ISAR/libmdbx
3
0
Fork 0
mirror of https://github.com/isar/libmdbx.git synced 2025-01-02 02:34:12 +08:00
Code Issues Packages Projects Releases Wiki Activity

mdbx: rework & clarify backlog for gc-update.

Browse Source 
Change-Id: I31f2b6919810b894e69af34bfee9a5b7f5a513fc
This commit is contained in:
Leonid Yuriev 2019-12-14 02:38:13 +03:00
parent b6a00a881e
commit 13d68a1200
1 changed files with 28 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
src/elements/core.c
View File

@ -5842,39 +5842,40 @@ static __always_inline unsigned backlog_size(MDBX_txn *txn) {
return MDBX_PNL_SIZE(txn->tw.reclaimed_pglist) + txn->tw.loose_count;
}
static __always_inline unsigned gctree_backlog(MDBX_txn *txn) {
return /* for split upto root page */ txn->mt_dbs[FREE_DBI].md_depth +
/* for rebalance */ 2 + /* for grow */ 1;
}
/* LY: Prepare a backlog of pages to modify GC itself,
* while reclaiming is prohibited. It should be enough to prevent search
* in mdbx_page_alloc() during a deleting, when GC tree is unbalanced. */
static int mdbx_prep_backlog(MDBX_txn *txn, MDBX_cursor *gc_cursor,
const size_t pnl_bytes) {
const unsigned linear = number_of_ovpages(
txn->mt_env,
pnl_bytes ? pnl_bytes : MDBX_PNL_SIZEOF(txn->tw.retired_pages));
const unsigned backlog = linear + gctree_backlog(txn);
const unsigned linear4list = number_of_ovpages(txn->mt_env, pnl_bytes);
const unsigned backlog4cow = txn->mt_dbs[FREE_DBI].md_depth;
const unsigned backlog4rebalance = backlog4cow + 1;
if (likely(
linear == 1 &&
backlog_size(txn) >
(pnl_bytes
? backlog
: backlog + /* for COW */ txn->mt_dbs[FREE_DBI].md_depth)))
if (likely(linear4list == 1 &&
backlog_size(txn) > (pnl_bytes
? backlog4rebalance
: (backlog4cow + backlog4rebalance))))
return MDBX_SUCCESS;
mdbx_trace(">> pnl_bytes %zu, backlog %u, 4list %u, 4cow %u, 4rebalance %u",
pnl_bytes, backlog_size(txn), linear4list, backlog4cow,
backlog4rebalance);
gc_cursor->mc_flags &= ~C_RECLAIMING;
int err = mdbx_cursor_touch(gc_cursor);
if (err == MDBX_SUCCESS && linear > 1)
err = mdbx_page_alloc(gc_cursor, linear, nullptr, MDBX_ALLOC_ALL);
mdbx_trace("== after-touch, backlog %u, err %d", backlog_size(txn), err);
while (err == MDBX_SUCCESS && backlog_size(txn) < backlog)
if (linear4list > 1 && err == MDBX_SUCCESS) {
err = mdbx_page_alloc(gc_cursor, linear4list, nullptr,
MDBX_ALLOC_GC | MDBX_ALLOC_CACHE);
mdbx_trace("== after-4linear, backlog %u, err %d", backlog_size(txn), err);
}
while (backlog_size(txn) < backlog4cow + linear4list && err == MDBX_SUCCESS)
err = mdbx_page_alloc(gc_cursor, 1, NULL, MDBX_ALLOC_GC);
gc_cursor->mc_flags |= C_RECLAIMING;
mdbx_trace("<< backlog %u, err %d", backlog_size(txn), err);
return (err != MDBX_NOTFOUND) ? err : MDBX_SUCCESS;
}
@ -5924,6 +5925,10 @@ retry:
goto bailout;
}
rc = mdbx_prep_backlog(txn, &mc, MDBX_PNL_SIZEOF(txn->tw.retired_pages));
if (unlikely(rc != MDBX_SUCCESS))
goto bailout;
unsigned settled = 0, cleaned_gc_slot = 0, reused_gc_slot = 0,
filled_gc_slot = ~0u;
txnid_t cleaned_gc_id = 0, gc_rid = txn->tw.last_reclaimed;
@ -6131,8 +6136,11 @@ retry:
mdbx_debug_extra_print(" %" PRIaPGNO, txn->tw.retired_pages[i]);
mdbx_debug_extra_print("%s", "\n");
}
if (unlikely(amount != MDBX_PNL_SIZE(txn->tw.reclaimed_pglist)))
if (unlikely(amount != MDBX_PNL_SIZE(txn->tw.reclaimed_pglist))) {
mdbx_trace("%s.reclaimed-list changed %u -> %u, retry", dbg_prefix_mode,
amount, (unsigned)MDBX_PNL_SIZE(txn->tw.reclaimed_pglist));
goto retry /* rare case, but avoids GC fragmentation and one loop. */;
}
continue;
}