Arm Mmu
Arm Mmu
11
7
7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.11 Introduction
MMU Program Accessible Registers Address Translation Translation Process Translating Section References Translating Small Page References Translating Large Page References MMU Faults and CPU Aborts Fault Address and Fault Status Registers (FAR and FSR) Fault Checking Sequence
7.10 Domain Access Control 7.12 Interaction of the MMU, IDC and Write Buffer 7.13 Effect of Reset
7-1
Preliminary
Preliminary
If the TLB contains a translated entry for the virtual address, the access control logic determines whether access is permitted. If access is permitted, the MMU outputs the appropriate physical address corresponding to the virtual address. If access is not permitted, the MMU signals the CPU to abort. If the TLB misses (it does not contain a translated entry for the virtual address), the translation table walk hardware is invoked to retrieve the translation information from a translation table in physical memory. Once retrieved, the translation information is placed into the TLB, possibly overwriting an existing value. The entry to be overwritten is chosen by cycling sequentially through the TLB locations. When the MMU is turned off (as happens on reset), the virtual address is output directly onto the physical address bus.
7-2
Register 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10
0 0 0 0 0
Control
Domain Access Control 9 8 7 6
0R S B 1 D P W C A M
0 0 0 0
Domain
Status
The registers not shown are reserved and should not be used. The Fault Status Register indicates the domain and type of access being attempted when an abort occurred. Bits 7:4 specify which of the sixteen domains (D15-D0) was being accessed when a fault occurred. Bits 3:1 indicate the type of access being attempted. The encoding of these bits is different for internal and external faults (as indicated by bit 0 in the register) and is shown in Table 7-4: Priority encoding of fault status on page 7-13. A write to this register ushes the TLB.
The Fault Address Register holds the virtual address of the access which was attempted when a fault occurred. A write to this register causes the data written to be treated as an address and, if it is found in the TLB, the entry is marked as invalid. (This operation is known as a TLB purge). The Fault Status Register and Fault Address Register are only updated for data faults, not for prefetch faults.
7-3
Preliminary
Preliminary
7-4
Table Index
Section Index
Translation Base
12 18
31 14 13 2 1 0
Translation Base
Table Index
0 0
7-5
Preliminary
Domain 1 C B 1 0 1 1
Preliminary
Value 00 01 10 11
Notes Generates a Section Translation Fault Indicates that this is a Page Descriptor Indicates that this is a Section Descriptor Reserved for future use
7-6
7-7
Preliminary
Virtual Address
31 20 19 0
Table Index
Section Index
Translation Base
Preliminary
12 18
31 14 13 2 1 0
Translation Base
Table Index
0 0
AP
Domain 1 C B 1 0
20
12
31
Physical Address
20 19 0
Section Index
7-8
0 0 Large Page Base Address Small Page Base Address ap3 ap2 ap1 ap0 C B 0 1 ap3 ap2 ap1 ap0 C B 1 0 1 1
Value 00 01 10 11
Notes Generates a Page Translation Fault Indicates that this is a 64 kB Page Indicates that this is a 4 kB Page Reserved for future use
7-9
Preliminary
The two least signicant bits indicate the page size and validity, and are interpreted as follows.
Table Index
12
L2 Table Index
8
Page Index
12
Translation Base
Preliminary
18
31 14 13 2 1 0
Translation Base
Table Index
0 0
Domain
0 1
31
10 9
L2 Table Index
0 0
Physical Address
31 12 11 0
Page Index
Table Index
12
L2 Table Index
8
Page Index
12
Translation Base
18
31 14 13 2 1 0
Translation Base
Table Index
0 0
Domain
0 1
31
10 9
L2 Table Index
0 0
Physical Address
31 16 15 0
Page Index
7-11
Preliminary
Preliminary
7-12
Highest
Alignment Bus Error (translation) Translation Domain Permission Bus Error (linefetch) level1 level2 Section Page Section Page Section Page Section Page Section Page
00x1 1100 1110 0101 0111 1001 1011 1101 1111 0100 0110 1000 1010
x x valid Note 2 valid valid valid valid valid valid valid valid valid
valid valid valid valid valid valid valid valid valid valid valid valid valid
Lowest
7-13
Preliminary
Source
FS[3210]
Domain[3:0]
FAR
31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10
15
14
13
12
11
10
Preliminary
00 01 10 11
7-14
Virtual Address
misaligned
Alignment Fault
invalid
invalid
no access(00) reserved(10)
no access(00) reserved(10)
client(01)
client(01)
violation
violation
Physical Address
Figure 7-10: Sequence for checking faults
7-15
Preliminary
Preliminary
There are two types of domain fault: section and page. In both cases the Level One descriptor holds the 4-bit Domain eld which selects one of the sixteen 2-bit domains in the Domain Access Control Register. The two bits of the specied domain are then checked for access permissions as detailed in Table 7-2: Interpreting access permission (AP) bits on page 7-7. In the case of a section, the domain is checked once the Level One descriptor is returned, and in the case of a page, the domain is checked once the Page Table Entry is returned. If the specied access is either No Access (00) or Reserved (10) then either a Section Domain Fault or Page Domain Fault occurs.
7-16
7-17
Preliminary
Preliminary
To enable the MMU: 1 2 3 Note Program the Translation Table Base and Domain Access Control Registers Program Level 1 and Level 2 page tables as required Enable the MMU by setting bit 0 in the Control Register.
Care must be taken if the translated address differs from the untranslated address as the two instructions following the enabling of the MMU will have been fetched using flat translation and enabling the MMU may be considered as a branch with delayed execution. A similar situation occurs when the MMU is disabled. Consider the following code sequence:
MOV R1, #0x1 MCR 15,0,R1,0,0 Fetch Flat Fetch Flat Fetch Translated To disable the MMU: 1 2 Disable the WB by clearing bit 3 in the Control Register. Disable the IDC by clearing bit 2 in the Control Register.
; Enable MMU
Note
3 Disable the MMU by clearing bit 0 in the Control Register. If the MMU is enabled, then disabled and subsequently re-enabled the contents of the TLB will have been preserved. If these are now invalid, the TLB should be flushed before re-enabling the MMU. Disabling of all three functions may be done simultaneously.
7-18
7-19
Preliminary
Preliminary
7-20
11
8
8.1 8.2 8.3
8-1
Preliminary
Preliminary
8-2
Port E data register. Port E data direction register. System control register System status flags. Expansion and ROM memory configuration register 1. Expansion and ROM memory configuration register 2. DRAM refresh period register. Interrupt status register. Interrupt mask register. LCD control register. Read-write data to TC1. Read-write data to TC2. Real time clock data register. Real time clock match register. DC to DC pump control register.
8-3
Preliminary
Preliminary
8-4
8-5
Preliminary
Preliminary
Bits set in this 4-bit read-write register will select the corresponding pin in port E to become an output, clearing bit sets the pin to input. All bits are cleared by a system reset so that port E is input by default.
TC2S
TC2M
TCIS
TCIM
Keyboard scan
15
14
13
12
11
10
SIREN
23
CDENRX CDENTX
22 21
LCDEN
20
DBGEN
19
BZMOD
18
BZTOG
17
UARTEN
16
IRTXM WAKEDIS
EXCKEN
ADCKSEL
8-6
Column All driven HIGH All driven LOW All Tristate Column 0 only driven HIGH Column 1 only driven HIGH Column 2 only driven HIGH Column 3 only driven HIGH Column 4 only driven HIGH Column 5 only driven HIGH Column 6 only driven HIGH Column 7 only driven HIGH
TC2M TC2S
Microwire / SPI peripheral clock speed select. This 2-bit eld selects the frequency of the ADC sample clock. This is twice the frequency of the synchronous serial ADC interface clock. Table 8-3: ADCCLK frequencies shows the available frequencies.
8-7
Preliminary
ADCKSEL 00 01 10 11
DBGEN
Table 8-3: ADCCLK frequencies Setting this bit enables debug/broadcast mode. In this mode, all internal accesses are output as if they were reads or writes to expansion memory addressed by CS6. CS6 will still be active in its standard address range. In addition the internal interrupt request and the fast interrupt request signals to the processor are output on port E bits 1 and 2 ie. in debug mode:
CS6 = CS6 / internal strobe PE1 = nIRQ PE2 = nFIQ
Preliminary
LCDEN CDENTX
Setting this bit enables the LCD controller. CODEC interface enable Tx bit. Setting this bit enables the CODEC interface for data transmission to an external CODEC device. CODEC interface enable Rx bit. Setting this bit enables the CODEC interface for data reception from an external CODEC device. SIR protocol encoding enable bit. This has no effect if the UART is not enabled. External expansion clock enable. If this bit is set the EXPCLK is enabled continuously with the same speed and phase as the CPU clock and will free run all the time the main oscillator is running. This bit should not be left set all the time for power consumption reasons. If the system enters the standby state the EXPCLK will become undened. If this bit is clear EXPCLK will be active during memory cycles to expansion slots that have external wait state generation enabled only. Switch on via the wakeup input is disabled if this bit is set. IrDA Tx mode bit. This bit controls the IrDA encoding strategy. Clearing it means each zero bit transmitted is represented as a pulse of width 3/16th of the bit rate period. Setting this bit means each zero bit is represented as a pulse of width 3/16th of the period of 115,000 bit rate clock ie. 1.6Sec regardless of the selected bit rate. Setting this bit will use less power but probably reduce transmission distances. Reserved. Write will have no effect, will always read zero.
CDENRX
SIREN EXCLKEN
WAKEDIS IRTXM
DID
15 14 13 12
WUON
11
WUDR
10
DCDET
9
MCDR
8
CLDFLG
23
PFFLG
22
RSTFLG
21
NBFLG
UBUSY
DCD
DSR
CTS
16
UTXFF
31
UTXFE
30 29 28
RTCDIV
27 26 25 24
VERID
CTXFF
CRXFE
DID
8-9
Preliminary
NBFLG
RSTFLG PFFLG
CLDFLG
RTCDIV
Preliminary
URXFE
UTXFF
8-10
31
24
23
16
15
nCS3 conguration
nCS2 conguration
nCS1 conguration
nCS0 conguration
The memory conguration register 2 is a 32-bit read-write register which sets the conguration of the four expansion and ROM selects CS[4:7]. Each select is congured with a one byte eld, starting with expansion select 4.
31
24
23
16
15
CS7 conguration
CS6 conguration
CS5 conguration
CS4 conguration
CLKEN
SQAEN
Bus Width
8-11
Preliminary
Preliminary
16-bit I/O accesses are not converted to 32-bit ARM word accesses. This means that D[16:31] will be invalid during ARM word accesses to this memory area.
A26 0 1 0 1 A27 0 0 1 1 Bus Width 8 Bits 16 Bits 8 Bits 16 Bits Word Bus Conversion Yes Yes Yes No PCMCIA Memory Area 8-bit attribute memory access 16-bit common memory access 8-bit I/O access 16-bit I/O access
8-12
Table 8-7: Values of the page mode access wait state eld
SQAEN Sequential access enable. Setting this bit enables sequential accesses that are on a quad word boundary to take advantage of faster access times from devices that support page mode. The sequential access will be faulted after four words, (to allow video refresh cycles to occur) even if the access is part of a longer sequential access. Expansion clock enable. Setting this bit enables the EXPCLK to be active during accesses to the selected expansion device. This provides a timing reference for devices that need to extend bus cycles using the EXPRDY input. Back to back (but not necessarily page mode) accesses result in a continuous clock.
CLKEN
For more details on bus timing, refer to Chapter 20, DC and AC Parameters.
8-13
Preliminary
RFSHEN
RFDIV
RFSHEN
DRAM refresh enable. Setting this bit enables periodic refresh cycles to be generated by ARM7100 at a rate set by the RFDIV eld. Setting this bit also enables self refresh mode when ARM7100 is in the standby state. This 7-bit eld sets the DRAM refresh rate. The refresh period is derived from a 128 KHz clock and is given by the following formula: Frequency (KHz) = 128/(RFDIV + 1) or RFDIV = (128/Refresh frequency (KHz) ) - 1
RFDIV
The maximum refresh frequency is 64 KHz, the minimum is 1KHz. The RFDIV eld should not be programmed with zero as this results in no refresh cycles being initiated.
Preliminary
EINT3
EINT2
EINT1
CSINT
MCINT
WEINT
BLINT
EXTFIQ
15
14
13
12
11
10
SSEOTI
UMSINT
URXINT
UTXINT
TINT
RTCMI
TC2OI
TC1OI
BLINT
8-14
MCINT
CSINT
EINT1
EINT2
EINT3
TC1OI
TC2OI
RTCMI
TINT
8-15
Preliminary
URXINT
UMSINT
Preliminary
SSEOTI
EINT3
EINT2
EINT1
CSINT
MCINT
WEINT
BLINT
EXTFIQ
15
14
13
12
11
10
SSEOTI
UMSINT
URXINT
UTXINT
TINT
RTCMI
TC2OI
TC1OI
8-16
31
30
29
25 24
19 18
13
12
GSMD
GSEN
Line length
The video buffer size eld is a 13-bit eld that sets the total number of bytes (*128 quad words) in the video display buffer. This is calculated from the following formula:
Video buffer size = (Total bytes in video buffer/128) - 1 For example, for a 640 x 240 LCD and 4 bits per pixel the size of the video buffer = 640 x 240 x 4 = 614400 bits. video buffer size eld = (614400/128)-1 = 4799 or 0x12BF Hex Line length
Line length
= ( No. pixels in line/16 ) - 1 For example: 640 x 240 LCD line length = (640/16)-1 = 39 or 0x27 Hex
Pixel prescale
The pixel prescale eld is a 6-bit number that sets the pixel rate prescale. The pixel rate is derived from a 36.864 MHz clock and is calculated from the following formula:
Pixel rate (MHz) = 36.864/ (Pixel prescale + 1) The pixel rate should be chosen to give a complete screen refresh frequency of approximately 70 Hz to avoid icker. Frequencies above 70 Hz should be avoided as this consumes additional power. The pixel prescale value can be expressed in terms of the LCD size by the following formula: Pixel prescale = (526628/Total pixels in display)-1 The value should be rounded down to the nearest whole number, and zero is illegal and results in no pixel clock. For example: 640 x 240 LCD pixel prescale Actual pixel rate Actual refresh frequency = 526628/(640x240)-1 = 2.428(2) = 36.864E6/2+1 =12.288MHz = 12.288E6/(640x240) = 80Hz 8-17
Preliminary
The line length eld is a 6-bit eld that sets the number of pixels in one complete line. This eld is calculated from the formula:
The AC prescale eld is a 5-bit number that sets LCD AC bias frequency. This frequency is the required AC bias frequency for a given manufacturers LCD plate. It is derived from the frequency of the line clock (CL1). The M signal toggles after n + 1 counts of the line clock (CL1) where n is the number programmed into the AC prescale eld. This number must be chosen to match the manufacturers recommendation. This is normally 13 but must not be exactly divisible by the number of lines in the display.
GSEN
Grey scale enable bit. Setting this bit enables grey scale output to the LCD. When it is cleared, each bit in the video map directly corresponds to a pixel in the display. Grey scale mode bit. Clearing this bit sets the controller to 2 bits per pixel (4 grey scales). Setting it sets it to 4 bits per pixel (15 grey scales).
GSMD
Preliminary
VL pump ratio
Vh from battery
This 4-bit eld controls the on time for the DC to DC pump for a Vh rail while the nEXTPWR input is HIGH. Setting these bits to 0 disables this pump. Setting them to 1 allows the pump to be driven in a 1:16 duty ratio, 2 in a 2:16 duty ratio etc. up to a 15:16 duty ratio. An 8:16 duty ratio results in a square wave of 96 KHz. This 4-bit eld controls the on time for the DC to DC pump for a Vh rail while the nEXTPWR input is LOW. Setting these bits to 0 disables this pump. Setting them to 1 allows the pump to be driven in a 1:16 duty ratio, 2 in a 2:16 duty ratio etc. up to a 15:16 duty ratio. An 8:16 duty ratio results in a square wave of 96 KHz. This 4-bit eld controls the on time for the DC to DC pump for the VL voltage rail. Setting these bits to 0 disables this pump. Setting them to 1 allows the pump to be driven in a 1:16 duty ratio, 2 in a 2:16 duty ratio etc. up to a 15:16 duty ratio. An 8:16 duty ratio results in a square wave of 96 KHz. The state of the output drive pin (drive 1) is latched during power on reset, this latched value is used to determine the polarity of the bias voltage. The sense of the DC to DC converter control lines is summarised in Table 8-8: Sense of DC to DC Converter Control Lines.
Vh from mains
VL pump ratio
8-19
Preliminary
10
OVERR
PARERR
FRMERR
Rx data
FRMERR
Preliminary
UART framing error. This bit is set if the UART detected an overrun or framing error while receiving the Rx data byte. UART parity error. This bit is set if the UART detected a parity error while receiving the Rx data byte. UART overrun error. This bit is set if more data is received by the UART and the FIFO is full. The overrun error bit is not associated with any single character and so is not stored in the FIFO. If this bit is set, the entire contents of the FIFO is invalid and should be cleared. This error bit is cleared by reading the UARTDR register.
PARERR OVERR
31
1 18
17
16
15
14
13
12
11
WRDLEN UFIFOEN
XSTOP
EVENPRT
PRTEN
BREAK
This 12-bit eld sets the bit rate. The bit rate divider is fed by a clock frequency of 3.6864 MHz. It is then further divided internally by 16 to give the bit rate. The following formula gives the divisor value for any bit rate: Divisor = (230400/bit rate)-1 A value of zero in this eld is illegal.
8-20
Setting this bit drives the Tx output active (HIGH) to generate a break. Parity enable bit. Setting this bit enables parity detection and generation. Even parity bit. Setting this bit sets parity generation and checking to even parity, clearing it sets odd parity. This bit has no effect if the PRTEN bit is clear. Extra stop bit. Setting this bit will cause the UART to transmit two stop bits. Clearing it sets one stop bit after each data byte. Set to enable FIFO buffering of Rx and Tx data. Clear to disable the FIFO, ie. set its depth to one byte. This two bit eld selects the word length according to Table 8-10: UART word length on page 8-21.
WRDLEN 00 01 10 11 Word length 5 bits 6 bits 7 bits 8 bits
XSTOP
FIFOEN WRDLEN
8-21
Preliminary
31 - 28
27 - 24
23 - 20
19 - 16
15 - 12
11 - 8
7-4
3-0
Preliminary
31 - 28
27 - 24
23 - 20
19 - 16
15 - 12
11 - 8
7-4
3-0
8-22
Duty cycle 0 1/9 1/5 4/15 3/9 2/5 4/9 1/2 1/2 5/9 3/5 6/9 11/15 4/5 8/9 1
% pixels lit 0 11.1 20 26.7 33.3 40 44.4 50 50 55.6 60 66.7 73.3 80 88.9 100
8-23
Preliminary
Reserved
TXFRMEN
SMCKEN
Frame length
SMCKEN
TXFRMEN
Preliminary
8-25
Preliminary
Preliminary
8-26
11
9
9.1 Interrupt Controller
Interrupt Controller
This chapter describes the interrupt controller. 9-2
9-1
Preliminary
Interrupt Controller
9.1 Interrupt Controller
The ARM 710a has two interrupt types: interrupt request (IRQ)
fast interrupt request (FIQ) The interrupt controller in ARM7100 controls interrupts from 16 different sources. Twelve interrupt sources are mapped to the IRQ input and four sources to the FIQ input. FIQs have a higher priority than IRQs and if two interrupts at the same priority are active, the priority they are serviced in must be resolved in software. All interrupts are level sensitive, ie. they must conform to the following sequence: 1 2 3 4 The device asserts the appropriate interrupt request line. If the appropriate bit is set in the interrupt mask register, either FIQ or IRQ is asserted by the interrupt controller. If interrupts are enabled, the processor jumps to the appropriate vector. Interrupt despatch software reads the interrupt control and status register to establish the source(s) of the interrupt and calls the appropriate interrupt service routine(s). Software in the interrupt service routine clears the interrupt source by an action specic to the device requesting the interrupt, eg. reading the UART Rx register. The interrupt service routine may then re-enable interrupts and any other pending interrupts will be serviced in a similar way, or return to the interrupt dispatch code which can check for any more pending interrupts and dispatch them accordingly.
Preliminary
9-2
Interrupt Controller
Table 9-1: Interrupt allocation shows the names and allocation of interrupts in ARM7100.
Interrupt FIQ FIQ FIQ FIQ IRQ IRQ IRQ IRQ IRQ IRQ IRQ IRQ IRQ IRQ IRQ IRQ Bit in Mask and ISR 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Name EXTFIQ BLINT WEINT MCINT CSINT EINT1 EINT2 EINT3 TC1OI TC2OI RTCMI TINT UTXINT URXINT UMSINT SSEOTI Comment External fast interrupt input Battery low interrupt Watch dog expired interrupt MEDCHG interrupt CODEC sound interrupt External interrupt input 1 External interrupt input 2 External interrupt input 3 TC1 under flow interrupt TC2 under flow interrupt
64 Hz tick interrupt Internal UART transmit FIFO empty interrupt Internal UART receive FIFO full interrupt Internal UART modem status changed interrupt Synchronous serial interface end of transfer interrupt
9-3
Preliminary
Interrupt Controller
Preliminary
9-4
11
10
10-1
Preliminary
execute from 0 to 4 wait states. In addition, bus cycles can be extended using the EXPRDY input signal. Page mode access is accomplished by running up to four accesses together. This can signicantly improve bus bandwidth to devices such as ROMs. Sequential burst mode access is always faulted (the bus returned to idle) after four accesses regardless of bus width to allow DMA and refresh cycles. See Chapter 8, ARM7100 Programmers Model for details of the expansion and ROM interface registers.
Preliminary
10-2
11
11
11.1 DRAM Controller
DRAM controller
This chapter describes the DRAM controller. 11-2
11-1
Preliminary
DRAM controller
11.1 DRAM Controller
The DRAM controller in ARM7100 provides connections allowing a direct interface to up to four banks of DRAM. Each bank is 32 bits wide and up to 256 Mb in size. Four RAS lines are provided (one per bank) and four CAS lines (one per byte line). The DRAM device size is not programmable if devices smaller than the largest size supported (1 Gbit) are used. This leads to a segmented memory map with each bank separated by 256 MBytes. Segments that are smaller than the bank size will repeat within the bank. Table 11-1: Physical to DRAM address mapping shows the mapping of physical address to DRAM row and column address. This mapping has been organised to support any DRAM device size from 4 Mbit to 1 Gbit with a square row and column conguration, ie. the number of column addresses is equal to the number of row addresses. If a non-square DRAM is used, further fragmentation of the memory map will occur. However the smallest contiguous segment will always be 1 Mb.
Memory Address 0
DRAM Row A10 A11 A12 A13 A14 A15 A16 A17 A18 A20 A22 A24 A26
Pin Name A[27]/DRA[0] A[26]/DRA[1] A[25]/DRA[2] A[24]/DRA[3] A[23]/DRA[4] A[22]/DRA[5] A[21]/DRA[6] A[20]/DRA[7] A[19]/DRA[8] A[18]/DRA[9] A[17]/DRA[10] A[16]/DRA[11] A[15]/DRA[12]
Preliminary
1 2 3 4 5 6 7 8 9 10 11 12
11-2
DRAM controller
Address Range of Segment(s) n000.0000 - n00F.FFFF n000.0000 - n03F.FFFF n000.0000 - n003.FFFF n010.0000 - n013.FFFF n040.0000 - n043.FFFF n050.0000 - n053.FFFF n100.0000 - n103.FFFF n110.0000 - n113.FFFF n140.0000 - n143.FFFF n150.0000 - n153.FFFF n400.0000 - n403.FFFF n410.0000 - n413.FFFF n440.0000 - n443.FFFF n450.0000 - n453.FFFF n500.0000 - n503.FFFF n510.0000 - n513.FFFF n540.0000 - n543.FFFF n550.0000 - n553.FFFF n000.0000 - n0FF.FFFF n000.0000 - n00F.FFFF n020.0000 - n02F.FFFF n080.0000 - n08F.FFFF n0A0.0000 - n0AF.FFFF n200.0000 - n20F.FFFF n220.0000 - n22F.FFFF n280.0000 - n28F.FFFF n2A0.0000 - n2AF.FFFF n800.0000 - n80F.FFFF n820.0000 - n82F.FFFF n880.0000 - n88F.FFFF n8A0.0000 - n8AF.FFFF nA00.0000 - nA0F.FFFF nA20.0000 - nA2F.FFFF nA80.0000 - nA8F.FFFF nAA0.0000 - nAAF.FFFF n000.0000 - n3FF.FFFF n000.0000 - nFFF.FFFF
Size of Segment(s) 1 Mbyte 4 MBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 256 KBytes 16 MBytes 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 1 MByte 64 MBytes 256 MBytes
64 Mbit 64 Mbit
16 MBytes 16 MBytes
11-3
Preliminary
DRAM controller
Preliminary
11-4