1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-30 15:32:52 +01:00
Commit Graph

9780 Commits

Author SHA1 Message Date
Andrew Lenharth
e831777542 turn on IEEE for compares
llvm-svn: 20425
2005-03-03 22:12:11 +00:00
Andrew Lenharth
e6dbf989b3 beter Select on FP
llvm-svn: 20424
2005-03-03 21:47:53 +00:00
Chris Lattner
4439f1686f Print -X like this:
double test(double l1_X) {
  return (-l1_X);
}

instead of like this:

double test(double l1_X) {
  return (-0x0p+0 - l1_X);
}

llvm-svn: 20423
2005-03-03 21:12:04 +00:00
Andrew Lenharth
b5ddbc074d LSR cleanup patch
llvm-svn: 20422
2005-03-03 19:03:21 +00:00
Chris Lattner
8074739aa2 Do not lower malloc's to pass "sizeof" expressions like this:
ltmp_0_7 = malloc(((unsigned )(&(((signed char (*)[784])/*NULL*/0)[1u]))));

Instead, just emit the literal constant, like this:

  ltmp_0_7 = malloc(784u);

This works around a bug in ICC 8.1 compiling the CBE generated code.  :-(

llvm-svn: 20415
2005-03-03 01:04:50 +00:00
Chris Lattner
4814696e7d Add an optional argument to lower to a specific constant value instead of
to a "sizeof" expression.

llvm-svn: 20414
2005-03-03 01:03:43 +00:00
Misha Brukman
5b587350ee Fix the spelling of the word `the'
llvm-svn: 20412
2005-03-02 23:17:31 +00:00
Chris Lattner
f9597dc689 Print the module ID as a comment.
llvm-svn: 20411
2005-03-02 23:12:40 +00:00
Chris Lattner
b205d87afe cleanup the cfg after lsr
llvm-svn: 20410
2005-03-02 21:56:00 +00:00
Andrew Lenharth
1e213c7924 remove 32 sign extend after 32 sextload and handle small negative constant
llvm-svn: 20408
2005-03-02 17:23:03 +00:00
Andrew Lenharth
8fc5ba2e06 Added LSR as a beta pass for alpha
llvm-svn: 20407
2005-03-02 17:21:38 +00:00
Chris Lattner
798b18474c Add a temporary option for llc-beta: -enable-lsr-for-ppc, which turns on
Loop Strength Reduction.

llvm-svn: 20399
2005-03-02 06:19:22 +00:00
Reid Spencer
a5fbf1d659 Be slightly more accurate in an error message.
llvm-svn: 20397
2005-03-02 05:45:56 +00:00
Chris Lattner
0bb2828efb Fix a nasty order of evaluation bug that Gabor Greif ran into. Here's an
explanation from IRC:

	|sabre|	I think it's an order of evaluation thing
	|sabre|	for me, the RHS of the assignment is evaluated first
	|sabre|	getTypeDescription checks to see if ConcreteTypeDescription[Ty] contains anything
	|sabre|	since it doesn't, it computes and returns the value
	|sabre|	this gets put into the map.
	|sabre|	For you, the LHS is evaluated first.
	|sabre|	Map[Ty] (aka ConcreteTypeDescriptions[Ty]) inserts an empty string into the map, returning a reference
	|sabre|	getTypeDesc then sees the empty string in the map
	|sabre|	and returns it
	|sabre|	bork :)

llvm-svn: 20394
2005-03-02 03:54:43 +00:00
Jeff Cohen
6d82d5b23e Fixed the following LSR bugs:
* Loop invariant code does not dominate the loop header, but rather
    the end of the loop preheader.

  * The base for a reduced GEP isn't a constant unless all of its
    operands (preceding the induction variable) are constant.

  * Allow induction variable elimination for the simple case after all.

Also made changes recommended by Chris for properly deleting
instructions.

llvm-svn: 20383
2005-03-01 03:46:11 +00:00
Alkis Evlogimenos
422af394b6 Lower llvm.isunordered(a, b) into a != a | b != b.
llvm-svn: 20382
2005-03-01 02:07:58 +00:00
Chris Lattner
9d57998cda Remove tabs from file.
llvm-svn: 20380
2005-02-28 19:36:15 +00:00
Chris Lattner
b2720f5b57 Add support to the C backend for llvm.prefetch. Patch contributed by
Justin Wick!

llvm-svn: 20378
2005-02-28 19:29:46 +00:00
Chris Lattner
82480f68d7 recognize llvm.prefetch. Patch contributed by Justin Wick!
llvm-svn: 20377
2005-02-28 19:28:00 +00:00
Chris Lattner
c4205a6b93 Verify llvm.prefetch.
llvm-svn: 20376
2005-02-28 19:27:42 +00:00
Chris Lattner
9ccfcab3db Lower prefetch to a noop, patch contributed by Justin Wick!
llvm-svn: 20375
2005-02-28 19:27:23 +00:00
Andrew Lenharth
7dc9ea9509 fix integer division and stuff
llvm-svn: 20372
2005-02-28 17:22:18 +00:00
Jeff Cohen
d5b1827c3f Fix crash in LSR due to attempt to remove original induction variable. However,
for reasons explained in the comments, I also deactivated this code as it needs
more thought.

llvm-svn: 20367
2005-02-28 00:08:56 +00:00
Jeff Cohen
fd9504c7d9 PHI nodes were incorrectly placed when more than one GEP is reduced in a loop.
llvm-svn: 20360
2005-02-27 21:08:04 +00:00
Jeff Cohen
6258d4a431 First pass at improved Loop Strength Reduction. Still not yet ready for prime time.
llvm-svn: 20358
2005-02-27 19:37:07 +00:00
Chris Lattner
b632a13aa7 Use const iterators where possible. Patch by Evan Jones!
llvm-svn: 20354
2005-02-27 19:06:10 +00:00
Chris Lattner
73d4556bb6 Teach globalopt how memset/cpy/move affect memory, to allow better optimization.
llvm-svn: 20352
2005-02-27 18:58:52 +00:00
Chris Lattner
a024984017 Fix spelling, patch contributed by Gabor Greif!
llvm-svn: 20343
2005-02-27 06:18:25 +00:00
Chris Lattner
cf3862ce8d Fix spelling, patch contributed by Gabor Greif
llvm-svn: 20342
2005-02-27 06:15:51 +00:00
Chris Lattner
a17076b771 Remove some stuff I checked in accidentally
llvm-svn: 20340
2005-02-27 04:32:35 +00:00
Chris Lattner
2311dcd08d DCE a dead function
llvm-svn: 20339
2005-02-26 23:36:45 +00:00
Reid Spencer
24b41ba78d Implement an isBytecodeArchive method to determine if an archive contains
bytecode file members or not.
Patch Contributed By Adam Treat

llvm-svn: 20338
2005-02-26 22:00:32 +00:00
Chris Lattner
cf3cda8125 1 + 100 + 51 == 152, not 52.
If we fold three constants together (c1+c2+c3), make sure to keep
LHSC updated, instead of reusing (in this case), the 1 instead of the
partial sum.

llvm-svn: 20337
2005-02-26 18:50:19 +00:00
Chris Lattner
14f720d625 remove extraneous cast
llvm-svn: 20334
2005-02-26 18:33:28 +00:00
Andrew Lenharth
b5331ffe0f make BB labels be exported for debuging, add fp negation optimization, further pecimise the FP instructions
llvm-svn: 20332
2005-02-25 22:55:15 +00:00
Chris Lattner
9340ba4bf9 Handle null a bit more carefully.
Actually teach dsa about select instructions.  This doesn't affect the
graph in any way other than not setting a spurious U marker on pointer
nodes that are selected.

llvm-svn: 20324
2005-02-25 01:27:48 +00:00
Chris Lattner
16f321bbe7 This instruction:
X = gep null, ...

Used to not create a scalar map entry for X, which caused clients to barf.
This is bad.

llvm-svn: 20316
2005-02-24 19:55:31 +00:00
Chris Lattner
085b39c9e0 Fix a bug introduced by revision 1.187 of this file.
llvm-svn: 20308
2005-02-24 18:48:07 +00:00
Andrew Lenharth
ef5f87784b fix Allocas. Really. I mean it this time.
llvm-svn: 20306
2005-02-24 18:36:32 +00:00
Chris Lattner
af54bd6050 Fix some problems where the verifier would crash on invalid input instead of
reporting the problem and exiting.

llvm-svn: 20302
2005-02-24 16:58:29 +00:00
Chris Lattner
7a434679c3 Implement Transforms/SimplifyCFG/switch_thread.ll
This does a simple form of "jump threading", which eliminates CFG edges that
are provably dead.  This triggers 90 times in the external tests, and
eliminating CFG edges is always always a good thing! :)

llvm-svn: 20300
2005-02-24 06:17:52 +00:00
Chris Lattner
902d9dc660 switch instructions only allow constantints for their values, be more specific.
llvm-svn: 20298
2005-02-24 05:32:09 +00:00
Chris Lattner
608a8c9f55 use more specific cast.
llvm-svn: 20297
2005-02-24 05:26:04 +00:00
Chris Lattner
ce949bdbfc add more checking
llvm-svn: 20296
2005-02-24 05:25:17 +00:00
Chris Lattner
f0863ee08c Do not read free'd memory when printing an error message.
llvm-svn: 20295
2005-02-24 04:59:49 +00:00
Chris Lattner
8044aa8d33 add a new method.
llvm-svn: 20293
2005-02-24 02:37:26 +00:00
Tanya Lattner
b640bb0d88 Only print out machine instructions before modulo scheduling if we are actually doing modulo scheduling! :)
llvm-svn: 20292
2005-02-24 02:14:44 +00:00
Andrew Lenharth
69a8320c0d Ah the problems you have to fix when you stray from the One True Way (TM)
llvm-svn: 20290
2005-02-23 17:33:42 +00:00
Chris Lattner
bfb6a94126 make this more efficient. Scan up to 16 nodes, not the whole list.
llvm-svn: 20289
2005-02-23 16:53:04 +00:00
Chris Lattner
a91c25c69b new method
llvm-svn: 20288
2005-02-23 16:51:11 +00:00
Chris Lattner
0ce5361846 Reduce the amount of searching this assertion does. On a testcase of mine,
this reduces the time for -simplifycfg in a debug build from 106s to 14.82s

llvm-svn: 20286
2005-02-23 07:09:08 +00:00
Chris Lattner
9838ab1271 Silence some uninit variable warnings.
llvm-svn: 20284
2005-02-23 05:57:21 +00:00
Tanya Lattner
a981a711aa Fixed bug in findAllcircuits. Fixed branch addition to schedule. Added debug information.
llvm-svn: 20280
2005-02-23 02:01:42 +00:00
Andrew Lenharth
889efe4fb3 oops
llvm-svn: 20278
2005-02-22 23:29:25 +00:00
Chris Lattner
1969249f13 Remove use of bind_obj, deleter, and finegrainify namespacification.
llvm-svn: 20277
2005-02-22 23:27:21 +00:00
Chris Lattner
b5256c157d Remove use of bind_obj
llvm-svn: 20276
2005-02-22 23:22:58 +00:00
Chris Lattner
d888514f0c C++ is not a functional programming language.
llvm-svn: 20274
2005-02-22 23:13:58 +00:00
Andrew Lenharth
d870103306 dynamic stack allocas
llvm-svn: 20273
2005-02-22 21:59:48 +00:00
Chris Lattner
4ba91f5168 Fix a bug in the 'store fpimm, ptr' -> 'store intimm, ptr' handling code.
Changing 'op' here caused us to not enter the store into a map, causing
reemission of the code!!  In practice, a simple loop like this:

no_exit:                ; preds = %no_exit, %entry
        %indvar = phi uint [ %indvar.next, %no_exit ], [ 0, %entry ]            ; <uint> [#uses=3]
        %tmp.4 = getelementptr "complex long double"* %P, uint %indvar, uint 0          ; <double*> [#uses=1]
        store double 0.000000e+00, double* %tmp.4
        %indvar.next = add uint %indvar, 1              ; <uint> [#uses=2]
        %exitcond = seteq uint %indvar.next, %N         ; <bool> [#uses=1]
        br bool %exitcond, label %return, label %no_exit

was being code gen'd to:

.LBBtest_1:     # no_exit
        movl %edx, %esi
        shll $4, %esi
        movl $0, 4(%eax,%esi)
        movl $0, (%eax,%esi)
        incl %edx
        movl $0, (%eax,%esi)
        movl $0, 4(%eax,%esi)
        cmpl %ecx, %edx
        jne .LBBtest_1  # no_exit

Note that we are doing 4 32-bit stores instead of 2.  Now we generate:

.LBBtest_1:     # no_exit
        movl %edx, %esi
        incl %esi
        shll $4, %edx
        movl $0, (%eax,%edx)
        movl $0, 4(%eax,%edx)
        cmpl %ecx, %esi
        movl %esi, %edx
        jne .LBBtest_1  # no_exit

This is much happier, though it would be even better if the increment of ESI
was scheduled after the compare :-/

llvm-svn: 20265
2005-02-22 07:23:39 +00:00
Andrew Lenharth
8ead0f13d3 no longer build as a shared library
llvm-svn: 20264
2005-02-22 04:58:26 +00:00
Chris Lattner
68c342b28f Fix problems running the HowToUseJIT on powerpc, and probably problems with
ANY program that does not have all functions internalized.

llvm-svn: 20258
2005-02-20 18:43:35 +00:00
Jeff Cohen
91e04e17a7 Fix silly mistake.
llvm-svn: 20256
2005-02-20 02:48:51 +00:00
Jeff Cohen
96558e2f93 Implement standard I/O redirection in ExecuteAndWait().
llvm-svn: 20255
2005-02-20 02:43:04 +00:00
Chris Lattner
15759af51d Add support for ".so" files compiled with LLVM which contain LLVM bytecode.
llvm-svn: 20253
2005-02-19 18:30:29 +00:00
Chris Lattner
042a54de90 Eliminate silly warnings from the linker of the form:
WARNING: Type conflict between types named 'union.._604.'.
    Src=' %union.._604.'.
   Dest=' %union.._604.'
llvm-svn: 20252
2005-02-19 17:52:37 +00:00
Jeff Cohen
b83e650f1f Change __MINGW to __MINGW32__. Patch submitted by Henrik Bach.
llvm-svn: 20243
2005-02-19 03:01:13 +00:00
Jeff Cohen
fa31c775b6 Make PreventCoreFiles() do the right thing on Windows.
llvm-svn: 20237
2005-02-18 07:05:18 +00:00
Misha Brukman
ea2c511191 Fix compilation errors with VS 2005, patch contributed by Aaron Gray.
llvm-svn: 20232
2005-02-17 21:40:27 +00:00
Misha Brukman
381d248dc6 Fix compilation errors with VS 2005, patch by Aaron Gray.
llvm-svn: 20231
2005-02-17 21:39:27 +00:00
Chris Lattner
89105cec43 Don't rely on doubles comparing identical to each other, which doesn't work
for 0.0 and -0.0.

llvm-svn: 20230
2005-02-17 20:17:32 +00:00
Chris Lattner
0de03b45ab Don't sink argument loads into loops or other bad places. This disables folding of argument loads with instructions that are not in the entry block.
llvm-svn: 20228
2005-02-17 19:40:32 +00:00
Chris Lattner
29126b65c4 Do not mark obviously unreachable blocks live when processing PHI nodes,
and handle incomplete control dependences correctly.  This fixes:

Regression/Transforms/ADCE/dead-phi-edge.ll
  -> a missed optimization

Regression/Transforms/ADCE/dead-phi-edge.ll
  -> a compiler crash distilled from QT4

llvm-svn: 20227
2005-02-17 19:28:49 +00:00
Chris Lattner
2198bb53ee Scary typo that fixes Regression/Transforms/IndVarsSimplify/2005-02-17-TruncateExprCrash.ll
and PR515.

llvm-svn: 20224
2005-02-17 16:54:16 +00:00
Jeff Cohen
72f7799517 Arg list already has program name in it.
llvm-svn: 20208
2005-02-16 04:43:45 +00:00
Tanya Lattner
a35f2f428e Fixed node deletion bug.
llvm-svn: 20207
2005-02-16 04:00:59 +00:00
Chris Lattner
6a86178364 Instead of doing a manual comparison loop, just use memcmp, thanks to JohnC
for the suggestion! :)

llvm-svn: 20203
2005-02-15 22:12:10 +00:00
Chris Lattner
4139a6ea3d Make this more efficient now that we know both files are the same length.
llvm-svn: 20202
2005-02-15 22:01:43 +00:00
Misha Brukman
5c9328b088 Fix spelling
llvm-svn: 20201
2005-02-15 21:59:53 +00:00
Reid Spencer
3c3c05bfbd Adjust DiffFilesWithTolerance to help poor cygwin's mmap facility by
handling zero length files a little more intelligently. If both files are
zero length then we return 0 (true) indicating a match. If only one of the
files is zero length then we return 1 (false) indicating that the files
differ. If the files don't agree in length then they can't match so we
skip the first loop that looks for a quick match.

llvm-svn: 20200
2005-02-15 21:47:02 +00:00
Chris Lattner
5174b9cb60 Fix a problem where the PPC backend lost track of the fact that it had
to save and restore the LR register on entry and exit of a leaf function
that needed to access globals or the constant pool.  This should hopefully
fix oscar from sending the PPC tester spinning out of control.

llvm-svn: 20197
2005-02-15 20:26:49 +00:00
Chris Lattner
c5646b3a15 Add a sanity check.
llvm-svn: 20195
2005-02-15 18:48:48 +00:00
Chris Lattner
e7d2b05fb0 Add a new method to make it easy to update graphs.
llvm-svn: 20194
2005-02-15 18:40:55 +00:00
Chris Lattner
bf72146607 Fix volatile load/store of pointers. Consider this testcase:
void %test(int** %P) {
  %A = volatile load int** %P
  ret void
}

void %test2(int*** %Q) {
  %P = load int*** %Q
  volatile store int** %P, int*** %Q
  ret void
}

instead of emitting:

void test(int **l1_P) {
  int *l2_A;

  l2_A = (int **((volatile int **)l1_P));
  return;
}
void test2(int ***l2_Q) {
  int **l1_P;

  l1_P = *l2_Q;
  *((volatile int ***)l2_Q) = l1_P;
  return;
}

... which is loading/storing volatile pointers, not through volatile pointers,
emit this (which is right):

void test(int **l1_P) {
  int *l3_A;

  l3_A = *((int * volatile*)l1_P);
  return;
}
void test2(int ***l2_Q) {
  int **l1_P;

  l1_P = *l2_Q;
  *((int ** volatile*)l2_Q) = l1_P;
  return;
}

llvm-svn: 20191
2005-02-15 05:52:14 +00:00
Chris Lattner
51c55602c9 Fix a bug in my previous change to this, which broke the build on sparcs.
llvm-svn: 20184
2005-02-14 21:42:10 +00:00
Chris Lattner
43b14db4d9 Print GEP offsets as signed values instead of unsigned values. On X86, this
prints:

getelementptr (int* %A, int -1)

as: "(A) - 4" instead of "(A) + 18446744073709551612", which makes the
assembler much happier.

This fixes test/Regression/CodeGen/X86/2005-02-14-IllegalAssembler.ll,
and Benchmarks/Prolangs-C/cdecl with LLC on X86.

llvm-svn: 20183
2005-02-14 21:40:26 +00:00
Chris Lattner
0a61115d67 Fix the second bug attached to PR504.
llvm-svn: 20181
2005-02-14 20:11:45 +00:00
Chris Lattner
786b30e148 Work around GCC PR19958, which causes programs to sometimes crash after
printing help output or version info.

llvm-svn: 20180
2005-02-14 19:17:29 +00:00
Misha Brukman
7b6a863954 Write out single characters as chars, not strings.
llvm-svn: 20179
2005-02-14 18:52:35 +00:00
Chris Lattner
ef836a918e Implement CodeGen/CBackend/2005-02-14-VolatileOperations.ll
Volatile loads and stores need to emit volatile pointer operations in C.

llvm-svn: 20177
2005-02-14 16:47:52 +00:00
Andrew Lenharth
f023ce8d97 fix setcc on floats, fixes singlesource:pi, perhaps others
llvm-svn: 20172
2005-02-14 05:41:43 +00:00
Chris Lattner
efacfe896c Fix the llvm bootstrap
llvm-svn: 20170
2005-02-13 23:37:09 +00:00
Chris Lattner
405367eb58 Move helper function here.
llvm-svn: 20168
2005-02-13 23:13:47 +00:00
Chris Lattner
535796ff26 If errno is zero strerror_r does not modify the buffer, leaving it unterminated.
This causes garbage to be printed out after error messages.

llvm-svn: 20165
2005-02-13 22:46:37 +00:00
Reid Spencer
a7a01f13df Make the check for global variables the same as the one for functions. In
both cases they are looking for non-external variables/functions that do
not have internal linkage. Using "!isExternal()" is a little more
understandable than "hasInitializer()"

llvm-svn: 20155
2005-02-13 18:12:20 +00:00
Chris Lattner
0ac02cee3c Nuke blank line.
llvm-svn: 20154
2005-02-13 17:54:21 +00:00
Chris Lattner
1194167daa Minor cleanup. No need to explicitly tell the compiler the template arguments.
llvm-svn: 20153
2005-02-13 17:50:16 +00:00
Chris Lattner
1788cf35ab Make sure to clear the LazyFunctionLoadMap after we ParseAllFunctionBodies.
Otherwise, clients who call ParseAllFunctionBodies will attempt to parse
the function bodies twice, which is (uh) very very bad (tm).

This fixes gccld on python.

llvm-svn: 20152
2005-02-13 17:48:18 +00:00
Chris Lattner
fdc746ac38 Do not put internal symbols into the symbol table. This shrinks the symbol
table for archives in common cases, and prevents trying to resolve a
external reference with an internal reference.  This shrinks the libpython.a
symbol table from 126302 to 19770 bytes.

llvm-svn: 20151
2005-02-13 17:42:11 +00:00
Chris Lattner
da614a0e0b Print something useful for gccld -v with an archive.
llvm-svn: 20148
2005-02-13 15:26:14 +00:00
Chris Lattner
63d47e902f Correct the recursive PHI node handling routines in a way that CANNOT induce
infinite loops (using the new replaceSymbolicValuesWithConcrete method).

This patch reverts this patch:
http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20050131/023830.html

... which was an attempted fix for this problem.  Unfortunately, that patch
caused test/Regression/Transforms/IndVarsSimplify/exit_value_tests.llx to fail
and slightly castrated the entire analysis.  This patch fixes it right.

This patch is dedicated to jeffc, for making me deal with this.  :)

llvm-svn: 20146
2005-02-13 04:37:18 +00:00
Andrew Lenharth
e398f7797e try to do better match for i32 adds
llvm-svn: 20143
2005-02-12 21:11:17 +00:00
Andrew Lenharth
089b56ae58 make FP conversion more conservative (matches gcc)
llvm-svn: 20142
2005-02-12 21:10:58 +00:00
Andrew Lenharth
b9c44170a5 oops, I was sure this had already gond though the nightly tester
llvm-svn: 20141
2005-02-12 20:42:09 +00:00
Andrew Lenharth
a12e5330bf added sign extend for boolean
llvm-svn: 20137
2005-02-12 19:35:12 +00:00
Chris Lattner
6e3279c0b4 Allow globals to be of different const'nesses when we link.
This finally resolves PR502, PR450,
and test/Regression/Linker/2005-02-12-ConstantGlobals{,-2}.ll correctly

llvm-svn: 20135
2005-02-12 19:20:28 +00:00
Chris Lattner
8f3f72f2bc Fix for testcase Transforms/IndVarsSimplify/2005-02-11-InvokeCrash.ll
and PR504.

llvm-svn: 20129
2005-02-12 03:26:49 +00:00
Andrew Lenharth
076faf95a8 fix a bunch of regressions due to call behavior
llvm-svn: 20110
2005-02-10 20:10:38 +00:00
Alkis Evlogimenos
82f384712f Localize globals if they are only used in main(). This replaces the
global with an alloca, which eventually gets promoted into a
register. This enables a lot of other optimizations later on.

llvm-svn: 20109
2005-02-10 18:36:30 +00:00
Tanya Lattner
04a78e13df Added new circuit finding alogrithm.
Fixed bug in graph so that phi ite diff edges are added.

llvm-svn: 20108
2005-02-10 17:02:58 +00:00
Tanya Lattner
cd54968a72 Allow modsched and local scheduling to both be run.
llvm-svn: 20107
2005-02-10 17:02:06 +00:00
Andrew Lenharth
56c441caf2 so, if you beat on it, you too can talk emacs into having a sane indenting policy... Also, optimize many function calls with pc-relative calls (partial prologue skipping for that case coming soon), try to fix the random jumps to strange places problem by pesimizing div et. al. register usage and fixing up GP before using, some calling convention tweaks, and make frame pointer unallocatable (not strickly necessary, but let's go for correctness first)
llvm-svn: 20106
2005-02-10 06:25:22 +00:00
Andrew Lenharth
6c28128e3e fix fp branch
llvm-svn: 20105
2005-02-10 05:17:38 +00:00
Misha Brukman
4bd124492b * Fix spelling of `volatile'
* Align comments with tablegen elements

llvm-svn: 20103
2005-02-10 01:52:22 +00:00
Chris Lattner
f1a9655358 Don't print a 'Total Execution Time' line for the 'Miscellaneous Ungrouped
Timers' section.  Since these are random timers in the program it doesn't
make sense to sum them up.

llvm-svn: 20090
2005-02-09 18:41:32 +00:00
Chris Lattner
343a831bab Fix test/Regression/Assembler/2005-02-09-AsmWriterStoreBug.ll
llvm-svn: 20089
2005-02-09 17:45:03 +00:00
Chris Lattner
82ef7e89fd Use new edge iterators to simplify some code.
llvm-svn: 20086
2005-02-09 03:20:43 +00:00
Andrew Lenharth
d42ae810cb BranchCC, nifty
llvm-svn: 20067
2005-02-08 00:40:03 +00:00
Andrew Lenharth
71fce71669 fix store issue and an FP conversion (segfault) issue
llvm-svn: 20066
2005-02-07 23:02:23 +00:00
Chris Lattner
bfd643f504 IndCallGraphMap is now a pointer to a new'd map.
llvm-svn: 20065
2005-02-07 16:09:15 +00:00
Andrew Lenharth
cf4f405e55 copytoreg fix
llvm-svn: 20063
2005-02-07 06:31:44 +00:00
Andrew Lenharth
d20853f420 copyfromreg fix
llvm-svn: 20062
2005-02-07 06:21:37 +00:00
Andrew Lenharth
80cf648100 fix load bug
llvm-svn: 20061
2005-02-07 05:55:55 +00:00
Andrew Lenharth
9f5502e40f more FP load store fixes and Load store simplifications
llvm-svn: 20060
2005-02-07 05:33:15 +00:00
Andrew Lenharth
bc6ddca09c clean up load and stores alot
llvm-svn: 20059
2005-02-07 05:18:02 +00:00
Andrew Lenharth
4416315969 teach all loads and stores about the stack
llvm-svn: 20058
2005-02-07 05:07:00 +00:00
Andrew Lenharth
1b8bf311d2 prefer FP scratch registers and more check in LowerArguments
llvm-svn: 20057
2005-02-06 21:07:31 +00:00
Andrew Lenharth
23ca0026fa fix oopso
llvm-svn: 20056
2005-02-06 16:22:15 +00:00
Andrew Lenharth
baa723abc0 smarter loads and stores. can now handle base+offset.
llvm-svn: 20055
2005-02-06 15:40:40 +00:00
Andrew Lenharth
9a2bc47fba fix build
llvm-svn: 20053
2005-02-05 19:46:51 +00:00
Andrew Lenharth
6bd554a11e clean up
llvm-svn: 20051
2005-02-05 17:41:39 +00:00
Andrew Lenharth
9fd7ce4bca fix f32 setcc, and fp select
llvm-svn: 20050
2005-02-05 16:41:03 +00:00
Andrew Lenharth
5447bb6596 added ugly support for fp compares
llvm-svn: 20049
2005-02-05 13:19:12 +00:00
Misha Brukman
75da90f127 Make the rest of file header comments consistent in format and style
llvm-svn: 20048
2005-02-05 02:24:26 +00:00
Chris Lattner
ec9411df83 Instead of initializing the volatile field, use accessors to set it.
llvm-svn: 20045
2005-02-05 01:38:38 +00:00
Chris Lattner
d73939a8ed Initialize new field.
llvm-svn: 20044
2005-02-05 01:37:58 +00:00
Misha Brukman
74be40e1d2 Make file header comment consistent: extend the whole 80 cols to fill the line
llvm-svn: 20039
2005-02-04 20:25:52 +00:00
Chris Lattner
43b0441ccb If we have an indirect call site that calls N functions, inline the N functions
into a temporary graph, remember it for later, then inline the tmp graph into
the call site.

In the case where there are other call sites to the same set of functions, this
permits us to just inline the temporary graph instead of all of the callees.

This turns N*M inlining situations into an N+M inlining situation.

llvm-svn: 20036
2005-02-04 19:59:49 +00:00
Chris Lattner
5f8b7b03b0 Split mergeInGraph into two methods.
llvm-svn: 20035
2005-02-04 19:58:28 +00:00
Chris Lattner
48828f062c Fix the Regression/Transforms/DSAnalysis/recursion.ll regression.
llvm-svn: 20031
2005-02-04 18:58:04 +00:00
Chris Lattner
c808a143af Fix a case where were incorrectly compiled cast from short to int on 64-bit
targets.

llvm-svn: 20030
2005-02-04 18:39:19 +00:00
Andrew Lenharth
e081ab1c69 alignment
llvm-svn: 20028
2005-02-04 14:09:38 +00:00
Andrew Lenharth
68f8792889 get alignment printing correctly and get rid of __main hack
llvm-svn: 20027
2005-02-04 14:01:21 +00:00
Andrew Lenharth
d2d24eee40 fix constant pointer outputing on 64 bit machines
llvm-svn: 20026
2005-02-04 13:47:16 +00:00
Andrew Lenharth
fa74ac60e6 FP fixes
llvm-svn: 20019
2005-02-03 21:01:15 +00:00
Chris Lattner
15151dbff8 Refactor getFunctionArgumentsForCall out of mergeInGraph.
llvm-svn: 20018
2005-02-03 18:40:25 +00:00
Chris Lattner
11f76e4975 This is no longer needed. Global variables with undef initializers can be
initialized to anything, including garbage.

llvm-svn: 20010
2005-02-02 20:50:50 +00:00
Andrew Lenharth
d5de7adf26 Store fix
llvm-svn: 20004
2005-02-02 17:32:39 +00:00
Andrew Lenharth
fef75b04f1 oops
llvm-svn: 20003
2005-02-02 17:01:31 +00:00
Andrew Lenharth
b4bf49a4ae prevent register allocator from using the stack pointer :)
llvm-svn: 20002
2005-02-02 17:00:21 +00:00
Andrew Lenharth
c3e3bd1c22 fix loading of floats
llvm-svn: 19997
2005-02-02 15:05:33 +00:00
Andrew Lenharth
2482a0ef99 marked mem* as not supported
llvm-svn: 19992
2005-02-02 05:49:42 +00:00
Alkis Evlogimenos
3521fcb18f Fix crash on MallocInsts of unsized types.
llvm-svn: 19988
2005-02-02 04:43:37 +00:00
Andrew Lenharth
a856b4db61 fix Load bug
llvm-svn: 19987
2005-02-02 04:35:44 +00:00
Chris Lattner
c3f476e9c2 Fix yet another memset issue.
llvm-svn: 19986
2005-02-02 03:44:41 +00:00
Andrew Lenharth
35ae745650 try to make a bug bugpointable, add yet more constant pool stuff, fixup constant loads for FP
llvm-svn: 19985
2005-02-02 03:36:35 +00:00
Andrew Lenharth
5a2bb3de8b better constant handling, should fix many remaining cases
llvm-svn: 19984
2005-02-02 00:51:15 +00:00
Chris Lattner
56e08cc9a3 Eliminate some duplicated debug code
llvm-svn: 19980
2005-02-01 21:55:40 +00:00
Chris Lattner
57cdba4154 Eliminate self-recursion as a special case.
llvm-svn: 19979
2005-02-01 21:49:43 +00:00
Chris Lattner
b35595d4c0 Eliminate use of DSCallSiteIterator in key loop. This is a half step to
a tasty speedup.

llvm-svn: 19978
2005-02-01 21:37:27 +00:00
Andrew Lenharth
172fc4b1fd fix FP arg passing bug, Add unsigned to/from int, fix SELECT, fix Constant pool
llvm-svn: 19976
2005-02-01 20:40:27 +00:00
Andrew Lenharth
9086064b72 Print the Constant pool
llvm-svn: 19975
2005-02-01 20:38:53 +00:00
Andrew Lenharth
540700124d Make cmov work right and loads for fp from constant pool
llvm-svn: 19974
2005-02-01 20:36:44 +00:00
Andrew Lenharth
bf88e70920 Correct stack stuff for FP
llvm-svn: 19973
2005-02-01 20:35:57 +00:00
Andrew Lenharth
810fa6d4f1 try to match alpha pattern
llvm-svn: 19972
2005-02-01 20:35:11 +00:00
Andrew Lenharth
b8e15cfe9c fix register names
llvm-svn: 19971
2005-02-01 20:34:29 +00:00
Chris Lattner
028a8a8f56 Signficantly speed up printing by not emitting the same file twice with
different names.  Large SCC's tend to be big, so this saves a lot of time.

llvm-svn: 19970
2005-02-01 19:10:48 +00:00
Chris Lattner
9cf60e3459 Fix some bugs andrew noticed legalizing memset for alpha
llvm-svn: 19969
2005-02-01 18:38:28 +00:00
Chris Lattner
26d4a4b2d2 Do not revisit nodes in the SCC traversal. This speeds up the BU pass a bit.
llvm-svn: 19968
2005-02-01 17:35:52 +00:00
Chris Lattner
78e8e86a9e Fix test/Regression/Assembler/2005-01-31-CallingAggregateFunction.ll
llvm-svn: 19966
2005-02-01 01:47:42 +00:00
Chris Lattner
2947b1be51 Apparently := confuses makellvm
llvm-svn: 19965
2005-02-01 01:47:12 +00:00
Andrew Lenharth
fa52a84802 pecimise loads, put indirect call addr in right register. still doesn't fix methcall
llvm-svn: 19963
2005-02-01 01:37:24 +00:00
Chris Lattner
839b0ef616 Updates for new use list changes.
llvm-svn: 19961
2005-02-01 01:24:21 +00:00
Chris Lattner
1cbca58b87 Update for API change.
llvm-svn: 19960
2005-02-01 01:24:01 +00:00
Chris Lattner
d04865822c API change.
llvm-svn: 19959
2005-02-01 01:23:49 +00:00
Chris Lattner
555ef32a44 Adjust to changes in APIs
llvm-svn: 19958
2005-02-01 01:23:31 +00:00
Chris Lattner
849a4118e6 Hacks to make this ugly ugly code work with the new use lists.
llvm-svn: 19957
2005-02-01 01:22:56 +00:00
Chris Lattner
d9b3839cec Fix a problem where we could infinitely recurse on phi nodes.
llvm-svn: 19955
2005-02-01 00:18:30 +00:00
Misha Brukman
1e8e58b829 Fix hyphenation in output comment
llvm-svn: 19954
2005-01-31 06:19:57 +00:00
Chris Lattner
c4884b553e Implement InstCombine/cast.ll:test25, a case that occurs many times
in spec

llvm-svn: 19953
2005-01-31 05:51:45 +00:00
Chris Lattner
6a431c25d1 Implement the trivial cases in InstCombine/store.ll
llvm-svn: 19950
2005-01-31 05:36:43 +00:00
Chris Lattner
a4ba8e9f0d Implement Transforms/InstCombine/cast-load-gep.ll, which allows us to devirtualize
11 indirect calls in perlbmk.

llvm-svn: 19947
2005-01-31 04:50:46 +00:00
Andrew Lenharth
648e85bb8a indirect call fix
llvm-svn: 19945
2005-01-31 03:19:31 +00:00
Andrew Lenharth
401bb5807b fp to int and back conversion sequences
llvm-svn: 19944
2005-01-31 01:44:26 +00:00
Chris Lattner
e4aaa4cf01 Fix the regressions my User changes introduced. Apparently some parts of
LLVM make the very reasonable assumption that constant expressions will
have at least one operand! :)

llvm-svn: 19943
2005-01-31 01:11:13 +00:00
Chris Lattner
11a25d99d7 Rename variables to work with VC++'s hokey scoping rules.
llvm-svn: 19942
2005-01-31 00:10:58 +00:00
Chris Lattner
86f89c506f Fix some scary bugs that VC++ detected.
llvm-svn: 19941
2005-01-31 00:10:45 +00:00
Chris Lattner
ee6bc42f3e * Make some methods more const correct.
* Change the FunctionCalls and AuxFunctionCalls vectors into std::lists.
  This makes many operations on these lists much more natural, and avoids
  *exteremely* expensive copying of DSCallSites (e.g. moving nodes around
  between lists, erasing a node from not the end of the vector, etc).

With a profile build of analyze, this speeds up BU DS from 25.14s to
12.59s on 176.gcc.  I expect that it would help TD even more, but I don't
have data for it.

This effectively eliminates removeIdenticalCalls and children from the
profile, going from 6.53 to 0.27s.

llvm-svn: 19939
2005-01-30 23:51:02 +00:00
Andrew Lenharth
43c294ffc3 added fp extend and removed a forgotten assert in more than 6 arg support (should break somewhere else now :) ) and fix an incorrect asm sequence for indirect calls
llvm-svn: 19938
2005-01-30 20:42:36 +00:00
Chris Lattner
c952d46d13 This code is really unreachable.
llvm-svn: 19934
2005-01-30 16:33:46 +00:00
Chris Lattner
d27e3639ba Fix warnings.
llvm-svn: 19933
2005-01-30 16:32:48 +00:00
Andrew Lenharth
60966b9bf0 support for larger calls
llvm-svn: 19932
2005-01-30 00:35:27 +00:00
Chris Lattner
382abe80a0 Improve conformance with the Misha spelling benchmark suite
llvm-svn: 19930
2005-01-30 00:09:23 +00:00
Tanya Lattner
560e1612df Make this work on systems where size_t == unsigned and where they are not
the same.

llvm-svn: 19929
2005-01-29 23:29:55 +00:00
Tanya Lattner
53173dac64 Make this work on systems where size_t is not the same as unsigned.
llvm-svn: 19928
2005-01-29 23:08:01 +00:00
Chris Lattner
fc8d0e9460 Unbreak the build :(
llvm-svn: 19926
2005-01-29 19:27:28 +00:00
Chris Lattner
8200976176 adjust to ilist changes.
llvm-svn: 19924
2005-01-29 18:41:25 +00:00
Chris Lattner
aa4ae8f5b3 Adjust to ilist changes.
llvm-svn: 19923
2005-01-29 18:41:12 +00:00
Chris Lattner
68040588d9 This file was schizophrenic when it came to representing sizes. In some
cases it represented them as 'unsigned's, which are not enough for 64-bit
hosts.  In other cases, it represented them as uint64_t's, which are
inefficient for 32-bit hosts.

This patch unifies all of the sizes to use size_t instead.

llvm-svn: 19918
2005-01-29 17:17:18 +00:00
Chris Lattner
af99e3993c After reading in a bc file, trim the resultant buffer down to what we
really need.  This reduces 4M of memory consumption reading 176.gcc.

llvm-svn: 19916
2005-01-29 17:05:56 +00:00
Chris Lattner
6bf1c6c6ae Finegrainify namespacification
llvm-svn: 19915
2005-01-29 16:53:02 +00:00
Andrew Lenharth
f426f1c0c9 first step towards a correct and complete stack. also add some forms for things that were getting stuck in the nightly tester.
llvm-svn: 19914
2005-01-29 15:42:07 +00:00
Chris Lattner
703dfdda2a Due to previous simplifications, we can simplify the data structures being
used here.

llvm-svn: 19913
2005-01-29 07:04:10 +00:00
Chris Lattner
0823ac4234 Properly handle volatile.
llvm-svn: 19912
2005-01-29 06:42:34 +00:00
Chris Lattner
8df4ba51b4 Remove some useless map operations. Loads/stores that are in the same
BB as the load are not included in the Cand* sets at all.

llvm-svn: 19911
2005-01-29 06:39:25 +00:00
Chris Lattner
af286e3600 Before doing expensive global analysis, check to make sure the pointer is
not invalidated on entry and on exit of the block.  This fixes some N^2
behavior in common cases, and speeds up gcc another 5% to 22.35s.

llvm-svn: 19910
2005-01-29 06:31:53 +00:00
Chris Lattner
43ccc5c945 Minor simplification/speedup. Replaces a set lookup with a pointer comparison.
This speeds up 176.gcc from 25.73s to 23.48s, which is 9.5%

llvm-svn: 19907
2005-01-29 06:20:55 +00:00
Chris Lattner
ae8d4bb675 Eliminate generality that is not buying us anything. In particular, this
will cause us to miss cases where the input pointer to a load could be value
numbered to another load.  Something like this:


  %X = load int* %P1
  %Y = load int* %P2

Those are obviously the same if P1/P2 are the same.  The code this patch
removes attempts to handle that.  However, since GCSE iterates, this doesn't
actually buy us anything: GCSE will first replace P1 or P2 with the other
one, then the load can be value numbered as equal.

Removing this code speeds up gcse a lot.  On 176.gcc in debug mode, this
speeds up gcse from 29.08s -> 25.73s, a 13% savings.

llvm-svn: 19906
2005-01-29 06:11:16 +00:00
Chris Lattner
68d73bed4a If we see:
%A = alloca int
  %V = load int* %A

value number %V to undef, not 0.

llvm-svn: 19905
2005-01-29 05:57:01 +00:00
Chris Lattner
3b3e7f7cc2 Memory used is a delta between memuse at the start of the time and the
memuse at the end, thus it is signed.

llvm-svn: 19904
2005-01-29 05:21:16 +00:00
Chris Lattner
35281b677a Make sure that we always grow a multiple of 2 operands.
llvm-svn: 19902
2005-01-29 01:05:12 +00:00
Chris Lattner
c29f25e260 Adjust to changes in instruction interfaces.
llvm-svn: 19900
2005-01-29 00:39:08 +00:00
Chris Lattner
7dab604f10 Switchinst takes a hint for the number of cases it will have.
llvm-svn: 19899
2005-01-29 00:38:45 +00:00
Chris Lattner
ccc0c99fae switchinst ctor now takes a hint for the number of cases that it will have.
llvm-svn: 19898
2005-01-29 00:38:26 +00:00
Chris Lattner
d87666619f Adjust Valuehandle to hold its operand directly in it.
llvm-svn: 19897
2005-01-29 00:37:36 +00:00
Chris Lattner
774d64469c Finegrainify namespacification.
Adjust TmpInstruction to work with the new User model.

llvm-svn: 19896
2005-01-29 00:36:59 +00:00
Chris Lattner
0b2dec0f59 add namespace qualifier
llvm-svn: 19895
2005-01-29 00:36:38 +00:00
Chris Lattner
031d5f649c Adjust to changes in User class and minor changes in instruction ctors.
llvm-svn: 19894
2005-01-29 00:36:19 +00:00
Chris Lattner
e0580122ab Adjust to slight changes in instruction interfaces.
llvm-svn: 19893
2005-01-29 00:35:55 +00:00
Chris Lattner
50d674e9da Adjust to changes in User class.
llvm-svn: 19892
2005-01-29 00:35:33 +00:00
Chris Lattner
12f9442dc3 Merge InstrTypes.cpp into this file
Adjust to changes in the User class, operand handling is very different.
PHI node and switch statements must handle explicit resizing of operand
lists.

llvm-svn: 19891
2005-01-29 00:35:16 +00:00
Chris Lattner
5244a770c3 Adjust to changes in User class. Aggregate constants now must explicitly
manage their operands.

llvm-svn: 19890
2005-01-29 00:34:39 +00:00
Chris Lattner
a0498b9190 This file is now merged into Instructions.cpp
llvm-svn: 19889
2005-01-29 00:33:32 +00:00
Andrew Lenharth
8a3a14d343 fix ExprMap, partially teach about add long
llvm-svn: 19882
2005-01-28 23:17:54 +00:00
Chris Lattner
f893f17907 Fix a nasty thinko in my previous commit.
llvm-svn: 19881
2005-01-28 23:17:27 +00:00
Chris Lattner
2755fb4171 Alpha doesn't have a native f32 extload instruction.
llvm-svn: 19880
2005-01-28 22:58:25 +00:00
Chris Lattner
da7b5277c1 implement legalization of truncates whose results and sources need to be
truncated, e.g. (truncate:i8 something:i16) on a 32 or 64-bit RISC.

llvm-svn: 19879
2005-01-28 22:52:50 +00:00
Chris Lattner
89cac82479 Get alpha working with memset/memcpy/memmove
llvm-svn: 19878
2005-01-28 22:29:18 +00:00
Chris Lattner
b035ec4a9c * add some DEBUG statements
* Properly compile this:

struct a {};
int test() {
  struct a b[2];
  if (&b[0] != &b[1])
    abort ();
  return 0;
}

to 'return 0', not abort().

llvm-svn: 19875
2005-01-28 19:32:01 +00:00
Chris Lattner
f24ea9cf5e Fix ConstProp/2005-01-28-SetCCGEP.ll: indexing over zero sized elements does
not change the address.

llvm-svn: 19874
2005-01-28 19:09:51 +00:00
Andrew Lenharth
9db35b0763 fix ExprMap and constant check in setcc
llvm-svn: 19870
2005-01-28 14:06:46 +00:00
Andrew Lenharth
4cfda09ee9 move FP into it's own select
llvm-svn: 19867
2005-01-28 06:57:18 +00:00
Chris Lattner
4134789c8f CopyFromReg produces two values. Make sure that we remember that both are
legalized, and actually return the correct result when we legalize the chain first.

llvm-svn: 19866
2005-01-28 06:27:38 +00:00
Chris Lattner
fadcc07232 Remove this code as it is currently completely broken and unmaintained.
If needed, this can be resurrected from CVS.

Note that several of the interfaces (e.g. the IPModRef ones) are supersumed
by generic AliasAnalysis interfaces that have been written since this code
was developed (and they are not DSA specific).

llvm-svn: 19864
2005-01-28 06:12:46 +00:00
Jeff Cohen
df055196d4 Properly close mapped files.
llvm-svn: 19863
2005-01-28 01:17:07 +00:00
Andrew Lenharth
c0cd77a1a0 stack frame fix and zero FP reg fix
llvm-svn: 19857
2005-01-27 08:31:19 +00:00
Andrew Lenharth
55eadc4772 Floating point instructions like Floating point registers
llvm-svn: 19856
2005-01-27 07:58:15 +00:00
Andrew Lenharth
67328d7fac int to float conversion and another setcc
llvm-svn: 19855
2005-01-27 07:50:35 +00:00
Misha Brukman
b0fce1668a Fix grammar
llvm-svn: 19854
2005-01-27 06:46:38 +00:00
Andrew Lenharth
4283bf1216 teach isel about comparison with constants and zero extending bits
llvm-svn: 19853
2005-01-27 03:49:45 +00:00
Jeff Cohen
1502dd24d9 Fix some Path bugs
llvm-svn: 19852
2005-01-27 03:49:03 +00:00
Andrew Lenharth
b539dd2c83 perhaps this will let me have calls again
llvm-svn: 19851
2005-01-27 01:22:48 +00:00
Andrew Lenharth
227bc0e21a minor bug fix
llvm-svn: 19850
2005-01-27 00:52:26 +00:00
Andrew Lenharth
ae8ce1856a minor bug fix
llvm-svn: 19849
2005-01-27 00:51:05 +00:00
Andrew Lenharth
11fc660a34 added instructions for fp to int to fp moves
llvm-svn: 19848
2005-01-26 23:56:48 +00:00
Andrew Lenharth
1f0b710fb6 initial fp support
llvm-svn: 19847
2005-01-26 21:54:09 +00:00
Andrew Lenharth
f9f01c190b hum, writing on one machine, testing on another...
llvm-svn: 19844
2005-01-26 02:53:56 +00:00
Andrew Lenharth
53ad9ac1db add some operations, fix others. should compile several more tests now
llvm-svn: 19843
2005-01-26 01:24:38 +00:00
Chris Lattner
ab92b92bc5 We can fold promoted and non-promoted loads into divs also!
llvm-svn: 19835
2005-01-25 20:35:10 +00:00
Chris Lattner
a9a0369879 Fold promoted loads into binary ops for FP, allowing us to generate m32 forms
of FP ops.

llvm-svn: 19834
2005-01-25 20:03:11 +00:00
Andrew Lenharth
4e10c17eeb problems with bools, and their work arounds
llvm-svn: 19833
2005-01-25 19:58:40 +00:00
Alkis Evlogimenos
eb6bfe9cee Add a dependency to the trace library so that it gets pulled in
automatically.

llvm-svn: 19828
2005-01-25 16:23:57 +00:00
Andrew Lenharth
3ae267eb3b more load choices, better add with imm
llvm-svn: 19821
2005-01-25 00:35:34 +00:00
Chris Lattner
f03b87704f Make -ds-aa more useful, allowing it to be updated as xforms hack on the program.
llvm-svn: 19818
2005-01-24 20:00:14 +00:00
Andrew Lenharth
3b44cfa26d Clean ups, and taught the instruction selector about immediate forms
llvm-svn: 19816
2005-01-24 19:44:07 +00:00
Andrew Lenharth
3c6e50e63b Alpha JIT prune
llvm-svn: 19815
2005-01-24 18:48:22 +00:00
Andrew Lenharth
ae874f0d85 include prune and JIT prune
llvm-svn: 19814
2005-01-24 18:45:41 +00:00
Andrew Lenharth
e3991f8256 Pruned includes
llvm-svn: 19813
2005-01-24 18:37:48 +00:00
Chris Lattner
88a4c43e67 Fix a spurious warning.
llvm-svn: 19799
2005-01-24 01:40:18 +00:00
Chris Lattner
6ff85c3152 Silence a warning.
llvm-svn: 19798
2005-01-23 23:20:06 +00:00
Chris Lattner
849899e193 Silence optimized warnings.
llvm-svn: 19797
2005-01-23 23:19:44 +00:00
Chris Lattner
94952e0947 Allow the FP stackifier to completely ignore functions that do not use FP at
all.  This should speed up the X86 backend fairly significantly on integer
codes.  Now if only we didn't have to compute livevar still... ;-)

llvm-svn: 19796
2005-01-23 23:13:59 +00:00
Chris Lattner
65fc8007cd Simplify/speedup the PEI by not having to scan for uses of the callee saved
registers.  This information is computed directly by the register allocator
now.

llvm-svn: 19795
2005-01-23 23:13:12 +00:00
Chris Lattner
556679b89d Update physregsused info.
llvm-svn: 19793
2005-01-23 22:55:45 +00:00
Chris Lattner
cc22be2981 Update this pass to set PhysRegsUsed info in MachineFunction.
llvm-svn: 19792
2005-01-23 22:51:56 +00:00
Chris Lattner
964297fc32 Update these register allocators to set the PhysRegUsed info in MachineFunction.
llvm-svn: 19791
2005-01-23 22:45:13 +00:00
Chris Lattner
6a6d5cf9eb Add support for the PhysRegsUsed array.
llvm-svn: 19789
2005-01-23 22:13:58 +00:00
Chris Lattner
c187b917f2 Speed this up a bit by making ModifiedRegs a vector<char> not vector<bool>
llvm-svn: 19787
2005-01-23 21:45:01 +00:00
Chris Lattner
ab2ab313d3 Get rid of a several dozen more and instructions in specint.
llvm-svn: 19786
2005-01-23 20:26:55 +00:00
Chris Lattner
ab1804175d Fix crash comparing empty file against nonempty file.
llvm-svn: 19782
2005-01-23 06:02:40 +00:00
Chris Lattner
b3a5fc3ec0 Adjust to changes in SelectionDAG interfaces
The first half of correct chain insertion for libcalls. This is not enough
to fix Fhourstones yet though.

llvm-svn: 19781
2005-01-23 04:42:50 +00:00
Chris Lattner
3165569ba9 Remove the 3 HACK HACK HACKs I put in before, fixing them properly with
the new TLI that is available.

Implement support for handling out of range shifts.  This allows us to
compile this code (a 64-bit rotate):

unsigned long long f3(unsigned long long x) {
  return (x << 32) | (x >> (64-32));
}

into this:

f3:
        mov %EDX, DWORD PTR [%ESP + 4]
        mov %EAX, DWORD PTR [%ESP + 8]
        ret

GCC produces this:

$ gcc t.c -masm=intel -O3 -S -o - -fomit-frame-pointer
..
f3:
        push    %ebx
        mov     %ebx, DWORD PTR [%esp+12]
        mov     %ecx, DWORD PTR [%esp+8]
        mov     %eax, %ebx
        mov     %edx, %ecx
        pop     %ebx
        ret

The Simple ISEL produces (eww gross):

f3:
        sub %ESP, 4
        mov DWORD PTR [%ESP], %ESI
        mov %EDX, DWORD PTR [%ESP + 8]
        mov %ECX, DWORD PTR [%ESP + 12]
        mov %EAX, 0
        mov %ESI, 0
        or %EAX, %ECX
        or %EDX, %ESI
        mov %ESI, DWORD PTR [%ESP]
        add %ESP, 4
        ret

llvm-svn: 19780
2005-01-23 04:39:44 +00:00
Chris Lattner
4c997d281c Adjust to changes in SelectionDAG interface.
llvm-svn: 19779
2005-01-23 04:36:26 +00:00
Chris Lattner
680bc75f7f Build Alpha by default.
llvm-svn: 19777
2005-01-23 04:34:46 +00:00
Reid Spencer
e48557f583 Fix alloca support for Cygwin. On cygwin its __alloca not __builtin_alloca
llvm-svn: 19776
2005-01-23 04:32:47 +00:00
Reid Spencer
5c7b6e83f0 Support Cygwin assembly generation. The cygwin version of Gnu ASsembler
doesn't support certain directives and symbols on cygwin are prefixed with
an underscore. This patch makes the necessary adjustments to the output.

llvm-svn: 19775
2005-01-23 03:52:14 +00:00
Chris Lattner
331670eb3e Make DiffFilesWithTolerance take sys::Path's instead of std::strings
Delete dead functions.

llvm-svn: 19771
2005-01-23 03:31:02 +00:00
Chris Lattner
ea7baf8d9f Fix a bug in previous checkin
llvm-svn: 19769
2005-01-23 03:19:13 +00:00
Chris Lattner
dcf389fb90 Add a new method, refactored out of fpcmp
llvm-svn: 19766
2005-01-23 03:13:43 +00:00
Andrew Lenharth
f5b9a8fe57 Let me introduce you to the early stages of the llvm backend for the alpha processor
llvm-svn: 19764
2005-01-22 23:41:55 +00:00
Chris Lattner
63ec3c402b Get this to work for 64-bit systems.
llvm-svn: 19763
2005-01-22 23:04:37 +00:00
Jeff Cohen
7311de2af2 Use binary mode for reading/writing bytecode files
llvm-svn: 19751
2005-01-22 17:36:17 +00:00
Jeff Cohen
66b5805c50 Fix destroyDirectory bug
llvm-svn: 19746
2005-01-22 16:28:33 +00:00
Chris Lattner
29d6389d78 Implicitly defined registers can clobber callee saved registers too!
This fixes the return-address-not-being-saved problem in the Alpha backend.

llvm-svn: 19741
2005-01-22 00:49:16 +00:00
Chris Lattner
97f35a7a07 More bugfixes for IA64 shifts.
llvm-svn: 19739
2005-01-22 00:33:03 +00:00
Chris Lattner
67deea9d05 Fix problems with non-x86 targets.
llvm-svn: 19738
2005-01-22 00:31:52 +00:00
Chris Lattner
42e239ed58 Add a nasty hack to fix Alpha/IA64 multiplies by a power of two.
llvm-svn: 19737
2005-01-22 00:20:42 +00:00
Chris Lattner
e724100870 Remove unneeded line.
llvm-svn: 19736
2005-01-21 23:43:12 +00:00
Chris Lattner
a974e215a5 test commit
llvm-svn: 19735
2005-01-21 23:38:56 +00:00
Chris Lattner
151c8e6390 Handle comparisons of gep instructions that have different typed indices
as long as they are the same size.

llvm-svn: 19734
2005-01-21 23:06:49 +00:00
Chris Lattner
b4cf4ffb04 Speed up folding operations into loads.
llvm-svn: 19733
2005-01-21 21:43:02 +00:00
Chris Lattner
fd4d7f71ae The ever-important vanity pass name :)
llvm-svn: 19731
2005-01-21 21:35:14 +00:00
Chris Lattner
24bf1ca350 If the interpreter tries to execute an external function, kill it. Of course
since we are dirty, special case __main.  This should fix the infinite loop
horrible stuff that happens on linux-alpha when configuring llvm-gcc.  It
might also help cygwin, who knows??

llvm-svn: 19729
2005-01-21 19:59:37 +00:00
Chris Lattner
5f2fbeaa69 Fix a FIXME: realize that argument stores are all independent (don't alias)
llvm-svn: 19728
2005-01-21 19:46:38 +00:00
Chris Lattner
392ddf430b Unary token factor nodes are unneeded.
llvm-svn: 19727
2005-01-21 18:01:22 +00:00
Chris Lattner
07c35617d5 Refactor libcall code a bit. Initial implementation of expanding int -> FP
operations for 64-bit integers.

llvm-svn: 19724
2005-01-21 06:05:23 +00:00
Chris Lattner
6258ec2e1d Simplify the shift-expansion code.
llvm-svn: 19721
2005-01-20 20:29:23 +00:00
Chris Lattner
febeb380ae Implement ADD_PARTS/SUB_PARTS so that 64-bit integer add/sub work. This
fixes most of the remaining llc-beta failures.

llvm-svn: 19716
2005-01-20 18:53:00 +00:00
Chris Lattner
c95c7c90c9 Expand add/sub into ADD_PARTS/SUB_PARTS instead of a non-existant libcall.
llvm-svn: 19715
2005-01-20 18:52:28 +00:00
Chris Lattner
4086a7a803 implement add_parts/sub_parts.
llvm-svn: 19714
2005-01-20 18:50:55 +00:00
Chris Lattner
e7ce5d0e4c Add missing entry.
llvm-svn: 19712
2005-01-20 17:32:28 +00:00
Chris Lattner
8b0a2a3251 Fix a crash compiling 134.perl.
llvm-svn: 19711
2005-01-20 16:50:16 +00:00
Chris Lattner
e5212a16a2 Support targets that do not use i8 shift amounts.
llvm-svn: 19707
2005-01-19 22:31:21 +00:00
Chris Lattner
5a9660aa71 Add two optimizations. The first folds (X+Y)-X -> Y
The second folds operations into selects, e.g. (select C, (X+Y), (Y+Z))
-> (Y+(select C, X, Z)

This occurs a few times across spec, e.g.

         select    add/sub
mesa:    83        0
povray:  5         2
gcc      4         2
parser   0         22
perlbmk  13        30
twolf    0         3

llvm-svn: 19706
2005-01-19 21:50:18 +00:00
Chris Lattner
0e7435bc5b Add an assertion that would have made more sense to duraid
llvm-svn: 19704
2005-01-19 21:32:07 +00:00
Chris Lattner
c662697319 Add support for targets that pass args in registers to calls.
llvm-svn: 19703
2005-01-19 20:24:35 +00:00
Chris Lattner
277ac2be70 Fold single use token factor nodes into other token factor nodes.
llvm-svn: 19701
2005-01-19 19:10:54 +00:00
Chris Lattner
85e0771f79 Realize the individual pieces of an expanded copytoreg/store/load are
independent of each other.

llvm-svn: 19700
2005-01-19 18:02:17 +00:00
Chris Lattner
027c97e93e Know some identities about tokenfactor nodes.
llvm-svn: 19699
2005-01-19 18:01:40 +00:00
Chris Lattner
7114e8a527 Know some simple identities. This improves codegen for (1LL << N).
llvm-svn: 19698
2005-01-19 17:29:49 +00:00
Chris Lattner
6534e1ede3 Fix a problem where were were literally selecting for INCREASED register
pressure, not decreases register pressure.  Fix problem where we accidentally
swapped the operands of SHLD, which caused fourinarow to fail.  This fixes
fourinarow.

llvm-svn: 19697
2005-01-19 17:24:34 +00:00
Chris Lattner
e97ed92617 Just in case, handle something that is both a use and a def.
llvm-svn: 19696
2005-01-19 17:11:51 +00:00
Chris Lattner
2cb11bd2b9 When an instruction moves, make sure to update the VarInfo::Kills list as
well as all of teh other stuff in livevar. This fixes the compiler crash
on fourinarow last night.

llvm-svn: 19695
2005-01-19 17:09:15 +00:00
Chris Lattner
b75589131d When commuting these instructions, make sure to actually swap the operands too.
llvm-svn: 19694
2005-01-19 16:55:52 +00:00
Chris Lattner
302ea8908d Fix 'raise' to work with packed types. Patch by Morten Ofstad.
llvm-svn: 19693
2005-01-19 16:16:35 +00:00
Chris Lattner
fde1a5688b Implement Regression/CodeGen/X86/rotate.ll: emit rotate instructions (which
typically cost 1 cycle) instead of shld/shrd instruction (which are typically
6 or more cycles).  This also saves code space.

For example, instead of emitting:

rotr:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %CL, BYTE PTR [%ESP + 8]
        shrd %EAX, %EAX, %CL
        ret
rotli:
        mov %EAX, DWORD PTR [%ESP + 4]
        shrd %EAX, %EAX, 27
        ret

Emit:

rotr32:
        mov %CL, BYTE PTR [%ESP + 8]
        mov %EAX, DWORD PTR [%ESP + 4]
        ror %EAX, %CL
        ret
rotli32:
        mov %EAX, DWORD PTR [%ESP + 4]
        ror %EAX, 27
        ret

We also emit byte rotate instructions which do not have a sh[lr]d counterpart
at all.

llvm-svn: 19692
2005-01-19 08:07:05 +00:00
Chris Lattner
34757ff939 Add rotate instructions.
llvm-svn: 19690
2005-01-19 07:50:03 +00:00
Chris Lattner
e539ce8223 Match 16-bit shld/shrd instructions as well, implementing shift-double.llx:test5
llvm-svn: 19689
2005-01-19 07:37:26 +00:00
Chris Lattner
9d5ee289d7 Improve coverage of the X86 instruction set by adding 16-bit shift doubles.
llvm-svn: 19687
2005-01-19 07:31:24 +00:00
Chris Lattner
c03f360215 Teach the code generator that shrd/shld is commutable if it has an immediate.
This allows us to generate this:

foo:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %EDX, DWORD PTR [%ESP + 8]
        shld %EDX, %EDX, 2
        shl %EAX, 2
        ret

instead of this:

foo:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %ECX, DWORD PTR [%ESP + 8]
        mov %EDX, %EAX
        shrd %EDX, %ECX, 30
        shl %EAX, 2
        ret

Note the magically transmogrifying immediate.

llvm-svn: 19686
2005-01-19 07:11:01 +00:00
Chris Lattner
408325ffdf Use the TargetInstrInfo::commuteInstruction method to commute instructions
instead of doing it manually.

llvm-svn: 19685
2005-01-19 07:08:42 +00:00
Chris Lattner
33efebcdc8 Finegrainify namespacification
Add default impl of commuteInstruction
Add notes about ugly V9 code.

llvm-svn: 19684
2005-01-19 06:53:34 +00:00
Chris Lattner
575e912fcf Codegen long >> 2 to this:
foo:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %EDX, DWORD PTR [%ESP + 8]
        shrd %EAX, %EDX, 2
        sar %EDX, 2
        ret

instead of this:

test1:
        mov %ECX, DWORD PTR [%ESP + 4]
        shr %ECX, 2
        mov %EDX, DWORD PTR [%ESP + 8]
        mov %EAX, %EDX
        shl %EAX, 30
        or %EAX, %ECX
        sar %EDX, 2
        ret

and long << 2 to this:

foo:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %ECX, DWORD PTR [%ESP + 8]
***     mov %EDX, %EAX
        shrd %EDX, %ECX, 30
        shl %EAX, 2
        ret

instead of this:

foo:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %ECX, %EAX
        shr %ECX, 30
        mov %EDX, DWORD PTR [%ESP + 8]
        shl %EDX, 2
        or %EDX, %ECX
        shl %EAX, 2
        ret

The extra copy (marked ***) can be eliminated when I teach the code generator
that shrd32rri8 is really commutative.

llvm-svn: 19681
2005-01-19 06:18:43 +00:00
Chris Lattner
743a36c818 Implement a way of expanding shifts. This applies to targets that offer
select operations or to shifts that are by a constant.  This automatically
implements (with no special code) all of the special cases for shift by 32,
shift by < 32 and shift by > 32.

llvm-svn: 19679
2005-01-19 04:19:40 +00:00
Chris Lattner
419a5d213b X86 shifts mask the amount.
llvm-svn: 19678
2005-01-19 03:36:30 +00:00
Chris Lattner
fbd1f8e4fd Add a hook to find out how the target handles shift amounts that are out of
range.  Either they are undefined (the default), they mask the shift amount
to the size of the register (X86, Alpha, etc), or they extend the shift (PPC).

This defaults to undefined, which is conservatively correct.

llvm-svn: 19677
2005-01-19 03:36:14 +00:00
Chris Lattner
0df1935505 Zero is cheaper than sign extend.
llvm-svn: 19675
2005-01-18 21:57:59 +00:00
Chris Lattner
6dec8cb829 Code to handle FP_EXTEND is dead now. X86 doesn't support any data types to
FP_EXTEND from!

llvm-svn: 19674
2005-01-18 20:05:56 +00:00
Chris Lattner
798e9c85d6 Remove more dead code.
llvm-svn: 19673
2005-01-18 19:50:08 +00:00
Chris Lattner
401814508f The selection dag code handles the promotions from F32 to F64 for us, so we
don't need to even think about F32 in the X86 code anymore.

llvm-svn: 19672
2005-01-18 19:46:54 +00:00
Chris Lattner
4360871e16 Fix some fixmes (promoting bools for select and brcond), fix promotion
of zero and sign extends.

llvm-svn: 19671
2005-01-18 19:27:06 +00:00
Chris Lattner
eea485de1f Keep track of the retval type as well.
llvm-svn: 19670
2005-01-18 19:26:36 +00:00
Chris Lattner
ff086f3016 Teach legalize to promote copy(from|to)reg, instead of making the isel pass
do it.  This results in better code on X86 for floats (because if strict
precision is not required, we can elide some more expensive double -> float
conversions like the old isel did), and allows other targets to emit
CopyFromRegs that are not legal for arguments.

llvm-svn: 19668
2005-01-18 17:54:55 +00:00
Chris Lattner
dc09e52b3e Fix 124.m88ksim.
llvm-svn: 19667
2005-01-18 17:35:28 +00:00
Chris Lattner
a04b1ee7a8 Do not emit loads multiple times, potentially in the wrong places.
llvm-svn: 19661
2005-01-18 04:18:32 +00:00
Tanya Lattner
d3459278f2 Minor changes.
llvm-svn: 19660
2005-01-18 04:15:41 +00:00
Chris Lattner
722ddeb86e Eliminate bad assertions.
llvm-svn: 19659
2005-01-18 04:00:54 +00:00
Chris Lattner
8f3a8d96e2 * Eliminate the TokenSet and just use the ExprMap for both tokens and values.
* Insert some really pedantic assertions that will notice when we emit the
  same loads more than one time, exposing bugs.  This turns a miscompilation in
  bzip2 into a compile-fail.  yaay.

llvm-svn: 19658
2005-01-18 03:51:59 +00:00
Chris Lattner
891aa537f7 Teach legalize to promote SetCC results.
llvm-svn: 19657
2005-01-18 02:59:52 +00:00
Chris Lattner
95307053ec Allow setcc operations to have nonbool types.
llvm-svn: 19656
2005-01-18 02:52:03 +00:00
Chris Lattner
b3edb09ede Rely on the code in MatchAddress to do this work. Otherwise we fail to
match (X+Y)+(Z << 1), because we match the X+Y first, consuming the index
register, then there is no place to put the Z.

llvm-svn: 19652
2005-01-18 02:25:52 +00:00
Chris Lattner
906541da95 Fix the completely broken FP constant folds for setcc's.
llvm-svn: 19651
2005-01-18 02:11:55 +00:00
Chris Lattner
ce2e0125dc Fix a problem where probing for addressing modes caused expressions to be
emitted too early.  In particular, this fixes
Regression/CodeGen/X86/regpressure.ll:regpressure3.

This also improves the 2nd basic block in 164.gzip:flush_block, which went from

.LBBflush_block_1:      # loopentry.1.i
        movzx %EAX, WORD PTR [dyn_ltree + 20]
        movzx %ECX, WORD PTR [dyn_ltree + 16]
        mov DWORD PTR [%ESP + 32], %ECX
        movzx %ECX, WORD PTR [dyn_ltree + 12]
        movzx %EDX, WORD PTR [dyn_ltree + 8]
        movzx %EBX, WORD PTR [dyn_ltree + 4]
        mov DWORD PTR [%ESP + 36], %EBX
        movzx %EBX, WORD PTR [dyn_ltree]
        add DWORD PTR [%ESP + 36], %EBX
        add %EDX, DWORD PTR [%ESP + 36]
        add %ECX, %EDX
        add DWORD PTR [%ESP + 32], %ECX
        add %EAX, DWORD PTR [%ESP + 32]
        movzx %ECX, WORD PTR [dyn_ltree + 24]
        add %EAX, %ECX
        mov %ECX, 0
        mov %EDX, %ECX

to

.LBBflush_block_1:      # loopentry.1.i
        movzx %EAX, WORD PTR [dyn_ltree]
        movzx %ECX, WORD PTR [dyn_ltree + 4]
        add %ECX, %EAX
        movzx %EAX, WORD PTR [dyn_ltree + 8]
        add %EAX, %ECX
        movzx %ECX, WORD PTR [dyn_ltree + 12]
        add %ECX, %EAX
        movzx %EAX, WORD PTR [dyn_ltree + 16]
        add %EAX, %ECX
        movzx %ECX, WORD PTR [dyn_ltree + 20]
        add %ECX, %EAX
        movzx %EAX, WORD PTR [dyn_ltree + 24]
        add %ECX, %EAX
        mov %EAX, 0
        mov %EDX, %EAX

... which results in less spilling in the function.

This change alone speeds up 164.gzip from 37.23s to 36.24s on apoc.  The
default isel takes 37.31s.

llvm-svn: 19650
2005-01-18 01:06:26 +00:00
Chris Lattner
a78f9ced61 Fix indentation.
llvm-svn: 19649
2005-01-17 23:25:45 +00:00
Chris Lattner
dff1e3e86f Don't bother using max here.
llvm-svn: 19647
2005-01-17 23:02:13 +00:00
Chris Lattner
2d86b43318 Do not give token factor nodes outrageous weights
llvm-svn: 19645
2005-01-17 22:56:09 +00:00
Chris Lattner
c0aca0d13c Non-volatile loads can be freely reordered against each other. This fixes
X86/reg-pressure.ll again, and allows us to do nice things in other cases.
For example, we now codegen this sort of thing:

int %loadload(int *%X, int* %Y) {
  %Z = load int* %Y
  %Y = load int* %X      ;; load between %Z and store
  %Q = add int %Z, 1
  store int %Q, int* %Y
  ret int %Y
}

Into this:

loadload:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %EAX, DWORD PTR [%EAX]
        mov %ECX, DWORD PTR [%ESP + 8]
        inc DWORD PTR [%ECX]
        ret

where we weren't able to form the 'inc [mem]' before.  This also lets the
instruction selector emit loads in any order it wants to, which can be good
for register pressure as well.

llvm-svn: 19644
2005-01-17 22:19:26 +00:00
Chris Lattner
f2878ce8ba Two changes:
1. Fold  [mem] += (1|-1) into inc [mem]/dec [mem] to save some icache space.
 2. Do not let token factor nodes prevent forming '[mem] op= val' folds.

llvm-svn: 19643
2005-01-17 22:10:42 +00:00
Chris Lattner
49291c4d96 Don't call SelectionDAG.getRoot() directly, go through a forwarding method.
llvm-svn: 19642
2005-01-17 19:43:36 +00:00
Chris Lattner
40c0fca632 Refactor load/op/store folding into it's own method, no functionality changes.
llvm-svn: 19641
2005-01-17 19:25:26 +00:00
Chris Lattner
88bbcfc893 Implement a target independent optimization to codegen arguments only into
the basic block that uses them if possible.  This is a big win on X86, as it
lets us fold the argument loads into instructions and reduce register pressure
(by not loading all of the arguments in the entry block).

For this (contrived to show the optimization) testcase:

int %argtest(int %A, int %B) {
        %X = sub int 12345, %A
        br label %L
L:
        %Y = add int %X, %B
        ret int %Y
}

we used to produce:

argtest:
        mov %ECX, DWORD PTR [%ESP + 4]
        mov %EAX, 12345
        sub %EAX, %ECX
        mov %EDX, DWORD PTR [%ESP + 8]
.LBBargtest_1:  # L
        add %EAX, %EDX
        ret


now we produce:

argtest:
        mov %EAX, 12345
        sub %EAX, DWORD PTR [%ESP + 4]
.LBBargtest_1:  # L
        add %EAX, DWORD PTR [%ESP + 8]
        ret

This also fixes the FIXME in the code.

BTW, this occurs in real code.  164.gzip shrinks from 8623 to 8608 lines of
.s file.  The stack frame in huft_build shrinks from 1644->1628 bytes,
inflate_codes shrinks from 116->108 bytes, and inflate_block from 2620->2612,
due to fewer spills.

Take that alkis. :-)

llvm-svn: 19639
2005-01-17 17:55:19 +00:00
Chris Lattner
2348abc421 Fix a major regression last night that prevented us from producing [mem] op= reg
operations.

The body of the if is less indented but unmodified in this patch.

llvm-svn: 19638
2005-01-17 17:49:14 +00:00
Chris Lattner
49a1f3a109 Refactor code into a new method.
llvm-svn: 19635
2005-01-17 17:15:02 +00:00
Chris Lattner
adb669ab1f Codegen this:
int %foo(int %X) {
        %T = add int %X, 13
        %S = mul int %T, 3
        ret int %S
}

as this:

        mov %ECX, DWORD PTR [%ESP + 4]
        lea %EAX, DWORD PTR [%ECX + 2*%ECX + 39]
        ret

instead of this:

        mov %ECX, DWORD PTR [%ESP + 4]
        mov %EAX, %ECX
        add %EAX, 13
        imul %EAX, %EAX, 3
        ret

llvm-svn: 19633
2005-01-17 06:48:02 +00:00
Tanya Lattner
5a10531cf8 Added tmp instructions to preserve ssa.
llvm-svn: 19632
2005-01-17 06:47:26 +00:00
Chris Lattner
51590b615c Fix test/Regression/CodeGen/X86/2005-01-17-CycleInDAG.ll and 132.ijpeg.
Do not fold a load into an operation if it will induce a cycle in the DAG.

Repeat after me: dAg.

llvm-svn: 19631
2005-01-17 06:26:58 +00:00
Chris Lattner
3402945d52 Delete PHI nodes that are not dead but are locked in a cycle of single
useness.

llvm-svn: 19629
2005-01-17 05:10:15 +00:00
Chris Lattner
de6b1ca556 Move code out of indentation one level to make it easier to read.
Disable the xform for < > cases.  It turns out that the following is being
miscompiled:

bool %test(sbyte %S) {
        %T = cast sbyte %S to uint
        %V = setgt uint %T, 255
        ret bool %V
}

llvm-svn: 19628
2005-01-17 03:20:02 +00:00
Chris Lattner
f1e85bec5a Do not fold a load into a comparison that is used by more than one place.
The comparison will probably be folded, so this is not ok to do.
This fixed 197.parser.

llvm-svn: 19624
2005-01-17 01:34:14 +00:00
Chris Lattner
1b8c8fe020 Do not codegen 'xor bool, true' as 'not reg'. not reg inverts the upper bits
of the bytereg.  This fixes yacr2, 300.twolf and probably others.

llvm-svn: 19622
2005-01-17 00:23:16 +00:00
Chris Lattner
46dac4394c Set up the shift and setcc types.
If we emit a load because we followed a token chain to get to it, try to
fold it into its single user if possible.

llvm-svn: 19620
2005-01-17 00:00:33 +00:00
Chris Lattner
4c88cc95ee Shift and setcc types default to the pointer type.
llvm-svn: 19619
2005-01-16 23:59:48 +00:00
Chris Lattner
ec55e3e529 Implement legalize of call nodes.
llvm-svn: 19617
2005-01-16 19:46:48 +00:00
Tanya Lattner
fea188af7e Added paramters to a few functions in order to allow me to change the functions to preserve SSA
llvm-svn: 19615
2005-01-16 08:51:10 +00:00
Chris Lattner
9ffc59287e * Adjust to changes in TargetLowering interfaces.
* Remove custom promotion for bool and byte select ops.  Legalize now
  promotes them for us.
* Allow folding ConstantPoolIndexes into EXTLOAD's, useful for float immediates.
* Declare which operations are not supported better.
* Add some hacky code for TRUNCSTORE to pretend that we have truncstore
  for i16 types.  This is useful for testing promotion code because I can
  just remove 16-bit registers all together and verify that programs work.

llvm-svn: 19614
2005-01-16 07:34:08 +00:00
Chris Lattner
0eca430af1 Revamp supported ops. Instead of just being supported or not, we now keep
track of how to deal with it, and provide the target with a hook that they
can use to legalize arbitrary operations in arbitrary ways.

Implement custom lowering for a couple of ops, implement promotion for select
operations (which x86 needs).

llvm-svn: 19613
2005-01-16 07:29:19 +00:00
Chris Lattner
835a5efef3 add method stub
llvm-svn: 19612
2005-01-16 07:28:41 +00:00
Chris Lattner
907534af24 Don't mash stuff together.
llvm-svn: 19611
2005-01-16 07:28:31 +00:00
Chris Lattner
b49d2a7b0f Use enums, move virtual dtor out of line.
llvm-svn: 19610
2005-01-16 07:28:11 +00:00
Chris Lattner
0f4f239899 Implement some more missing promotions.
llvm-svn: 19606
2005-01-16 05:06:12 +00:00
Chris Lattner
e88e660817 Fix bugpoint
llvm-svn: 19605
2005-01-16 04:23:22 +00:00
Chris Lattner
be2a427f51 cycles_t -> CycleCount_t
llvm-svn: 19604
2005-01-16 04:20:30 +00:00
Chris Lattner
742b77f9af Clarify assertion.
llvm-svn: 19597
2005-01-16 02:23:34 +00:00
Chris Lattner
4517b8af97 Add assertions.
llvm-svn: 19596
2005-01-16 02:23:22 +00:00
Chris Lattner
9f8589f4b3 Add support for promoted registers being live across blocks.
llvm-svn: 19595
2005-01-16 02:23:07 +00:00
Reid Spencer
afa1cb9e11 Rename BUILD_* to PROJ_*
llvm-svn: 19592
2005-01-16 02:21:29 +00:00
Tanya Lattner
66cf1a6f82 Fixed a couple of instructions that broke SSA.
llvm-svn: 19587
2005-01-16 02:14:17 +00:00
Chris Lattner
605b9a23a2 Improve compatiblity with HPUX on Itanium, patch by Duraid Madina
llvm-svn: 19586
2005-01-16 01:31:31 +00:00
Chris Lattner
06c297f8ca Set up identity transforms.
llvm-svn: 19584
2005-01-16 01:20:18 +00:00
Chris Lattner
01e2ce8a4c Move some information into the TargetLowering object.
llvm-svn: 19583
2005-01-16 01:11:45 +00:00
Chris Lattner
9762070e50 Use the new TLI method to get this.
llvm-svn: 19582
2005-01-16 01:11:19 +00:00
Chris Lattner
1d0e1ffe02 Move some information out of LegalizeDAG into the generic Target interface.
llvm-svn: 19581
2005-01-16 01:10:58 +00:00
Chris Lattner
0777f84d53 legalize a bunch of operations that I missed.
llvm-svn: 19580
2005-01-16 00:38:00 +00:00
Chris Lattner
1de18d422e Add support for targets that require promotions.
llvm-svn: 19579
2005-01-16 00:37:38 +00:00
Chris Lattner
8c4c81d6b3 Fix some serious bugs in promotion.
llvm-svn: 19578
2005-01-16 00:17:42 +00:00
Chris Lattner
9785def2cd Eliminate unneeded extensions.
llvm-svn: 19577
2005-01-16 00:17:20 +00:00
Chris Lattner
df02c93d90 Implement promotion of a whole bunch more operators. I think that this is
basically everything.

llvm-svn: 19576
2005-01-15 22:16:26 +00:00
Chris Lattner
f3fd0c6a93 Print extra type for nodes with extra type info.
llvm-svn: 19575
2005-01-15 21:11:37 +00:00
Chris Lattner
1ab9009270 Add support for legalizing FP_ROUND_INREG, SIGN_EXTEND_INREG, and
ZERO_EXTEND_INREG for targets that don't support them.

llvm-svn: 19573
2005-01-15 07:15:18 +00:00
Chris Lattner
191ac9c589 Common code factored out.
llvm-svn: 19572
2005-01-15 07:14:32 +00:00
Chris Lattner
3b20db54f3 implement these methods.
llvm-svn: 19571
2005-01-15 06:52:40 +00:00
Chris Lattner
fdd07b4092 Add support for promoting ADD/MUL.
Add support for new SIGN_EXTEND_INREG, ZERO_EXTEND_INREG, and FP_ROUND_INREG operators.
Realize that if we do any promotions, we need to iterate SelectionDAG
construction.

llvm-svn: 19569
2005-01-15 06:18:18 +00:00
Chris Lattner
2f65e8798f Add new SIGN_EXTEND_INREG, ZERO_EXTEND_INREG, and FP_ROUND_INREG operators.
llvm-svn: 19568
2005-01-15 06:17:04 +00:00
Chris Lattner
98611ce291 Add a new target-independent code generator flag.
llvm-svn: 19567
2005-01-15 06:00:32 +00:00
Chris Lattner
f3d950e816 Add support for truncstore and *extload.
llvm-svn: 19566
2005-01-15 05:22:24 +00:00
Chris Lattner
94b8a3e50c Add intitial support for promoting some operators.
llvm-svn: 19565
2005-01-15 05:21:40 +00:00
Reid Spencer
ad96095c97 We don't distribute the operating system specific directories any more.
llvm-svn: 19563
2005-01-14 22:43:01 +00:00
Chris Lattner
2dfbc4fddd Adjust to CopyFromReg changes, implement deletion of truncating/extending
stores/loads.

llvm-svn: 19562
2005-01-14 22:38:01 +00:00
Chris Lattner
27c91fac94 Adjust to CopyFromREg changes.
llvm-svn: 19561
2005-01-14 22:37:41 +00:00
Chris Lattner
0974002024 Start implementing truncating stores and extending loads.
llvm-svn: 19559
2005-01-14 22:08:15 +00:00
Chris Lattner
c032990335 Fix Regression/CodeGen/PowerPC/2005-01-14-UndefLong.ll
llvm-svn: 19557
2005-01-14 20:22:02 +00:00
Chris Lattner
b0b49268c4 Fix: Regression/CodeGen/PowerPC/2005-01-14-SetSelectCrash.ll
llvm-svn: 19555
2005-01-14 19:31:00 +00:00
Chris Lattner
708ff662ba Fix some bugs in an xform added yesterday. This fixes Prolangs-C/allroots.
llvm-svn: 19553
2005-01-14 17:35:12 +00:00
Chris Lattner
13fd87be57 Fix a compile crash on spiff
llvm-svn: 19552
2005-01-14 17:17:59 +00:00
Chris Lattner
2087f3c8e9 Improve compatibility with acc
llvm-svn: 19549
2005-01-14 15:54:24 +00:00
Chris Lattner
1e5620dfe1 Make this compatible with the HP/intel compiler. Fix by Duraid, thanks!
llvm-svn: 19548
2005-01-14 15:53:26 +00:00
Jeff Cohen
7dfbb46f7f Fix and improve win32 path validation.
llvm-svn: 19545
2005-01-14 04:09:39 +00:00
Reid Spencer
4e90250e81 Make asctime_r work for HP/UX.
llvm-svn: 19544
2005-01-14 00:50:50 +00:00
Chris Lattner
6b519e3314 if two gep comparisons only differ by one index, compare that index directly.
This allows us to better optimize begin() -> end() comparisons in common cases.

llvm-svn: 19542
2005-01-14 00:20:05 +00:00
Chris Lattner
283b7d9809 Do not overrun iterators. This fixes a 176.gcc crash
llvm-svn: 19541
2005-01-13 23:26:48 +00:00
Chris Lattner
b3dfd0aecd Turn select C, (X+Y), (X-Y) --> (X+(select C, Y, (-Y))). This occurs in
the 'sim' program and probably elsewhere.  In sim, it comes up for cases
like this:

#define round(x) ((x)>0.0 ? (x)+0.5 : (x)-0.5)
double G;
void T(double X) { G = round(X); }

(it uses the round macro a lot).  This changes the LLVM code from:

        %tmp.1 = setgt double %X, 0.000000e+00          ; <bool> [#uses=1]
        %tmp.4 = add double %X, 5.000000e-01            ; <double> [#uses=1]
        %tmp.6 = sub double %X, 5.000000e-01            ; <double> [#uses=1]
        %mem_tmp.0 = select bool %tmp.1, double %tmp.4, double %tmp.6
        store double %mem_tmp.0, double* %G

to:

        %tmp.1 = setgt double %X, 0.000000e+00          ; <bool> [#uses=1]
        %mem_tmp.0.p = select bool %tmp.1, double 5.000000e-01, double -5.000000e-01
        %mem_tmp.0 = add double %mem_tmp.0.p, %X
        store double %mem_tmp.0, double* %G
        ret void

llvm-svn: 19537
2005-01-13 22:52:24 +00:00
Chris Lattner
e59c6d1cbe Implement an optimization for == and != comparisons like this:
_Bool test2(int X, int Y) {
  return &arr[X][Y] == arr;
}

instead of generating this:

bool %test2(int %X, int %Y) {
        %tmp.3.idx = mul int %X, 160            ; <int> [#uses=1]
        %tmp.3.idx1 = shl int %Y, ubyte 2               ; <int> [#uses=1]
        %tmp.3.offs2 = sub int 0, %tmp.3.idx            ; <int> [#uses=1]
        %tmp.7 = seteq int %tmp.3.idx1, %tmp.3.offs2            ; <bool> [#uses=1]
        ret bool %tmp.7
}


generate this:

bool %test2(int %X, int %Y) {
        seteq int %X, 0         ; <bool>:0 [#uses=1]
        seteq int %Y, 0         ; <bool>:1 [#uses=1]
        %tmp.7 = and bool %0, %1                ; <bool> [#uses=1]
        ret bool %tmp.7
}

This idiom occurs in C++ programs when iterating from begin() to end(),
in a vector or array.  For example, we now compile this:

void test(int X, int Y) {
  for (int *i = arr; i != arr+100; ++i)
    foo(*i);
}

to this:

no_exit:                ; preds = %entry, %no_exit
	...
        %exitcond = seteq uint %indvar.next, 100                ; <bool> [#uses=1]
        br bool %exitcond, label %return, label %no_exit



instead of this:

no_exit:                ; preds = %entry, %no_exit
	...
        %inc5 = getelementptr [100 x [40 x int]]* %arr, int 0, int 0, int %inc.rec              ; <int*> [#uses=1]
        %tmp.8 = seteq int* %inc5, getelementptr ([100 x [40 x int]]* %arr, int 0, int 100, int 0)              ; <bool> [#uses=1]
        %indvar.next = add uint %indvar, 1              ; <uint> [#uses=1]
        br bool %tmp.8, label %return, label %no_exit

llvm-svn: 19536
2005-01-13 22:25:21 +00:00
Chris Lattner
7a8788c9ac Add new ImplicitDef node, rename CopyRegSDNode class to RegSDNode.
llvm-svn: 19535
2005-01-13 20:50:02 +00:00
Chris Lattner
ee469241c3 Fix some bugs in code I didn't mean to check in.
llvm-svn: 19534
2005-01-13 20:40:58 +00:00
Chris Lattner
aebad4db9a Fix a crash compiling 129.compress
llvm-svn: 19533
2005-01-13 20:14:25 +00:00
Chris Lattner
fce6a5439d Codegen factor nodes more intelligently according to perceived register pressure.
llvm-svn: 19532
2005-01-13 19:56:00 +00:00
Chris Lattner
9cc534f2dc Don't forget the existing root.
llvm-svn: 19531
2005-01-13 19:53:14 +00:00
Chris Lattner
cb4359465a Initial trivial (but stupid) codegen for this node.
llvm-svn: 19529
2005-01-13 18:01:36 +00:00
Chris Lattner
160fdb384b Codegen independent ops as being independent.
llvm-svn: 19528
2005-01-13 17:59:43 +00:00
Chris Lattner
37a5de6eb0 Legalize new node, add assertion.
llvm-svn: 19527
2005-01-13 17:59:25 +00:00
Chris Lattner
86b19c5605 Print new node.
llvm-svn: 19526
2005-01-13 17:59:10 +00:00
Chris Lattner
9a70166615 Add some really pedantic assertions to the load folding code. Fix a bunch
of cases where we accidentally emitted a load folded once and unfolded
elsewhere.

llvm-svn: 19522
2005-01-13 05:53:16 +00:00
Chris Lattner
93cb0148f8 Do not fold (zero_ext (sign_ext V)) -> (sign_ext V), they are not the same.
This fixes llvm-test/SingleSource/Regression/C/casts.c

llvm-svn: 19519
2005-01-12 18:51:15 +00:00
Chris Lattner
2ab70aafe0 We can only fold a load into an op if there is exactly one use of the value.
Checking to see if the load has two uses is not equivalent, as the chain
value may have zero uses.

llvm-svn: 19518
2005-01-12 18:38:26 +00:00
Chris Lattner
e97b0e1358 New method
llvm-svn: 19517
2005-01-12 18:37:47 +00:00
Chris Lattner
1b3b24f116 Fix sign extend to long. When coming from sbyte, we used to generate:
movsbl 4(%esp), %eax
        movl %eax, %edx
        sarl $7, %edx

Now we generate:

        movsbl 4(%esp), %eax
        movl %eax, %edx
        sarl $31, %edx

Which is right.

llvm-svn: 19515
2005-01-12 18:19:52 +00:00
Chris Lattner
4b03f0f99e Try both ways to fold an add together. This allows us to generate this code
imul %EAX, %EAX, 400
        add %ECX, %EAX
        add %ESI, DWORD PTR [%ECX + 4*%EDX]
        inc %EDX
        cmp %EDX, 100

instead of this:

        imul %EAX, %EAX, 400
        add %ECX, %EAX
        mov %EAX, %EDX
        shl %EAX, 2
        add %ECX, %EAX
        add %ESI, DWORD PTR [%ECX]
        inc %EDX
        cmp %EDX, 100

llvm-svn: 19513
2005-01-12 18:08:53 +00:00
Reid Spencer
c8c50250a1 Shut up warnings with GCC 3.4.3 about uninitialized variables.
llvm-svn: 19512
2005-01-12 14:53:45 +00:00
Chris Lattner
61c572eb7f Fix a major miscompilation where we were overwriting the scale reg.
llvm-svn: 19511
2005-01-12 07:33:20 +00:00
Chris Lattner
5816f1a302 Do not use the type of the RHS constant to determine the type of the operation.
This fails for shifts because the constant is always 8 bits.

llvm-svn: 19508
2005-01-12 05:22:07 +00:00
Chris Lattner
89d6b21ae6 Do not lose the offset from teh global when peephole optimizing instructions.
This fixes FreeBench/pcompress

llvm-svn: 19507
2005-01-12 05:17:28 +00:00
Chris Lattner
c9b64b9749 Silence VC++ warnings.
llvm-svn: 19506
2005-01-12 04:51:37 +00:00
Jeff Cohen
614a5ec22a Fix C++ more compilatiom errors
llvm-svn: 19504
2005-01-12 04:29:05 +00:00
Chris Lattner
5ef92f3a40 Fix a compile error with VC++, which things that static const arrays need
to be dynamically initialized. :(

llvm-svn: 19503
2005-01-12 04:23:22 +00:00
Chris Lattner
627c64e5e5 Fix a bug that caused us to crash on povray. We weren't emitting an FP_REG_KILL into a block that had a successor with a FP PHI node.
llvm-svn: 19502
2005-01-12 04:21:28 +00:00
Chris Lattner
a5f0ba59a0 Print a load of a null pointer (in intel mode) like this:
mov %AX, WORD PTR [0]

instead of like this:

        mov %AX, WORD PTR []

llvm-svn: 19501
2005-01-12 04:07:11 +00:00
Chris Lattner
360988bae2 Print a load of a null pointer like this:
movw 0, %ax

instead of like this:

        movw , %ax

llvm-svn: 19500
2005-01-12 04:05:19 +00:00
Chris Lattner
3c85c67c97 Fix a crash compiling povray on UINT_TO_FP from i16.
llvm-svn: 19499
2005-01-12 04:00:00 +00:00
Chris Lattner
e7945a2e2e Add an option to view the selection dags as they are generated.
llvm-svn: 19498
2005-01-12 03:41:21 +00:00
Chris Lattner
4e72a2a000 There are no [mem] op= reg instructions for FP, so remove their entries.
llvm-svn: 19496
2005-01-12 03:16:09 +00:00
Chris Lattner
00cb0ace9b Fix a bug where we didn't insert FP_REG_KILL instructions into MBB's that
contain FP PHI nodes but no other FP defining instructions.  This fixes
183.equake

llvm-svn: 19495
2005-01-12 02:57:10 +00:00
Chris Lattner
92166ed1df Fold TRUNCATE (LOAD P) into a smaller load from P.
llvm-svn: 19494
2005-01-12 02:19:06 +00:00
Chris Lattner
258b23bd9d Be more careful about order of arg evalution for CopyToReg nodes. This shrinks
256.bzip2 from 7142 to 7103 lines of .s file.

Second, add initial support for folding loads into compares, though this code
is dynamically dead for now. :(

llvm-svn: 19493
2005-01-12 02:02:48 +00:00
Chris Lattner
604416e8f4 Fold some more [mem] op= val operators. This allows us to things like this
several times in 256.bzip2:

        mov %EAX, DWORD PTR [%ESP + 204]
-       mov %EAX, DWORD PTR [%EAX]
-       or %EAX, 2097152
-       mov %ECX, DWORD PTR [%ESP + 204]
-       mov DWORD PTR [%ECX], %EAX
+       or DWORD PTR [%EAX], 2097152

llvm-svn: 19492
2005-01-12 01:28:00 +00:00
Chris Lattner
e83ae1063f Fold loads into sign/zero extends. instead of:
mov %AL, BYTE PTR [%EDX + l18_length_code]
  movzx %EAX, %AL

Emit:

  movzx %EAX, BYTE PTR [%EDX + l18_length_code]

llvm-svn: 19489
2005-01-11 23:33:00 +00:00
Chris Lattner
87a38bd4a8 Comment out debug code :)
Select [mem] += Val operations.  For constants, we used to get:

  mov %ECX, -32768
  add %ECX, DWORD PTR [l4_match_start]
  mov DWORD PTR [l4_match_start], %ECX

Now we get:

  add DWORD PTR [l4_match_start], -32768

For other values we used to get:

  mov %EBP, %EDI   ;; because the add destroys the value
  add %EBP, DWORD PTR [l4_input_len]
  mov DWORD PTR [l4_input_len], %EBP

now we get:

  add DWORD PTR [l4_input_len], %EDI

Both of these use less registers than the alternative, are faster and smaller.

llvm-svn: 19488
2005-01-11 23:21:30 +00:00
Chris Lattner
282473a25d Handle the global address case here, not just the offset case.
llvm-svn: 19487
2005-01-11 22:58:43 +00:00
Chris Lattner
9eb2cc700b Treat int constants as not requiring a register, since they are almost always
folded into an instruction.

llvm-svn: 19486
2005-01-11 22:29:12 +00:00
Chris Lattner
74fcfd5148 Print the value types in the nodes of the graph
llvm-svn: 19485
2005-01-11 22:21:04 +00:00
Chris Lattner
f588cdd51e add an assertion, avoid creating copyfromreg/copytoreg pairs that are the
same for PHI nodes.

llvm-svn: 19484
2005-01-11 22:03:46 +00:00
Chris Lattner
7cb2220907 * Factor a bunch of binary operator cases into shared code.
* Fold loads into Add, sub, and, or, xor and mul when possible.
* Codegen shl X, 1 as add X, X

llvm-svn: 19483
2005-01-11 21:19:59 +00:00
Chris Lattner
b1a72cb39a Clear the whole array, always.
llvm-svn: 19482
2005-01-11 20:25:26 +00:00
Chris Lattner
b838c9748e Fold multiplies by 3,5,9 into addressing modes when possible.
llvm-svn: 19480
2005-01-11 19:37:02 +00:00
Chris Lattner
8de5a27681 Squelch optimized warning.
llvm-svn: 19475
2005-01-11 17:46:49 +00:00
Chris Lattner
e7b1130b01 Instead of generating stuff like this:
mov %ECX, %EAX
        add %ECX, 32768
        mov %SI, WORD PTR [2*%ECX + l13_prev]

Generate this:

        mov %SI, WORD PTR [2*%ECX + l13_prev + 65536]

This occurs when you have a GEP instruction where an index is
"something + imm".

llvm-svn: 19472
2005-01-11 06:36:20 +00:00
Chris Lattner
bb63a09cd1 Implement MEMCPY natively in terms of rep movs*
llvm-svn: 19468
2005-01-11 06:19:26 +00:00
Chris Lattner
b2b08a8bc1 Implement memset -> rep stos*
llvm-svn: 19467
2005-01-11 06:14:36 +00:00
Chris Lattner
58816a9e81 Announce that we don't support mem ops yet.
llvm-svn: 19466
2005-01-11 05:57:36 +00:00
Chris Lattner
963af6652b Teach legalize to lower MEMSET/MEMCPY/MEMMOVE operations if the target
does not support them.

llvm-svn: 19465
2005-01-11 05:57:22 +00:00
Chris Lattner
6b9082114f Print new operations.
llvm-svn: 19464
2005-01-11 05:57:01 +00:00
Chris Lattner
7cde8a2658 Turn memset/memcpy/memmove into the corresponding operations.
llvm-svn: 19463
2005-01-11 05:56:49 +00:00
Chris Lattner
f867443d7e Teach the address selector to make 'reg+reg' addressing modes.
llvm-svn: 19457
2005-01-11 04:40:19 +00:00
Reid Spencer
7e9642515c Add the LOADABLE_MODULE=1 directive to indicate that this shared library is
intended to be a dlopenable module and not a "plain" shared library.

llvm-svn: 19456
2005-01-11 04:33:32 +00:00
Chris Lattner
edf06be50e Emit NOT instructions.
llvm-svn: 19455
2005-01-11 04:31:30 +00:00
Chris Lattner
2eacd11a86 shift X, 0 -> X
llvm-svn: 19453
2005-01-11 04:25:13 +00:00
Chris Lattner
4e4bef2d6c Fix a bug emitting branches that broke a lot of programs.
llvm-svn: 19452
2005-01-11 04:06:27 +00:00
Chris Lattner
4b51297a94 Be more careful where we set ContainsFPCode. We were missing a set in the
int -> FP casting code.  Note that we don't have to set it for FP operations
that take FP values as operands: whatever produces the FP value will set the
flag.

llvm-svn: 19451
2005-01-11 03:50:45 +00:00
Chris Lattner
0c4c4094e3 Fix a major bug in setcc/cmov folding, where we accidentally
inverted the sense of the comparison.

llvm-svn: 19450
2005-01-11 03:37:59 +00:00
Chris Lattner
d188e03011 Take register pressure into account when we have to decide whether to
evaluate the LHS or the RHS of an operation first.  This causes good things
to happen.  For example, instead of compiling a loop to this:

.LBBstrength_result7_1: # loopentry
        movl 16(%esp), %edi
        movl (%edi), %edi             ;;; LOAD
        movl (%ecx), %ebx
        movl $2, (%eax,%ebx,4)
        movl (%edx), %ebx
        movl %esi, %ebp
        addl $21, %ebp
        addl $42, %esi
        cmpl $0, %edi                 ;;; USE
        cmovne %esi, %ebp
        cmpl %ebp, %ebx
        movl %ebp, %esi
        jg .LBBstrength_result7_1

We now compile it to this:

.LBBstrength_result7_1: # loopentry
        movl %edi, %ebx
        addl $42, %ebx
        addl $21, %edi
        movl (%ecx), %ebp              ;; LOAD
        cmpl $0, %ebp                  ;; USE
        cmovne %ebx, %edi
        movl (%edx), %ebx
        movl $2, (%eax,%ebx,4)
        movl (%esi), %ebx
        cmpl %edi, %ebx
        jg .LBBstrength_result7_1

Which reduces register pressure enough (in this case) to avoid spilling in the
loop.

As another example, consider the CodeGen/X86/regpressure.ll testcase.  We
used to generate this code for both cases:

regpressure1:
        subl $32, %esp
        movl %esi, 12(%esp)
        movl %edi, 8(%esp)
        movl %ebx, 4(%esp)
        movl %ebp, (%esp)
        movl 36(%esp), %ecx
        movl (%ecx), %eax
        movl 4(%ecx), %edx
        movl %edx, 24(%esp)
        movl 8(%ecx), %edx
        movl %edx, 16(%esp)
        movl 12(%ecx), %edx
        movl 16(%ecx), %esi
        movl 20(%ecx), %edi
        movl 24(%ecx), %ebx
        movl %ebx, 28(%esp)
        movl 28(%ecx), %ebx
        movl 32(%ecx), %ebp
        movl %ebp, 20(%esp)
        movl 36(%ecx), %ecx
        imull 24(%esp), %eax
        imull 16(%esp), %eax
        imull %edx, %eax
        imull %esi, %eax
        imull %edi, %eax
        imull 28(%esp), %eax
        imull %ebx, %eax
        imull 20(%esp), %eax
        imull %ecx, %eax
        movl (%esp), %ebp
        movl 4(%esp), %ebx
        movl 8(%esp), %edi
        movl 12(%esp), %esi
        addl $32, %esp
        ret

This code is basically trying to do all of the loads first, then execute all
of the multiplies.  Because we run out of registers, lots of spill code happens.
We now generate this code for both cases:

regpressure1:
        movl 4(%esp), %ecx
        movl (%ecx), %eax
        movl 4(%ecx), %edx
        imull %edx, %eax
        movl 8(%ecx), %edx
        imull %edx, %eax
        movl 12(%ecx), %edx
        imull %edx, %eax
        movl 16(%ecx), %edx
        imull %edx, %eax
        movl 20(%ecx), %edx
        imull %edx, %eax
        movl 24(%ecx), %edx
        imull %edx, %eax
        movl 28(%ecx), %edx
        imull %edx, %eax
        movl 32(%ecx), %edx
        imull %edx, %eax
        movl 36(%ecx), %ecx
        imull %ecx, %eax
        ret

which is much nicer (when we fold loads into the muls it will be even better).
The old instruction selector used to produce the good code for regpressure1
but not for regpressure2, as it depended on the order of operations in the
LLVM code.

llvm-svn: 19449
2005-01-11 03:11:44 +00:00
Chris Lattner
07a3ade230 Print SelectionDAGs bottom up, include extra info in the node labels
llvm-svn: 19447
2005-01-11 00:34:33 +00:00
Chris Lattner
1c273d3a14 Add a marker for the graph root.
llvm-svn: 19445
2005-01-10 23:52:04 +00:00
Chris Lattner
daa052a97e Put the operation name in each node, put the function name on the graph.
llvm-svn: 19444
2005-01-10 23:26:00 +00:00
Chris Lattner
0307506841 Split out SDNode::getOperationName into its own method.
llvm-svn: 19443
2005-01-10 23:25:25 +00:00
Chris Lattner
8c13447254 Implement initial selectiondag printing support. This gets us a nice
graph with no labels! :)

llvm-svn: 19441
2005-01-10 23:08:40 +00:00
Chris Lattner
497e24c885 Fold setcc instructions into selects.
llvm-svn: 19438
2005-01-10 22:10:13 +00:00
Chris Lattner
65d007ab62 Add conditional moves for the parity flag.
llvm-svn: 19437
2005-01-10 22:09:33 +00:00
Chris Lattner
5433d8de29 Lower to the correct functions. This fixes FreeBench/fourinarow
llvm-svn: 19436
2005-01-10 21:02:37 +00:00
Chris Lattner
d61491dea2 Implement 8-bit multiply for X86.
llvm-svn: 19435
2005-01-10 20:55:48 +00:00
Chris Lattner
b35b30c283 Rework constant pool handling so that function constant pools are no longer
leaked to the system.  Now they are destroyed with the JITMemoryManager is
destroyed.

llvm-svn: 19434
2005-01-10 18:23:22 +00:00
Jeff Cohen
8b03a55724 Apply feedback from Chris.
llvm-svn: 19432
2005-01-10 04:23:32 +00:00
Jeff Cohen
a7f1ae5dc0 Apply feed back from Chris:
1. Rename createLoaderPass to CreateProfileLoaderPass
  2. Opt shouldn't use the pass registered in CodeGen.

llvm-svn: 19431
2005-01-10 03:56:27 +00:00
Chris Lattner
02236df007 Implement a couple of more simplifications. This lets us codegen:
int test2(int * P, int* Q, int A, int B) {
        return P+A == P;
}

into:

test2:
        movl 4(%esp), %eax
        movl 12(%esp), %eax
        shll $2, %eax
        cmpl $0, %eax
        sete %al
        movzbl %al, %eax
        ret

instead of:

test2:
        movl 4(%esp), %eax
        movl 12(%esp), %ecx
        leal (%eax,%ecx,4), %ecx
        cmpl %eax, %ecx
        sete %al
        movzbl %al, %eax
        ret

ICC is producing worse code:

test2:
        movl      4(%esp), %eax                                 #8.5
        movl      12(%esp), %edx                                #8.5
        lea       (%edx,%edx), %ecx                             #9.9
        addl      %ecx, %ecx                                    #9.9
        addl      %eax, %ecx                                    #9.9
        cmpl      %eax, %ecx                                    #9.16
        movl      $0, %eax                                      #9.16
        sete      %al                                           #9.16
        ret                                                     #9.16

as is GCC (looks like our old code):

test2:
        movl    4(%esp), %edx
        movl    12(%esp), %eax
        leal    (%edx,%eax,4), %ecx
        cmpl    %edx, %ecx
        sete    %al
        movzbl  %al, %eax
        ret

llvm-svn: 19430
2005-01-10 02:03:02 +00:00
Chris Lattner
8d09b03ed1 Fix incorrect constant folds, fixing Stepanov after the SHR patch.
llvm-svn: 19429
2005-01-10 01:16:03 +00:00
Chris Lattner
9d479d4a34 Constant fold shifts, turning this loop:
.LBB_Z5test0PdS__3:     # no_exit.1
        fldl data(,%eax,8)
        fldl 24(%esp)
        faddp %st(1)
        fstl 24(%esp)
        incl %eax
        movl $16000, %ecx
        sarl $3, %ecx
        cmpl %eax, %ecx
        fstpl 16(%esp)
        #FP_REG_KILL
        jg .LBB_Z5test0PdS__3   # no_exit.1

into:

.LBB_Z5test0PdS__3:     # no_exit.1
        fldl data(,%eax,8)
        fldl 24(%esp)
        faddp %st(1)
        fstl 24(%esp)
        incl %eax
        cmpl $2000, %eax
        fstpl 16(%esp)
        #FP_REG_KILL
        jl .LBB_Z5test0PdS__3   # no_exit.1

llvm-svn: 19427
2005-01-10 00:07:15 +00:00
Reid Spencer
283688b80d Rename Unix/*.cpp and Win32/*.cpp to have a *.inc suffix so that the silly
gdb debugger doesn't get confused on which file it is reading (the one in
lib/System or the one in lib/System/{Win32,Unix})

llvm-svn: 19426
2005-01-09 23:29:00 +00:00
Chris Lattner
59d7066da8 Add some folds for == and != comparisons. This allows us to
codegen this loop in stepanov:

no_exit.i:              ; preds = %entry, %no_exit.i, %then.i, %_Z5checkd.exit
        %i.0.0 = phi int [ 0, %entry ], [ %i.0.0, %no_exit.i ], [ %inc.0, %_Z5checkd.exit ], [ %inc.012, %then.i ]              ; <int> [#uses=3]
        %indvar = phi uint [ %indvar.next, %no_exit.i ], [ 0, %entry ], [ 0, %then.i ], [ 0, %_Z5checkd.exit ]          ; <uint> [#uses=3]
        %result_addr.i.0 = phi double [ %tmp.4.i.i, %no_exit.i ], [ 0.000000e+00, %entry ], [ 0.000000e+00, %then.i ], [ 0.000000e+00, %_Z5checkd.exit ]          ; <double> [#uses=1]
        %first_addr.0.i.2.rec = cast uint %indvar to int                ; <int> [#uses=1]
        %first_addr.0.i.2 = getelementptr [2000 x double]* %data, int 0, uint %indvar           ; <double*> [#uses=1]
        %inc.i.rec = add int %first_addr.0.i.2.rec, 1           ; <int> [#uses=1]
        %inc.i = getelementptr [2000 x double]* %data, int 0, int %inc.i.rec            ; <double*> [#uses=1]
        %tmp.3.i.i = load double* %first_addr.0.i.2             ; <double> [#uses=1]
        %tmp.4.i.i = add double %result_addr.i.0, %tmp.3.i.i            ; <double> [#uses=2]
        %tmp.2.i = seteq double* %inc.i, getelementptr ([2000 x double]* %data, int 0, int 2000)                ; <bool> [#uses=1]
        %indvar.next = add uint %indvar, 1              ; <uint> [#uses=1]
        br bool %tmp.2.i, label %_Z10accumulateIPddET0_T_S2_S1_.exit, label %no_exit.i

To this:

.LBB_Z4testIPddEvT_S1_T0__1:    # no_exit.i
        fldl data(,%eax,8)
        fldl 16(%esp)
        faddp %st(1)
        fstpl 16(%esp)
        incl %eax
        movl %eax, %ecx
        shll $3, %ecx
        cmpl $16000, %ecx
        #FP_REG_KILL
        jne .LBB_Z4testIPddEvT_S1_T0__1 # no_exit.i

instead of this:

.LBB_Z4testIPddEvT_S1_T0__1:    # no_exit.i
        fldl data(,%eax,8)
        fldl 16(%esp)
        faddp %st(1)
        fstpl 16(%esp)
        incl %eax
        leal data(,%eax,8), %ecx
        leal data+16000, %edx
        cmpl %edx, %ecx
        #FP_REG_KILL
        jne .LBB_Z4testIPddEvT_S1_T0__1 # no_exit.i

llvm-svn: 19425
2005-01-09 20:52:51 +00:00
Jeff Cohen
f692cd303d Add last four createXxxPass functions
llvm-svn: 19424
2005-01-09 20:42:52 +00:00
Jeff Cohen
91dd6d2d20 Fix VC++ compilation error
llvm-svn: 19423
2005-01-09 20:41:56 +00:00
Chris Lattner
fa06762d0e Print the DAG out more like a DAG in nested format.
llvm-svn: 19422
2005-01-09 20:38:33 +00:00
Chris Lattner
e3b9f22967 Print out nodes sorted by their address to make it easier to find them in a list.
llvm-svn: 19421
2005-01-09 20:26:36 +00:00
Chris Lattner
fcab5f75c0 Codegen (Reg|imm)+&GV as an LEA, because we cannot put it into the immediate field
of an ADDri (due to current restrictions on MachineOperand :( ).  This allows
us to generate:

        leal Data+16000, %edx

instead of:

        movl $Data, %edx
        addl $16000, %edx

llvm-svn: 19420
2005-01-09 20:20:29 +00:00
Chris Lattner
82caa0dc2e Add a simple transformation. This allows us to compile one of the inner
loops in stepanov to this:

.LBB_Z5test0PdS__2:     # no_exit.1
        fldl data(,%eax,8)
        fldl 24(%esp)
        faddp %st(1)
        fstl 24(%esp)
        incl %eax
        cmpl $2000, %eax
        fstpl 16(%esp)
        #FP_REG_KILL
        jl .LBB_Z5test0PdS__2

instead of this:

.LBB_Z5test0PdS__2:     # no_exit.1
        fldl data(,%eax,8)
        fldl 24(%esp)
        faddp %st(1)
        fstl 24(%esp)
        incl %eax
        movl $data, %ecx
        movl %ecx, %edx
        addl $16000, %edx
        subl %ecx, %edx
        movl %edx, %ecx
        sarl $2, %ecx
        shrl $29, %ecx
        addl %ecx, %edx
        sarl $3, %edx
        cmpl %edx, %eax
        fstpl 16(%esp)
        #FP_REG_KILL
        jl .LBB_Z5test0PdS__2

The old instruction selector produced:

.LBB_Z5test0PdS__2:     # no_exit.1
        fldl 24(%esp)
        faddl data(,%eax,8)
        fstl 24(%esp)
        movl %eax, %ecx
        incl %ecx
        incl %eax
        leal data+16000, %edx
        movl $data, %edi
        subl %edi, %edx
        movl %edx, %edi
        sarl $2, %edi
        shrl $29, %edi
        addl %edi, %edx
        sarl $3, %edx
        cmpl %edx, %ecx
        fstpl 16(%esp)
        #FP_REG_KILL
        jl .LBB_Z5test0PdS__2   # no_exit.1

Which is even worse!

llvm-svn: 19419
2005-01-09 20:09:57 +00:00
Chris Lattner
35375c11bf Fix copy and pasto's for FP -> Int. This fixes fldry
llvm-svn: 19418
2005-01-09 19:49:59 +00:00
Chris Lattner
d674d08230 Fix a bug legalizing call instructions (make sure to remember all result
values), and eliminate some switch statements.

llvm-svn: 19417
2005-01-09 19:43:23 +00:00
Chris Lattner
ac23355362 Fix a minor bug legalizing dynamic_stackalloc. This allows us to compile
std::__pad<wchar_t, std::char_traits<wchar_t> >::_S_pad(std::ios_base&, wchar_t, wchar_t*, wchar_t const*, int, int, bool)

from libstdc++

llvm-svn: 19416
2005-01-09 19:07:54 +00:00
Chris Lattner
b3e31c6def Teach legalize to deal with DYNAMIC_STACKALLOC (aka a dynamic llvm alloca)
llvm-svn: 19415
2005-01-09 19:03:49 +00:00
Chris Lattner
45155a3dee Initial implementation of FP->INT and INT->FP casts
Also, fix zero_extend from bool to i8, which fixes Shootout/objinst.

llvm-svn: 19414
2005-01-09 18:52:44 +00:00
Jeff Cohen
6827f061cc Get lib/Analysis/DataStructure to compile with VC++
llvm-svn: 19412
2005-01-09 04:18:28 +00:00
Chris Lattner
9ca9b20447 Fix a subtle bug involving constant expr casts from int to fp
llvm-svn: 19410
2005-01-09 01:49:29 +00:00
Chris Lattner
cc18c057cf Handle static alloca arguments to PHI nodes.
llvm-svn: 19409
2005-01-09 01:16:24 +00:00
Chris Lattner
c5e53c07fd Implement varargs and returnaddress/frameaddress intrinsics. With this
patch, all of SingleSource/UnitTests passes.

llvm-svn: 19408
2005-01-09 00:01:27 +00:00