Commit graph

64 commits

Author SHA1 Message Date
Gareth Rees
f6901a9e23 Complete design.mps.type and design.mps.interface-c and move them to "current" design in the manual.
Remove unused enumeration FormatVariety from mpmtypes.
Improve wording of manual entry on MPS_RM_PROT.

Copied from Perforce
 Change: 185259
 ServerID: perforce.ravenbrook.com
2014-04-06 20:00:31 +01:00
Richard Brooksby
6e46485dce Catch-up merge from master, mainly to pick up gcbench fixes.
Copied from Perforce
 Change: 184504
 ServerID: perforce.ravenbrook.com
2014-02-25 01:29:12 +00:00
Richard Brooksby
ba1a4f78cb Declaring result codes using a relational macro, so that tests can print more meaningful output, and to eliminate redundant code.
Copied from Perforce
 Change: 184382
 ServerID: perforce.ravenbrook.com
2014-02-19 11:55:43 +00:00
Richard Brooksby
4dff43a172 Catch-up merge from master, mainly to pick up spare-ring merge.
Copied from Perforce
 Change: 184365
 ServerID: perforce.ravenbrook.com
2014-02-18 13:47:49 +00:00
Richard Brooksby
199a1261ed Catch-up merge from master, mainly to pick up chain-zones merge.
Copied from Perforce
 Change: 184343
 ServerID: perforce.ravenbrook.com
2014-02-14 15:23:36 +00:00
Richard Brooksby
2b5e126c48 Further tidying up of the page descriptor and naming.
Copied from Perforce
 Change: 184323
 ServerID: perforce.ravenbrook.com
2014-02-11 14:14:16 +00:00
Richard Brooksby
142063f059 Abolishing vmarenaalloc and its horrible relatives.
Removing redundant SegPref structures from pools that only ever used the default.
Abolishing the VMNZ arena class, to be replaced with an option to ArenaCreate.

Copied from Perforce
 Change: 184293
 ServerID: perforce.ravenbrook.com
2014-02-04 18:10:53 +00:00
Richard Brooksby
35ccb414a2 Adding arena extension back to the arena allocation policy.
Copied from Perforce
 Change: 184280
 ServerID: perforce.ravenbrook.com
2014-01-31 22:47:29 +00:00
Richard Brooksby
d206bd35b9 Changes from review comments made in <https://info.ravenbrook.com/mail/2014/01/31/13-40-25/0/>.
ArenaSpareCommitExceeded changed to ArenaPurgeSpare so that the arena can properly reduce the amount of spare memory when setting the commit limit etc.
pagesMarkAllocated promptly unmaps page table pages in its error path.
Miscellaneous clarifications and extra checking.

Copied from Perforce
 Change: 184267
 ServerID: perforce.ravenbrook.com
2014-01-31 15:21:21 +00:00
Richard Brooksby
070f7784dd Refactored implementation of preferenced allocation policy from vm arena to general purpose arena.
Eliminating SegPrefExpress in most places, especially where there wasn’t really any preference.
Eliminating special case knowledge about garbage collection from the arena.

Copied from Perforce
 Change: 184279
 ServerID: perforce.ravenbrook.com
2014-01-30 18:08:10 +00:00
Richard Brooksby
d0a93e6caf Merging from branch //info.ravenbrook.com/project/mps/branch/2014-01-29/mps-chain-zones/... in order to eliminate generational considerations from the vm arena's placement policy.
Copied from Perforce
 Change: 184233
 ServerID: perforce.ravenbrook.com
2014-01-30 15:29:46 +00:00
Richard Brooksby
d8e81ae3c3 Abolishing the redundant generation to zone map in the vm arena, and replacing it with chainalloc, which allocates segments within a generation.
Generations from different chains will no longer share the same zone sets, and will not be collected together.  To get this behaviour, use the same chain, as documented.

Copied from Perforce
 Change: 184213
 ServerID: perforce.ravenbrook.com
2014-01-29 12:54:05 +00:00
Richard Brooksby
7b54fc9926 Creating and tearing down the arena freecbs that will hold free address ranges.
Copied from Perforce
 Change: 184049
 ServerID: perforce.ravenbrook.com
2014-01-18 16:25:08 +00:00
Gareth Rees
bcfad58d3c Merge custom cet mainline into the master sources. this includes fixes for job003519 (clock values in text telemetry logs on windows have top 32 bits zero), job003536 (cet office can't replicate 100k snowmen in reasonable time), job003651 (stack overflows are not reliably handled).
Copied from Perforce
 Change: 183926
 ServerID: perforce.ravenbrook.com
2014-01-10 14:16:08 +00:00
Richard Brooksby
f9724af3e0 Catch-up merge from custom/cet/main to master, to pick up arena extension callbacks.
Copied from Perforce
 Change: 182701
 ServerID: perforce.ravenbrook.com
2013-06-12 15:33:49 +01:00
Gareth Rees
4753e1ba3a Improve design docs following review from rb.
Copied from Perforce
 Change: 182606
 ServerID: perforce.ravenbrook.com
2013-06-07 14:50:22 +01:00
Gareth Rees
e7559cf22f Rename the enumeration cbsfinddelete to finddelete in anticipation of it being shared between cbs and new freelist module.
Since ABQDisposition enumeration only has two values, it's better to use Bool.

Copied from Perforce
 Change: 182347
 ServerID: perforce.ravenbrook.com
2013-05-31 00:09:33 +01:00
Richard Brooksby
1ab2cfe210 Eliminating unnecessary use of varargs in segment split and merge functions.
Copied from Perforce
 Change: 181842
 ServerID: perforce.ravenbrook.com
2013-05-16 03:02:20 +01:00
Richard Brooksby
a9ed62e39a Converting varags on buffer and ap creation into keyword argument lists.
Copied from Perforce
 Change: 181680
 ServerID: perforce.ravenbrook.com
2013-05-09 18:25:53 +01:00
Richard Brooksby
8b430778b6 Replacing varargs passed to seginit with keyword argument list.
Copied from Perforce
 Change: 181679
 ServerID: perforce.ravenbrook.com
2013-05-09 17:34:01 +01:00
Richard Brooksby
5751150183 Moving arena varags parsing into arena class method, similar to pools. much simpler. all tests pass.
Copied from Perforce
 Change: 181643
 ServerID: perforce.ravenbrook.com
2013-05-08 23:00:00 +01:00
Richard Brooksby
e67939cabb Pushing varargs decoding into a pool class method. much simpler.
Copied from Perforce
 Change: 181642
 ServerID: perforce.ravenbrook.com
2013-05-08 22:43:51 +01:00
Richard Brooksby
a8afd8e6f4 Converting pools to use keyword arguments, mostly, and so far inconsistently, but at least it compiles.
Copied from Perforce
 Change: 181635
 ServerID: perforce.ravenbrook.com
2013-05-08 16:21:12 +01:00
Richard Brooksby
098bf44a5e First draft of keyword arguments. mainly checking in in order to try working with git fusion on this branch.
Copied from Perforce
 Change: 181538
 ServerID: perforce.ravenbrook.com
2013-05-02 17:55:07 +01:00
Nick Barnes
292d0108a5 change mps_clock_t to word.
Copied from Perforce
 Change: 180151
 ServerID: perforce.ravenbrook.com
2012-10-30 08:43:39 +00:00
Richard Brooksby
6a71263677 Merging recent improvements from custom/cet/main to masters, but carefully excluding inclusion of configura-specific modules in mps.c.
Copied from Perforce
 Change: 179473
 ServerID: perforce.ravenbrook.com
2012-09-13 17:33:36 +01:00
Richard Brooksby
eeef8ade97 Eliminating interface puns that break the strict aliasing rule. the last of the structure puns and casts, i think.
Copied from Perforce
 Change: 179453
 ServerID: perforce.ravenbrook.com
2012-09-12 20:00:04 +01:00
Richard Brooksby
9f3c6c49bb Further elimination of type puns that violate the strict aliasing rule. this change eliminates puns on the format methods.
Copied from Perforce
 Change: 179452
 ServerID: perforce.ravenbrook.com
2012-09-12 19:45:27 +01:00
Richard Brooksby
1a103e0a5a Further reduction of interface puns that violate the strict aliasing rule, this time in thread function types and the corresponding threadstruct.
Copied from Perforce
 Change: 179451
 ServerID: perforce.ravenbrook.com
2012-09-12 19:32:37 +01:00
Richard Brooksby
dd039bfbdc Fixing bugs introduced by type puns passed through scanning methods, discovered by offsetting mps_ss_s from the beginning of scanstatestruct.
Adding a pseudo-target "find-puns" which, on FreeBSD mostly, prints useful warnings about strict aliasing rule violations.
Fixing more strict aliasing rule violations found by the above.

Copied from Perforce
 Change: 179358
 ServerID: perforce.ravenbrook.com
2012-09-07 23:48:14 +01:00
Richard Brooksby
f748ad2b28 Eliminating type puns on scan states, location dependencies, and allocation points through the mps interface.
Now that we're recommending inlining with client code and optimising with -O2 or -O3, we can't afford any bug introduced by the strict aliasing rule.

Copied from Perforce
 Change: 179322
 ServerID: perforce.ravenbrook.com
2012-09-07 12:58:57 +01:00
Richard Brooksby
4c51d72b56 Partial merge of branch/2012-07-23/cet-transform, excluding cet-specific parts.
Copied from Perforce
 Change: 179309
 ServerID: perforce.ravenbrook.com
2012-09-06 17:17:18 +01:00
Richard Brooksby
eb9e12cacb All tests listed under the "all" target in gc.gmk are now building, and all but sacss do not find defects.
Copied from Perforce
 Change: 178321
 ServerID: perforce.ravenbrook.com
2012-05-24 17:39:13 +01:00
Richard Brooksby
82c29cfd3e Fixing up occurences of "unsigned long" in the main mps and test code so that we no longer assume that "unsigned long" is large enough for a pointer or size, or that it's the longest available integer type, since these are not true on 64-bit windows.
Copied from Perforce
 Change: 178021
 ServerID: perforce.ravenbrook.com
2012-04-23 16:20:40 +01:00
Richard Brooksby
98f3e3bdd2 Consistently passing closure sizes as size_t and not casting back and forth to size or anything else.
Copied from Perforce
 Change: 177817
 ServerID: perforce.ravenbrook.com
2012-03-27 18:12:52 +01:00
Richard Brooksby
49c6d6b2f9 Introducing ulongest type. see design/types/#ulongest.
Fixing WriteF to take ULongest rather than unsigned long, and fixing intrnal assumption that Word was as big as unsigned long.

Copied from Perforce
 Change: 177814
 ServerID: perforce.ravenbrook.com
2012-03-27 16:55:31 +01:00
Richard Kistruck
33d05b4195 mps br/vmem: simple-chunk-return:
arenavm.c -- move chunk-return into new function "VMCompact".
 (also, in VMArenaFinish, null out arena->primary in VMArenaFinish, so it is not left dangling).

 arena.c, arenavm.c, mpm.h, mpmst.h, mpmtypes.h:
 arena->class->compact: ArenaCompact, ArenaTrivCompact, VMCompact

 trace.c -- traceReclaim calls ArenaCompact!

Copied from Perforce
 Change: 170095
 ServerID: perforce.ravenbrook.com
2010-03-23 01:49:17 +00:00
Richard Kistruck
beb476fd52 Mps br/timing: traceanc.c -- all this ancillary stuff was making
trace.c very cluttered.  Put it here instead.

Copied from Perforce
 Change: 166911
 ServerID: perforce.ravenbrook.com
2008-12-09 16:55:22 +00:00
Richard Kistruck
2b4cf6ca29 Mps br/timing: new type "clock", internal equivalent of mps_clock_t.
Used for a field in messages.  (More uses to follow).

Copied from Perforce
 Change: 166697
 ServerID: perforce.ravenbrook.com
2008-11-11 16:57:30 +00:00
David Jones
e48dcde8fe Mps: removing unused traceband stuff introduced by
branch/2007-04-18/diag/...
       This was merged onto master by change 162839, but I, drj, felt
       it was simpler to remove it in a separate change.

Copied from Perforce
 Change: 162841
 ServerID: perforce.ravenbrook.com
2007-07-10 14:42:11 +01:00
Richard Kistruck
16ddbe30bc Mps br/diag: trace: new field in tracestruct: traceband band for
current band (eg via-F) of trace.  Fix for job001658.

Copied from Perforce
 Change: 162696
 ServerID: perforce.ravenbrook.com
2007-06-29 17:17:00 +01:00
Richard Kistruck
b44832b7f6 Mps br/diag: rename #def "diagnostics" as "statistics", as all the
macros it controls are called "STATISTIC_blah", and same replacement 
in comments.  The term "diagnostic" means something else now: see 
design/diag/.

Copied from Perforce
 Change: 162692
 ServerID: perforce.ravenbrook.com
2007-06-29 14:08:00 +01:00
Richard Kistruck
06d906a04b Mps mpmtypes.h (comment only) messagetypegcstart matches mps_message_type_gc_start
Copied from Perforce
 Change: 161205
 ServerID: perforce.ravenbrook.com
2006-12-12 10:06:27 +00:00
Richard Kistruck
6e7b73614b Mps integrate gcgenmsg into master:
- new mps_message_type_gc_start() explains what triggered a collection;
  - design/message: add guide.

Copied from Perforce
 Change: 161204
 ServerID: perforce.ravenbrook.com
2006-12-12 09:56:09 +00:00
Richard Kistruck
1999a785c5 Mps gcgenmsg: 2006 copyrights
Copied from Perforce
 Change: 161202
 ServerID: perforce.ravenbrook.com
2006-12-12 08:47:10 +00:00
Richard Kistruck
908c81804f Mps _gc_start: make tracestartmessage reasons specific trigger names, not broad categories.
Make _why text more explanatory.

Copied from Perforce
 Change: 161196
 ServerID: perforce.ravenbrook.com
2006-12-11 16:17:31 +00:00
Richard Kistruck
17c26e4657 Mps: (comment only) fix typo, and clarify messagetypelimit is not a message type
Copied from Perforce
 Change: 161191
 ServerID: perforce.ravenbrook.com
2006-12-11 15:33:11 +00:00
Richard Kistruck
4b5383fbc4 Mps: (comments only) add cross-refs from mpmtypes.h to must-match identifiers in mps.h: eg. mps_message_type_gc
Copied from Perforce
 Change: 155507
 ServerID: perforce.ravenbrook.com
2005-10-21 14:51:26 +01:00
David Jones
eb7fb7e384 Mps: added (untested) internals for messagetypegcstart
Copied from Perforce
 Change: 39768
 ServerID: perforce.ravenbrook.com
2003-02-19 13:27:16 +00:00
Richard Brooksby
33db57fe2e Fixed various tagging and cross-referencing errors.
Copied from Perforce
 Change: 30455
 ServerID: perforce.ravenbrook.com
2002-06-24 13:16:30 +01:00