project-base/server-skynet-source-3rd-jemalloc
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2,701 Commits 2 Branches 0 Tags 13 MiB
497836dbc8
Commit Graph

1158 Commits

Author SHA1 Message Date
David Goldblatt
909c501b07 Cache_bin: Shouldn't know about tcache. 
Instead, have it take the cache_bin_info_ts to use by pointer.  While we're
here, add a src file for the cache bin.
 2020-03-12 11:54:19 -07:00
David Goldblatt
79f1ee2fc0 Move junking out of arena/tcache code. 
This is debug only and we keep it off the fast path.  Moving it here simplifies
the internal logic.

This never tries to junk on regions that were shrunk via xallocx.  I think this
is fine for two reasons:
- The shrunk-with-xallocx case is rare.
- We don't always do that anyway before this diff (it depends on the opt
  settings and extent hooks in effect).
 2020-03-12 11:54:19 -07:00
David T. Goldblatt
6c3491ad31 Tcache: Unify bin flush logic. 
The small and large pathways share most of their logic, even if some of the
individual operations are different.  We pull out the common logic into a
force-inlined function, and then specialize twice, once for each value of
"small".
 2020-02-25 10:21:03 -08:00
David T. Goldblatt
9f4fc27389 Ehooks: Fix a build warning. 
We wrote `return some_void_func()` in a function returning void, which is
confusing and triggers warnings on MSVC.
 2020-02-25 10:21:03 -08:00
David T. Goldblatt
162c2bcf31 Background thread: take base as a parameter. 2020-02-18 11:22:09 -08:00
David T. Goldblatt
29436fa056 Break prof and tcache knowledge of b0. 2020-02-18 11:22:09 -08:00
David T. Goldblatt
a0c1f4ac57 Rtree: take the base allocator as a parameter. 
This facilitates better testing by avoiding mixing of the "real" base with the
base used by the rtree under test.
 2020-02-18 11:22:09 -08:00
David T. Goldblatt
7013716aaa Emap: Take (and propagate) a zeroed parameter. 
Rtree needs this, and we should really treat them similarly.
 2020-02-18 11:22:09 -08:00
David T. Goldblatt
182192f83c Base: Pull into a single header. 2020-02-18 11:22:09 -08:00
David T. Goldblatt
34b7165fde Put szind_t, pszind_t in sz.h. 2020-02-18 11:22:09 -08:00
David Goldblatt
7e6c8a7286 Emap: Standardize naming. 
Namespace everything under emap_, always specify what it is we're looking up
(emap_lookup -> emap_edata_lookup), and use "ctx" over "info".
 2020-02-17 10:50:51 -08:00
David Goldblatt
ac50c1e44b Emap: Remove direct access to emap internals. 
In the process, we do a few local cleanups and optimizations.  In particular,
the size safety check on tcache flush no longer does a redundant load.
 2020-02-17 10:50:51 -08:00
David Goldblatt
06e42090f7 Make jemalloc.c use the emap interface. 
While we're here, we'll also clean up some style nits.
 2020-02-17 10:50:51 -08:00
David Goldblatt
f7d9c6c42d Emap: Move in alloc_ctx lookup functionality. 2020-02-17 10:50:51 -08:00
David Goldblatt
65a54d7714 Emap: Move in szind and slab modifications. 2020-02-17 10:50:51 -08:00
David Goldblatt
9b5d105fc3 Emap: Move in iealloc. 
This is logically scoped to the emap.
 2020-02-17 10:50:51 -08:00
David Goldblatt
1d449bd9a6 Emap: Internal rtree context setting. 
The only time sharing an rtree context saves across extent operations isn't a
no-op is when tsd is unavailable.  But this happens only in situations like
thread death or initialization, and we don't care about shaving off every
possible cycle in such scenarios.
 2020-02-17 10:50:51 -08:00
David Goldblatt
08eb1e6c31 Emap: Comments and cleanup 
Document some of the public interface, and hide the functions that are no longer
used outside of the emap module.
 2020-02-17 10:50:51 -08:00
David Goldblatt
231d1477e5 Rename emap_split_prepare_t -> emap_prepare_t. 
Both the split and merge functions use it.
 2020-02-17 10:50:51 -08:00
David Goldblatt
0586a56f39 Emap: Move in merge functionality. 2020-02-17 10:50:51 -08:00
David Goldblatt
040eac77cc Tell edatas their creation arena immediately. 
This avoids having to pass it in anywhere else.
 2020-02-17 10:50:51 -08:00
David Goldblatt
7c7b702064 Emap: Move over metadata splitting logic. 2020-02-17 10:50:51 -08:00
David Goldblatt
44f5f53605 Emap: Move over deregistration functions. 2020-02-17 10:50:51 -08:00
David Goldblatt
6513d9d923 Emap: Move over deregistration boundary functions. 2020-02-17 10:50:51 -08:00
David Goldblatt
9b5ca0b09d Emap: Move in slab interior registration. 2020-02-17 10:50:51 -08:00
David Goldblatt
d05b61db4a Emap: Move extent boundary registration in. 2020-02-17 10:50:51 -08:00
David Goldblatt
ca21ce4071 Emap: Move in write_acquired from extent. 2020-02-17 10:50:51 -08:00
David Goldblatt
01f255161c Add emap, for tracking extent locking. 2020-02-17 10:50:51 -08:00
Qi Wang
ba0e35411c Rework the bin locking around tcache refill / flush. 
Previously, tcache fill/flush (as well as small alloc/dalloc on the arena) may
potentially drop the bin lock for slab_alloc and slab_dalloc.  This commit
refactors the logic so that the slab calls happen in the same function / level
as the bin lock / unlock.  The main purpose is to be able to use flat combining
without having to keep track of stack state.

In the meantime, this change reduces the locking, especially for slab_dalloc
calls, where nothing happens after the call.
 2020-02-13 23:31:54 -08:00
Kamil Rytarowski
7fd22f7b2e Fix Undefined Behavior in hash.h 
hash.h:200:27, left shift of 250 by 24 places cannot be represented in type 'int'
 2020-02-13 12:25:26 -08:00
Yinan Zhang
9cac3fa8f5 Encapsulate buffer allocation in buffered writer 2020-02-04 13:21:58 -08:00
Yinan Zhang
bdc08b5158 Better naming buffered writer 2020-02-04 13:21:58 -08:00
Qi Wang
c6bfe55857 Update the tsd description. 2020-02-04 13:07:05 -08:00
Qi Wang
e896522616 Abbreviate thread-event to te. 2020-02-04 13:07:05 -08:00
Qi Wang
5e500523a0 Remove thread_event_boot(). 2020-02-04 00:18:15 -08:00
Qi Wang
97dd79db6c Implement deallocation events. 
Make the event module to accept two event types, and pass around the event
context.  Use bytes-based events to trigger tcache GC on deallocation, and get
rid of the tcache ticker.
 2020-02-04 00:18:15 -08:00
Qi Wang
974222c626 Add safety check on sdallocx slow / sampled path. 2020-01-31 00:04:22 -08:00
Qi Wang
88d9eca848 Enforce page alignment for sampled allocations. 
This allows sampled allocations to be checked through alignment, therefore
enable sized deallocation regardless of cache_oblivious.
 2020-01-31 00:04:22 -08:00
Qi Wang
0f552ed673 Don't purge huge extents when decay is off. 2020-01-30 14:40:38 -08:00
Qi Wang
38a48e5741 Set reentrancy to 1 for tsd_state_purgatory. 
Reentrancy is already set for other non-nominal tsd states (reincarnated and
minimal_initialized).  Add purgatory to be safe and consistent.
 2020-01-30 13:55:20 -08:00
Qi Wang
88b0e03a4e Implement opt.stats_interval and the _opts options. 
Add options stats_interval and stats_interval_opts to allow interval based stats
printing.  This provides an easy way to collect stats without code changes,
because opt.stats_print may not work (some binaries never exit).
 2020-01-29 09:57:55 -08:00
Qi Wang
d71a145ec1 Chagne prof_accum_t to counter_accum_t for general purpose. 2020-01-29 09:57:55 -08:00
Yinan Zhang
f81341a48b Fallback to unbuffered printing if OOM 2020-01-21 17:09:44 -08:00
David Goldblatt
bd3be8e0b1 Remove commit parameter to ecache functions. 
No caller ever wants uncommitted memory.
 2020-01-17 10:54:56 -08:00
Qi Wang
dab81bd315 Rework and fix the assertions on malloc fastpath. 
The first half of the malloc fastpath may execute before malloc_init.  Make the
assertions work in that case.
 2020-01-14 15:00:41 -08:00
Yinan Zhang
2b604a3016 Record request size in prof recent entries 2020-01-10 12:01:01 -08:00
Yinan Zhang
40a391408c Define constructor for buffered writer argument 2020-01-10 11:59:02 -08:00
Yinan Zhang
6d8e616902 Make buffered writer an independent module 2020-01-10 11:59:02 -08:00
Yinan Zhang
6b6b4709b3 Unify buffered writer naming 2020-01-09 14:31:31 -08:00
Yinan Zhang
9a60cf54ec Last-N profiling mode 2019-12-30 15:58:57 -08:00
Yinan Zhang
7a27a05940 Delete tdata states used for cleanup 2019-12-30 15:58:57 -08:00
Yinan Zhang
e98ddf7987 Fix unlikely condition in arena_prof_info_get() 2019-12-30 15:58:57 -08:00
Yinan Zhang
3fa142cf39 Remove _externs from prof internal header names 2019-12-23 11:14:15 -08:00
Yinan Zhang
112dc36dd5 Handle log_mtx during forking 2019-12-20 17:17:48 -08:00
Yinan Zhang
ea42174d07 Refactor profiling headers 2019-12-20 17:17:48 -08:00
David Goldblatt
6342da0970 Ehooks: Further optimize default merge case. 
This avoids the cost of an iealloc in cases where the user uses the default
merge hook without using the default extent hooks.
 2019-12-20 10:18:40 -08:00
David Goldblatt
e210ccc57e Move extent2 -> extent. 
Eventually, we may fully break off the extent module; but not for some time.  If
it's going to live on in a non-transitory state, it might as well have the nicer
name.
 2019-12-20 10:18:40 -08:00
David Goldblatt
2f4fa80414 Rename extents -> ecache. 2019-12-20 10:18:40 -08:00
David Goldblatt
56cc56b692 Break extent split dependence on arena. 2019-12-20 10:18:40 -08:00
David Goldblatt
0aa9769fb0 Break commit functions' arena dependence 2019-12-20 10:18:40 -08:00
David Goldblatt
576d7047ab Ecache: Should know its arena_ind. 
What we call an arena_ind is really the index associated with some particular
set of ehooks; the arena is just the user-visible portion of that.  Making this
explicit, and reframing checks in terms of that, makes the code simpler and
cleaner, and helps us avoid passing the arena itself all throughout extent code.

This lets us put back an arena-specific assert.
 2019-12-20 10:18:40 -08:00
David Goldblatt
372042a082 Remove merge dependence on the arena. 2019-12-20 10:18:40 -08:00
David Goldblatt
9cad5639ff Ehooks: remove arena_ind parameter. 
This lives within the ehooks_t now, so that callers don't need to know it.
 2019-12-20 10:18:40 -08:00
David Goldblatt
57fe99d4be Move relevant index into the ehooks_t itself. 
It's always passed into the ehooks; keeping it colocated lets us avoid passing
the arena everywhere.
 2019-12-20 10:18:40 -08:00
David Goldblatt
c792f3e4ab edata_cache: Remember the associated base_t. 
This will save us some trouble down the line when we stop passing arena pointers
everywhere; we won't have to pass around a base_t pointer either.
 2019-12-20 10:18:40 -08:00
David Goldblatt
ae23e5f426 Unify extent_alloc_wrapper with the other wrappers. 
Previously, it was really more like extents_alloc (it looks in an ecache for an
extent to reuse as its primary allocation pathway).  Make that pathway more
explciitly like extents_alloc, and rename extent_alloc_wrapper_hard accordingly.
 2019-12-20 10:18:40 -08:00
David Goldblatt
d8b0b66c6c Put extent_state_t into ecache as well as eset. 2019-12-20 10:18:40 -08:00
David Goldblatt
98eb40e563 Move delay_coalesce from the eset to the ecache. 2019-12-20 10:18:40 -08:00
David Goldblatt
bb70df8e5b Extent refactor: Introduce ecache module. 
This will eventually completely wrap the eset, and handle concurrency,
allocation, and deallocation.  For now, we only pull out the mutex from the
eset.
 2019-12-20 10:18:40 -08:00
David Goldblatt
0704516245 Ehooks: Add head tracking. 2019-12-20 10:18:40 -08:00
David Goldblatt
09475bf8ac extent_may_dalloc -> ehooks_dalloc_will_fail 2019-12-20 10:18:40 -08:00
David Goldblatt
7859184179 Pull out edata_t caching into its own module. 2019-12-20 10:18:40 -08:00
David Goldblatt
a7862df616 Rename extent_t to edata_t. 
This frees us up from the unfortunate extent/extent2 naming collision.
 2019-12-20 10:18:40 -08:00
David Goldblatt
865debda22 Rename extent.h -> edata.h. 
This name is slightly pithier; a full-on rename will come shortly.
 2019-12-20 10:18:40 -08:00
David Goldblatt
a738a66b5c Ehooks: Add some debug zero and addr checks. 
These help make sure that the ehooks return properly zeroed memory when required
to.
 2019-12-20 10:18:40 -08:00
David Goldblatt
4b2e5ee8b9 Ehooks: Add a "zero" ehook. 
This is the first API expansion.  It lets the hooks pick where and how to purge
within themselves.
 2019-12-20 10:18:40 -08:00
David Goldblatt
403f2d1664 Extents: Split out introspection functionality. 
This isn't really part of the core extent allocation facilities.  Especially as
this module grows, having it in its own place may come in handy.
 2019-12-20 10:18:40 -08:00
David Goldblatt
92a511d385 Make extent module hermetic. 
In the form of extent2.h.  The naming leaves something to be desired, but I'll
leave that for a later diff.
 2019-12-20 10:18:40 -08:00
David Goldblatt
39fdc690a0 Ehooks comments and cleanup. 2019-12-20 10:18:40 -08:00
David Goldblatt
c8dae890c8 Extent -> Ehooks: Move over default hooks. 2019-12-20 10:18:40 -08:00
David Goldblatt
2fe5108263 Extent -> Ehooks: Move merge hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
1fff4d2ee3 Extent -> Ehooks: Move split hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
a5b42a1a10 Extent -> Ehooks: Move purge_forced hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
368baa42ef Extent -> Ehooks: Move purge_lazy hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
d78fe241ac Extent -> Ehooks: Move commit and decommit hooks. 2019-12-20 10:18:40 -08:00
David Goldblatt
5459ec9dae Extent -> Ehooks: Move destroy hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
bac8e2e5a6 Extent -> Ehooks: Move dalloc hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
dc8b4e6e13 Extent -> Ehooks: Move alloc hook. 2019-12-20 10:18:40 -08:00
David Goldblatt
703fbc0ff5 Introduce unsafe reentrancy guards. 
We have to work to circumvent the safety checks in pre_reentrancy when going
down extent hook pathways.  Instead, let's explicitly have checked and unchecked
guards.
 2019-12-20 10:18:40 -08:00
David Goldblatt
ae0d8e8591 Move extent ehook calls into ehooks 2019-12-20 10:18:40 -08:00
David Goldblatt
ba8b9ecbcb Add ehooks module 2019-12-20 10:18:40 -08:00
David Goldblatt
837119a948 base_structs.h: Remove some mid-line tabs. 2019-12-20 10:18:40 -08:00
David Goldblatt
9f6eb09585 Extents: Eagerly initialize extent hooks. 
When deferred initialization was added, initializing required copying
sizeof(extent_hooks_t) bytes after a pointer chase. Today, it's just a single
pointer loaded from the base_t. In subsequent diffs, we'll get rid of even that.
 2019-12-20 10:18:40 -08:00
David Goldblatt
4278f84603 Move extent hook getters/setters to arena.c 
This is where they're logically scoped; they access arena data.
 2019-12-20 10:18:40 -08:00
Qi Wang
d5031ea824 Allow dallocx and sdallocx after tsd destruction. 
After a thread turns into purgatory / reincarnated state, still allow dallocx
and sdallocx to function normally.
 2019-12-19 11:17:03 -08:00
Yinan Zhang
4afd709d1f Restructure setters for profiling info 
Explicitly define three setters:

- `prof_tctx_reset()`: set `prof_tctx` to `1U`, if we don't know in
advance whether the allocation is large or not;
- `prof_tctx_reset_sampled()`: set `prof_tctx` to `1U`, if we already
know in advance that the allocation is large;
- `prof_info_set()`: set a real `prof_tctx`, and also set other
profiling info e.g. the allocation time.

Code structure wise, the prof level is kept as a thin wrapper, the
large level only provides low level setter APIs, and the arena level
carries out the main logic.
 2019-12-17 10:01:28 -08:00
Yinan Zhang
1d01e4c770 Initialization utilities for nstime 2019-12-16 16:08:56 -08:00
Qi Wang
dd649c9485 Optimize away the tsd_fast() check on fastpath. 
Fold the tsd_state check onto the event threshold check.  The fast threshold is
set to 0 when tsd switch to non-nominal.

The fast_threshold can be reset by remote threads, to refect the non nominal tsd
state change.
 2019-12-11 23:44:20 -08:00
Yinan Zhang
45836d7fd3 Pass nstime_t pointer for profiling 2019-12-11 11:38:16 -08:00
Yinan Zhang
7d2bac5a38 Refactor destroy code path for prof_tctx 2019-12-10 16:31:05 -08:00