• IDT70V07 easily expands data bus width to 16 bits or
more using the Master/Slave select when cascading
more than one device
•M/S = H for
M/S = L for
BUSY
output flag on Master
BUSY
input on Slave
• Busy and Interrupt Flags
FUNCTIONAL BLOCK DIAGRAM
OE
L
CE
L
R/
W
L
• On-chip port arbitration logic
• Full on-chip hardware support of semaphore signaling
between ports
• Fully asynchronous operation from either port
• Devices are capable of withstanding greater than 2001V
electrostatic discharge
• LVTTL-compatible, single 3.3V (±0.3V) power supply
• Available in 68-pin PGA, 68-pin PLCC, and a 64-pin
TQFP
DESCRIPTION:
The IDT70V07 is a high-speed 32K x 8 Dual-Port Static
RAM. The IDT70V07 is designed to be used as a stand-alone
Dual-Port RAM or as a combination MASTER/SLAVE DualPort RAM for 16-bit-or-more word systems. Using the IDT
MASTER/SLAVE Dual-Port RAM approach in 16-bit or wider
memory system applications results in full-speed, error-free
operation without the need for additional discrete logic.
For latest information contact IDT’s web site at www.idt.com or fax-on-demand at 408-492-8391.
6.37
1
IDT70V07S/L
INDEX
I/O
2L
V
CC
GND
GND
A
4R
BUSY
L
BUSY
R
GND
M/
S
OE
L
I/O
1L
R/
W
L
CE
L
SEM
L
V
CC
OE
R
CE
R
R/
W
R
SEM
R
A
12R
GND
I/O
3L
I/O
4L
I/O
5L
I/O
6L
I/O
7L
I/O
0R
I/O
1R
I/O
2R
V
CC
I/O
3R
I/O
4R
I/O
5R
I/O
7R
A
11R
A
10R
A
9R
A
8R
A
7R
A
6R
A
5R
A
3R
A
2R
A
1R
A
0R
A
0L
A
1L
A
2L
A
3L
A
4L
A
6L
A
7L
A
8L
A
9L
A
10L
A
11L
A
12L
I/O
0L
2943 drw 03
A
13R
A
13L
70V07
PN80-1
TQFP
TOP
VIEW
(3)
8
9
10
11
12
13
14
15
16
1
2
3
4
5
6
7
46
45
44
43
42
41
56
55
54
53
52
51
50
47
48
49
32
31
30
29
28
27
26
25
24
23
22
21
63
62
61
64
33
34
35
36
37
38
39
40
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
N/C
N/C
A
14L
N/C
N/C
N/C
N/C
A
14R
N/C
N/C
17
18
19
20
57
58
5960A
5L
N/C
INT
L
INT
R
N/C
N/C
N/C
I/O
6R
N/C
N/C
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
This device provides two independent ports with separate
control, address, and I/O pins that permit independent,
asynchronous access for reads or writes to any location in
memory. An automatic power down feature controlled by
CE
permits the on-chip circuitry of each port to enter a very low
standby power mode.
Fabricated using IDT’s CMOS high-performance technol-
ogy, these devices typically operate on only 450mW of power.
The IDT70V07 is packaged in a ceramic 68-pin PGA, a 68-
pin PLCC, and a 80-pin thin plastic quad flatpack (TQFP).
A5L
4L
A
A3L
A2L
A1L
A0L
INT
L
BUSY
L
GND
M/
S
R
BUSY
INT
R
A0R
A1R
A2R
A3R
A4R
NOTES:
1. All Vcc pins must be connected to the power supply.
2. All GND pins must be connected to the ground supply.
3. This text does not indicate the actual part marking.
6.372
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
PIN CONFIGURATIONS (CONT'D)
(1,2)
11
10
09
08
07
06
05
04
03
02
01
515048464442403836
M/
L
BUSY
S
R
53
55
57
59
A
4L
A
2L
A
0L
INT
L
BUSY
GND
A
5L
52
A
7L
A
6L
4745434134
493937
A
3L
A
1L
54
A
9L
A
8L
56
A
11L
A
10L
58
V
CC
A
12L
IDT70V07
G68-1
60
61
A
14L
13L
62
L
CE
L
64
L
R/
W
L
A
63
SEM
65
OE
676866
I/O
0L
N/C
13579
I/O
1L
I/O
2L
I/O
GNDGND
4L
246810121416
I/O
5L
I/O
I/O
3L
6L
68-PIN PGA
TOP VIEW
I/O
7L
I/O
V
CC
0R
(3)
111315
I/O
1R
I/O2RI/O3RI/O5RI/O
INT
A
V
0R
CC
1R
A
A
R
3R
35
A
2R
4R
32
A
7R
30
A
9R
28
A
11R
26
GND
24
A
14R
22
SEM
20
OE
A
5R
33
A
6R
31
A
8R
29
A
10R
27
A
12R
25
A
13R
23
R
CE
R
21
R
R/
W
R
A
1819
I/O
7R
I/O
4R
N/C
17
6R
ABCDEFGHJ
INDEX
NOTES:
CC pins must be connected to power supply.
1. All V
2. All GND pins must be connected to ground supply.
3. This text does not indicate orientation of the actual part-marking.
PIN NAMES
Left PortRight PortNames
CE
L
R/
W
LR/WRRead/Write Enable
OE
L
A
0L – A14LA0R – A14RAddress
I/O
0L – I/O7LI/O0R – I/O7RData Input/Output
SEM
L
INT
L
BUSY
L
CE
RChip Enable
OE
ROutput Enable
SEM
RSemaphore Enable
INT
RInterrupt Flag
BUSY
RBusy Flag
M/
S
V
CCPower
Master or Slave Select
GNDGround
2943 tbl 01
K
L
2943 drw 04
6.373
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
TRUTH TABLE I – NON-CONTENTION READ/WRITE CONTROL
Inputs
CECE
CE
CECE
R/
WW
W
WW
HXXHHigh-ZDeselected: Power-Down
LLXHDATA
LHLHDATA
XXHXHigh-ZOutputs Disabled
NOTE:
1. A0L — A14L≠ A0R — A14R.
(1)
OEOE
OE
OEOE
SEMSEM
SEM
SEMSEM
Outputs
0-7 Mode
I/O
INWrite to Memory
OUTRead Memory
2943 tbl 02
TRUTH TABLE II – SEMAPHORE READ/WRITE CONTROL
(1)
InputsOutputs
CECE
CE
CECE
HHLLDATA
HXLDATA
R/
WW
W
WW
OEOE
OE
OEOE
SEMSEM
SEM
SEMSEM
0-7 Mode
I/O
OUTRead Data in Semaphore Flag
INWrite I/O0 into Semaphore Flag
LXXL—Not Allowed
NOTE:2943 tbl 03
1. There are eight semaphore flags written to via I/O0 and read from all I/O's (I/O0-I/O7). These eight semaphores are addressed by A0 - A2.
ABSOLUTE MAXIMUM RATINGS
SymbolRatingCommercial Unit
(2)
V
TERM
T
AOperating0 to +70°C
BIASTemperature–55 to +125°C
T
T
STGStorage–55 to +125°C
OUTDC Output50mA
I
NOTES:2943 tbl 04
1. Stresses greater than those listed under ABSOLUTE MAXIMUM
RATINGS may cause permanent damage to the device. This is a stress
rating only and functional operation of the device at these or any other
conditions above those indicated in the operational sections of this
specification is not implied. Exposure to absolute maximum rating
conditions for extended periods may affect reliability.
TERM must not exceed Vcc + 0.3V for more than 25% of the cycle time
2. V
or 10ns maximum, and is limited to
+ 0.3V.
Terminal Voltage–0.5 to +4.6V
with Respect
to GND
Temperature
Under Bias
Temperature
Current
< 20mA for the period of VTERM > Vcc
(1)
RECOMMENDED OPERATING
TEMPERATURE AND SUPPLY VOLTAGE
Ambient
GradeTemperatureGNDV
CC
Commercial0°C to +70°C0V3.3V ± 0.3V
2943 tbl 05
RECOMMENDED DC OPERATING
CONDITIONS
SymbolParameterMin.Typ.Max. Unit
CCSupply Voltage3.03.33.6V
V
GNDSupply Voltage000V
IHInput High Voltage2.0—VCC+0.3 V
V
ILInput Low Voltage–0.3
V
NOTES:2943 tbl 06
1. VIL > -1.5V for pulse width less than 10ns.
TERM must not exceed Vcc + 0.3V.
2. V
CAPACITANCE
(TA = +25°C, f = 1.0MHz)TQFP ONLY
SymbolParameterConditions
INInput CapacitanceVIN = 3dV9pF
C
OUTOutputVOUT = 3dV10pF
C
(2)
(1)
(1)
—0.8V
(2)
Max.Unit
Capacitance
NOTES:2943 tbl 07
1. This parameter is determined by device characterization but is not
production tested.
2. 3dV represents the interpolated capacitance when the input and output
signals switch from 0V to 3V or from 3V to 0V.
6.374
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
DC ELECTRICAL CHARACTERISTICS OVER THE
OPERATING TEMPERATURE AND SUPPLY VOLTAGE RANGE
AOEOutput Enable Access Time—15—20—30ns
OHOutput Hold from Address Change3—3—3—ns
t
t
LZOutput Low-Z Time
t
HZOutput High-Z Time
PUChip Enable to Power Up Time
t
t
PDChip Disable to Power Down Time
t
SOPSemaphore Flag Update Pulse (
SAASemaphore Address Access Time—35—45—65ns
t
NOTES:2943 tbl 11
1. Transition is measured ±200mV from Low or High-impedance voltage with Output Test Load (Figure 2).
2. This parameter is guaranteed by device characterization, but is not production tested.
3. To access RAM, CE = V
4. "X" in part numbers indicates power rating (S or L).
IL and
SEM
(3)
(1, 2)
(1, 2)
(2)
(2)
OE
or
SEM
)15—15—15—ns
= VIH. To access semaphore, CE = VIH and
—25—35 —55ns
3—3— 3—ns
—15—20 —25ns
0—0— 0—ns
—25—35 —50ns
SEM
= VIL.
(4)
TIMING OF POWER-UP POWER-DOWN
CE
t
I
CC
I
SB
PU
t
PD
50%50%
2943 drw 07
6.376
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
WAVEFORM OF READ CYCLES
(5)
t
RC
ADDR
(4)
t
AA
(4)
t
CE
ACE
t
AOE
(4)
OE
R/
W
t
(1)
t
LZ
DATA
OUT
BUSY
OUT
(3, 4)
t
BDD
NOTES:
1. Timing depends on which signal is asserted last, OE or CE.
2. Timing depends on which signal is de-asserted first, CE or OE.
3. t
BDD delay is required only in cases where the opposite port is completing a write operation to the same address location. For simultaneous read operations
BUSY
4. Start of valid data depends on which timing becomes effective last t
5.
has no relation to valid output data.
= V
IH.
SEM
AOE, tACE, tAA or tBDD.
VALID DATA
(4)
OH
(2)
t
HZ
2943 drw 08
AC ELECTRICAL CHARACTERISTICS OVER THE
OPERATING TEMPERATURE AND SUPPLY VOLTAGE
1. "X" in part numbers indicates power rating (S or L).
(1)
2942 tbl 14
6.3712
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
WAVEFORM OF INTERRUPT TIMING
ADDR
CE
R/
INT
ADDR
CE
W
"A"
"A"
"A"
"B"
"B"
t
"B"
INTERRUPT SET ADDRESS
(3)
t
AS
(3)
t
INS
INTERRUPT CLEAR ADDRESS
(3)
AS
(1)
t
t
WC
RC
(2)
(2)
t
WR
(4)
2943 drw 17
OE
"B"
(3)
t
INR
INT
"B"
2943 drw 18
NOTES:
1. All timing is the same for left and right ports. Port “A” may be either the left or right port. Port “B” is the port opposite from port “A”.
2. See Interrupt truth table.
3. Timing depends on which enable signal (CE or R/W) is asserted last.
4. Timing depends on which enable signal (CE or R/W) is de-asserted first.
TRUTH TABLES
TRUTH TABLE III — INTERRUPT FLAG
Left PortRight Port
CECE
WW
CE
L
R/
W
CECE
WW
LLX7FFFXXXXXL
XXXXXXLL7FFFH
XXXXL
XLL7FFEH
NOTES:2942 tbl 15
1. Assumes
2. If
3. If
BUSY
L = VIL, then no change.
BUSY
R = VIL, then no change.
BUSY
L
L =
OEOE
OE
OEOE
BUSY
L A14L-A0L
R =VIH.
INTINT
INT
INTINT
LR/
(3)
(2)
(1)
CECE
WW
CE
W
R
CECE
WW
LLX7FFEXSet Left
OEOE
OE
R
R A14R-A 0R
OEOE
INTINT
INT
RFunction
INTINT
(2)
Set Right
(3)
Reset Right
INT
XXXXXReset Left
INT
INT
R Flag
INT
L Flag
L Flag
R Flag
6.3713
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
TRUTH TABLE IV —
ADDRESS BUSYARBITRATION
InputsOutputs
0L-A14L
A
CECE
CECE
CE
CE
L
CECE
XX
HX
XH
LL
NOTES:2943 tbl 16
1. Pins
IDT7007 are push-pull, not open drain outputs. On slaves the
2. "L" if the inputs to the opposite port were stable prior to the address and enable inputs of this port. "H" if the inputs to the opposite port became stable
after the address and enable inputs of this port. If t
simultaneously.
3. Writes to the left port are internally ignored when
internally ignored when
R are both outputs when the part is configured as a master. Both are inputs when configured as a slave.
BUSY
TRUTH TABLE V — EXAMPLE OF SEMAPHORE PROCUREMENT SEQUENCE
FunctionsD0 - D7 LeftD0 - D7 RightStatus
No Action11Semaphore free
Left Port Writes "0" to Semaphore01Left port has semaphore token
Right Port Writes "0" to Semaphore01No change. Right side has no write access to semaphore
Left Port Writes "1" to Semaphore10Right port obtains semaphore token
Left Port Writes "0" to Semaphore10No change. Left port has no write access to semaphore
Right Port Writes "1" to Semaphore01Left port obtains semaphore token
Left Port Writes "1" to Semaphore11Semaphore free
Right Port Writes "0" to Semaphore10Right port has semaphore token
Right Port Writes "1" to Semaphore11Semaphore free
Left Port Writes "0" to Semaphore01Left port has semaphore token
Left Port Writes "1" to Semaphore11Semaphore free
(1)
BUSYBUSY
BUSY
L
BUSYBUSY
R outputs are driving LOW regardless of actual logic level on the pin.
(1)
BUSYBUSY
BUSY
R
BUSYBUSY
APS is not met, either
BUSY
Function
(3)
input internally inhibits writes.
BUSY
BUSY
L or
BUSY
R = LOW will result.
L outputs are driving LOW regardless of actual logic level on the pin. Writes to the right port are
BUSY
L and
BUSY
outputs on the
BUSY
R outputs can not be LOW
(1,2)
NOTES:2943 tbl 17
1. This table denotes a sequence of events for only one of the eight semaphores on the IDT70V07.
2. There are eight semaphore flags written to via I/O
0 and read from all I/O's (I/O0-I/O7). These eight semaphores are addressed by A0 - A2.
FUNCTIONAL DESCRIPTION
The IDT70V07 provides two ports with separate control,
address and I/O pins that permit independent access for reads
or writes to any location in memory. The IDT70V07 has an
automatic power down feature controlled by CE. The
controls on-chip power down circuitry that permits the
respective port to go into a standby mode when not selected
(CE HIGH). When a port is enabled, access to the entire
memory array is permitted.
7FFF location 7FFF. The message (8 bits) at 7FFE or 7FFF
is user-defined since it is an addressable SRAM location. If
the interrupt function is not used, address locations 7FFE and
7FFF are not used as mail boxes, but as part of the random
CE
access memory. Refer to Truth Table for the interrupt
operation.
BUSY LOGIC
INTERRUPTS
If the user chooses to use the interrupt function, a memory
location (mail box or message center) is assigned to each port.
The left port interrupt flag (
writes to memory location 7FFE (HEX), where a write is
defined as CE = R/W = VIL per the Truth Table. The left port
clears the interrupt through access of address location 7FFE
when
CE
R = OER = VIL, R/
right port interrupt flag (
writes to memory location 7FFF (HEX) and to clear the
interrupt flag (
INT
R), the right port must read the memory
INT
L) is asserted when the right port
W
is a "don't care". Likewise, the
INT
R) is asserted when the left port
Busy Logic provides a hardware indication that both ports
of the RAM have accessed the same location at the same
time. It also allows one of the two accesses to proceed and
signals the other side that the RAM is “Busy”. The busy pin can
then be used to stall the access until the operation on the other
side is completed. If a write operation has been attempted
from the side that receives a busy indication, the write signal
is gated internally to prevent the write from proceeding.
The use of busy logic is not required or desirable for all
applications. In some cases it may be useful to logically OR
the busy outputs together and use any busy indication as an
6.3714
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
interrupt source to flag the event of an illegal or illogical
operation. If the write inhibit function of busy logic is not
desirable, the busy logic can be disabled by placing the part
in slave mode with the M/S pin. Once in slave mode the
BUSY
pin operates solely as a write inhibit input pin. Normal operation can be programmed by tying the
BUSY
pins high. If
desired, unintended write operations can be prevented to a
port by tying the busy pin for that port low.
The busy outputs on the IDT 70V07 RAM in master mode,
are push-pull type outputs and do not require pull up resistors
to operate. If these RAMs are being expanded in depth, then
the busy indication for the resulting array requires the use of
an external AND gate.
WIDTH EXPANSION WITH BUSY LOGIC
MASTER/SLAVE ARRAYS
When expanding an IDT70V07 RAM array in width while
using busy logic, one master part is used to decide which side
of the RAM array will receive a busy indication, and to output
that indication. Any number of slaves to be addressed in the
MASTER
Dual Port
RAM
BUSY
L
MASTER
Dual Port
RAM
BUSY
BUSY
L
Figure 3. Busy and chip enable routing for both width and depth
L
expansion with IDT70V07 RAMs.
BUSY
BUSY
CE
CE
SLAVE
Dual Port
RAM
BUSY
R
R
L
SLAVE
Dual Port
RAM
BUSY
L
CE
BUSY
CE
BUSY
R
R
DECODER
BUSY
2943 drw 19
R
same address range as the master, use the busy signal as a
write inhibit signal. Thus on the IDT70V07 RAM the busy pin
is an output if the part is used as a master (M/S pin = H), and
the busy pin is an input if the part used as a slave (M/S pin =
L) as shown in Figure 3.
If two or more master parts were used when expanding in
width, a split decision could result with one master indicating
busy on one side of the array and another master indicating
busy on one other side of the array. This would inhibit the write
operations from one port for part of a word and inhibit the write
operations from the other port for the other part of the word.
The busy arbitration, on a master, is based on the chip
enable and address signals only. It ignores whether an access
is a read or write. In a master/slave array, both address and
chip enable must be valid long enough for a busy flag to be
output from the master before the actual write pulse can be
initiated with the R/W signal. Failure to observe this timing can
result in a glitched internal write inhibit signal and corrupted
data in the slave.
SEMAPHORES
The IDT70V07 is an extremely fast Dual-Port 32K 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 RAM 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 RAM. These devices have an automatic
power-down feature controlled by CE, the Dual-Port RAM
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 where
and
SEM
are both high.
CE
Systems which can best use the IDT70V07 contain multiple processors or controllers and are typically very highspeed systems which are software controlled or software
intensive. These systems can benefit from a performance
increase offered by the IDT70V07'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 IDT70V07 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 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 has set the
latch first, has the token and is using the shared resource. The
6.3715
IDT70V07S/L
D
2943 drw 20
0
D
Q
WRITE
D0
D
Q
WRITE
SEMAPHORE
REQUEST FLIP FLOP
SEMAPHORE
REQUEST FLIP FLOP
L PORT
R PORT
SEMAPHORE
READ
SEMAPHORE
READ
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
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 IDT70V07 in
a separate memory space from the Dual-Port RAM. This
address space is accessed by placing a low input on the
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 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 side
(see Table III). 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 discussing 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 (
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 (
to go inactive or the output will never change.
A sequence 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 one, a
fact which the processor will verify by the subsequent read
(see Table III). As an example, assume a processor writes a
zero to 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 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.
SEM
Figure 4. IDT70V07 Semaphore Logic
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 4. Two semaphore request latches feed into a
semaphore flag. Whichever latch is first to present a zero to
the semaphore flag will force its side of the semaphore flag
low and the other side high. This condition will continue until
a one is written to the same semaphore request latch. Should
the other side’s semaphore request latch have been written to
a zero in the meantime, the semaphore flag will flip over to the
other side as soon as a one is written into the first side’s
request latch. The second side’s flag will now stay low until its
semaphore request latch is written to a one. From this it is
easy to understand that, if a semaphore is requested and the
processor which requested it no longer needs the resource,
the entire system can hang up until a one is written into that
semaphore request latch.
The critical case of semaphore timing is when both sides
request a single token by attempting to write a zero into it at
the same time. The semaphore logic is specially designed to
resolve this problem. If simultaneous requests are made, the
SEM
) and output enable (OE) signals go
logic guarantees that only one side receives the token. If one
side is earlier than the other in making the request, the first
side to make the request will receive the token. If both
requests arrive at the same time, the assignment will be
SEM
or OE)
arbitrarily made to one port or the other.
One caution that should be noted when using semaphores
is that semaphores alone do not guarantee that access to a
resource is secure. As with any powerful programming technique, if semaphores are misused or misinterpreted, a software error can easily happen.
Initialization of the semaphores is not automatic and must
be handled via the initialization program at power-up. Since
any semaphore request flag which contains a zero must be
reset to a one, all semaphores on both sides should have a
one written into them at initialization from both sides to assure
that they will be free when needed.
USING SEMAPHORES—SOME EXAMPLES
Perhaps the simplest application of semaphores is their
application as resource markers for the IDT70V07’s Dual-Port
RAM. Say the 32K x 8 RAM was to be divided into two 16K
6.3716
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE
x 8 blocks which were to be dedicated at any one time to
servicing either the left or right port. Semaphore 0 could be
used to indicate the side which would control the lower section
of memory, and Semaphore 1 could be defined as the indicator for the upper section of memory.
To take a resource, in this example the lower 16K of
Dual-Port RAM, the processor on the left port could write and
then read a zero in to Semaphore 0. If this task were
successfully completed (a zero was read back rather than a
one), the left processor would assume control of the lower
16K. Meanwhile the right processor was attempting to gain
control of the resource after the left processor, it would read
back a one in response to the zero it had attempted to write into
Semaphore 0. At this point, the software could choose to try
and gain control of the second 16K section by writing, then
reading a zero into Semaphore 1. If it succeeded in gaining
control, it would lock out the left side.
Once the left side was finished with its task, it would write
a one to Semaphore 0 and may then try to gain access to
Semaphore 1. If Semaphore 1 was still occupied by the right
side, the left side could undo its semaphore request and
perform other tasks until it was able to write, then read a zero
into Semaphore 1. If the right processor performs a similar
task with Semaphore 0, this protocol would allow the two
processors to swap 16K blocks of Dual-Port RAM with each
other.
The blocks do not have to be any particular size and can
even be variable, depending upon the complexity of the
software using the semaphore flags. All eight semaphores
could be used to divide the Dual-Port RAM or other shared
resources into eight parts. Semaphores can even be assigned
different meanings on different sides rather than being given
a common meaning as was shown in the example above.
Semaphores are a useful form of arbitration in systems like
disk interfaces where the CPU must be locked out of a section
of memory during a transfer and the I/O device cannot tolerate
any wait states. With the use of semaphores, once the two
devices has determined which memory area was “off-limits” to
the CPU, both the CPU and the I/O devices could access their
assigned portions of memory continuously without any wait
states.
Semaphores are also useful in applications where no
memory “WAIT” state is available on one or both sides. Once
a semaphore handshake has been performed, both processors can access their assigned RAM segments at full speed.
Another application is in the area of complex data structures. In this case, block arbitration is very important. For this
application one processor may be responsible for building and
updating a data structure. The other processor then reads
and interprets that data structure. If the interpreting processor
reads an incomplete data structure, a major error condition
may exist. Therefore, some sort of arbitration must be used
between the two different processors. The building processor
arbitrates for the block, locks it and then is able to go in and
update the data structure. When the update is completed, the
data structure block is released. This allows the interpreting
processor to come back and read the complete data structure,
thereby guaranteeing a consistent data structure.
6.3717
IDT70V07S/L
HIGH-SPEED 3.3V 32K x 8 DUAL-PORT STATIC RAMCOMMERCIAL TEMPERATURE RANGE