| 8 |
The form of an entry should be: |
The form of an entry should be: |
| 9 |
|
|
| 10 |
Name: |
Name: |
| 11 |
Date: |
Date: yyyy/mm/dd |
| 12 |
Tag: <post-commit CVS tag> |
Tag: <post-commit CVS tag> |
| 13 |
Description: |
Description: |
| 14 |
|
|
| 15 |
---------------------------------------------------------------------- |
---------------------------------------------------------------------- |
| 16 |
|
Name: Matthias Blume |
| 17 |
|
Date: 2002/02/13 22:40:00 EST |
| 18 |
|
Tag: blume-20020213-fptr-rtti |
| 19 |
|
Description: |
| 20 |
|
|
| 21 |
|
Fixed a bug in ml-nlffigen that was introduced with one of the previous |
| 22 |
|
updates. |
| 23 |
|
|
| 24 |
|
---------------------------------------------------------------------- |
| 25 |
|
Name: Matthias Blume |
| 26 |
|
Date: 2002/02/13 16:41:00 EST |
| 27 |
|
Tag: blume-20020213-cmlpq |
| 28 |
|
Description: |
| 29 |
|
|
| 30 |
|
Added new priority queue export symbols (which have just been added to |
| 31 |
|
smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would |
| 32 |
|
not compile and the installer would choke.) |
| 33 |
|
|
| 34 |
|
---------------------------------------------------------------------- |
| 35 |
|
Name: Matthias Blume |
| 36 |
|
Date: 2002/02/13 16:15:00 EST |
| 37 |
|
Tag: blume-20020213-various |
| 38 |
|
Description: |
| 39 |
|
|
| 40 |
|
1. More tweaks to ml-nlffigen: |
| 41 |
|
|
| 42 |
|
- better internal datastructures (resulting in slight speedup) |
| 43 |
|
- "-match" option requires exact match |
| 44 |
|
- "localized" gensym counters (untagged structs/unions nested within |
| 45 |
|
other structs/unions or within typedefs get a fresh counter; their |
| 46 |
|
tag will be prefixed by a concatenation of their parents' tags) |
| 47 |
|
- bug fixes (related to calculation of transitive closure of types |
| 48 |
|
to be included in the output) |
| 49 |
|
|
| 50 |
|
2. Minor Basis updates: |
| 51 |
|
|
| 52 |
|
- added implementations for List.collate and Option.app |
| 53 |
|
|
| 54 |
|
---------------------------------------------------------------------- |
| 55 |
|
Name: Matthias Blume |
| 56 |
|
Date: 2002/02/11 15:55:00 EST |
| 57 |
|
Tag: blume-20020211-gensym |
| 58 |
|
Description: |
| 59 |
|
|
| 60 |
|
Added a "-gensym" option to command line of ml-nlffigen. This can be |
| 61 |
|
used to specify a "stem" -- a string that is inserted in all "gensym'd" |
| 62 |
|
names (ML structure names that correspond to unnamed C structs, unions, |
| 63 |
|
and enums), so that separate runs of ml-nlffigen do not clash. |
| 64 |
|
|
| 65 |
|
---------------------------------------------------------------------- |
| 66 |
|
Name: Matthias Blume |
| 67 |
|
Date: 2002/02/11 12:05:00 EST |
| 68 |
|
Tag: blume-20020211-gensml |
| 69 |
|
Description: |
| 70 |
|
|
| 71 |
|
A quick fix for a problem with GenSML (in the pgraph-util library): |
| 72 |
|
Make generation of toplevel "local" optional. (Strictly speaking, |
| 73 |
|
signature definitions within "local" are not legal SML.) |
| 74 |
|
|
| 75 |
|
Other than that: updates to INSTALL and cm/TODO. |
| 76 |
|
|
| 77 |
|
---------------------------------------------------------------------- |
| 78 |
|
Name: Matthias Blume |
| 79 |
|
Date: 2002/02/08 15:00:00 EST |
| 80 |
|
Tag: blume-20020208-uniquepid |
| 81 |
|
Description: |
| 82 |
|
|
| 83 |
|
0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!! |
| 84 |
|
|
| 85 |
|
1. The installer (config/install.sh) has gotten smarter: |
| 86 |
|
|
| 87 |
|
- Configuration options are a bit easier to specify now |
| 88 |
|
(in config/targets). |
| 89 |
|
- Bug in recognizing .tar.bz2 files fixed. |
| 90 |
|
- Installer automatically resolves dependencies between |
| 91 |
|
configuration options (e.g., if you ask for eXene, you will |
| 92 |
|
also get cml -- regardless whether you asked for it or not). |
| 93 |
|
- Installer can run in "quieter mode" by setting the environment |
| 94 |
|
variable INSTALL_QUIETLY to "true". "Quieter" does not mean |
| 95 |
|
"completely silent", though. |
| 96 |
|
- Build HashCons library as part of smlnj-lib. |
| 97 |
|
|
| 98 |
|
2. A new scheme for assigning persistent identifiers to compilation |
| 99 |
|
units (and, by extension, to types etc.) has been put into place. |
| 100 |
|
This fixes a long-standing bug where types and even dynamic values |
| 101 |
|
can get internally confused, thereby compromising type safety |
| 102 |
|
(abstraction) and dynamic correctness. See |
| 103 |
|
|
| 104 |
|
http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz |
| 105 |
|
|
| 106 |
|
for an example of how things could go wrong until now. |
| 107 |
|
|
| 108 |
|
The downside of the new scheme is that pids are not quite as |
| 109 |
|
persistent as they used to be: CM will generate a fresh pid |
| 110 |
|
for every compilation unit that it thinks it sees for the first |
| 111 |
|
time. That means that if you compile starting from a clean, fresh |
| 112 |
|
source tree at two different times, you end up with different |
| 113 |
|
binaries. |
| 114 |
|
|
| 115 |
|
Cutoff recompilation, however, has not been compromised because |
| 116 |
|
CM keeps pid information in special caches between runs. |
| 117 |
|
|
| 118 |
|
---------------------------------------------------------------------- |
| 119 |
|
Name: Lal George |
| 120 |
|
Date: 2002/02/07 15:34:13 EST 2002 |
| 121 |
|
Tag: <none> |
| 122 |
|
Description: |
| 123 |
|
|
| 124 |
|
Compilers that generate assembly code may produce global labels |
| 125 |
|
whose value is resolved at link time. The various peephole optimization |
| 126 |
|
modules did not take this in account. |
| 127 |
|
|
| 128 |
|
TODO. The Labels.addrOf function should really return an option |
| 129 |
|
type so that clients are forced to deal with this issue, rather |
| 130 |
|
than an exception being raised. |
| 131 |
|
|
| 132 |
|
---------------------------------------------------------------------- |
| 133 |
|
Name: Lal George |
| 134 |
|
Date: 2002/02/06 13:55:02 EST |
| 135 |
|
Tag: george-20020206-ra-breakup |
| 136 |
|
Description: |
| 137 |
|
|
| 138 |
|
1. A bug fix from Allen. |
| 139 |
|
|
| 140 |
|
A typo causes extra fstp %st(0)'s to be generated at compensation |
| 141 |
|
edges, which might cause stack underflow traps at runtime. This |
| 142 |
|
occurs in fft where there are extraneous fstps right before the 'into' |
| 143 |
|
trap instruction (in this case they are harmless since none of the |
| 144 |
|
integers overflow.) |
| 145 |
|
|
| 146 |
|
2. Pulled out various utility modules that were embedded in the modules |
| 147 |
|
of the register allocator. I need these modules for other purposes, but |
| 148 |
|
they are not complete enough to put into a library (just yet). |
| 149 |
|
---------------------------------------------------------------------- |
| 150 |
|
Name: Matthias Blume |
| 151 |
|
Date: 2002/01/31 16:05:00 EST |
| 152 |
|
Tag: blume-20020131-sparc-ccalls |
| 153 |
|
Description: |
| 154 |
|
|
| 155 |
|
1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes) |
| 156 |
|
of extra stack space by mistake. Fixed. |
| 157 |
|
|
| 158 |
|
2. Bug in logic of handling of command-line options in ml-nlffigen fixed. |
| 159 |
|
|
| 160 |
|
---------------------------------------------------------------------- |
| 161 |
|
Name: Allen Leung |
| 162 |
|
Date: 2002/01/30 |
| 163 |
|
Tag: leunga-20020130-nowhere-bug-fix |
| 164 |
|
Description: |
| 165 |
|
|
| 166 |
|
MLRISC bug fixes: |
| 167 |
|
1. Fixed a bindings computation bug in the 'nowhere' program generator tool. |
| 168 |
|
2. MachineInt.fromString was negating its value. |
| 169 |
|
|
| 170 |
|
---------------------------------------------------------------------- |
| 171 |
|
Name: Matthias Blume |
| 172 |
|
Date: 2002/01/29 |
| 173 |
|
Tag: blume-20020129-INSTALL |
| 174 |
|
Description: |
| 175 |
|
|
| 176 |
|
- Added somewhat detailed installation instructions (file INSTALL). |
| 177 |
|
- Fixed curl-detection bug in config/install.sh. |
| 178 |
|
- It is now possible to select the URL getter using the URLGETTER |
| 179 |
|
environment variable: |
| 180 |
|
|
| 181 |
|
not set / "unknown" --> automatic detection (script tries wget, |
| 182 |
|
curl, and lynx) |
| 183 |
|
"wget" / "curl" / "lynx" --> use the specified program (script "knows" |
| 184 |
|
how to properly invoke them) |
| 185 |
|
other --> use $URLGETTER directly, it must take |
| 186 |
|
precisely two command-line arguments |
| 187 |
|
(source URL and destination file name) |
| 188 |
|
|
| 189 |
|
---------------------------------------------------------------------- |
| 190 |
|
Name: Matthias Blume |
| 191 |
|
Date: 2002/01/28 |
| 192 |
|
Tag: blume-20020128-sparc-ccalls |
| 193 |
|
Description: |
| 194 |
|
|
| 195 |
|
- Fixed problem with calculation of "used" registers in sparc-c-calls. |
| 196 |
|
- Make use of the allocParam argument in sparc-c-calls. |
| 197 |
|
|
| 198 |
|
---------------------------------------------------------------------- |
| 199 |
|
Name: Matthias Blume |
| 200 |
|
Date: 2002/01/28 |
| 201 |
|
Tag: blume-20020128-allocParam |
| 202 |
|
Description: |
| 203 |
|
|
| 204 |
|
John Reppy: Changes c-calls API to accept client-callback for |
| 205 |
|
allocating extra stack space. |
| 206 |
|
me: Corresponding changes to mlriscGen (using a dummy argument that |
| 207 |
|
does not change the current behavior). |
| 208 |
|
|
| 209 |
|
---------------------------------------------------------------------- |
| 210 |
|
Name: Matthias Blume |
| 211 |
|
Date: 2002/01/28 12:00:00 |
| 212 |
|
Tag: Release_110_38 |
| 213 |
|
Description: |
| 214 |
|
|
| 215 |
|
This time for real!!! |
| 216 |
|
|
| 217 |
|
---------------------------------------------------------------------- |
| 218 |
|
Name: Matthias Blume |
| 219 |
|
Date: 2002/01/28 10:56:00 EST |
| 220 |
|
Tag: blume-20020128-retraction |
| 221 |
|
Description: |
| 222 |
|
|
| 223 |
|
0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced |
| 224 |
|
with blume-Release_110_38-retracted.) |
| 225 |
|
|
| 226 |
|
1. Fixed a problem with incorrect rounding modes in real64.sml. |
| 227 |
|
(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.) |
| 228 |
|
|
| 229 |
|
2. A bug in ml-nlffigen related to the handling of unnamed structs, unions, |
| 230 |
|
and enums fixed. The naming of corresponding ML identifiers should |
| 231 |
|
now be consistent again. |
| 232 |
|
|
| 233 |
|
---------------------------------------------------------------------- |
| 234 |
|
Name: Allen Leung |
| 235 |
|
Date: 2002/01/27 |
| 236 |
|
Tag: leunga-20020127-nowhere |
| 237 |
|
Description: |
| 238 |
|
|
| 239 |
|
Added a target called nowhere in the configuration scripts. |
| 240 |
|
Enabling this will build the MLRISC 'nowhere' tool (for translating |
| 241 |
|
programs with where-clauses into legal SML code) during installation. |
| 242 |
|
|
| 243 |
|
---------------------------------------------------------------------- |
| 244 |
|
Name: Matthias Blume |
| 245 |
|
Date: 2002/01/25 21:27:00 EST |
| 246 |
|
Tag: blume-Release_110_38-retracted |
| 247 |
|
Description: |
| 248 |
|
|
| 249 |
|
Call it a (working) release! Version is 110.38. Bootfiles are ready. |
| 250 |
|
|
| 251 |
|
README will be added later. |
| 252 |
|
|
| 253 |
|
!!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag |
| 254 |
|
(Release_110_38) removed. Reason: Last-minute bug fixes. |
| 255 |
|
|
| 256 |
|
---------------------------------------------------------------------- |
| 257 |
|
Name: Matthias Blume |
| 258 |
|
Date: 2002/01/25 |
| 259 |
|
Tag: blume-20020125-ffi |
| 260 |
|
Description: |
| 261 |
|
|
| 262 |
|
A large number of tweaks and improvements to ml-nlffi-lib and |
| 263 |
|
ml-nlffigen: |
| 264 |
|
|
| 265 |
|
- ML represenation types have been streamlined |
| 266 |
|
- getter and setter functions work with concrete values, not abstract |
| 267 |
|
ones where possible |
| 268 |
|
- ml-nlffigen command line more flexible (see README file there) |
| 269 |
|
- some bugs have been fixed (hopefully) |
| 270 |
|
|
| 271 |
|
---------------------------------------------------------------------- |
| 272 |
|
Name: Lal George |
| 273 |
|
Date: 2002/01/24 |
| 274 |
|
Tag: george-20020124-risc-ra-interface |
| 275 |
|
Description: |
| 276 |
|
|
| 277 |
|
There is a dramatic simplification in the interface to the |
| 278 |
|
register allocator for RISC architectures as a result of making |
| 279 |
|
parallel copy instructions explicit. |
| 280 |
|
|
| 281 |
|
---------------------------------------------------------------------- |
| 282 |
|
Name: Matthias Blume |
| 283 |
|
Date: 2002/01/22 |
| 284 |
|
Tag: blume-20020122-x86-ccalls |
| 285 |
|
Description: |
| 286 |
|
|
| 287 |
|
Bug fix for c-calls on x86 (having to do with how char- and |
| 288 |
|
short-arguments are being handled). |
| 289 |
|
|
| 290 |
|
---------------------------------------------------------------------- |
| 291 |
|
Name: Matthias Blume |
| 292 |
|
Date: 2002/01/21 |
| 293 |
|
Tag: blume-20020121-ff |
| 294 |
|
Description: |
| 295 |
|
|
| 296 |
|
Another day of fiddling with the FFI... |
| 297 |
|
|
| 298 |
|
1. Bug fix/workaround: CKIT does not complain about negative array |
| 299 |
|
dimensions, so ml-nlffigen has to guard itself against this possibility. |
| 300 |
|
(Otherwise a negative dimension would send it into an infinite loop.) |
| 301 |
|
|
| 302 |
|
2. Some of the abstract types (light objects, light pointers, most "base" |
| 303 |
|
types) in structure C are now eqtypes. |
| 304 |
|
|
| 305 |
|
3. Added constructors and test functions for NULL function pointers. |
| 306 |
|
|
| 307 |
|
---------------------------------------------------------------------- |
| 308 |
|
Name: Matthias Blume |
| 309 |
|
Date: 2002/01/18 |
| 310 |
|
Tag: blume-20020118-ready-for-new-release |
| 311 |
|
Description: |
| 312 |
|
|
| 313 |
|
Made config/srcarchiveurl point to a new place. (Will provide boot |
| 314 |
|
files shortly.) |
| 315 |
|
|
| 316 |
|
Maybe we christen this to be 110.38? |
| 317 |
|
|
| 318 |
|
---------------------------------------------------------------------- |
| 319 |
|
Name: Matthias Blume |
| 320 |
|
Date: 2002/01/18 |
| 321 |
|
Tag: blume-20020118-more-ffifiddle |
| 322 |
|
Description: |
| 323 |
|
|
| 324 |
|
Today's FFI fiddling: |
| 325 |
|
|
| 326 |
|
- Provided a structure CGetSet with "convenient" versions of C.Get.* and |
| 327 |
|
C.Set.* that use concrete (MLRep.*) arguments and results instead |
| 328 |
|
of abstract ones. |
| 329 |
|
|
| 330 |
|
- Provided word-style bit operations etc. for "int" representation |
| 331 |
|
types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short, |
| 332 |
|
and Long. |
| 333 |
|
|
| 334 |
|
---------------------------------------------------------------------- |
| 335 |
|
Name: Matthias Blume |
| 336 |
|
Date: 2002/01/18 |
| 337 |
|
Tag: blume-20020118-use-x86-fp |
| 338 |
|
Description: |
| 339 |
|
|
| 340 |
|
Now that x86-fast-fp seems to be working, I turned it back on again |
| 341 |
|
by default. (Seems to work fine now, even with the FFI.) |
| 342 |
|
|
| 343 |
|
Other than that, I added some documentation about the FFI to |
| 344 |
|
src/ml-nlffigen/README and updated the FFI test examples in |
| 345 |
|
src/ml-nlffi-lib/Tests/*. |
| 346 |
|
|
| 347 |
|
---------------------------------------------------------------------- |
| 348 |
|
Name: Allen Leung |
| 349 |
|
Date: 2002/01/17 |
| 350 |
|
Tag: leunga-20020117-x86-fast-fp-call |
| 351 |
|
Description: |
| 352 |
|
|
| 353 |
|
1. Fixed a problem with handling return fp values when x86's fast fp |
| 354 |
|
mode is turned on. |
| 355 |
|
|
| 356 |
|
2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead |
| 357 |
|
of %f32. |
| 358 |
|
|
| 359 |
|
3. Added a constructor INT32lit to the ast of MLRISC tools. |
| 360 |
|
|
| 361 |
|
---------------------------------------------------------------------- |
| 362 |
|
Name: Matthias Blume |
| 363 |
|
Date: 2002/01/16 |
| 364 |
|
Tag: blume-20020116-ffifiddle |
| 365 |
|
Description: |
| 366 |
|
|
| 367 |
|
More fiddling with the FFI interface: |
| 368 |
|
|
| 369 |
|
- Make constness 'c instead of rw wherever possible. This eliminates |
| 370 |
|
the need for certain explicit coercions. (However, due to ML's |
| 371 |
|
value polymorphism, there will still be many cases where explicit |
| 372 |
|
coercions are necessary. Phantom types are not the whole answer |
| 373 |
|
to modeling a subtyping relationship in ML.) |
| 374 |
|
|
| 375 |
|
- ro/rw coersions for pointers added. (Avoids the detour through */&.) |
| 376 |
|
|
| 377 |
|
- "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates |
| 378 |
|
clumsy workaround for varargs problem.) |
| 379 |
|
|
| 380 |
|
---------------------------------------------------------------------- |
| 381 |
|
Name: Lal George |
| 382 |
|
Date: 2002/01/15 |
| 383 |
|
Tag: <none> |
| 384 |
|
Description: |
| 385 |
|
|
| 386 |
|
1. Since COPY instructions are no longer native to the architecture, |
| 387 |
|
a generic functor can be used to implement the expandCopies function. |
| 388 |
|
|
| 389 |
|
2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a |
| 390 |
|
TEXT segment. |
| 391 |
|
|
| 392 |
|
---------------------------------------------------------------------- |
| 393 |
|
Name: Matthias Blume |
| 394 |
|
Date: 2002/01/15 |
| 395 |
|
Tag: blume-20020115-ffiupdates |
| 396 |
|
Description: |
| 397 |
|
|
| 398 |
|
1. Fix for bug resulting in single-precision float values being returned |
| 399 |
|
incorrectly from FFI calls. |
| 400 |
|
|
| 401 |
|
2. Small modifications to C FFI API: |
| 402 |
|
|
| 403 |
|
- memory-allocation routines return straight objects (no options) |
| 404 |
|
and raise an exception in out-of-memory situations |
| 405 |
|
- unsafe extensions to cast between function pointers and pointers |
| 406 |
|
from/to ints |
| 407 |
|
- added structure C_Debug as an alternative to structure C where |
| 408 |
|
pointer-dereferencing (|*| and |*!) always check for null-pointers |
| 409 |
|
- added open_lib' to DynLinkage; open_lib' works like open_lib |
| 410 |
|
but also takes a (possibly empty) list of existing library handles |
| 411 |
|
that the current library depends on |
| 412 |
|
|
| 413 |
|
---------------------------------------------------------------------- |
| 414 |
|
Name: Matthias Blume |
| 415 |
|
Date: 2002/01/10 |
| 416 |
|
Tag: blume-20020110-newffigen |
| 417 |
|
Description: |
| 418 |
|
|
| 419 |
|
1. Updates to portable graph code. |
| 420 |
|
|
| 421 |
|
2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much |
| 422 |
|
more scalable now so that even huge interfaces such as the one |
| 423 |
|
for GTK compile in finite time and space. :-) |
| 424 |
|
See src/ml-nlffigen/README for details on what's new. |
| 425 |
|
|
| 426 |
|
---------------------------------------------------------------------- |
| 427 |
|
Name: Lal George |
| 428 |
|
Date: 2001/01/09 14:31:35 EST 2002 |
| 429 |
|
Tag: george-20011206-rm-native-copy |
| 430 |
|
Description: |
| 431 |
|
|
| 432 |
|
Removed the native COPY and FCOPY instructions |
| 433 |
|
from all the architectures and replaced it with the |
| 434 |
|
explicit COPY instruction from the previous commit. |
| 435 |
|
|
| 436 |
|
It is now possible to simplify many of the optimizations |
| 437 |
|
modules that manipulate copies. This has not been |
| 438 |
|
done in this change. |
| 439 |
|
|
| 440 |
|
---------------------------------------------------------------------- |
| 441 |
|
Name: Lal George |
| 442 |
|
Date: 2001/12/06 16:50:13 EST 2001 |
| 443 |
|
Tag: george-20011206-mlrisc-instruction |
| 444 |
|
Description: |
| 445 |
|
|
| 446 |
|
Changed the representation of instructions from being fully abstract |
| 447 |
|
to being partially concrete. That is to say: |
| 448 |
|
|
| 449 |
|
from |
| 450 |
|
type instruction |
| 451 |
|
|
| 452 |
|
to |
| 453 |
|
type instr (* machine instruction *) |
| 454 |
|
|
| 455 |
|
datatype instruction = |
| 456 |
|
LIVE of {regs: C.cellset, spilled: C.cellset} |
| 457 |
|
| KILL of {regs: C.cellset, spilled: C.cellset} |
| 458 |
|
| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list} |
| 459 |
|
| ANNOTATION of {i: instruction, a: Annotations.annotation} |
| 460 |
|
| INSTR of instr |
| 461 |
|
|
| 462 |
|
This makes the handling of certain special instructions that appear on |
| 463 |
|
all architectures easier and uniform. |
| 464 |
|
|
| 465 |
|
LIVE and KILL say that a list of registers are live or killed at the |
| 466 |
|
program point where they appear. No spill code is generated when an |
| 467 |
|
element of the 'regs' field is spilled, but the register is moved to |
| 468 |
|
the 'spilled' (which is present, more for debugging than anything else). |
| 469 |
|
|
| 470 |
|
LIVE replaces the (now deprecated) DEFFREG instruction on the alpha. |
| 471 |
|
We used to generate: |
| 472 |
|
|
| 473 |
|
DEFFREG f1 |
| 474 |
|
f1 := f2 + f3 |
| 475 |
|
trapb |
| 476 |
|
|
| 477 |
|
but now generate: |
| 478 |
|
|
| 479 |
|
f1 := f2 + f3 |
| 480 |
|
trapb |
| 481 |
|
LIVE {regs=[f1,f2,f3], spilled=[]} |
| 482 |
|
|
| 483 |
|
Furthermore, the DEFFREG (hack) required that all floating point instruction |
| 484 |
|
use all registers mentioned in the instruction. Therefore f1 := f2 + f3, |
| 485 |
|
defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting |
| 486 |
|
in a cleaner alpha implementation. (Hopefully, intel will not get rid of |
| 487 |
|
this architecture). |
| 488 |
|
|
| 489 |
|
COPYXXX is intended to replace the parallel COPY and FCOPY available on |
| 490 |
|
all the architectures. This will result in further simplification of the |
| 491 |
|
register allocator that must be aware of them for coalescing purposes, and |
| 492 |
|
will also simplify certain aspects of the machine description that provides |
| 493 |
|
callbacks related to parallel copies. |
| 494 |
|
|
| 495 |
|
ANNOTATION should be obvious, and now INSTR represents the honest to God |
| 496 |
|
machine instruction set! |
| 497 |
|
|
| 498 |
|
The <arch>/instructions/<arch>Instr.sml files define certain utility |
| 499 |
|
functions for making porting easier -- essentially converting upper case |
| 500 |
|
to lower case. All machine instructions (of type instr) are in upper case, |
| 501 |
|
and the lower case form generates an MLRISC instruction. For example on |
| 502 |
|
the alpha we have: |
| 503 |
|
|
| 504 |
|
datatype instr = |
| 505 |
|
LDA of {r:cell, b:cell, d:operand} |
| 506 |
|
| ... |
| 507 |
|
|
| 508 |
|
val lda : {r:cell, b:cell, d:operand} -> instruction |
| 509 |
|
... |
| 510 |
|
|
| 511 |
|
where lda is just (INSTR o LDA), etc. |
| 512 |
|
|
| 513 |
|
---------------------------------------------------------------------- |
| 514 |
|
Name: Matthias Blume |
| 515 |
|
Date: 2001/11/22 21:40:00 EST |
| 516 |
|
Tag: Release_110_37 |
| 517 |
|
Description: |
| 518 |
|
|
| 519 |
|
Release 110.37. This time for real. |
| 520 |
|
|
| 521 |
|
---------------------------------------------------------------------- |
| 522 |
|
Name: Matthias Blume |
| 523 |
|
Date: 2001/11/21 16:35:00 EST |
| 524 |
|
Tag: blume-20011121-foot-in-mouth |
| 525 |
|
Description: |
| 526 |
|
|
| 527 |
|
Removed the "Release_110_37" tag because of a serious bug. |
| 528 |
|
This will be re-tagged once the bug is fixed. |
| 529 |
|
|
| 530 |
|
---------------------------------------------------------------------- |
| 531 |
|
Name: Matthias Blume |
| 532 |
|
Date: 2001/11/21 16:14:00 EST |
| 533 |
|
Tag: blume-20011121-forgottenfile |
| 534 |
|
Description: |
| 535 |
|
|
| 536 |
|
Forgot to add a file. (Just a .tex-file -- part of |
| 537 |
|
the CM manual source.) |
| 538 |
|
|
| 539 |
|
---------------------------------------------------------------------- |
| 540 |
|
Name: Matthias Blume |
| 541 |
|
Date: 2001/11/21 16:10:00 EST |
| 542 |
|
Tag: blume-20011121-invalid_110_37 |
| 543 |
|
Description: |
| 544 |
|
|
| 545 |
|
Note: I removed the original tag "Release_110_37" from this commit |
| 546 |
|
because we found a serious bug in all non-x86 backends. |
| 547 |
|
- Matthias |
| 548 |
|
|
| 549 |
|
1. Modifications to the SML/NJ code generator and to the runtime system |
| 550 |
|
so that code object name strings are directly inserted into code |
| 551 |
|
objects at code generation time. The only business the runtime system |
| 552 |
|
has with this is now to read the name strings on occasions. |
| 553 |
|
(The encoding of the name string has also changed somewhat.) |
| 554 |
|
|
| 555 |
|
2. CM now implements a simple "set calculus" for specifying export lists. |
| 556 |
|
In particular, it is now possible to refer to the export lists of |
| 557 |
|
other libraries/groups/sources and form unions as well as differences. |
| 558 |
|
See the latest CM manual for details. |
| 559 |
|
|
| 560 |
|
3. An separate notion of "proxy" libraries has again be eliminated from |
| 561 |
|
CM's model. (Proxy libraries are now simply a special case of using |
| 562 |
|
the export list calculus.) |
| 563 |
|
|
| 564 |
|
4. Some of the existing libraries now take advantage of the new set |
| 565 |
|
calculus. |
| 566 |
|
(Notice that not all libraries have been converted because some |
| 567 |
|
of the existing .cm-files are supposed to be backward compatible |
| 568 |
|
with 110.0.x.) |
| 569 |
|
|
| 570 |
|
5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use |
| 571 |
|
"General.exnMessage"! The former relies on a certain hook to be |
| 572 |
|
initialized, and that often does not happen in the stand-alone case.) |
| 573 |
|
|
| 574 |
|
---------------------------------------------------------------------- |
| 575 |
|
Name: Lal George |
| 576 |
|
Date: 2001/11/21 13:56:18 EST |
| 577 |
|
Tag: george-2001121-pseudo-ops |
| 578 |
|
Description: |
| 579 |
|
|
| 580 |
|
Implemented a complete redesign of MLRISC pseudo-ops. Now there |
| 581 |
|
ought to never be any question of incompatabilities with |
| 582 |
|
pseudo-op syntax expected by host assemblers. |
| 583 |
|
|
| 584 |
|
For now, only modules supporting GAS syntax are implemented |
| 585 |
|
but more should follow, such as MASM, and vendor assembler |
| 586 |
|
syntax, e.g. IBM as, Sun as, etc. |
| 587 |
|
|
| 588 |
|
---------------------------------------------------------------------- |
| 589 |
|
Name: Matthias Blume |
| 590 |
|
Date: 2001/11/14 11:52:00 EST |
| 591 |
|
Tag: blume-20011114-srcname |
| 592 |
|
Description: |
| 593 |
|
|
| 594 |
|
1. Routed the name of the current source file to mlriscgen where it |
| 595 |
|
should be directly emitted into the code object. (This last part |
| 596 |
|
is yet to be done.) |
| 597 |
|
|
| 598 |
|
2. Some cleanup of the pgraph code to make it match the proposal that |
| 599 |
|
I put out the other day. (The proposal notwithstanding, things are |
| 600 |
|
still in flux here.) |
| 601 |
|
|
| 602 |
|
---------------------------------------------------------------------- |
| 603 |
Name: Lal George |
Name: Lal George |
| 604 |
Date: 2001/11/14 09:44:04 EST |
Date: 2001/11/14 09:44:04 EST |
| 605 |
Tag: |
Tag: |
| 767 |
|
|
| 768 |
---------------------------------------------------------------------- |
---------------------------------------------------------------------- |
| 769 |
Name: Matthias Blume |
Name: Matthias Blume |
|
>>>>>>> 1.169 |
|
| 770 |
Date: 2001/09/18 15:35:00 EDT |
Date: 2001/09/18 15:35:00 EDT |
| 771 |
Tag: blume-20010918-readme11036 |
Tag: blume-20010918-readme11036 |
| 772 |
Description: |
Description: |
Parent Directory
| 
Revision Log
| 
Patch
