Table of Contents

PHP RFC: 64 bit platform improvements for string length and integer in zval

Introduction

Current PHP zval datatype implementation uses long datatype to handle signed integer and int datatype to handle string length. The most 64 bit capable platforms PHP is used on are LP64 (Linux/most Unix-like), LLP64 (Windows), ILP64 (SPARC64). The data model used in PHP for the relevant datatypes on those platforms looks as follows

string size signed integer
Platform int long
LP64 32 bit 64 bit
LLP64 32 bit 32 bit
ILP64 64 bit 64 bit


Regarding PHP that means today, even on 64 bit platforms the following features lack on consistency


Thus, the current situation contains a roadblock on the way to the overall consistent 64 bit platforms support and further improvement of PHP.

To bring everything inline, a dynamic types model is being suggested. Besides the platform inconsistency elimination, this will enable several further moves (see the Future scope). Performance improvements for 64 bit platforms, just to name one, could be then the subject of a new RFC and would have the base for the further development.

Proposal

The following datatypes are suggested for use in zval

Platform string size signed integer
LP64 size_t int64_t
LLP64 size_t __int64
ILP64 size_t int64_t


Those datatypes are guaranteed to be 64 bit which makes PHP work consistent on any 64 bit platforms. The usage of this datatypes is integrated into the appropriate places across PHP. The size of zval, with default field alignment, is expected to grow by 4 bytes on LLP64 platforms only.

For the consistent LFS support, a set of portable datatypes and macros was invented and integrated.

For consistent 64 bit numeric hash keys support, the appropriate 64 bit unsigned was integrated.

The usage of long datatype continues on 32 bit platforms. The only change there is the usage of size_t for the string length (which is effectively a 32 bit unsigned).

No further configure options are needed, the platform and compiler will be automatically recognized and an appropriate set of datatypes and macros will be activated.

Relevant headers

New portable datatypes

Old datatype New datatype Comment
int, uint, zend_uint, size_t, long zend_size_t Overall datatype to handle object size and string length in zval. Aliased as php_size_t in php.h
long zend_int_t Overall datatype to handle integers in zval. Aliased with php_int_t in php.h
ulong, unsigned long zend_uint_t Overall datatype to handle numeric hash indexes and other situations with need on an unsigned. Aliased with php_uint_t in php.h
off_t, _off_t zend_off_t Overall datatype to handle file offsets. Corresponding portable macros have to be used.
struct stat, struct _stat, struct _stat64 zend_stat_t Overall datatype to handle the FS info. Corresponding portable macros have to be used.


New portable macros for LFS support

function(s) Alias Comment
stat, _stat64 zend_stat for use with zend_stat_t
fstat, _fstat64 zend_fstat for use with zend_stat_t
lseek, _lseeki64 zend_lseek for use with zend_off_t
ftell, _ftelli64 zend_ftell for use with zend_off_t
fseek, _fseeki64 zend_fseek for use with zend_off_t


New portable macros for integers

function(s) Alias Comment
snprintf with “%ld” or “%lld”, _ltoa_s, _i64toa_s ZEND_ITOA for use with zend_int_t
atol, atoll, _atoi64 ZEND_ATOI for use with zend_int_t
strtol, strtoll, _strtoi64 ZEND_STRTOI for use with zend_int_t
strtoul, strtoull, _strtoui64 ZEND_STRTOUIfor use with zend_int_t
abs, llabs, _abs64 ZEND_ABS for use with zend_int_t
- ZEND_INT_MAX Aliased with PHP_INT_MAX in php.h, replaces LONG_MAX where appropriate
- ZEND_INT_MIN Aliased with PHP_INT_MIN in php.h, replaces LONG_MIN where appropriate
- ZEND_UINT_MAX ULONG_MAX
- SIZEOF_ZEND_INT Replaces SIZEOF_ZEND_LONG where appropriate
- ZEND_SIZE_MAX Max value of zend_size_t


Semantical macro renamings

Old New Comment
Z_STRLEN Z_STRSIZE as well the whole Z_STRLEN_* family
IS_LONG IS_INT
RETURN_LONG RETURN_INT
RETVAL_LONG RETVAL_INT
Z_LVAL Z_IVAL as well the whole Z_LVAL_* family
LITERAL_LONG LITERAL_INT
REGISTER_LONG_CONSTANT REGISTER_INT_CONSTANT
REGISTER_MAIN_LONG_CONSTANT REGISTER_MAIN_INT_CONSTANT
ZEND_SIGNED_MULTIPLY_LONG ZEND_SIGNED_MULTIPLY_INT
... ...

Generally speaking, every occurence mentioning “long” in macros or function names should be replaced with a corresponding neutral keyword, suggested “int”, in further like “lval” with “ival”, etc.


Accepting values with zend_parse_parameters()

Old New Comment
“s” “S” accept string argument, the length has to be declared as php_size_t (or zend_size_t)
“p” “P” accept path argument, the length has to be declared as php_size_t (or zend_size_t)
“l” “i” to accept integer argument, the internal var has to be declared as php_int_t (inside PHP) or zend_int_t (inside Zend)
“L” “I” to accept integer argument with range check, the internal var has to be declared as php_int_t (inside PHP) or zend_int_t (inside Zend)


Clang PHP checker

Johannes Schlüter did a great work on developing a clang plugin allowing static analyze of the zpp calls. Steps was undertaken to make it available on windows as well.


spprintf formats

New spprintf modifier 'p' was implemented to platform independently output php_int_t datatype. That modifier can be used with 'd', 'u', 'x' and 'o' printf format specs with spprintf, snprintf and the wrapping printf implementations.

Portable macros to use with printf

Format spec Macros Comment
%I64d, “%” PRId64, %ld ZEND_INT_FMT for use with zend_int_t
%I64u, “%” PRIu64, %lu ZEND_UINT_FMT for use with zend_uint_t

This modifier is of course available in all the spprintf/snprintf derivatives. Any of the introduced new datatypes can be used with the appropriate format spec.


Backward Incompatible Changes

Proposed PHP Version

PHP6 (or whatever next major is called)

Impacts

Open Issues

Some dead SAPIs are present in the core. They was not ported. A decision based on whether the authors are willing to support them has to be met. Then porting or removal of those SAPIs can be scheduled. The separate RFC https://wiki.php.net/rfc/removal_of_dead_sapis was created to handle this issue.

Unaffected PHP Functionality

It has to do with squeezing anything possible from the 64 bit platforms, for maximal PHP benefit. No real features are going to be changed, removed or added to the PHP language.

Performance

Performance test done on the base of master and str_size_and_int64 can be found here http://windows.php.net/downloads/snaps/ostc/pftt/perf/results-20140411-masterr6c2f7bc-str_size_and_int64raa0c920.html. There is no performance difference between two (between 0-2%).

/*

Some performance comparsion

PHP Version Wordpress Drupal Joomla
str_size_and_int64-x86 NoCache: 68 NoCache: 70 NoCache: 53
- Cache: 284 Cache: 393 Cache: 127
php-5.5.8-nts-Win32-VC11-x86 NoCache: 67 NoCache: 69 NoCache: 53
- Cache: 280 Cache: 390 Cache: 125
str_size_and_int64-x64 NoCache: 58 NoCache: 64 NoCache: 50
- Cache: 313* Cache: 348* Cache: 100*
php-5.5.8-nts-Win32-VC11-x64 NoCache: 59 NoCache: 65 NoCache: 51
- Cache: 270* Cache: ** Cache: **

The numbers here are the test scores one already might have seen in the other performance tests.

*, ** Some issues with the x64 versions of 5.5.8 and str_size_and_int64 when testing with opcache enabled. However issues of this kind are well known on windows (for instance #64926) and are due to some unluckily choosen memory address. So the cause persists in the mainstream and is not because of this patch. \\*/

Migration path for PECL extensions

Tutorial, tools and compatibility header to ease the migration of the PECL extensions are available. The goal is to make the same source in the new semantic compatible with older PHP versions.

Example on accepting parameters with zpp

	php_int_t i0, i1;
	char *s0, p0;
	php_size_t s0_len, p0_len;
 
	if(zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "iISP", &i0, &i1, &s0, &s0_len, &p0, &p0_len) == FAILURE) {
		return;
	}

Example on printf specs usage

	php_int_t i0;
 
	if(zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "i", &i0) == FAILURE) {
		return;
	}
 
	if (INT_MAX < i0 || INT_MIN > i0) {
		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Value '" ZEND_INT_FMT "' is out of range", i0);
		return;
	}

Example on printf specs usage (no BC)

	php_error_docref(NULL TSRMLS_CC, E_WARNING, "Value '%pd' is out of range", i0);

Example proper check of string size

	char *s0;
	php_size_t s0_len;
	php_int_t max_len;
 
	if(zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "Si", &s0, &s0_len, &max_len) == FAILURE) {
		return;
	}
 
	if (max_len >= 0 && s0_len > max_len) {
		return;
	}


Example with some renamed macros

char *dup_substr(zval *s, zval *i)
{
	php_size_t len;
	php_int_t max;
	char ret;
 
	convert_to_string(s);
	convert_to_int(i);
 
	len = Z_STRSIZE_P(s);
	max = Z_IVAL(i);
 
	if (max < 0 || max >= 0 && max > len)
	{
	return NULL;
	}
 
	ret = emalloc((max + 1) * sizeof(char));
 
	if (!ret) {
		return NULL;
	}
 
	memmove(ret, Z_STRVAL_P(s), max);
	ret[max] = '\0';
 
	return ret;
}

Future Scope

Vote

Accept this RFC for PHP6 (or whatever next major is called)
Real name Yes No
ab (ab)  
adobkin (adobkin)  
ajf (ajf)  
andi (andi)  
arpad (arpad)  
auroraeosrose (auroraeosrose)  
bjori (bjori)  
brianlmoon (brianlmoon)  
bwoebi (bwoebi)  
chobieeee (chobieeee)  
chx (chx)  
crodas (crodas)  
danielc (danielc)  
datibbaw (datibbaw)  
davey (davey)  
derick (derick)  
dm (dm)  
dmitry (dmitry)  
felipe (felipe)  
frozenfire (frozenfire)  
googleguy (googleguy)  
guilhermeblanco (guilhermeblanco)  
gwynne (gwynne)  
iliaa (iliaa)  
indeyets (indeyets)  
jedibc (jedibc)  
jpauli (jpauli)  
jwage (jwage)  
kalle (kalle)  
klaussilveira (klaussilveira)  
krakjoe (krakjoe)  
kriscraig (kriscraig)  
laruence (laruence)  
levim (levim)  
lstrojny (lstrojny)  
malukenho (malukenho)  
mattficken (mattficken)  
mbeccati (mbeccati)  
nikic (nikic)  
pajoye (pajoye)  
philip (philip)  
rasmus (rasmus)  
salathe (salathe)  
szarkos (szarkos)  
thorstenr (thorstenr)  
treffynnon (treffynnon)  
tyrael (tyrael)  
uw (uw)  
yohgaki (yohgaki)  
zeev (zeev)  
zhangzhenyu (zhangzhenyu)  
Final result: 28 23
This poll has been closed.


Merge strategy
Real name after the vote, master phpng
ab (ab)  
adobkin (adobkin)  
ajf (ajf)
andrey (andrey)  
auroraeosrose (auroraeosrose)  
bjori (bjori)  
brianlmoon (brianlmoon)  
bwoebi (bwoebi)  
chobieeee (chobieeee)  
chx (chx)  
crodas (crodas)  
davey (davey)  
derick (derick)  
dm (dm)  
dmitry (dmitry)  
googleguy (googleguy)  
guilhermeblanco (guilhermeblanco)  
gwynne (gwynne)  
iliaa (iliaa)  
indeyets (indeyets)  
jedibc (jedibc)
jgmdev (jgmdev)  
jpauli (jpauli)
jwage (jwage)  
kalle (kalle)
klaussilveira (klaussilveira)  
krakjoe (krakjoe)  
kriscraig (kriscraig)
laruence (laruence)  
levim (levim)  
lstrojny (lstrojny)
malukenho (malukenho)  
mattficken (mattficken)  
mbeccati (mbeccati)  
nicolas (nicolas)  
nikic (nikic)  
pajoye (pajoye)
philip (philip)  
philstu (philstu)  
pollita (pollita)  
rasmus (rasmus)  
salathe (salathe)  
stas (stas)  
szarkos (szarkos)  
thorstenr (thorstenr)  
tony2001 (tony2001)  
treffynnon (treffynnon)
tyrael (tyrael)  
uw (uw)  
Final result: 17 40
This poll has been closed.


The RFC is considered approved with 50%+1 acceptance.
The vote starts on May 13th 2014 and ends on May 20th 2014 due to the issues with the mailing lists, the vote period is extended till May 26th 2014.

Patches and Tests

Feature branch
Backport branch of PHP 5.6 (for comparisons only)
Windows builds
Test reports

References

Patch progress page
PECL porting docs'n'tools

Initial discussion brought up by Anthony Ferrara
Discussion after implementation start
Discussions of the previous RFC version

Implementation

After the project is implemented, this section should contain

  1. the version(s) it was merged to
  2. a link to the git commit(s)
  3. a link to the PHP manual entry for the feature

Rejected Features

-