git/reftable
Patrick Steinhardt 6fdfaf15a0 reftable/stack: use stat info to avoid re-reading stack list
Whenever we call into the refs interfaces we potentially have to reload
refs in case they have been concurrently modified, either in-process or
externally. While this happens somewhat automatically for loose refs
because we simply try to re-read the files, the "packed" backend will
reload its snapshot of the packed-refs file in case its stat info has
changed since last reading it.

In the reftable backend we have a similar mechanism that is provided by
`reftable_stack_reload()`. This function will read the list of stacks
from "tables.list" and, if they have changed from the currently stored
list, reload the stacks. This is heavily inefficient though, as we have
to check whether the stack is up-to-date on basically every read and
thus keep on re-reading the file all the time even if it didn't change
at all.

We can do better and use the same stat(3P)-based mechanism that the
"packed" backend uses. Instead of reading the file, we will only open
the file descriptor, fstat(3P) it, and then compare the info against the
cached value from the last time we have updated the stack. This should
always work alright because "tables.list" is updated atomically via a
rename, so even if the ctime or mtime wasn't granular enough to identify
a change, at least the inode number or file size should have changed.

This change significantly speeds up operations where many refs are read,
like when using git-update-ref(1). The following benchmark creates N
refs in an otherwise-empty repository via `git update-ref --stdin`:

  Benchmark 1: update-ref: create many refs (refcount = 1, revision = HEAD~)
    Time (mean ± σ):       5.1 ms ±   0.2 ms    [User: 2.4 ms, System: 2.6 ms]
    Range (min … max):     4.8 ms …   7.2 ms    109 runs

  Benchmark 2: update-ref: create many refs (refcount = 100, revision = HEAD~)
    Time (mean ± σ):      19.1 ms ±   0.9 ms    [User: 8.9 ms, System: 9.9 ms]
    Range (min … max):    18.4 ms …  26.7 ms    72 runs

  Benchmark 3: update-ref: create many refs (refcount = 10000, revision = HEAD~)
    Time (mean ± σ):      1.336 s ±  0.018 s    [User: 0.590 s, System: 0.724 s]
    Range (min … max):    1.314 s …  1.373 s    10 runs

  Benchmark 4: update-ref: create many refs (refcount = 1, revision = HEAD)
    Time (mean ± σ):       5.1 ms ±   0.2 ms    [User: 2.4 ms, System: 2.6 ms]
    Range (min … max):     4.8 ms …   7.2 ms    109 runs

  Benchmark 5: update-ref: create many refs (refcount = 100, revision = HEAD)
    Time (mean ± σ):      14.8 ms ±   0.2 ms    [User: 7.1 ms, System: 7.5 ms]
    Range (min … max):    14.2 ms …  15.2 ms    82 runs

  Benchmark 6: update-ref: create many refs (refcount = 10000, revision = HEAD)
    Time (mean ± σ):     927.6 ms ±   5.3 ms    [User: 437.8 ms, System: 489.5 ms]
    Range (min … max):   919.4 ms … 936.4 ms    10 runs

  Summary
    update-ref: create many refs (refcount = 1, revision = HEAD) ran
      1.00 ± 0.07 times faster than update-ref: create many refs (refcount = 1, revision = HEAD~)
      2.89 ± 0.14 times faster than update-ref: create many refs (refcount = 100, revision = HEAD)
      3.74 ± 0.25 times faster than update-ref: create many refs (refcount = 100, revision = HEAD~)
    181.26 ± 8.30 times faster than update-ref: create many refs (refcount = 10000, revision = HEAD)
    261.01 ± 12.35 times faster than update-ref: create many refs (refcount = 10000, revision = HEAD~)

Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2024-01-11 12:10:59 -08:00
..
basics.c reftable: utility functions 2021-10-08 10:45:48 -07:00
basics.h reftable: utility functions 2021-10-08 10:45:48 -07:00
basics_test.c reftable: utility functions 2021-10-08 10:45:48 -07:00
block.c reftable/block: reuse buffer to compute record keys 2023-12-11 07:23:17 -08:00
block.h reftable/block: reuse buffer to compute record keys 2023-12-11 07:23:17 -08:00
block_test.c reftable/block: introduce macro to initialize struct block_iter 2023-12-11 07:23:17 -08:00
blocksource.c reftable: handle interrupted reads 2023-12-11 07:23:16 -08:00
blocksource.h reftable: add blocksource, an abstraction for random access reads 2021-10-08 10:45:48 -07:00
constants.h reftable: (de)serialization for the polymorphic record type. 2021-10-08 10:45:48 -07:00
dump.c hash-ll.h: split out of hash.h to remove dependency on repository.h 2023-04-24 12:47:32 -07:00
error.c reftable: ensure git-compat-util.h is the first (indirect) include 2023-04-24 12:47:33 -07:00
generic.c reftable: make assignments portable to AIX xlc v12.01 2022-03-28 13:58:10 -07:00
generic.h reftable: generic interface to tables 2021-10-08 10:45:48 -07:00
iter.c reftable: make reftable_record a tagged union 2022-01-20 11:31:53 -08:00
iter.h reftable/block: introduce macro to initialize struct block_iter 2023-12-11 07:23:17 -08:00
LICENSE
merged.c reftable/merged: reuse buffer to compute record keys 2023-12-11 07:23:16 -08:00
merged.h reftable/merged: reuse buffer to compute record keys 2023-12-11 07:23:16 -08:00
merged_test.c reftable tests: avoid "int" overflow, use "uint64_t" 2022-01-13 13:39:09 -08:00
pq.c reftable: use a pointer for pq_entry param 2022-09-15 11:32:37 -07:00
pq.h reftable: use a pointer for pq_entry param 2022-09-15 11:32:37 -07:00
pq_test.c reftable: use a pointer for pq_entry param 2022-09-15 11:32:37 -07:00
publicbasics.c reftable: ensure git-compat-util.h is the first (indirect) include 2023-04-24 12:47:33 -07:00
reader.c reftable/block: introduce macro to initialize struct block_iter 2023-12-11 07:23:17 -08:00
reader.h reftable: read reftable files 2021-10-08 10:45:48 -07:00
readwrite_test.c reftable/stack: fix use of unseeded randomness 2023-12-11 07:23:16 -08:00
record.c reftable: add print functions to the record types 2022-01-20 11:31:53 -08:00
record.h reftable: add print functions to the record types 2022-01-20 11:31:53 -08:00
record_test.c reftable: make assignments portable to AIX xlc v12.01 2022-03-28 13:58:10 -07:00
refname.c reftable: implement refname validation 2021-10-08 10:45:48 -07:00
refname.h reftable: implement refname validation 2021-10-08 10:45:48 -07:00
refname_test.c reftable: implement refname validation 2021-10-08 10:45:48 -07:00
reftable-blocksource.h reftable: add blocksource, an abstraction for random access reads 2021-10-08 10:45:48 -07:00
reftable-error.h reftable: signal overflow 2021-12-23 12:28:34 -08:00
reftable-generic.h reftable: generic interface to tables 2021-10-08 10:45:48 -07:00
reftable-iterator.h reftable: generic interface to tables 2021-10-08 10:45:48 -07:00
reftable-malloc.h reftable: utility functions 2021-10-08 10:45:48 -07:00
reftable-merged.h reftable: add merged table view 2021-10-08 10:45:48 -07:00
reftable-reader.h reftable: read reftable files 2021-10-08 10:45:48 -07:00
reftable-record.h reftable: make reftable-record.h function signatures const correct 2022-01-20 11:31:53 -08:00
reftable-stack.h reftable: implement stack, a mutable database of reftable files. 2021-10-08 10:45:48 -07:00
reftable-tests.h reftable: add a heap-based priority queue for reftable records 2021-10-08 10:45:48 -07:00
reftable-writer.h reftable: rename writer_stats to reftable_writer_stats 2022-02-23 13:36:26 -08:00
stack.c reftable/stack: use stat info to avoid re-reading stack list 2024-01-11 12:10:59 -08:00
stack.h reftable/stack: use stat info to avoid re-reading stack list 2024-01-11 12:10:59 -08:00
stack_test.c reftable/stack: perform auto-compaction with transactional interface 2023-12-11 07:23:16 -08:00
system.h reftable/stack: use stat info to avoid re-reading stack list 2024-01-11 12:10:59 -08:00
test_framework.c reftable: utility functions 2021-10-08 10:45:48 -07:00
test_framework.h reftable: wrap EXPECT macros in do/while 2023-12-11 07:23:15 -08:00
tree.c reftable: ensure git-compat-util.h is the first (indirect) include 2023-04-24 12:47:33 -07:00
tree.h reftable: a generic binary tree implementation 2021-10-08 10:45:48 -07:00
tree_test.c reftable: ensure git-compat-util.h is the first (indirect) include 2023-04-24 12:47:33 -07:00
writer.c Merge branch 'ep/maint-equals-null-cocci' 2022-05-20 15:26:59 -07:00
writer.h reftable: write reftable files 2021-10-08 10:45:48 -07:00