The break is outside the if, so if it succeeds or not this will always
stop after the first loop iteration instead of trying more allocators if
the first one fails.
Closes GH-11306.
Avoid missing possible candidates due to the large address range of the free segment.
Eg,
48000000-49400000 r-xs 08000000 00:0f 39322841 segment1
7ffff2ec8000-7ffff2f49000 rw-p 00000000 00:00 0 segment2
7ffff6fae000-7ffff735c000 r-xp 00200000 08:02 11538515 /usr/local/sbin/php-fpm
original code will miss the opportunity between [7ffff2ec** - 7ffff2ec8000].
Fix issue #11265.
Signed-off-by: Long, Tao <tao.long@intel.com>
Signed-off-by: Dmitry Stogov <dmitrystogov@gmail.com>
The block optimizer pass allows the use of sources of the preceding
block if the block is a follower and not a target. This causes issues
when trying to remove FREE instructions: if the source is not in the
block of the FREE, then the FREE and source are still removed. Therefore
the other successor blocks, which must consume or FREE the temporary,
will still contain the FREE opline. This opline will now refer to a
temporary that doesn't exist anymore, which most of the time results in
a crash. For these kind of non-local scenarios, we'll let the SSA
based optimizations handle those cases.
Closes GH-11251.
Normally, we add classes without parents (and no interfaces or traits) directly
to the class map, early binding the class. However, if the same class has
already been registered, we would instead just add a ZEND_DECLARE_CLASS
instruction and let the handler throw a duplicate class declaration exception.
However, with opcache, if on the next request the files are included in the
opposite order, we won't perform early binding. To fix this, create a
ZEND_DECLARE_CLASS_DELAYED instruction instead and handle classes without
parents accordingly, skipping any linking for classes that are already linked in
delayed early binding.
Fixes GH-8846
Once code is emitted to JIT buffer, hint the hardware to
demote the corresponding cache lines to more distant level
so other CPUs can access them more quickly.
This gets nearly 1% performance gain on our workload.
Signed-off-by: Xue,Wang <xue1.wang@intel.com>
Signed-off-by: Tao,Su <tao.su@intel.com>
Signed-off-by: Hu,chen <hu1.chen@intel.com>
If we bind the class to the runtime slot even if we're not the ones who have
performed early binding we'll miss the redeclaration error in the
ZEND_DECLARE_CLASS_DELAYED handler.
Closes GH-11226
There is a typo which causes the AND and OR range inference to infer a
wider range than necessary. Fix this typo. There are many ranges for
which the inference is too wide, I just picked one for AND and one for
OR that I found through symbolic execution.
In this example test, the previous range inferred for test_or was [-27..-1]
instead of [-20..-1].
And the previous range inferred for test_and was [-32..-25]
instead of [-28..-25].
Closes GH-11170.
Max length of a single trace. A long trace generates long JITTed
code, which influences the performance slightly.
opcache.jit_max_trace_length range is [4,1024], the default value
is 1024.
Reviewed-by: Su, Tao <tao.su@intel.com>
Signed-off-by: Wang, Xue <xue1.wang@intel.com>
* ext/opcache/config.m4: new --with-opcache-capstone flag.
Until now, libcapstone has been detected "automagically" and used for
JIT disassembly whenever it is available on the system used to compile
PHP. This can have some unintended consequences, however: many users
have capstone installed for some other purpose, and are surprised to
find that PHP breaks when capstone is later uninstalled.
To address this, we have introduced a new --with-opcache-capstone flag
that is disabled by default, and that makes the user's preference
explicit. It is ignored unless the JIT is enabled.
* ext/opcache: drop support for the oprofile JIT profiler.
Recently we have replaced the "automagic" detection of capstone at
build time with a --with-opcache-capstone flag. The detection of
oprofile causes similar problems and would likely have the same
solution; however, it was suggested that we might remove oprofile
altogether. So, this commit removes it:
* Remove the detection bits from ext/opcache/config.m4.
* Drop HAVE_OPROFILE ifdef blocks.
* Delete ext/opcache/jit/zend_jit_oprofile.c.
* Undefine the ZEND_JIT_DEBUG_OPROFILE constant.
Commit a21195650e fixed a leak by adding a TSRM destructor for the
JIT globals in ZTS mode. In case the main thread shuts down the TSRM, it
will call all the destructors. The JIT globals destructor will be
invoked, but will always access the main thread globals using JIT_G.
This means that instead of freeing the JIT globals in the different
threads, the one in the main thread is freed repeatedly over and over,
crashing PHP. Fix it by always passing the pointer instead of relying on
JIT_G.
Closes GH-10835.
