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 838, Tue Jun 5 19:10:21 2001 UTC revision 843, Tue Jun 19 21:53:04 2001 UTC
# Line 14  Line 14 
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2001/06/19 17:55:00 EDT
18    Tag: blume-20010619-instantiate
19    Description:
21    This un-breaks the fix for bug 1432.
22    (The bug was originally fixed in 110.9 but I broke it again some
23    time after that.)
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2001/06/19 17:25:00 EDT
28    Tag: blume-20010619-signals
29    Description:
31    This should (hopefully) fix the long-standing signal handling bug.
32    (The runtime system was constructing a continuation record with an
33    incorrect descriptor which would cause the GC to drop data on the floor...)
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2001/06/15 15:05:00 EDT
38    Tag: blume-20010615-moresparc
39    Description:
41    Here is a short late-hour update related to Sparc c-calls:
43     -- made handling of double-word arguments a bit smarter
45     -- instruction selection phase tries to collapse certain clumsily
46        constructed ML-Trees; typical example:
48            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
50        This currently has no further impact on SML/NJ since mlriscGen does
51        not seem to generate such patterns in the first place, and c-calls
52        (which did generate them in the beginning) has meanwhile been fixed
53        so as to avoid them as well.
55    ----------------------------------------------------------------------
56    Name: Matthias Blume
57    Date: 2001/06/15 15:05:00 EDT
58    Tag: blume-20010615-sparc
59    Description:
61    The purpose of this update is to provide an implementation of NLFFI
62    on Sparc machines.
64    Here are the changes in detail:
66       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
67       the Sparc implementation of the c-calls API.
68       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
69       for accessing the ML frame.  Thus, we have a real frame pointer and
70       can freely modify %sp without need for an omit-frame-ptr phase.
71       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
72       this case.
73       * ml-nlffigen has been taught to produce code for different architectures
74       and calling conventions.
75       * In a way similar to what was done in the x86 case, the Sparc
76       backend uses its own specific extension to mltree.  (For example,
77       it needs to be able to generate UNIMP instructions which are part
78       of the calling convention.)
79       * ml-nlffi-lib was reorganized to make it more modular (in particular,
80       to make it easier to plug in new machine- and os-dependent parts).
82    There are some other fairly unrelated bug fixes and cleanups as well:
84       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
85       that they properly share their libraries with existing SML/NJ libraries.
86       * I fixed a minor cosmetic bug in CM, supressing certain spurious
87       follow-up error messages.
88       * Updates to CM/CMB documentation.
90    TODO items:
92       * MLRISC should use a different register as its asmTemp on the Sparc.
93         (The current %o2 is a really bad choice because it is part of the
94         calling conventions, so things might interfere in unexpected ways.)
96    ----------------------------------------------------------------------
97    Name: Matthias Blume
98    Date: 2001/06/07
99    Tag: blume-20010607-calls
100    Description:
102    A number of internal changes related to C calls and calling conventions:
104    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
105       number of bytes popped implicitly (by the callee).  In most cases
106       this field is 0 but on x86/win32 it is some non-zero value.  This
107       is information provided for the benefit of the "omit-frameptr" pass.
108    2. The CALL instruction on the x86 carries a similar "pops" field.
109       The instruction selection phase copies its value from the ML-Tree
110       CALL statement.
111    3. On all other architectures, the instruction selection phase checks
112       whether "pops=0" and complains if not.
113    4. The c-calls implementation for x86 now accepts two calling conventions:
114       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
115       up after the call and pops is set to 0.  For "stdcall", the caller
116       does nothing, leaving the cleanup to the callee; pops is set to
117       the number of bytes that were pushed onto the stack.
118    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
119       distinguish between "ccall" and "stdcall".
120    6. The UNIMP instruction has been added to the supported Sparc instruction
121       set. (This is needed for implementing the official C calling convention
122       on this architecture.)
123    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
124       work with the latest CM.
126    ----------------------------------------------------------------------
127    Name: Matthias Blume
128  Date: 2001/06/05 15:10:00 EDT  Date: 2001/06/05 15:10:00 EDT
129  Tag: blume-20010605-cm-index  Tag: blume-20010605-cm-index
130  Description:  Description:

Removed from v.838  
changed lines
  Added in v.843

ViewVC Help
Powered by ViewVC 1.0.0