0450e203db
regressions/ck_ring: reduce buffer size for CI.
...
CI machines are struggling with current buffer sizes (hits
timeout).
6 years ago
dbfe282866
regressions/ck_epoch: fix other record read-reclaim races.
6 years ago
1cb2e6b281
regressions/ck_epoch_section_2: improve logging of failure conditions.
...
Certain environments don't provide us core dumps to work with,
unfortunately.
6 years ago
474dfc6eb1
regressions: fix ck_pr make clean and .gitignore
6 years ago
27c5f6f05f
travis: run regression test (limited due to hardware available) + test with ppc64le ( #117 )
...
build: add linux-ppc64le target.
There appears to be a regression on the target localized to epoch section optimization. I will need to investigate further.
6 years ago
1c2469358e
Implement ck_pr_dec_is_zero family of functions ( #115 )
...
* Implement ck_pr_dec_is_zero family of functions
* include/ck_pr.h: add ck_pr_{dec,inc}_is_zero and implement ck_pr_{dec,inc}_zero in terms of the new functions. Convert the architecture-specific implementations of ck_pr_foo_zero for x86 and x86-64 to ck_pr_foo_is_zero.
* regressions/ck_pr/validate: add smoke tests for ck_pr_dec_{,is_}zero and ck_pr_inc_{,is_}zero
* doc: document ck_pr_inc_is_zero
7 years ago
b19ed4c6a5
regressions/ck_pr: Cast -1 to intptr_t before casting it to void *.
7 years ago
db5db446d2
regression/ck_pr: Add tests for ck_pr_load_ptr() and ck_pr_store_ptr().
7 years ago
0bd946d6a0
regressions/ck_cc: Don't forget to add a Makefile.
7 years ago
afe01108d1
ck_cc: add a disable builtin flag for the FreeBSD kernel.
...
This primarily affects the FreeBSD kernel, where the popcount builtin
can be problematic (relies on compiler-provided libraries). See the
history of __POPCNT__ for details [1].
- A new flag, CK_MD_CC_BUILTIN_DISABLE, can be set to indicate that CK
should not rely on compiler builtins when possible.
- ck_cc_clz has been removed, it was unused.
- ck_internal_bsf has been removed, it was duplicate of ck_cc_ffs but broken,
replaced in favor of ck_cc_ffs. Previous consumers were using the bsf
instruction, eitherway.
- ck_{rhs,hs,ht} have been updated to use ck_cc_ffs*.
If FreeBSD requires the builtins for performance reasons, we will lift the
appropriate detection into ck_md (at least, bt* bs* family of functions don't
have the same problems on most targets unlike popcount).
1: https://lists.freebsd.org/pipermail/svn-src-head/2015-March/069663.html
7 years ago
f18ec01c16
regressions/ck_cc: basic coverage for ck_cc.
7 years ago
a8b9f7f411
regressions: add ck_pr_fence for basic validation of fence definitions.
7 years ago
3b86f1a4fd
[whitespace] regressions: Fix repeated typo in license header.
7 years ago
0d1e86d18e
ck_hs: add ck_hs_next_spmc
...
Memoize the map into ck_hs_iterator_t to make iteration more safe in the face of growth or shrinkage of the map. Tests for same.
Work from Riley Berton.
7 years ago
362f58e6b6
regressions: update ck_epoch usage.
8 years ago
db4554c686
regressions: update ck_epoch usage.
8 years ago
ffe9cbf9ca
ck_epoch: ck_epoch_end returns true if forward progress has been made.
8 years ago
8391338299
regressions/ck_epoch: fix up message.
8 years ago
cce21287fe
ck_epoch_call: basic coverage for call_strict.
8 years ago
8580ccc580
ck_epoch: introduce synchronize_wait that allows blocking synchronize operation.
8 years ago
8c12481577
regressions/ck_epoch: n_dispatch is now unsigned int.
8 years ago
77fc933229
regressions/epoch/torture: include synchronize latency.
8 years ago
24d26965d1
regressions/ck_epoch_section_2: Force exit after timer.
8 years ago
696ae6e995
Add benchmarks for ck_pr_add_64(), ck_pr_faa_64() and ck_pr_neg_64().
8 years ago
ab8f2b0344
Add AArch64 timer support to rdtsc().
...
Add support for RDTSC-like timer on AArch64. This is only used by the
internal regression/benchmark test suite.
8 years ago
1b461ca57a
regressions/ck_spinlock: Add ck_hclh to the check target.
8 years ago
2552550256
regressions/ck_hp: Make sure stores are visible before attempting CAS.
8 years ago
b6d82a481a
Ht: regressions: validate serial: check that our hash function was actually called.
...
Of course, assumes that there won't be an overflow caueing
it to be exactly 0 in the end...
The test fails.
9 years ago
03bd7a4158
regressions/ck_epoch: Change formatting.
9 years ago
f97d3da5c3
regressions/ck_sequence: Use flag rather than counter value for first set.
...
Prevents resetting exit barrier on overflow. Otherwise, hangs occur
on higher-performing systems.
9 years ago
ece3a24a4f
regressions/ck_epoch: Switch to heap allocated records for torture test.
9 years ago
5b64342668
regressions/ck_epoch: Remove counter dependencies from torture parity test.
9 years ago
f22f921576
regressions/ck_epoch_section_2: Switch to heap allocated records.
9 years ago
5999d8a243
regressions/ck_epoch: Shorten stack walk length for ck_epoch_synchronize.
...
This test was taking excessively long (read: hours) on some targets.
9 years ago
3a3562b96c
regressions/epoch: Add new torture test.
...
Verified with fault injection into ck_epoch_synchronize.
9 years ago
275908d246
regressions: Add ck_test convenience macro.
9 years ago
c845efcb0a
regressions/epoch: Various clean-up including fixes for Windows.
9 years ago
968b56598b
regressions/epoch: Additional coverage for epoch sections.
9 years ago
2e75aefc4a
ck_epoch: Allow for forward progress in concurrent epoch sections.
...
This work is derived directly from the work of John Esmet and Paul
Khuong ({jesmet,pkhuong}@appnexus.com) and PR34.
9 years ago
0d6d384f3f
ck_epoch: Cache associated epoch state machine in record.
...
This ends up removing arguments (and vector for error) for several
functions on the common path of ck_epoch.
9 years ago
ea9658d8d8
regressions: Switch to volatile load on ready flag.
9 years ago
cc600e7d4f
ck_ring: Add MPMC and MPSC variants.
9 years ago
7d1e3a6723
regressions/ck_cohort: make it pass with CORES=1
9 years ago
dfd8b2fb96
Merge branch 'master' of github.com:concurrencykit/ck
9 years ago
da4b956499
regressions/ck_ht: Remove dependency on ck_pr_add_64.
9 years ago
576e65fc74
ck_ht: Fallback to use 32bits values when 64bits load/store isn't supported.
...
That way we can use ck_ht on i386.
9 years ago
360bee1013
regressions/ck_pr/fp: Remove redundant load.
10 years ago
93c3d452a2
regressions/ck_pr: Add fence_lock coverage to fast-path test.
10 years ago
b46703ec5c
regressions: Add stupid fast-path benchmark for fences.
...
This is far from comprehensive but needed for some upcoming
tests.
10 years ago
7ffe069973
regressions/spinlock: Do not rely on RMW for counter update.
...
This makes it more likely to trigger load-store/store-load re-ordering
across critical sections.
10 years ago
Samy Al Bahra
Kevin Bowling
Daniel Black
Theo Schlossnagle
Olivier Houchard
Olivier Houchard
Riley Berton
Alexey Kopytov
Roman Lebedev
Samy Al Bahra
Michael Haberler