linux-stable/lib/sort.c

273 lines
8.4 KiB
C
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 14:07:57 +00:00
// SPDX-License-Identifier: GPL-2.0
/*
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
* A fast, small, non-recursive O(n log n) sort for the Linux kernel
*
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
* This performs n*log2(n) + 0.37*n + o(n) comparisons on average,
* and 1.5*n*log2(n) + O(n) in the (very contrived) worst case.
*
* Glibc qsort() manages n*log2(n) - 1.26*n for random inputs (1.63*n
* better) at the expense of stack usage and much larger code to avoid
* quicksort's O(n^2) worst case.
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/types.h>
#include <linux/export.h>
#include <linux/sort.h>
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
/**
* is_aligned - is this pointer & size okay for word-wide copying?
* @base: pointer to data
* @size: size of each element
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
* @align: required alignment (typically 4 or 8)
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
*
* Returns true if elements can be copied using word loads and stores.
* The size must be a multiple of the alignment, and the base address must
* be if we do not have CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS.
*
* For some reason, gcc doesn't know to optimize "if (a & mask || b & mask)"
* to "if ((a | b) & mask)", so we do that by hand.
*/
__attribute_const__ __always_inline
static bool is_aligned(const void *base, size_t size, unsigned char align)
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
{
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
unsigned char lsbits = (unsigned char)size;
(void)base;
#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
lsbits |= (unsigned char)(uintptr_t)base;
#endif
return (lsbits & (align - 1)) == 0;
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
}
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
/**
* swap_words_32 - swap two elements in 32-bit chunks
* @a: pointer to the first element to swap
* @b: pointer to the second element to swap
* @n: element size (must be a multiple of 4)
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
*
* Exchange the two objects in memory. This exploits base+index addressing,
* which basically all CPUs have, to minimize loop overhead computations.
*
* For some reason, on x86 gcc 7.3.0 adds a redundant test of n at the
* bottom of the loop, even though the zero flag is stil valid from the
* subtract (since the intervening mov instructions don't alter the flags).
* Gcc 8.1.0 doesn't have that problem.
*/
static void swap_words_32(void *a, void *b, size_t n)
{
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
do {
u32 t = *(u32 *)(a + (n -= 4));
*(u32 *)(a + n) = *(u32 *)(b + n);
*(u32 *)(b + n) = t;
} while (n);
}
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
/**
* swap_words_64 - swap two elements in 64-bit chunks
* @a: pointer to the first element to swap
* @b: pointer to the second element to swap
* @n: element size (must be a multiple of 8)
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
*
* Exchange the two objects in memory. This exploits base+index
* addressing, which basically all CPUs have, to minimize loop overhead
* computations.
*
* We'd like to use 64-bit loads if possible. If they're not, emulating
* one requires base+index+4 addressing which x86 has but most other
* processors do not. If CONFIG_64BIT, we definitely have 64-bit loads,
* but it's possible to have 64-bit loads without 64-bit pointers (e.g.
* x32 ABI). Are there any cases the kernel needs to worry about?
*/
static void swap_words_64(void *a, void *b, size_t n)
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
{
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
do {
#ifdef CONFIG_64BIT
u64 t = *(u64 *)(a + (n -= 8));
*(u64 *)(a + n) = *(u64 *)(b + n);
*(u64 *)(b + n) = t;
#else
/* Use two 32-bit transfers to avoid base+index+4 addressing */
u32 t = *(u32 *)(a + (n -= 4));
*(u32 *)(a + n) = *(u32 *)(b + n);
*(u32 *)(b + n) = t;
t = *(u32 *)(a + (n -= 4));
*(u32 *)(a + n) = *(u32 *)(b + n);
*(u32 *)(b + n) = t;
#endif
} while (n);
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
}
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
/**
* swap_bytes - swap two elements a byte at a time
* @a: pointer to the first element to swap
* @b: pointer to the second element to swap
* @n: element size
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
*
* This is the fallback if alignment doesn't allow using larger chunks.
*/
static void swap_bytes(void *a, void *b, size_t n)
{
do {
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
char t = ((char *)a)[--n];
((char *)a)[n] = ((char *)b)[n];
((char *)b)[n] = t;
} while (n);
}
/*
* The values are arbitrary as long as they can't be confused with
* a pointer, but small integers make for the smallest compare
* instructions.
*/
#define SWAP_WORDS_64 (swap_func_t)0
#define SWAP_WORDS_32 (swap_func_t)1
#define SWAP_BYTES (swap_func_t)2
/*
* The function pointer is last to make tail calls most efficient if the
* compiler decides not to inline this function.
*/
static void do_swap(void *a, void *b, size_t size, swap_func_t swap_func)
{
if (swap_func == SWAP_WORDS_64)
swap_words_64(a, b, size);
else if (swap_func == SWAP_WORDS_32)
swap_words_32(a, b, size);
else if (swap_func == SWAP_BYTES)
swap_bytes(a, b, size);
else
swap_func(a, b, (int)size);
}
#define _CMP_WRAPPER ((cmp_r_func_t)0L)
static int do_cmp(const void *a, const void *b, cmp_r_func_t cmp, const void *priv)
{
if (cmp == _CMP_WRAPPER)
return ((cmp_func_t)(priv))(a, b);
return cmp(a, b, priv);
}
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
/**
* parent - given the offset of the child, find the offset of the parent.
* @i: the offset of the heap element whose parent is sought. Non-zero.
* @lsbit: a precomputed 1-bit mask, equal to "size & -size"
* @size: size of each element
*
* In terms of array indexes, the parent of element j = @i/@size is simply
* (j-1)/2. But when working in byte offsets, we can't use implicit
* truncation of integer divides.
*
* Fortunately, we only need one bit of the quotient, not the full divide.
* @size has a least significant bit. That bit will be clear if @i is
* an even multiple of @size, and set if it's an odd multiple.
*
* Logically, we're doing "if (i & lsbit) i -= size;", but since the
* branch is unpredictable, it's done with a bit of clever branch-free
* code instead.
*/
__attribute_const__ __always_inline
static size_t parent(size_t i, unsigned int lsbit, size_t size)
{
i -= size;
i -= size & -(i & lsbit);
return i / 2;
}
/**
* sort_r - sort an array of elements
* @base: pointer to data to sort
* @num: number of elements
* @size: size of each element
* @cmp_func: pointer to comparison function
* @swap_func: pointer to swap function or NULL
* @priv: third argument passed to comparison function
*
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
* This function does a heapsort on the given array. You may provide
* a swap_func function if you need to do something more than a memory
* copy (e.g. fix up pointers or auxiliary data), but the built-in swap
* avoids a slow retpoline and so is significantly faster.
*
* Sorting time is O(n log n) both on average and worst-case. While
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
* quicksort is slightly faster on average, it suffers from exploitable
* O(n*n) worst-case behavior and extra memory requirements that make
* it less suitable for kernel use.
*/
void sort_r(void *base, size_t num, size_t size,
cmp_r_func_t cmp_func,
swap_func_t swap_func,
const void *priv)
{
/* pre-scale counters for performance */
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
size_t n = num * size, a = (num/2) * size;
const unsigned int lsbit = size & -size; /* Used to find parent */
if (!a) /* num < 2 || size == 0 */
return;
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
if (!swap_func) {
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
if (is_aligned(base, size, 8))
swap_func = SWAP_WORDS_64;
lib/sort: make swap functions more generic Patch series "lib/sort & lib/list_sort: faster and smaller", v2. Because CONFIG_RETPOLINE has made indirect calls much more expensive, I thought I'd try to reduce the number made by the library sort functions. The first three patches apply to lib/sort.c. Patch #1 is a simple optimization. The built-in swap has special cases for aligned 4- and 8-byte objects. But those are almost never used; most calls to sort() work on larger structures, which fall back to the byte-at-a-time loop. This generalizes them to aligned *multiples* of 4 and 8 bytes. (If nothing else, it saves an awful lot of energy by not thrashing the store buffers as much.) Patch #2 grabs a juicy piece of low-hanging fruit. I agree that nice simple solid heapsort is preferable to more complex algorithms (sorry, Andrey), but it's possible to implement heapsort with far fewer comparisons (50% asymptotically, 25-40% reduction for realistic sizes) than the way it's been done up to now. And with some care, the code ends up smaller, as well. This is the "big win" patch. Patch #3 adds the same sort of indirect call bypass that has been added to the net code of late. The great majority of the callers use the builtin swap functions, so replace the indirect call to sort_func with a (highly preditable) series of if() statements. Rather surprisingly, this decreased code size, as the swap functions were inlined and their prologue & epilogue code eliminated. lib/list_sort.c is a bit trickier, as merge sort is already close to optimal, and we don't want to introduce triumphs of theory over practicality like the Ford-Johnson merge-insertion sort. Patch #4, without changing the algorithm, chops 32% off the code size and removes the part[MAX_LIST_LENGTH+1] pointer array (and the corresponding upper limit on efficiently sortable input size). Patch #5 improves the algorithm. The previous code is already optimal for power-of-two (or slightly smaller) size inputs, but when the input size is just over a power of 2, there's a very unbalanced final merge. There are, in the literature, several algorithms which solve this, but they all depend on the "breadth-first" merge order which was replaced by commit 835cc0c8477f with a more cache-friendly "depth-first" order. Some hard thinking came up with a depth-first algorithm which defers merges as little as possible while avoiding bad merges. This saves 0.2*n compares, averaged over all sizes. The code size increase is minimal (64 bytes on x86-64, reducing the net savings to 26%), but the comments expanded significantly to document the clever algorithm. TESTING NOTES: I have some ugly user-space benchmarking code which I used for testing before moving this code into the kernel. Shout if you want a copy. I'm running this code right now, with CONFIG_TEST_SORT and CONFIG_TEST_LIST_SORT, but I confess I haven't rebooted since the last round of minor edits to quell checkpatch. I figure there will be at least one round of comments and final testing. This patch (of 5): Rather than having special-case swap functions for 4- and 8-byte objects, special-case aligned multiples of 4 or 8 bytes. This speeds up most users of sort() by avoiding fallback to the byte copy loop. Despite what ca96ab859ab4 ("lib/sort: Add 64 bit swap function") claims, very few users of sort() sort pointers (or pointer-sized objects); most sort structures containing at least two words. (E.g. drivers/acpi/fan.c:acpi_fan_get_fps() sorts an array of 40-byte struct acpi_fan_fps.) The functions also got renamed to reflect the fact that they support multiple words. In the great tradition of bikeshedding, the names were by far the most contentious issue during review of this patch series. x86-64 code size 872 -> 886 bytes (+14) With feedback from Andy Shevchenko, Rasmus Villemoes and Geert Uytterhoeven. Link: http://lkml.kernel.org/r/f24f932df3a7fa1973c1084154f1cea596bcf341.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Dave Chinner <dchinner@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:49 +00:00
else if (is_aligned(base, size, 4))
swap_func = SWAP_WORDS_32;
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
else
swap_func = SWAP_BYTES;
lib/sort: Add 64 bit swap function In case the call side is not providing a swap function, we either use a 32 bit or a generic swap function. When swapping around pointers on 64 bit architectures falling back to use the generic swap function seems like an unnecessary waste. There at least 9 users ('sort' is of difficult to grep for) of sort() and all of them use the sort function without a customized swap function. Furthermore, they are all using pointers to swap around: arch/x86/kernel/e820.c:sanitize_e820_map() arch/x86/mm/extable.c:sort_extable() drivers/acpi/fan.c:acpi_fan_get_fps() fs/btrfs/super.c:btrfs_descending_sort_devices() fs/xfs/libxfs/xfs_dir2_block.c:xfs_dir2_sf_to_block() kernel/range.c:clean_sort_range() mm/memcontrol.c:__mem_cgroup_usage_register_event() sound/pci/hda/hda_auto_parser.c:snd_hda_parse_pin_defcfg() sound/pci/hda/hda_auto_parser.c:sort_pins_by_sequence() Obviously, we could improve the swap for other sizes as well but this is overkill at this point. A simple test shows sorting a 400 element array (try to stay in one page) with either with u32_swap() or u64_swap() show that the theory actually works. This test was done on a x86_64 (Intel Xeon E5-4610) machine. - swap_32: NumSamples = 100; Min = 48.00; Max = 49.00 Mean = 48.320000; Variance = 0.217600; SD = 0.466476; Median 48.000000 each * represents a count of 1 48.0000 - 48.1000 [ 68]: ******************************************************************** 48.1000 - 48.2000 [ 0]: 48.2000 - 48.3000 [ 0]: 48.3000 - 48.4000 [ 0]: 48.4000 - 48.5000 [ 0]: 48.5000 - 48.6000 [ 0]: 48.6000 - 48.7000 [ 0]: 48.7000 - 48.8000 [ 0]: 48.8000 - 48.9000 [ 0]: 48.9000 - 49.0000 [ 32]: ******************************** - swap_64: NumSamples = 100; Min = 44.00; Max = 63.00 Mean = 48.250000; Variance = 18.687500; SD = 4.322904; Median 47.000000 each * represents a count of 1 44.0000 - 45.9000 [ 15]: *************** 45.9000 - 47.8000 [ 37]: ************************************* 47.8000 - 49.7000 [ 39]: *************************************** 49.7000 - 51.6000 [ 0]: 51.6000 - 53.5000 [ 0]: 53.5000 - 55.4000 [ 0]: 55.4000 - 57.3000 [ 0]: 57.3000 - 59.2000 [ 1]: * 59.2000 - 61.1000 [ 3]: *** 61.1000 - 63.0000 [ 5]: ***** - swap_72: NumSamples = 100; Min = 53.00; Max = 71.00 Mean = 55.070000; Variance = 21.565100; SD = 4.643824; Median 53.000000 each * represents a count of 1 53.0000 - 54.8000 [ 73]: ************************************************************************* 54.8000 - 56.6000 [ 9]: ********* 56.6000 - 58.4000 [ 9]: ********* 58.4000 - 60.2000 [ 0]: 60.2000 - 62.0000 [ 0]: 62.0000 - 63.8000 [ 0]: 63.8000 - 65.6000 [ 0]: 65.6000 - 67.4000 [ 1]: * 67.4000 - 69.2000 [ 4]: **** 69.2000 - 71.0000 [ 4]: **** - test program: static int cmp_32(const void *a, const void *b) { u32 l = *(u32 *)a; u32 r = *(u32 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_64(const void *a, const void *b) { u64 l = *(u64 *)a; u64 r = *(u64 *)b; if (l < r) return -1; if (l > r) return 1; return 0; } static int cmp_72(const void *a, const void *b) { u32 l = get_unaligned((u32 *) a); u32 r = get_unaligned((u32 *) b); if (l < r) return -1; if (l > r) return 1; return 0; } static void init_array32(void *array) { u32 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array64(void *array) { u64 *a = array; int i; a[0] = 3821; for (i = 1; i < ARRAY_ELEMENTS; i++) a[i] = next_pseudo_random32(a[i-1]); } static void init_array72(void *array) { char *p; u32 v; int i; v = 3821; for (i = 0; i < ARRAY_ELEMENTS; i++) { p = (char *)array + (i * 9); put_unaligned(v, (u32*) p); v = next_pseudo_random32(v); } } static void sort_test(void (*init)(void *array), int (*cmp) (const void *, const void *), void *array, size_t size) { ktime_t start, stop; int i; for (i = 0; i < 10000; i++) { init(array); local_irq_disable(); start = ktime_get(); sort(array, ARRAY_ELEMENTS, size, cmp, NULL); stop = ktime_get(); local_irq_enable(); if (i > 10000 - 101) pr_info("%lld\n", ktime_to_us(ktime_sub(stop, start))); } } static void *create_array(size_t size) { void *array; array = kmalloc(ARRAY_ELEMENTS * size, GFP_KERNEL); if (!array) return NULL; return array; } static int perform_test(size_t size) { void *array; array = create_array(size); if (!array) return -ENOMEM; pr_info("test element size %d bytes\n", (int)size); switch (size) { case 4: sort_test(init_array32, cmp_32, array, size); break; case 8: sort_test(init_array64, cmp_64, array, size); break; case 9: sort_test(init_array72, cmp_72, array, size); break; } kfree(array); return 0; } static int __init sort_tests_init(void) { int err; err = perform_test(sizeof(u32)); if (err) return err; err = perform_test(sizeof(u64)); if (err) return err; err = perform_test(sizeof(u64)+1); if (err) return err; return 0; } static void __exit sort_tests_exit(void) { } module_init(sort_tests_init); module_exit(sort_tests_exit); MODULE_LICENSE("GPL v2"); MODULE_AUTHOR("Daniel Wagner"); MODULE_DESCRIPTION("sort perfomance tests"); Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-06-25 22:02:14 +00:00
}
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
/*
* Loop invariants:
* 1. elements [a,n) satisfy the heap property (compare greater than
* all of their children),
* 2. elements [n,num*size) are sorted, and
* 3. a <= b <= c <= d <= n (whenever they are valid).
*/
for (;;) {
size_t b, c, d;
if (a) /* Building heap: sift down --a */
a -= size;
else if (n -= size) /* Sorting: Extract root to --n */
do_swap(base, base + n, size, swap_func);
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
else /* Sort complete */
break;
/*
* Sift element at "a" down into heap. This is the
* "bottom-up" variant, which significantly reduces
* calls to cmp_func(): we find the sift-down path all
* the way to the leaves (one compare per level), then
* backtrack to find where to insert the target element.
*
* Because elements tend to sift down close to the leaves,
* this uses fewer compares than doing two per level
* on the way down. (A bit more than half as many on
* average, 3/4 worst-case.)
*/
for (b = a; c = 2*b + size, (d = c + size) < n;)
b = do_cmp(base + c, base + d, cmp_func, priv) >= 0 ? c : d;
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
if (d == n) /* Special case last leaf with no sibling */
b = c;
/* Now backtrack from "b" to the correct location for "a" */
while (b != a && do_cmp(base + a, base + b, cmp_func, priv) >= 0)
lib/sort: use more efficient bottom-up heapsort variant This uses fewer comparisons than the previous code (approaching half as many for large random inputs), but produces identical results; it actually performs the exact same series of swap operations. Specifically, it reduces the average number of compares from 2*n*log2(n) - 3*n + o(n) to n*log2(n) + 0.37*n + o(n). This is still 1.63*n worse than glibc qsort() which manages n*log2(n) - 1.26*n, but at least the leading coefficient is correct. Standard heapsort, when sifting down, performs two comparisons per level: one to find the greater child, and a second to see if the current node should be exchanged with that child. Bottom-up heapsort observes that it's better to postpone the second comparison and search for the leaf where -infinity would be sent to, then search back *up* for the current node's destination. Since sifting down usually proceeds to the leaf level (that's where half the nodes are), this does O(1) second comparisons rather than log2(n). That saves a lot of (expensive since Spectre) indirect function calls. The one time it's worse than the previous code is if there are large numbers of duplicate keys, when the top-down algorithm is O(n) and bottom-up is O(n log n). For distinct keys, it's provably always better, doing 1.5*n*log2(n) + O(n) in the worst case. (The code is not significantly more complex. This patch also merges the heap-building and -extracting sift-down loops, resulting in a net code size savings.) x86-64 code size 885 -> 767 bytes (-118) (I see the checkpatch complaint about "else if (n -= size)". The alternative is significantly uglier.) Link: http://lkml.kernel.org/r/2de8348635a1a421a72620677898c7fd5bd4b19d.1552704200.git.lkml@sdf.org Signed-off-by: George Spelvin <lkml@sdf.org> Acked-by: Andrey Abramov <st5pub@yandex.ru> Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Daniel Wagner <daniel.wagner@siemens.com> Cc: Dave Chinner <dchinner@redhat.com> Cc: Don Mullis <don.mullis@gmail.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:42:52 +00:00
b = parent(b, lsbit, size);
c = b; /* Where "a" belongs */
while (b != a) { /* Shift it into place */
b = parent(b, lsbit, size);
do_swap(base + b, base + c, size, swap_func);
}
}
}
EXPORT_SYMBOL(sort_r);
void sort(void *base, size_t num, size_t size,
cmp_func_t cmp_func,
swap_func_t swap_func)
{
return sort_r(base, num, size, _CMP_WRAPPER, swap_func, cmp_func);
}
EXPORT_SYMBOL(sort);