Latches

What is a latch?
Latches are locking mechanisms used to protect shared memory structures from potential corruption due to concurrent access. In other words,latches ensure exclusive access to the shared data structures in the SGA. Access to SGA structures is seralized, using latches. Latches, unlike locks or enqueues are not used to protect database objects.
Latches are used in almost all operations. For example
1. When a session reads a block from disk, it must modify a free block in the buffer cache and adjust the buffer cache LRU chain.
2. When a session reads a block from the SGA, it will modify the LRU chain.
3. When a new SQL statement is parsed, it will be added to the library cache within the SGA.
4. As modifications are made to blocks, entries are placed in the redo buffer.
5. The database writer periodically writes buffers from the cache to disk (and must update their status from “dirty” to “clean”).
The redo log writer writes entries from the redo buffer to the redo logs.

Latch Types

There are three types of latches: parent, child, and solitary latches. The parent and solitary latches are fixed in the Oracle kernel code. Child latches are created at instance startup.

1. Latch operation

2. How are latches acquired?

A process may request a latch in the willing-to-wait or no-wait (immediate) mode. If the latch is available on the first request, the process acquires it. Before modifying the protected data structure, the process writes the recovery information in the latch recovery area so that PMON knows what to clean up if the process dies while holding the latch.

If the latch is not available, the process spins on the CPU for a short while and retries for the latch. This spin and retry activity can be repeated up to the value specified in the intialization parameter _SPIN_COUNT value (default 2000). The values of initialization parameters _MAX_EXPONENTIAL_SLEEP and _MAX_SLEEP_HOLDING_LATCH are used at arriving the spin duration.

Latches are very light and on most systems, a single machine instruction called “test and set” is used to see if the latch is taken (by looking at a specific memory address) and if not, acquire it (by changing the value in the memory address).

---Wakeup mechanisms

1. Timeout

The operating system signals (wakes up) the process when a set alarm is triggered

2. Latch wait posting

The next process to free the required latch will wake up the process waiting for the latch. This is initiated by the requesting

process before going to sleep by putting itself in a latch wait list