| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The reftable library aborts with a bug in case `reftable_record_cmp()`
is invoked with two records of differing types. This would cause the
program to die without the caller being able to handle the error, which
is not something we want in the context of library code. And it ties us
to the Git codebase.
Refactor the code such that `reftable_record_cmp()` returns an error
code separate from the actual comparison result. This requires us to
also adapt some callers up the callchain in a similar fashion.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We're aborting the program via `BUG()` in case `reftable_record_init()`
was invoked with an unknown record type. This is bad because we may now
die in library code, and because it makes us depend on the Git codebase.
Refactor the code such that `reftable_record_init()` can return an error
code to the caller. Adapt any callers accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Drop our use of `COPY_ARRAY()`, replacing it with an open-coded variant
thereof. This is done to reduce our dependency on the Git library.
While at it, guard the whole array copy logic so that we only copy it in
case there actually is anything to be copied. Otherwise, we may end up
trying to allocate a zero-sized array, which will return a NULL pointer
and thus cause us to return an `REFTABLE_OUT_OF_MEMORY_ERROR`.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We use `xmmap()` to map reftables into memory. This function has two
problems:
- It causes us to die in case the mmap fails.
- It ties us to the Git codebase.
Refactor the code to use mmap(3p) instead with manual error checking.
Note that this function may not be the system-provided mmap(3p), but may
point to our `git_mmap()` wrapper that emulates the syscall on systems
that do not have mmap(3p) available.
Fix `reftable_block_source_from_file()` to properly bubble up the error
code in case the map(3p) call fails.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Similar to the preceding commit, drop our use of `write_in_full()` and
implement a new wrapper `reftable_write_full()` that handles this logic
for us. This is done to reduce our dependency on the Git library.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
There is a single callsite of `read_in_full()` in the reftable library.
Open-code the function to reduce our dependency on the Git library.
Note that we only partially port over the logic from `read_in_full()`
and its underlying `xread()` helper. Most importantly, the latter also
knows to handle `EWOULDBLOCK` via `handle_nonblock()`. This logic is
irrelevant for us though because the reftable library never sets the
`O_NONBLOCK` option in the first place.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Code clean-up.
* kn/reflog-migration-fix-followup:
reftable: prevent 'update_index' changes after adding records
refs: use 'uint64_t' for 'ref_update.index'
refs: mark `ref_transaction_update_reflog()` as static
|
|
The code paths to interact with zlib has been cleaned up in
preparation for building with zlib-ng.
* ps/zlib-ng:
ci: make "linux-musl" job use zlib-ng
ci: switch linux-musl to use Meson
compat/zlib: allow use of zlib-ng as backend
git-zlib: cast away potential constness of `next_in` pointer
compat/zlib: provide stubs for `deflateSetHeader()`
compat/zlib: provide `deflateBound()` shim centrally
git-compat-util: move include of "compat/zlib.h" into "git-zlib.h"
compat: introduce new "zlib.h" header
git-compat-util: drop `z_const` define
compat: drop `uncompress2()` compatibility shim
|
|
We include "compat/zlib.h" in "git-compat-util.h", which is
unnecessarily broad given that we only have a small handful of files
that use the zlib library. Move the header into "git-zlib.h" instead and
adapt users of zlib to include that header.
One exception is the reftable library, as we don't want to use the
Git-specific wrapper of zlib there, so we include "compat/zlib.h"
instead. Furthermore, we move the include into "reftable/system.h" so
that users of the library other than Git can wire up zlib themselves.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Introduce a new "compat/zlib-compat.h" header that we include instead of
including <zlib.h> directly. This will allow us to wire up zlib-ng as an
alternative backend for zlib compression in a subsequent commit.
Note that we cannot just call the file "compat/zlib.h", as that may
otherwise cause us to include that file instead of <zlib.h>.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The reftable/ library code has been made -Wsign-compare clean.
* ps/reftable-sign-compare:
reftable: address trivial -Wsign-compare warnings
reftable/blocksource: adjust `read_block()` to return `ssize_t`
reftable/blocksource: adjust type of the block length
reftable/block: adjust type of the restart length
reftable/block: adapt header and footer size to return a `size_t`
reftable/basics: adjust `hash_size()` to return `uint32_t`
reftable/basics: adjust `common_prefix_size()` to return `size_t`
reftable/record: handle overflows when decoding varints
reftable/record: drop unused `print` function pointer
meson: stop disabling -Wsign-compare
|
|
The function `reftable_writer_set_limits()` allows updating the
'min_update_index' and 'max_update_index' of a reftable writer. These
values are written to both the writer's header and footer.
Since the header is written during the first block write, any subsequent
changes to the update index would create a mismatch between the header
and footer values. The footer would contain the newer values while the
header retained the original ones.
To protect against this bug, prevent callers from updating these values
after any record is written. To do this, modify the function to return
an error whenever the limits are modified after any record adds. Check
for record adds within `reftable_writer_set_limits()` by checking the
`last_key` and `next` variable. The former is updated after each record
added, but is reset at certain points. The latter is set after writing
the first block.
Modify all callers of the function to anticipate a return type and
handle it accordingly. Add a unit test to also ensure the function
returns the error as expected.
Helped-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Address the last couple of trivial -Wsign-compare warnings in the
reftable library and remove the DISABLE_SIGN_COMPARE_WARNINGS macro that
we have in "reftable/system.h".
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The `block_source_read_block()` function and its implementations return
an integer as a result that reflects either the number of bytes read, or
an error. As such its return type, a signed integer, isn't wrong, but it
doesn't give the reader a good hint what it actually returns.
Refactor the function to return an `ssize_t` instead, which is typical
for functions similar to read(3p) and should thus give readers a better
signal what they can expect as a result.
Adjust callers to better handle the returned value to avoid warnings
with -Wsign-compare. One of these callers is `reader_get_block()`, whose
return value is only ever used by its callers to figure out whether or
not the read was successful. So instead of bubbling up the `ssize_t`
there, too, we adapt it to only indicate success or errors.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The block length is used to track the number of bytes available in a
specific block. As such, it is never set to a negative value, but is
still represented by a signed integer.
Adjust the type of the variable to be `size_t`.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The restart length is tracked as a positive integer even though it
cannot ever be negative. Furthermore, it is effectively capped via the
MAX_RESTARTS variable.
Adjust the type of the variable to be `uint32_t`. While this type is
excessive given that MAX_RESTARTS fits into an `uint16_t`, other places
already use 32 bit integers for restarts, so this type is being more
consistent.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The functions `header_size()` and `footer_size()` return a positive
integer representing the size of the header and footer, respectively,
dependent on the version of the reftable format. Similar to the
preceding commit, these functions return a signed integer though, which
is nonsensical given that there is no way for these functions to return
negative.
Adapt the functions to return a `size_t` instead to fix a couple of sign
comparison warnings.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The `hash_size()` function returns the number of bytes used by the hash
function. Weirdly enough though, it returns a signed integer for its
size even though the size obviously cannot ever be negative. The only
case where it could be negative is if the function returned an error
when asked for an unknown hash, but we assert(3p) instead.
Adjust the type of `hash_size()` to be `uint32_t` and adapt all places
that use signed integers for the hash size to follow suit. This also
allows us to get rid of a couple asserts that we had which verified that
the size was indeed positive, which further stresses the point that this
refactoring makes sense.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The `common_prefix_size()` function computes the length of the common
prefix between two buffers. As such its return value will always be an
unsigned integer, as the length cannot be negative. Regardless of that,
the function returns a signed integer, which is nonsensical and causes a
couple of -Wsign-compare warnings all over the place.
Adjust the function to return a `size_t` instead.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The logic to decode varints isn't able to detect integer overflows: as
long as the buffer still has more data available, and as long as the
current byte has its 0x80 bit set, we'll continue to add up these values
to the result. This will eventually cause the `uint64_t` to overflow, at
which point we'll return an invalid result.
Refactor the function so that it is able to detect such overflows. The
implementation is basically copied from Git's own `decode_varint()`,
which already knows to handle overflows. The only adjustment is that we
also take into account the string view's length in order to not overrun
it. The reftable documentation explicitly notes that those two encoding
schemas are supposed to be the same:
Varint encoding
^^^^^^^^^^^^^^^
Varint encoding is identical to the ofs-delta encoding method used
within pack files.
Decoder works as follows:
....
val = buf[ptr] & 0x7f
while (buf[ptr] & 0x80) {
ptr++
val = ((val + 1) << 7) | (buf[ptr] & 0x7f)
}
....
While at it, refactor `put_var_int()` in the same way by copying over
the implementation of `encode_varint()`. While `put_var_int()` doesn't
have an issue with overflows, it generates warnings with -Wsign-compare.
The implementation of `encode_varint()` doesn't, is battle-tested and at
the same time way simpler than what we currently have.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
In 42c424d69d (t/helper: inline printing of reftable records,
2024-08-22) we stopped using the `print` function of the reftable record
vtable and instead moved its implementation into the single user of it.
We didn't remove the function itself from the vtable though. Drop it.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The reftable library uses randomness in two call paths:
- When reading a stack in case some of the referenced tables
disappears. The randomness is used to delay the next read by a
couple of milliseconds.
- When writing a new table, where the randomness gets appended to the
table name (e.g. "0x000000000001-0x000000000002-0b1d8ddf.ref").
In neither of these cases do we need strong randomness.
Unfortunately though, we have observed test failures caused by the
former case. In t0610 we have a test that spawns a 100 processes at
once, all of which try to write a new table to the stack. And given that
all of the processes will require randomness, it can happen that these
processes make the entropy pool run dry, which will then cause us to
die:
+ test_seq 100
+ printf %s commit\trefs/heads/branch-%s\n
68d032e9edd3481ac96382786ececc37ec28709e 1
+ printf %s commit\trefs/heads/branch-%s\n
68d032e9edd3481ac96382786ececc37ec28709e 2
...
+ git update-ref refs/heads/branch-98 HEAD
+ git update-ref refs/heads/branch-97 HEAD
+ git update-ref refs/heads/branch-99 HEAD
+ git update-ref refs/heads/branch-100 HEAD
fatal: unable to get random bytes
fatal: unable to get random bytes
fatal: unable to get random bytes
fatal: unable to get random bytes
fatal: unable to get random bytes
fatal: unable to get random bytes
fatal: unable to get random bytes
The report was for NonStop, which uses OpenSSL as the backend for
randomness. In the preceding commit we have adapted that backend to also
return randomness in case the entropy pool is empty and the caller
passes the `CSPRNG_BYTES_INSECURE` flag. Do so to fix the issue.
Reported-by: Randall S. Becker <rsbecker@nexbridge.com>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The `csprng_bytes()` function generates randomness and writes it into a
caller-provided buffer. It abstracts over a couple of implementations,
where the exact one that is used depends on the platform.
These implementations have different guarantees: while some guarantee to
never fail (arc4random(3)), others may fail. There are two significant
failures to distinguish from one another:
- Systemic failure, where e.g. opening "/dev/urandom" fails or when
OpenSSL doesn't have a provider configured.
- Entropy failure, where the entropy pool is exhausted, and thus the
function cannot guarantee strong cryptographic randomness.
While we cannot do anything about the former, the latter failure can be
acceptable in some situations where we don't care whether or not the
randomness can be predicted.
Introduce a new `CSPRNG_BYTES_INSECURE` flag that allows callers to opt
into weak cryptographic randomness. The exact behaviour of the flag
depends on the underlying implementation:
- `arc4random_buf()` never returns an error, so it doesn't change.
- `getrandom()` pulls from "/dev/urandom" by default, which never
blocks on modern systems even when the entropy pool is empty.
- `getentropy()` seems to block when there is not enough randomness
available, and there is no way of changing that behaviour.
- `GtlGenRandom()` doesn't mention anything about its specific
failure mode.
- The fallback reads from "/dev/urandom", which also returns bytes in
case the entropy pool is drained in modern Linux systems.
That only leaves OpenSSL with `RAND_bytes()`, which returns an error in
case the returned data wouldn't be cryptographically safe. This function
is replaced with a call to `RAND_pseudo_bytes()`, which can indicate
whether or not the returned data is cryptographically secure via its
return value. If it is insecure, and if the `CSPRNG_BYTES_INSECURE` flag
is set, then we ignore the insecurity and return the data regardless.
It is somewhat questionable whether we really need the flag in the first
place, or whether we wouldn't just ignore the potentially-insecure data.
But the risk of doing that is that we might have or grow callsites that
aren't aware of the potential insecureness of the data in places where
it really matters. So using a flag to opt-in to that behaviour feels
like the more secure choice.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Check the final reallocation for adding the terminating NULL and handle
it just like those in the loop. Simply use REFTABLE_ALLOC_GROW instead
of keeping the REFTABLE_REALLOC_ARRAY call and adding code to preserve
the original pointer value around it.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When realloc(3) fails, it returns NULL and keeps the original allocation
intact. REFTABLE_ALLOC_GROW overwrites both the original pointer and
the allocation count variable in that case, simultaneously leaking the
original allocation and misrepresenting the number of storable items.
parse_names() avoids the leak by keeping the original pointer if
reallocation fails, but still increase the allocation count in such a
case as if it succeeded. That's OK, because the error handling code
just frees everything and doesn't look at names_cap anymore.
reftable_buf_add() does the same, but here it is a problem as it leaves
the reftable_buf in a broken state, with ->alloc being roughly twice as
big as the actually allocated memory, allowing out-of-bounds writes in
subsequent calls.
Reimplement REFTABLE_ALLOC_GROW to avoid leaks, keep allocation counts
in sync and still signal failures to callers while avoiding code
duplication in callers. Make it an expression that evaluates to 0 if no
reallocation is needed or it succeeded and 1 on failure while keeping
the original pointer and allocation counter values.
Adjust REFTABLE_ALLOC_GROW_OR_NULL to the new calling convention for
REFTABLE_ALLOC_GROW, but keep its support for non-size_t alloc variables
for now.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
When realloc(3) fails, it returns NULL and keeps the original allocation
intact. REFTABLE_ALLOC_GROW overwrites both the original pointer and
the allocation count variable in that case, simultaneously leaking the
original allocation and misrepresenting the number of storable items.
parse_names() and reftable_buf_add() avoid leaking by restoring the
original pointer value on failure, but all other callers seem to be OK
with losing the old allocation. Add a new variant of the macro,
REFTABLE_ALLOC_GROW_OR_NULL, which plugs the leak and zeros the
allocation counter. Use it for those callers.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Start working to make the codebase buildable with -Wsign-compare.
* ps/build-sign-compare:
t/helper: don't depend on implicit wraparound
scalar: address -Wsign-compare warnings
builtin/patch-id: fix type of `get_one_patchid()`
builtin/blame: fix type of `length` variable when emitting object ID
gpg-interface: address -Wsign-comparison warnings
daemon: fix type of `max_connections`
daemon: fix loops that have mismatching integer types
global: trivial conversions to fix `-Wsign-compare` warnings
pkt-line: fix -Wsign-compare warning on 32 bit platform
csum-file: fix -Wsign-compare warning on 32-bit platform
diff.h: fix index used to loop through unsigned integer
config.mak.dev: drop `-Wno-sign-compare`
global: mark code units that generate warnings with `-Wsign-compare`
compat/win32: fix -Wsign-compare warning in "wWinMain()"
compat/regex: explicitly ignore "-Wsign-compare" warnings
git-compat-util: introduce macros to disable "-Wsign-compare" warnings
|
|
Reftable backend adds check for upper limit of log's update_index.
* kn/reftable-writer-log-write-verify:
reftable/writer: ensure valid range for log's update_index
|
|
Recent reftable updates mistook a NULL return from a request for
0-byte allocation as OOM and died unnecessarily, which has been
corrected.
* ps/reftable-alloc-failures-zalloc-fix:
reftable/basics: return NULL on zero-sized allocations
reftable/stack: fix zero-sized allocation when there are no readers
reftable/merged: fix zero-sized allocation when there are no readers
reftable/stack: don't perform auto-compaction with less than two tables
|
|
In the preceding commits we have fixed a couple of issues when
allocating zero-sized objects. These issues were masked by
implementation-defined behaviour. Quoting malloc(3p):
If size is 0, either:
* A null pointer shall be returned and errno may be set to an
implementation-defined value, or
* A pointer to the allocated space shall be returned. The
application shall ensure that the pointer is not used to access an
object.
So it is perfectly valid that implementations of this function may or
may not return a NULL pointer in such a case.
Adapt both `reftable_malloc()` and `reftable_realloc()` so that they
return NULL pointers on zero-sized allocations. This should remove any
implementation-defined behaviour in our allocators and thus allows us to
detect such platform-specific issues more easily going forward.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Similar as the preceding commit, we may try to do a zero-sized
allocation when reloading a reftable stack that ain't got any tables.
It is implementation-defined whether malloc(3p) returns a NULL pointer
in that case or a zero-sized object. In case it does return a NULL
pointer though it causes us to think we have run into an out-of-memory
situation, and thus we return an error.
Fix this by only allocating arrays when they have at least one entry.
Reported-by: Randall S. Becker <rsbecker@nexbridge.com>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
It was reported [1] that Git started to fail with an out-of-memory error
when initializing repositories with the reftable backend on NonStop
platforms. A bisect led to 802c0646ac (reftable/merged: handle
allocation failures in `merged_table_init_iter()`, 2024-10-02), which
changed how we allocate memory when initializing a merged table.
The root cause of this seems to be that NonStop returns a `NULL` pointer
when doing a zero-sized allocation. This would've already happened
before the above change, but we never noticed because we did not check
the result. Now we do notice and thus return an out-of-memory error to
the caller.
Fix the issue by skipping the allocation altogether in case there are no
readers.
[1]: <00ad01db5017$aa9ce340$ffd6a9c0$@nexbridge.com>
Reported-by: Randall S. Becker <rsbecker@nexbridge.com>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
In order to compact tables we need at least two tables. Bail out early
from `reftable_stack_auto_compact()` in case we have less than two
tables.
In the original, `stack_table_sizes_for_compaction()` yields an array
that has the same length as the number of tables. This array is then
passed on to `suggest_compaction_segment()`, which returns an empty
segment in case we have less than two tables. The segment is then passed
to `segment_size()`, which will return `0` because both start and end of
the segment are `0`. And because we only call `stack_compact_range()` in
case we have a positive segment size we don't perform auto-compaction at
all. Consequently, this change does not result in a user-visible change
in behaviour when called with a single table.
But when called with no tables this protects us against a potential
out-of-memory error: `stack_table_sizes_for_compaction()` would try to
allocate a zero-byte object when there aren't any tables, and that may
lead to a `NULL` pointer on some platforms like NonStop which causes us
to bail out with an out-of-memory error.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Optimize reading random references out of the reftable backend by
allowing reuse of iterator objects.
* ps/reftable-iterator-reuse:
refs/reftable: reuse iterators when reading refs
reftable/merged: drain priority queue on reseek
reftable/stack: add mechanism to notify callers on reload
refs/reftable: refactor reflog expiry to use reftable backend
refs/reftable: refactor reading symbolic refs to use reftable backend
refs/reftable: read references via `struct reftable_backend`
refs/reftable: figure out hash via `reftable_stack`
reftable/stack: add accessor for the hash ID
refs/reftable: handle reloading stacks in the reftable backend
refs/reftable: encapsulate reftable stack
|
|
Isolates the reftable subsystem from the rest of Git's codebase by
using fewer pieces of Git's infrastructure.
* ps/reftable-detach:
reftable/system: provide thin wrapper for lockfile subsystem
reftable/stack: drop only use of `get_locked_file_path()`
reftable/system: provide thin wrapper for tempfile subsystem
reftable/stack: stop using `fsync_component()` directly
reftable/system: stop depending on "hash.h"
reftable: explicitly handle hash format IDs
reftable/system: move "dir.h" to its only user
|
|
Each reftable addition has an associated update_index. While writing
refs, the update_index is verified to be within the range of the
reftable writer, i.e. `writer.min_update_index <= ref.update_index` and
`writer.max_update_index => ref.update_index`.
The corresponding check for reflogs in `reftable_writer_add_log` is
however missing. Add a similar check, but only check for the upper
limit. This is because reflogs are treated a bit differently than refs.
Each reflog entry in reftable has an associated update_index and we also
allow expiring entries in the middle, which is done by simply writing a
new reflog entry with the same update_index. This means, writing reflog
entries with update_index lesser than the writer's update_index is an
expected scenario.
Add a new unit test to check for the limits and fix some of the existing
tests, which were setting arbitrary values for the update_index by
ensuring they stay within the now checked limits.
Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Mark code units that generate warnings with `-Wsign-compare`. This
allows for a structured approach to get rid of all such warnings over
time in a way that can be easily measured.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
In 5bf96e0c39 (reftable/generic: move seeking of records into the
iterator, 2024-05-13) we have refactored the reftable codebase such that
iterators can be initialized once and then re-seeked multiple times.
This feature is used by 1869525066 (refs/reftable: wire up support for
exclude patterns, 2024-09-16) in order to skip records based on exclude
patterns provided by the caller.
The logic to re-seek the merged iterator is insufficient though because
we don't drain the priority queue on a re-seek. This means that the
queue may contain stale entries and thus reading the next record in the
queue will return the wrong entry. While this is an obvious bug, it is
harmless in the context of above exclude patterns:
- If the queue contained stale entries that match the pattern then the
caller would already know to filter out such refs. This is because
our codebase is prepared to handle backends that don't have a way to
efficiently implement exclude patterns.
- If the queue contained stale entries that don't match the pattern
we'd eventually filter out any duplicates. This is because the
reftable code discards items with the same ref name and sorts any
remaining entries properly.
So things happen to work in this context regardless of the bug, and
there is no other use case yet where we re-seek iterators. We're about
to introduce a caching mechanism though where iterators are reused by
the reftable backend, and that will expose the bug.
Fix the issue by draining the priority queue when seeking and add a
testcase that surfaces the issue.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Reftable stacks are reloaded in two cases:
- When calling `reftable_stack_reload()`, if the stat-cache tells us
that the stack has been modified.
- When committing a reftable addition.
While callers can figure out the second case, they do not have a
mechanism to figure out whether `reftable_stack_reload()` led to an
actual reload of the on-disk data. All they can do is thus to assume
that data is always being reloaded in that case.
Improve the situation by introducing a new `on_reload()` callback to the
reftable options. If provided, the function will be invoked every time
the stack has indeed been reloaded. This allows callers to invalidate
data that depends on the current stack data.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Add an accessor function that allows callers to access the hash ID of a
reftable stack. This function will be used in a subsequent commit.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Both `struct block_writer` and `struct reftable_writer` have a `buf`
member that is being reused to optimize the number of allocations.
Rename the variable to `scratch` to clarify its intend and provide a
comment explaining why it exists.
Suggested-by: Christian Couder <christian.couder@gmail.com>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The block writer needs to compute the key for every record that one adds
to the writer. The buffer for this key is stored on the stack and thus
reallocated on every call to `block_writer_add()`, which is inefficient.
Refactor the code so that we store the buffer in the `block_writer`
struct itself so that we can reuse it. This reduces the number of
allocations when writing many refs, e.g. when migrating one million refs
from the "files" backend to the "reftable backend. Before this change:
HEAP SUMMARY:
in use at exit: 80,048 bytes in 49 blocks
total heap usage: 3,025,864 allocs, 3,025,815 frees, 372,746,291 bytes allocated
After this change:
HEAP SUMMARY:
in use at exit: 80,048 bytes in 49 blocks
total heap usage: 2,013,250 allocs, 2,013,201 frees, 347,543,583 bytes allocated
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Adapt the name of the `block_writer::buf` variable to instead be called
`block`. This aligns it with the existing `block_len` variable, which
tracks the length of this buffer, and is generally a bit more tied to
the actual context where this variable gets used.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
Both `writer_add_record()` and `reftable_writer_add_ref()` get executed
for every single ref record we're adding to the reftable writer. And as
both functions use a local buffer to write data, the allocations we have
to do here add up during larger transactions.
Refactor the code to use a scratch buffer part of the `reftable_writer`
itself such that we can reuse it. This signifcantly reduces the number
of allocations during large transactions, e.g. when migrating refs from
the "files" backend to the "reftable" backend. Before this change:
HEAP SUMMARY:
in use at exit: 80,048 bytes in 49 blocks
total heap usage: 5,032,171 allocs, 5,032,122 frees, 418,792,092 bytes allocated
After this change:
HEAP SUMMARY:
in use at exit: 80,048 bytes in 49 blocks
total heap usage: 3,025,864 allocs, 3,025,815 frees, 372,746,291 bytes allocated
This also translate into a small speedup:
Benchmark 1: migrate files:reftable (refcount = 1000000, revision = HEAD~)
Time (mean ± σ): 827.2 ms ± 16.5 ms [User: 689.4 ms, System: 124.9 ms]
Range (min … max): 809.0 ms … 924.7 ms 50 runs
Benchmark 2: migrate files:reftable (refcount = 1000000, revision = HEAD)
Time (mean ± σ): 813.6 ms ± 11.6 ms [User: 679.0 ms, System: 123.4 ms]
Range (min … max): 786.7 ms … 833.5 ms 50 runs
Summary
migrate files:reftable (refcount = 1000000, revision = HEAD) ran
1.02 ± 0.02 times faster than migrate files:reftable (refcount = 1000000, revision = HEAD~)
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We use the lockfile subsystem to write lockfiles for "tables.list". As
with the tempfile subsystem, the lockfile subsystem also hooks into our
infrastructure to prune stale locks via atexit(3p) or signal handlers.
Furthermore, the lockfile subsystem also handles locking timeouts, which
do add quite a bit of logic. Having to reimplement that in the context
of Git wouldn't make a whole lot of sense, and it is quite likely that
downstream users of the reftable library may have a better idea for how
exactly to implement timeouts.
So again, provide a thin wrapper for the lockfile subsystem instead such
that the compatibility shim is fully self-contained.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We've got a single callsite where we call `get_locked_file_path()`. As
we're about to convert our usage of the lockfile subsystem to instead be
used via a compatibility shim we'd have to implement more logic for this
single callsite. While that would be okay if Git was the only supposed
user of the reftable library, it's a bit more awkward when considering
that we have to reimplement this functionality for every user of the
library eventually.
Refactor the code such that we don't call `get_locked_file_path()`
anymore.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We use the tempfile subsystem to write temporary tables, but given that
we're in the process of converting the reftable library to become
standalone we cannot use this subsystem directly anymore. While we could
in theory convert the code to use mkstemp(3p) instead, we'd lose access
to our infrastructure that automatically prunes tempfiles via atexit(3p)
or signal handlers.
Provide a thin wrapper for the tempfile subsystem instead. Like this,
the compatibility shim is fully self-contained in "reftable/system.c".
Downstream users of the reftable library would have to implement their
own tempfile shims by replacing "system.c" with a custom version.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We're executing `fsync_component()` directly in the reftable library so
that we can fsync data to disk depending on "core.fsync". But as we're
in the process of converting the reftable library to become standalone
we cannot use that function in the library anymore.
Refactor the code such that users of the library can inject a custom
fsync function via the write options. This allows us to get rid of the
dependency on "write-or-die.h".
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
We include "hash.h" in "reftable/system.h" such that we can use hash
format IDs as well as the raw size of SHA1 and SHA256. As we are in the
process of converting the reftable library to become standalone we of
course cannot rely on those constants anymore.
Introduce a new `enum reftable_hash` to replace internal uses of the
hash format IDs and new constants that replace internal uses of the hash
size. Adapt the reftable backend to set up the correct hash function.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
|
The hash format IDs are used for two different things across the
reftable codebase:
- They are used as a 32 bit unsigned integer when reading and writing
the header in order to identify the hash function.
- They are used internally to identify which hash function is in use.
When one only considers the second usecase one might think that one can
easily change the representation of those hash IDs. But because those
IDs end up in the reftable header and footer on disk it is important
that those never change.
Create separate constants `REFTABLE_FORMAT_ID_*` and use them in
contexts where we read or write reftable headers. This serves multiple
purposes:
- It allows us to more easily discern cases where we actually use
those constants for the on-disk format.
- It detangles us from the same constants that are defined in
libgit.a, which is another required step to convert the reftable
library to become standalone.
- It makes the next step easier where we stop using `GIT_*_FORMAT_ID`
constants in favor of a custom enum.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
|
