• Mathieu Desnoyers's avatar
    Add cmpxchg_local to asm-generic for per cpu atomic operations · 068fbad2
    Mathieu Desnoyers authored
    
    
    Emulates the cmpxchg_local by disabling interrupts around variable modification.
    This is not reentrant wrt NMIs and MCEs. It is only protected against normal
    interrupts, but this is enough for architectures without such interrupt sources
    or if used in a context where the data is not shared with such handlers.
    
    It can be used as a fallback for architectures lacking a real cmpxchg
    instruction.
    
    For architectures that have a real cmpxchg but does not have NMIs or MCE,
    testing which of the generic vs architecture specific cmpxchg is the fastest
    should be done.
    
    asm-generic/cmpxchg.h defines a cmpxchg that uses cmpxchg_local. It is meant to
    be used as a cmpxchg fallback for architectures that do not support SMP.
    
    * Patch series comments
    
    Using cmpxchg_local shows a performance improvements of the fast path goes from
    a 66% speedup on a Pentium 4 to a 14% speedup on AMD64.
    
    In detail:
    Tested-by: default avatarMathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
    Measurements on a Pentium4, 3GHz, Hyperthread.
    SLUB Performance testing
    ========================
    1. Kmalloc: Repeatedly allocate then free test
    
    * slub HEAD, test 1
    kmalloc(8) = 201 cycles         kfree = 351 cycles
    kmalloc(16) = 198 cycles        kfree = 359 cycles
    kmalloc(32) = 200 cycles        kfree = 381 cycles
    kmalloc(64) = 224 cycles        kfree = 394 cycles
    kmalloc(128) = 285 cycles       kfree = 424 cycles
    kmalloc(256) = 411 cycles       kfree = 546 cycles
    kmalloc(512) = 480 cycles       kfree = 619 cycles
    kmalloc(1024) = 623 cycles      kfree = 750 cycles
    kmalloc(2048) = 686 cycles      kfree = 811 cycles
    kmalloc(4096) = 482 cycles      kfree = 538 cycles
    kmalloc(8192) = 680 cycles      kfree = 734 cycles
    kmalloc(16384) = 713 cycles     kfree = 843 cycles
    
    * Slub HEAD, test 2
    kmalloc(8) = 190 cycles         kfree = 351 cycles
    kmalloc(16) = 195 cycles        kfree = 360 cycles
    kmalloc(32) = 201 cycles        kfree = 370 cycles
    kmalloc(64) = 245 cycles        kfree = 389 cycles
    kmalloc(128) = 283 cycles       kfree = 413 cycles
    kmalloc(256) = 409 cycles       kfree = 547 cycles
    kmalloc(512) = 476 cycles       kfree = 616 cycles
    kmalloc(1024) = 628 cycles      kfree = 753 cycles
    kmalloc(2048) = 684 cycles      kfree = 811 cycles
    kmalloc(4096) = 480 cycles      kfree = 539 cycles
    kmalloc(8192) = 661 cycles      kfree = 746 cycles
    kmalloc(16384) = 741 cycles     kfree = 856 cycles
    
    * cmpxchg_local Slub test
    kmalloc(8) = 83 cycles          kfree = 363 cycles
    kmalloc(16) = 85 cycles         kfree = 372 cycles
    kmalloc(32) = 92 cycles         kfree = 377 cycles
    kmalloc(64) = 115 cycles        kfree = 397 cycles
    kmalloc(128) = 179 cycles       kfree = 438 cycles
    kmalloc(256) = 314 cycles       kfree = 564 cycles
    kmalloc(512) = 398 cycles       kfree = 615 cycles
    kmalloc(1024) = 573 cycles      kfree = 745 cycles
    kmalloc(2048) = 629 cycles      kfree = 816 cycles
    kmalloc(4096) = 473 cycles      kfree = 548 cycles
    kmalloc(8192) = 659 cycles      kfree = 745 cycles
    kmalloc(16384) = 724 cycles     kfree = 843 cycles
    
    2. Kmalloc: alloc/free test
    
    * slub HEAD, test 1
    kmalloc(8)/kfree = 322 cycles
    kmalloc(16)/kfree = 318 cycles
    kmalloc(32)/kfree = 318 cycles
    kmalloc(64)/kfree = 325 cycles
    kmalloc(128)/kfree = 318 cycles
    kmalloc(256)/kfree = 328 cycles
    kmalloc(512)/kfree = 328 cycles
    kmalloc(1024)/kfree = 328 cycles
    kmalloc(2048)/kfree = 328 cycles
    kmalloc(4096)/kfree = 678 cycles
    kmalloc(8192)/kfree = 1013 cycles
    kmalloc(16384)/kfree = 1157 cycles
    
    * Slub HEAD, test 2
    kmalloc(8)/kfree = 323 cycles
    kmalloc(16)/kfree = 318 cycles
    kmalloc(32)/kfree = 318 cycles
    kmalloc(64)/kfree = 318 cycles
    kmalloc(128)/kfree = 318 cycles
    kmalloc(256)/kfree = 328 cycles
    kmalloc(512)/kfree = 328 cycles
    kmalloc(1024)/kfree = 328 cycles
    kmalloc(2048)/kfree = 328 cycles
    kmalloc(4096)/kfree = 648 cycles
    kmalloc(8192)/kfree = 1009 cycles
    kmalloc(16384)/kfree = 1105 cycles
    
    * cmpxchg_local Slub test
    kmalloc(8)/kfree = 112 cycles
    kmalloc(16)/kfree = 103 cycles
    kmalloc(32)/kfree = 103 cycles
    kmalloc(64)/kfree = 103 cycles
    kmalloc(128)/kfree = 112 cycles
    kmalloc(256)/kfree = 111 cycles
    kmalloc(512)/kfree = 111 cycles
    kmalloc(1024)/kfree = 111 cycles
    kmalloc(2048)/kfree = 121 cycles
    kmalloc(4096)/kfree = 650 cycles
    kmalloc(8192)/kfree = 1042 cycles
    kmalloc(16384)/kfree = 1149 cycles
    Tested-by: default avatarMathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
    Measurements on a AMD64 2.0 GHz dual-core
    
    In this test, we seem to remove 10 cycles from the kmalloc fast path.
    On small allocations, it gives a 14% performance increase. kfree fast
    path also seems to have a 10 cycles improvement.
    
    1. Kmalloc: Repeatedly allocate then free test
    
    * cmpxchg_local slub
    kmalloc(8) = 63 cycles      kfree = 126 cycles
    kmalloc(16) = 66 cycles     kfree = 129 cycles
    kmalloc(32) = 76 cycles     kfree = 138 cycles
    kmalloc(64) = 100 cycles    kfree = 288 cycles
    kmalloc(128) = 128 cycles   kfree = 309 cycles
    kmalloc(256) = 170 cycles   kfree = 315 cycles
    kmalloc(512) = 221 cycles   kfree = 357 cycles
    kmalloc(1024) = 324 cycles  kfree = 393 cycles
    kmalloc(2048) = 354 cycles  kfree = 440 cycles
    kmalloc(4096) = 394 cycles  kfree = 330 cycles
    kmalloc(8192) = 523 cycles  kfree = 481 cycles
    kmalloc(16384) = 643 cycles kfree = 649 cycles
    
    * Base
    kmalloc(8) = 74 cycles      kfree = 113 cycles
    kmalloc(16) = 76 cycles     kfree = 116 cycles
    kmalloc(32) = 85 cycles     kfree = 133 cycles
    kmalloc(64) = 111 cycles    kfree = 279 cycles
    kmalloc(128) = 138 cycles   kfree = 294 cycles
    kmalloc(256) = 181 cycles   kfree = 304 cycles
    kmalloc(512) = 237 cycles   kfree = 327 cycles
    kmalloc(1024) = 340 cycles  kfree = 379 cycles
    kmalloc(2048) = 378 cycles  kfree = 433 cycles
    kmalloc(4096) = 399 cycles  kfree = 329 cycles
    kmalloc(8192) = 528 cycles  kfree = 624 cycles
    kmalloc(16384) = 651 cycles kfree = 737 cycles
    
    2. Kmalloc: alloc/free test
    
    * cmpxchg_local slub
    kmalloc(8)/kfree = 96 cycles
    kmalloc(16)/kfree = 97 cycles
    kmalloc(32)/kfree = 97 cycles
    kmalloc(64)/kfree = 97 cycles
    kmalloc(128)/kfree = 97 cycles
    kmalloc(256)/kfree = 105 cycles
    kmalloc(512)/kfree = 108 cycles
    kmalloc(1024)/kfree = 105 cycles
    kmalloc(2048)/kfree = 107 cycles
    kmalloc(4096)/kfree = 390 cycles
    kmalloc(8192)/kfree = 626 cycles
    kmalloc(16384)/kfree = 662 cycles
    
    * Base
    kmalloc(8)/kfree = 116 cycles
    kmalloc(16)/kfree = 116 cycles
    kmalloc(32)/kfree = 116 cycles
    kmalloc(64)/kfree = 116 cycles
    kmalloc(128)/kfree = 116 cycles
    kmalloc(256)/kfree = 126 cycles
    kmalloc(512)/kfree = 126 cycles
    kmalloc(1024)/kfree = 126 cycles
    kmalloc(2048)/kfree = 126 cycles
    kmalloc(4096)/kfree = 384 cycles
    kmalloc(8192)/kfree = 749 cycles
    kmalloc(16384)/kfree = 786 cycles
    Tested-by: default avatarChristoph Lameter <clameter@sgi.com>
    I can confirm Mathieus' measurement now:
    
    Athlon64:
    
    regular NUMA/discontig
    
    1. Kmalloc: Repeatedly allocate then free test
    10000 times kmalloc(8) -> 79 cycles kfree -> 92 cycles
    10000 times kmalloc(16) -> 79 cycles kfree -> 93 cycles
    10000 times kmalloc(32) -> 88 cycles kfree -> 95 cycles
    10000 times kmalloc(64) -> 124 cycles kfree -> 132 cycles
    10000 times kmalloc(128) -> 157 cycles kfree -> 247 cycles
    10000 times kmalloc(256) -> 200 cycles kfree -> 257 cycles
    10000 times kmalloc(512) -> 250 cycles kfree -> 277 cycles
    10000 times kmalloc(1024) -> 337 cycles kfree -> 314 cycles
    10000 times kmalloc(2048) -> 365 cycles kfree -> 330 cycles
    10000 times kmalloc(4096) -> 352 cycles kfree -> 240 cycles
    10000 times kmalloc(8192) -> 456 cycles kfree -> 340 cycles
    10000 times kmalloc(16384) -> 646 cycles kfree -> 471 cycles
    2. Kmalloc: alloc/free test
    10000 times kmalloc(8)/kfree -> 124 cycles
    10000 times kmalloc(16)/kfree -> 124 cycles
    10000 times kmalloc(32)/kfree -> 124 cycles
    10000 times kmalloc(64)/kfree -> 124 cycles
    10000 times kmalloc(128)/kfree -> 124 cycles
    10000 times kmalloc(256)/kfree -> 132 cycles
    10000 times kmalloc(512)/kfree -> 132 cycles
    10000 times kmalloc(1024)/kfree -> 132 cycles
    10000 times kmalloc(2048)/kfree -> 132 cycles
    10000 times kmalloc(4096)/kfree -> 319 cycles
    10000 times kmalloc(8192)/kfree -> 486 cycles
    10000 times kmalloc(16384)/kfree -> 539 cycles
    
    cmpxchg_local NUMA/discontig
    
    1. Kmalloc: Repeatedly allocate then free test
    10000 times kmalloc(8) -> 55 cycles kfree -> 90 cycles
    10000 times kmalloc(16) -> 55 cycles kfree -> 92 cycles
    10000 times kmalloc(32) -> 70 cycles kfree -> 91 cycles
    10000 times kmalloc(64) -> 100 cycles kfree -> 141 cycles
    10000 times kmalloc(128) -> 128 cycles kfree -> 233 cycles
    10000 times kmalloc(256) -> 172 cycles kfree -> 251 cycles
    10000 times kmalloc(512) -> 225 cycles kfree -> 275 cycles
    10000 times kmalloc(1024) -> 325 cycles kfree -> 311 cycles
    10000 times kmalloc(2048) -> 346 cycles kfree -> 330 cycles
    10000 times kmalloc(4096) -> 351 cycles kfree -> 238 cycles
    10000 times kmalloc(8192) -> 450 cycles kfree -> 342 cycles
    10000 times kmalloc(16384) -> 630 cycles kfree -> 546 cycles
    2. Kmalloc: alloc/free test
    10000 times kmalloc(8)/kfree -> 81 cycles
    10000 times kmalloc(16)/kfree -> 81 cycles
    10000 times kmalloc(32)/kfree -> 81 cycles
    10000 times kmalloc(64)/kfree -> 81 cycles
    10000 times kmalloc(128)/kfree -> 81 cycles
    10000 times kmalloc(256)/kfree -> 91 cycles
    10000 times kmalloc(512)/kfree -> 90 cycles
    10000 times kmalloc(1024)/kfree -> 91 cycles
    10000 times kmalloc(2048)/kfree -> 90 cycles
    10000 times kmalloc(4096)/kfree -> 318 cycles
    10000 times kmalloc(8192)/kfree -> 483 cycles
    10000 times kmalloc(16384)/kfree -> 536 cycles
    
    Changelog:
    - Ran though checkpatch.
    Signed-off-by: default avatarMathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
    Cc: <linux-arch@vger.kernel.org>
    Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
    Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
    068fbad2