(Probably fixes the ITS, definitely fixes a bug) when collapsing
the root page, fixups of other cursors was incomplete.

Leonid Yuriev authored Oct 17, 2014 and Hallvard Furuseth committed Jan 14, 2015

Inconsistent reading & SIGSEGV due to the race condition.

Leonid Yuriev authored Oct 17, 2014 and Hallvard Furuseth committed Jan 14, 2015

Globally shared fields of meta-data were not 'volatile'.
Code from 9a80a8a8,
except the unportable __synchronize().

Check for ext3/ext4 fs, then check kernel version.

This reverts commit e95c9231.

Revert "Hack for potential ext3/ext4 corruption issue"

This reverts commit 91155b9d.

Use regular fsync() if we think this commit grew the DB file.

More fallout from 4d02c741:
Don't modify another thread's write txn. Reinit me_txn0 fully.

Howard Chu authored Oct 08, 2014 and Hallvard Furuseth committed Dec 01, 2014

fstat returns -1, not an errno. Move code into its own function
for reuse.

Howard Chu authored Oct 04, 2014 and Hallvard Furuseth committed Dec 01, 2014

More fallout from 4d02c741

Jean-Christophe DUBOIS authored Oct 03, 2014 and Hallvard Furuseth committed Dec 01, 2014

Signed-off-by: Jean-Christophe DUBOIS <jcd@tribudubois.net>

When reusing the main write txn

Free writer txn on env_close

Change retry to num times 60.  Testing shows that on a known dataset, this has the same growth behavior as 2.4.39, while num times 20 resulted in significant growth.

Leave it in main env struct, it doesn't need to be rolled back
with nested txns.

Don't walk the readers table unless we really need to.

we already checked for data == NULL, drop unnecessary if

Move CACHEFLUSH, avoid unnecessary if()

Of course, MacOSX only has posix_memalign, and only in recent versions.
(Older versions could use valloc instead. That precludes ever changing
the LMDB pagesize to be different from the OS pagesize.) Not all that
crucial since it only affects mdb_env_copyfd1.

Leaving them in dirty_list caused breakage.  Instead merge them
into me_pghead at the end, when no more pages will be allocated.

Don't let dummy go out of scope

Everything has it. Solaris doesn't have posix_memalign.