GGLinnk/bees
1
0
Fork 0
mirror of https://github.com/Zygo/bees.git synced 2025-05-17 21:35:45 +02:00
Code Packages

hash: clean up comments, audit for bugs

Browse Source 
We stopped supporting shared hash tables a long time ago.  Remove comments
describing the behavior of shared hash tables.

Add an event counter for pushing a hash to the front when it is already at
the front.

Audited the code for a bug related to bucket handling that impairs space
efficiency when the bucket size is greater than 1.  Didn't find one.

Signed-off-by: Zygo Blaxell <bees@furryterror.org>
This commit is contained in:
Zygo Blaxell 2018-12-02 15:52:00 -05:00
parent 570b3f7de0
commit f17cf084e6
1 changed files with 30 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
src/bees-hash.cc
View File

@ -35,7 +35,7 @@ dump_bucket_locked(BeesHashTable::Cell *p, BeesHashTable::Cell *q)
} }
#endif #endif
const bool VERIFY_CLEARS_BUGS = false; static const bool VERIFY_CLEARS_BUGS = false;
bool bool
verify_cell_range(BeesHashTable::Cell *p, BeesHashTable::Cell *q, bool clear_bugs = VERIFY_CLEARS_BUGS) verify_cell_range(BeesHashTable::Cell *p, BeesHashTable::Cell *q, bool clear_bugs = VERIFY_CLEARS_BUGS)
@ -396,10 +396,10 @@ BeesHashTable::fetch_missing_extent_by_index(uint64_t extent_index)
catch_all([&]() { catch_all([&]() {
BEESTOOLONG("pread(fd " << m_fd << " '" << name_fd(m_fd)<< "', length " << to_hex(dirty_extent_end - dirty_extent) << ", offset " << to_hex(dirty_extent - m_byte_ptr) << ")"); BEESTOOLONG("pread(fd " << m_fd << " '" << name_fd(m_fd)<< "', length " << to_hex(dirty_extent_end - dirty_extent) << ", offset " << to_hex(dirty_extent - m_byte_ptr) << ")");
pread_or_die(m_fd, dirty_extent, dirty_extent_end - dirty_extent, dirty_extent - m_byte_ptr); pread_or_die(m_fd, dirty_extent, dirty_extent_end - dirty_extent, dirty_extent - m_byte_ptr);
});
// Only count extents successfully read // Only count extents successfully read
BEESCOUNT(hash_extent_in); BEESCOUNT(hash_extent_in);
});
} }
void void
@ -425,11 +425,9 @@ BeesHashTable::find_cell(HashType hash)
return rv; return rv;
} }
// Move an entry to the end of the list. Used after an attempt to resolve /// Remove a hash from the table, leaving an empty space on the list
// an address in the hash table fails. Probably more correctly called /// where the hash used to be. Used when an invalid address is found
// push_back_hash_addr, except it never inserts. Shared hash tables /// because lookups on invalid addresses really hurt.
// never erase anything, since there is no way to tell if an entry is
// out of date or just belonging to the wrong filesystem.
void void
BeesHashTable::erase_hash_addr(HashType hash, AddrType addr) BeesHashTable::erase_hash_addr(HashType hash, AddrType addr)
{ {
@ -441,7 +439,6 @@ BeesHashTable::erase_hash_addr(HashType hash, AddrType addr)
Cell *ip = find(er.first, er.second, mv); Cell *ip = find(er.first, er.second, mv);
bool found = (ip < er.second); bool found = (ip < er.second);
if (found) { if (found) {
// Lookups on invalid addresses really hurt us. Kill it with fire!
*ip = Cell(0, 0); *ip = Cell(0, 0);
set_extent_dirty_locked(hash_to_extent_index(hash)); set_extent_dirty_locked(hash_to_extent_index(hash));
BEESCOUNT(hash_erase); BEESCOUNT(hash_erase);
@ -450,14 +447,17 @@ BeesHashTable::erase_hash_addr(HashType hash, AddrType addr)
BEESLOGDEBUG("while erasing hash " << hash << " addr " << addr); BEESLOGDEBUG("while erasing hash " << hash << " addr " << addr);
} }
#endif #endif
} else {
BEESCOUNT(hash_erase_miss);
} }
} }
// If entry is already present in list, move it to the front of the /// Insert a hash entry at the head of the list. If entry is already
// list without dropping any entries, and return true. If entry is not /// present in list, move it to the front of the list without dropping
// present in list, insert it at the front of the list, possibly dropping /// any entries, and return true. If entry is not present in list,
// the last entry in the list, and return false. Used to move duplicate /// insert it at the front of the list, possibly dropping the last entry
// hash blocks to the front of the list. /// in the list, and return false. Used to move duplicate hash blocks
/// to the front of the list.
bool bool
BeesHashTable::push_front_hash_addr(HashType hash, AddrType addr) BeesHashTable::push_front_hash_addr(HashType hash, AddrType addr)
{ {
@ -481,7 +481,7 @@ BeesHashTable::push_front_hash_addr(HashType hash, AddrType addr)
auto dp = ip; auto dp = ip;
--sp; --sp;
// If we are deleting the last entry then don't copy it // If we are deleting the last entry then don't copy it
if (ip == er.second) { if (dp == er.second) {
--sp; --sp;
--dp; --dp;
BEESCOUNT(hash_evict); BEESCOUNT(hash_evict);
@ -495,6 +495,8 @@ BeesHashTable::push_front_hash_addr(HashType hash, AddrType addr)
er.first[0] = mv; er.first[0] = mv;
set_extent_dirty_locked(hash_to_extent_index(hash)); set_extent_dirty_locked(hash_to_extent_index(hash));
BEESCOUNT(hash_front); BEESCOUNT(hash_front);
} else {
BEESCOUNT(hash_front_already);
} }
#if 0 #if 0
if (verify_cell_range(er.first, er.second)) { if (verify_cell_range(er.first, er.second)) {
@ -504,11 +506,12 @@ BeesHashTable::push_front_hash_addr(HashType hash, AddrType addr)
return found; return found;
} }
// If entry is already present in list, returns true and does not /// Insert a hash entry at some unspecified point in the list.
// modify list. If entry is not present in list, returns false and /// If entry is already present in list, returns true and does not
// inserts at a random position in the list, possibly evicting the entry /// modify list. If entry is not present in list, returns false and
// at the end of the list. Used to insert new unique (not-yet-duplicate) /// inserts at a random position in the list, possibly evicting the entry
// blocks in random order. /// at the end of the list. Used to insert new unique (not-yet-duplicate)
/// blocks in random order.
bool bool
BeesHashTable::push_random_hash_addr(HashType hash, AddrType addr) BeesHashTable::push_random_hash_addr(HashType hash, AddrType addr)
{ {
@ -573,7 +576,12 @@ BeesHashTable::push_random_hash_addr(HashType hash, AddrType addr)
} }
// Evict something and insert at pos // Evict something and insert at pos
move_backward(er.first + pos, er.second - 1, er.second); // move_backward(er.first + pos, er.second - 1, er.second);
ip = er.second - 1;
while (ip > er.first + pos) {
auto dp = ip;
*dp = *--ip;
}
er.first[pos] = mv; er.first[pos] = mv;
BEESCOUNT(hash_evict); BEESCOUNT(hash_evict);
case_cond = 5; case_cond = 5;