Commit Graph

3336 Commits

Author SHA1 Message Date
barracuda156
70e3735f3a jemalloc: fix PowerPC definitions in quantum.h 2022-05-26 10:51:10 -07:00
Alex Lapenkou
5b1f2cc5d7 Implement pvalloc replacement
Despite being an obsolete function, pvalloc is still present in GLIBC and should
work correctly when jemalloc replaces libc allocator.
2022-05-18 17:01:09 -07:00
Qi Wang
cd5aaf308a Improve the failure message upon opt_experimental_infallible_new. 2022-05-17 16:07:40 -07:00
Yuriy Chernyshov
70d4102f48 Fix compiling edata.h with MSVC
At the time an attempt to compile jemalloc 5.3.0 with MSVC 2019 results in the followin error message:

> jemalloc/include/jemalloc/internal/edata.h:660: error C4576: a parenthesized type followed by an initializer list is a non-standard explicit type conversion syntax
2022-05-09 14:51:07 -07:00
Qi Wang
54eaed1d8b Merge branch 'dev' 2022-05-06 11:28:25 -07:00
Qi Wang
304c919829 Update ChangeLog for 5.3.0. 2022-05-06 11:24:21 -07:00
Qi Wang
8cb814629a Make the default option of zero realloc match the system allocator. 2022-05-05 17:11:18 -07:00
Qi Wang
66c889500a Make test/unit/background_thread_enable more conservative.
To avoid resource exhaustion on 32-bit platforms.
2022-05-04 15:32:57 -07:00
Qi Wang
a7d73dd4c9 Update TUNING.md to include the new tcache_max option. 2022-05-04 10:59:40 -07:00
Qi Wang
254b011915 Small doc tweak of opt.trust_madvise.
Avoid quoted enabled and disabled because it's a bool type instead of char *.
2022-04-28 21:16:25 -07:00
Qi Wang
f5e840bbf0 Minor typo fix in doc. 2022-04-27 20:25:29 -07:00
Qi Wang
ceca07d2ca Correct the name of stats.mutexes.prof_thds_data in doc. 2022-04-25 20:12:58 -07:00
Qi Wang
391bad4b95 Avoid abort() in test/integration/cpp/infallible_new_true.
Allow setting the safety check abort hook through mallctl, which avoids abort()
and core dumps.
2022-04-25 11:29:32 -07:00
cuishuang
9a242f16d9 fix some typos
Signed-off-by: cuishuang <imcusg@gmail.com>
2022-04-25 11:29:00 -07:00
Qi Wang
0e29ad4efa Rename zero_realloc option "strict" to "alloc".
With realloc(ptr, 0) being UB per C23, the option name "strict" makes less sense
now.  Rename to "alloc" which describes the behavior.
2022-04-20 10:27:25 -07:00
Qi Wang
5841b6dbe7 Update FreeBSD image to 12.3 for cirrus ci. 2022-04-19 15:29:30 -07:00
Qi Wang
ed5fc14b28 Use volatile to workaround buffer overflow false positives.
In test/integration/rallocx, full usable size is checked which may confuse
overflow detection.
2022-04-04 12:16:46 -07:00
Alex Lapenkou
25517b852e Reoreder TravisCI jobs to optimize CI time
Sorting jobs by descending expected runtime helps to utilize concurrency
better.
2022-03-29 11:58:27 -07:00
Alex Lapenkou
8a49b62e78 Enable TravisCI for Windows 2022-03-29 11:58:27 -07:00
Alex Lapenkou
fdb6c10162 Add FreeBSD to TravisCI
Implement the generation of Travis jobs for FreeBSD. The generated jobs
replicate the existing CirrusCI config.
2022-03-29 11:58:27 -07:00
Alex Lapenkou
a93931537e Do not disable SEC by default for 64k pages platforms
Default SEC max_alloc option value was 32k, disabling SEC for platforms with
lg-page=16. This change enables SEC for all platforms, making minimum max_alloc
value equal to PAGE.
2022-03-24 22:05:35 -07:00
Charles
eaaa368bab Add comments and use meaningful vars in sz_psz2ind. 2022-03-24 16:56:59 -07:00
Alex Lapenkou
5bf03f8ce5 Implement PAGE_FLOOR macro 2022-03-22 17:45:55 -07:00
Alex Lapenkou
52631c90f6 Fix size class calculation for sec
Due to a bug in sec initialization, the number of cached size classes
was equal to 198. The bug caused the creation of more than a hundred of
unused bins, although it didn't affect the caching logic.
2022-03-22 17:45:55 -07:00
Qi Wang
7ae0f15c59 Add a default page size when cross-compile for Apple M1.
When cross-compile for M1 and no page size specified, use the default 16K and
skip detecting the page size (which is likely incorrect).
2022-03-21 14:30:48 -07:00
Alex Lapenkov
eb65d1b078 Fix FreeBSD system jemalloc TSD cleanup
Before this commit, in case FreeBSD libc jemalloc was overridden by another
jemalloc, proper thread shutdown callback was involved only for the overriding
jemalloc. A call to _malloc_thread_cleanup from libthr would be redirected to
user jemalloc, leaving data about dead threads hanging in system jemalloc. This
change tackles the issue in two ways. First, for current and old system
jemallocs, which we can not modify, the overriding jemalloc would locate and
invoke system cleanup routine. For upcoming jemalloc integrations, the cleanup
registering function will also be redirected to user jemalloc, which means that
system jemalloc's cleanup routine will be registered in user's jemalloc and a
single call to _malloc_thread_cleanup will be sufficient to invoke both
callbacks.
2022-03-02 10:10:27 -08:00
Azat Khuzhin
78b58379c8 Fix possible "nmalloc >= ndalloc" assertion.
It is possible that ndalloc will be updated before nmalloc, in
arena_large_ralloc_stats_update(), fix this by reorder those calls.

It was found by ClickHouse CI, that periodically hits this assertion [1].

  [1]: https://github.com/ClickHouse/ClickHouse/issues/31531

That issue contains lots of examples, with core dump and some gdb output [2].

  [2]: https://s3.amazonaws.com/clickhouse-test-reports/34951/96390a9263cb5af3d6e42a84988239c9ae87ce32/stress_test__debug__actions_.html

Here you can find binaries for that particular report [3] you need
clickhouse debug build [4].

  [3]: https://s3.amazonaws.com/clickhouse-builds/34951/96390a9263cb5af3d6e42a84988239c9ae87ce32/clickhouse_build_check_(actions)/report.html
  [4]: https://s3.amazonaws.com/clickhouse-builds/34951/96390a9263cb5af3d6e42a84988239c9ae87ce32/package_debug/clickhouse

Brief info from that report:

    2 0x000000002ad6dbfe in arena_stats_merge (tsdn=0x7f2399abdd20, arena=0x7f241ce01080, nthreads=0x7f24e4360958, dss=0x7f24e4360960, dirty_decay_ms=0x7f24e4360968, muzzy_decay_ms=0x7f24e4360970, nactive=0x7f24e4360978, ndirty=0x7f24e43
    e4360988, astats=0x7f24e4360998, bstats=0x7f24e4363310, lstats=0x7f24e4364990, estats=0x7f24e4366e50, hpastats=0x7f24e43693a0, secstats=0x7f24e436a020) at ../contrib/jemalloc/src/arena.c:138
            ndalloc = 226
            nflush = 0
            curlextents = 0
            nmalloc = 225
            nrequests = 0

Here you can see that they differs only by 1.

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-03-01 12:28:28 -08:00
Alex Lapenkou
ca709c3139 Fix failed assertion due to racy memory access
While calculating the number of stashed pointers, multiple variables
potentially modified by a concurrent thread were used for the
calculation.  This led to some inconsistencies, correctly detected by
the assertions.  The change eliminates some possible inconsistencies by
using unmodified variables and only once a concurrently modified one.
The assertions are omitted for the cases where we acknowledge potential
inconsistencies too.
2022-02-17 09:35:52 -08:00
Qi Wang
063d134aeb Properly detect background thread support on Darwin.
When cross-compile, the host type / abi should be checked to determine
background thread compatibility.
2022-02-15 10:10:11 -08:00
Alex Lapenkou
a4e81221cc Document 'make uninstall'
Update INSTALL.md, reflecting the addition of 'uninstall' target.
2022-01-31 14:55:00 -08:00
Qi Wang
20f9802e4f Avoid overflow warnings in test/unit/safety_check. 2022-01-27 10:29:54 -08:00
Qi Wang
8c59c44ffa Add a dependency checking step at the end of malloc_conf_init.
Currently only prof_leak_error and prof_final are checked.
2022-01-26 17:17:48 -08:00
Qi Wang
efc539c040 Initialize prof_leak during prof init.
Otherwise, prof_leak may get set after prof_leak_error, and disagree with each
other.
2022-01-26 17:17:48 -08:00
Alex Lapenkou
002f0e9397 Disable TravisCI jobs generation for Windows
These jobs take about 20 minutes to complete. We don't want to enable
them until we switch to unlimited concurrency plan, otherwise the builds
will take way too long.
2022-01-26 10:16:57 -08:00
Alex Lapenkou
01a293fc08 Add Windows to TravisCI
Implement the generation of Travis jobs for Windows. Currently, the
generated jobs replicate Appveyor setup and complete successfully. There
is support for MinGW GCC and MSVC compilers as well as 64 and 32 bit
compilation. Linux and MacOS jobs behave identically, but some
environment variables change - CROSS_COMPILE_32BIT=yes is added for
builds with cross compilation, empty COMPILER_FLAGS are not set anymore.
2022-01-26 10:16:57 -08:00
yunxu
b798fabdf7 Add prof_leak_error option
The option makes the process to exit with error code 1 if a memory leak
is detected. This is useful for implementing automated tools that rely
on leak detection.
2022-01-21 16:24:20 -08:00
Alex Lapenkou
eafd2ac39f Forbid spaces in prefix and exec_prefix
Spaces in these are also not handled correctly by Make, so there's sense
in not allowing that.
2022-01-19 12:28:16 -08:00
Alex Lapenkou
36a09ba2c7 Forbid spaces in install suffix
To avoid potential issues with removing unintended files after 'make
uninstall', spaces are no longer allowed in install suffix. It's worth
mentioning, that with GNU Make on Linux spaces in install suffix didn't
work anyway, leading to errors in the Makefile. But being verbose
about this restriction makes it more transparent for the developers.
2022-01-19 12:28:16 -08:00
Shuduo Sang
640c3c72e6 Add support for 'make uninstall' 2022-01-19 12:28:16 -08:00
Alex Lapenkou
f15d8f3b41 Echo installed files via verbose 'install' command
It's not necessary to manually echo all install commands, similar effect
is achieved via 'install -v'
2022-01-19 12:28:16 -08:00
Charles
eb196815d6 Avoid calculating size of size class twice & delete sc_data_global. 2022-01-18 11:54:12 -08:00
Qi Wang
011449f17b Fix doc build with install-suffix. 2022-01-11 21:15:24 -08:00
Qi Wang
8b49eb132e Fix the HELP_STRING of --enable-doc. 2022-01-11 21:15:24 -08:00
Qi Wang
ddb170b1d9 Simplify arena_migrate() to take arena_t* instead of indices.
This makes debugging slightly easier and avoids the confusion of "should we
create new arenas" here.
2022-01-11 16:59:22 -08:00
Qi Wang
648b3b9f76 Lower the num_threads in the stress test of test/unit/prof_recent
This takes a fair amount of resources.  Under high concurrency it was causing
resource exhaustion such as pthread_create and mmap failures.
2022-01-11 16:58:56 -08:00
Qi Wang
d66162e032 Fix the extent state checking on the merge error path.
With DSS as primary, the default merge impl will (correctly) decline to merge
when one of the extent is non-dss.  The error path should tolerate the
not-merged extent being in a merging state.
2022-01-11 16:58:47 -08:00
Craig Leres
c9946fa7e6 FreeBSD also needs the OS-X "don't declare system functions as
nothrow" fix since it also has jemalloc in the base system
2022-01-11 11:53:25 -08:00
Jonathan Swinney
89fe8ee6bf Use the isb instruction instead of yield for spin locks on arm
isb introduces a small delay which is closer to the x86 pause instruction.
2022-01-10 15:29:56 -08:00
Qi Wang
6230cc88b6 Add background thread sleep retry in test/unit/hpa_background_thread
Under high concurrency / heavy test load (e.g. using run_tests.sh), the
background thread may not get scheduled for a longer period of time.  Retry 100
times max before bailing out.
2022-01-07 10:28:28 -08:00
Qi Wang
61978bbe69 Purge all if the last thread migrated away from an arena. 2022-01-06 19:02:26 -08:00