summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
| * Add rtree lookup path caching.Jason Evans2016-06-067-95/+268
| | | | | | | | | | | | | | | | | | rtree-based extent lookups remain more expensive than chunk-based run lookups, but with this optimization the fast path slowdown is ~3 CPU cycles per metadata lookup (on Intel Core i7-4980HQ), versus ~11 cycles prior. The path caching speedup tends to degrade gracefully unless allocated memory is spread far apart (as is the case when using a mixture of sbrk() and mmap()).
| * Make tsd cleanup functions optional, remove noop cleanup functions.Jason Evans2016-06-0611-81/+23
| |
| * Remove some unnecessary locking.Jason Evans2016-06-061-20/+2
| |
| * Reduce NSZS, since NSIZES (was nsizes) can not be so large.Jason Evans2016-06-061-1/+1
| |
| * Fix rallocx() sampling code to not eagerly commit sampler update.Jason Evans2016-06-061-3/+3
| | | | | | | | | | | | rallocx() for an alignment-constrained request may end up with a smaller-than-worst-case size if in-place reallocation succeeds due to serendipitous alignment. In such cases, sampling may not happen.
| * Add a missing prof_alloc_rollback() call.Jason Evans2016-06-061-0/+1
| | | | | | | | | | | | In the case where prof_alloc_prep() is called with an over-estimate of allocation size, and sampling doesn't end up being triggered, the tctx must be discarded.
| * Miscellaneous s/chunk/extent/ updates.Jason Evans2016-06-068-19/+17
| |
| * Relax NBINS constraint (max 255 --> max 256).Jason Evans2016-06-061-4/+2
| |
| * Relax opt_lg_chunk clamping constraints.Jason Evans2016-06-061-10/+2
| |
| * Remove obsolete stats.arenas.<i>.metadata.mapped mallctl.Jason Evans2016-06-068-73/+34
| | | | | | | | | | Rename stats.arenas.<i>.metadata.allocated mallctl to stats.arenas.<i>.metadata .
| * Better document --enable-ivsalloc.Jason Evans2016-06-063-6/+14
| |
| * Rename most remaining *chunk* APIs to *extent*.Jason Evans2016-06-0619-1159/+1151
| |
| * s/chunk_lookup/extent_lookup/g, s/chunks_rtree/extents_rtree/gJason Evans2016-06-068-44/+55
| |
| * s/CHUNK_HOOKS_INITIALIZER/EXTENT_HOOKS_INITIALIZER/gJason Evans2016-06-065-16/+16
| |
| * Rename chunks_{cached,retained,mtx} to extents_{cached,retained,mtx}.Jason Evans2016-06-064-34/+35
| |
| * Rename chunk_*_t hooks to extent_*_t.Jason Evans2016-06-063-126/+129
| |
| * s/chunk_hook/extent_hook/gJason Evans2016-06-0612-191/+200
| |
| * Rename huge to large.Jason Evans2016-06-0637-626/+587
| |
| * Update private symbols.Jason Evans2016-06-062-13/+21
| |
| * Move slabs out of chunks.Jason Evans2016-06-0621-2327/+591
| |
| * Improve interval-based profile dump triggering.Jason Evans2016-06-062-1/+15
| | | | | | | | | | | | | | | | | | | | | | When an allocation is large enough to trigger multiple dumps, use modular math rather than subtraction to reset the interval counter. Prior to this change, it was possible for a single allocation to cause many subsequent allocations to all trigger profile dumps. When updating usable size for a sampled object, try to cancel out the difference between LARGE_MINCLASS and usable size from the interval counter.
| * Use huge size class infrastructure for large size classes.Jason Evans2016-06-0634-1975/+459
| |
| * Implement cache-oblivious support for huge size classes.Jason Evans2016-06-0312-170/+298
| |
| * Allow chunks to not be naturally aligned.Jason Evans2016-06-0311-268/+105
| | | | | | | | | | Precisely size extents for huge size classes that aren't multiples of chunksize.
| * Remove CHUNK_ADDR2BASE() and CHUNK_ADDR2OFFSET().Jason Evans2016-06-036-183/+190
| |
| * Make extent_prof_tctx_[gs]et() atomic.Jason Evans2016-06-031-3/+7
| |
| * Add extent_dirty_[gs]et().Jason Evans2016-06-036-10/+34
| |
| * Convert rtree from per chunk to per page.Jason Evans2016-06-035-52/+94
| | | | | | | | Refactor [de]registration to maintain interior rtree entries for slabs.
| * Refactor chunk_purge_wrapper() to take extent argument.Jason Evans2016-06-034-12/+10
| |
| * Refactor chunk_[de]commit_wrapper() to take extent arguments.Jason Evans2016-06-033-16/+14
| |
| * Refactor chunk_dalloc_{cache,wrapper}() to take extent arguments.Jason Evans2016-06-0310-198/+147
| | | | | | | | | | | | Rename arena_extent_[d]alloc() to extent_[d]alloc(). Move all chunk [de]registration responsibility into chunk.c.
| * Add/use chunk_split_wrapper().Jason Evans2016-06-037-519/+563
| | | | | | | | | | | | Remove redundant ptr/oldsize args from huge_*(). Refactor huge/chunk/arena code boundaries.
| * Add/use chunk_merge_wrapper().Jason Evans2016-06-036-93/+101
| |
| * Add/use chunk_commit_wrapper().Jason Evans2016-06-034-30/+44
| |
| * Add/use chunk_decommit_wrapper().Jason Evans2016-06-034-7/+20
| |
| * Merge chunk_alloc_base() into its only caller.Jason Evans2016-06-034-23/+9
| |
| * Replace extent_tree_szad_* with extent_heap_*.Jason Evans2016-06-039-103/+332
| |
| * Use rtree rather than [sz]ad trees for chunk split/coalesce operations.Jason Evans2016-06-036-197/+233
| |
| * Dodge ivsalloc() assertion in test code.Jason Evans2016-06-031-1/+16
| |
| * Remove redundant chunk argument from chunk_{,de,re}register().Jason Evans2016-06-034-25/+25
| |
| * Fix opt_zero-triggered in-place huge reallocation zeroing.Jason Evans2016-06-031-4/+4
| | | | | | | | | | | | Fix huge_ralloc_no_move_expand() to update the extent's zeroed attribute based on the intersection of the previous value and that of the newly merged trailing extent.
| * Add extent_past_get().Jason Evans2016-06-032-0/+9
| |
| * Replace extent_achunk_[gs]et() with extent_slab_[gs]et().Jason Evans2016-06-038-33/+33
| |
| * Add extent_active_[gs]et().Jason Evans2016-06-037-21/+37
| | | | | | | | Always initialize extents' runs_dirty and chunks_cache linkage.
| * Move *PAGE* definitions to pages.h.Jason Evans2016-06-032-15/+15
| |
| * Set/unset rtree node for last chunk of extents.Jason Evans2016-06-031-4/+41
| | | | | | | | | | Set/unset rtree node for last chunk of extents, so that the rtree can be used for chunk coalescing.
| * Add rtree element witnesses.Jason Evans2016-06-0310-40/+241
| |
| * Refactor rtree to always use base_alloc() for node allocation.Jason Evans2016-06-0315-217/+315
| |
| * Use rtree-based chunk lookups rather than pointer bit twiddling.Jason Evans2016-06-0314-504/+548
| | | | | | | | | | | | | | | | | | | | | | Look up chunk metadata via the radix tree, rather than using CHUNK_ADDR2BASE(). Propagate pointer's containing extent. Minimize extent lookups by doing a single lookup (e.g. in free()) and propagating the pointer's extent into nearly all the functions that may need it.
| * Add element acquire/release capabilities to rtree.Jason Evans2016-06-036-136/+303
| | | | | | | | | | | | | | | | This makes it possible to acquire short-term "ownership" of rtree elements so that it is possible to read an extent pointer *and* read the extent's contents with a guarantee that the element will not be modified until the ownership is released. This is intended as a mechanism for resolving rtree read/write races rather than as a way to lock extents.
generated by cgit v0.12 at 2025-12-29 06:06:13 (GMT)