Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 952, Tue Oct 9 14:00:40 2001 UTC revision 1127, Fri Mar 8 01:35:33 2002 UTC
# Line 8  Line 8 
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: Allen Leung
17    Date: 2002/03/07 20:45:00 EST
18    Tag: leunga-20020307-x86-cmov
19    Description:
20    
21       Bug fixes for CMOVcc on x86.
22    
23       1. Added machine code generation for CMOVcc
24       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
25       3. CMOVcc cannot have an immediate operand as argument.
26    
27    ----------------------------------------------------------------------
28    Name: Matthias Blume
29    Date: 2002/03/07 16:15:00 EST
30    Tag: blume-20020307-controls
31    Description:
32    
33    This is a very large but mostly boring patch which makes (almost)
34    every tuneable compiler knob (i.e., pretty much everything under
35    Control.* plus a few other things) configurable via both the command
36    line and environment variables in the style CM did its configuration
37    until now.
38    
39    Try starting sml with '-h' (or, if you are brave, '-H')
40    
41    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
42    implements the underlying generic mechanism.
43    
44    The interface to some of the existing such facilities has changed somewhat.
45    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
46    (The getFoo interface is still there for backward-compatibility, but its
47    use is deprecated.)
48    
49    The ml-build script passes -Cxxx=yyy command-line arguments through so
50    that one can now twiddle the compiler settings when using this "batch"
51    compiler.
52    
53    TODO items:
54    
55    We should go through and throw out all controls that are no longer
56    connected to anything.  Moreover, we should go through and provide
57    meaningful (and correct!) documentation strings for those controls
58    that still are connected.
59    
60    Currently, multiple calls to Controls.new are accepted (only the first
61    has any effect).  Eventually we should make sure that every control
62    is being made (via Controls.new) exactly once.  Future access can then
63    be done using Controls.acc.
64    
65    Finally, it would probably be a good idea to use the getter-setter
66    interface to controls rather than ref cells.  For the time being, both
67    styles are provided by the Controls module, but getter-setter pairs are
68    better if thread-safety is of any concern because they can be wrapped.
69    
70    *****************************************
71    
72    One bug fix: The function blockPlacement in three of the MLRISC
73    backpatch files used to be hard-wired to one of two possibilities at
74    link time (according to the value of the placementFlag).  But (I
75    think) it should rather sense the flag every time.
76    
77    *****************************************
78    
79    Other assorted changes (by other people who did not supply a HISTORY entry):
80    
81    1. the cross-module inliner now works much better (Monnier)
82    2. representation of weights, frequencies, and probabilities in MLRISC
83       changed in preparation of using those for weighted block placement
84       (Reppy, George)
85    
86    ----------------------------------------------------------------------
87    Name: Lal George
88    Date: 2002/03/07 14:44:24 EST 2002
89    Tag: george-20020307-weighted-block-placement
90    
91    Tested the weighted block placement optimization on all architectures
92    (except the hppa) using AMPL to generate the block and edge frequencies.
93    Changes were required in the machine properties to correctly
94    categorize trap instructions. There is an MLRISC flag
95    "weighted-block-placement" that can be used to enable weighted block
96    placement, but this will be ineffective without block/edge
97    frequencies (coming soon).
98    
99    
100    ----------------------------------------------------------------------
101    Name: Lal George
102    Date: 2002/03/05 17:24:48 EST
103    Tag: george-20020305-linkage-cluster
104    
105    In order to support the block placement optimization, a new cluster
106    is generated as the very first cluster (called the linkage cluster).
107    It contains a single jump to the 'real' entry point for the compilation
108    unit. Block placement has no effect on the linkage cluster itself, but
109    all the other clusters  have full freedom in the manner in which they
110    reorder blocks or functions.
111    
112    On the x86 the typical linkage code that is generated is:
113       ----------------------
114            .align 2
115       L0:
116            addl    $L1-L0, 72(%esp)
117            jmp     L1
118    
119    
120            .align  2
121       L1:
122       ----------------------
123    
124    72(%esp) is the memory location for the stdlink register. This
125    must contain the address of the CPS function being called. In the
126    above example, it contains the address of  L0; before
127    calling L1 (the real entry point for the compilation unit), it
128    must contain the address for L1, and hence
129    
130            addl $L1-L0, 72(%esp)
131    
132    I have tested this on all architectures except the hppa.The increase
133    in code size is of course negligible
134    
135    ----------------------------------------------------------------------
136    Name: Allen Leung
137    Date: 2002/03/03 13:20:00 EST
138    Tag: leunga-20020303-mlrisc-tools
139    
140      Added #[ ... ] expressions to mlrisc tools
141    
142    ----------------------------------------------------------------------
143    Name: Matthias Blume
144    Date: 2002/02/27 12:29:00 EST
145    Tag: blume-20020227-cdebug
146    Description:
147    
148    - made types in structure C and C_Debug to be equal
149    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
150    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume
154    Date: 2002/02/26 12:00:00 EST
155    Tag: blume-20020226-ffi
156    Description:
157    
158    1. Fixed a minor bug in CM's "noweb" tool:
159       If numbering is turned off, then truly don't number (i.e., do not
160       supply the -L option to noweb).  The previous behavior was to supply
161       -L'' -- which caused noweb to use the "default" line numbering scheme.
162       Thanks to Chris Richards for pointing this out (and supplying the fix).
163    
164    2. Once again, I reworked some aspects of the FFI:
165    
166       A. The incomplete/complete type business:
167    
168       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
169         gone!
170       - ML types representing an incomplete type are now *equal* to
171         ML types representing their corresponding complete types (just like
172         in C).  This is still safe because ml-nlffigen will not generate
173         RTTI for incomplete types, nor will it generate functions that
174         require access to such RTTI.   But when ML code generated from both
175         incomplete and complete versions of the C type meet, the ML types
176         are trivially interoperable.
177    
178         NOTE:  These changes restore the full generality of the translation
179         (which was previously lost when I eliminated functorization)!
180    
181       B. Enum types:
182    
183       - Structure C now has a type constructor "enum" that is similar to
184         how the "su" constructor works.  However, "enum" is not a phantom
185         type because each "T enum" has values (and is isomorphic to
186         MLRep.Signed.int).
187       - There are generic access operations for enum objects (using
188         MLRep.Signed.int).
189       - ml-nlffigen will generate a structure E_foo for each "enum foo".
190         * The structure contains the definition of type "mlrep" (the ML-side
191         representation type of the enum).  Normally, mlrep is the same
192         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
193         then mlrep will be defined as a datatype -- thus facilitating
194         pattern matching on mlrep values.
195         ("-ec" will be suppressed if there are duplicate values in an
196          enumeration.)
197         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
198         will be generated for each C enum constant xxx.
199         * Conversion functions m2i and i2m convert between mlrep and
200         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
201         * Coversion functions c and ml convert between mlrep and "tag enum".
202         * Access functions (get/set) fetch and store mlrep values.
203       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
204         enumerations are merged into one single enumeration represented by
205         structure E_'.
206    
207    ----------------------------------------------------------------------
208    Name: Allen Leung
209    Date: 2002/02/25 04:45:00 EST
210    Tag: leunga-20020225-cps-spill
211    
212    This is a new implementation of the CPS spill phase.
213    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
214    In case of problems, replace it with the old file spill.sml
215    
216    The current compiler runs into some serious performance problems when
217    constructing a large record.  This can happen when we try to compile a
218    structure with many items.  Even a very simple structure like the following
219    makes the compiler slow down.
220    
221        structure Foo = struct
222           val x_1 = 0w1 : Word32.int
223           val x_2 = 0w2 : Word32.int
224           val x_3 = 0w3 : Word32.int
225           ...
226           val x_N = 0wN : Word32.int
227        end
228    
229    The following table shows the compile time, from N=1000 to N=4000,
230    with the old compiler:
231    
232    N
233    1000   CPS 100 spill                           0.04u  0.00s  0.00g
234           MLRISC ra                               0.06u  0.00s  0.05g
235              (spills = 0 reloads = 0)
236           TOTAL                                   0.63u  0.07s  0.21g
237    
238    1100   CPS 100 spill                           8.25u  0.32s  0.64g
239           MLRISC ra                               5.68u  0.59s  3.93g
240              (spills = 0 reloads = 0)
241           TOTAL                                   14.71u  0.99s  4.81g
242    
243    1500   CPS 100 spill                           58.55u  2.34s  1.74g
244           MLRISC ra                               5.54u  0.65s  3.91g
245              (spills = 543 reloads = 1082)
246           TOTAL                                   65.40u  3.13s  6.00g
247    
248    2000   CPS 100 spill                           126.69u  4.84s  3.08g
249           MLRISC ra                               0.80u  0.10s  0.55g
250              (spills = 42 reloads = 84)
251           TOTAL                                   129.42u  5.10s  4.13g
252    
253    3000   CPS 100 spill                           675.59u  19.03s  11.64g
254           MLRISC ra                               2.69u  0.27s  1.38g
255              (spills = 62 reloads = 124)
256           TOTAL                                   682.48u  19.61s  13.99g
257    
258    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
259           MLRISC ra                               4.96u  0.27s  2.72g
260              (spills = 85 reloads = 170)
261           TOTAL                                   2375.26u  57.21s  48.00g
262    
263    As you can see the old cps spill module suffers from some serious
264    performance problem.  But since I cannot decipher the old code fully,
265    instead of patching the problems up, I'm reimplementing it
266    with a different algorithm.  The new code is more modular,
267    smaller when compiled, and substantially faster
268    (O(n log n) time and O(n) space).  Timing of the new spill module:
269    
270    4000  CPS 100 spill                           0.02u  0.00s  0.00g
271          MLRISC ra                               0.25u  0.02s  0.15g
272             (spills=1 reloads=3)
273          TOTAL                                   7.74u  0.34s  1.62g
274    
275    Implementation details:
276    
277    As far as I can tell, the purpose of the CPS spill module is to make sure the
278    number of live variables at any program point (the bandwidth)
279    does not exceed a certain limit, which is determined by the
280    size of the spill area.
281    
282    When the bandwidth is too large, we decrease the register pressure by
283    packing live variables into spill records.  How we achieve this is
284    completely different than what we did in the old code.
285    
286    First, there is something about the MLRiscGen code generator
287    that we should be aware of:
288    
289    o MLRiscGen performs code motion!
290    
291       In particular, it will move floating point computations and
292       address computations involving only the heap pointer to
293       their use sites (if there is only a single use).
294       What this means is that if we have a CPS record construction
295       statement
296    
297           RECORD(k,vl,w,e)
298    
299       we should never count the new record address w as live if w
300       has only one use (which is often the case).
301    
302       We should do something similar to floating point, but the transformation
303       there is much more complex, so I won't deal with that.
304    
305    Secondly, there are now two new cps primops at our disposal:
306    
307     1. rawrecord of record_kind option
308        This pure operator allocates some uninitialized storage from the heap.
309        There are two forms:
310    
311         rawrecord NONE [INT n]  allocates a tagless record of length n
312         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
313                                     and initializes the tag.
314    
315     2. rawupdate of cty
316          rawupdate cty (v,i,x)
317          Assigns to x to the ith component of record v.
318          The storelist is not updated.
319    
320    We use these new primops for both spilling and increment record construction.
321    
322     1. Spilling.
323    
324        This is implemented with a linear scan algorithm (but generalized
325        to trees).  The algorithm will create a single spill record at the
326        beginning of the cps function and use rawupdate to spill to it,
327        and SELECT or SELp to reload from it.  So both spills and reloads
328        are fine-grain operations.  In contrast, in the old algorithm
329        "spills" have to be bundled together in records.
330    
331        Ideally, we should sink the spill record construction to where
332        it is needed.  We can even split the spill record into multiple ones
333        at the places where they are needed.  But CPS is not a good
334        representation for global code motion, so I'll keep it simple and
335        am not attempting this.
336    
337     2. Incremental record construction (aka record splitting).
338    
339        Long records with many component values which are simulatenously live
340        (recall that single use record addresses are not considered to
341         be live) are constructed with rawrecord and rawupdate.
342        We allocate space on the heap with rawrecord first, then gradually
343        fill it in with rawupdate.  This is the technique suggested to me
344        by Matthias.
345    
346        Some restrictions on when this is applicable:
347        1. It is not a VECTOR record.  The code generator currently does not handle
348           this case. VECTOR record uses double indirection like arrays.
349        2. All the record component values are defined in the same "basic block"
350           as the record constructor.  This is to prevent speculative
351           record construction.
352    
353    ----------------------------------------------------------------------
354    Name: Allen Leung
355    Date: 2002/02/22 01:02:00 EST
356    Tag: leunga-20020222-mlrisc-tools
357    
358    Minor bug fixes in the parser and rewriter
359    
360    ----------------------------------------------------------------------
361    Name: Allen Leung
362    Date: 2002/02/21 20:20:00 EST
363    Tag: leunga-20020221-peephole
364    
365    Regenerated the peephole files.  Some contained typos in the specification
366    and some didn't compile because of pretty printing bugs in the old version
367    of 'nowhere'.
368    
369    ----------------------------------------------------------------------
370    Name: Allen Leung
371    Date: 2002/02/19 20:20:00 EST
372    Tag: leunga-20020219-mlrisc-tools
373    Description:
374    
375       Minor bug fixes to the mlrisc-tools library:
376    
377       1.  Fixed up parsing colon suffixed keywords
378       2.  Added the ability to shut the error messages up
379       3.  Reimplemented the pretty printer and fixed up/improved
380           the pretty printing of handle and -> types.
381       4.  Fixed up generation of literal symbols in the nowhere tool.
382       5.  Added some SML keywords to to sml.sty
383    
384    ----------------------------------------------------------------------
385    Name: Matthias Blume
386    Date: 2002/02/19 16:20:00 EST
387    Tag: blume-20020219-cmffi
388    Description:
389    
390    A wild mix of changes, some minor, some major:
391    
392    * All C FFI-related libraries are now anchored under $c:
393        $/c.cm      --> $c/c.cm
394        $/c-int.cm  --> $c/internals/c-int.cm
395        $/memory.cm --> $c/memory/memory.cm
396    
397    * "make" tool (in CM) now treats its argument pathname slightly
398      differently:
399        1. If the native expansion is an absolute name, then before invoking
400           the "make" command on it, CM will apply OS.Path.mkRelative
401           (with relativeTo = OS.FileSys.getDir()) to it.
402        2. The argument will be passed through to subsequent phases of CM
403           processing without "going native".  In particular, if the argument
404           was an anchored path, then "make" will not lose track of that anchor.
405    
406    * Compiler backends now "know" their respective C calling conventions
407      instead of having to be told about it by ml-nlffigen.  This relieves
408      ml-nlffigen from one of its burdens.
409    
410    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
411    
412    * Export C_DEBUG and C_Debug from $c/c.cm.
413    
414    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
415      subtyping relationship between incomplete pointers and their complete
416      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
417      with the convention of instantiating 'o with ('t, 'c) obj whenever
418      the pointer target type is complete.  In the incomplete case, 'o
419      will be instantiated with some "'c iobj" -- a type obtained by
420      using one of the functors PointerToIncompleteType or PointerToCompleteType.
421    
422      Operations that work on both incomplete and complete pointer types are
423      typed as taking an 'o ptr while operations that require the target to
424      be known are typed as taking some ('t, 'c) obj ptr.
425    
426      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
427      where void is an eqtype without any values.  This makes it possible
428      to work on voidptr values using functions meant to operate on light
429      incomplete pointers.
430    
431    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
432      been vastly simplified.
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume
436    Date: 2002/02/19 10:48:00 EST
437    Tag: blume-20020219-pqfix
438    Description:
439    
440    Applied Chris Okasaki's bug fix for priority queues.
441    
442    ----------------------------------------------------------------------
443    Name: Matthias Blume
444    Date: 2002/02/15 17:05:00
445    Tag: Release_110_39
446    Description:
447    
448    Last-minute retagging is becoming a tradition... :-(
449    
450    This is the working release 110.39.
451    
452    ----------------------------------------------------------------------
453    Name: Matthias Blume
454    Date: 2002/02/15 16:00:00 EST
455    Tag: Release_110_39-orig
456    Description:
457    
458    Working release 110.39.  New bootfiles.
459    
460    (Update: There was a small bug in the installer so it wouldn't work
461    with all shells.  So I retagged. -Matthias)
462    
463    ----------------------------------------------------------------------
464    Name: Matthias Blume
465    Date: 2002/02/15 14:17:00 EST
466    Tag: blume-20020215-showbindings
467    Description:
468    
469    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
470    the latter can be useful for exploring what bindings are available at
471    the interactive prompt.  (The first function returns only the list
472    of symbols that are really bound, the second prints those but also the
473    ones that CM's autoloading mechanism knows about.)
474    
475    ----------------------------------------------------------------------
476    Name: Matthias Blume
477    Date: 2002/02/15 12:08:00 EST
478    Tag: blume-20020215-iptrs
479    Description:
480    
481    Two improvements to ml-nlffigen:
482    
483      1. Write files only if they do not exist or if their current contents
484         do not coincide with what's being written.  (That is, avoid messing
485         with the time stamps unless absolutely necessary.)
486    
487      2. Implement a "repository" mechanism for generated files related
488         to "incomplete pointer types".   See the README file for details.
489    
490    ----------------------------------------------------------------------
491    Name: Matthias Blume
492    Date: 2002/02/14 11:50:00 EST
493    Tag: blume-20020214-quote
494    Description:
495    
496    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
497    because of the new and improved tag generation scheme.  (Thanks to Allen
498    Leung for pointing it out.)
499    
500    ----------------------------------------------------------------------
501    Name: Lal George
502    Date: 2002/02/14 09:55:27 EST 2002
503    Tag: george-20020214-isabelle-bug
504    Description:
505    
506    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
507    of Isabelle on the x86.
508    
509    From Allen:
510    -----------
511     I've found the problem:
512    
513         in ra-core.sml, I use the counter "blocked" to keep track of the
514         true number of elements in the freeze queue.  When the counter goes
515         to zero, I skip examining the queue.  But I've messed up the
516         bookkeeping in combine():
517    
518             else ();
519             case !ucol of
520               PSEUDO => (if !cntv > 0 then
521                     (if !cntu > 0 then blocked := !blocked - 1 else ();
522                                        ^^^^^^^^^^^^^^^^^^^^^^^
523                      moveu := mergeMoveList(!movev, !moveu)
524                     )
525                  else ();
526    
527         combine() is called to coalesce two nodes u and v.
528         I think I was thinking that if the move counts of u and v are both
529         greater than zero then after they are coalesced then one node is
530         removed from the freeze queue.  Apparently I was thinking that
531         both u and v are of low degree, but that's clearly not necessarily true.
532    
533    
534    02/12/2002:
535        Here's the patch.  HOL now compiles.
536    
537        I don't know how this impact on performance (compile
538        time or runtime).  This bug caused the RA (especially on the x86)
539        to go thru the potential spill phase when there are still nodes on the
540        freeze queue.
541    
542    
543    
544    
545    ----------------------------------------------------------------------
546    Name: Matthias Blume
547    Date: 2002/02/13 22:40:00 EST
548    Tag: blume-20020213-fptr-rtti
549    Description:
550    
551    Fixed a bug in ml-nlffigen that was introduced with one of the previous
552    updates.
553    
554    ----------------------------------------------------------------------
555    Name: Matthias Blume
556    Date: 2002/02/13 16:41:00 EST
557    Tag: blume-20020213-cmlpq
558    Description:
559    
560    Added new priority queue export symbols (which have just been added to
561    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
562    not compile and the installer would choke.)
563    
564    ----------------------------------------------------------------------
565    Name: Matthias Blume
566    Date: 2002/02/13 16:15:00 EST
567    Tag: blume-20020213-various
568    Description:
569    
570    1. More tweaks to ml-nlffigen:
571    
572       - better internal datastructures (resulting in slight speedup)
573       - "-match" option requires exact match
574       - "localized" gensym counters (untagged structs/unions nested within
575         other structs/unions or within typedefs get a fresh counter; their
576         tag will be prefixed by a concatenation of their parents' tags)
577       - bug fixes (related to calculation of transitive closure of types
578         to be included in the output)
579    
580    2. Minor Basis updates:
581    
582       - added implementations for List.collate and Option.app
583    
584    ----------------------------------------------------------------------
585    Name: Matthias Blume
586    Date: 2002/02/11 15:55:00 EST
587    Tag: blume-20020211-gensym
588    Description:
589    
590    Added a "-gensym" option to command line of ml-nlffigen.  This can be
591    used to specify a "stem" -- a string that is inserted in all "gensym'd"
592    names (ML structure names that correspond to unnamed C structs, unions,
593    and enums), so that separate runs of ml-nlffigen do not clash.
594    
595    ----------------------------------------------------------------------
596    Name: Matthias Blume
597    Date: 2002/02/11 12:05:00 EST
598    Tag: blume-20020211-gensml
599    Description:
600    
601    A quick fix for a problem with GenSML (in the pgraph-util library):
602    Make generation of toplevel "local" optional.  (Strictly speaking,
603    signature definitions within "local" are not legal SML.)
604    
605    Other than that: updates to INSTALL and cm/TODO.
606    
607    ----------------------------------------------------------------------
608    Name: Matthias Blume
609    Date: 2002/02/08 15:00:00 EST
610    Tag: blume-20020208-uniquepid
611    Description:
612    
613    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
614    
615    1. The installer (config/install.sh) has gotten smarter:
616    
617         - Configuration options are a bit easier to specify now
618           (in config/targets).
619         - Bug in recognizing .tar.bz2 files fixed.
620         - Installer automatically resolves dependencies between
621           configuration options (e.g., if you ask for eXene, you will
622           also get cml -- regardless whether you asked for it or not).
623         - Installer can run in "quieter mode" by setting the environment
624           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
625           "completely silent", though.
626         - Build HashCons library as part of smlnj-lib.
627    
628    2. A new scheme for assigning persistent identifiers to compilation
629       units (and, by extension, to types etc.) has been put into place.
630       This fixes a long-standing bug where types and even dynamic values
631       can get internally confused, thereby compromising type safety
632       (abstraction) and dynamic correctness.  See
633    
634         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
635    
636       for an example of how things could go wrong until now.
637    
638       The downside of the new scheme is that pids are not quite as
639       persistent as they used to be: CM will generate a fresh pid
640       for every compilation unit that it thinks it sees for the first
641       time.  That means that if you compile starting from a clean, fresh
642       source tree at two different times, you end up with different
643       binaries.
644    
645       Cutoff recompilation, however, has not been compromised because
646       CM keeps pid information in special caches between runs.
647    
648    ----------------------------------------------------------------------
649    Name: Lal George
650    Date: 2002/02/07 15:34:13 EST 2002
651    Tag: <none>
652    Description:
653    
654    Compilers that generate assembly code may produce  global labels
655    whose value is resolved at link time. The various peephole optimization
656    modules did not take this in account.
657    
658    TODO. The Labels.addrOf function should really return an option
659    type so that clients are forced to deal with this issue, rather
660    than an exception being raised.
661    
662    ----------------------------------------------------------------------
663    Name: Lal George
664    Date: 2002/02/06 13:55:02 EST
665    Tag: george-20020206-ra-breakup
666    Description:
667    
668    1. A bug fix from Allen.
669    
670        A typo causes extra fstp %st(0)'s to be generated at compensation
671        edges, which might cause stack underflow traps at runtime.  This
672        occurs in fft where there are extraneous fstps right before the 'into'
673        trap instruction (in this case they are harmless since none of the
674        integers overflow.)
675    
676    2. Pulled out various utility modules that were embedded in the modules
677       of the register allocator. I need these modules for other purposes, but
678       they are not complete enough to put into a library (just yet).
679    ----------------------------------------------------------------------
680    Name: Matthias Blume
681    Date: 2002/01/31 16:05:00 EST
682    Tag: blume-20020131-sparc-ccalls
683    Description:
684    
685    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
686       of extra stack space by mistake.  Fixed.
687    
688    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
689    
690    ----------------------------------------------------------------------
691    Name: Allen Leung
692    Date: 2002/01/30
693    Tag: leunga-20020130-nowhere-bug-fix
694    Description:
695    
696       MLRISC bug fixes:
697       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
698       2. MachineInt.fromString was negating its value.
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume
702    Date: 2002/01/29
703    Tag: blume-20020129-INSTALL
704    Description:
705    
706    - Added somewhat detailed installation instructions (file INSTALL).
707    - Fixed curl-detection bug in config/install.sh.
708    - It is now possible to select the URL getter using the URLGETTER
709      environment variable:
710    
711          not set / "unknown"      --> automatic detection (script tries wget,
712                                       curl, and lynx)
713          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
714                                       how to properly invoke them)
715          other                    --> use $URLGETTER directly, it must take
716                                       precisely two command-line arguments
717                                       (source URL and destination file name)
718    
719    ----------------------------------------------------------------------
720    Name: Matthias Blume
721    Date: 2002/01/28
722    Tag: blume-20020128-sparc-ccalls
723    Description:
724    
725    - Fixed problem with calculation of "used" registers in sparc-c-calls.
726    - Make use of the allocParam argument in sparc-c-calls.
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume
730    Date: 2002/01/28
731    Tag: blume-20020128-allocParam
732    Description:
733    
734    John Reppy:  Changes c-calls API to accept client-callback for
735    allocating extra stack space.
736    me: Corresponding changes to mlriscGen (using a dummy argument that
737        does not change the current behavior).
738    
739    ----------------------------------------------------------------------
740    Name: Matthias Blume
741    Date: 2002/01/28 12:00:00
742    Tag: Release_110_38
743    Description:
744    
745    This time for real!!!
746    
747    ----------------------------------------------------------------------
748    Name: Matthias Blume
749    Date: 2002/01/28 10:56:00 EST
750    Tag: blume-20020128-retraction
751    Description:
752    
753    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
754       with blume-Release_110_38-retracted.)
755    
756    1. Fixed a problem with incorrect rounding modes in real64.sml.
757       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
758    
759    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
760       and enums fixed.  The naming of corresponding ML identifiers should
761       now be consistent again.
762    
763    ----------------------------------------------------------------------
764    Name: Allen Leung
765    Date: 2002/01/27
766    Tag: leunga-20020127-nowhere
767    Description:
768    
769       Added a target called nowhere in the configuration scripts.
770       Enabling this will build the MLRISC 'nowhere' tool (for translating
771       programs with where-clauses into legal SML code) during installation.
772    
773    ----------------------------------------------------------------------
774    Name: Matthias Blume
775    Date: 2002/01/25 21:27:00 EST
776    Tag: blume-Release_110_38-retracted
777    Description:
778    
779    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
780    
781    README will be added later.
782    
783    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
784    (Release_110_38) removed.  Reason: Last-minute bug fixes.
785    
786    ----------------------------------------------------------------------
787    Name: Matthias Blume
788    Date: 2002/01/25
789    Tag: blume-20020125-ffi
790    Description:
791    
792    A large number of tweaks and improvements to ml-nlffi-lib and
793    ml-nlffigen:
794    
795       - ML represenation types have been streamlined
796       - getter and setter functions work with concrete values, not abstract
797         ones where possible
798       - ml-nlffigen command line more flexible (see README file there)
799       - some bugs have been fixed (hopefully)
800    
801    ----------------------------------------------------------------------
802    Name: Lal George
803    Date: 2002/01/24
804    Tag: george-20020124-risc-ra-interface
805    Description:
806    
807       There is a dramatic simplification in the interface to the
808       register allocator for RISC architectures as a result of making
809       parallel copy instructions explicit.
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume
813    Date: 2002/01/22
814    Tag: blume-20020122-x86-ccalls
815    Description:
816    
817    Bug fix for c-calls on x86 (having to do with how char- and
818    short-arguments are being handled).
819    
820    ----------------------------------------------------------------------
821    Name: Matthias Blume
822    Date: 2002/01/21
823    Tag: blume-20020121-ff
824    Description:
825    
826    Another day of fiddling with the FFI...
827    
828    1. Bug fix/workaround:  CKIT does not complain about negative array
829       dimensions, so ml-nlffigen has to guard itself against this possibility.
830       (Otherwise a negative dimension would send it into an infinite loop.)
831    
832    2. Some of the abstract types (light objects, light pointers, most "base"
833       types) in structure C are now eqtypes.
834    
835    3. Added constructors and test functions for NULL function pointers.
836    
837    ----------------------------------------------------------------------
838    Name: Matthias Blume
839    Date: 2002/01/18
840    Tag: blume-20020118-ready-for-new-release
841    Description:
842    
843    Made config/srcarchiveurl point to a new place.  (Will provide boot
844    files shortly.)
845    
846    Maybe we christen this to be 110.38?
847    
848    ----------------------------------------------------------------------
849    Name: Matthias Blume
850    Date: 2002/01/18
851    Tag: blume-20020118-more-ffifiddle
852    Description:
853    
854    Today's FFI fiddling:
855    
856      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
857        C.Set.* that use concrete (MLRep.*) arguments and results instead
858        of abstract ones.
859    
860      - Provided word-style bit operations etc. for "int" representation
861        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
862        and Long.
863    
864    ----------------------------------------------------------------------
865    Name: Matthias Blume
866    Date: 2002/01/18
867    Tag: blume-20020118-use-x86-fp
868    Description:
869    
870    Now that x86-fast-fp seems to be working, I turned it back on again
871    by default.  (Seems to work fine now, even with the FFI.)
872    
873    Other than that, I added some documentation about the FFI to
874    src/ml-nlffigen/README and updated the FFI test examples in
875    src/ml-nlffi-lib/Tests/*.
876    
877    ----------------------------------------------------------------------
878    Name: Allen Leung
879    Date: 2002/01/17
880    Tag: leunga-20020117-x86-fast-fp-call
881    Description:
882    
883       1. Fixed a problem with handling return fp values when x86's fast fp
884          mode is turned on.
885    
886       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
887          of %f32.
888    
889       3. Added a constructor INT32lit to the ast of MLRISC tools.
890    
891    ----------------------------------------------------------------------
892    Name: Matthias Blume
893    Date: 2002/01/16
894    Tag: blume-20020116-ffifiddle
895    Description:
896    
897    More fiddling with the FFI interface:
898    
899     - Make constness 'c instead of rw wherever possible.  This eliminates
900       the need for certain explicit coercions.  (However, due to ML's
901       value polymorphism, there will still be many cases where explicit
902       coercions are necessary.  Phantom types are not the whole answer
903       to modeling a subtyping relationship in ML.)
904    
905     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
906    
907     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
908       clumsy workaround for varargs problem.)
909    
910    ----------------------------------------------------------------------
911    Name: Lal George
912    Date: 2002/01/15
913    Tag: <none>
914    Description:
915    
916    1. Since COPY instructions are no longer native to the architecture,
917       a generic functor can be used to implement the expandCopies function.
918    
919    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
920       TEXT segment.
921    
922    ----------------------------------------------------------------------
923    Name: Matthias Blume
924    Date: 2002/01/15
925    Tag: blume-20020115-ffiupdates
926    Description:
927    
928    1. Fix for bug resulting in single-precision float values being returned
929       incorrectly from FFI calls.
930    
931    2. Small modifications to C FFI API:
932    
933        - memory-allocation routines return straight objects (no options)
934          and raise an exception in out-of-memory situations
935        - unsafe extensions to cast between function pointers and pointers
936          from/to ints
937        - added structure C_Debug as an alternative to structure C where
938          pointer-dereferencing (|*| and |*!) always check for null-pointers
939        - added open_lib' to DynLinkage;  open_lib' works like open_lib
940          but also takes a (possibly empty) list of existing library handles
941          that the current library depends on
942    
943    ----------------------------------------------------------------------
944    Name: Matthias Blume
945    Date: 2002/01/10
946    Tag: blume-20020110-newffigen
947    Description:
948    
949    1. Updates to portable graph code.
950    
951    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
952       more scalable now so that even huge interfaces such as the one
953       for GTK compile in finite time and space. :-)
954       See src/ml-nlffigen/README for details on what's new.
955    
956    ----------------------------------------------------------------------
957    Name: Lal George
958    Date: 2001/01/09 14:31:35 EST 2002
959    Tag: george-20011206-rm-native-copy
960    Description:
961    
962            Removed the native COPY and FCOPY instructions
963            from all the architectures and replaced it with the
964            explicit COPY instruction from the previous commit.
965    
966            It is now possible to simplify many of the optimizations
967            modules that manipulate copies. This has not been
968            done in this change.
969    
970    ----------------------------------------------------------------------
971    Name: Lal George
972    Date: 2001/12/06 16:50:13 EST 2001
973    Tag: george-20011206-mlrisc-instruction
974    Description:
975    
976    Changed the representation of instructions from being fully abstract
977    to being partially concrete. That is to say:
978    
979      from
980            type instruction
981    
982      to
983            type instr                              (* machine instruction *)
984    
985            datatype instruction =
986                LIVE of {regs: C.cellset, spilled: C.cellset}
987              | KILL of {regs: C.cellset, spilled: C.cellset}
988              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
989              | ANNOTATION of {i: instruction, a: Annotations.annotation}
990              | INSTR of instr
991    
992    This makes the handling of certain special instructions that appear on
993    all architectures easier and uniform.
994    
995    LIVE and KILL say that a list of registers are live or killed at the
996    program point where they appear. No spill code is generated when an
997    element of the 'regs' field is spilled, but the register is moved to
998    the 'spilled' (which is present, more for debugging than anything else).
999    
1000    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1001    We used to generate:
1002    
1003            DEFFREG f1
1004            f1 := f2 + f3
1005            trapb
1006    
1007    but now generate:
1008    
1009            f1 := f2 + f3
1010            trapb
1011            LIVE {regs=[f1,f2,f3], spilled=[]}
1012    
1013    Furthermore, the DEFFREG (hack) required that all floating point instruction
1014    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1015    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1016    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1017    this architecture).
1018    
1019    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1020    all the architectures. This will result in further simplification of the
1021    register allocator that must be aware of them for coalescing purposes, and
1022    will also simplify certain aspects of the machine description that provides
1023    callbacks related to parallel copies.
1024    
1025    ANNOTATION should be obvious, and now INSTR represents the honest to God
1026    machine instruction set!
1027    
1028    The <arch>/instructions/<arch>Instr.sml files define certain utility
1029    functions for making porting easier -- essentially converting upper case
1030    to lower case. All machine instructions (of type instr) are in upper case,
1031    and the lower case form generates an MLRISC instruction. For example on
1032    the alpha we have:
1033    
1034      datatype instr =
1035         LDA of {r:cell, b:cell, d:operand}
1036       | ...
1037    
1038      val lda : {r:cell, b:cell, d:operand} -> instruction
1039        ...
1040    
1041    where lda is just (INSTR o LDA), etc.
1042    
1043    ----------------------------------------------------------------------
1044    Name: Matthias Blume
1045    Date: 2001/11/22 21:40:00 EST
1046    Tag: Release_110_37
1047    Description:
1048    
1049    Release 110.37.  This time for real.
1050    
1051    ----------------------------------------------------------------------
1052    Name: Matthias Blume
1053    Date: 2001/11/21 16:35:00 EST
1054    Tag: blume-20011121-foot-in-mouth
1055    Description:
1056    
1057    Removed the "Release_110_37" tag because of a serious bug.
1058    This will be re-tagged once the bug is fixed.
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume
1062    Date: 2001/11/21 16:14:00 EST
1063    Tag: blume-20011121-forgottenfile
1064    Description:
1065    
1066    Forgot to add a file.  (Just a .tex-file -- part of
1067    the CM manual source.)
1068    
1069    ----------------------------------------------------------------------
1070    Name: Matthias Blume
1071    Date: 2001/11/21 16:10:00 EST
1072    Tag: blume-20011121-invalid_110_37
1073    Description:
1074    
1075    Note: I removed the original tag "Release_110_37" from this commit
1076          because we found a serious bug in all non-x86 backends.
1077          - Matthias
1078    
1079    1. Modifications to the SML/NJ code generator and to the runtime system
1080       so that code object name strings are directly inserted into code
1081       objects at code generation time.  The only business the runtime system
1082       has with this is now to read the name strings on occasions.
1083       (The encoding of the name string has also changed somewhat.)
1084    
1085    2. CM now implements a simple "set calculus" for specifying export lists.
1086       In particular, it is now possible to refer to the export lists of
1087       other libraries/groups/sources and form unions as well as differences.
1088       See the latest CM manual for details.
1089    
1090    3. An separate notion of "proxy" libraries has again be eliminated from
1091       CM's model.  (Proxy libraries are now simply a special case of using
1092       the export list calculus.)
1093    
1094    4. Some of the existing libraries now take advantage of the new set
1095       calculus.
1096       (Notice that not all libraries have been converted because some
1097       of the existing .cm-files are supposed to be backward compatible
1098       with 110.0.x.)
1099    
1100    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
1101       "General.exnMessage"!  The former relies on a certain hook to be
1102       initialized, and that often does not happen in the stand-alone case.)
1103    
1104    ----------------------------------------------------------------------
1105    Name: Lal George
1106    Date: 2001/11/21  13:56:18 EST
1107    Tag: george-2001121-pseudo-ops
1108    Description:
1109    
1110      Implemented a complete redesign of MLRISC pseudo-ops. Now there
1111      ought to never be any question of incompatabilities with
1112      pseudo-op syntax expected by host assemblers.
1113    
1114      For now, only modules supporting GAS syntax are implemented
1115      but more should follow, such as MASM, and vendor assembler
1116      syntax, e.g. IBM as, Sun as, etc.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume
1120    Date: 2001/11/14 11:52:00 EST
1121    Tag: blume-20011114-srcname
1122    Description:
1123    
1124    1. Routed the name of the current source file to mlriscgen where it
1125       should be directly emitted into the code object.  (This last part
1126       is yet to be done.)
1127    
1128    2. Some cleanup of the pgraph code to make it match the proposal that
1129       I put out the other day.  (The proposal notwithstanding, things are
1130       still in flux here.)
1131    
1132    ----------------------------------------------------------------------
1133    Name: Lal George
1134    Date: 2001/11/14 09:44:04 EST
1135    Tag:
1136    Description:
1137    
1138      Fix for a backpatching bug reported by Allen.
1139    
1140      Because the boundary between short and long span-dependent
1141      instructions is +/- 128, there are an astounding number of
1142      span-dependent instructions whose size is over estimated.
1143    
1144      Allen came up with the idea of letting the size of span
1145      dependent instructions be non-monotonic, for a maxIter
1146      number of times, after which the size must be monotonically
1147      increasing.
1148    
1149      This table shows the number of span-dependent instructions
1150      whose size was over-estimated as a function of maxIter, for the
1151      file Parse/parse/ml.grm.sml:
1152    
1153         maxIter            # of instructions:
1154            10                      687
1155            20                      438
1156            30                      198
1157            40                        0
1158    
1159      In compiling the compiler, there is no significant difference in
1160      compilation speed between maxIter=10 and maxIter=40. Actually,
1161      my measurements showed that maxIter=40 was a tad faster than
1162      maxIter=10! Also 96% of the  files in the compiler reach a fix
1163      point within 13 iterations, so fixing maxIter at 40, while high,
1164      is okay.
1165    
1166    ----------------------------------------------------------------------
1167    Name: Matthias Blume
1168    Date: 2001/10/31 15:25:00 EST
1169    Tag: blume-20011031-pgraph
1170    Description:
1171    
1172    CKIT:
1173    * Changed the "Function" constructor of type Ast.ctype to carry optional
1174      argument identifiers.
1175    * Changed the return type of TypeUtil.getFunction accordingly.
1176    * Type equality ignores the argument names.
1177    * TypeUtil.composite tries to preserve argument names but gives up quickly
1178      if there is a mismatch.
1179    
1180    installation script:
1181    * attempts to use "curl" if available (unless "wget" is available as well)
1182    
1183    CM:
1184    * has an experimental implementation of "portable graphs" which I will
1185      soon propose as an implementation-independent library format
1186    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1187    
1188    NLFFI-LIB:
1189    * some cleanup (all cosmetic)
1190    
1191    NLFFIGEN:
1192    * temporarily disabled the mechanism that suppresses ML output for
1193      C definitions whose identifiers start with an underscore character
1194    * generate val bindings for enum constants
1195    * user can request that only one style (light or heavy) is being used;
1196      default is to use both (command-line arguments: -heavy and -light)
1197    * fixed bug in handling of function types involving incomplete pointers
1198    * generate ML entry points that take record arguments (i.e., using
1199      named arguments) for C functions that have a prototype with named
1200      arguments
1201      (see changes to CKIT)
1202    
1203    ----------------------------------------------------------------------
1204    Name: Allen Leung
1205    Date: 2001/10/27 20:34:00 EDT
1206    Tag: leunga-20011027-x86-fast-fp-call
1207    Description:
1208    
1209       Fixed the bug described in blume-20010920-slowfp.
1210    
1211       The fix involves
1212          1. generating FCOPYs in FSTP in ia32-svid
1213          2. marking a CALL with the appropriate annotation
1214    
1215    ----------------------------------------------------------------------
1216    Name: Matthias Blume
1217    Date: 2001/10/16 11:32:00 EDT
1218    Tag: blume-20011016-netbsd
1219    Description:
1220    
1221    Underscore patch from Chris Richards (fixing problem with compiling
1222    runtime system under recent NetBSD).
1223    
1224    ----------------------------------------------------------------------
1225    Name: Allen Leung
1226    Date: 2001/10/12 17:18:32 EDT 2001
1227    Tag: leung-20011012-x86-printflowgraph
1228    Description:
1229    
1230    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1231    
1232    ----------------------------------------------------------------------
1233    Name: Lal George
1234    Date: 2001/10/11 23:51:34 EDT
1235    Tag: george-20011011-too-many-instrs
1236    Description:
1237    
1238    The representation of a program point never expected to see more
1239    than 65536 instructions in a basic block!
1240    
1241    ----------------------------------------------------------------------
1242  Name: Lal George  Name: Lal George
1243  Date: 2001/10/09 09:41:37 EDT  Date: 2001/10/09 09:41:37 EDT
1244  Tag: george-20011008-mlrisc-labels  Tag: george-20011008-mlrisc-labels
# Line 71  Line 1297 
1297    
1298  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1299  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
1300  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
1301  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
1302  Description:  Description:

Legend:
Removed from v.952  
changed lines
  Added in v.1127

root@smlnj-gforge.cs.uchicago.edu
ViewVC Help
Powered by ViewVC 1.0.0