6.42
IDT71342SA/LA
High-Speed 4K x 8 Dual-Port Static RAM with Semaphore Industrial and Commercial Temperature Ranges
11
FUNCTIONAL DESCRIPTION
The IDT71342 is an extremely fast Dual-Port 4K x 8 CMOS Static
RAM with an additional 8 address locations dedicated to binary
semaphore flags. These flags allow either processor on the left or right
side of the Dual-Port RAM to claim a privilege over the other processor
for functions defined by the system designer’s software. As an example,
the semaphore can be used by one processor to inhibit the other from
accessing a portion of the Dual-Port RAM or any other shared
resource.
The Dual-Port RAM features a fast access time, and both ports are
completely independent of each other. This means that the activity on
the left port in no way slows the access time of the right port. Both ports
are identical in function to standard CMOS Static RAMs and can be
read from or written to at the same time, with the only possible conflict
arising from the simultaneous writing of, or a simultaneous READ/
WRITE of, a non-semaphore location. Semaphores are protected
against such ambiguous situations and may be used by the system
program to avoid any conflicts in the non-semaphore portion of the
Dual-Port SRAM. These devices have an automatic power-down
feature controlled by CE, the Dual-Port SRAM enable, and SEM, the
semaphore enable. The CE and SEM pins control on-chip power down
circuitry that permits the respective port to go into standby mode when
not selected. This is the condition which is shown in Truth Table I
where CE and SEM are both HIGH.
Systems which can best use the IDT71342 contain multiple
processors or controllers and are typically very high-speed systems
which are software controlled or software intensive. These systems
can benefit from a performance increase offered by the IDT71342’s
hardware semaphores, which provide a lockout mechanism without
requiring complex programming.
Software handshaking between processors offers the maximum in
system flexibility by permitting shared resources to be allocated in
varying configurations. The IDT71342 does not use its semaphore
flags to control any resources through hardware, thus allowing the
system designer total flexibility in system architecture.
An advantage of using semaphores rather than the more common
methods of hardware arbitration is that wait states are never incurred
in either processor. This can prove to be a major advantage in very
high-speed systems.
How the Semaphore Flags Work
The semaphore logic is a set of eight latches which are independent
of the Dual-Port RAM. These latches can be used to pass a flag, or
token, from one port to the other to indicate that a shared resource is
in use. The semaphores provide a hardware assist for a use assignment
method called “Token Passing Allocation.” In this method, the state of
a semaphore latch is used as a token indicating that a shared resource
is in use. If the left processor wants to use this resource, it requests the
token by setting the latch. This processor then verifies its success in
setting the latch by reading it. If it was successful, it proceeds to
assume control over the shared resource. If it was not successful in
setting the latch, it determines that the right side processor had set the
latch first, has the token and is using the shared resource. The left
processor can then either repeatedly request that semaphore’s status
or remove its request for that semaphore to perform another task and
occasionally attempt again to gain control of the token via the set and
test sequence. Once the right side has relinquished the token, the left
side should succeed in gaining control.
The semaphore flags are active LOW. A token is requested by
writing a zero into a semaphore latch and is released when the same
side writes a one to that latch.
The eight semaphore flags reside within the IDT71342 in a separate
memory space from the Dual-Port RAM. This address space is
accessed by placing a LOW input on the SEM pin (which acts as a chip
select for the semaphore flags) and using the other control pins
(Address, OE, and R/W) as they would be used in accessing
a standard Static RAM. Each of the flags has a unique address
which can be accessed by either side through the address pins A0–A2.
When accessing the semaphores, none of the other address pins has
any effect.
When writing to a semaphore, only data pin D0 is used. If a LOW
level is written into an unused semaphore location, that flag will be set
to a zero on that side and a one on the other (see Truth Table II). That
semaphore can now only be modified by the side showing the zero.
When a one is written into the same location from the same side, the
flag will be set to a one for both sides (unless a semaphore request
from the other side is pending) and then can be written to by both sides.
The fact that the side which is able to write a zero into a semaphore
subsequently locks out writes from the other side is what makes
semaphore flags useful in interprocessor communications. (A thorough
discussion on the use of this feature follows shortly.) A zero written into
the same location from the other side will be stored in the semaphore
request latch for that side until the semaphore is freed by the first side.
When a semaphore flag is read, its value is spread into all data bits
so that a flag that is a one reads as a one in all data bits and a flag
containing a zero reads as all zeros. The read value is latched into one
side’s output register when that side’s semaphore select (SEM) and
output enable (OE) signals go active. This serves to disallow the
semaphore from changing state in the middle of a read cycle due to a
write cycle from the other side. Because of this latch, a repeated read
of a semaphore in a test loop must cause either signal (SEM or OE) to
go inactive or the output will never change.
A sequence of WRITE/READ must be used by the semaphore in
order to guarantee that no system level contention will occur. A
processor requests access to shared resources by attempting to write
a zero into a semaphore location. If the semaphore is already in use,
the semaphore request latch will contain a zero, yet the semaphore
flag will appear as a one, a fact which the processor will verify by the
subsequent read (see Truth Table II). As an example, assume a
processor writes a zero in the left port at a free semaphore location. On
a subsequent read, the processor will verify that it has written
successfully to that location and will assume control over the resource
in question. Meanwhile, if a processor on the right side attempts to
write a zero to the same semaphore flag it will fail, as will be verified
by the fact that a one will be read from that semaphore on the right side
during a subsequent read. Had a sequence of READ/WRITE been
used instead, system contention problems could have occurred during
the gap between the read and write cycles.
It is important to note that a failed semaphore request must be
followed by either repeated reads or by writing a one into the same
location. The reason for this is easily understood by looking at the
simple logic diagram of the semaphore flag in Figure 3. Two semaphore