -> gpr moves that executes faster, but we currently don't
generate it.
llvm-svn: 21314
2005-04-18 02:43:24 +00:00
ed5189da51
Add support for targets that require stubs for external functions.
...
llvm-svn: 21313
2005-04-18 01:44:27 +00:00
f281f791b5
Handle ExternalSymbol operands in the PPC JIT
...
llvm-svn: 21312
2005-04-18 00:46:10 +00:00
f24225d4d7
Update dejagnu tests to use the new pattern isel flag
...
llvm-svn: 21311
2005-04-16 04:25:48 +00:00
df4f480498
Add CondPropagate.cpp to VC++ Transforms project
...
llvm-svn: 21310
2005-04-16 02:51:44 +00:00
779c5cbb44
Make pattern isel default for ppc
...
Add new ppc beta option related to using condition registers
Make pattern isel control flag (-enable-pattern-isel) global and tristate
0 == off
1 == on
2 == target default
llvm-svn: 21309
2005-04-15 22:12:16 +00:00
e0a9d042e2
new pass
...
llvm-svn: 21307
2005-04-15 21:13:16 +00:00
16a50fd0a0
a new simple pass, which will be extended to be more useful in the future.
...
This pass forward branches through conditions when it can show that the
conditions is either always true or false for a predecessor. This currently
only handles the most simple cases of this, but is successful at threading
across 2489 branches and 65 switch instructions in 176.gcc, which isn't bad.
llvm-svn: 21306
2005-04-15 19:28:32 +00:00
d041242256
add a new prototype
...
llvm-svn: 21305
2005-04-15 19:24:49 +00:00
8579c2c474
new testcase
...
llvm-svn: 21304
2005-04-15 19:24:36 +00:00
00ce283b3f
fix calls
...
llvm-svn: 21303
2005-04-14 17:34:20 +00:00
7ae3aba5aa
a 21264 fix, and fix the operator precidence on an and -> zap check (should fix hundreds of test cases
...
llvm-svn: 21302
2005-04-14 16:24:00 +00:00
10b7bceb8e
added a random and mask test
...
llvm-svn: 21301
2005-04-14 16:17:49 +00:00
0a7c2b9078
print negative 64 bit immediates as negative numbers, makes things a little
...
easier on the eyes, not that numbers like 18446744073709541376 are bad or
anything
llvm-svn: 21300
2005-04-14 10:08:01 +00:00
dfbbcc098b
oops, this stopped us turning movl r4=0xFFFFFFFF;; and rX, r4 into zxt4
...
llvm-svn: 21299
2005-04-14 10:06:35 +00:00
53d3eccbe2
Implement multi-way branches through logical ops on condition registers.
...
This can generate considerably shorter code, reducing the size of crafty
by almost 1%. Also fix the printing of mcrf. The code is currently
disabled until it gets a bit more testing, but should work as-is.
llvm-svn: 21298
2005-04-14 09:45:08 +00:00
80c095f422
Add a couple missing transforms in getSetCC that were triggering assertions
...
in the PPC Pattern ISel
llvm-svn: 21297
2005-04-14 08:56:52 +00:00
f6b666fb06
we have zextloads, not sextloads!
...
llvm-svn: 21296
2005-04-14 08:37:32 +00:00
65a82c562e
Add the necessary support to codegen condition register logical ops with
...
register allocated condition registers. Make sure that the printed
output is gas compatible.
llvm-svn: 21295
2005-04-14 03:20:38 +00:00
be21011911
Start allocating condition registers. Almost all explicit uses of CR0 are
...
now gone. Next step is to get rid of the remaining ones and then start
allocating bools to CRs where appropriate.
llvm-svn: 21294
2005-04-13 23:15:44 +00:00
Reid Spencer