2014-06-13 16:19:36 +04:00
|
|
|
#include "cache.h"
|
|
|
|
#include "split-index.h"
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
#include "ewah/ewok.h"
|
2014-06-13 16:19:36 +04:00
|
|
|
|
|
|
|
struct split_index *init_split_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
if (!istate->split_index) {
|
|
|
|
istate->split_index = xcalloc(1, sizeof(*istate->split_index));
|
|
|
|
istate->split_index->refcount = 1;
|
|
|
|
}
|
|
|
|
return istate->split_index;
|
|
|
|
}
|
|
|
|
|
|
|
|
int read_link_extension(struct index_state *istate,
|
|
|
|
const void *data_, unsigned long sz)
|
|
|
|
{
|
|
|
|
const unsigned char *data = data_;
|
|
|
|
struct split_index *si;
|
2014-06-13 16:19:41 +04:00
|
|
|
int ret;
|
|
|
|
|
2018-05-02 03:25:43 +03:00
|
|
|
if (sz < the_hash_algo->rawsz)
|
2014-06-13 16:19:36 +04:00
|
|
|
return error("corrupt link extension (too short)");
|
|
|
|
si = init_split_index(istate);
|
2018-05-02 03:25:43 +03:00
|
|
|
hashcpy(si->base_oid.hash, data);
|
|
|
|
data += the_hash_algo->rawsz;
|
|
|
|
sz -= the_hash_algo->rawsz;
|
2014-06-13 16:19:41 +04:00
|
|
|
if (!sz)
|
|
|
|
return 0;
|
|
|
|
si->delete_bitmap = ewah_new();
|
|
|
|
ret = ewah_read_mmap(si->delete_bitmap, data, sz);
|
|
|
|
if (ret < 0)
|
|
|
|
return error("corrupt delete bitmap in link extension");
|
|
|
|
data += ret;
|
|
|
|
sz -= ret;
|
|
|
|
si->replace_bitmap = ewah_new();
|
|
|
|
ret = ewah_read_mmap(si->replace_bitmap, data, sz);
|
|
|
|
if (ret < 0)
|
|
|
|
return error("corrupt replace bitmap in link extension");
|
|
|
|
if (ret != sz)
|
2014-06-13 16:19:36 +04:00
|
|
|
return error("garbage at the end of link extension");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int write_link_extension(struct strbuf *sb,
|
|
|
|
struct index_state *istate)
|
|
|
|
{
|
|
|
|
struct split_index *si = istate->split_index;
|
2018-05-02 03:25:43 +03:00
|
|
|
strbuf_add(sb, si->base_oid.hash, the_hash_algo->rawsz);
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
if (!si->delete_bitmap && !si->replace_bitmap)
|
|
|
|
return 0;
|
2015-03-08 13:12:32 +03:00
|
|
|
ewah_serialize_strbuf(si->delete_bitmap, sb);
|
|
|
|
ewah_serialize_strbuf(si->replace_bitmap, sb);
|
2014-06-13 16:19:36 +04:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mark_base_index_entries(struct index_state *base)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
/*
|
|
|
|
* To keep track of the shared entries between
|
|
|
|
* istate->base->cache[] and istate->cache[], base entry
|
|
|
|
* position is stored in each base entry. All positions start
|
2016-05-06 15:36:46 +03:00
|
|
|
* from 1 instead of 0, which is reserved to say "this is a new
|
2014-06-13 16:19:36 +04:00
|
|
|
* entry".
|
|
|
|
*/
|
|
|
|
for (i = 0; i < base->cache_nr; i++)
|
|
|
|
base->cache[i]->index = i + 1;
|
|
|
|
}
|
|
|
|
|
2014-06-13 16:19:44 +04:00
|
|
|
void move_cache_to_base_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
struct split_index *si = istate->split_index;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
/*
|
block alloc: allocate cache entries from mem_pool
When reading large indexes from disk, a portion of the time is
dominated in malloc() calls. This can be mitigated by allocating a
large block of memory and manage it ourselves via memory pools.
This change moves the cache entry allocation to be on top of memory
pools.
Design:
The index_state struct will gain a notion of an associated memory_pool
from which cache_entries will be allocated from. When reading in the
index from disk, we have information on the number of entries and
their size, which can guide us in deciding how large our initial
memory allocation should be. When an index is discarded, the
associated memory_pool will be discarded as well - so the lifetime of
a cache_entry is tied to the lifetime of the index_state that it was
allocated for.
In the case of a Split Index, the following rules are followed. 1st,
some terminology is defined:
Terminology:
- 'the_index': represents the logical view of the index
- 'split_index': represents the "base" cache entries. Read from the
split index file.
'the_index' can reference a single split_index, as well as
cache_entries from the split_index. `the_index` will be discarded
before the `split_index` is. This means that when we are allocating
cache_entries in the presence of a split index, we need to allocate
the entries from the `split_index`'s memory pool. This allows us to
follow the pattern that `the_index` can reference cache_entries from
the `split_index`, and that the cache_entries will not be freed while
they are still being referenced.
Managing transient cache_entry structs:
Cache entries are usually allocated for an index, but this is not always
the case. Cache entries are sometimes allocated because this is the
type that the existing checkout_entry function works with. Because of
this, the existing code needs to handle cache entries associated with an
index / memory pool, and those that only exist transiently. Several
strategies were contemplated around how to handle this:
Chosen approach:
An extra field was added to the cache_entry type to track whether the
cache_entry was allocated from a memory pool or not. This is currently
an int field, as there are no more available bits in the existing
ce_flags bit field. If / when more bits are needed, this new field can
be turned into a proper bit field.
Alternatives:
1) Do not include any information about how the cache_entry was
allocated. Calling code would be responsible for tracking whether the
cache_entry needed to be freed or not.
Pro: No extra memory overhead to track this state
Con: Extra complexity in callers to handle this correctly.
The extra complexity and burden to not regress this behavior in the
future was more than we wanted.
2) cache_entry would gain knowledge about which mem_pool allocated it
Pro: Could (potentially) do extra logic to know when a mem_pool no
longer had references to any cache_entry
Con: cache_entry would grow heavier by a pointer, instead of int
We didn't see a tangible benefit to this approach
3) Do not add any extra information to a cache_entry, but when freeing a
cache entry, check if the memory exists in a region managed by existing
mem_pools.
Pro: No extra memory overhead to track state
Con: Extra computation is performed when freeing cache entries
We decided tracking and iterating over known memory pool regions was
less desirable than adding an extra field to track this stae.
Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 22:49:37 +03:00
|
|
|
* If there was a previous base index, then transfer ownership of allocated
|
|
|
|
* entries to the parent index.
|
2014-06-13 16:19:44 +04:00
|
|
|
*/
|
block alloc: allocate cache entries from mem_pool
When reading large indexes from disk, a portion of the time is
dominated in malloc() calls. This can be mitigated by allocating a
large block of memory and manage it ourselves via memory pools.
This change moves the cache entry allocation to be on top of memory
pools.
Design:
The index_state struct will gain a notion of an associated memory_pool
from which cache_entries will be allocated from. When reading in the
index from disk, we have information on the number of entries and
their size, which can guide us in deciding how large our initial
memory allocation should be. When an index is discarded, the
associated memory_pool will be discarded as well - so the lifetime of
a cache_entry is tied to the lifetime of the index_state that it was
allocated for.
In the case of a Split Index, the following rules are followed. 1st,
some terminology is defined:
Terminology:
- 'the_index': represents the logical view of the index
- 'split_index': represents the "base" cache entries. Read from the
split index file.
'the_index' can reference a single split_index, as well as
cache_entries from the split_index. `the_index` will be discarded
before the `split_index` is. This means that when we are allocating
cache_entries in the presence of a split index, we need to allocate
the entries from the `split_index`'s memory pool. This allows us to
follow the pattern that `the_index` can reference cache_entries from
the `split_index`, and that the cache_entries will not be freed while
they are still being referenced.
Managing transient cache_entry structs:
Cache entries are usually allocated for an index, but this is not always
the case. Cache entries are sometimes allocated because this is the
type that the existing checkout_entry function works with. Because of
this, the existing code needs to handle cache entries associated with an
index / memory pool, and those that only exist transiently. Several
strategies were contemplated around how to handle this:
Chosen approach:
An extra field was added to the cache_entry type to track whether the
cache_entry was allocated from a memory pool or not. This is currently
an int field, as there are no more available bits in the existing
ce_flags bit field. If / when more bits are needed, this new field can
be turned into a proper bit field.
Alternatives:
1) Do not include any information about how the cache_entry was
allocated. Calling code would be responsible for tracking whether the
cache_entry needed to be freed or not.
Pro: No extra memory overhead to track this state
Con: Extra complexity in callers to handle this correctly.
The extra complexity and burden to not regress this behavior in the
future was more than we wanted.
2) cache_entry would gain knowledge about which mem_pool allocated it
Pro: Could (potentially) do extra logic to know when a mem_pool no
longer had references to any cache_entry
Con: cache_entry would grow heavier by a pointer, instead of int
We didn't see a tangible benefit to this approach
3) Do not add any extra information to a cache_entry, but when freeing a
cache entry, check if the memory exists in a region managed by existing
mem_pools.
Pro: No extra memory overhead to track state
Con: Extra computation is performed when freeing cache entries
We decided tracking and iterating over known memory pool regions was
less desirable than adding an extra field to track this stae.
Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 22:49:37 +03:00
|
|
|
if (si->base &&
|
|
|
|
si->base->ce_mem_pool) {
|
|
|
|
|
|
|
|
if (!istate->ce_mem_pool)
|
|
|
|
mem_pool_init(&istate->ce_mem_pool, 0);
|
|
|
|
|
|
|
|
mem_pool_combine(istate->ce_mem_pool, istate->split_index->base->ce_mem_pool);
|
|
|
|
}
|
|
|
|
|
2014-06-13 16:19:44 +04:00
|
|
|
si->base = xcalloc(1, sizeof(*si->base));
|
|
|
|
si->base->version = istate->version;
|
|
|
|
/* zero timestamp disables racy test in ce_write_index() */
|
|
|
|
si->base->timestamp = istate->timestamp;
|
|
|
|
ALLOC_GROW(si->base->cache, istate->cache_nr, si->base->cache_alloc);
|
|
|
|
si->base->cache_nr = istate->cache_nr;
|
block alloc: allocate cache entries from mem_pool
When reading large indexes from disk, a portion of the time is
dominated in malloc() calls. This can be mitigated by allocating a
large block of memory and manage it ourselves via memory pools.
This change moves the cache entry allocation to be on top of memory
pools.
Design:
The index_state struct will gain a notion of an associated memory_pool
from which cache_entries will be allocated from. When reading in the
index from disk, we have information on the number of entries and
their size, which can guide us in deciding how large our initial
memory allocation should be. When an index is discarded, the
associated memory_pool will be discarded as well - so the lifetime of
a cache_entry is tied to the lifetime of the index_state that it was
allocated for.
In the case of a Split Index, the following rules are followed. 1st,
some terminology is defined:
Terminology:
- 'the_index': represents the logical view of the index
- 'split_index': represents the "base" cache entries. Read from the
split index file.
'the_index' can reference a single split_index, as well as
cache_entries from the split_index. `the_index` will be discarded
before the `split_index` is. This means that when we are allocating
cache_entries in the presence of a split index, we need to allocate
the entries from the `split_index`'s memory pool. This allows us to
follow the pattern that `the_index` can reference cache_entries from
the `split_index`, and that the cache_entries will not be freed while
they are still being referenced.
Managing transient cache_entry structs:
Cache entries are usually allocated for an index, but this is not always
the case. Cache entries are sometimes allocated because this is the
type that the existing checkout_entry function works with. Because of
this, the existing code needs to handle cache entries associated with an
index / memory pool, and those that only exist transiently. Several
strategies were contemplated around how to handle this:
Chosen approach:
An extra field was added to the cache_entry type to track whether the
cache_entry was allocated from a memory pool or not. This is currently
an int field, as there are no more available bits in the existing
ce_flags bit field. If / when more bits are needed, this new field can
be turned into a proper bit field.
Alternatives:
1) Do not include any information about how the cache_entry was
allocated. Calling code would be responsible for tracking whether the
cache_entry needed to be freed or not.
Pro: No extra memory overhead to track this state
Con: Extra complexity in callers to handle this correctly.
The extra complexity and burden to not regress this behavior in the
future was more than we wanted.
2) cache_entry would gain knowledge about which mem_pool allocated it
Pro: Could (potentially) do extra logic to know when a mem_pool no
longer had references to any cache_entry
Con: cache_entry would grow heavier by a pointer, instead of int
We didn't see a tangible benefit to this approach
3) Do not add any extra information to a cache_entry, but when freeing a
cache entry, check if the memory exists in a region managed by existing
mem_pools.
Pro: No extra memory overhead to track state
Con: Extra computation is performed when freeing cache entries
We decided tracking and iterating over known memory pool regions was
less desirable than adding an extra field to track this stae.
Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 22:49:37 +03:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The mem_pool needs to move with the allocated entries.
|
|
|
|
*/
|
|
|
|
si->base->ce_mem_pool = istate->ce_mem_pool;
|
|
|
|
istate->ce_mem_pool = NULL;
|
|
|
|
|
2016-09-25 10:24:03 +03:00
|
|
|
COPY_ARRAY(si->base->cache, istate->cache, istate->cache_nr);
|
2014-06-13 16:19:44 +04:00
|
|
|
mark_base_index_entries(si->base);
|
|
|
|
for (i = 0; i < si->base->cache_nr; i++)
|
|
|
|
si->base->cache[i]->ce_flags &= ~CE_UPDATE_IN_BASE;
|
|
|
|
}
|
|
|
|
|
2014-06-13 16:19:41 +04:00
|
|
|
static void mark_entry_for_delete(size_t pos, void *data)
|
|
|
|
{
|
|
|
|
struct index_state *istate = data;
|
|
|
|
if (pos >= istate->cache_nr)
|
|
|
|
die("position for delete %d exceeds base index size %d",
|
|
|
|
(int)pos, istate->cache_nr);
|
|
|
|
istate->cache[pos]->ce_flags |= CE_REMOVE;
|
|
|
|
istate->split_index->nr_deletions = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void replace_entry(size_t pos, void *data)
|
|
|
|
{
|
|
|
|
struct index_state *istate = data;
|
|
|
|
struct split_index *si = istate->split_index;
|
|
|
|
struct cache_entry *dst, *src;
|
2014-06-13 16:19:43 +04:00
|
|
|
|
2014-06-13 16:19:41 +04:00
|
|
|
if (pos >= istate->cache_nr)
|
|
|
|
die("position for replacement %d exceeds base index size %d",
|
|
|
|
(int)pos, istate->cache_nr);
|
|
|
|
if (si->nr_replacements >= si->saved_cache_nr)
|
|
|
|
die("too many replacements (%d vs %d)",
|
|
|
|
si->nr_replacements, si->saved_cache_nr);
|
|
|
|
dst = istate->cache[pos];
|
|
|
|
if (dst->ce_flags & CE_REMOVE)
|
|
|
|
die("entry %d is marked as both replaced and deleted",
|
|
|
|
(int)pos);
|
|
|
|
src = si->saved_cache[si->nr_replacements];
|
2014-06-13 16:19:43 +04:00
|
|
|
if (ce_namelen(src))
|
|
|
|
die("corrupt link extension, entry %d should have "
|
|
|
|
"zero length name", (int)pos);
|
2014-06-13 16:19:41 +04:00
|
|
|
src->index = pos + 1;
|
|
|
|
src->ce_flags |= CE_UPDATE_IN_BASE;
|
2014-06-13 16:19:43 +04:00
|
|
|
src->ce_namelen = dst->ce_namelen;
|
|
|
|
copy_cache_entry(dst, src);
|
2018-07-02 22:49:31 +03:00
|
|
|
discard_cache_entry(src);
|
2014-06-13 16:19:41 +04:00
|
|
|
si->nr_replacements++;
|
|
|
|
}
|
|
|
|
|
2014-06-13 16:19:36 +04:00
|
|
|
void merge_base_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
struct split_index *si = istate->split_index;
|
2014-06-13 16:19:41 +04:00
|
|
|
unsigned int i;
|
2014-06-13 16:19:36 +04:00
|
|
|
|
|
|
|
mark_base_index_entries(si->base);
|
2014-06-13 16:19:41 +04:00
|
|
|
|
|
|
|
si->saved_cache = istate->cache;
|
|
|
|
si->saved_cache_nr = istate->cache_nr;
|
|
|
|
istate->cache_nr = si->base->cache_nr;
|
|
|
|
istate->cache = NULL;
|
|
|
|
istate->cache_alloc = 0;
|
2014-06-13 16:19:36 +04:00
|
|
|
ALLOC_GROW(istate->cache, istate->cache_nr, istate->cache_alloc);
|
2016-09-25 10:24:03 +03:00
|
|
|
COPY_ARRAY(istate->cache, si->base->cache, istate->cache_nr);
|
2014-06-13 16:19:41 +04:00
|
|
|
|
|
|
|
si->nr_deletions = 0;
|
|
|
|
si->nr_replacements = 0;
|
|
|
|
ewah_each_bit(si->replace_bitmap, replace_entry, istate);
|
|
|
|
ewah_each_bit(si->delete_bitmap, mark_entry_for_delete, istate);
|
|
|
|
if (si->nr_deletions)
|
|
|
|
remove_marked_cache_entries(istate);
|
|
|
|
|
|
|
|
for (i = si->nr_replacements; i < si->saved_cache_nr; i++) {
|
2014-06-13 16:19:43 +04:00
|
|
|
if (!ce_namelen(si->saved_cache[i]))
|
|
|
|
die("corrupt link extension, entry %d should "
|
|
|
|
"have non-zero length name", i);
|
2014-06-13 16:19:41 +04:00
|
|
|
add_index_entry(istate, si->saved_cache[i],
|
|
|
|
ADD_CACHE_OK_TO_ADD |
|
2014-06-13 16:19:42 +04:00
|
|
|
ADD_CACHE_KEEP_CACHE_TREE |
|
2014-06-13 16:19:41 +04:00
|
|
|
/*
|
|
|
|
* we may have to replay what
|
|
|
|
* merge-recursive.c:update_stages()
|
|
|
|
* does, which has this flag on
|
|
|
|
*/
|
|
|
|
ADD_CACHE_SKIP_DFCHECK);
|
|
|
|
si->saved_cache[i] = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
ewah_free(si->delete_bitmap);
|
|
|
|
ewah_free(si->replace_bitmap);
|
2017-06-16 02:15:49 +03:00
|
|
|
FREE_AND_NULL(si->saved_cache);
|
2014-06-13 16:19:41 +04:00
|
|
|
si->delete_bitmap = NULL;
|
|
|
|
si->replace_bitmap = NULL;
|
|
|
|
si->saved_cache_nr = 0;
|
2014-06-13 16:19:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
void prepare_to_write_split_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
struct split_index *si = init_split_index(istate);
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
struct cache_entry **entries = NULL, *ce;
|
|
|
|
int i, nr_entries = 0, nr_alloc = 0;
|
|
|
|
|
|
|
|
si->delete_bitmap = ewah_new();
|
|
|
|
si->replace_bitmap = ewah_new();
|
|
|
|
|
|
|
|
if (si->base) {
|
|
|
|
/* Go through istate->cache[] and mark CE_MATCHED to
|
|
|
|
* entry with positive index. We'll go through
|
|
|
|
* base->cache[] later to delete all entries in base
|
2016-10-23 12:26:30 +03:00
|
|
|
* that are not marked with either CE_MATCHED or
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
* CE_UPDATE_IN_BASE. If istate->cache[i] is a
|
|
|
|
* duplicate, deduplicate it.
|
|
|
|
*/
|
|
|
|
for (i = 0; i < istate->cache_nr; i++) {
|
|
|
|
struct cache_entry *base;
|
|
|
|
/* namelen is checked separately */
|
|
|
|
const unsigned int ondisk_flags =
|
|
|
|
CE_STAGEMASK | CE_VALID | CE_EXTENDED_FLAGS;
|
|
|
|
unsigned int ce_flags, base_flags, ret;
|
|
|
|
ce = istate->cache[i];
|
|
|
|
if (!ce->index)
|
|
|
|
continue;
|
|
|
|
if (ce->index > si->base->cache_nr) {
|
|
|
|
ce->index = 0;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
ce->ce_flags |= CE_MATCHED; /* or "shared" */
|
|
|
|
base = si->base->cache[ce->index - 1];
|
|
|
|
if (ce == base)
|
|
|
|
continue;
|
|
|
|
if (ce->ce_namelen != base->ce_namelen ||
|
|
|
|
strcmp(ce->name, base->name)) {
|
|
|
|
ce->index = 0;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
ce_flags = ce->ce_flags;
|
|
|
|
base_flags = base->ce_flags;
|
|
|
|
/* only on-disk flags matter */
|
|
|
|
ce->ce_flags &= ondisk_flags;
|
|
|
|
base->ce_flags &= ondisk_flags;
|
|
|
|
ret = memcmp(&ce->ce_stat_data, &base->ce_stat_data,
|
|
|
|
offsetof(struct cache_entry, name) -
|
|
|
|
offsetof(struct cache_entry, ce_stat_data));
|
|
|
|
ce->ce_flags = ce_flags;
|
|
|
|
base->ce_flags = base_flags;
|
|
|
|
if (ret)
|
|
|
|
ce->ce_flags |= CE_UPDATE_IN_BASE;
|
2018-07-02 22:49:31 +03:00
|
|
|
discard_cache_entry(base);
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
si->base->cache[ce->index - 1] = ce;
|
|
|
|
}
|
|
|
|
for (i = 0; i < si->base->cache_nr; i++) {
|
|
|
|
ce = si->base->cache[i];
|
|
|
|
if ((ce->ce_flags & CE_REMOVE) ||
|
|
|
|
!(ce->ce_flags & CE_MATCHED))
|
|
|
|
ewah_set(si->delete_bitmap, i);
|
|
|
|
else if (ce->ce_flags & CE_UPDATE_IN_BASE) {
|
|
|
|
ewah_set(si->replace_bitmap, i);
|
2014-06-13 16:19:43 +04:00
|
|
|
ce->ce_flags |= CE_STRIP_NAME;
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
ALLOC_GROW(entries, nr_entries+1, nr_alloc);
|
|
|
|
entries[nr_entries++] = ce;
|
|
|
|
}
|
2018-01-08 01:30:14 +03:00
|
|
|
if (is_null_oid(&ce->oid))
|
|
|
|
istate->drop_cache_tree = 1;
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < istate->cache_nr; i++) {
|
|
|
|
ce = istate->cache[i];
|
|
|
|
if ((!si->base || !ce->index) && !(ce->ce_flags & CE_REMOVE)) {
|
2014-06-13 16:19:43 +04:00
|
|
|
assert(!(ce->ce_flags & CE_STRIP_NAME));
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
ALLOC_GROW(entries, nr_entries+1, nr_alloc);
|
|
|
|
entries[nr_entries++] = ce;
|
|
|
|
}
|
|
|
|
ce->ce_flags &= ~CE_MATCHED;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* take cache[] out temporarily, put entries[] in its place
|
|
|
|
* for writing
|
|
|
|
*/
|
|
|
|
si->saved_cache = istate->cache;
|
2014-06-13 16:19:36 +04:00
|
|
|
si->saved_cache_nr = istate->cache_nr;
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
istate->cache = entries;
|
|
|
|
istate->cache_nr = nr_entries;
|
2014-06-13 16:19:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
void finish_writing_split_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
struct split_index *si = init_split_index(istate);
|
split-index: the writing part
prepare_to_write_split_index() does the major work, classifying
deleted, updated and added entries. write_link_extension() then just
writes it down.
An observation is, deleting an entry, then adding it back is recorded
as "entry X is deleted, entry X is added", not "entry X is replaced".
This is simpler, with small overhead: a replaced entry is stored
without its path, a new entry is store with its path.
A note about unpack_trees() and the deduplication code inside
prepare_to_write_split_index(). Usually tracking updated/removed
entries via read-cache API is enough. unpack_trees() manipulates the
index in a different way: it throws the entire source index out,
builds up a new one, copying/duplicating entries (using dup_entry)
from the source index over if necessary, then returns the new index.
A naive solution would be marking the entire source index "deleted"
and add their duplicates as new. That could bring $GIT_DIR/index back
to the original size. So we try harder and memcmp() between the
original and the duplicate to see if it needs updating.
We could avoid memcmp() too, by avoiding duplicating the original
entry in dup_entry(). The performance gain this way is within noise
level and it complicates unpack-trees.c. So memcmp() is the preferred
way to deal with deduplication.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-06-13 16:19:40 +04:00
|
|
|
|
|
|
|
ewah_free(si->delete_bitmap);
|
|
|
|
ewah_free(si->replace_bitmap);
|
|
|
|
si->delete_bitmap = NULL;
|
|
|
|
si->replace_bitmap = NULL;
|
|
|
|
free(istate->cache);
|
|
|
|
istate->cache = si->saved_cache;
|
2014-06-13 16:19:36 +04:00
|
|
|
istate->cache_nr = si->saved_cache_nr;
|
|
|
|
}
|
|
|
|
|
|
|
|
void discard_split_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
struct split_index *si = istate->split_index;
|
|
|
|
if (!si)
|
|
|
|
return;
|
|
|
|
istate->split_index = NULL;
|
|
|
|
si->refcount--;
|
|
|
|
if (si->refcount)
|
|
|
|
return;
|
|
|
|
if (si->base) {
|
|
|
|
discard_index(si->base);
|
|
|
|
free(si->base);
|
|
|
|
}
|
|
|
|
free(si);
|
|
|
|
}
|
2014-06-13 16:19:38 +04:00
|
|
|
|
|
|
|
void save_or_free_index_entry(struct index_state *istate, struct cache_entry *ce)
|
|
|
|
{
|
|
|
|
if (ce->index &&
|
|
|
|
istate->split_index &&
|
|
|
|
istate->split_index->base &&
|
|
|
|
ce->index <= istate->split_index->base->cache_nr &&
|
|
|
|
ce == istate->split_index->base->cache[ce->index - 1])
|
|
|
|
ce->ce_flags |= CE_REMOVE;
|
|
|
|
else
|
2018-07-02 22:49:31 +03:00
|
|
|
discard_cache_entry(ce);
|
2014-06-13 16:19:38 +04:00
|
|
|
}
|
2014-06-13 16:19:39 +04:00
|
|
|
|
|
|
|
void replace_index_entry_in_base(struct index_state *istate,
|
2018-02-14 21:59:48 +03:00
|
|
|
struct cache_entry *old_entry,
|
|
|
|
struct cache_entry *new_entry)
|
2014-06-13 16:19:39 +04:00
|
|
|
{
|
2018-02-14 21:59:48 +03:00
|
|
|
if (old_entry->index &&
|
2014-06-13 16:19:39 +04:00
|
|
|
istate->split_index &&
|
|
|
|
istate->split_index->base &&
|
2018-02-14 21:59:48 +03:00
|
|
|
old_entry->index <= istate->split_index->base->cache_nr) {
|
|
|
|
new_entry->index = old_entry->index;
|
|
|
|
if (old_entry != istate->split_index->base->cache[new_entry->index - 1])
|
2018-07-02 22:49:31 +03:00
|
|
|
discard_cache_entry(istate->split_index->base->cache[new_entry->index - 1]);
|
2018-02-14 21:59:48 +03:00
|
|
|
istate->split_index->base->cache[new_entry->index - 1] = new_entry;
|
2014-06-13 16:19:39 +04:00
|
|
|
}
|
|
|
|
}
|
2017-02-27 21:00:01 +03:00
|
|
|
|
|
|
|
void add_split_index(struct index_state *istate)
|
|
|
|
{
|
|
|
|
if (!istate->split_index) {
|
|
|
|
init_split_index(istate);
|
|
|
|
istate->cache_changed |= SPLIT_INDEX_ORDERED;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void remove_split_index(struct index_state *istate)
|
|
|
|
{
|
2017-06-24 22:02:39 +03:00
|
|
|
if (istate->split_index) {
|
|
|
|
/*
|
block alloc: allocate cache entries from mem_pool
When reading large indexes from disk, a portion of the time is
dominated in malloc() calls. This can be mitigated by allocating a
large block of memory and manage it ourselves via memory pools.
This change moves the cache entry allocation to be on top of memory
pools.
Design:
The index_state struct will gain a notion of an associated memory_pool
from which cache_entries will be allocated from. When reading in the
index from disk, we have information on the number of entries and
their size, which can guide us in deciding how large our initial
memory allocation should be. When an index is discarded, the
associated memory_pool will be discarded as well - so the lifetime of
a cache_entry is tied to the lifetime of the index_state that it was
allocated for.
In the case of a Split Index, the following rules are followed. 1st,
some terminology is defined:
Terminology:
- 'the_index': represents the logical view of the index
- 'split_index': represents the "base" cache entries. Read from the
split index file.
'the_index' can reference a single split_index, as well as
cache_entries from the split_index. `the_index` will be discarded
before the `split_index` is. This means that when we are allocating
cache_entries in the presence of a split index, we need to allocate
the entries from the `split_index`'s memory pool. This allows us to
follow the pattern that `the_index` can reference cache_entries from
the `split_index`, and that the cache_entries will not be freed while
they are still being referenced.
Managing transient cache_entry structs:
Cache entries are usually allocated for an index, but this is not always
the case. Cache entries are sometimes allocated because this is the
type that the existing checkout_entry function works with. Because of
this, the existing code needs to handle cache entries associated with an
index / memory pool, and those that only exist transiently. Several
strategies were contemplated around how to handle this:
Chosen approach:
An extra field was added to the cache_entry type to track whether the
cache_entry was allocated from a memory pool or not. This is currently
an int field, as there are no more available bits in the existing
ce_flags bit field. If / when more bits are needed, this new field can
be turned into a proper bit field.
Alternatives:
1) Do not include any information about how the cache_entry was
allocated. Calling code would be responsible for tracking whether the
cache_entry needed to be freed or not.
Pro: No extra memory overhead to track this state
Con: Extra complexity in callers to handle this correctly.
The extra complexity and burden to not regress this behavior in the
future was more than we wanted.
2) cache_entry would gain knowledge about which mem_pool allocated it
Pro: Could (potentially) do extra logic to know when a mem_pool no
longer had references to any cache_entry
Con: cache_entry would grow heavier by a pointer, instead of int
We didn't see a tangible benefit to this approach
3) Do not add any extra information to a cache_entry, but when freeing a
cache entry, check if the memory exists in a region managed by existing
mem_pools.
Pro: No extra memory overhead to track state
Con: Extra computation is performed when freeing cache entries
We decided tracking and iterating over known memory pool regions was
less desirable than adding an extra field to track this stae.
Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 22:49:37 +03:00
|
|
|
* When removing the split index, we need to move
|
|
|
|
* ownership of the mem_pool associated with the
|
|
|
|
* base index to the main index. There may be cache entries
|
|
|
|
* allocated from the base's memory pool that are shared with
|
|
|
|
* the_index.cache[].
|
2017-06-24 22:02:39 +03:00
|
|
|
*/
|
block alloc: allocate cache entries from mem_pool
When reading large indexes from disk, a portion of the time is
dominated in malloc() calls. This can be mitigated by allocating a
large block of memory and manage it ourselves via memory pools.
This change moves the cache entry allocation to be on top of memory
pools.
Design:
The index_state struct will gain a notion of an associated memory_pool
from which cache_entries will be allocated from. When reading in the
index from disk, we have information on the number of entries and
their size, which can guide us in deciding how large our initial
memory allocation should be. When an index is discarded, the
associated memory_pool will be discarded as well - so the lifetime of
a cache_entry is tied to the lifetime of the index_state that it was
allocated for.
In the case of a Split Index, the following rules are followed. 1st,
some terminology is defined:
Terminology:
- 'the_index': represents the logical view of the index
- 'split_index': represents the "base" cache entries. Read from the
split index file.
'the_index' can reference a single split_index, as well as
cache_entries from the split_index. `the_index` will be discarded
before the `split_index` is. This means that when we are allocating
cache_entries in the presence of a split index, we need to allocate
the entries from the `split_index`'s memory pool. This allows us to
follow the pattern that `the_index` can reference cache_entries from
the `split_index`, and that the cache_entries will not be freed while
they are still being referenced.
Managing transient cache_entry structs:
Cache entries are usually allocated for an index, but this is not always
the case. Cache entries are sometimes allocated because this is the
type that the existing checkout_entry function works with. Because of
this, the existing code needs to handle cache entries associated with an
index / memory pool, and those that only exist transiently. Several
strategies were contemplated around how to handle this:
Chosen approach:
An extra field was added to the cache_entry type to track whether the
cache_entry was allocated from a memory pool or not. This is currently
an int field, as there are no more available bits in the existing
ce_flags bit field. If / when more bits are needed, this new field can
be turned into a proper bit field.
Alternatives:
1) Do not include any information about how the cache_entry was
allocated. Calling code would be responsible for tracking whether the
cache_entry needed to be freed or not.
Pro: No extra memory overhead to track this state
Con: Extra complexity in callers to handle this correctly.
The extra complexity and burden to not regress this behavior in the
future was more than we wanted.
2) cache_entry would gain knowledge about which mem_pool allocated it
Pro: Could (potentially) do extra logic to know when a mem_pool no
longer had references to any cache_entry
Con: cache_entry would grow heavier by a pointer, instead of int
We didn't see a tangible benefit to this approach
3) Do not add any extra information to a cache_entry, but when freeing a
cache entry, check if the memory exists in a region managed by existing
mem_pools.
Pro: No extra memory overhead to track state
Con: Extra computation is performed when freeing cache entries
We decided tracking and iterating over known memory pool regions was
less desirable than adding an extra field to track this stae.
Signed-off-by: Jameson Miller <jamill@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 22:49:37 +03:00
|
|
|
mem_pool_combine(istate->ce_mem_pool, istate->split_index->base->ce_mem_pool);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The split index no longer owns the mem_pool backing
|
|
|
|
* its cache array. As we are discarding this index,
|
|
|
|
* mark the index as having no cache entries, so it
|
|
|
|
* will not attempt to clean up the cache entries or
|
|
|
|
* validate them.
|
|
|
|
*/
|
|
|
|
if (istate->split_index->base)
|
|
|
|
istate->split_index->base->cache_nr = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We can discard the split index because its
|
|
|
|
* memory pool has been incorporated into the
|
|
|
|
* memory pool associated with the the_index.
|
|
|
|
*/
|
|
|
|
discard_split_index(istate);
|
|
|
|
|
2017-06-24 22:02:39 +03:00
|
|
|
istate->cache_changed |= SOMETHING_CHANGED;
|
|
|
|
}
|
2017-02-27 21:00:01 +03:00
|
|
|
}
|