wine/gecko
0
0
Fork 0
mirror of https://gitlab.winehq.org/wine/wine-gecko.git synced 2024-09-13 09:24:08 -07:00
Code Issues Packages Projects Releases Wiki Activity
18,746 Commits 1 Branch 166 Tags 1.2 GiB
1d44de3cb6
Commit Graph

846 Commits

Author SHA1 Message Date
Andreas Gal
13a4a98604 Merge. 2008-08-14 16:04:14 -07:00
Andreas Gal
1edbe7e6bf Don't over-compensate sp_adj (exclude the callee). 2008-08-14 16:02:33 -07:00
Vladimir Vukicevic
61a89af8e6 b=450176; trace parseInt and parseFloat; r=gal 2008-08-14 15:13:39 -07:00
David Anderson
9f7d261303 Fixed the known native arg handler from accidentally using 'continue' to break out of a loop one level higher. This was causing crashes when the arg didn't match the expected types (bug 450530). 2008-08-14 11:48:43 -07:00
Andreas Gal
bfcf6a784d Write back outer tree frames, but exclude the current frame (which the next tree will do). This still crashes the trace-tests.js test case with nesting enabled and for the life of me I can't figure out why (try with TRACEMONKEY=nesting). 2008-08-14 01:44:32 -07:00
Andreas Gal
fa52d00c5a Backed out changeset 089406b2b0aa 2008-08-14 00:45:39 -07:00
Andreas Gal
55bc4df432 Add an explicit start frame parameter to the FORALL macros, which allows FlushNativeStackFrame to be used for other frames than just the topmost N ones. 2008-08-13 23:20:44 -07:00
Andreas Gal
a5e83b133c Fix outrageously incorrect comment. 2008-08-13 22:54:23 -07:00
Andreas Gal
0530506440 Restore state of the outer frames in case of a nested exit. Should be feature complete but needs more debugging. 2008-08-13 21:34:41 -07:00
Andreas Gal
8d4c97dbef FlushNativeStackFrame currently only handles the innermost nested trees writeback, so at least make it use the proper adjusted stack base for that. 2008-08-13 19:23:28 -07:00
Andreas Gal
01efe3882f Update inlineCallCount with the total call stack height, which is the sum of rp_adj and any adjustments nested trees added. Also make sure to read all stack adjustment information from the tree we exit on, not the tree we entered (might be different in case of nesting.) 2008-08-13 19:09:05 -07:00
Andreas Gal
30b82dbcbd Print relative stack instead of absolute stack address upon side exit (debug mode). 2008-08-13 18:28:43 -07:00
Andreas Gal
9350282579 Implement nested side exits. 2008-08-13 18:10:19 -07:00
Andreas Gal
222e746fe7 Merge. If I had a cent for every merge. Sigh. 2008-08-13 17:48:04 -07:00
Andreas Gal
f1d87d6234 Reshuffle code lines in the side-exit return path (do asserts closer to the GlobalFrame writeback.) 2008-08-13 17:47:18 -07:00
David Anderson
4d25c18771 Merge. 2008-08-13 17:46:58 -07:00
David Anderson
87d9c2089d Merge. 2008-08-13 17:45:50 -07:00
Brendan Eich
fc782418ee Set rval_ins with initializing=true because nothing will get the callee, so the tracker won't have tracked it yet if it's in virgin stack. 2008-08-13 17:45:49 -07:00
David Anderson
3652f7c126 Fix from brendan for propertyIsEnumerable having an inconsistent return type. 2008-08-13 17:44:26 -07:00
Andreas Gal
eac5c3ed47 Merge. Lame. 2008-08-13 17:13:53 -07:00
Andreas Gal
201341cfc5 Add nestedExit to InterpState and comment its fields. 2008-08-13 17:12:55 -07:00
David Anderson
a9922db939 Merge. 2008-08-13 17:12:08 -07:00
David Anderson
4c7aaa248b Merge. 2008-08-13 17:11:22 -07:00
Brendan Eich
ca1e2a283a Fix prototype hit case in prop to advance obj and obj_ins up the proto chain. 2008-08-13 17:10:18 -07:00
David Anderson
fe66e966ca Fixed from brendan for counting disparity between nativeStackSlots and FORALL_SLOTS_IN_PENDING_FRAMES. 
Added debug code to catch future disparities.
 2008-08-13 17:09:16 -07:00
Andreas Gal
ee55743358 Don't print LeaveFrame debug info when falling out of the global frame. 2008-08-13 16:57:14 -07:00
Andreas Gal
9c95a20382 Merge. 2008-08-13 16:43:46 -07:00
Andreas Gal
e702dc5ac4 When re-importing register values after a tree call, make sure to use the inner tree's nativeStackBase, since we also use the inner tree's sp. 2008-08-13 16:29:59 -07:00
Brendan Eich
7fc14c40cf Fix bogus varval assert. 2008-08-13 16:14:22 -07:00
Andreas Gal
27c128f706 Merge. 2008-08-13 15:50:40 -07:00
Andreas Gal
e68f9ee859 Read back registers used by inner tree relative to inner_sp (the adjusted sp value). Print frames we enter into and return to in Enter/LeaveFrame. 2008-08-13 15:50:11 -07:00
Brendan Eich
3d21588b2c Major and winning overhaul to for-in codegen (mad props to Andreas for advice). 2008-08-13 14:02:35 -07:00
Andreas Gal
892137794b Don't call inner trees if they use global slots since we don't support those in inner trees yet. 2008-08-13 14:00:37 -07:00
Andreas Gal
1d809d7adf Merge. 2008-08-13 13:52:21 -07:00
Andreas Gal
fa09943222 Flexible call stack allocation with proper guarding for call stack overflows. 2008-08-13 13:51:59 -07:00
David Anderson
ff9bb9ea9d Fixed ifop predicting NaNs as true when they should be false. 2008-08-13 12:25:50 -07:00
Andreas Gal
fa3085ca3d Merge. Nothing to see here. Move along. 2008-08-13 03:55:24 -07:00
Andreas Gal
5d7914a698 Blacklist a trace if we hit a global shape mismatch. This makes us suck less on date-format-tofte until we find a way to fix the property cache misses. 2008-08-13 03:54:54 -07:00
Andreas Gal
2a60d7a27b Only trash the tree, not the entire cash, on global shape mismatch. 2008-08-13 03:50:53 -07:00
Brendan Eich
ed5f8343a2 Fix deep for-in loop bug (450334). 2008-08-13 00:41:13 -07:00
Andreas Gal
335fb1edd8 Merge. 2008-08-12 23:31:12 -07:00
Andreas Gal
91e81a361d Don't cache loads in import into the nativeFrameTracker, since when coming back from a nested tree we re-load the frame state into registers using import() based on the called tree's state, which pushes loads into the nativeFrameTracker that are relative to the inner tree's call depth. Only cache on writes now, which should be always safe. 2008-08-12 23:30:49 -07:00
Brendan Eich
17a4dd5c46 Remove bogus assertion. 2008-08-12 23:30:38 -07:00
Andreas Gal
8af146d25b Merge. 2008-08-12 23:25:40 -07:00
Andreas Gal
a271a0ee0e When re-importing registers after a tree call, make sure to use the inner trees calldepth at its side exit, not the call depths of the calling tree. 2008-08-12 23:25:17 -07:00
Brendan Eich
452bea0c69 Fix missing argument stack offset computation (in both places: FORALL_FRAME_SLOTS and nativeStackOffset). Clear missing args in nativeFrameTracker. 2008-08-12 23:21:52 -07:00
Andreas Gal
58cd4dc419 Introduce a large fixed-size native frame stack and store its ceiling in state->eos. 2008-08-12 21:39:44 -07:00
Andreas Gal
3fd013cd9b Merge. 2008-08-12 20:19:05 -07:00
Andreas Gal
ed3125b915 Hands down the hardest bug I had to debug in TM so far. Make sure to read back any registers an inner tree might have changed before writing out the typemap for the nested_exit guard, otherwise we might be pointing to old stale pre-(inner-)loop state and pick an incorrect (in this case too narrow) type. fannkuch=2.8x with this. 2008-08-12 20:18:29 -07:00
Brendan Eich
967ec56341 - Add String match and three replace overloadings, and allow known native matching to continue in search of exact match (not best, and not abort on first mismatch). 
- Add CallGetter built-in and use it for regexp class-getter-implemented prototype properties.
- Add BUILTIN5 support (this should be "it" ;-).
 2008-08-12 18:52:28 -07:00
Brendan Eich
48181da4e7 Use aobj consistently in test_property_cache (450317). 2008-08-12 17:21:32 -07:00
Andreas Gal
d56662f293 Enumerate missing argument slots on the caller's stack and initialize them to undefined (450304). 2008-08-12 16:51:55 -07:00
Andreas Gal
1819586877 js_obj_hasOwnProperty is supposed to return a boolean but was incorrectly generating code to return a number, resulting on a failure in the boolean comparison due to an unexpected i2f (450304). 2008-08-12 14:28:15 -07:00
Brendan Eich
64fa74acf5 More guard argument formatting. 2008-08-12 11:36:52 -07:00
Brendan Eich
827b341985 Consolidate common name (global slot) addressing code. 2008-08-12 09:42:29 -07:00
Brendan Eich
26c7507913 Put trailing args on their own lines when they otherwise would seem to associate with a nested call in a previous arg position. 2008-08-12 09:27:54 -07:00
Brendan Eich
fd08308ba0 Fix goto over init whinage. 2008-08-12 09:18:19 -07:00
shaver@mozilla.org
1813fb2fa7 NAMEINC, INCNAME, NAMEDEC, DECNAME 2008-08-12 11:33:40 -04:00
Andreas Gal
b94f5433e5 Merge. 2008-08-12 03:40:16 -07:00
Andreas Gal
b4f7639e6e Don't read type from typemap if we already have it in a local variable. 2008-08-12 02:00:05 -07:00
Andreas Gal
7e95e972d8 Don't re-record traces over and over in case of excessive type mismatches. Instead, blacklist the fragment and slowly back away from recording it. 2008-08-12 01:31:28 -07:00
Brendan Eich
8d77aad49c JSOP_FORNAME. 2008-08-12 01:06:02 -07:00
Brendan Eich
d2e6e9e641 Factor forInOp from JSOP_FORLOCAL, use it from JSOP_FORARG's recorder too; tighten up JSOP_SETNAME. 2008-08-12 00:34:43 -07:00
Brendan Eich
e90731a3b6 Comments for JSOP_{,STRICT}{EQ,NE} about the layering and constraints on evolution of the non-strict equality ops' recorders. 2008-08-11 23:44:43 -07:00
shaver@mozilla.org
cfc93a30d9 cmp over number and (string or bool or undefined) 2008-08-12 01:05:33 -04:00
shaver@mozilla.org
2b16f54740 reland eb01b1d55d9b after the mismerge 2008-08-12 00:37:16 -04:00
shaver@mozilla.org
ba5a56aa89 reland 3ea1e1317707 after the mismerge 2008-08-12 00:36:25 -04:00
Brendan Eich
2cbfd79ea0 Restore lost shaver rev 0e50c89c476b -- how did that get lost? 2008-08-11 20:27:47 -07:00
Brendan Eich
a63e9a65ba Merge, sigh. 2008-08-11 19:21:42 -07:00
Brendan Eich
bba0ea3906 Rename js_DestroyJIT to js_FinishJIT to match Init/Finish vs. New/Destroy naming scheme used elsewhere; use same #if conditions around #include jstracer.h as around code depending on it. 2008-08-11 19:17:30 -07:00
shaver@mozilla.org
16a7af3ef4 JSOP_EQ/NE for objects 2008-08-11 22:11:31 -04:00
Brendan Eich
9dd296d70a Merge, in the name of all that which does not suck\! 2008-08-11 18:53:21 -07:00
David Anderson
9ce9cb42b5 Merge. 2008-08-11 18:49:34 -07:00
David Anderson
9c0315ccf7 Fragmento lifetime is now associated with the thread/runtime instead of the context. 2008-08-11 18:47:22 -07:00
Brendan Eich
7fbc9732fd JSOP_LENGTH handles dependent strings now. 2008-08-11 18:41:06 -07:00
shaver@mozilla.org
6a35d8134e implement JSOP_ADD(str, num) 2008-08-11 21:32:26 -04:00
David Anderson
c0b377bb8b Fixed some explicit deallocation bugs that cropped up running in the browser. 2008-08-11 18:20:10 -07:00
Andreas Gal
ade8388ef5 Merge. 2008-08-11 18:03:31 -07:00
Andreas Gal
20318a6ce0 Trash tree if we see an obsolete treeinfo object on record. 2008-08-11 18:03:13 -07:00
Brendan Eich
5137a68ca5 Merge. 2008-08-11 17:54:59 -07:00
Brendan Eich
d004346b52 Object.prototype.{hasOwnProperty,propertyIsEnumerable}. 2008-08-11 17:47:05 -07:00
Brendan Eich
9849ef7434 Handle multi-level property cache hits; put BRANCH_EXIT on its own line so it stands out as the third param to guard in ifop. 2008-08-11 17:40:26 -07:00
David Anderson
c5a85f7524 Merge. 2008-08-11 16:14:29 -07:00
David Anderson
1b623f7d25 Added explicit deallocation (bug 443500, r=gal) 2008-08-11 16:01:21 -07:00
Andreas Gal
020e68ab1e Remove default parameter for guard, specify MISMATCH_EXIT explicitly. 2008-08-11 15:40:04 -07:00
Andreas Gal
00e5631b45 Remove dead code, add a debug printf. 2008-08-11 15:23:37 -07:00
Andreas Gal
7bcf550b11 If we get a series of tree type mismatches, trash the tree (and all dependent trees in case of nesting.) 2008-08-11 15:16:34 -07:00
Andreas Gal
120117749e Don't try to access the name of arguments beyond nargs (r=brendan). 2008-08-11 12:47:18 -07:00
Andreas Gal
285acf8b9c Merge. 2008-08-11 11:43:59 -07:00
Andreas Gal
ca11b2e069 Don't use asprintf (446508). 2008-08-11 11:37:35 -07:00
Brendan Eich
13cfac8f97 Rework GCF_DONT_BLOCK to be cx->gcDontBlock, assert it's set, set before entering trace and clear on exit. 2008-08-11 11:18:16 -07:00
Brendan Eich
a56862bdff Make js_ConcatStrings JS_FASTCALL and use directly as a built-in; remove gcflag param from it and from js_NewString. 2008-08-11 11:03:54 -07:00
Brendan Eich
56fac8fe90 Back out extra gcflag param change (859b9a23adbf), I set bad precedent and will back out previous such shortly. Want cx->gcflags instead. 2008-08-11 10:33:23 -07:00
Brendan Eich
fced00a77b - Trace JSOP_NEWINIT/INITPROP/INITELEM/ENDINIT, which required extending the FastNewObject builtin to create a dense Array or a new Object (easy to discriminate on the constructor function's u.n.clasp member). 
- Fix record_JSOP_SETELEM to avoid storing if a JSOP_POP that will be skipped follows.
 2008-08-11 10:24:47 -07:00
shaver@mozilla.org
ee67ea9db4 extend js_NewObjectWithGivenProto to permit specifying additional new-thing flags (such as GCX_DONT_BLOCK) 2008-08-10 22:56:10 -04:00
Brendan Eich
0d83651633 - Skip dense array object, try prototype, in all JSOP_GET*PROP variants, interpreter and tracer. 
- Test entry instead of retesting aobj->map->ops->getProperty == js_GetProperty in JSOP_CALLPROP.
- Handle object and null types as well as booleans in record_JSOP_NOT.
- Trivially implement JSOP_STRICT{EQ,NE} recording by forwarding to the non-strict record methods. This works so long as the latter trace only same-type operands and do no conversions.
 2008-08-11 00:05:39 -07:00
Brendan Eich
43649cd01e Record JSOP_LOOKUPSWITCH. 2008-08-10 23:20:24 -07:00
Brendan Eich
64fc3b0367 - Fix shapeless callee guarding to guard on function object value. 
- Add JSOP_NULLTHIS to help the tracer guard shapeless callees (see trace-tests.js)
- Culled bogus record_JSOP_CALLGVAR left-over forwarding to record_JSOP_GETGVAR.
- Better shapeless callee tests.
 2008-08-10 22:36:48 -07:00
Brendan Eich
a67dad8f4d Reoptimize call and new based on shape guards and branded method-ful scopes -- no need to emit function-is-interpreted guards. 2008-08-10 15:26:39 -07:00
Brendan Eich
0099753ff4 Add Array (generic) join builtin, plus (not yet used) optional this-class guarding for builtins. 2008-08-10 14:50:31 -07:00
Brendan Eich
5bbc609a8b Implement JSOP_OBJECT. 2008-08-10 12:01:03 -07:00
Brendan Eich
e5b0039eaa Match native map guard specializations in tracer to interpreter, particularly to skip up the proto chain from a dense array in JSOP_CALLPROP (json2.js showed an abort here, pointing to the tracer/interpreter discrepancy). 2008-08-10 11:54:27 -07:00
Brendan Eich
18248a701c Fix JSOP_FORLOCAL to push false when guarded conditions are false while recording (449961). 2008-08-10 10:42:21 -07:00
Brendan Eich
2991c66416 - Avoid unnecessary FASTCALL builtin wrappers for existing js_* library-private or friend functions, which could be (and now are) fastcall (JS_FASTCALL). A couple of builtins avoid name collisions by using js_Fast instead of just js_ as their name prefix. 
- Use GCF_DONT_BLOCK from fastcall builtins that call into the GC allocator (mostly; a few places need help still, or better: need to avoid allocation).
- Cope with C clients of jscntxt.h and its include files (these hacks should go away).
- Naming convention and function definition line-break style nit-picks.
 2008-08-10 00:39:18 -07:00
Brendan Eich
f2b84e6ea0 - Test JSVAL_BOOLEAN tag in TraceRecorder::record_JSOP_TYPEOF, assert no holes or other pseudo-booleans, to unify false/true/undefined testing. 
- JS_NOT_REACHED("...") not JS_ASSERT(0 && "...").
-
 2008-08-09 22:27:58 -07:00
shaver@mozilla.org
bd5846e102 JSOP_ANONFUNOBJ 2008-08-09 23:12:32 -04:00
shaver@mozilla.org
a937f98d5b JSOP_TYPEOF and JSOP_TYPEOFEXPR 2008-08-09 22:54:01 -04:00
shaver@mozilla.org
c4ca6755bf Builtin for String.prototype.concat (single int-arg version) 2008-08-09 22:15:21 -04:00
shaver@mozilla.org
3341bf1c11 Implement cmp for string-on-string. 
Add < and > tests for strings.
Condition the top-level truthies test on a synthetic test name, so it can be
excluded.
 2008-08-09 21:50:52 -04:00
Robert Sayre
c8dd484645 Bug 449978 – TM: trace some more builtins for string-validate-input 2008-08-09 21:47:52 -04:00
Robert Sayre
2ac505035a Bug 449666 – TM: Assertion failure: JSSTRING_IS_FLAT during trace recording. r=brendan 2008-08-09 18:20:18 -04:00
shaver@mozilla.org
d431dd547d improve naming of anonymous functions and excess args 2008-08-09 15:05:34 -04:00
Brendan Eich
c90cc76482 Prettier spacing knownNatives table. 2008-08-08 22:43:31 -07:00
Andreas Gal
36eb8da275 Merge. 2008-08-08 19:24:50 -07:00
Andreas Gal
66dd6aab15 Add an assert to protect against nested exits (not implemented yet.) 2008-08-08 19:23:28 -07:00
Brendan Eich
eaf4254c15 Merge from mozilla-central. 2008-08-08 18:58:04 -07:00
Andreas Gal
8244eafaff Support calling of nested tree from within inlined frames. When we call a tree that tree expects to be the top-level tree, which means it assumes callDepth=0. If the call of the tree is within an inlined frame, thats not true (callDepth > 0). We adjust the native stack pointer accordingly before calling the tree, and then restore the value of the stack pointer after the call. Fancy stuff. 2008-08-08 18:52:08 -07:00
Andreas Gal
3f90159e8b Remove ip from InterpState. Calculate the new pc after a trace side exit relative to lr->from->root->ip, which also works if we side exit on a different tree than we entered (which can happen in case of nested trees.) 2008-08-08 18:20:20 -07:00
Brendan Eich
f028f26f52 Merge. 2008-08-08 16:40:45 -07:00
Brendan Eich
072c7b59a5 Guard against shapeless callees, with tests (more to do here, can't break the unknown callee case yet, but it should be breakable). 2008-08-08 16:37:01 -07:00
Andreas Gal
c7be42f2f3 Add very preliminary nesting for trees. This is disabled by default. To enable set TRACEMONKEY=nesting in the environment. 2008-08-08 15:26:31 -07:00
Brendan Eich
de40e589da Merge. 2008-08-08 14:45:27 -07:00
Brendan Eich
153e03bb32 Get operator new on interpreted functions working, along with JSOP_SETPROP on an unmutated object, or one of the right shape but where the setprop is adding the next property, and it's not in the object yet. 2008-08-08 14:38:44 -07:00
Andreas Gal
f2972efa5c Remove tracking of outer trees and merging of globals since globals will be maintained centrally soon, not attached to trees. 2008-08-07 19:40:41 -07:00
Brendan Eich
43f60608f4 Fix bug 449627. 2008-08-07 18:12:31 -07:00
Andreas Gal
f77dcbe3f6 We don't deallocate JIT-related data structures when the JIT is shut down. Add a TODO for this. 2008-08-07 17:52:48 -07:00
Andreas Gal
2d0c9c41ab Hand the script in when updating the oracle regarding global variable slots even if we don't use it right now. This will be useful for caching information across runtimes. 2008-08-07 17:43:13 -07:00
Andreas Gal
3ac96b066b Use a typedef SlotList instead of Queue<uint16>. 2008-08-07 17:22:21 -07:00
Andreas Gal
a21818bc72 Teach typemaps how to capture the types of all slots in a slot list. 2008-08-07 17:20:49 -07:00
Andreas Gal
d9159ecfb2 When merging globals from inner trees to outer trees, make sure to push the expected type upstream as well. 2008-08-07 17:08:27 -07:00
Andreas Gal
afda8f7095 When we trash a tree, we have to trash all outer trees that call it as well. 2008-08-07 17:01:00 -07:00
Andreas Gal
d988d3523a Recognize loop edges that hit a nested tree. 2008-08-07 16:41:20 -07:00
Andreas Gal
bbe7acfbaf Fix typo (missing memcmp in statement). 2008-08-07 16:27:26 -07:00
Andreas Gal
600b9c3e14 Remove type-map hash code. A direct memcmp seems to be faster (probably some SSE magic behind it in gcc/glibc). 2008-08-07 16:26:56 -07:00
Andreas Gal
27e8ef817a Factor out the code to capture the current stack type map and move around the type map method implementations to make sure they can see the FORALL macros. 2008-08-07 16:23:50 -07:00
Andreas Gal
01e3af1f79 We expect no interned globals in a new tree. Assert on that. 2008-08-07 15:44:11 -07:00
Andreas Gal
00b12bbaca Always trash TreeInfo when we recompile. This avoids the much dreaded JS_ASSERT(insInt32) errors and we can do this now because the demotion information is provided by the oracle and no longer stored in the typemap directly. 2008-08-07 15:38:54 -07:00
Andreas Gal
95f5262e4f Use an oracle to predict when slots are demotable. This allows trashing the TreeInfo* object in the root fragment more aggressively. Rebuild the stack type map at every compile to avert type map mismatches. Purge all global slot info when we re-record a tree. 2008-08-07 15:28:43 -07:00
Andreas Gal
e09041e2a5 If we run into an error during compilation, blacklist that fragment. If we run out of memory, flush the cache. 2008-08-06 22:26:20 -07:00
Andreas Gal
0fe0cf9a98 Flush JIT cache for all contexts. Clear nanojit error state when we start compiling. 2008-08-06 22:18:33 -07:00
Andreas Gal
e9733ceed9 Remove dead code. 2008-08-06 22:10:30 -07:00
Andreas Gal
9385e7842c Add a helper to flush the JIT code cache (and the fragment lookup quick cache). On a global shape mismatch trash the entire cache (might be a bit overly aggressive). Similary, during GC flush the code cache. 2008-08-06 21:56:25 -07:00
Andreas Gal
d1a1a76121 Tracker outer trees for every tree and merge globals of inner trees into all outer trees as we register inner trees with outer trees. 2008-08-06 19:25:24 -07:00
Andreas Gal
3785f2a5af Fix misleading debug text. 2008-08-06 19:04:41 -07:00
Andreas Gal
ff22344594 Fix tracing of code inside methods that were called with arity mismatch. 2008-08-06 18:51:26 -07:00
Brendan Eich
f23efc77b2 Replace bogus assertion with runtime test for correct abort test (getter or setter, given sprop hit from propcache). 2008-08-06 17:54:36 -07:00
Andreas Gal
d41941073e Disable outerlining, to be replaced with nesting. 2008-08-06 17:34:06 -07:00
Andreas Gal
1bf2697257 Add a helper to emit a tree-call into the currently recording trace. 2008-08-06 16:24:29 -07:00
Andreas Gal
27de48d1f9 Cleanup stack offset calculation and eliminate the stack offset fiddling that was necessary for nanojit prior to adding the getTop hook. 2008-08-06 15:04:02 -07:00
Brendan Eich
0f1f29e9a4 Merge again, I lost to Andreas\! 2008-08-06 13:44:07 -07:00
Andreas Gal
2acb2afdec Always trashing the tree on a type mismatch can't work since the first iteration comes in as undefined for loop-outputs and then kills the main tree that deals with the proper stable types. Have to find a different way to deal with this. 2008-08-06 13:41:04 -07:00
Andreas Gal
fcaab6d007 Pull recompile flag out of the state and pass in as argument. Trash the tree if a secondary trace can't be connected to the loop header. This is very aggressive and might need more tinkering. Trashing the tree on every mismatch doesn't seem to work well, so thats currently disabled. 2008-08-06 13:37:29 -07:00
Brendan Eich
6e215765df - Add builtins to support for-in loops, both iterating and getting/setter properties by name using o[i] instead of o.p where i is 'p'. 
- Record JSOP_ITER and JSOP_ENDITER, which must be handled since tracing could start in between them (and always will, until we nest or unless we outerline).
- Add a null state guard in JSOP_FORVAR's recorder method.
- Pick include order, indentation, and trailing space nits.
 2008-08-06 13:36:29 -07:00
Andreas Gal
a253b66a09 If we stop recording, immediately attempt to trigger the tree. 2008-08-06 13:18:30 -07:00
Andreas Gal
141f71186c Properly deal with trace entry with non-empty stack. 2008-08-06 12:38:17 -07:00
Andreas Gal
0ed65746a7 Split side exit handling from js_ExecuteTree. 2008-08-06 06:34:16 -07:00
Brendan Eich
03b511f9e4 - Use JSStackFrame* fp over, don't declare another JSStackFrame* f, in the FORALL macros and clones (f is canonical variable name for nanojit::Fragment*, fp for JSStackFrame*). 
- Nit-pick include order (alphabetical within each group).
- Trailing space and indentation no-knock raid.
 2008-08-03 22:55:28 -07:00
Brendan Eich
da42a8c6f7 Strength-reduced unsigned modulus in the fragment quick cache, don't use signed % which requires a branch and less-than-zero test. 2008-08-03 22:35:15 -07:00
Brendan Eich
df36d9e36f Use INS_CONST to addName a few lir->insImm immediates. 2008-08-02 20:01:36 -07:00
Brendan Eich
5b9f299115 Fix TraceRecorder::ifop backward logic bug in the OBJECT case, and implement the STRING case. Add tests for truthy and falsy strings. 2008-08-02 16:51:22 -07:00
Brendan Eich
607e259b69 Fix return NULL in bool to return false. 2008-08-02 16:05:29 -07:00
Andreas Gal
3106dc8845 Use quick cache to bypass fragmento is possible. 2008-08-01 22:56:52 -07:00
Andreas Gal
12f1c06952 Factor out most of the remaining code in js_LoopEdge into js_RecordTree and move the code to attach new branches into js_ExecuteTree. 2008-08-01 22:47:15 -07:00
Andreas Gal
746cb91802 Factor out js_ContinueRecording from js_LoopEdge. 2008-08-01 22:39:52 -07:00
Andreas Gal
3fea3f5d53 Split up FORALL_SLOTS_IN_PENDING_FRAMES macro in a macro that processes a frame, and a macro that uses that macro to process all pending frames. 2008-08-01 22:27:39 -07:00
Brendan Eich
3691962a2f Extend ifop to handle undefined tests; implement JSOP_OR and JSOP_AND (so much work\!). 2008-08-01 12:06:04 -07:00
Brendan Eich
dc1072fe55 Style police raid: function names start in column 1, with one blank line between functions. 2008-08-01 11:03:52 -07:00
Brendan Eich
2cfdd9a55a 1. Don't store thisp literally in state.rp, get it from argv[-1]. 2. Fix DEBUG localNames code. 2008-08-01 10:40:38 -07:00
Brendan Eich
8f837e0784 - Implement interpreter frame reconstruction (js_SynthesizeFrame). 
- Fix MONITOR_BRANCH to restore frame-dependent interpreter locals (we want fewer of these if they can be removed with no perf loss).
- Fix FORALL_SLOTS_IN_PENDING_FRAMES not to scan argv in callee when it has scanned operand stack in caller.
- Fix import to take the localFrame from whose fun its localNames parameter was computed -- it was using cx->fp which is wrong when FORALL_SLOTS_IN_PENDING_FRAMES iterates other than the top of stack frame.
- A few interval tests that were double-ended are single-ended now.
- Add call.js mini-test (more cases needed).
 2008-08-01 08:26:32 -07:00
Andreas Gal
4eba8e4012 Remove entryStackDepth. Calculate sp_adj relative to entryNativeStackSlots. 2008-08-01 02:33:54 -07:00
shaver@mozilla.org
3221386525 add specialized StringToInt32 and filter for it 2008-07-31 11:35:08 -07:00
shaver@mozilla.org
12614d0334 Coerce strings to numbers for appropriate ops. 
(Wants a specialized StringToInt32 and a filter to put it in place.)
 2008-07-31 07:41:58 -07:00
shaver@mozilla.org
d200eb863e merge backout of d24e6005ee4c to fix the world 2008-07-31 06:36:53 -07:00
shaver@mozilla.org
f2c0c93637 Backed out changeset d24e6005ee4c (causing major array-fail). 2008-07-31 06:35:11 -07:00
Andreas Gal
e5a87ade9f Merge. 2008-07-31 01:35:56 -07:00
Andreas Gal
6f4cfc2628 Remove ANY_TYPE and move debug printf around to print entry point even if we can't enter due to type mismatch. 2008-07-31 01:35:18 -07:00
shaver@mozilla.org
a3bc7296d9 Make guard return expected, so we can trace alternate cases easily. 
Better diagnostic for non-global scope chain head.
 2008-07-30 22:59:13 -07:00
Andreas Gal
86aa1f7ff7 If trees are not enabled, don't try to reuse state and param1. 2008-07-30 21:40:21 -07:00
David Anderson
ad3e892e92 Re-use initial parameters on tree fragments. 2008-07-30 23:30:58 -05:00
shaver@mozilla.org
959dd9184f Implement JSOP_EQ and JSOP_NE over strings, plus JSOP_STRING. 2008-07-30 17:28:59 -07:00
Andreas Gal
e91eb4bc47 Merge. 2008-07-30 16:37:35 -07:00
Brendan Eich
fe3a6127c5 - Export JSSLOT_ITER_* from jsiter.cpp to jsiter.h, for jstracer.cpp to use. 
- Fix OBJ_GET_SLOT to be STOBJ_GET_SLOT in jsiter.cpp, no thread safety here (bug on file).
- Move JSNativeEnumerator from jsobj.cpp to jsobj.h for jstracer.cpp as well.
- Rename JOF_2BYTE JOF_UINT8 for consistency, and actually decompile it.
- Trace JSOP_FORVAR (can't do anything in JSOP_ITER, it comes before the loop).
- Shortened some guardMyLongSummerVacationWithinBounds names ;-).
- Removed/refactored to avoid dslots_ins obligation on all callers.
 2008-07-30 16:32:33 -07:00
Andreas Gal
698d3fb5cc Make the global frame layout match the slot layout in the global object. This will allow leaving global values in place when switching trees as long both inner and outer tree use the same value. 2008-07-30 16:28:48 -07:00
Andreas Gal
e8043788a7 Merge. 2008-07-30 16:06:11 -07:00
Andreas Gal
9b9d60583c Use isGlobal to distinguish whether a value is a global slot instead of scanning the table every time. 2008-07-30 16:05:51 -07:00
shaver@mozilla.org
124e9ae44b trace Math.random 2008-07-30 15:51:44 -07:00
shaver@mozilla.org
b1df95f721 Trace String.fromCharCode. 
Handle failure signals of < 0 or NULL from traceable natives.
 2008-07-30 15:19:25 -07:00
Andreas Gal
266a7c534f Clean up the living room a bit since people are going to come by to look at it. 2008-07-30 13:34:22 -07:00
Andreas Gal
3a278cdf00 Add isGlobal to check whether a value is a slot of the global object. 2008-07-30 12:15:53 -07:00
Andreas Gal
40696cff65 Note to self: hacking after 4am is detrimental to my spelling. 2008-07-30 04:20:48 -07:00
Andreas Gal
f1a5494d05 Major shakeup of the interning code for globals. Globals are now detected on demand as they are used and the slots are noted in treeInfo->globalSlots. At the same time the type is recorded in treeInfo->globalTypeMap. The stack type-map is maintained separately in treeInfo->stackTypeMap. All these structures are lists and are maintained as List<T> objects. Imports for globals can appear at the top (if we have already seen some imports for the loop header and are recompiling), or on the fly for lazily found values. We no longer intern all global properties that happen to match a name in the current function, and we also support inlining of functions that touch globals that are not used in the method where the trace started in. 2008-07-30 04:17:22 -07:00
Andreas Gal
7290f281a0 Use List<T> to maintain global slot list in TreeInfo. 2008-07-30 01:38:21 -07:00
Andreas Gal
7beb3cd77f Trash entire tree with all the information associated with it when we have a typemap conflict or the global shape changes. 2008-07-30 01:29:13 -07:00
Andreas Gal
afc7936599 Remember number of global slots known at that point in the trace in every side exit (forward-looking change to cope with dynamic collection of interned globals.) 2008-07-30 00:15:07 -07:00
Andreas Gal
90eee755f0 Merge. 2008-07-29 23:51:43 -07:00
Andreas Gal
f60c785137 Add a generic list data structure and fix side exit handling to always pick the right typemap to work with (exit map, not entry map). 2008-07-29 23:48:39 -07:00
David Anderson
60181b7fd2 Fixed some MSVC whinings, implemented rdtsc on win32 2008-07-29 19:19:51 -05:00
Brendan Eich
48a3285a82 Merge again. 2008-07-29 16:02:53 -07:00
Brendan Eich
73e208808e Guard property cache hits by shape(s). 2008-07-29 16:01:00 -07:00
Andreas Gal
4d5b44839c Rename gslots to globalSlots. 2008-07-29 15:28:23 -07:00
Andreas Gal
437f15d428 Remove global frame transition code. Obsoleted by the new on-demand global loading code which we are about to add. 2008-07-29 15:16:35 -07:00
Brendan Eich
cc79bb34c5 Merge. 2008-07-29 14:52:22 -07:00
Andreas Gal
5a7ac301e6 Limit tree growth to side exits that expicitly declare that they want to be grown. 2008-07-29 11:13:41 -07:00
Brendan Eich
636b8ed6b6 Fix a couple of comments. 2008-07-29 10:53:58 -07:00
shaver@mozilla.org
ce85bd8955 JSOP_ADD over strings 2008-07-29 07:53:31 -07:00
shaver@mozilla.org
2a1fd601c4 signs, signs, everywhere signs 2008-07-29 07:38:04 -07:00
shaver@mozilla.org
a8b963d4a4 Trace String.prototype.substring for two-arg case. 
* Export str_substring as js_str_substring.
* Add basic String_p_substring builtin (only handle end > begin, both in range).
* Add String_p_substring_1 builtin for the missing-end case.
* INS_CONST for named constants in traces.
* Support boxing of strings.
* Support CALLPROP with primitive this.
* Support traceable natives which require cx and this.
* Support fallible traceable natives.
* Fix JSOP_LENGTH to use i2f on result (need that everything-is-doubles T-shirt).
* Add strings test.
 2008-07-29 07:32:18 -07:00
shaver@mozilla.org
98caf383a5 Trace JSOP_LENGTH over flat strings. (Need some cmovery for dep strings.) 2008-07-29 06:51:27 -07:00
Andreas Gal
eaeee57c9b Fix spelling. 2008-07-29 01:02:06 -07:00
Andreas Gal
1d8df7fd3c Cleanup global frame switching and add delayed write-back code for doubles. 2008-07-29 01:00:50 -07:00
Brendan Eich
747b8fab0e Regularize loop update in SwitchNativeGlobalFrame. 2008-07-29 00:13:59 -07:00
Brendan Eich
0a6546fb7a Fix synthesizeFrame's newifp->callerRegs/frame.regs update to pass along the pointer to the precious js_Interpret regs local and update it. 2008-07-29 00:06:29 -07:00
Brendan Eich
f2ca9496f4 Try to fix SwitchNativeGlobalFrame, still studying it but these changes seem necessary. 2008-07-28 23:59:29 -07:00
Brendan Eich
3e4d484622 Fix synthesizeFrame parameterization, and have it reconstruct stack depth; warning and space fixes. 2008-07-28 23:47:20 -07:00
Andreas Gal
28deb1f5b0 Merge. 2008-07-28 23:31:39 -07:00
Andreas Gal
0bb95b6c46 Switch from one global frame to another by walking the two sorted gslots lists. 2008-07-28 23:30:16 -07:00
Brendan Eich
9894a22d65 Fix uninitialized nbytes in synthesize_frame. 2008-07-28 22:25:13 -07:00
Andreas Gal
ed2b81df60 Compilation fix for MSVC. 2008-07-28 21:17:43 -07:00
Andreas Gal
a9e5f84f42 Cleanup and split native frame reading/writing into stack and global part. 2008-07-28 18:06:34 -07:00
Andreas Gal
52bda2ec2c Sort interned global slots in ascending order for fast comparison of two different global frames when switching between them. 2008-07-28 16:49:13 -07:00
shaver@mozilla.org
9c4c8495f9 not-yet-working beginnings of frame reconstruction 2008-07-28 09:05:55 -04:00
Andreas Gal
cd39ba2407 Trash the interned globals of a tree if we experience a global shape mismatch. 2008-07-27 21:44:08 -07:00
Andreas Gal
4df581e3b3 Trash the typemap if a change of the global shape forces us to throw away a tree. 2008-07-27 19:18:51 -07:00
Andreas Gal
b798d08035 Abort trace if we inline too deeply. 2008-07-27 16:28:09 -07:00
Andreas Gal
4710877c9a Merge. 2008-07-27 16:19:15 -07:00
Andreas Gal
6126f5fa5c Track the type of guards and react accordingly if we bail out on them. Guards that protect against out-of-memory conditions don't try to grow the tree. Instead we just resume the interpreter. 2008-07-27 16:18:53 -07:00
Andreas Gal
51b608badc Don't use ABORT_TRACE outside the recorder. 2008-07-27 15:40:34 -07:00
shaver@mozilla.org
a4af3457f4 do setelem in a builtin, so we don't abort trace every 8 times when growing 2008-07-27 18:34:23 -04:00
Andreas Gal
e6cc02f5cc If we see a f2i(UnboxDouble) chain, simplify it to UnboxInt32 which does the conversion internally. This also enables a fastpath to read 31-bit jsval integers from arrays. 2008-07-27 14:55:26 -07:00
Andreas Gal
7837aae392 Merge. 2008-07-27 14:29:24 -07:00
Andreas Gal
a3eb60b77e Add limited outerlining. Much of this will be subsumed by nested trees. 2008-07-27 14:28:56 -07:00
shaver@mozilla.org
e69c5b7239 remove rval tracking, since it's no longer necessary 2008-07-27 15:47:43 -04:00
Andreas Gal
ca563057ee Remove guardCount. No longer needed. 2008-07-27 12:05:01 -07:00
Andreas Gal
8a7bb9c550 Smarter speculative demotion of numbers to integers and promotion of the trace seems to require actual doubles. If the number at entry looks like an int we make the slot an int and compile as such. If the loop-tail proves the slot to be a double, we recompile the trace. Currently such miss-speculation cannot be handled on secondary traces since we are currently unable to recompile the primary trace. Such secondary traces are blacklisted. 2008-07-27 02:15:17 -07:00
Andreas Gal
a4fc19765e Don't try to demote slots on secondary traces (we have to recompile the primary trace as well for that, which we currently don't do.) 2008-07-25 19:30:33 -07:00
Andreas Gal
059abfa50f Added sunspider tests to t/ for dvander. 2008-07-25 18:43:19 -07:00
Andreas Gal
812989bb8a Eliminate EntryRegs. Its not safe to keep a reference to the entry SP since we might extend the tree from a different outer stack frame. Instead just store the entryStackDepth. 2008-07-25 18:22:15 -07:00
Andreas Gal
bbb1bc89c0 Print real recording point, not entryRegs and add an assert that makes sure the guard we come out of is associated with the tree we entered into. 2008-07-25 18:13:57 -07:00
Andreas Gal
24447ee8fb Merge. 2008-07-25 16:54:14 -07:00
Andreas Gal
1b7930d39d The VP engineering broke JSOP_NAME! We have to check whether the slot is actually interned and otherwise abort. 2008-07-25 16:51:42 -07:00
Brendan Eich
a8c6046d8c Avoid JSOP_POPV in global scripts from load(), etc. 2008-07-26 01:23:12 +02:00
David Anderson
4560ed961f Fixed entry typemap having wrong allocation size. 2008-07-25 13:30:52 -07:00
Andreas Gal
d952f192ba Can't use lr->from. Seems to not get set in certain cases. Go figure. The nanojit tree code is really weird. 2008-07-25 02:29:36 -07:00
Andreas Gal
dee805421b Factor out AttemptToGrowTree. 2008-07-25 02:00:02 -07:00
Andreas Gal
ecb79b1640 Determine ahead of time whether a side exit is a loop-terminating side exit instead of re-determining this at every side exit. 2008-07-25 01:44:40 -07:00
Andreas Gal
f3573b9940 Move trace activation code into js_ExecuteTree(). 2008-07-25 01:20:40 -07:00
Andreas Gal
3999210c12 Make demotion threshold optional (0=off, 32=default). Turn off to debug the type assert bug. 2008-07-24 17:33:34 -07:00
Andreas Gal
e838b138b8 Don't demote slots in overly long traces (based on counting the number of guards, current threshold=32). 2008-07-24 16:25:18 -07:00
Andreas Gal
f0399ffea7 Cleanup iteration over all slots to use a single macro to avoid code duplication. 2008-07-24 14:51:14 -07:00
Andreas Gal
e7f92ecf31 Add shark support to md5.js directly, no longer around the recorder. 2008-07-24 11:08:37 -07:00