Junio noticed that 'non-trivial' pushes were failing if executed
using the sliding window mmap changes. This was somewhat difficult
to track down as the failure was appearing randomly.
It turns out this was a failure caused by the delta base reference
(either ref or offset format) spanning over the end of a mmap window.
The error in pack-objects was we were not recalling use_pack
after the object header was unpacked, and therefore we did not
get the promise of at least 20 bytes in the buffer for the delta
base parsing. This would case later memcmp() calls to walk into
unassigned address space at the end of the window.
The reason Junio and I had hard time tracking this down in current
Git repositories is we were both probably packing with offset deltas,
which minimized the odds of the delta base reference spanning over
the end of the mmap window. Stepping back and repacking with
version 1.3.3 (which only supported reference deltas) increased
the likelyhood of seeing the bug.
The correct technique (as used in sha1_file.c) is to invoke
use_pack() after unpack_object_header_gently to ensure we have
enough data available for the delta base decoding.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
When multiple mmaps start getting used for all pack file access it
is not possible to get all data associated with a specific object
in one contiguous memory region. This limitation prevents simply
passing a single address and length to SHA1_Update or to inflate.
Instead we need to loop until we have processed all data of interest.
As we loop over the data we are always interested in reusing the same
window 'cursor', as the prior window will no longer be of any use
to us. This allows the use_pack() call to automatically decrement
the use count of the prior window before setting up access for us
to the next window.
Within each loop we need to make use of the available length output
parameter of use_pack() to tell us how many bytes are available in
the current memory region, as we cannot tell otherwise.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Part of the implementation concept of the sliding mmap window for
pack access is to permit multiple windows per pack to be mapped
independently. Since the inuse_cnt is associated with the mmap and
not with the file, this value is in struct pack_window and needs to
be incremented/decremented for each pack_window accessed by any code.
To faciliate that implementation we need to replace all uses of
use_packed_git() and unuse_packed_git() with a different API that
follows struct pack_window objects rather than struct packed_git.
The way this works is when we need to start accessing a pack for
the first time we should setup a new window 'cursor' by declaring
a local and setting it to NULL:
struct pack_windows *w_curs = NULL;
To obtain the memory region which contains a specific section of
the pack file we invoke use_pack(), supplying the address of our
current window cursor:
unsigned int len;
unsigned char *addr = use_pack(p, &w_curs, offset, &len);
the returned address `addr` will be the first byte at `offset`
within the pack file. The optional variable len will also be
updated with the number of bytes remaining following the address.
Multiple calls to use_pack() with the same window cursor will
update the window cursor, moving it from one window to another
when necessary. In this way each window cursor variable maintains
only one struct pack_window inuse at a time.
Finally before exiting the scope which originally declared the window
cursor we must invoke unuse_pack() to unuse the current window (which
may be different from the one that was first obtained from use_pack):
unuse_pack(&w_curs);
This implementation is still not complete with regards to multiple
windows, as only one window per pack file is supported right now.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The idea behind the sliding mmap window pack reader implementation
is to have multiple mmap regions active against the same pack file,
thereby allowing the process to mmap in only the active/hot sections
of the pack and reduce overall virtual address space usage.
To implement this we need to refactor the mmap related data
(pack_base, pack_use_cnt) out of struct packed_git and move them
into a new struct pack_window.
We are refactoring the code to support a single struct pack_window
per packfile, thereby emulating the prior behavior of mmap'ing the
entire pack file.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This adds a new option --reflog to pack-objects and revision
machinery; do not bother documenting it for now, since this is
only useful for local repacking.
When the option is passed, objects reachable from reflog entries
are marked as interesting while computing the set of objects to
pack.
Signed-off-by: Junio C Hamano <junkio@cox.net>
This is a mechanical clean-up of the way *.c files include
system header files.
(1) sources under compat/, platform sha-1 implementations, and
xdelta code are exempt from the following rules;
(2) the first #include must be "git-compat-util.h" or one of
our own header file that includes it first (e.g. config.h,
builtin.h, pkt-line.h);
(3) system headers that are included in "git-compat-util.h"
need not be included in individual C source files.
(4) "git-compat-util.h" does not have to include subsystem
specific header files (e.g. expat.h).
Signed-off-by: Junio C Hamano <junkio@cox.net>
The final 'nr_result' and 'written' values must always be the same
otherwise we're in deep trouble. So let's remove a redundent report.
And for paranoia sake let's make sure those two variables are actually
equal after all objects are written (one never knows).
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The heuristics to give up deltification when both the source and the
target are both in the same pack affects negatively when we are
repacking the subset of objects in the existing pack. This caused
any incremental updates to use suboptimal packs. Tweak the heuristics
to avoid this problem.
Signed-off-by: Junio C Hamano <junkio@cox.net>
This adds documentation for --progress and --all-progress, remove a
duplicate --progress handling and make usage string more readable.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Currently git-push displays progress status for the local packing of
objects to send, but nothing once it starts to push it over the
connection. Having progress status in that later case is especially
nice when pushing lots of objects over a slow network link.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This is the missing part to git-pack-objects allowing it to reuse delta
data to/from any of the two delta types. It can reuse delta from any
type, and it outputs base offsets when --allow-delta-base-offset is
provided and the base is also included in the pack. Otherwise it
outputs base sha1 references just like it always did.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This is enabled with --delta-base-offset only, and doesn't work with
pack data reuse yet.
The idea is to allow for the fetch protocol to use an extension flag
to notify the remote end that --delta-base-offset can be used with
git-pack-objects. Eventually git-repack will always provide this flag.
With this, all delta base objects are now pushed before deltas that depend
on them. This is a requirements for OBJ_OFS_DELTA. This is not a
requirement for OBJ_REF_DELTA but always doing so makes the code simpler.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This adds a new object, namely OBJ_OFS_DELTA, renames OBJ_DELTA to
OBJ_REF_DELTA to better make the distinction between those two delta
objects, and adds support for the handling of those new delta objects
in sha1_file.c only.
The OBJ_OFS_DELTA contains a relative offset from the delta object's
position in a pack instead of the 20-byte SHA1 reference to identify
the base object. Since the base is likely to be not so far away, the
relative offset is more likely to have a smaller encoding on average
than an absolute offset. And for those delta objects the base must
always be stored first because there is no way to know the distance of
later objects when streaming a pack. Hence this relative offset is
always meant to be negative.
The offset encoding is slightly denser than the one used for object
size -- credits to <linux@horizon.com> (whoever this is) for bringing
it to my attention.
This allows for pack size reduction between 3.2% (Linux-2.6) to over 5%
(linux-historic). Runtime pack access should be faster too since delta
replay does skip a search in the pack index for each delta in a chain.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Those cleanups are mainly to set the table for the support of deltas
with base objects referenced by offsets instead of sha1. This means
that many pack lookup functions are converted to take a pack/offset
tuple instead of a sha1.
This eliminates many struct pack_entry usages since this structure
carried redundent information in many cases, and it increased stack
footprint needlessly for a couple recursively called functions that used
to declare a local copy of it for every recursion loop.
In the process, packed_object_info_detail() has been reorganized as well
so to look much saner and more amenable to deltas with offset support.
Finally the appropriate adjustments have been made to functions that
depend on the above changes. But there is no functionality changes yet
simply some code refactoring at this point.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This teaches the internal rev-list logic to understand options
that are needed for pack handling: --all, --unpacked, and --thin.
It also moves two functions from builtin-rev-list to list-objects
so that the two programs can share more code.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Instead of piping the rev-list output from its standard input,
you can say:
pack-objects --all --unpacked --revs pack
and feed the rev parameters you would otherwise give the
rev-list on its command line from the standard input.
In other words:
echo 'master..next' | pack-objects --revs pack
and
rev-list --objects master..next | pack-objects pack
are equivalent.
Signed-off-by: Junio C Hamano <junkio@cox.net>
When copying from an existing pack and when copying from a loose
object with new style header, the code makes sure that the piece
we are going to copy out inflates well and inflate() consumes
the data in full while doing so.
The check to see if the xdelta really apply is quite expensive
as you described, because you would need to have the image of
the base object which can be represented as a delta against
something else.
Signed-off-by: Junio C Hamano <junkio@cox.net>
When revalidating an entry from an existing pack entry->size and
entry->type are not necessarily the size of the final object
when the entry is deltified, but for base objects they must
match.
Signed-off-by: Junio C Hamano <junkio@cox.net>
When reusing data from an existing pack and from a new style
loose objects, we used to just copy it staight into the
resulting pack. Instead make sure they are not corrupt, but
do so only when we are not streaming to stdout, in which case
the receiving end will do the validation either by unpacking
the stream or by constructing the .idx file.
Signed-off-by: Junio C Hamano <junkio@cox.net>
This abstracts away the size of the hash values when copying them
from memory location to memory location, much as the introduction
of hashcmp abstracted away hash value comparsion.
A few call sites were using char* rather than unsigned char* so
I added the cast rather than open hashcpy to be void*. This is a
reasonable tradeoff as most call sites already use unsigned char*
and the existing hashcmp is also declared to be unsigned char*.
[jc: Splitted the patch to "master" part, to be followed by a
patch for merge-recursive.c which is not in "master" yet.
Fixed the cast in the latter hunk to combine-diff.c which was
wrong in the original.
Also converted ones left-over in combine-diff.c, diff-lib.c and
upload-pack.c ]
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Introduces global inline:
hashcmp(const unsigned char *sha1, const unsigned char *sha2)
Uses memcmp for comparison and returns the result based on the length of
the hash name (a future runtime decision).
Acked-by: Alex Riesen <raa.lkml@gmail.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
[jc: I needed to hand merge the changes to the updated codebase,
so the result needs to be checked.]
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
When packing an object without deltifying, if the data is stored in
a loose object that is encoded with a new style header, copy it without
inflating and deflating.
Signed-off-by: Junio C Hamano <junkio@cox.net>
For some repositories, deltas simply don't make sense. One can disable
them for git-repack by adding --window, but git-push insists on making
the deltas which can be very CPU-intensive for little benefit.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The only visible change is that git-blame doesn't understand
"--compability" anymore, but it does accept "--compatibility" instead,
which is already documented.
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
If no delta is attempted on some objects then it is useless to load them
in memory, neither create any delta index for them. The best thing to
do is therefore to load and index them only when really needed.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Without this there would never be a chance to improve packing for
previously undeltified objects.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
In the repacking window, if both objects we are looking at already came
from the same (old) pack-file, don't bother delta'ing them against each
other.
That means that we'll still always check for better deltas for (and
against!) _unpacked_ objects, but assuming incremental repacks, you'll
avoid the delta creation 99% of the time.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This does not implement sideband for propagating the status to
the downloader yet, but add code to capture the standard error
output from the pack-objects process in preparation for sending
it off to the client when the protocol extension allows us to do
so.
Signed-off-by: Junio C Hamano <junkio@cox.net>
ANSI C99 doesn't allow void-pointer arithmetic. This patch fixes this in
various ways. Usually the strategy that required the least changes was used.
Signed-off-by: Florian Forster <octo@verplant.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This trivial patch not only simplifies the name hashing, it actually
improves packing for both git and the kernel.
The git archive pack shrinks from 6824090->6622627 bytes (a 3%
improvement), and the kernel pack shrinks from 108756213 to 108219021 (a
mere 0.5% improvement, but still, it's an improvement from making the
hashing much simpler!)
We just create a 32-bit hash, where we "age" previous characters by two
bits, so the last characters in a filename count most. So when we then
compare the hashes in the sort routine, filenames that end the same way
sort the same way.
It takes the subdirectory into account (unless the filename is > 16
characters), but files with the same name within the same subdirectory
will obviously sort closer than files in different subdirectories.
And, incidentally (which is why I tried the hash change in the first
place, of course) builtin-rev-list.c will sort fairly close to rev-list.c.
And no, it's not a "good hash" in the sense of being secure or unique, but
that's not what we're looking for. The whole "hash" thing is misnamed
here. It's not so much a hash as a "sorting number".
[jc: rolled in simplification for computing the sorting number
computation for thin pack base objects]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This adds a "tree_entry()" function that combines the common operation of
doing a "tree_entry_extract()" + "update_tree_entry()".
It also has a simplified calling convention, designed for simple loops
that traverse over a whole tree: the arguments are pointers to the tree
descriptor and a name_entry structure to fill in, and it returns a boolean
"true" if there was an entry left to be gotten in the tree.
This allows tree traversal with
struct tree_desc desc;
struct name_entry entry;
desc.buf = tree->buffer;
desc.size = tree->size;
while (tree_entry(&desc, &entry) {
... use "entry.{path, sha1, mode, pathlen}" ...
}
which is not only shorter than writing it out in full, it's hopefully less
error prone too.
[ It's actually a tad faster too - we don't need to recalculate the entry
pathlength in both extract and update, but need to do it only once.
Also, some callers can avoid doing a "strlen()" on the result, since
it's returned as part of the name_entry structure.
However, by now we're talking just 1% speedup on "git-rev-list --objects
--all", and we're definitely at the point where tree walking is no
longer the issue any more. ]
NOTE! Not everybody wants to use this new helper function, since some of
the tree walkers very much on purpose do the descriptor update separately
from the entry extraction. So the "extract + update" sequence still
remains as the core sequence, this is just a simplified interface.
We should probably add a silly two-line inline helper function for
initializing the descriptor from the "struct tree" too, just to cut down
on the noise from that common "desc" initializer.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This provides a linear decrement on the penalty related to delta depth
instead of being an 1/x function. With this another 5% reduction is
observed on packs for both the GIT repo and the Linux kernel repo, as
well as fixing a pack size regression in another sample repo I have.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Apparently <stdint.h> is not enough for uint32_t on OpenBSD; use
"unsigned int" -- hopefully that would stay 32-bit on every
platform we care about, at least until we update the pack-index
file format.
Our sha1 routines optimized for architectures use uint32_t and
expects '#include <stdint.h>' to be enough, so OpenBSD on arm or
ppc might have similar issues down the road, I dunno.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Avoid creating a delta index for objects with maximum depth since they
are not going to be used as delta base anyway. This also reduce peak
memory usage slightly as the current object's delta index is not useful
until the next object in the loop is considered for deltification. This
saves a bit more than 1% on CPU usage.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Given that the early eviction of objects with maximum delta depth
may exhibit bad packing on its own, why not considering a bias against
deep base objects in try_delta() to mitigate that bad behavior.
This patch adjust the MAX_size allowed for a delta based on the depth of
the base object as well as enabling the early eviction of max depth
objects from the object window. When used separately, those two things
produce slightly better and much worse results respectively. But their
combined effect is a surprising significant packing improvement.
With this really simple patch the GIT repo gets nearly 15% smaller, and
the Linux kernel repo about 5% smaller, with no significantly measurable
CPU usage difference.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The offset of an object in the pack is recorded as a 4-byte integer
in the index file. When reading the offset from the mmap'ed index
in prepare_pack_revindex(), the address is dereferenced as a long*.
This works fine as long as the long type is four bytes wide. On
NetBSD/sparc64, however, a long is 8 bytes wide and so dereferencing
the offset produces garbage.
[jc: taking suggestion by Linus to use uint32_t]
Signed-off-by: Dennis Stosberg <dennis@stosberg.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
One of my post-update scripts runs a git-fetch into a separate
repository and sends the results back to me (2>&1); I end up
getting this in the mail:
Generating pack...
Done counting 180 objects.
Result has 131 objects.
Deltifying 131 objects.
0% (0/131) done^M 1% (2/131) done^M...
This defaults not to do the progress report when not on a tty.
You could give --progress to force the progress report, but
let's not bother even documenting it nor mentioning it in the
usage string.
Signed-off-by: Junio C Hamano <junkio@cox.net>
We used to omit delta base candidates that is much bigger than
the target, but delta size does not grow when we delete more, so
that was not a very good heuristics.
Signed-off-by: Junio C Hamano <junkio@cox.net>
This patch allows for computing the delta index for each base object
only once and reuse it when trying to find the best delta match.
This should set the mark and pave the way for possibly better delta
generator algorithms.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The input line has 40 _chars_ of sha1 and no 20 _bytes_. It should also
account for the space before the pathname, and the terminating \n and \0.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Because we sort the delta window by name-hash and then size,
hitting an object that is too small to consider as a delta base
for the current object does not mean we do not have better
candidate in the window beyond it.
Noticed by Shawn Pearce, analyzed by Nico, Linus and me.
Signed-off-by: Junio C Hamano <junkio@cox.net>
It appears the fingerprinting itself is too expensive to be worth doing
for this purpose. A failed experiment.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Jens Axboe noticed that recent "git push" has become very slow
since we made --thin transfer the default.
Thin pack generation to push a handful revisions that touch
relatively small number of paths out of huge tree was stupid; it
registered _everything_ from the excluded revisions. As a
result, "Counting objects" phase was unnecessarily expensive.
This changes the logic to register the blobs and trees from
excluded revisions only for paths we are actually going to send
to the other end.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Shawn O. Pearce