Age | Commit message (Collapse) | Author | Files | Lines |
|
(This used to be commit 40c0919aaa9c1b14bbaebb95ecce53eb0380fdbb)
|
|
(This used to be commit 84b468b2f8f2dffda89593f816e8bc6a8b6d42ac)
|
|
this ensures that having the global lock also implies the transaction
lock
(This used to be commit 9dbb2633d7781fcc5d15b175ef36bfda5eb199bb)
|
|
(This used to be commit bffb8f88006fd2530cad270df642d9a23a01b34b)
|
|
(This used to be commit a0ff739bcab32065d9d3957eb1d93f7791f84f04)
|
|
metze
(This used to be commit 8f24f6b38e967075589529a08c68a1a56f9f0499)
|
|
samba-technical@samba.org. If
you do a tdb_set_max_dead(tdb, n), then for this tdb a delete operation will
only mark a record as dead and re-use it if a new record is created. The
parameter n allows for at most n dead records per hash chain. If this number
is exceeded, all dead records are put on the central freelist.
Volker
(This used to be commit 98a27ab28a3cd554e370a9a0e3652f4dea8749e9)
|
|
(This used to be commit afe7d7855841066b88859976ac748cbf438a9a9f)
|
|
in-memory
mirrors of the hash chain locks to a dynamically allocated one.
Jeremy, I count on you to revert it if the build farm freaks out, it's after
midnight here :-)
Volker
(This used to be commit 7b5db2e472c7e27231fa432d3930789e708abd09)
|
|
call which does a read lock on all chains. These will be used to make
ldb searches more efficient
(This used to be commit de664ec1f8cf179f1d650563272c0de3f7636e2b)
|
|
(This used to be commit 025b142ff268498cfb36fc0e7e9e25a2c5963d38)
|
|
isolation of our portability environment from the main code, and also
simplifies the includes system (no separate #ifdef _SAMBA_BUILD for
tdb. ldb etc now)
(This used to be commit 77d1a468e06290aba789e2f3affc769fc5159a21)
|
|
HAVE_PREAD_DECL/HAVE_PWRITE_DECL
and common/io.h already defines pread and pwrite as static if they are not.
remove unneded defines
(This used to be commit 941f680453a081d51f6499f9b5dc06c7e6640334)
|
|
This clearly links the log function with its private pointer, and
makes the argument list for tdb_open_ex a bit shorter.
Andrew Bartlett
(This used to be commit 5d5503e8d8a10ead3ef21a5ffda52cadb9a07727)
|
|
as part of ldb.
This allows tdb failures to be passed all the way up to Samba's DEBUG
system, which allowed easier debugging.
Unfortunately I had to extend the tdb API, as the logging function
didn't have a context pointer.
I've worked over the 'debug levels' in TDB. Most of them were 0,
which didn't seem right, as some were trace-like messages. We didn't
see any of these previously, except when accessing TDB directly.
Andrew Bartlett
(This used to be commit 58898092c1ce043f6d698db5065f372b79109e22)
|
|
for an extremely lightweight test to see if a tdb has possibly
changed.
(This used to be commit f325ba605ccceca63712c0f2c98961e35e437b3d)
|
|
the discard_const_p() was causing problems on openbsd where intptr_t is not defined
(This used to be commit f02a92787b179a4070227d87e37b7ac092d9e6eb)
|
|
(This used to be commit a11762e7bc139b0280e457e682722d955b81d8e5)
|
|
(This used to be commit 144cc3da5eadf0dcb28ef722feeed813c033e08a)
|
|
EDEADLK even when progress can be made. This is not a good solution,
but I can't find anything better.
(This used to be commit 980dd17f7d0a622cd772afc9ba15e50007ad9c6e)
|
|
(This used to be commit 196995ea38414ce5a57031c4cdf61b3ed548196a)
|
|
have it. Overestimating is harmless.
(This used to be commit ab953c8c72060c496876b6f39d388ad2f7e3c7e0)
|
|
(This used to be commit e615f36733966f9992670e3650fbe9b20d7d218c)
|
|
don't end up doing a mmap read only
(This used to be commit 294ccfd46a0c4e1af9365d028acdabec03c41ad3)
|
|
(This used to be commit 03942dd54b31e7a4cf9c6270b26ccf68a3234b40)
|
|
allowing searches to proceed while another process is in a
transaction, then only upgrading the transaction lock to a write lock
on commit.
The solution is:
- split tdb_traverse() into two calls, called tdb_traverse() and
tdb_traverse_read(). The _read() version only gets read locks, and
will fail any write operations made in the callback from the
traverse.
- the normal tdb_traverse() call allows for read or write operations
in the callback, but gets the transaction lock, preventing
transastions from starting inside the traverse
In addition we enforce the following rule that you may not start a
transaction within a traverse callback, although you can start a
traverse within a transaction
With these rules in place I believe all the deadlock possibilities are
removed, and we can now allow for searches to happen in parallel with
transactions
(This used to be commit 7dd31288a701d772e45b1960ac4ce4cc1be782ed)
|
|
samba-technical posting for more details on the transactions design.
This also adds a number of command line arguments to tdbtorture,
making it more flexible, and fixes some lock deadlock conditions in
the tdbtorture code.
(This used to be commit 06bd8abba942ec9f1e23f5c5d546cbb71ca3a701)
|
|
just involves splitting up the core tdb.c code into separate files on
logical boundaries, but there are some minor functional changes as well:
- move the 'struct tdb_context' into tdb_private.h, hiding it from
users. This was done to allow the structure to change without
breaking code that uses tdb.
- added accessor functions tdb_fd(), tdb_name(), and tdb_log_fn() to
access the elements of struct tdb_context that were used by
external code but are no longer visible
- simplied tdb_append() to use tdb_fetch()/tdb_store(), which is just
as good due to the way tdb locks work
- changed some of the types (such as tdb_off to tdb_off_t) to make
syntax highlighting work better
- removed the old optional spinlock code. It was a bad idea.
- fixed a bug in tdb_reopen_all() that caused tdbtorture to sometimes
fail or report nasty looking errors. This is the only real bug
fixed in this commit. Jeremy/Jerry, you might like to pickup this
change for Samba3, as that could definately affect smbd in
Samba3.
The aim of all of these changes is to make the tdb
transactions/journaling code I am working on easier to write. I
started to write it on top of the existing tdb.c code and it got very
messy. Splitting up the code makes it much easier to follow.
There are more cleanups we could do in tdb, such as using uint32_t
instead of u32 (suggested by metze). I'll leave those for another day.
(This used to be commit 4673cdd0d261614e707b72a7a348bb0e7dbb2482)
|