bdiGDB enhances the GNU debugger (GDB), with JTAG/COP debugging for QorIQ P4 based targets. With the built-in Ethernet interface you get a very fast code download speed. No target communication channel (e.g. serial line) is wasted for debugging purposes. Even better, you can use fast
Ethernet debugging with target systems without network capability. The host to BDI communication
uses the standard GDB remote protocol.
An additional Telnet interface is available for special debug tasks (e.g. force a hardware reset,
program flash memory).
The following figure shows how the BDI3000 interface is connected between the host and the target:
Target System
P4080
COP Interface
BDI3000
GNU Debugger
(GDB)
Ethernet (10/100 BASE-T)
1.1 BDI3000
The BDI3000 is the main part of the bdiGDB system. This small box implements the interface between the JTAG pins of the target CPU and a 10/100Base-T Ethernet connector. The firmware of the
BDI3000 can be updated by the user with a simple Linux/Windows configuration program or interactively via Telnet/TFTP. The BDI3000 supports 1.2 – 5.0 Volts target systems.
As an initial setup, the IP address of the BDI3000, the IP address of the host with the configuration
file and the name of the configuration file is stored within the flash of the BDI3000.
Every time the BDI3000 is powered on, it reads the configuration file via TFTP.
Following an example of a typical configuration file:
; Release cores for booting
WM32 0xfe0E00E4 0x00000003 ;BRR: release core 0 and 1
;
;
[TARGET]
; common parameters
POWERUP 5000 ;start delay after power-up detected in ms
JTAGCLOCK 1 ;use 16 MHz JTAG clock
WAKEUP 1000 ;give reset time to complete
;
; CoreID#0 parameters (active vCPU after reset)
#0 CPUTYPE P4080 0 0 ;Core0 / SOC0
#0 STARTUP HALT ;halt at the reset vector (this halts all cores !!!)
;
; CoreID#1 parameters
#1 CPUTYPE P4080 1 0 ;Core1 / SOC0
#1 STARTUP HALT ;halt at the reset vector
;
[HOST]
IP 151.120.25.112
FILE E:\temp\dump1024k.bin
FORMAT BIN 0x10000
;
#0 PROMPT P4080#0>
#1 PROMPT P4080#1>
;
[FLASH]
; only to test execution of target code
WORKSPACE 0x80001000 ;workspace in CPC1/SRAM
CHIPTYPE AM29BX16 ;Flash type
CHIPSIZE 0x00200000 ;The size of one flash chip in bytes
BUSWIDTH 16 ;The width of the flash memory bus in bits
FILE E:\temp\dump16k.bin
FORMAT BIN 0x00300000
[REGS]
FILE $regP4080.def
Based on the information in the configuration file, the target is automatically initialized after every reset.
The cable to the target system is a 16 pin flat ribbon cable. In case where the target system has an
appropriate connector, the cable can be directly connected. The pin assignment is in accordance with
the COP connector specification.
!
In order to ensure reliable operation of the BDI (EMC, runtimes, etc.) the target cable length must not
exceed 20 cm (8").
Target System
P4080
1
15
COP/JTAG Connector
1 - TDO
2 - NC (
3 - TDI
4 -
5 - NC (
6 - Vcc Target
7 - TCK
8 -
9 - TMS
10 - NC
11 12 - GROUND
13 -
14 - NC (key)
15 -
16 - GROUND
QACK)
TRST
QREQ)
CHKSTP_IN
SRESET
HRESET
CHKSTP_OUT
BDI3000
2
TARGET A TARGET B
BDI
TRGT
The green LED «TRGT» marked light up when target is powered up
MODE
16
151
16
2
For BDI TARGET B connector signals see table on next page.
Note:
For critical designs (long traces on the target board) there is a shorter target cable available (p/n
90020-S).This may improve JTAG communication reliability. But best is to keep the JTAG traces on
the board as short as possible.
Warning:
Before you can use the BDI3000 with an other target processor type (e.g. PPC <--> ARM), a new
setup has to be done (see chapter 2.5). During this process the target cable must be disconnected
from the target system.
!
To avoid conflicts between data lines, the BDI3000 must be disconnected from the target system while programming a new firmware for an other target CPU.
This input to the BDI3000 connects to the target TDO pin.
2IO2General purpose I/O
Currently not used.
3TDIJTAG Test Data In
This output of the BDI3000 connects to the target TDI pin.
4
TRSTJTAG Test Reset
This output of the BDI3000 resets the JTAG TAP controller on the target.
5IN0General purpose Input
Currently not used.
6Vcc Target1.2 – 5.0V:
This is the target reference voltage. It indicates that the target has power and it is also used
to create the logic-level reference for the input comparators. It also controls the output logic
levels to the target. It is normally connected to Vdd I/O on the target board.
7TCKJTAG Test Clock
This output of the BDI3000 connects to the target TCK pin.
8IO8General purpose I/O
This output of the BDI3000 connects to the target CKSTP_IN pin. Currently not used.
9TMSJTAG Test Mode Select
This output of the BDI3000 connects to the target TMS line.
10IO10General purpose I/O
Currently not used.
11
12GROUNDSystem Ground
13
14<reseved>
15IN1General purpose Input
16GROUNDSystem Ground
SRESETSoft-Reset
This open collector output of the BDI3000 connects to the target HRESET pin.
HRESETHard-Reset
This open collector output of the BDI3000 connects to the target PORESET pin.
This input to the BDI3000 connects to the target CKSTP_OUT pin. Currently not used.
The BDI3000 needs to be supplied with the enclosed power supply from Abatron (5VDC).
!
Before use, check if the mains voltage is in accordance with the input voltage printed on power
supply. Make sure that, while operating, the power supply is not covered up and not situated near
a heater or in direct sun light. Dry location use only.
!
For error-free operation, the power supply to the BDI3000 must be between 4.75V and 5.25V DC.
The maximal tolerable supply voltage is 5.25 VDC. Any higher voltage or a wrong polarity
might destroy the electronics.
+5 VDC GND
RS232POWER
casing connected to ground terminal
TARGET A TARGET B
BDI
TRGT
The green LED «BDI» marked light up when 5V power is connected to the BDI3000
MODE
Please switch on the system in the following sequence:
Serial line communication is only used for the initial configuration of the bdiGDB system.
The host is connected to the BDI through the serial interface (COM1...COM4). The communication
cable (included) between BDI and Host is a serial cable. There is the same connector pinout for the
BDI and for the Host side (Refer to Figure below).
The BDI3000 has a built-in 10/100 BASE-T Ethernet interface (see figure below). Connect an UTP
(Unshielded Twisted Pair) cable to the BD3000. Contact your network administrator if you have questions about the network.
Target System
10/100 BASE-T
18
Connector
1 - TD+
2 - TD 3 - RD+
4 - NC
5 - NC
6 - RD-
7 - NC
8 - NC
RS232 POWER
PC / Unix
Host
The following explains the meanings of the built-in LED lights:
LED1 LED2
Ethernet (10/100 BASE-T)
P4080
BDI3000
LEDFunctionDescription
LED 1
(green)
LED 2
(amber)
Link / ActivityWhen this LED light is ON, data link is successful between the UTP port
of the BDI3000 and the hub to which it is connected.
The LED blinks when the BDI3000 is receiving or transmitting data.
SpeedWhen this LED light is ON, 100Mb/s mode is selected (default).
When this LED light is OFF, 10Mb/s mode is selected
On the enclosed diskette you will find the BDI configuration software and the firmware required for
the BDI3000. For Windows users there is also a TFTP server included.
The following files are on the diskette.
b30qp4gd.exeWindows Configuration program
b30qp4gd.xxxFirmware for the BDI3000
tftpsrv.exeTFTP server for Windows (WIN32 console application)
*.cfgConfiguration files
*.defRegister definition files
bdisetup.zipZIP Archive with the Setup Tool sources for Linux / UNIX hosts.
Overview of an installation / configuration process:
• Create a new directory on your hard disk
• Copy the entire contents of the enclosed diskette into this directory
• Linux only: extract the setup tool sources and build the setup tool
• Use the setup tool or Telnet (default IP) to load/update the BDI firmware
Note: A new BDI has no firmware loaded.
• Use the setup tool or Telnet (default IP) to load the initial configuration parameters
- IP address of the BDI.
- IP address of the host with the configuration file.
- Name of the configuration file. This file is accessed via TFTP.
The BDI can get the network configuration and the name of the configuration file also via BOOTP.
For this simple enter 0.0.0.0 as the BDI’s IP address (see following chapters). If present, the subnet
mask and the default gateway (router) is taken from the BOOTP vendor-specific field as defined in
RFC 1533.
With the Linux setup tool, simply use the default parameters for the -c option:
The MAC address is derived from the serial number as follows:
MAC: 00-0C-01-xx-xx-xx , replace the xx-xx-xx with the 6 left digits of the serial number
Example: SN# 33123407 ==>> 00-0C-01-33-12-34
Default IP: 192.168.53.72
Before the BDI is configured the first time, it has a default IP of 192.168.53.72 that allows an initial
configuration via Ethernet (Telnet or Setup Tools). If your host is not able to connect to this default
IP, then the initial configuration has to be done via the serial connection.
The firmware update and the initial configuration of the BDI3000 is done with a command line utility.
In the ZIP Archive bdisetup.zip are all sources to build this utility. More information about this utility
can be found at the top in the bdisetup.c source file. There is also a make file included.
Starting the tool without any parameter displays information about the syntax and parameters.
!
To avoid data line conflicts, the BDI3000 must be disconnected from the target system while
programming the firmware for an other target CPU family.
Following the steps to bring-up a new BDI3000:
1. Build the setup tool:
The setup tool is delivered only as source files. This allows to build the tool on any Linux / Unix host.
To build the tool, simply start the make utility.
[root@LINUX_1 bdisetup]# make
cc -O2 -c -o bdisetup.o bdisetup.c
cc -O2 -c -o bdicnf.o bdicnf.c
cc -O2 -c -o bdidll.o bdidll.c
cc -s bdisetup.o bdicnf.o bdidll.o -o bdisetup
2. Check the serial connection to the BDI:
With "bdisetup -v" you may check the serial connection to the BDI. The BDI will respond with information about the current loaded firmware and network configuration.
Note: Login as root, otherwise you probably have no access to the serial port.
$ ./bdisetup -v -p/dev/ttyS0 -b115
BDI Type : BDI3000 (SN: 30000154)
Loader : V1.00
Firmware : unknown
MAC : ff-ff-ff-ff-ff-ff
IP Addr : 255.255.255.255
Subnet : 255.255.255.255
Gateway : 255.255.255.255
Host IP : 255.255.255.255
Config : яяяяяяя........
3. Load/Update the BDI firmware:
With "bdisetup -u" the firmware is programmed into the BDI3000 flash memory. This configures the
BDI for the target you are using. Based on the parameters -a and -t, the tool selects the correct firmware file. If the firmware file is in the same directory as the setup tool, there is no need to enter a -d
parameter.
$ ./bdisetup -u -p/dev/ttyS0 -b115 -aGDB -tP4080
Connecting to BDI loader
Programming firmware with ./b30qp4gd.100
With "bdisetup -c" the configuration parameters are written to the flash memory within the BDI.
The following parameters are used to configure the BDI:
BDI IP AddressThe IP address for the BDI3000. Ask your network administrator for as-
signing an IP address to this BDI3000. Every BDI3000 in your network
needs a different IP address.
Subnet MaskThe subnet mask of the network where the BDI is connected to. A subnet
mask of 255.255.255.255 disables the gateway feature. Ask your network
administrator for the correct subnet mask. If the BDI and the host are in
the same subnet, it is not necessary to enter a subnet mask.
Default GatewayEnter the IP address of the default gateway. Ask your network administra-
tor for the correct gateway IP address. If the gateway feature is disabled,
you may enter 255.255.255.255 or any other value.
Config - Host IP Address Enter the IP address of the host with the configuration file. The configura-
tion file is automatically read by the BDI3000 after every start-up.
Configuration fileEnter the full path and name of the configuration file. This file is read via
TFTP. Keep in mind that TFTP has it’s own root directory (usual /tftpboot).
You can simply copy the configuration file to this directory and the use the
file name without any path.
For more information about TFTP use "man tftpd".
The BDI is in loader mode when there is no valid firmware loaded or you connect to it with the setup
tool. While in loader mode, the Mode LED is blinking. The BDI will not respond to network requests
while in loader mode. To exit loader mode, the "bdisetup -v -s" can be used. You may also power-off
the BDI, wait some time (1min.) and power-on it again to exit loader mode.
$ ./bdisetup -v -p/dev/ttyS0 -b115 -s
BDI Type : BDI3000 (SN: 30000154)
Loader : V1.00
Firmware : V1.00 bdiGDB for P4080
MAC : 00-0c-01-30-00-01
IP Addr : 151.120.25.102
Subnet : 255.255.255.255
Gateway : 255.255.255.255
Host IP : 151.120.25.112
Config : /bdi3000/mytarget.cfg
The Mode LED should go off, and you can try to connect to the BDI via Telnet.
First make sure that the BDI is properly connected (see Chapter 2.1 to 2.4).
!
To avoid data line conflicts, the BDI3000 must be disconnected from the target system while
programming the firmware for an other target CPU family.
dialog box «BDI3000 Update/Setup»
Before you can use the BDI3000 together with the GNU debugger, you must store the initial configuration parameters in the BDI3000 flash memory. The following options allow you to do this:
PortSelect the communication port where the BDI3000 is connected during
this setup session. If you select Network, make sure the Loader is already
active (Mode LED blinking). If there is already a firmware loaded and running, use the Telnet command "boot loader" to activate Loader Mode.
SpeedSelect the baudrate used to communicate with the BDI3000 loader during
this setup session.
ConnectClick on this button to establish a connection with the BDI3000 loader.
Once connected, the BDI3000 remains in loader mode until it is restarted
or this dialog box is closed.
CurrentPress this button to read back the current loaded BDI3000 firmware ver-
sion. The current firmware version will be displayed.
ErasePress this button to erase the current loaded firmware.
UpdateThis button is only active if there is a newer firmware version present in the
execution directory of the bdiGDB setup software. Press this button to
write the new firmware into the BDI3000 flash memory.
BDI IP AddressEnter the IP address for the BDI3000. Use the following format:
xxx.xxx.xxx.xxx e.g.151.120.25.101
Ask your network administrator for assigning an IP address to this
BDI3000. Every BDI3000 in your network needs a different IP address.
Subnet MaskEnter the subnet mask of the network where the BDI is connected to.
Use the following format: xxx.xxx.xxx.xxxe.g.255.255.255.0
A subnet mask of 255.255.255.255 disables the gateway feature.
Ask your network administrator for the correct subnet mask.
Default GatewayEnter the IP address of the default gateway. Ask your network administra-
tor for the correct gateway IP address. If the gateway feature is disabled,
you may enter 255.255.255.255 or any other value.
Config - Host IP Address Enter the IP address of the host with the configuration file. The configura-
tion file is automatically read by the BDI3000 after every start-up.
Configuration fileEnter the full path and name of the configuration file. This name is trans-
mitted to the TFTP server when reading the configuration file.
TransmitClick on this button to store the configuration in the BDI3000 flash
memory.
Note:
Using this setup tool via the Network channel is only possible if the BDI3000 is already in Loader
mode (Mode LED blinking). To force Loader mode, enter "boot loader" at the Telnet. The setup tool
tries first to establish a connection to the Loader via the IP address present in the "BDI IP Address"
entry field. If there is no connection established after a time-out, it tries to connect to the default IP
(192.168.53.72).
The firmware update and the initial configuration of the BDI3000 can also be done interactively via a
Telnet connection and a running TFTP server on the host with the firmware file. In cases where it is
not possible to connect to the default IP, the initial setup has to be done via a serial connection.
!
To avoid data line conflicts, the BDI3000 must be disconnected from the target system while
programming the firmware for an other target CPU family.
Following the steps to bring-up a new BDI3000 or updating the firmware.
Connect to the BDI Loader via Telnet.
If a firmware is already running enter "boot loader" and reconnect via Telnet.
$ telnet 192.168.53.72
or
$ telnet <your BDI IP address>
Update the network parameters so it matches your needs:
LDR>network
BDI MAC : 00-0c-01-30-00-01
BDI IP : 192.168.53.72
BDI Subnet : 255.255.255.0
BDI Gateway : 255.255.255.255
Config IP : 255.255.255.255
Config File :
After the initial setup is done, you can test the communication between the host and the BDI3000.
There is no need for a target configuration file and no TFTP server is needed on the host.
• If not already done, connect the BDI3000 system to the network.
• Power-up the BDI3000.
• Start a Telnet client on the host and connect to the BDI3000 (the IP address you entered during initial configuration).
• If everything is okay, a sign on message like «BDI Debugger for Embedded PowerPC» and
a list of the available commands should be displayed in the Telnet window.
2.7 TFTP server for Windows
The bdiGDB system uses TFTP to access the configuration file and to load the application program.
Because there is no TFTP server bundled with Windows, Abatron provides a TFTP server application
tftpsrv.exe. This WIN32 console application runs as normal user application (not as a system service).
Command line syntax:tftpsrv [p] [w] [dRootDirectory]
Without any parameter, the server starts in read-only mode. This means, only read access request
from the client are granted. This is the normal working mode. The bdiGDB system needs only read
access to the configuration and program files.
The parameter [p] enables protocol output to the console window. Try it.
The parameter [w] enables write accesses to the host file system.
The parameter [d] allows to define a root directory.
tftpsrv pStarts the TFTP server and enables protocol output
tftpsrv p wStarts the TFTP server, enables protocol output and write accesses are
allowed.
tftpsrv dC:\tftp\Starts the TFTP server and allows only access to files in C:\tftp and its
subdirectories. As file name, use relative names.
For example "bdi\mpc8548.cfg" accesses "C:\tftp\bdi\mpc8548.cfg"
You may enter the TFTP server into the Startup group so the server is started every time you login.
The firmware within the BDI handles the GDB request and accesses the target memory or registers
via the JTAG interface. There is no need for any debug software on the target system. After loading
the code via TFTP, debugging can begin at the very first assembler statement.
Whenever the BDI system is powered-up the following sequence starts:
Numeric parameters can be entered as decimal (e.g. 700) or as hexadecimal (0x80000).
Note about how to enter 64bit values:
The syntax for 64 bit parameters is :[<high word>_]<low word>
Hex values may also be entered as:0xnnnnnnnnnnnnnnnn
The "high word" (optional) and "low word" can be entered as decimal or hexadecimal. They are handled as two separate values concatenated with an underscore.
The part [INIT] defines a list of commands which should be executed every time the target comes out
of reset. The commands are used to get the target ready for loading the program file.
WGPR register valueWrite value to the selected general purpose register.
registerthe register number 0 .. 31
valuethe value to write into the register
Example: WGPR 0 5
WSPR register valueWrite value to the selected special purpose register.
registerthe register number
valuethe value to write into the register
Example: WSPR 27 0x00001002 ; SRR1 : ME,RI
WREG name valueWrite value to the selected register/memory by name
namethe case sensitive register name from the reg def file
valuethe value to write to the register/memory
Example: WREG pc 0x00001000
WDCSR address valueWrite value to the selected register in DCSR space
addressaddress / offset in DCSR space
valuethe value to write into the register
Example: WDCSR 0x20c 0x0000000e ;CGCR1: Core Group 1
DELAY valueDelay for the selected time. A delay may be necessary to let the clock PLL
lock again after a new clock rate is selected.
valuethe delay time in milliseconds (1...30000)
Example: DELAY 500 ; delay for 0.5 seconds
WM8 address valueWrite a byte (8bit) to the selected memory place.
addressthe memory address
valuethe value to write to the target memory
Example: WM8 0xFFFFFA21 0x04 ; SYPCR: watchdog disable ...
WM16 address valueWrite a half word (16bit) to the selected memory place.
addressthe memory address
valuethe value to write to the target memory
Example: WM16 0x02200200 0x0002 ; TBSCR
WM32 address valueWrite a word (32bit) to the selected memory place.
addressthe memory address
valuethe value to write to the target memory
Example: WM32 0x02200000 0x01632440 ; SIUMCR
WM64 address valueWrite a double word (64bit) to the selected memory place.
addressthe memory address
valuethe value to write to the target memory
Example: WM64 0xFFF00000 0x123456789abcdef0
TSZ1 start endDefines a memory range with 1 byte maximal transfer size.
Normally when the BDI reads or writes a memory block, it tries to access
the memory with a burst access. The TSZx entry allows to define a maximal transfer size for up to 8 address ranges.
startthe start address of the memory range
endthe end address of the memory range
Example: TSZ1 0xFF000000 0xFFFFFFFF ; PCI ROM space
TSZ2 start endDefines a memory range with 2 byte maximal transfer size.
TSZ4 start endDefines a memory range with 4 byte maximal transfer size.
TSZ8 start endDefines a memory range with 8 byte maximal transfer size.
MMAP start endBecause a memory access to an invalid memory space via JTAG can lead
to a deadlock, this entry can be used to define up to 32 valid memory ranges. If at least one memory range is defined, the BDI checks against this
range(s) and avoids accessing of not mapped memory ranges.
startthe start address of a valid memory range
endthe end address of this memory range
Example: MMAP 0xFFE00000 0xFFFFFFFF ;Boot ROM
EXEC [n_]opcode [data] This entry causes the processor to execute one instruction. If a load in-
struction should get the data from the JTAG port instead from normal
memory, the optional n_ and data parameters are used.
n0 = load data from normal memory
1 = load 8/16/32-bit data from JTAG port
2 = load 64-bit data from JTAG port
opcodethe opcode of the PowerPC instruction
datathe data to present to the instruction
Example: EXEC 0_0x3c60aba4 ;load GPR 3 via lui
Adds an entry to the TLB0 or TLB1 array. The two 64-bit values of an init
list entry are used to define MAS0 (upper 16 bits), MAS1, MAS2, MAS3
and MAS7 (lower 16 bits) before a tlbwe instruction is executed. If other
MASx registers needs a special value, use the WSPR init list entry. A TLB
entry can also be addd via a Telnet command (enter WTLB at the telnet
for a description).
mas1value to load into MAS1
mas2value to load into MAS2 (only lower 32 bits)
mas3value to load into MAS3
mas0value to load into upper 16-bits of MAS0
mas7value to load into lower 16-bits of MAS7
POWERUP delayWhen the BDI detects target power-up, HRESET is forced immediately.
This way no code from a boot ROM is executed after power-up. The value
entered in this configuration line is the delay time in milliseconds the BDI
waits before it begins JTAG communication. This time should be longer
than the on-board reset circuit asserts HRESET.
delaythe power-up start delay in milliseconds
Example:POWERUP 5000 ;start delay after power-up
RESET type [time]Normally the BDI drives the HRESET line during startup. If reset type is
NONE, the BDI does not assert a hardware reset during startup. This entry
can also be used to change the default reset time.
typeNONE
HARD (default)
KEEP keep HRESET assert during target power-up
timeThe time in milliseconds the BDI assert the reset signal.
Example: RESET NONE ; no reset during startup
RESET HARD 1000 ; assert RESET for 1 second
EDBCR0 listThis parameter allows to change the default EDBCR0 value. By default
the EDM, DNH and EFT bits are set.
listdefines the bits to set (EDM, DNH and EFT)
Example: EDBCR0 EDM DNH EFT ;this is the default
WAKEUP timeThis entry in the init list allows to define a delay time (in ms) the BDI inserts
between releasing the COP-HRESET line and starting communicating
with the target. This init list entry may be necessary if COP-HRESET is delayed on its way to the PowerPC reset pin.
timethe delay time in milliseconds
Example: WAKEUP 3000 ; insert 3sec wake-up time
STARTUP mode [runtime]
This parameter selects the target startup mode.
The following modes are supported:
HALTThis mode forces the target to debug mode immediately
out of reset. If HALT for core number 0 is defined, then
all cores within an SOC will be halted immediatelly out of
reset.
STOPIn this mode, the BDI lets the target execute code for
"runtime" milliseconds after reset. This mode is useful
when monitor code should initialize the target system.
RUNAfter reset, the target executes code until stopped by the
Telnet "halt" command.
Example: STARTUP STOP 3000 ; let the CPU run for 3 seconds
BREAKMODE modeThis parameter defines how breakpoints are implemented. The current
mode can also be changed via the Telnet interface
SOFTThis is the normal mode. Breakpoints are implemented
by replacing code with a DNH instruction.
HARDIn this mode, the target breakpoint hardware is used.
Only 2 breakpoints at a time is supported.
LOOPIn this mode, breakpoints are implemented by replacing
code with an endless loop (0x48000000). Maybe useful
for special debug tasks. The processor does not auto-
matically enter debug mode, it has to be halted manually
via Telnet or GDB.
Example: BREAKMODE HARD
STEPMODE modeThis parameter defines how single step (instruction step) is implemented.
The alternate step mode (HWBP) may be useful when stepping instructions that causes a TLB miss exception.
In case BREAKMODE LOOP is selected, this parameter is ignored and
single step is implemented by replacing the code of the next instruction(s)
with an endless loop (0x48000000).
ICMPThis is the mode, single step is implemented via the in-
struction complete (ICMP) debug event.
HWBPIn this mode, a hardware breakpoint on the next instruc-
tion is used to implement single stepping.
Example: STEPMODE HWBP
MMU XLAT [kb]In order to support Linux kernel debugging when MMU is on, the BDI
translates effective (virtual) to physical addresses. This translation is done
based on the current MMU configuration (page tables). If this configuration
line is present, the BDI translates the addresses received from GDB before it accesses physical memory. The optional parameter defines the kernel virtual base address (default is 0xC0000000) and is used for default
address translation. For more information see also chapter "Embedded
Linux MMU Support". Addresses entered at the Telnet are never translated. Translation can be probed with the Telnet command PHYS.
If not zero, the 12 lower bits of "kb" defines the position of the page present
bit in a page table entry. By default 0x800 is assumed for the page present
bit. The position may depend on the Linux kernel version.
A "kb" value of 0xFFFFFFFF disables the default translation.
PTBASE addr [64BIT]This parameter defines the physical memory address where the BDI looks
for the virtual/physical address of the array with the two page table pointers. For more information see also chapter "Embedded Linux MMU Support". If this parameter is not defined, the BDI searches TLB0 in order to
translate a virtual address (TLB1 is always searched).
If the additional "64BIT" option is present, the BDI assume a 64-bit PTE.
addrPhysical address of the memory used to store the virtual
address of the array with the two page table pointers.
Example: PTBASE 0xf0
SIO port [baudrate]When this line is present, a TCP/IP channel is routed to the BDI’s RS232
connector. The port parameter defines the TCP port used for this BDI to
host communication. You may choose any port except 0 and the default
Telnet port (23). On the host, open a Telnet session using this port. Now
you should see the UART output in this Telnet session. You can use the
normal Telnet connection to the BDI in parallel, they work completely independent. Also input to the UART is implemented.
portThe TCP/IP port used for the host communication.
baudrateThe BDI supports 2400 ... 115200 baud
Example: SIO 7 9600 ;TCP port for virtual IO
MEMACCES mode [attr] There are two possible ways to access memory. Via the current core by
executing ld/st instructions or via the System Access Port (SAP).
See also Telnet chapter. The following modes are supported:
SAPMemory access via SAP (default). The attr is a delay
sometimes necessary when accessing slow memory.
COREMemory access via current core. The optional attr pa-
rameter is explained in the Telnet chapter.
Example: MEMACCES CORE ; access via core
REGLIST listThis parameter defines the transferred GDB registers packet. By default
STD and FPR are read and transferred in 32-bit mode. The following
names are use to select a register group and register size:
STDThe standard register block. The FP registers are not
read from the target. Placeholders are transferred.
FPRThe floating point registers are read and transferred.
64BITThe register packet is sent as expected by GDB for a 64-
bit PowerPC target.
Example: REGLIST STD ; standard registers in 32-bit mode
REGLIST STD FPR 64BIT ;use 64-bit mode
Daisy chained JTAG devices:
The BDI can also handle systems with multiple devices connected to the JTAG scan chain. In order
to put the other devices into BYPASS mode and to count for the additional bypass registers, the BDI
needs some information about the scan chain layout. Enter the number (count) and total instruction
register (irlen) length of the devices present before the PowerPC chip (Predecessor). Enter the appropriate information also for the devices following the PowerPC chip (Successor):
SCANPRED count irlenThis value gives the BDI information about JTAG devices present before
the PowerPC chip in the JTAG scan chain.
countThe number of preceding devices
irlenThe sum of the length of all preceding instruction regis-
ters (IR).
Example: SCANPRED 1 8 ; one device with an IR length of 8
SCANSUCC count irlenThis value gives the BDI information about JTAG devices present after the
PowerPC chip in the JTAG scan chain.
countThe number of succeeding devices
irlenThe sum of the length of all succeeding instruction reg-
isters (IR).
Example: SCANSUCC 2 12 ; two device with an IR length of 8+4
Overriding Reset Configuration Word (RCW):
The BDI supports overriding the RCW Source and also overriding individual RCW values. If there is
no valid RCW present at the currently via pin selected RCW Source a Hard-coded RCW should be
selected. Never override the PLL configuration in RCW[0]-RCW[3] !
The part [HOST] defines some host specific values.
IP ipaddressThe IP address of the host.
ipaddressthe IP address in the form xxx.xxx.xxx.xxx
Example: IP 151.120.25.100
FILE filenameThe default name of the file that is loaded into RAM using the Telnet ’load’
command. This name is used to access the file via TFTP. If the filename
starts with a $, this $ is replace with the path of the configuration file name.
filenamethe filename including the full path or $ for relative path.
Example:FILE F:\gnu\demo\ppc\test.elf
FILE $test.elf
FORMAT format [offset]The format of the image file and an optional load address offset. If the im-
age is already stored in ROM on the target, select ROM as the format. The
optional parameter "offset" is added to any load address read from the image file.
formatSREC, BIN, AOUT, ELF or ROM
Example: FORMAT ELF
FORMAT ELF 0x10000
LOAD modeIn Agent mode, this parameters defines if the code is loaded automatically
after every reset.
modeAUTO, MANUAL
Example: LOAD MANUAL
START addressThe address where to start the program file. If this value is not defined and
the core is not in ROM, the address is taken from the image file. If this value is not defined and the core is already in ROM, the PC will not be set
before starting the program file. This means, the program starts at the normal reset address (0xFFF00100).
addressthe address where to start the program file
Example:START 0x1000
DEBUGPORT port [RECONNECT]
The TCP port GDB uses to access the target. If the RECONNECT parameter is present, an open TCP/IP connection (Telnet/GDB) will be closed if
there is a connect request from the same host (same IP address).
portthe TCP port number (default = 2001)
Example:DEBUGPORT 2001
PROMPT stringThis entry defines a new Telnet prompt. The current prompt can also be
The Telnet interface supports programming and erasing of flash memories. The bdiGDB system has
to know which type of flash is used, how the chip(s) are connected to the CPU and which sectors to
erase in case the ERASE command is entered without any parameter.
CHIPTYPE typeThis parameter defines the type of flash used. It is used to select the cor-
CHIPSIZE sizeThe size of one flash chip in bytes (e.g. AM29F010 = 0x20000). This value
is used to calculate the starting address of the current flash memory bank.
sizethe size of one flash chip in bytes
Example:CHIPSIZE 0x80000
BUSWIDTH widthEnter the width of the memory bus that leads to the flash chips. Do not en-
ter the width of the flash chip itself. The parameter CHIPTYPE carries the
information about the number of data lines connected to one flash chip.
For example, enter 16 if you are using two AM29F010 to build a 16bit flash
memory bank.
withthe width of the flash memory bus in bits (8 | 16 | 32 | 64)
Example:BUSWIDTH 16
FILE filenameThe default name of the file that is programmed into flash using the Telnet
’prog’ command. This name is used to access the file via TFTP. If the filename starts with a $, this $ is replace with the path of the configuration file
name. This name may be overridden interactively at the Telnet interface.
filenamethe filename including the full path or $ for relative path.
Example:FILE F:\gnu\ppc\bootrom.hex
FILE $bootrom.hex
FORMAT format [offset]The format of the file and an optional address offset. The optional param-
eter "offset" is added to any load address read from the program file.
You get the best programming performance when using a binary format
(BIN, AOUT, ELF or IMAGE).
formatSREC, BIN, AOUT, ELF or IMAGE
Example: FORMAT BIN 0x10000
WORKSPACE addressIf a workspace is defined, the BDI uses a faster programming algorithm
that runs out of RAM on the target system. Otherwise, the algorithm is processed within the BDI. The workspace is used for a 1kByte data buffer and
to store the algorithm code. There must be at least 2kBytes of RAM available for this purpose.
addressthe address of the RAM area
Example:WORKSPACE 0x00000000
The flash memory may be individually erased or unlocked via the Telnet
interface. In order to make erasing of multiple flash sectors easier, you can
enter an erase list. All entries in the erase list will be processed if you enter
ERASE at the Telnet prompt without any parameter. This list is also used
if you enter UNLOCK at the Telnet without any parameters. With the "increment" and "count" option you can erase multiple equal sized sectors
with one entry in the erase list.
addressAddress of the flash sector, block or chip to erase
incrementIf present, the address offset to the next flash sector
countIf present, the number of equal sized sectors to erase
modeBLOCK, CHIP, UNLOCK
Without this optional parameter, the BDI executes a sec-
tor erase. If supported by the chip, you can also specify
a block or chip erase. If UNLOCK is defined, this entry is
also part of the unlock list. This unlock list is processed
if the Telnet UNLOCK command is entered without any
parameters.
Note: Chip erase does not work for large chips because
the BDI time-outs after 3 minutes. Use block erase.
waitThe wait time in ms is only used for the unlock mode. Af-
ter starting the flash unlock, the BDI waits until it pro-
cesses the next entry.
Example:ERASE 0xff040000 ;erase sector 4 of flash
ERASE 0xff060000 ;erase sector 6 of flash
ERASE 0xff000000 CHIP ;erase whole chip(s)
ERASE 0xff010000 UNLOCK 100 ;unlock, wait 100ms
ERASE 0xff000000 0x10000 7 ; erase 7 sectors
Example for the ADS8260 flash memory:
[FLASH]
CHIPTYPE I28BX8 ;Flash type
CHIPSIZE 0x200000 ;The size of one flash chip in bytes (e.g. AM29F010 = 0x20000)
BUSWIDTH 32 ;The width of the flash memory bus in bits (8 | 16 | 32 | 64)
WORKSPACE 0x04700000 ;workspace in dual port RAM
FILE E:\gnu\demo\ads8260\bootrom.hex ;The file to program
ERASE 0xFF900000 ;erase sector 4 of flash SIMM (LH28F016SCT)
ERASE 0xFF940000 ;erase sector 5 of flash SIMM
ERASE 0xFF980000 ;erase sector 6 of flash SIMM
ERASE 0xFF9c0000 ;erase sector 7 of flash SIMM
the above erase list maybe replaces with:
ERASE 0xFF900000 0x40000 4 ; erase sector 4 to 7 of flash SIMM
There are different flash algorithm supported. Almost all currently available parallel NOR flash memories can be programmed with one of these algorithm. The flash type selects the appropriate algorithm and gives additional information about the used flash.
On our web site (www.abatron.ch -> Debugger Support -> GNU Support -> Flash Support) there is a
PDF document available that shows the supported parallel NOR flash memories.
Some newer Spansion MirrorBit flashes cannot be programmed with the MIRRORX16 algorithm because of the used unlock address offset. Use S29M32X16 for these flashes.
The AMD and AT49 algorithm are almost the same. The only difference is, that the AT49 algorithm
does not check for the AMD status bit 5 (Exceeded Timing Limits).
Only the AMD and AT49 algorithm support chip erase. Block erase is only supported with the AT49
algorithm. If the algorithm does not support the selected mode, sector erase is performed. If the chip
does not support the selected mode, erasing will fail. The erase command sequence is different only
in the 6th write cycle. Depending on the selected mode, the following data is written in this cycle (see
also flash data sheets): 0x10 for chip erase, 0x30 for sector erase, 0x50 for block erase.
To speed up programming of Intel Strata Flash and AMD MirrorBit Flash, an additional algorithm is
implemented that makes use of the write buffer. The Strata algorithm needs a workspace, otherwise
the standard Intel algorithm is used.
Some Intel flash chips (e.g. 28F800C3, 28F160C3, 28F320C3) power-up with all blocks in locked
state. In order to erase/program those flash chips, use the init list to unlock the appropriate blocks:
addrThis is the address of the first sector to erase or unlock.
stepThis value is added to the last used address in order to get to the next sec-
tor. In other words, this is the size of one sector in bytes.
countThe number of sectors to erase or unlock.
The following example unlocks all 256 sectors of an Intel Strata flash (28F256K3) that is mapped to
0x00000000. In case there are two flash chips to get a 32bit system, double the "step" parameter.
In order to make it easier to access target registers via the Telnet interface, the BDI can read in a
register definition file. In this file, the user defines a name for the register and how the BDI should
access it (e.g. as memory mapped, memory mapped with offset, ...). The name of the register definition file and information for different registers type has to be defined in the configuration file. The
register name, type, address/offset/number and size are defined in a separate register definition file.
An entry in the register definition file has the following syntax:
nametypeaddr[size [SWAP]]
nameThe name of the register (max. 15 characters)
typeThe register type
GPRGeneral purpose register
SPRSpecial purpose register
PMRPerformance monitor register
CCSRRelative to CCSRBAR memory mapped register.
DCSRMemory mapped register in DCSR space
MMAbsolute direct memory mapped register
DMM1...DMM4Relative direct memory mapped register
IMM1...IMM4Indirect memory mapped register
addrThe address, offset or number of the register
sizeThe size (8, 16, 32) of the register (default is 32)
SWAPIf present, the bytes of a 16bit or 32bit register are swapped. This is useful
to access little endian ordered registers (e.g. PCI bridge configuration registers).
The following entries are supported in the [REGS] part of the configuration file:
FILE filenameThe name of the register definition file. This name is used to access the
file via TFTP. The file is loaded once during BDI startup.
filenamethe filename including the full path
Example:FILE C:\bdi\regs\regP4080.def
DMMn baseThis defines the base address of direct memory mapped registers. This
base address is added to the individual offset of the register.
basethe base address
Example:DMM1 0x01000
IMMn addr dataThis defines the addresses of the memory mapped address and data reg-
isters of indirect memory mapped registers. The address of a IMMn register is first written to "addr" and then the register value is access using
"data" as address.
addrthe address of the Address register
datathe address of the Data register
Example:DMM1 0x04700000
Remark:
The registers msr, cr, pc, pc64, iar, iar64 and fpscr and are predefined.
Because the GDB server runs within the BDI, no debug support has to be linked to your application.
There is also no need for any BDI specific changes in the application sources.
3.3.1 Target setup
Target initialization may be done at two places. First with the BDI configuration file, second within the
application. The setup in the configuration file must at least enable access to the target memory
where the application will be loaded. Disable the watchdog and setting the CPU clock rate should
also be done with the BDI configuration file. Application specific initializations like setting the timer
rate are best located in the application startup sequence.
3.3.2 Connecting to the target
As soon as the target comes out of reset, BDI initializes it and optionally loads your application code.
BDI now waits for GDB request from the debugger running on the host.
After starting the debugger, it must be connected to the remote target. This can be done with the following command at the GDB prompt:
(gdb)target remote bdi3000:2001
bdi3000This stands for an IP address. The HOST file must have an appropriate
entry. You may also use an IP address in the form xxx.xxx.xxx.xxx
2001This is the TCP port used to communicate with the BDI
If not already suspended, this stops the execution of application code and the target CPU changes
to background debug mode.
Remember, every time the application is suspended, the target CPU is freezed. During this time, no
hardware interrupts will be processed.
Note: For convenience, the GDB detach command triggers a target reset sequence in the BDI.
(gdb)detach
... Wait until BDI has resetet the target and reloaded the image
(gdb)target remote bdi3000:2001
3.3.3 GDB monitor command
The BDI supports the GDB V5.x "monitor" command. Telnet commands are executed and the Telnet
output is returned to GDB. This way you can for example switch the BDI breakpoint mode from within
your GDB session.
(gdb) target remote bdi3000:2001
Remote debugging using bdi3000:2001
0x10b2 in start ()
(gdb) monitor break
Breakpoint mode is SOFT
(gdb) mon break hard
A RS232 port of the target can be connected to the RS232 port of the BDI3000. This way it is possible
to access the target’s serial I/O via a TCP/IP channel. For example, you can connect a Telnet session
to the appropriate BDI3000 port. Connecting GDB to a GDB server (stub) running on the target
should also be possible.
Target System
RS232 Connector
1 - NC
2 - RXD
3 - TXD
4 - NC
5 - GROUND
6 - NC
7 - NC
8 - NC
9 - NC
54321
9876
RS232 POWER
P4080
RS232
BDI3000
Ethernet (10/100 BASE-T)
The configuration parameter "SIO" is used to enable this serial I/O routing.
The used framing parameters are 8 data, 1 stop and not parity.
[TARGET]
....
SIO79600;Enable SIO via TCP port 7 at 9600 baud
Warning!!!
Once SIO is enabled, connecting with the setup tool to update the firmware will fail. In this case either
disable SIO first or disconnect the BDI from the LAN while updating the firmware.
The bdiGDB system supports Linux kernel debugging when MMU is on. The MMU configuration parameter enables this mode of operation. In this mode, all addresses received from GDB are assumed
to be virtual. Before the BDI accesses memory, it translates this address into a physical one based
on information found in the TLB’s or kernel/user page table.
If PTBASE is not defined, the BDI does TLB1, TLB0 and if enabled default translation (in this order).
In order to search the page tables, the BDI needs to know the start addresses of the first level page
table. The configuration parameter PTBASE defines the physical address where the BDI looks for
the virtual/physical address of an array with two virtual/physical addresses of first level page tables.
The first one points normally to the kernel page table, the second one can point to the current user
page table. As long as the base pointer or the first entry is zero, the BDI does only L2 CAM (L2 TLB1)
and default translation. Default translation maps a 256 Mbyte range starting at KERNELBASE to
0x00000000. The second page table is only searched if its address is not zero and there was no
match in the first one.
The pointer structure is as follows:
PTBASE (physical address) ->
PTE pointer pointer(virtual or physical address) ->
PTE kernel pointer (virtual or physical address)
PTE user pointer (virtual or physical address)
The pointers are assumed virtual if they are >= KERNELBASE. In that case, default translation is applied to get the physical address.
Newer versions of "arch/ppc/kernel/head.S" support the automatic update of the BDI page table information structure. Search "head.S" for "abatron" and you will find the BDI specific extensions.
Extract from the configuration file:
[INIT]
......
WM32 0x000000f0 0x00000000 ;invalidate page table base
[TARGET]
....
MMUXLAT;translate effective to physical address
PTBASE 0x000000f0 ;here is the pointer to the page table pointers
To debug the Linux kernel when MMU is enabled you may use the following load and startup sequence:
• Load the compressed linux image
• Set a hardware breakpoint with the Telnet at a point where MMU is enabled. For example at
"start_kernel".
BDI> BI 0xC0061550
• Start the code with GO at the Telnet
• The Linux kernel is decompressed and started
• The system should stop at the hardware breakpoint (e.g. at start_kernel)
• Disable the hardware breakpoint with the Telnet command CI.
• If not automatically done by the kernel, setup the page table pointers for the BDI.
• Start GDB with vmlinux as parameter
• Attach to the target
• Now you should be able to debug the Linux kernel
To setup the BDI page table information structure manually, set a hardware breakpoint at
"start_kernel" and use the Telnet to write the address of "swapper_pg_dir" to the appropriate place.
BDI>bi 0xc0061550/* set breakpoint at start_kernel */
BDI>go
../* target stops at start_kernel */
BDI>ci
BDI>mm 0xf0 0xc00000f8/* Let PTBASE point to an array of two pointers*/
BDI>mm 0xf8 0xc0057000/* write address of swapper_pg_dir to first pointer */
BDI>mm 0xfc 0x00000000/* clear second (user) pointer */
A Telnet server is integrated within the BDI. The Telnet channel is used by the BDI to output error
messages and other information. Also some basic debug commands can be executed.
Telnet Debug features:
• Display and modify memory locations
• Display and modify general and special purpose registers
• Single step a code sequence
• Set hardware breakpoints
• Load a code file from any host
• Start / Stop program execution
• Programming and Erasing Flash memory
During debugging with GDB, the Telnet is mainly used to reboot the target (generate a hardware reset and reload the application code). It may be also useful during the first installation of the bdiGDB
system or in case of special debug needs.
Example of a Telnet session:
P4080#0>info
Target CPU : P4080 Core#0
Core state : halted
Debug entry cause : device event
Current PC : 0xfffffffc
Current CR : 0x00000000
Current MSR : 0x00000000
Current LR : 0x00000000
Current CCSRBAR : 0x0_fe000000
P4080#0>rd
GPR00: c11bc002 06278553 80028188 aba40000
GPR04: 609db195 ad1944b2 deadbeef 002883a6
GPR08: 20119520 2032dc90 94110404 29038003
GPR12: 9422cf8e 0c105000 613b00b0 4e0d4548
GPR16: 0f15d163 3820d4a3 806b42d8 4c005402
GPR20: b0010949 846310d8 c0d53502 4c41d854
GPR24: c0602409 4443cd98 a8911575 e0021810
GPR28: 200842c0 c890cc15 2c2390ce 604bc0c1
CR : 00000000 MSR: 00000000
P4080#0>md 0
The DUMP command uses TFTP to write a binary image to a host file. Writing via TFTP on a Linux/
Unix system is only possible if the file already exists and has public write access. Use "man tftpd" to
get more information about the TFTP server on your host.
"PHYS <address> converts an effective to a physical address",
"MD [<address>] [<count>] display target memory as word (32bit)",
"MDD [<address>] [<count>] display target memory as double word (64bit)",
"MDH [<address>] [<count>] display target memory as half word (16bit)",
"MDB [<address>] [<count>] display target memory as byte (8bit)",
"DUMP <addr> <size> [<file>] dump target memory to a file",
"MM <addr> <value> [<cnt>] modify word(s) (32bit) in target memory",
"MMD <addr> <value> [<cnt>] modify double word(s) (64bit) in target memory",
"MMH <addr> <value> [<cnt>] modify half word(s) (16bit) in target memory",
"MMB <addr> <value> [<cnt>] modify byte(s) (8bit) in target memory",
"MT <addr> <count>[<loop>] memory test",
"MC [<address>] [<count>] calculates a checksum over a memory range",
"MV verifies the last calculated checksum",
"RD [<name>] display general purpose or user defined register",
"RDUMP [<file>] dump all user defined register to a file",
"RDSPR <number> display special purpose register",
"RDPMR <number> display performance monitor register",
"RM {<nbr>|<name>} <value> modify general purpose or user defined register",
"RMSPR <number> <value> modify special purpose register",
"RMPMR <number> <value> modify performance monitor register",
"RDCSR <addr> [<count>] display register(s) in DCSR space (Run Control)",
"WDCSR <addr> <value> write to a register in DCSR space (Run Control)",
"RESET [HALT | RUN [time]] reset the target system, change startup mode",
"BREAK [SOFT | HARD] display or set current breakpoint mode",
"GO [<pc>] set PC and start current core",
"CONT <cores> restart multiple cores (<cores> = core bit map)",
"TI [<pc>] trace on instruction (single step)",
"TC [<pc>] trace on change of flow",
"HALT [<cores>] force core(s) to debug mode (<cores> = core bit map)",
"BI <addr> set instruction hardware breakpoint",
"CI [<id>] clear instruction hardware breakpoint(s)",
"BD [R|W] <addr> set data watchpoint",
"CD [<id>] clear data watchpoint(s)",
"INFO display information about the current core",
"STATE display information about all cores",
"LOAD [<offset>] [<file> [<format>]] load program file to target memory",
"VERIFY [<offset>] [<file> [<format>]] verify a program file to target memory",
"PROG [<offset>] [<file> [<format>]] program flash memory",
" <format> : SREC, BIN, AOUT or ELF",
"ERASE [<address> [<mode>]] erase a flash memory sector, chip or block",
" <mode> : CHIP, BLOCK or SECTOR (default is sector)",
"ERASE <addr> <step> <count> erase multiple flash sectors",
"UNLOCK [<addr> [<delay>]] unlock a flash sector",
"UNLOCK <addr> <step> <count> unlock multiple flash sectors",
"FLASH <type> <size> <bus> change flash configuration",
"DELAY <ms> delay for a number of milliseconds",
"MEMACC {CORE | SAP} [<attr>] select memory access mode (normally SAP)",
"SELECT <core> change the current core",
"HOST <ip> change IP address of program file host",
"PROMPT <string> defines a new prompt string",
"CONFIG display or update BDI configuration",
"CONFIG <file> [<hostIP> [<bdiIP> [<gateway> [<mask>]]]]",
"UPDATE reload the configuration without a reboot",
"HELP display command list",
"JTAG switch to JTAG command mode",
"BOOT [loader] reboot the BDI and reload the configuration",
"QUIT terminate the Telnet session"
There are two memory access modes implemented. The default is via System Access Port (SAP).
Via SAP physical addresses are used and the access does not make use of any of the cores. SAP
accesses memory like an additional bus master. If memory access CORE is selected, the current
core executes load/store instructions in its current context. In this mode MMU translation takes place
unless the use of real addresses is forced.
For memory accesses via core the <attr> parameter in the Telnet MEMACC command has the following meaning (default is 0):
R : 1 = Real Addressing Mode (no MMU translation)
WIMGE: Page attributes for load/store instruction, only used if R = 1
For example to access the real address 0x3_8400_0000 with I and G set via the current core:
BDI> memacc core 0x94
BDI> md 0x384000000 1
For memory accesses via SAP the <attr> defines a delay sometimes necessary when accessing
slow memory. The following example define a 100us delay during memory accesses via SAP.
BDI> memacc sap 100
Memory access mode is a global selection. It is not possible to select different modes for different
cores.
Note:
For information about the registers in DCSR space please contact Freescale.
The bdiGDB system supports concurrent debugging of up to 8 e500mc cores. For every core you can
start its own GDB session. The default port numbers used to attach the remote targets are 2001 ...
2008. In the Telnet you switch between the cores with the command "select <0..7>". In the configuration file, simply begin the line with the appropriate core number. If there is no #n in front of a line,
the BDI assumes core #0.
The following example defines 4 cores for debugging. For a complete example, look at the configuration examples.
[TARGET]
; common parameters
POWERUP 5000 ;start delay after power-up detected in ms
JTAGCLOCK 1 ;use 16 MHz JTAG clock
WAKEUP 200 ;give reset time to complete
;
;========================================================
; !!!! define the core ID (the #x) without any holes !!!!
; !!!! no need that core ID matches the core number !!!!
; !!!! A valid example is: #1 CPUTYPE P4080 5 0 !!!!
;========================================================
;
; Core#0 parameters (active vCPU after reset)
#0 CPUTYPE P4080 0 0 ;Core0 / SOC0
#0 STARTUP STOP 5000 ;let U-boot setup the system
#0 BREAKMODE HARD ;SOFT or HARD, HARD uses PPC hardware breakpoint
#0 STEPMODE HWBP ;ICMP or HWBP, HWBP uses a hardware breakpoint
;
; Core#1 parameters
#1 CPUTYPE P4080 1 0 ;Core1 / SOC0
#1 STARTUP RUN ;let core run
;
; Core#2 parameters
#2 CPUTYPE P4080 2 0 ;Core2 / SOC0
#2 STARTUP RUN ;let core run
;
; Core#3 parameters
#3 CPUTYPE P4080 3 0 ;Core3 / SOC0
#3 STARTUP RUN ;let core run
;
Be aware that via Telnet you select the e500mc core by its BDI core ID (#n). This BDI core ID is not
necessary the core number within the SOC. Assuming there are two P4040 daisy chained, you may
use BDI core ID #4 to select core 0 in the second P4040.
HALT [<cores>]Force one or multiple cores to debug mode. If there is no <cores> param-
eter, the currently selected core is forced to debug mode.
<cores>core bit map
Example:halt 0x00ff ; halt 8 cores #0...#7
Telnet session:
P4080#0>info
Target CPU : P4080 Core#0
Core state : halted
Debug entry cause : debug halt
Current PC : 0x7ff74c34
Current CR : 0x22000084
Current MSR : 0x00029200
Current LR : 0x7ff74c38
Current CCSRBAR : 0x0_fe000000
P4080#0>state
Core#0: halted 0x7ff74c34 debug halt
Core#1: running
Core#2: running
Core#3: running
Core#4: running
Core#5: running
Core#6: running
Core#7: running
P4080#0>halt 0xf0
- TARGET: core #4 has entered debug mode
- TARGET: core #5 has entered debug mode
- TARGET: core #6 has entered debug mode
- TARGET: core #7 has entered debug mode
P4080#0>state
Core#0: halted 0x7ff74c34 debug halt
Core#1: running
Core#2: running
Core#3: running
Core#4: halted 0xfffff0f8 debug halt
Core#5: halted 0xfffff0f8 debug halt
Core#6: halted 0xfffff0f8 debug halt
Core#7: halted 0xfffff0f8 debug halt
P4080#0>cont 0x30
P4080#0>state
Core#0: halted 0x7ff74c34 debug halt
Core#1: running
Core#2: running
Core#3: running
Core#4: running
Core#5: running
Core#6: halted 0xfffff0f8 debug halt
Core#7: halted 0xfffff0f8 debug halt
P4080#0>
ABATRON Switzerland warrants the Hardware to be free of defects in materials and workmanship
for a period of 3 years following the date of purchase when used under normal conditions. In the
event of notification within the warranty period of defects in material or workmanship, ABATRON will
repair or replace the defective hardware. The cost for the shipment to Abatron must be paid by the
customer. Failure in handling which leads to defects are not covered under this warranty. The warranty is void under any self-made repair operation.
7.2 Software
License
Against payment of a license fee the client receives a usage license for this software product, which
is not exclusive and cannot be transferred.
Copies
The client is entitled to make copies according to the number of licenses purchased. Copies
exceeding this number are allowed for storage purposes as a replacement for defective storage
mediums.
Update and Support
The agreement includes free software maintenance (update and support) for one year from date of
purchase. After this period the client may purchase software maintenance for an additional year.
7.3 Warranty and Disclaimer
ABATRON AND ITS SUPPLIERS HEREBY DISCLAIMS AND EXCLUDES, TO THE EXTENT
PERMITTED BY APPLICABLE LAW, ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT.
7.4 Limitation of Liability
IN NO EVENT SHALL ABATRON OR ITS SUPPLIERS BE LIABLE TO YOU FOR ANY DAMAGES,
INCLUDING, WITHOUT LIMITATION, ANY SPECIAL, INDIRECT, INCIDENTAL OR
CONSEQUENTIAL DAMAGES, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
PERFORMANCE OF THE HARDWARE AND/OR SOFTWARE, INCLUDING WITHOUT
LIMITATION, LOSS OF PROFITS, BUSINESS, DATA, GOODWILL, OR ANTICIPATED SAVINGS,
EVEN IF ADVISED OF THE POSSIBILITY OF THOSE DAMAGES.
The hardware and software product with all its parts, copyrights and any other rights remain in possession of ABATRON. Any dispute, which may arise in connection with the present agreement shall
be submitted to Swiss Law in the Court of Zug to which both parties hereby assign competence.