searching for reviewer(nnethercote)
e5090c9f6be394214f2efbfa87496084c391a51d: Bug 609502 - Fix edge cases in charCodeAt and Math.abs (r=nnethercote,gal, a=blocker)
Paul Biggar <pbiggar@mozilla.com> - Mon, 13 Dec 2010 16:22:59 -0800 - rev 59235
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 609502 - Fix edge cases in charCodeAt and Math.abs (r=nnethercote,gal, a=blocker)
25fd3451c0ae6de4f182bd48e02768369ed08c06: Bug 609905 - reTempAlloc: 512k heap allocation for no purpose. r=nnethercote.
Julian Seward <jseward@acm.org> - Thu, 02 Dec 2010 22:26:24 -0800 - rev 58715
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 609905 - reTempAlloc: 512k heap allocation for no purpose. r=nnethercote.
cf37bdb2b7947d76fab7a87ca3450e4421427dfa: Bug 594296 - Fix the implementation of Assembler::asm_cmov() in the SH4 backend (r=nnethercote r=rreitmai sr=edwsmith)
Cédric VINCENT <cedric.vincent> - Wed, 08 Sep 2010 01:50:00 -0700 - rev 58057
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 594296 - Fix the implementation of Assembler::asm_cmov() in the SH4 backend (r=nnethercote r=rreitmai sr=edwsmith)
a23a09d2a9b6758e9ee3365e640c2afce219ebdc: Bug 602786 - lirasm: fix trivial Windows test failure (r=nnethercote)
Rick Reitmaier <rreitmai> - Mon, 01 Nov 2010 17:09:00 -0700 - rev 57694
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 602786 - lirasm: fix trivial Windows test failure (r=nnethercote)
322fe3e6482e13e682b28f9d84ddcc6b25e27148: Bug 602387 - nanojit: make Register a non-numeric type on PPC (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Tue, 02 Nov 2010 15:43:39 -0400 - rev 57682
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 602387 - nanojit: make Register a non-numeric type on PPC (r=nnethercote+) This patch removes REGINC(), and adds several operator overloads to make working with Register values more convenient without introducing implicit downcasts to int.
33da20109d8e063e0aeb9de9da267e981324fec9: Bug 603890 add big endian test cases r=nnethercote
Ginn Chen <ginn.chen@oracle.com> - Fri, 22 Oct 2010 19:22:13 +0800 - rev 56580
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 603890 add big endian test cases r=nnethercote
ca83ee059cf80cbdbb9e80dd5d04b58d0ba64f9d: Bug 603890 add --endianness to lirasm, make tests endianness aware r=nnethercote
Ginn Chen <ginn.chen@oracle.com> - Fri, 22 Oct 2010 16:53:42 +0800 - rev 56579
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 603890 add --endianness to lirasm, make tests endianness aware r=nnethercote
1b8e50da7f24e43a9a98e6d5b0e8dbbb4b109351: Bug 565571 - nanojit: improve X64 codegen for stores of integer immediates. r=nnethercote,edwsmith.
Werner Sharp <wsharp@adobe.com> - Thu, 21 Oct 2010 20:45:34 -0700 - rev 56557
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 565571 - nanojit: improve X64 codegen for stores of integer immediates. r=nnethercote,edwsmith.
cc0c87e2dd2887d3b6f07c7558c1ea5bdde8b2fe: Bug 602388 make Register a non-numeric type on SPARC r=nnethercote
Ginn Chen <ginn.chen@oracle.com> - Wed, 20 Oct 2010 13:32:16 +0800 - rev 56545
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 602388 make Register a non-numeric type on SPARC r=nnethercote
dd3ec76eacea439997114f5493a8495f834aa4db: Bug 603560 Workaround for Solaris Studio compiler, it failed to compile static const int x = { 1 }; r=edwsmith,nnethercote
Ginn Chen <ginn.chen@oracle.com> - Mon, 18 Oct 2010 19:46:04 +0800 - rev 56197
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 603560 Workaround for Solaris Studio compiler, it failed to compile static const int x = { 1 }; r=edwsmith,nnethercote
81d0ca612cc899707eb307d109d2d8b945f28c24: Bug 599214 - JM: Add a fast path for object equality (r=dvander,nnethercote)
Bill McCloskey <wmccloskey@mozilla.com> - Mon, 18 Oct 2010 17:04:43 -0700 - rev 56192
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 599214 - JM: Add a fast path for object equality (r=dvander,nnethercote)
a1989f4fa64b68598437abed0333d690c7943dde: Bug 600414 - TM: Use a more precise check for negative zero when multiplying (r=nnethercote)
Bill McCloskey <wmccloskey@mozilla.com> - Mon, 18 Oct 2010 15:13:55 -0700 - rev 56185
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 600414 - TM: Use a more precise check for negative zero when multiplying (r=nnethercote)
817760a6153a2d53026a7a4747cf95f49334a1b6: Bug 600016 - TM: Recognize that the result of Math.floor is an integer (r=nnethercote)
Bill McCloskey <wmccloskey@mozilla.com> - Mon, 18 Oct 2010 15:10:52 -0700 - rev 56184
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 600016 - TM: Recognize that the result of Math.floor is an integer (r=nnethercote)
df05d03542f711f92a9ea2faab04cd2b3f2af82c: Bug 572798 - add LIR_callv (r=nnethercote)
Edwin Smith <edwsmith@adobe.com> - Thu, 23 Sep 2010 15:08:15 -0400 - rev 54741
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 572798 - add LIR_callv (r=nnethercote) Adds LIR_callv for calls to helper functions that return void. Added a ValidateWriter check that LIR_callv to be paired with ARGTYPE_V, plus checks for the other obvious pairings, plus a check that callv must not call a _pure=1 function. getCallOpcode() returns LIR_callv for ARGTYPE_V, as expected. This means that some calls will return LTy_V from LIns::retType(), as expected, but unlike before. This in turn can cause a ValidateWriter error if an instruction uses the result of a void call. (after all, that's the point). Each backend was modified to not assign a register or save the result of a void call.
f9a5fbc24118fdf25fe62ae443ac9dff88e03539: Add more calling-convention tests. [Bug 596545] [r=nnethercote]
Jacob Bramley <Jacob.Bramley@arm.com> - Thu, 23 Sep 2010 10:31:16 +0100 - rev 54740
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Add more calling-convention tests. [Bug 596545] [r=nnethercote]
793fa56233c04283d89db3525ba089af51e9ee37: Bug 542905 - cse chains should be cleared more selectively in case of labels. r=nnethercote.
William Maddox <wmaddox@adobe.com> - Wed, 22 Sep 2010 19:04:48 -0700 - rev 54584
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 542905 - cse chains should be cleared more selectively in case of labels. r=nnethercote.
b95bdfa39e4e7de54a4eb0b6f4223c0a73d98231: Fix long-range offsets on ARM. [Bug 596923] [r=nnethercote]
Jacob Bramley <Jacob.Bramley@arm.com> - Mon, 20 Sep 2010 07:58:53 +0100 - rev 54444
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Fix long-range offsets on ARM. [Bug 596923] [r=nnethercote]
e6a9a04ed86a0b97cc705c86273969101318406c: Remove dependency on Bash 3 regex support and simplify test selection for soft-float platforms. [Bug 535709] [r=nnethercote]
Jacob Bramley <Jacob.Bramley@arm.com> - Thu, 16 Sep 2010 16:27:45 +0100 - rev 54443
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Remove dependency on Bash 3 regex support and simplify test selection for soft-float platforms. [Bug 535709] [r=nnethercote]
27cb341980f2960f0f606abdc2d0f9b286d7a35f: Bug 590553 - asm_ui2d on x86 generates illegal writes below ESP, caught by valgrind (r=nnethercote)
Edwin Smith <edwsmith@adobe.com> - Mon, 13 Sep 2010 09:13:44 -0400 - rev 54440
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 590553 - asm_ui2d on x86 generates illegal writes below ESP, caught by valgrind (r=nnethercote)
58bbf336fea2b5ca016e1b35e9b123a4016509ec: Fix inaccurate comment in ExprFilter::insBranch() (r=nnethercote+)
William Maddox <wmaddox@adobe.com> - Thu, 26 Aug 2010 17:44:45 -0700 - rev 54432
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Fix inaccurate comment in ExprFilter::insBranch() (r=nnethercote+)
3cf3d3f676e816f364d169a66de4ac92a682331e: Add commentary to further explain usage of the x87 FPU stack (r=nnethercote+)
William Maddox <wmaddox@adobe.com> - Tue, 24 Aug 2010 18:17:45 -0700 - rev 54429
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Add commentary to further explain usage of the x87 FPU stack (r=nnethercote+)
62545b0da5bcab0bad0eb7312a7835bdbb904b59: Bug 596026: PIC-like construct for GETELEM on trace. (r=jorendorff, a=nnethercote,cdleary)
Chris Leary <cdleary@mozilla.com> - Wed, 15 Sep 2010 21:33:17 -0700 - rev 54172
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 596026: PIC-like construct for GETELEM on trace. (r=jorendorff, a=nnethercote,cdleary)
b8e76c9b23d1e4a25d66adfbf974808a135bfb0a: Bug 586387 - Problem with Typed Float32 Arrays and canonicalizeNaNs() / LIR_cmovd. r=nnethercote,edwsmith.
Andreas Gal <gal@uci.edu> - Tue, 24 Aug 2010 16:31:48 -0700 - rev 51466
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 586387 - Problem with Typed Float32 Arrays and canonicalizeNaNs() / LIR_cmovd. r=nnethercote,edwsmith.
31ce22a238acd9af591f8c1725506d6abe13ef75: Bug 587916 - Cleanup of X87 FP stack code (r=nnethercote+)
William Maddox <wmaddox@adobe.com> - Tue, 24 Aug 2010 11:44:17 -0700 - rev 51463
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 587916 - Cleanup of X87 FP stack code (r=nnethercote+) 1) The "register" FST0 is the sole member of the x87regs register class. In many places, however, the code is written so as to strongly suggest that there might be multiple such registers. This patch removes such conceits, replacing expressions such as (rmask(r) & x87regs) with (r == FST0), etc. 2) prepareResultReg() has been slightly refactored to make the x87 stack fiddling a bit easier to follow and to remove a fragile assumption. 3) Do not pass the "pop" argument to asm_spill() on non-IA32 platforms. 4) Remove redundant normalization of boolean values. 5) Comment the FPU stack depth consistency check.
736c6cd5354161cfb16cc02c9f2ad75ef852632d: Bug 584935 - Speed up register iteration loops by using BSF instruction (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Thu, 19 Aug 2010 16:04:19 -0400 - rev 51131
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 584935 - Speed up register iteration loops by using BSF instruction (r=nnethercote+) Changed all the register iteration loops to use lsbSet/msbSet functions that use fast find-first-bit intrinsics when available. Typical loops of the form: for (Register r = FirstReg; r <= LastReg; r = nextReg(r)) if (predicate(r)) /* use r */ were transformed by replacing the per-iteration predicate with a single mask calculation, then iterating through only the 1 bits in the mask: RegisterMask set = /* calculate predicate with bitmask ops */; for (Register r = lsReg(set); set; r = lsNextReg(set)) /* use r */ Iteration can be low-to-hi with lsReg/lsNextReg, or hi-to-low with msReg/msNextReg. Primitives are provided for 32 and 64-bit masks. PPC and MIPS need a 64-bit mask, for example, even on 32-bit systems. Refactoring details: I renamed msbSet() to msbSet32() as part of adding [msb|lsb]Set[32|64], which affected the AccSet code trivially. I used if (sizeof(RegisterMask) == 4) to choose between 32 and 64bit implementations, counting on a sane compiler to strip out the provably dead path. An alternative would be to move the definitions of lsReg() and msReg() to NativeXXX.h, after the RegisterMask typedef, allowing backends to hardcode the choice. Given we have six backends and one more on the way, it seemed better to centralize the code and also avoid more ifdefs. I moved the definitions of msbSet/lsbSet to nanojit.h, where other such helpers already live. It didn't seem appropriate to keep adding to LIR.h since the helpers will now be used in several places in nanojit. RegAlloc::managed is now set in Assembler.cpp instead of each backend; six lines of code replaced by one. prevreg() was dead after these changes. Additionally, I hand-inlined nextreg() in the other backends, because the usage was highly specialized -- those call sites depended on nextreg being reg+1, (or reg+2) not some generic iteration. I removed RegAlloc::countActive() since the only case was testing countActive() == 0, which is equivalent to activeMask() == 0.
077c0836d5db5f094490f36f6972cd6805ae1355: Bug 584219 should use "signed bit" for signed bit-field r=nnethercote
Ginn Chen <ginn.chen@sun.com> - Thu, 19 Aug 2010 16:08:53 +0800 - rev 51130
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 584219 should use "signed bit" for signed bit-field r=nnethercote
4b999a6aca4eaf7419a1490a3ca27b91f19a52eb: b=586387; Problem with Typed Float32 Arrays and canonicalizeNaNs() / LIR_cmovd; r=nnethercote/edwsmith
Andreas Gal <gal@uci.edu> - Fri, 13 Aug 2010 14:57:50 -0700 - rev 51051
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
b=586387; Problem with Typed Float32 Arrays and canonicalizeNaNs() / LIR_cmovd; r=nnethercote/edwsmith
6b2ac4552e1afc1d28e702a80f596e0320ab2165: b=586387; Problem with Typed Float32 Arrays and canonicalizeNaNs() / LIR_cmovd; r=nnethercote/edwsmith
Andreas Gal <gal@uci.edu> - Sat, 14 Aug 2010 10:52:08 -0700 - rev 50602
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
b=586387; Problem with Typed Float32 Arrays and canonicalizeNaNs() / LIR_cmovd; r=nnethercote/edwsmith
62e0a9c018299549d4060b80b99d359982312fe7: Bug 580534 - implement LIR_cmovd. r=nnethercote.
Andreas Gal <gal@mozilla.com> - Tue, 03 Aug 2010 20:39:21 -0700 - rev 48819
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 580534 - implement LIR_cmovd. r=nnethercote.
ab3e0cc11532b3bab0340021e64a2126a483eb33: Fixed initialization of Assembler::nHints (bug 581702, r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Wed, 28 Jul 2010 11:24:25 -0400 - rev 48816
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Fixed initialization of Assembler::nHints (bug 581702, r=nnethercote+)
adce1efdf651eb1283bdb0063d95916d7fcecbdd: Bug 576363 - Qualify accesses to dependent base in js. r=nnethercote.
Eli Friedman <sharparrow1@yahoo.com> - Sun, 04 Jul 2010 19:12:15 -0700 - rev 47590
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 576363 - Qualify accesses to dependent base in js. r=nnethercote.
b2434099c8fe45c9701c17c603cb23854bdfdd58: Add a test for LIR_std2f (and LIR_ldf2d). [Bug 572679] [r=nnethercote]
Jacob Bramley <Jacob.Bramley@arm.com> - Fri, 18 Jun 2010 11:23:32 +0100 - rev 47554
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Add a test for LIR_std2f (and LIR_ldf2d). [Bug 572679] [r=nnethercote]
cad8520cb159589de9a48c37b01889e95a2b5bbf: Add LIR_(add|sub|mul)jovi support to the ARM back-end. [Bug 571202] [r=nnethercote]
Jacob Bramley <Jacob.Bramley@arm.com> - Mon, 14 Jun 2010 15:42:46 +0100 - rev 47551
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Add LIR_(add|sub|mul)jovi support to the ARM back-end. [Bug 571202] [r=nnethercote]
650a162fbb0045cdf623e5d8fa7f9200fb11ac9e: bug 572678 - valgrind integration with conservative GC. r=nnethercote
Igor Bukanov <igor@mir2.org> - Mon, 21 Jun 2010 14:22:32 +0200 - rev 47512
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
bug 572678 - valgrind integration with conservative GC. r=nnethercote
47136f7328b5a8fd42f8dae8f5210a5f8b935891: Bug 571595: make sure dense_grow always gets inlined, r=nnethercote
David Mandelin <dmandelin@mozilla.com> - Mon, 21 Jun 2010 11:23:39 -0700 - rev 47481
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 571595: make sure dense_grow always gets inlined, r=nnethercote
47e02e48437a9ed3f2afd43b4abe14e3a320adb6: Bug 562152 - Remove typedef LInsp (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Thu, 10 Jun 2010 11:22:18 -0400 - rev 47432
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 562152 - Remove typedef LInsp (r=nnethercote+)
25b12e7a9983bfd4952a2467db2bb22905c1079c: Extend the inappropriately-limited FSTD and FSTS range. [Bug 570694] [r=nnethercote]
Jacob Bramley <Jacob.Bramley@arm.com> - Wed, 09 Jun 2010 09:10:19 +0100 - rev 47422
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Extend the inappropriately-limited FSTD and FSTS range. [Bug 570694] [r=nnethercote]
42fc5e98b20eaa00d1b00f302e8c814dc4cbcc4e: Bug 527512 - ExprFilter missing trivial optimizations for ugt, ult (r=edwsmith,nnethercote)
Steven Johnson <stejohns@adobe.com> - Thu, 03 Jun 2010 10:35:26 -0700 - rev 47395
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 527512 - ExprFilter missing trivial optimizations for ugt, ult (r=edwsmith,nnethercote)
9b7c90737b113570b9e52d30d5a69fbd069b69d5: Bug 527512 - ExprFilter missing trivial optimizations for ugt, ult (r=edwsmith,nnethercote)
Steven Johnson <stejohns@adobe.com> - Thu, 03 Jun 2010 10:35:26 -0700 - rev 43276
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 527512 - ExprFilter missing trivial optimizations for ugt, ult (r=edwsmith,nnethercote)
0090cce519265f90fded8aa329ce169894440c53: Bug 560926 - Add support for arithmetic with branch on overflow (r=nnethercote+ edwsmith+ wmaddox+)
Edwin Smith <edwsmith@adobe.com> - Wed, 26 May 2010 14:03:14 -0400 - rev 43268
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 560926 - Add support for arithmetic with branch on overflow (r=nnethercote+ edwsmith+ wmaddox+) Patch to add 'j' (branch) variants of the exit-on-overflow instructions, plus a few dependents: 32-bit: LIR_addjovi LIR_subjovi LIR_muljovi 64-bit: LIR_subq LIR_addjovq LIR_subjovq
e41c02431fa33836bcac4115d69f7f04ba087ec4: Bug 566759 - Remove LOpcode aliases once they are no longer used (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Wed, 19 May 2010 09:34:07 -0400 - rev 42757
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 566759 - Remove LOpcode aliases once they are no longer used (r=nnethercote+)
900dfe71c47624a7897cb954303a4986fdfd429f: Bug 564941 - Rename LOpcode enums in lirasm.cpp (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Mon, 10 May 2010 21:00:15 -0400 - rev 42664
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 564941 - Rename LOpcode enums in lirasm.cpp (r=nnethercote+)
00131cfadbd64b5b9394b4824031ac55deb7e358: Bug 559972 - Rename LOpcode enums in NativeMIPS.cpp (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Mon, 10 May 2010 20:10:47 -0400 - rev 42663
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 559972 - Rename LOpcode enums in NativeMIPS.cpp (r=nnethercote+)
b0919f5cbe4bfdcba36a9ce8cac9b7cfc0637770: Bug 559974 - Rename LOpcode enums in NativeSparc.cpp (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Mon, 10 May 2010 20:09:05 -0400 - rev 42662
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 559974 - Rename LOpcode enums in NativeSparc.cpp (r=nnethercote+)
63c4aabbc6158cc9644d308c1f4d874ac60bdc17: Bug 559973 - Rename LOpcode enums in NativePPC.cpp (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Fri, 07 May 2010 08:39:56 -0400 - rev 42661
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 559973 - Rename LOpcode enums in NativePPC.cpp (r=nnethercote+)
e090f2679b68883ad2bbbdf856b52256c10435a7: Bug 559971 - Rename LOpcode enums in NativeARM.cpp (r=nnethercote+ f=jacob.bramley+)
Edwin Smith <edwsmith@adobe.com> - Thu, 06 May 2010 09:32:06 -0400 - rev 42658
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 559971 - Rename LOpcode enums in NativeARM.cpp (r=nnethercote+ f=jacob.bramley+)
f49dca159319739302bcf51719906a5af290eb18: Rename LIR opcodes in Nativei386.cpp (bug 559977 r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Tue, 27 Apr 2010 20:05:58 -0400 - rev 41812
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Rename LIR opcodes in Nativei386.cpp (bug 559977 r=nnethercote+)
a653310f9fd40ba0b2a65ff54d3419c84b8efd0d: Rematerialize add(r,imm) operations on x64 and i386 (bug 555255 r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Tue, 27 Apr 2010 10:00:27 -0400 - rev 41806
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Rematerialize add(r,imm) operations on x64 and i386 (bug 555255 r=nnethercote+) This patch detects when LEA can be used to rematerialize an add operation instead of spilling it. Sub and lsh could be supported too, if it would help (see notes in canRematLEA()).
7061e03640b8b592348569e19c69dfaced0e58e3: Bug 560578 - canRemat() and asm_restore() should agree (r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Fri, 23 Apr 2010 10:26:12 -0400 - rev 41787
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Bug 560578 - canRemat() and asm_restore() should agree (r=nnethercote+) Added comments clarifying the contract between canRemat() and asm_restore(), and fixed the ARM, MIPS, PPC, and Sparc backends so canRemat() doesn't return true for instructions that asm_restore() doesn't handle.
ec40d8d90ddb86bbe78bfb004ee99d6dbf2637a6: Rename LIR opcodes once more, using new C convention (bugs 555633, bug 557877 r=nnethercote+)
Edwin Smith <edwsmith@adobe.com> - Wed, 21 Apr 2010 21:13:17 -0400 - rev 41256
Push 1 by shaver@mozilla.com at Tue, 04 Jan 2011 17:58:04 +0000
Rename LIR opcodes once more, using new C convention (bugs 555633, bug 557877 r=nnethercote+) This patch un-does the intel names that have been used so far within nanojit, updates the aliases in LIR.h, and fixes the names used in lirasm tests. Host VM's will need to run rename_LIR.pl from bug 504506 if they have already begun using the intel names; aliases are not provided for them in LIR.h.