| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Optimize out repeated rename detection in a sequence of mergy
operations.
* en/ort-perf-batch-11:
merge-ort, diffcore-rename: employ cached renames when possible
merge-ort: handle interactions of caching and rename/rename(1to1) cases
merge-ort: add helper functions for using cached renames
merge-ort: preserve cached renames for the appropriate side
merge-ort: avoid accidental API mis-use
merge-ort: add code to check for whether cached renames can be reused
merge-ort: populate caches of rename detection results
merge-ort: add data structures for in-memory caching of rename detection
t6429: testcases for remembering renames
fast-rebase: write conflict state to working tree, index, and HEAD
fast-rebase: change assert() to BUG()
Documentation/technical: describe remembering renames optimization
t6423: rename file within directory that other side renamed
|
|
Doc fix.
* tl/fix-packfile-uri-doc:
packfile-uri.txt: fix blobPackfileUri description
|
|
Fix the 'uploadpack.blobPackfileUri' description in packfile-uri.txt
and the correct format also can be seen in t5702.
Signed-off-by: Teng Long <dyroneteng@gmail.com>
Reviewed-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We will soon be adding an optimization that caches (in memory only,
never written to disk) upstream renames during a sequence of merges such
as occurs during a cherry-pick or rebase operation. Add several tests
meant to stress such an implementation to ensure it does the right
thing, and include a test whose outcome we will later change due to this
optimization as well.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Remembering renames on the upstream side of history in an early merge of
a rebase or cherry-pick for re-use in a latter merge of the same
operation makes pretty good intuitive sense. However, trying to show
that it doesn't cause some subtle behavioral difference or some funny
edge or corner case is much more involved. And, in fact, it does
introduce a subtle behavioral change.
Document all the assumptions, special cases, and logic involved in such
an optimization, and describe why this optimization is safe under the
current optimizations/features/etc. -- even when the subtle behavioral
change is triggered.
Part of the point of adding this document that goes over the
optimization in such laborious detail, is that it is possible that
significant future changes (optimizations or feature changes) could
interact with this optimization in interesting ways; this document is
here to help folks making big changes sanity check that the assumptions
and arguments underlying this optimization are still valid. (As a side
note, creating this document forced me to review things in sufficient
detail that I found I was not properly caching directory-rename-induced
renames, resulting in the code not being aware of those renames and
causing unnecessary diffcore_rename_extended() calls in subsequent
merges.)
A subsequent commit will add several testcases based on this document
meant to stress-test the implementation and also document the case with
the subtle behavioral change.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
"git push" learns to discover common ancestor with the receiving
end over protocol v2.
* jt/push-negotiation:
send-pack: support push negotiation
fetch: teach independent negotiation (no packfile)
fetch-pack: refactor command and capability write
fetch-pack: refactor add_haves()
fetch-pack: refactor process_acks()
|
|
Over-the-wire protocol learns a new request type to ask for object
sizes given a list of object names.
* ba/object-info:
object-info: support for retrieving object info
|
|
Currently, the packfile negotiation step within a Git fetch cannot be
done independent of sending the packfile, even though there is at least
one application wherein this is useful. Therefore, make it possible for
this negotiation step to be done independently. A subsequent commit will
use this for one such application - push negotiation.
This feature is for protocol v2 only. (An implementation for protocol v0
would require a separate implementation in the fetch, transport, and
transport helper code.)
In the protocol, the main hindrance towards independent negotiation is
that the server can unilaterally decide to send the packfile. This is
solved by a "wait-for-done" argument: the server will then wait for the
client to say "done". In practice, the client will never say it; instead
it will cease requests once it is satisfied.
In the client, the main change lies in the transport and transport
helper code. fetch_refs_via_pack() performs everything needed - protocol
version and capability checks, and the negotiation itself.
There are 2 code paths that do not go through fetch_refs_via_pack() that
needed to be individually excluded: the bundle transport (excluded
through requiring smart_options, which the bundle transport doesn't
support) and transport helpers that do not support takeover. If or when
we support independent negotiation for protocol v0, we will need to
modify these 2 code paths to support it. But for now, report failure if
independent negotiation is requested in these cases.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The checkout machinery has been taught to perform the actual
write-out of the files in parallel when able.
* mt/parallel-checkout-part-2:
parallel-checkout: add design documentation
parallel-checkout: support progress displaying
parallel-checkout: add configuration options
parallel-checkout: make it truly parallel
unpack-trees: add basic support for parallel checkout
|
|
Builds on top of the sparse-index infrastructure to mark operations
that are not ready to mark with the sparse index, causing them to
fall back on fully-populated index that they always have worked with.
* ds/sparse-index-protections: (47 commits)
name-hash: use expand_to_path()
sparse-index: expand_to_path()
name-hash: don't add directories to name_hash
revision: ensure full index
resolve-undo: ensure full index
read-cache: ensure full index
pathspec: ensure full index
merge-recursive: ensure full index
entry: ensure full index
dir: ensure full index
update-index: ensure full index
stash: ensure full index
rm: ensure full index
merge-index: ensure full index
ls-files: ensure full index
grep: ensure full index
fsck: ensure full index
difftool: ensure full index
commit: ensure full index
checkout: ensure full index
...
|
|
Sometimes it is useful to get information of an object without having to
download it completely.
Add the "object-info" capability that lets the client ask for
object-related information with their full hexadecimal object names.
Only sizes are returned for now.
Signed-off-by: Bruno Albuquerque <bga@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Documentation updates, with unrelated comment updates, too.
* ab/usage-error-docs:
api docs: document that BUG() emits a trace2 error event
api docs: document BUG() in api-error-handling.txt
usage.c: don't copy/paste the same comment three times
|
|
Doc updte.
* hn/reftable-tables-doc-update:
reftable: document an alternate cleanup method on Windows
|
|
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Edit and expand the sparse-index design document with the plan for
guarding index operations with ensure_full_index().
Notably, the plan has changed to not have an expand_to_path() method in
favor of checking for a sparse-directory hit inside of the
index_path_pos() API.
The changes that follow this one will incrementally add
ensure_full_index() guards to iterations over all cache entries. Some
iterations over the cache entries are not protected due to a few
categories listed in the document. Since these are not being modified,
here is a short list of the files and methods that will not receive
these guards:
Looking for non-zero stage:
* builtin/add.c:chmod_pathspec()
* builtin/merge.c:count_unmerged_entries()
* merge-ort.c:record_conflicted_index_entries()
* read-cache.c:unmerged_index()
* rerere.c:check_one_conflict(), find_conflict(), rerere_remaining()
* revision.c:prepare_show_merge()
* sequencer.c:append_conflicts_hint()
* wt-status.c:wt_status_collect_changes_initial()
Looking for submodules:
* builtin/submodule--helper.c:module_list_compute()
* submodule.c: several methods
* worktree.c:validate_no_submodules()
Part of the index API:
* name-hash.c: lazy init methods
* preload-index.c:preload_thread(), preload_index()
* read-cache.c: file format methods
Checking for correct order of cache entries:
* read-cache.c:check_ce_order()
Ignores SKIP_WORKTREE entries or already aware:
* unpack-trees.c:mark_new_skip_worktree()
* wt-status.c:wt_status_check_sparse_checkout()
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Correct documentation added in e544221d97a (trace2:
Documentation/technical/api-trace2.txt, 2019-02-22) to state that
calling BUG() also emits an "error" event. See ee4512ed481 (trace2:
create new combined trace facility, 2019-02-22) for the initial
implementation.
The BUG() function did not emit an event then however, that was only
changed later in 0a9dde4a04c (usage: trace2 BUG() invocations,
2021-02-05), that commit changed the code, but didn't update any of
the docs.
Let's also add a cross-reference from api-error-handling.txt.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When the BUG() function was added in d8193743e08 (usage.c: add BUG()
function, 2017-05-12) these docs added in 1f23cfe0ef5 (doc: document
error handling functions and conventions, 2014-12-03) were not
updated. Let's do that.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The new method uses the update_index counter, which isn't susceptible to clock
inaccuracies.
Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
An on-disk reverse-index to map the in-pack location of an object
back to its object name across multiple packfiles is introduced.
* tb/reverse-midx:
midx.c: improve cache locality in midx_pack_order_cmp()
pack-revindex: write multi-pack reverse indexes
pack-write.c: extract 'write_rev_file_order'
pack-revindex: read multi-pack reverse indexes
Documentation/technical: describe multi-pack reverse indexes
midx: make some functions non-static
midx: keep track of the checksum
midx: don't free midx_name early
midx: allow marking a pack as preferred
t/helper/test-read-midx.c: add '--show-objects'
builtin/multi-pack-index.c: display usage on unrecognized command
builtin/multi-pack-index.c: don't enter bogus cmd_mode
builtin/multi-pack-index.c: split sub-commands
builtin/multi-pack-index.c: define common usage with a macro
builtin/multi-pack-index.c: don't handle 'progress' separately
builtin/multi-pack-index.c: inline 'flags' with options
|
|
A simple IPC interface gets introduced to build services like
fsmonitor on top.
* jh/simple-ipc:
t0052: add simple-ipc tests and t/helper/test-simple-ipc tool
simple-ipc: add Unix domain socket implementation
unix-stream-server: create unix domain socket under lock
unix-socket: disallow chdir() when creating unix domain sockets
unix-socket: add backlog size option to unix_stream_listen()
unix-socket: eliminate static unix_stream_socket() helper function
simple-ipc: add win32 implementation
simple-ipc: design documentation for new IPC mechanism
pkt-line: add options argument to read_packetized_to_strbuf()
pkt-line: add PACKET_READ_GENTLE_ON_READ_ERROR option
pkt-line: do not issue flush packets in write_packetized_*()
pkt-line: eliminate the need for static buffer in packet_write_gently()
|
|
As a prerequisite to implementing multi-pack bitmaps, motivate and
describe the format and ordering of the multi-pack reverse index.
The subsequent patch will implement reading this format, and the patch
after that will implement writing it while producing a multi-pack index.
Co-authored-by: Jeff King <peff@peff.net>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When multiple packs in the multi-pack index contain the same object, the
MIDX machinery must make a choice about which pack it associates with
that object. Prior to this patch, the lowest-ordered[1] pack was always
selected.
Pack selection for duplicate objects is relatively unimportant today,
but it will become important for multi-pack bitmaps. This is because we
can only invoke the pack-reuse mechanism when all of the bits for reused
objects come from the reuse pack (in order to ensure that all reused
deltas can find their base objects in the same pack).
To encourage the pack selection process to prefer one pack over another
(the pack to be preferred is the one a caller would like to later use as
a reuse pack), introduce the concept of a "preferred pack". When
provided, the MIDX code will always prefer an object found in a
preferred pack over any other.
No format changes are required to store the preferred pack, since it
will be able to be inferred with a corresponding MIDX bitmap, by looking
up the pack associated with the object in the first bit position (this
ordering is described in detail in a subsequent commit).
[1]: the ordering is specified by MIDX internals; for our purposes we
can consider the "lowest ordered" pack to be "the one with the
most-recent mtime.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The index format does not currently allow for sparse directory entries.
This violates some expectations that older versions of Git or
third-party tools might not understand. We need an indicator inside the
index file to warn these tools to not interact with a sparse index
unless they are aware of sparse directory entries.
Add a new _required_ index extension, 'sdir', that indicates that the
index may contain sparse directory entries. This allows us to continue
to use the differences in index formats 2, 3, and 4 before we create a
new index version 5 in a later change.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
This begins a long effort to update the index format to allow sparse
directory entries. This should result in a significant improvement to
Git commands when HEAD contains millions of files, but the user has
selected many fewer files to keep in their sparse-checkout definition.
Currently, the index format is only updated in the presence of
extensions.sparseIndex instead of increasing a file format version
number. This is temporary, and index v5 is part of the plan for future
work in this area.
The design document details many of the reasons for embarking on this
work, and also the plan for completing it safely.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Brief design documentation for new IPC mechanism allowing
foreground Git client to talk with an existing daemon process
at a known location using a named pipe or unix domain socket.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Documentation update.
* hn/reftable-tables-doc-update:
doc/reftable: document how to handle windows
|
|
The common code to deal with "chunked file format" that is shared
by the multi-pack-index and commit-graph files have been factored
out, to help codepaths for both filetypes to become more robust.
* ds/chunked-file-api:
commit-graph.c: display correct number of chunks when writing
chunk-format: add technical docs
chunk-format: restore duplicate chunk checks
midx: use 64-bit multiplication for chunk sizes
midx: use chunk-format read API
commit-graph: use chunk-format read API
chunk-format: create read chunk API
midx: use chunk-format API in write_midx_internal()
midx: drop chunk progress during write
midx: return success/failure in chunk write methods
midx: add num_large_offsets to write_midx_context
midx: add pack_perm to write_midx_context
midx: add entries to write_midx_context
midx: use context in write_midx_pack_names()
midx: rename pack_info to write_midx_context
commit-graph: use chunk-format write API
chunk-format: create chunk format write API
commit-graph: anonymize data in chunk_write_fn
|
|
A handful of multi-word configuration variable names in
documentation that are spelled in all lowercase have been corrected
to use the more canonical camelCase.
* dl/doc-config-camelcase:
index-format doc: camelCase core.excludesFile
blame-options.txt: camelcase blame.blankBoundary
i18n.txt: camel case and monospace "i18n.commitEncoding"
|
|
Docfix.
* js/doc-proto-v2-response-end:
doc: fix naming of response-end-pkt
|
|
* ds/chunked-file-api:
commit-graph.c: display correct number of chunks when writing
chunk-format: add technical docs
chunk-format: restore duplicate chunk checks
midx: use 64-bit multiplication for chunk sizes
midx: use chunk-format read API
commit-graph: use chunk-format read API
chunk-format: create read chunk API
midx: use chunk-format API in write_midx_internal()
midx: drop chunk progress during write
midx: return success/failure in chunk write methods
midx: add num_large_offsets to write_midx_context
midx: add pack_perm to write_midx_context
midx: add entries to write_midx_context
midx: use context in write_midx_pack_names()
midx: rename pack_info to write_midx_context
commit-graph: use chunk-format write API
chunk-format: create chunk format write API
commit-graph: anonymize data in chunk_write_fn
|
|
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
On Windows we can't delete or overwrite files opened by other processes. Here we
sketch how to handle this situation.
We propose to use a random element in the filename. It's possible to design an
alternate solution based on counters, but that would assign semantics to the
filenames that complicates implementation.
Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Update formatting and grammar of the hash transition plan
documentation, plus some updates.
* ta/hash-function-transition-doc:
doc: use https links
doc hash-function-transition: move rationale upwards
doc hash-function-transition: fix incomplete sentence
doc hash-function-transition: use upper case consistently
doc hash-function-transition: use SHA-1 and SHA-256 consistently
doc hash-function-transition: fix asciidoc output
|
|
The chunk-based file format is now an API in the code, but we should
also take time to document it as a file format. Specifically, it matches
the CHUNK LOOKUP sections of the commit-graph and multi-pack-index
files, but there are some commonalities that should be grouped in this
document.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
"git clone" tries to locally check out the branch pointed at by
HEAD of the remote repository after it is done, but the protocol
did not convey the information necessary to do so when copying an
empty repository. The protocol v2 learned how to do so.
* jt/clone-unborn-head:
clone: respect remote unborn HEAD
connect, transport: encapsulate arg in struct
ls-refs: report unborn targets of symrefs
|
|
The commit-graph learned to use corrected commit dates instead of
the generation number to help topological revision traversal.
* ak/corrected-commit-date:
doc: add corrected commit date info
commit-reach: use corrected commit dates in paint_down_to_common()
commit-graph: use generation v2 only if entire chain does
commit-graph: implement generation data chunk
commit-graph: implement corrected commit date
commit-graph: return 64-bit generation number
commit-graph: add a slab to store topological levels
t6600-test-reach: generalize *_three_modes
commit-graph: consolidate fill_commit_graph_info
revision: parse parent in indegree_walk_step()
commit-graph: fix regression when computing Bloom filters
|
|
Git Protocol version 2[1] defines 0002 as a Message Packet that indicates
the end of a response for stateless connections.
Change the naming of the 0002 Packet to 'Response End' to match the
parsing introduced in Wireshark's MR !1922 for consistency. A subsequent
MR in Wireshark will address additional mismatches.
[1] kernel.org/pub/software/scm/git/docs/technical/protocol-v2.html
[2] gitlab.com/wireshark/wireshark/-/merge_requests/1922
Signed-off-by: Joey Salazar <jgsal@protonmail.com>
Reviewed-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Introduce an on-disk file to record revindex for packdata, which
traditionally was always created on the fly and only in-core.
* tb/pack-revindex-on-disk:
t5325: check both on-disk and in-memory reverse index
pack-revindex: ensure that on-disk reverse indexes are given precedence
t: support GIT_TEST_WRITE_REV_INDEX
t: prepare for GIT_TEST_WRITE_REV_INDEX
Documentation/config/pack.txt: advertise 'pack.writeReverseIndex'
builtin/pack-objects.c: respect 'pack.writeReverseIndex'
builtin/index-pack.c: write reverse indexes
builtin/index-pack.c: allow stripping arbitrary extensions
pack-write.c: prepare to write 'pack-*.rev' files
packfile: prepare for the existence of '*.rev' files
|
|
Doc update.
* ma/doc-pack-format-varint-for-sizes:
pack-format.txt: document sizes at start of delta data
|
|
Document, clean-up and optimize the code around the cache-tree
extension in the index.
* ds/cache-tree-basics:
cache-tree: speed up consecutive path comparisons
cache-tree: use ce_namelen() instead of strlen()
index-format: discuss recursion of cache-tree better
index-format: update preamble to cache tree extension
index-format: use 'cache tree' over 'cached tree'
cache-tree: trace regions for prime_cache_tree
cache-tree: trace regions for I/O
cache-tree: use trace2 in cache_tree_update()
unpack-trees: add trace2 regions
tree-walk: report recursion counts
|
|
Doc fix for packfile URI feature.
* jt/packfile-as-uri-doc:
Doc: clarify contents of packfile sent as URI
|
|
When cloning, we choose the default branch based on the remote HEAD.
But if there is no remote HEAD reported (which could happen if the
target of the remote HEAD is unborn), we'll fall back to using our local
init.defaultBranch. Traditionally this hasn't been a big deal, because
most repos used "master" as the default. But these days it is likely to
cause confusion if the server and client implementations choose
different values (e.g., if the remote started with "main", we may choose
"master" locally, create commits there, and then the user is surprised
when they push to "master" and not "main").
To solve this, the remote needs to communicate the target of the HEAD
symref, even if it is unborn, and "git clone" needs to use this
information.
Currently, symrefs that have unborn targets (such as in this case) are
not communicated by the protocol. Teach Git to advertise and support the
"unborn" feature in "ls-refs" (by default, this is advertised, but
server administrators may turn this off through the lsrefs.unborn
config). This feature indicates that "ls-refs" supports the "unborn"
argument; when it is specified, "ls-refs" will send the HEAD symref with
the name of its unborn target.
This change is only for protocol v2. A similar change for protocol v0
would require independent protocol design (there being no analogous
position to signal support for "unborn") and client-side plumbing of the
data required, so the scope of this patch set is limited to protocol v2.
The client side will be updated to use this in a subsequent commit.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Use only https links for lore.kernel.org.
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Move rationale for new hash function to beginning of document
so that it appears before the concrete move to SHA-256 is described.
Remove some of the details about SHA-1 weaknesses and add references
to the details on how the new hash function was chosen instead.
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Use upper case consistently in Document History.
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Use SHA-1 and SHA-256 instead of sha1 and sha256 when referring
to the hash type.
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Asciidoc requires lists to start with an empty line and uses
different characters for indentation levels ("-", "*", "**", ...).
For special symbols like a dash "--" has to be used and there is
no double arrow "<->", so a left and right arrow "<-->" has to be
combined for that. Lastly for verbatim output a newline followed
by an indentation has to be used.
Fix asciidoc output for lists, special characters and verbatim
text while retaining the readabilty of the original text file.
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Doc fix for packfile URI feature.
* jt/packfile-as-uri-doc:
Doc: clarify contents of packfile sent as URI
|
|
Specify the format of the on-disk reverse index 'pack-*.rev' file, as
well as prepare the code for the existence of such files.
The reverse index maps from pack relative positions (i.e., an index into
the array of object which is sorted by their offsets within the
packfile) to their position within the 'pack-*.idx' file. Today, this is
done by building up a list of (off_t, uint32_t) tuples for each object
(the off_t corresponding to that object's offset, and the uint32_t
corresponding to its position in the index). To convert between pack and
index position quickly, this array of tuples is radix sorted based on
its offset.
This has two major drawbacks:
First, the in-memory cost scales linearly with the number of objects in
a pack. Each 'struct revindex_entry' is sizeof(off_t) +
sizeof(uint32_t) + padding bytes for a total of 16.
To observe this, force Git to load the reverse index by, for e.g.,
running 'git cat-file --batch-check="%(objectsize:disk)"'. When asking
for a single object in a fresh clone of the kernel, Git needs to
allocate 120+ MB of memory in order to hold the reverse index in memory.
Second, the cost to sort also scales with the size of the pack.
Luckily, this is a linear function since 'load_pack_revindex()' uses a
radix sort, but this cost still must be paid once per pack per process.
As an example, it takes ~60x longer to print the _size_ of an object as
it does to print that entire object's _contents_:
Benchmark #1: git.compile cat-file --batch <obj
Time (mean ± σ): 3.4 ms ± 0.1 ms [User: 3.3 ms, System: 2.1 ms]
Range (min … max): 3.2 ms … 3.7 ms 726 runs
Benchmark #2: git.compile cat-file --batch-check="%(objectsize:disk)" <obj
Time (mean ± σ): 210.3 ms ± 8.9 ms [User: 188.2 ms, System: 23.2 ms]
Range (min … max): 193.7 ms … 224.4 ms 13 runs
Instead, avoid computing and sorting the revindex once per process by
writing it to a file when the pack itself is generated.
The format is relatively straightforward. It contains an array of
uint32_t's, the length of which is equal to the number of objects in the
pack. The ith entry in this table contains the index position of the
ith object in the pack, where "ith object in the pack" is determined by
pack offset.
One thing that the on-disk format does _not_ contain is the full (up to)
eight-byte offset corresponding to each object. This is something that
the in-memory revindex contains (it stores an off_t in 'struct
revindex_entry' along with the same uint32_t that the on-disk format
has). Omit it in the on-disk format, since knowing the index position
for some object is sufficient to get a constant-time lookup in the
pack-*.idx file to ask for an object's offset within the pack.
This trades off between the on-disk size of the 'pack-*.rev' file for
runtime to chase down the offset for some object. Even though the lookup
is constant time, the constant is heavier, since it can potentially
involve two pointer walks in v2 indexes (one to access the 4-byte offset
table, and potentially a second to access the double wide offset table).
Consider trying to map an object's pack offset to a relative position
within that pack. In a cold-cache scenario, more page faults occur while
switching between binary searching through the reverse index and
searching through the *.idx file for an object's offset. Sure enough,
with a cold cache (writing '3' into '/proc/sys/vm/drop_caches' after
'sync'ing), printing out the entire object's contents is still
marginally faster than printing its size:
Benchmark #1: git.compile cat-file --batch-check="%(objectsize:disk)" <obj >/dev/null
Time (mean ± σ): 22.6 ms ± 0.5 ms [User: 2.4 ms, System: 7.9 ms]
Range (min … max): 21.4 ms … 23.5 ms 41 runs
Benchmark #2: git.compile cat-file --batch <obj >/dev/null
Time (mean ± σ): 17.2 ms ± 0.7 ms [User: 2.8 ms, System: 5.5 ms]
Range (min … max): 15.6 ms … 18.2 ms 45 runs
(Numbers taken in the kernel after cheating and using the next patch to
generate a reverse index). There are a couple of approaches to improve
cold cache performance not pursued here:
- We could include the object offsets in the reverse index format.
Predictably, this does result in fewer page faults, but it triples
the size of the file, while simultaneously duplicating a ton of data
already available in the .idx file. (This was the original way I
implemented the format, and it did show
`--batch-check='%(objectsize:disk)'` winning out against `--batch`.)
On the other hand, this increase in size also results in a large
block-cache footprint, which could potentially hurt other workloads.
- We could store the mapping from pack to index position in more
cache-friendly way, like constructing a binary search tree from the
table and writing the values in breadth-first order. This would
result in much better locality, but the price you pay is trading
O(1) lookup in 'pack_pos_to_index()' for an O(log n) one (since you
can no longer directly index the table).
So, neither of these approaches are taken here. (Thankfully, the format
is versioned, so we are free to pursue these in the future.) But, cold
cache performance likely isn't interesting outside of one-off cases like
asking for the size of an object directly. In real-world usage, Git is
often performing many operations in the revindex (i.e., asking about
many objects rather than a single one).
The trade-off is worth it, since we will avoid the vast majority of the
cost of generating the revindex that the extra pointer chase will look
like noise in the following patch's benchmarks.
This patch describes the format and prepares callers (like in
pack-revindex.c) to be able to read *.rev files once they exist. An
implementation of the writer will appear in the next patch, and callers
will gradually begin to start using the writer in the patches that
follow after that.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
Junio C Hamano
Teng Long
Elijah Newren
Jonathan Tan
Bruno Albuquerque
Matheus Tavares
Derrick Stolee
Ævar Arnfjörð Bjarmason
Han-Wen Nienhuys
Taylor Blau
Jeff Hostetler
Joey Salazar
Thomas Ackermann