1
0
mirror of https://github.com/php/php-src.git synced 2026-04-18 13:31:27 +02:00
Commit Graph

3885 Commits

Author SHA1 Message Date
Dmitry Stogov
e868ded6ef Don't call _zend_hash_index_find() for packed arrays 2021-11-08 13:13:23 +03:00
Dmitry Stogov
7bf63243e1 JIT: Fixed incorrect guard elimination 2021-11-08 12:28:36 +03:00
Dmitry Stogov
34d22acc7e Tracing JIT: Fixed incorrect guard elimination 2021-11-03 23:37:00 +03:00
Dmitry Stogov
66d93c63ff Fixed type inference for FETCH_DIM_W without use 2021-11-03 19:44:38 +03:00
Dmitry Stogov
2605911272 Tracing JIT: Fixed reference counting when escape because of IS_UNDEF element (test) 2021-11-01 22:19:10 +03:00
Dmitry Stogov
d7ac39b02d Tracing JIT: Fixed reference counting when escape because of IS_UNDEF element 2021-11-01 22:14:12 +03:00
Dmitry Stogov
5c43e0c0ee Fixed incorrect assumption about reference counting 2021-11-01 20:28:05 +03:00
Dmitry Stogov
f1e6a7d83d JIT: Fixed numeric string index handling 2021-11-01 18:50:32 +03:00
Dmitry Stogov
cd32b47235 JIT: Fixed register allocation in case of integer overflow 2021-11-01 16:46:41 +03:00
Nikita Popov
2e29817005 Handle FETCH_DIM_R after FETCH_DIM_FUNC_ARG in inference
This can happen if a call is optimized, but FETCH_DIM_FUNC_ARG
cannot be converted to FETCH_DIM_R because it uses an UNUSED op2,
which is not supported by FETCH_DIM_R.

Fixes oss-fuzz 6144185837682688.
2021-11-01 14:26:36 +01:00
Nikita Popov
f8ec5a1d4c Fix range inference hang
We shouldn't switch from range to no range for ZEND_DIV and instead
explicitly return an overflowing range. Otherwise the range will
not actually get updated during widening, and we'll perform
essentially infinite narrowing.

Fixes oss-fuzz #40566.
2021-11-01 12:31:34 +01:00
Dmitry Stogov
29c8c1e955 Tracing JIT: Fixed possible endless loop when escape from ZEND_CALL_TOP frame 2021-10-25 12:10:25 +03:00
Dmitry Stogov
d3251632ad Tracing JIT: Fixed incorrect assumption about in-memeory zval type 2021-10-22 14:38:38 +03:00
Dmitry Stogov
c7e974fc4e Tracing JIT: Fixed incorrect assumtion about temporary variable types clobbered by *_ROPE instructions 2021-10-21 21:35:57 +03:00
Dmitry Stogov
b69c6ba9df Use right frame 2021-10-20 23:50:33 +03:00
Dmitry Stogov
95c0dfc57f JIT: Fixed megamorphic call detection 2021-10-20 22:16:25 +03:00
Dmitry Stogov
38626dbaad JIT: Fixed memory leak 2021-10-20 10:38:47 +03:00
Dmitry Stogov
6bd5271c62 Fixed incorrect optimization that may cause JIT assertion 2021-10-18 23:18:52 +03:00
Dmitry Stogov
d3063c02c6 Tracing JIT: Fixed memory leak 2021-10-18 22:14:53 +03:00
Dmitry Stogov
e59d0a7151 JIT: Fixed DASM_S_RANGE_VREG error 2021-10-18 17:58:42 +03:00
Dmitry Stogov
13d1244575 Tracing JIT: Fixed ADD with two references to arrays 2021-10-18 17:26:36 +03:00
Dmitry Stogov
6021e72143 Tracinf JIT: Prevnt generation code for instruction that with cyclic dependency
e.g. $a[] += $a is compiled into ASSIGN_DIM_OP+OP_DATA, where OP_DATA.op1_use depends on ASSIGN_DIM_OP.op1_def
2021-10-18 16:53:45 +03:00
Nikita Popov
9ebe8494b8 Don't replace tmp with cv in YIELD argument
For by-ref generators, these may have different behavior.

Fixes oss-fuzz 6059739298004992.
2021-10-12 14:26:53 +02:00
Dmitry Stogov
a4c41d45cd Fixed type inference for COPY_TMP 2021-10-11 16:56:07 +03:00
Dmitry Stogov
ccdda69a18 JIT: Fixed usage of incorrect specialized zval destructor 2021-10-11 13:00:23 +03:00
Dmitry Stogov
ed8ec9d71e Fixed type inference (ASSIGN_OP with typed reference may cause type conversion) 2021-10-11 10:48:49 +03:00
Dmitry Stogov
bbd875268a JIT: Fixed incorrect double/long register hinting 2021-10-08 15:46:49 +03:00
Dmitry Stogov
b279de416e JIT: Fixed extension handling 2021-10-08 14:06:32 +03:00
Dmitry Stogov
45b127ad4a Flush error messages emitted during preloading 2021-10-08 12:34:25 +03:00
Dmitry Stogov
7cea20f937 JIT: Fixed FETCH_LIST_R on string 2021-10-07 22:13:39 +03:00
Dmitry Stogov
b47a48ff80 Fixed bug #81512 (Unexpected behavior with arrays and JIT) 2021-10-07 21:24:38 +03:00
Nikita Popov
b564113756 Check exception after undef var notice
For simple assignments, we may not emit an exception check elsewhere,
so perform it directly after the undef var notice.
2021-10-07 16:54:26 +02:00
Nikita Popov
08c29a657d Handle undefined dim in assign_dim_helper
Not only the value can be undefined here, but the offset as well.
2021-10-07 15:35:16 +02:00
Nikita Popov
455837139e Set opline before calling undef op helper 2021-10-07 14:38:02 +02:00
Dmitry Stogov
a743fd7633 JIT: Fixed wrong comparison skip 2021-10-07 15:27:56 +03:00
Dmitry Stogov
6ab36fb6f9 JIT: Fixed incorrect reference handling in PRE_INC/DEC_OBJ 2021-10-07 14:34:18 +03:00
Dmitry Stogov
cd45bd0a81 JIT: Fixed incorrect code generation for JMPZ 2021-10-07 13:48:28 +03:00
Dmitry Stogov
8a0873020a JIT: Fixed NaN handling 2021-10-06 16:35:33 +03:00
Dmitry Stogov
f681f9075c JIT: Fixed register clobbering 2021-10-06 12:10:39 +03:00
Nikita Popov
f890c9c5e3 Fix return type verification with undef var
This was loading EG(uninitialized_value) into r0 rather than
FCARG1a.

However, if we fix this issue an existing test fails because
the undef var warning promoted to exception is not caught early
enough, so we need to explicitly check for the exception before
performing the type check.
2021-10-05 15:30:55 +02:00
Dmitry Stogov
22ef1fb832 Fixed SSA construction 2021-10-05 16:14:39 +03:00
Nikita Popov
c958adce70 POST_INC_OBJ/POST_DEC_OBJ don't have optional result
Possibly we should be encoding this in the opcode metadata, rather
than maintaining a separate list.

Fixes oss-fuzz #39503.
2021-10-05 12:05:39 +02:00
Nikita Popov
d0860f67ca Fix cache slot assignment for ASSIGN_OBJ_OP
ASSIGN_OBJ_OP stores the cache slot in OP_DATA, so this ended up
overwriting the binop opcode instread.
2021-09-30 14:34:12 +02:00
Dmitry Stogov
512dfaba3a JIT: Fixed incorrect named parameter handling 2021-09-30 14:51:43 +03:00
Dmitry Stogov
4b31cb3eb8 JIT: Fixed memory leak 2021-09-30 13:01:56 +03:00
Dmitry Stogov
7710047ed1 JIT x86: Fixed NaN handling 2021-09-28 23:49:26 +03:00
Dmitry Stogov
c30298bf57 Tracing JIT: Fixed possible memory-leak or missed destructor call.
Reference-counter of return_value may be indirectly decremented to 1.
2021-09-28 22:40:04 +03:00
Nikita Popov
97b5eeeb6c Fix leak with ASSIGN_OBJ on null
We still need to free op1 in this case.

Fixes oss-fuzz 5782176231194624 (part of #38542).
2021-09-28 16:33:11 +02:00
Dmitry Stogov
3fd1f6cdad JIT: Fixed label mess 2021-09-28 15:21:11 +03:00
Nikita Popov
1bb7ee3207 Fix ASSIGN_DIM result inference with typed refs
Same issue as with ASSIGN. Also make the handling for ASSIGN more
precise, we can only have conversions between scalar values.
2021-09-28 14:14:01 +02:00