project-base/server-skynet-source-3rd-jemalloc
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
bd87b01999
server-skynet-source-3rd-je.../include/jemalloc/internal/size_classes.sh

122 lines
3.2 KiB
Bash
Raw Normal View History

Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
#!/bin/sh
# The following limits are chosen such that they cover all supported platforms.
# Range of quanta.
lg_qmin=3
lg_qmax=4
# The range of tiny size classes is [2^lg_tmin..2^(lg_q-1)].
lg_tmin=3
# Range of page sizes.
lg_pmin=12
lg_pmax=16
Remove bashism. Submitted by Mike Hommey.
2012-03-13 02:33:59 +08:00
pow2() {
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
e=$1
pow2_result=1
while [ ${e} -gt 0 ] ; do
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
pow2_result=$((${pow2_result} + ${pow2_result}))
e=$((${e} - 1))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
done
}
cat <<EOF
/* This file was automatically generated by size_classes.sh. */
/******************************************************************************/
#ifdef JEMALLOC_H_TYPES
EOF
lg_q=${lg_qmin}
while [ ${lg_q} -le ${lg_qmax} ] ; do
lg_t=${lg_tmin}
while [ ${lg_t} -le ${lg_q} ] ; do
lg_p=${lg_pmin}
while [ ${lg_p} -le ${lg_pmax} ] ; do
Use echo instead of cat in loops in size_classes.sh This avoids fork/exec()ing in loops, as echo is a builtin, and makes size_classes.sh much faster (from > 10s to < 0.2s on mingw on my machine).
2012-04-16 22:30:23 +08:00
echo "#if (LG_TINY_MIN == ${lg_t} && LG_QUANTUM == ${lg_q} && LG_PAGE == ${lg_p})"
echo "#define SIZE_CLASSES_DEFINED"
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
pow2 ${lg_q}; q=${pow2_result}
pow2 ${lg_t}; t=${pow2_result}
pow2 ${lg_p}; p=${pow2_result}
bin=0
psz=0
sz=${t}
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
delta=$((${sz} - ${psz}))
Use echo instead of cat in loops in size_classes.sh This avoids fork/exec()ing in loops, as echo is a builtin, and makes size_classes.sh much faster (from > 10s to < 0.2s on mingw on my machine).
2012-04-16 22:30:23 +08:00
echo "/* SIZE_CLASS(bin, delta, sz) */"
echo "#define SIZE_CLASSES \\"
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
# Tiny size classes.
while [ ${sz} -lt ${q} ] ; do
Use echo instead of cat in loops in size_classes.sh This avoids fork/exec()ing in loops, as echo is a builtin, and makes size_classes.sh much faster (from > 10s to < 0.2s on mingw on my machine).
2012-04-16 22:30:23 +08:00
echo " SIZE_CLASS(${bin}, ${delta}, ${sz}) \\"
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
bin=$((${bin} + 1))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
psz=${sz}
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
sz=$((${sz} + ${sz}))
delta=$((${sz} - ${psz}))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
done
# Quantum-multiple size classes. For each doubling of sz, as many as 4
# size classes exist. Their spacing is the greater of:
# - q
# - sz/4, where sz is a power of 2
while [ ${sz} -lt ${p} ] ; do
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
if [ ${sz} -ge $((${q} * 4)) ] ; then
i=$((${sz} / 4))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
else
i=${q}
fi
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
next_2pow=$((${sz} * 2))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
while [ ${sz} -lt $next_2pow ] ; do
Use echo instead of cat in loops in size_classes.sh This avoids fork/exec()ing in loops, as echo is a builtin, and makes size_classes.sh much faster (from > 10s to < 0.2s on mingw on my machine).
2012-04-16 22:30:23 +08:00
echo " SIZE_CLASS(${bin}, ${delta}, ${sz}) \\"
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
bin=$((${bin} + 1))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
psz=${sz}
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
sz=$((${sz} + ${i}))
delta=$((${sz} - ${psz}))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
done
done
Use echo instead of cat in loops in size_classes.sh This avoids fork/exec()ing in loops, as echo is a builtin, and makes size_classes.sh much faster (from > 10s to < 0.2s on mingw on my machine).
2012-04-16 22:30:23 +08:00
echo
echo "#define NBINS ${bin}"
echo "#define SMALL_MAXCLASS ${psz}"
echo "#endif"
echo
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
lg_p=$((${lg_p} + 1))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
done
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
lg_t=$((${lg_t} + 1))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
done
Use $((...)) instead of expr. Use $((...)) for math in size_classes.h rather than expr, because it is much faster. This is not supported syntax in the classic Bourne shell, but all modern sh implementations support it, including bash, zsh, and ash.
2012-04-04 04:20:21 +08:00
lg_q=$((${lg_q} + 1))
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
done
cat <<EOF
#ifndef SIZE_CLASSES_DEFINED
# error "No size class definitions match configuration"
#endif
#undef SIZE_CLASSES_DEFINED
/*
* The small_size2bin lookup table uses uint8_t to encode each bin index, so we
* cannot support more than 256 small size classes. Further constrain NBINS to
Remove support for non-prof-promote heap profiling metadata. Make promotion of sampled small objects to large objects mandatory, so that profiling metadata can always be stored in the chunk map, rather than requiring one pointer per small region in each small-region page run. In practice the non-prof-promote code was only useful when using jemalloc to track all objects and report them as leaks at program exit. However, Valgrind is at least as good a tool for this particular use case. Furthermore, the non-prof-promote code is getting in the way of some optimizations that will make heap profiling much cheaper for the predominant use case (sampling a small representative proportion of all allocations).
2014-04-12 05:24:51 +08:00
* 255 since all small size classes, plus a "not small" size class must be
* stored in 8 bits of arena_chunk_map_t's bits field.
Simplify small size class infrastructure. Program-generate small size class tables for all valid combinations of LG_TINY_MIN, LG_QUANTUM, and PAGE_SHIFT. Use the appropriate table to generate all relevant data structures, and remove the distinction between tiny/quantum/cacheline/subpage bins. Remove --enable-dynamic-page-shift. This option didn't prove useful in practice, and it prevented optimizations. Add Tilera architecture support.
2012-02-29 08:50:47 +08:00
*/
#if (NBINS > 255)
# error "Too many small size classes"
#endif
#endif /* JEMALLOC_H_TYPES */
/******************************************************************************/
#ifdef JEMALLOC_H_STRUCTS
#endif /* JEMALLOC_H_STRUCTS */
/******************************************************************************/
#ifdef JEMALLOC_H_EXTERNS
#endif /* JEMALLOC_H_EXTERNS */
/******************************************************************************/
#ifdef JEMALLOC_H_INLINES
#endif /* JEMALLOC_H_INLINES */
/******************************************************************************/
EOF
Reference in New Issue Copy Permalink