mirrors/emacs
1
Fork 0
mirror of git://git.sv.gnu.org/emacs.git synced 2026-03-24 07:41:54 -07:00
Code Issues Projects Releases Wiki Activity
6319 commits 248 branches 194 tags 1.1 GiB
Commit graph

2414 commits

Author SHA1 Message Date
Gareth Rees
bcfbaaf3e4 Initialization and checking of mutatorcontext data structures. 
New files prmcix.c and prmcxc.c avoid duplicated code.

Copied from Perforce
 Change: 192562
 ServerID: perforce.ravenbrook.com
 2016-10-13 23:13:40 +01:00
Gareth Rees
d82fc68bde Use protsgix.c on linux and delete protli.c. 
Copied from Perforce
 Change: 192556
 ServerID: perforce.ravenbrook.com
 2016-10-13 21:24:04 +01:00
Gareth Rees
1d65260d5b In protsgix.c, construct a mutatorcontext object and pass it to arenaaccess. 
Copied from Perforce
 Change: 192550
 ServerID: perforce.ravenbrook.com
 2016-10-13 20:12:32 +01:00
Gareth Rees
6903784501 Fix "make testmmqa" on freebsd (the shell doesn't do {a,b,c} expansion). 
Copied from Perforce
 Change: 192545
 ServerID: perforce.ravenbrook.com
 2016-10-13 19:31:25 +01:00
Gareth Rees
626f880a7e Fix comment: ll = clang/llvm, not gcc. 
Copied from Perforce
 Change: 192540
 ServerID: perforce.ravenbrook.com
 2016-10-13 19:15:27 +01:00
Gareth Rees
fe1f07321b Fix makefile, accidentally broken by change 192512. 
Copied from Perforce
 Change: 192539
 ServerID: perforce.ravenbrook.com
 2016-10-13 19:13:31 +01:00
Gareth Rees
d31012f13a Implement mutatorcontextsp and mutatorcontextscan for platforms w3i3 and w3i6. 
This means that ThreadScan becomes identical on these two platforms and can be moved to thw3.c.
This means that thw3.h, thw3i3.c and thw3i6.c become redundant and can be deleted.

Copied from Perforce
 Change: 192534
 ServerID: perforce.ravenbrook.com
 2016-10-13 16:28:50 +01:00
Gareth Rees
91551170b1 Remove return statements that have no effect. 
Copied from Perforce
 Change: 192529
 ServerID: perforce.ravenbrook.com
 2016-10-13 15:17:50 +01:00
Gareth Rees
5c31ca3cab Rename the "protection mutator context" module to "mutator context" (this module handles mutator context decoding for both the protection module and the thread module). 
Rename functions Prot{Can,}StepInstruction to MutatorContext{Can,}StepInstruction so that they follow the naming convention in guide.implc.naming.prefix.program.
Move mutator context declarations out of prot.h into new header prmc.h.
Correct .assume.null in a couple of places -- it's not safe for MutatorContextStepInstruction to return ResUNIMPL, instead MutatorContextCanStepInstruction should return FALSE.

Copied from Perforce
 Change: 192528
 ServerID: perforce.ravenbrook.com
 2016-10-13 15:06:14 +01:00
Gareth Rees
c35e544d73 Rename mutatorfaultcontext to mutatorcontext because this data structure is not only used to store the context of a fault, but also to store the context of a thread that has been suspended. 
Copied from Perforce
 Change: 192523
 ServerID: perforce.ravenbrook.com
 2016-10-13 14:23:39 +01:00
Gareth Rees
5a580660f6 Remove undocumented global mps_exception_info. now that the mps is open source, there is no difficulty in debugging the mutator context. 
Copied from Perforce
 Change: 192518
 ServerID: perforce.ravenbrook.com
 2016-10-13 13:45:31 +01:00
Gareth Rees
8b002dc4d5 Fix testmmqa target in the hot variety. 
Copied from Perforce
 Change: 192515
 ServerID: perforce.ravenbrook.com
 2016-10-13 13:31:58 +01:00
Gareth Rees
3dbe56a8b7 Rename prot{i3,i6}.c to prmc{i3,i6}.c because these files are part of the protection mutator context module. 
Rename prmc{i3,i6}{fr,li,w3,xc}.c to prmc{fr,li,w3,xc}{i3,i6}.c for consistency of ordering of platform codes (OS before AR before CT).

Copied from Perforce
 Change: 192512
 ServerID: perforce.ravenbrook.com
 2016-10-13 13:24:47 +01:00
Gareth Rees
904ac5beb6 Revert incorrect change to globalspreparetodestroy (see change 192365) that made the mps assert when destroying an arena with uncollected finalization messages. 
Copied from Perforce
 Change: 192473
 ServerID: perforce.ravenbrook.com
 2016-09-27 11:57:19 +01:00
Gareth Rees
d44d50dfd0 The mmqa runset command now takes multiple test-sets. 
Copied from Perforce
 Change: 192418
 ServerID: perforce.ravenbrook.com
 2016-09-14 16:24:34 +01:00
Gareth Rees
2b3d00c1ee Update following release 1.116.0. 
Copied from Perforce
 Change: 192406
 ServerID: perforce.ravenbrook.com
 2016-09-14 14:18:17 +01:00
Gareth Rees
4d5ae93957 Don't set stackprobedepth to 500 if config_pf_ansi is defined -- this leads to an assertion failure in span.c. 
Copied from Perforce
 Change: 192393
 ServerID: perforce.ravenbrook.com
 2016-09-14 13:03:27 +01:00
Gareth Rees
3b03393381 Account allocations in amc segments between whiten and reclaim as condemned, improving the accuracy of the mortality calculation and avoiding assertion failure. 
Copied from Perforce
 Change: 192387
 ServerID: perforce.ravenbrook.com
 2016-09-14 12:49:51 +01:00
Gareth Rees
5f86690a08 Update mps_release in preparation for forthcoming release 1.116.0. 
Copied from Perforce
 Change: 192370
 ServerID: perforce.ravenbrook.com
 2016-09-13 17:43:39 +01:00
Gareth Rees
4cc81bb371 Merge branch/2016-09-06/job004006. 
Copied from Perforce
 Change: 192365
 ServerID: perforce.ravenbrook.com
 2016-09-13 17:32:07 +01:00
Gareth Rees
63a44e3a91 Fix problems noted in review. 
Copied from Perforce
 Change: 192357
 ServerID: perforce.ravenbrook.com
 2016-09-13 17:19:07 +01:00
Gareth Rees
3cd7835844 Catch-up merge from custom/cet/branch/2016-09-13/job004006 to branch/2016-09-06/job004006. 
Copied from Perforce
 Change: 192351
 ServerID: perforce.ravenbrook.com
 2016-09-13 15:33:31 +01:00
Gareth Rees
3f13486e08 Usability improvements for mmqa: 
* -i option (MPS_INCLUDE_DIR) now defaults to the code directory relative to the test harness location.
* -p option (PLATFORM) is now a six-character platform code for consistency with the rest of the MPS.
* New -v option (VARIETY) for specifying cool/hot/rash.
* -l option (MPS_LINK_OBJ) now defaults to the appropriate object file for the platform and variety.
* The object file is built automatically.
* If $Id$ of a test file is longer than 70 characters, cut characters out of the middle rather than off the end.
* Eliminate obsolete option -prod (MPS_PRODUCT).
* Support for FreeBSD.

Copied from Perforce
 Change: 192285
 ServerID: perforce.ravenbrook.com
 2016-09-12 11:55:36 +01:00
Gareth Rees
8340e390a5 In order to be able to test mps_arena_busy in all build configurations, we must always take and release the arena lock, even in the single-threaded case. we gain some additional checking by doing this, at low cost (since in this build configuration the locks are generic and so just increment/decrement a count of claims). 
In the CONFIG_POLL_NONE we configuration, we would still like to check that no traces are busy when leaving the arena, but since we now call ArenaLeave in this configurations, move the assertion to ShieldLeave.
In ArenaDestroy, call ArenaLeave, don't just release the lock.

Copied from Perforce
 Change: 192267
 ServerID: perforce.ravenbrook.com
 2016-09-09 13:01:25 +01:00
Gareth Rees
7869db227e Improve documentation. 
Copied from Perforce
 Change: 192258
 ServerID: perforce.ravenbrook.com
 2016-09-09 11:18:12 +01:00
Gareth Rees
cf6c484a89 Lockisheld implementation for generic ("ansi") locks. 
When CONFIG_THREAD_SINGLE is defined, use the generic lock module rather than compiling out all lock calls via lock.h. (Otherwise test cases that check LockIsHeld will fail.)

Copied from Perforce
 Change: 192254
 ServerID: perforce.ravenbrook.com
 2016-09-09 11:01:04 +01:00
Gareth Rees
d8c32f0d3f New functions mps_arena_busy and mps_arena_postmortem, plus test coverage. 
Copied from Perforce
 Change: 192247
 ServerID: perforce.ravenbrook.com
 2016-09-08 13:55:43 +01:00
Gareth Rees
8523b6f666 Remove support for linux 2.4 and 2.5 (these versions used linuxthreads). linux is now supported from version 2.6 onwards (these versions use posix threads). 
Copied from Perforce
 Change: 192240
 ServerID: perforce.ravenbrook.com
 2016-09-08 11:36:02 +01:00
Gareth Rees
7473718e0a Catch-up merge from master sources to branch/2016-04-13/mortality. 
Copied from Perforce
 Change: 192192
 ServerID: perforce.ravenbrook.com
 2016-09-05 16:17:04 +01:00
Gareth Rees
331ec20beb Catch-up merge from master sources to branch/2016-04-12/job004000. 
Copied from Perforce
 Change: 192187
 ServerID: perforce.ravenbrook.com
 2016-09-05 16:01:17 +01:00
Gareth Rees
90a16b56b1 Catch-up merge from master sources to branch/2016-04-21/amswalk. 
Copied from Perforce
 Change: 192181
 ServerID: perforce.ravenbrook.com
 2016-09-05 15:48:47 +01:00
Gareth Rees
c07c5d6b8f Catch-up merge from the master sources to branch/2016-04-23/inst-methods. 
Copied from Perforce
 Change: 192176
 ServerID: perforce.ravenbrook.com
 2016-09-05 15:31:34 +01:00
Gareth Rees
baf2277b30 Catch-up merge from the master sources. 
Copied from Perforce
 Change: 192170
 ServerID: perforce.ravenbrook.com
 2016-09-05 15:21:33 +01:00
Gareth Rees
18e0381952 When injecting allocation failures to test the fail-over land, don't also inject failures into the arena's free land's block pool. 
Copied from Perforce
 Change: 192164
 ServerID: perforce.ravenbrook.com
 2016-09-05 13:17:59 +01:00
Gareth Rees
b9b1ad9074 Correct the logic in mvfreesize -- lost spans are not accounted as free. only sum the free space in the spans in checking varieties, otherwise just return the cached sum. 
Copied from Perforce
 Change: 192159
 ServerID: perforce.ravenbrook.com
 2016-09-05 12:03:46 +01:00
Gareth Rees
d7e36f8dbb Don't pick a grain size that's so large that you can't avoid hitting the commit limit. 
Copied from Perforce
 Change: 192154
 ServerID: perforce.ravenbrook.com
 2016-09-05 11:37:19 +01:00
Gareth Rees
49198a35ee Add _critical to assertions on the critical path for manual-allocation-bound programs using mvff. this brings the hot/rash ratio for "djbench mvff" down from over 150% to about 130%. 
Copied from Perforce
 Change: 192148
 ServerID: perforce.ravenbrook.com
 2016-09-04 19:20:05 +01:00
Gareth Rees
4dd6e899f5 Build and test the toy scheme interpreter during continuous integration, to reduce the risk of accidentally breaking it. 
Copied from Perforce
 Change: 192133
 ServerID: perforce.ravenbrook.com
 2016-09-04 14:47:08 +01:00
Gareth Rees
8005cb5f77 Initialize in/out parameter old_exception_count before passing it to thread_swap_exception_ports. 
Copied from Perforce
 Change: 192128
 ServerID: perforce.ravenbrook.com
 2016-09-04 14:28:28 +01:00
Gareth Rees
0400178bf4 Support registering a thread with the mps multiple times on os x. 
Restore test case.

Copied from Perforce
 Change: 192127
 ServerID: perforce.ravenbrook.com
 2016-09-04 14:23:07 +01:00
Gareth Rees
08bd2f3474 New preprocessor constants config_pthreadext_sigsuspend and config_pthreadext_sigresume for configuring the signals used to suspend and resume threads. 
Copied from Perforce
 Change: 192122
 ServerID: perforce.ravenbrook.com
 2016-09-04 13:55:39 +01:00
Gareth Rees
89b1b02ee5 Add a case for rootthread to rootcheck. 
Add test coverage for mps_root_create_thread_scanned and mps_root_create_area; improve coverage for mps_root_create_area_tagged, mps_scan_area_tagged, mps_scan_area.

Copied from Perforce
 Change: 192111
 ServerID: perforce.ravenbrook.com
 2016-09-04 11:56:09 +01:00
Gareth Rees
7d5e5ac0e0 Avoid strict aliasing errors in the mps test suite, so that it compiles and runs using gcc 6.1. 
Update the manual entries for mps_alloc and mps_sac_alloc with notes highlighting the problem.

Copied from Perforce
 Change: 192104
 ServerID: perforce.ravenbrook.com
 2016-09-02 19:20:58 +01:00
Gareth Rees
55adbe705d Fix assertion text in the manual. 
Place a cross-reference by each assertion the appears in the manual, to assist in keeping the manual up to date.

Copied from Perforce
 Change: 192097
 ServerID: perforce.ravenbrook.com
 2016-09-02 14:11:08 +01:00
Gareth Rees
9b81dcb5c4 New program mpseventpy emits python data structures representing telemetry events. 
Copied from Perforce
 Change: 191895
 ServerID: perforce.ravenbrook.com
 2016-06-02 15:02:05 +01:00
Gareth Rees
6f196185a1 Make shield assertions robust against deep checking -- when shieldcover is called from shieldraise there is one unsynced segment that has not yet been queued. record this fact in a new queuepending flag in the shield structure. 
Remove incorrect shield assertion "shield->depth == 0 || shield->suspended" -- depth may be increased without suspending the mutator if the segment did not need protecting.

Copied from Perforce
 Change: 191812
 ServerID: perforce.ravenbrook.com
 2016-05-03 17:25:27 +01:00
Gareth Rees
11944d44c2 Patch the mfs pool's alloc method instead of rewriting its class -- the latter violates the class invariants and so asserts when deep checking is turned on. 
Copied from Perforce
 Change: 191779
 ServerID: perforce.ravenbrook.com
 2016-05-02 13:49:43 +01:00
Gareth Rees
fa2fced306 Clear the emergency before removing the trace from busytraces, to avoid violating <code/global.c#emergency.invariant>. 
Copied from Perforce
 Change: 191774
 ServerID: perforce.ravenbrook.com
 2016-05-02 13:06:07 +01:00
Richard Brooksby
e26f0b64e5 Replacing segbuffer with version that returns bool and the buffer, and disallowing null to segsetbuffer, adding setunsetbuffer for that case instead. 
Copied from Perforce
 Change: 191695
 ServerID: perforce.ravenbrook.com
 2016-04-27 15:47:26 +01:00
Richard Brooksby
9a6981ae41 Temporarily introducing seggetbuffer to eliminate use of "null" to mean "no buffer". 
Copied from Perforce
 Change: 191694
 ServerID: perforce.ravenbrook.com
 2016-04-27 15:27:47 +01:00