We currently iterate over objects within a pack in .idx
order, which uses the object hashes. That means that it
is effectively random with respect to the location of the
object within the pack. If you're going to access the actual
object data, there are two reasons to move linearly through
the pack itself:
1. It improves the locality of access in the packfile. In
the cold-cache case, this may mean fewer disk seeks, or
better usage of disk cache.
2. We store related deltas together in the packfile. Which
means that the delta base cache can operate much more
efficiently if we visit all of those related deltas in
sequence, as the earlier items are likely to still be
in the cache. Whereas if we visit the objects in
random order, our cache entries are much more likely to
have been evicted by unrelated deltas in the meantime.
So in general, if you're going to access the object contents
pack order is generally going to end up more efficient.
But if you're simply generating a list of object names, or
if you're going to end up sorting the result anyway, you're
better off just using the .idx order, as finding the pack
order means generating the in-memory pack-revindex.
According to the numbers in 8b8dfd5132 (pack-revindex:
radix-sort the revindex, 2013-07-11), that takes about 200ms
for linux.git, and 20ms for git.git (those numbers are a few
years old but are still a good ballpark).
That makes it a good optimization for some cases (we can
save tens of seconds in git.git by having good locality of
delta access, for a 20ms cost), but a bad one for others
(e.g., right now "cat-file --batch-all-objects
--batch-check="%(objectname)" is 170ms in git.git, so adding
20ms to that is noticeable).
Hence this patch makes it an optional flag. You can't
actually do any interesting timings yet, as it's not plumbed
through to any user-facing tools like cat-file. That will
come in a later patch.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The is_descendant_of method previously used in_merge_bases() to check if
the commit can reach any of the commits in the provided list. This had
two performance problems:
1. The performance is quadratic in worst-case.
2. A single in_merge_bases() call requires walking beyond the target
commit in order to find the full set of boundary commits that may be
merge-bases.
The can_all_from_reach method avoids this quadratic behavior and can
limit the search beyond the target commits using generation numbers. It
requires a small prototype adjustment to stop using commit-date as a
cutoff, as that optimization is no longer appropriate here.
Since in_merge_bases() uses paint_down_to_common(), is_descendant_of()
naturally found cutoffs to avoid walking the entire commit graph. Since
we want to always return the correct result, we cannot use the
min_commit_date cutoff in can_all_from_reach. We then rely on generation
numbers to provide the cutoff.
Since not all repos will have a commit-graph file, nor will we always
have generation numbers computed for a commit-graph file, create a new
method, generation_numbers_enabled(), that checks for a commit-graph
file and sees if the first commit in the file has a non-zero generation
number. In the case that we do not have generation numbers, use the old
logic for is_descendant_of().
Performance was meausured on a copy of the Linux repository using the
'test-tool reach is_descendant_of' command using this input:
A:v4.9
X:v4.10
X:v4.11
X:v4.12
X:v4.13
X:v4.14
X:v4.15
X:v4.16
X:v4.17
X.v3.0
Note that this input is tailored to demonstrate the quadratic nature of
the previous method, as it will compute merge-bases for v4.9 versus all
of the later versions before checking against v4.1.
Before: 0.26 s
After: 0.21 s
Since we previously used the is_descendant_of method in the ref_newer
method, we also measured performance there using
'test-tool reach ref_newer' with this input:
A:v4.9
B:v3.19
Before: 0.10 s
After: 0.08 s
By adding a new commit with parent v3.19, we test the non-reachable case
of ref_newer:
Before: 0.09 s
After: 0.08 s
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a struct repository argument to the functions in commit-graph.h that
read the commit graph. (This commit does not affect functions that write
commit graphs.)
Because the commit graph functions can now read the commit graph of any
repository, the global variable core_commit_graph has been removed.
Instead, the config option core.commitGraph is now read on the first
time in a repository that a commit is attempted to be parsed using its
commit graph.
This commit includes a test that exercises the functionality on an
arbitrary repository that is not the_repository.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of storing commit graphs in static variables, store them in
struct object_store. There are no changes to the signatures of existing
functions - they all still only support the_repository, and support for
other instances of struct repository will be added in a subsequent
commit.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Two functions in the code (1) check if the repository is configured for
commit graphs, (2) call prepare_commit_graph(), and (3) check if the
graph exists. Move (1) and (3) into prepare_commit_graph(), reducing
duplication of code.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a repository argument to allow callers of lookup_commit to be more
specific about which repository to handle. This is a small mechanical
change; it doesn't change the implementation to handle repositories
other than the_repository yet.
As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a repository argument to allow callers of
lookup_commit_reference_gently to be more specific about which
repository to handle. This is a small mechanical change; it doesn't
change the implementation to handle repositories other than
the_repository yet.
As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a repository argument to allow the callers of lookup_tree
to be more specific about which repository to act on. This is a small
mechanical change; it doesn't change the implementation to handle
repositories other than the_repository yet.
As with the previous commits, use a macro to catch callers passing a
repository other than the_repository at compile time.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When writing commit-graph files, it can be convenient to ask for all
reachable commits (starting at the ref set) in the resulting file. This
is particularly helpful when writing to stdin is complicated, such as a
future integration with 'git gc'.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The commit-graph file ends with a SHA1 hash of the previous contents. If
a commit-graph file has errors but the checksum hash is correct, then we
know that the problem is a bug in Git and not simply file corruption
after-the-fact.
Compute the checksum right away so it is the first error that appears,
and make the message translatable since this error can be "corrected" by
a user by simply deleting the file and recomputing. The rest of the
errors are useful only to developers.
Be sure to continue checking the rest of the file data if the checksum
is wrong. This is important for our tests, as we break the checksum as
we modify bytes of the commit-graph file.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While iterating through the commit parents, perform the generation
number calculation and compare against the value stored in the
commit-graph.
The tests demonstrate that having a different set of parents affects
the generation number calculation, and this value propagates to
descendants. Hence, we drop the single-line condition on the output.
Since Git will ship with the commit-graph feature without generation
numbers, we need to accept commit-graphs with all generation numbers
equal to zero. In this case, ignore the generation number calculation.
However, verify that we should never have a mix of zero and non-zero
generation numbers. Create a test that sets one commit to generation
zero and all following commits report a failure as they have non-zero
generation in a file that contains generation number zero.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The commit-graph file stores parents in a two-column portion of the
commit data chunk. If there is only one parent, then the second column
stores 0xFFFFFFFF to indicate no second parent.
The 'verify' subcommand checks the parent list for the commit loaded
from the commit-graph and the one parsed from the object database. Test
these checks for corrupt parents, too many parents, and wrong parents.
Add a boundary check to insert_parent_or_die() for when the parent
position value is out of range.
The octopus merge will be tested in a later commit.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'verify' subcommand must compare the commit content parsed from the
commit-graph against the content in the object database. Use
lookup_commit() and parse_commit_in_graph_one() to parse the commits
from the graph and compare against a commit that is loaded separately
and parsed directly from the object database.
Add checks for the root tree OID.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the 'verify' subcommand, load commits directly from the object
database to ensure they exist. Parse by skipping the commit-graph.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the commit-graph file, the OID fanout chunk provides an index into
the OID lookup. The 'verify' subcommand should find incorrect values
in the fanout.
Similarly, the 'verify' subcommand should find out-of-order values in
the OID lookup.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The commit-graph file requires the following three chunks:
* OID Fanout
* OID Lookup
* Commit Data
If any of these are missing, then the 'verify' subcommand should
report a failure. This includes the chunk IDs malformed or the
chunk count is truncated.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If the commit-graph file becomes corrupt, we need a way to verify
that its contents match the object database. In the manner of
'git fsck' we will implement a 'git commit-graph verify' subcommand
to report all issues with the file.
Add the 'verify' subcommand to the 'commit-graph' builtin and its
documentation. The subcommand is currently a no-op except for
loading the commit-graph into memory, which may trigger run-time
errors that would be caught by normal use. Add a simple test that
ensures the command returns a zero error code.
If no commit-graph file exists, this is an acceptable state. Do
not report any errors.
Helped-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When lazy-loading a tree for a commit, it will be important to select
the tree from a specific struct commit_graph. Create a new method that
specifies the commit-graph file and use that in
get_commit_tree_in_graph().
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before verifying a commit-graph file against the object database, we
need to parse all commits from the given commit-graph file. Create
parse_commit_in_graph_one() to target a given struct commit_graph.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The GRAPH_MIN_SIZE macro should be the smallest size of a parsable
commit-graph file. However, the minimum number of chunks was wrong.
It is possible to write a commit-graph file with zero commits, and
that violates this macro's value.
Rewrite the macro, and use extra macros to better explain the magic
constants.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update to ds/generation-numbers topic.
* ds/commit-graph-lockfile-fix:
commit-graph: fix UX issue when .lock file exists
commit-graph.txt: update design document
merge: check config before loading commits
commit: use generation number in remove_redundant()
commit: add short-circuit to paint_down_to_common()
commit: use generation numbers for in_merge_bases()
ref-filter: use generation number for --contains
commit-graph: always load commit-graph information
commit: use generations in paint_down_to_common()
commit-graph: compute generation numbers
commit: add generation number to struct commit
ref-filter: fix outdated comment on in_commit_list
The codepath around object-info API has been taught to take the
repository object (which in turn tells the API which object store
the objects are to be located).
* sb/oid-object-info:
cache.h: allow oid_object_info to handle arbitrary repositories
packfile: add repository argument to cache_or_unpack_entry
packfile: add repository argument to unpack_entry
packfile: add repository argument to read_object
packfile: add repository argument to packed_object_info
packfile: add repository argument to packed_to_object_type
packfile: add repository argument to retry_bad_packed_offset
cache.h: add repository argument to oid_object_info
cache.h: add repository argument to oid_object_info_extended
The code has been taught to use the duplicated information stored
in the commit-graph file to learn the tree object name for a commit
to avoid opening and parsing the commit object when it makes sense
to do so.
* ds/lazy-load-trees:
coccinelle: avoid wrong transformation suggestions from commit.cocci
commit-graph: lazy-load trees for commits
treewide: replace maybe_tree with accessor methods
commit: create get_commit_tree() method
treewide: rename tree to maybe_tree
We use the lockfile API to avoid multiple Git processes from writing to
the commit-graph file in the .git/objects/info directory. In some cases,
this directory may not exist, so we check for its existence.
The existing code does the following when acquiring the lock:
1. Try to acquire the lock.
2. If it fails, try to create the .git/object/info directory.
3. Try to acquire the lock, failing if necessary.
The problem is that if the lockfile exists, then the mkdir fails, giving
an error that doesn't help the user:
"fatal: cannot mkdir .git/objects/info: File exists"
While technically this honors the lockfile, it does not help the user.
Instead, do the following:
1. Check for existence of .git/objects/info; create if necessary.
2. Try to acquire the lock, failing if necessary.
The new output looks like:
fatal: Unable to create
'<dir>/.git/objects/info/commit-graph.lock': File exists.
Another git process seems to be running in this repository, e.g.
an editor opened by 'git commit'. Please make sure all processes
are terminated then try again. If it still fails, a git process
may have crashed in this repository earlier:
remove the file manually to continue.
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Most code paths load commits using lookup_commit() and then
parse_commit(). In some cases, including some branch lookups, the commit
is parsed using parse_object_buffer() which side-steps parse_commit() in
favor of parse_commit_buffer().
With generation numbers in the commit-graph, we need to ensure that any
commit that exists in the commit-graph file has its generation number
loaded.
Create new load_commit_graph_info() method to fill in the information
for a commit that exists only in the commit-graph file. Call it from
parse_commit_buffer() after loading the other commit information from
the given buffer. Only fill this information when specified by the
'check_graph' parameter.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While preparing commits to be written into a commit-graph file, compute
the generation numbers using a depth-first strategy.
The only commits that are walked in this depth-first search are those
without a precomputed generation number. Thus, computation time will be
relative to the number of new commits to the commit-graph file.
If a computed generation number would exceed GENERATION_NUMBER_MAX, then
use GENERATION_NUMBER_MAX instead.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The generation number of a commit is defined recursively as follows:
* If a commit A has no parents, then the generation number of A is one.
* If a commit A has parents, then the generation number of A is one
more than the maximum generation number among the parents of A.
Add a uint32_t generation field to struct commit so we can pass this
information to revision walks. We use three special values to signal
the generation number is invalid:
GENERATION_NUMBER_INFINITY 0xFFFFFFFF
GENERATION_NUMBER_MAX 0x3FFFFFFF
GENERATION_NUMBER_ZERO 0
The first (_INFINITY) means the generation number has not been loaded or
computed. The second (_MAX) means the generation number is too large to
store in the commit-graph file. The third (_ZERO) means the generation
number was loaded from a commit graph file that was written by a version
of git that did not support generation numbers.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Precompute and store information necessary for ancestry traversal
in a separate file to optimize graph walking.
* ds/commit-graph:
commit-graph: implement "--append" option
commit-graph: build graph from starting commits
commit-graph: read only from specific pack-indexes
commit: integrate commit graph with commit parsing
commit-graph: close under reachability
commit-graph: add core.commitGraph setting
commit-graph: implement git commit-graph read
commit-graph: implement git-commit-graph write
commit-graph: implement write_commit_graph()
commit-graph: create git-commit-graph builtin
graph: add commit graph design document
commit-graph: add format document
csum-file: refactor finalize_hashfile() method
csum-file: rename hashclose() to finalize_hashfile()
The commit-graph file provides quick access to commit data, including
the OID of the root tree for each commit in the graph. When performing
a deep commit-graph walk, we may not need to load most of the trees
for these commits.
Delay loading the tree object for a commit loaded from the graph
until requested via get_commit_tree(). Do not lazy-load trees for
commits not in the graph, since that requires duplicate parsing
and the relative peformance improvement when trees are not needed
is small.
On the Linux repository, performance tests were run for the following
command:
git log --graph --oneline -1000
Before: 0.92s
After: 0.66s
Rel %: -28.3%
Adding '-- kernel/' to the command requires loading the root tree
for every commit that is walked. There was no measureable performance
change as a result of this patch.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In anticipation of making trees load lazily, create a Coccinelle
script (contrib/coccinelle/commit.cocci) to ensure that all
references to the 'maybe_tree' member of struct commit are either
mutations or accesses through get_commit_tree() or
get_commit_tree_oid().
Apply the Coccinelle script to create the rest of the patch.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Using the commit-graph file to walk commit history removes the large
cost of parsing commits during the walk. This exposes a performance
issue: lookup_tree() takes a large portion of the computation time,
even when Git never uses those trees.
In anticipation of lazy-loading these trees, rename the 'tree' member
of struct commit to 'maybe_tree'. This serves two purposes: it hints
at the future role of possibly being NULL even if the commit has a
valid tree, and it allows for unambiguous transformation from simple
member access (i.e. commit->maybe_tree) to method access.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach git-commit-graph to add all commits from the existing
commit-graph file to the file about to be written. This should be
used when adding new commits without performing garbage collection.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach git-commit-graph to read commits from stdin when the
--stdin-commits flag is specified. Commits reachable from these
commits are added to the graph. This is a much faster way to construct
the graph than inspecting all packed objects, but is restricted to
known tips.
For the Linux repository, 700,000+ commits were added to the graph
file starting from 'master' in 7-9 seconds, depending on the number
of packfiles in the repo (1, 24, or 120).
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach git-commit-graph to inspect the objects only in a certain list
of pack-indexes within the given pack directory. This allows updating
the commit graph iteratively.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach Git to inspect a commit graph file to supply the contents of a
struct commit when calling parse_commit_gently(). This implementation
satisfies all post-conditions on the struct commit, including loading
parents, the root tree, and the commit date.
If core.commitGraph is false, then do not check graph files.
In test script t5318-commit-graph.sh, add output-matching conditions on
read-only graph operations.
By loading commits from the graph instead of parsing commit buffers, we
save a lot of time on long commit walks. Here are some performance
results for a copy of the Linux repository where 'master' has 678,653
reachable commits and is behind 'origin/master' by 59,929 commits.
| Command | Before | After | Rel % |
|----------------------------------|--------|--------|-------|
| log --oneline --topo-order -1000 | 8.31s | 0.94s | -88% |
| branch -vv | 1.02s | 0.14s | -86% |
| rev-list --all | 5.89s | 1.07s | -81% |
| rev-list --all --objects | 66.15s | 58.45s | -11% |
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach write_commit_graph() to walk all parents from the commits
discovered in packfiles. This prevents gaps given by loose objects or
previously-missed packfiles.
Also automatically add commits from the existing graph file, if it
exists.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach git-commit-graph to read commit graph files and summarize their contents.
Use the read subcommand to verify the contents of a commit graph file in the
tests.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach Git to write a commit graph file by checking all packed objects
to see if they are commits, then store the file in the given object
directory.
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
