| blob | b89768f17917f9bfcf79ad9c462346f6835db639 |
1 # SPDX-License-Identifier: GPL-2.0+
2 #
3 # (C) Copyright 2000 - 2013
4 # Wolfgang Denk, DENX Software Engineering, wd@denx.de.
6 Summary:
7 ========
9 This directory contains the source code for U-Boot, a boot loader for
10 Embedded boards based on PowerPC, ARM, MIPS and several other
11 processors, which can be installed in a boot ROM and used to
12 initialize and test the hardware or to download and run application
13 code.
15 The development of U-Boot is closely related to Linux: some parts of
16 the source code originate in the Linux source tree, we have some
17 header files in common, and special provision has been made to
18 support booting of Linux images.
20 Some attention has been paid to make this software easily
21 configurable and extendable. For instance, all monitor commands are
22 implemented with the same call interface, so that it's very easy to
23 add new commands. Also, instead of permanently adding rarely used
24 code (for instance hardware test utilities) to the monitor, you can
25 load and run it dynamically.
28 Status:
29 =======
31 In general, all boards for which a default configuration file exists in the
32 configs/ directory have been tested to some extent and can be considered
33 "working". In fact, many of them are used in production systems.
35 In case of problems you can use
37 scripts/get_maintainer.pl <path>
39 to identify the people or companies responsible for various boards and
40 subsystems. Or have a look at the git log.
43 Where to get help:
44 ==================
46 In case you have questions about, problems with or contributions for
47 U-Boot, you should send a message to the U-Boot mailing list at
48 <u-boot@lists.denx.de>. There is also an archive of previous traffic
49 on the mailing list - please search the archive before asking FAQ's.
50 Please see https://lists.denx.de/pipermail/u-boot and
51 https://marc.info/?l=u-boot
53 Where to get source code:
54 =========================
56 The U-Boot source code is maintained in the Git repository at
57 https://source.denx.de/u-boot/u-boot.git ; you can browse it online at
58 https://source.denx.de/u-boot/u-boot
60 The "Tags" links on this page allow you to download tarballs of
61 any version you might be interested in. Official releases are also
62 available from the DENX file server through HTTPS or FTP.
63 https://ftp.denx.de/pub/u-boot/
64 ftp://ftp.denx.de/pub/u-boot/
67 Where we come from:
68 ===================
70 - start from 8xxrom sources
71 - create PPCBoot project (https://sourceforge.net/projects/ppcboot)
72 - clean up code
73 - make it easier to add custom boards
74 - make it possible to add other [PowerPC] CPUs
75 - extend functions, especially:
76 * Provide extended interface to Linux boot loader
77 * S-Record download
78 * network boot
79 * ATA disk / SCSI ... boot
80 - create ARMBoot project (https://sourceforge.net/projects/armboot)
81 - add other CPU families (starting with ARM)
82 - create U-Boot project (https://sourceforge.net/projects/u-boot)
83 - current project page: see https://www.denx.de/wiki/U-Boot
86 Names and Spelling:
87 ===================
89 The "official" name of this project is "Das U-Boot". The spelling
90 "U-Boot" shall be used in all written text (documentation, comments
91 in source files etc.). Example:
93 This is the README file for the U-Boot project.
95 File names etc. shall be based on the string "u-boot". Examples:
97 include/asm-ppc/u-boot.h
99 #include <asm/u-boot.h>
101 Variable names, preprocessor constants etc. shall be either based on
102 the string "u_boot" or on "U_BOOT". Example:
104 U_BOOT_VERSION u_boot_logo
105 IH_OS_U_BOOT u_boot_hush_start
108 Software Configuration:
109 =======================
111 Selection of Processor Architecture and Board Type:
112 ---------------------------------------------------
114 For all supported boards there are ready-to-use default
115 configurations available; just type "make <board_name>_defconfig".
117 Example: For a TQM823L module type:
119 cd u-boot
120 make TQM823L_defconfig
122 Note: If you're looking for the default configuration file for a board
123 you're sure used to be there but is now missing, check the file
124 doc/README.scrapyard for a list of no longer supported boards.
126 Sandbox Environment:
127 --------------------
129 U-Boot can be built natively to run on a Linux host using the 'sandbox'
130 board. This allows feature development which is not board- or architecture-
131 specific to be undertaken on a native platform. The sandbox is also used to
132 run some of U-Boot's tests.
134 See doc/arch/sandbox/sandbox.rst for more details.
137 Board Initialisation Flow:
138 --------------------------
140 This is the intended start-up flow for boards. This should apply for both
141 SPL and U-Boot proper (i.e. they both follow the same rules).
143 Note: "SPL" stands for "Secondary Program Loader," which is explained in
144 more detail later in this file.
146 At present, SPL mostly uses a separate code path, but the function names
147 and roles of each function are the same. Some boards or architectures
148 may not conform to this. At least most ARM boards which use
149 CONFIG_SPL_FRAMEWORK conform to this.
151 Execution typically starts with an architecture-specific (and possibly
152 CPU-specific) start.S file, such as:
154 - arch/arm/cpu/armv7/start.S
155 - arch/powerpc/cpu/mpc83xx/start.S
156 - arch/mips/cpu/start.S
158 and so on. From there, three functions are called; the purpose and
159 limitations of each of these functions are described below.
161 lowlevel_init():
162 - purpose: essential init to permit execution to reach board_init_f()
163 - no global_data or BSS
164 - there is no stack (ARMv7 may have one but it will soon be removed)
165 - must not set up SDRAM or use console
166 - must only do the bare minimum to allow execution to continue to
167 board_init_f()
168 - this is almost never needed
169 - return normally from this function
171 board_init_f():
172 - purpose: set up the machine ready for running board_init_r():
173 i.e. SDRAM and serial UART
174 - global_data is available
175 - stack is in SRAM
176 - BSS is not available, so you cannot use global/static variables,
177 only stack variables and global_data
179 Non-SPL-specific notes:
180 - dram_init() is called to set up DRAM. If already done in SPL this
181 can do nothing
183 SPL-specific notes:
184 - you can override the entire board_init_f() function with your own
185 version as needed.
186 - preloader_console_init() can be called here in extremis
187 - should set up SDRAM, and anything needed to make the UART work
188 - there is no need to clear BSS, it will be done by crt0.S
189 - for specific scenarios on certain architectures an early BSS *can*
190 be made available (via CONFIG_SPL_EARLY_BSS by moving the clearing
191 of BSS prior to entering board_init_f()) but doing so is discouraged.
192 Instead it is strongly recommended to architect any code changes
193 or additions such to not depend on the availability of BSS during
194 board_init_f() as indicated in other sections of this README to
195 maintain compatibility and consistency across the entire code base.
196 - must return normally from this function (don't call board_init_r()
197 directly)
199 Here the BSS is cleared. For SPL, if CONFIG_SPL_STACK_R is defined, then at
200 this point the stack and global_data are relocated to below
201 CONFIG_SPL_STACK_R_ADDR. For non-SPL, U-Boot is relocated to run at the top of
202 memory.
204 board_init_r():
205 - purpose: main execution, common code
206 - global_data is available
207 - SDRAM is available
208 - BSS is available, all static/global variables can be used
209 - execution eventually continues to main_loop()
211 Non-SPL-specific notes:
212 - U-Boot is relocated to the top of memory and is now running from
213 there.
215 SPL-specific notes:
216 - stack is optionally in SDRAM, if CONFIG_SPL_STACK_R is defined and
217 CONFIG_SYS_FSL_HAS_CCI400
219 Defined For SoC that has cache coherent interconnect
220 CCN-400
222 CONFIG_SYS_FSL_HAS_CCN504
224 Defined for SoC that has cache coherent interconnect CCN-504
226 The following options need to be configured:
228 - CPU Type: Define exactly one, e.g. CONFIG_MPC85XX.
230 - Board Type: Define exactly one, e.g. CONFIG_MPC8540ADS.
232 - 85xx CPU Options:
233 CONFIG_SYS_PPC64
235 Specifies that the core is a 64-bit PowerPC implementation (implements
236 the "64" category of the Power ISA). This is necessary for ePAPR
237 compliance, among other possible reasons.
239 CONFIG_SYS_FSL_ERRATUM_A004510
241 Enables a workaround for erratum A004510. If set,
242 then CONFIG_SYS_FSL_ERRATUM_A004510_SVR_REV and
243 CFG_SYS_FSL_CORENET_SNOOPVEC_COREONLY must be set.
245 CONFIG_SYS_FSL_ERRATUM_A004510_SVR_REV
246 CONFIG_SYS_FSL_ERRATUM_A004510_SVR_REV2 (optional)
248 Defines one or two SoC revisions (low 8 bits of SVR)
249 for which the A004510 workaround should be applied.
251 The rest of SVR is either not relevant to the decision
252 of whether the erratum is present (e.g. p2040 versus
253 p2041) or is implied by the build target, which controls
254 whether CONFIG_SYS_FSL_ERRATUM_A004510 is set.
256 See Freescale App Note 4493 for more information about
257 this erratum.
259 CFG_SYS_FSL_CORENET_SNOOPVEC_COREONLY
261 This is the value to write into CCSR offset 0x18600
262 according to the A004510 workaround.
264 CONFIG_SYS_FSL_SINGLE_SOURCE_CLK
265 Single Source Clock is clocking mode present in some of FSL SoC's.
266 In this mode, a single differential clock is used to supply
267 clocks to the sysclock, ddrclock and usbclock.
269 - Generic CPU options:
271 CONFIG_SYS_FSL_DDR
272 Freescale DDR driver in use. This type of DDR controller is
273 found in mpc83xx, mpc85xx as well as some ARM core SoCs.
275 CFG_SYS_FSL_DDR_ADDR
276 Freescale DDR memory-mapped register base.
278 CONFIG_SYS_FSL_IFC_CLK_DIV
279 Defines divider of platform clock(clock input to IFC controller).
281 CONFIG_SYS_FSL_LBC_CLK_DIV
282 Defines divider of platform clock(clock input to eLBC controller).
284 CFG_SYS_FSL_DDR_SDRAM_BASE_PHY
285 Physical address from the view of DDR controllers. It is the
286 same as CFG_SYS_DDR_SDRAM_BASE for all Power SoCs. But
287 it could be different for ARM SoCs.
289 - ARM options:
290 CFG_SYS_EXCEPTION_VECTORS_HIGH
292 Select high exception vectors of the ARM core, e.g., do not
293 clear the V bit of the c1 register of CP15.
295 COUNTER_FREQUENCY
296 Generic timer clock source frequency.
298 COUNTER_FREQUENCY_REAL
299 Generic timer clock source frequency if the real clock is
300 different from COUNTER_FREQUENCY, and can only be determined
301 at run time.
303 - Linux Kernel Interface:
304 CONFIG_OF_LIBFDT
306 New kernel versions are expecting firmware settings to be
307 passed using flattened device trees (based on open firmware
308 concepts).
310 CONFIG_OF_LIBFDT
311 * New libfdt-based support
312 * Adds the "fdt" command
313 * The bootm command automatically updates the fdt
315 OF_TBCLK - The timebase frequency.
317 boards with QUICC Engines require OF_QE to set UCC MAC
318 addresses
320 CONFIG_OF_IDE_FIXUP
322 U-Boot can detect if an IDE device is present or not.
323 If not, and this new config option is activated, U-Boot
324 removes the ATA node from the DTS before booting Linux,
325 so the Linux IDE driver does not probe the device and
326 crash. This is needed for buggy hardware (uc101) where
327 no pull down resistor is connected to the signal IDE5V_DD7.
329 - vxWorks boot parameters:
331 bootvx constructs a valid bootline using the following
332 environments variables: bootdev, bootfile, ipaddr, netmask,
333 serverip, gatewayip, hostname, othbootargs.
334 It loads the vxWorks image pointed bootfile.
336 Note: If a "bootargs" environment is defined, it will override
337 the defaults discussed just above.
339 - Cache Configuration for ARM:
340 CFG_SYS_PL310_BASE - Physical base address of PL310
341 controller register space
343 - Serial Ports:
344 CFG_PL011_CLOCK
346 If you have Amba PrimeCell PL011 UARTs, set this variable to
347 the clock speed of the UARTs.
349 CFG_PL01x_PORTS
351 If you have Amba PrimeCell PL010 or PL011 UARTs on your board,
352 define this to a list of base addresses for each (supported)
353 port. See e.g. include/configs/versatile.h
355 CONFIG_SERIAL_HW_FLOW_CONTROL
357 Define this variable to enable hw flow control in serial driver.
358 Current user of this option is drivers/serial/nsl16550.c driver
360 - Removal of commands
361 If no commands are needed to boot, you can disable
362 CONFIG_CMDLINE to remove them. In this case, the command line
363 will not be available, and when U-Boot wants to execute the
364 boot command (on start-up) it will call board_run_command()
365 instead. This can reduce image size significantly for very
366 simple boot procedures.
368 - Regular expression support:
369 CONFIG_REGEX
370 If this variable is defined, U-Boot is linked against
371 the SLRE (Super Light Regular Expression) library,
372 which adds regex support to some commands, as for
373 example "env grep" and "setexpr".
375 - Watchdog:
376 CFG_SYS_WATCHDOG_FREQ
377 Some platforms automatically call WATCHDOG_RESET()
378 from the timer interrupt handler every
379 CFG_SYS_WATCHDOG_FREQ interrupts. If not set by the
380 board configuration file, a default of CONFIG_SYS_HZ/2
381 (i.e. 500) is used. Setting CFG_SYS_WATCHDOG_FREQ
382 to 0 disables calling WATCHDOG_RESET() from the timer
383 interrupt.
385 - GPIO Support:
386 The CFG_SYS_I2C_PCA953X_WIDTH option specifies a list of
387 chip-ngpio pairs that tell the PCA953X driver the number of
388 pins supported by a particular chip.
390 Note that if the GPIO device uses I2C, then the I2C interface
391 must also be configured. See I2C Support, below.
393 - I/O tracing:
394 When CONFIG_IO_TRACE is selected, U-Boot intercepts all I/O
395 accesses and can checksum them or write a list of them out
396 to memory. See the 'iotrace' command for details. This is
397 useful for testing device drivers since it can confirm that
398 the driver behaves the same way before and after a code
399 change. Currently this is supported on sandbox and arm. To
400 add support for your architecture, add '#include <iotrace.h>'
401 to the bottom of arch/<arch>/include/asm/io.h and test.
403 Example output from the 'iotrace stats' command is below.
404 Note that if the trace buffer is exhausted, the checksum will
405 still continue to operate.
407 iotrace is enabled
408 Start: 10000000 (buffer start address)
409 Size: 00010000 (buffer size)
410 Offset: 00000120 (current buffer offset)
411 Output: 10000120 (start + offset)
412 Count: 00000018 (number of trace records)
413 CRC32: 9526fb66 (CRC32 of all trace records)
415 - Timestamp Support:
417 When CONFIG_TIMESTAMP is selected, the timestamp
418 (date and time) of an image is printed by image
419 commands like bootm or iminfo. This option is
420 automatically enabled when you select CONFIG_CMD_DATE .
422 - Partition Labels (disklabels) Supported:
423 Zero or more of the following:
424 CONFIG_MAC_PARTITION Apple's MacOS partition table.
425 CONFIG_ISO_PARTITION ISO partition table, used on CDROM etc.
426 CONFIG_EFI_PARTITION GPT partition table, common when EFI is the
427 bootloader. Note 2TB partition limit; see
428 disk/part_efi.c
429 CONFIG_SCSI) you must configure support for at
430 least one non-MTD partition type as well.
432 - NETWORK Support (PCI):
433 CONFIG_E1000_SPI
434 Utility code for direct access to the SPI bus on Intel 8257x.
435 This does not do anything useful unless you set at least one
436 of CONFIG_CMD_E1000 or CONFIG_E1000_SPI_GENERIC.
438 CONFIG_NATSEMI
439 Support for National dp83815 chips.
441 CONFIG_NS8382X
442 Support for National dp8382[01] gigabit chips.
444 - NETWORK Support (other):
445 CONFIG_CALXEDA_XGMAC
446 Support for the Calxeda XGMAC device
448 CONFIG_LAN91C96
449 Support for SMSC's LAN91C96 chips.
451 CONFIG_LAN91C96_USE_32_BIT
452 Define this to enable 32 bit addressing
454 CFG_SYS_DAVINCI_EMAC_PHY_COUNT
455 Define this if you have more then 3 PHYs.
457 CONFIG_FTGMAC100
458 Support for Faraday's FTGMAC100 Gigabit SoC Ethernet
460 CONFIG_FTGMAC100_EGIGA
461 Define this to use GE link update with gigabit PHY.
462 Define this if FTGMAC100 is connected to gigabit PHY.
463 If your system has 10/100 PHY only, it might not occur
464 wrong behavior. Because PHY usually return timeout or
465 useless data when polling gigabit status and gigabit
466 control registers. This behavior won't affect the
467 correctnessof 10/100 link speed update.
469 CONFIG_SH_ETHER
470 Support for Renesas on-chip Ethernet controller
472 CFG_SH_ETHER_USE_PORT
473 Define the number of ports to be used
475 CFG_SH_ETHER_PHY_ADDR
476 Define the ETH PHY's address
478 CFG_SH_ETHER_CACHE_WRITEBACK
479 If this option is set, the driver enables cache flush.
481 - TPM Support:
482 CONFIG_TPM
483 Support TPM devices.
485 CONFIG_TPM_TIS_INFINEON
486 Support for Infineon i2c bus TPM devices. Only one device
487 per system is supported at this time.
489 CONFIG_TPM_TIS_I2C_BURST_LIMITATION
490 Define the burst count bytes upper limit
492 CONFIG_TPM_ST33ZP24
493 Support for STMicroelectronics TPM devices. Requires DM_TPM support.
495 CONFIG_TPM_ST33ZP24_I2C
496 Support for STMicroelectronics ST33ZP24 I2C devices.
497 Requires TPM_ST33ZP24 and I2C.
499 CONFIG_TPM_ST33ZP24_SPI
500 Support for STMicroelectronics ST33ZP24 SPI devices.
501 Requires TPM_ST33ZP24 and SPI.
503 CONFIG_TPM_ATMEL_TWI
504 Support for Atmel TWI TPM device. Requires I2C support.
506 CONFIG_TPM_TIS_LPC
507 Support for generic parallel port TPM devices. Only one device
508 per system is supported at this time.
510 CONFIG_TPM
511 Define this to enable the TPM support library which provides
512 functional interfaces to some TPM commands.
513 Requires support for a TPM device.
515 CONFIG_TPM_AUTH_SESSIONS
516 Define this to enable authorized functions in the TPM library.
517 Requires CONFIG_TPM and CONFIG_SHA1.
519 - USB Support:
520 At the moment only the UHCI host controller is
521 supported (PIP405, MIP405); define
522 CONFIG_USB_UHCI to enable it.
523 define CONFIG_USB_KEYBOARD to enable the USB Keyboard
524 and define CONFIG_USB_STORAGE to enable the USB
525 storage devices.
526 Note:
527 Supported are USB Keyboards and USB Floppy drives
528 (TEAC FD-05PUB).
530 CONFIG_USB_DWC2_REG_ADDR the physical CPU address of the DWC2
531 HW module registers.
533 - USB Device:
534 Define the below if you wish to use the USB console.
535 Once firmware is rebuilt from a serial console issue the
536 command "setenv stdin usbtty; setenv stdout usbtty" and
537 attach your USB cable. The Unix command "dmesg" should print
538 it has found a new device. The environment variable usbtty
539 can be set to gserial or cdc_acm to enable your device to
540 appear to a USB host as a Linux gserial device or a
541 Common Device Class Abstract Control Model serial device.
542 If you select usbtty = gserial you should be able to enumerate
543 a Linux host by
544 # modprobe usbserial vendor=0xVendorID product=0xProductID
545 else if using cdc_acm, simply setting the environment
546 variable usbtty to be cdc_acm should suffice. The following
547 might be defined in YourBoardName.h
549 If you have a USB-IF assigned VendorID then you may wish to
550 define your own vendor specific values either in BoardName.h
551 or directly in usbd_vendor_info.h. If you don't define
552 CONFIG_USBD_MANUFACTURER, CONFIG_USBD_PRODUCT_NAME,
553 CONFIG_USBD_VENDORID and CONFIG_USBD_PRODUCTID, then U-Boot
554 should pretend to be a Linux device to it's target host.
556 CONFIG_USBD_MANUFACTURER
557 Define this string as the name of your company for
558 - CONFIG_USBD_MANUFACTURER "my company"
560 CONFIG_USBD_PRODUCT_NAME
561 Define this string as the name of your product
562 - CONFIG_USBD_PRODUCT_NAME "acme usb device"
564 CONFIG_USBD_VENDORID
565 Define this as your assigned Vendor ID from the USB
566 Implementors Forum. This *must* be a genuine Vendor ID
567 to avoid polluting the USB namespace.
568 - CONFIG_USBD_VENDORID 0xFFFF
570 CONFIG_USBD_PRODUCTID
571 Define this as the unique Product ID
572 for your device
573 - CONFIG_USBD_PRODUCTID 0xFFFF
575 - ULPI Layer Support:
576 The ULPI (UTMI Low Pin (count) Interface) PHYs are supported via
577 the generic ULPI layer. The generic layer accesses the ULPI PHY
578 via the platform viewport, so you need both the genric layer and
579 the viewport enabled. Currently only Chipidea/ARC based
580 viewport is supported.
581 To enable the ULPI layer support, define CONFIG_USB_ULPI and
582 CONFIG_USB_ULPI_VIEWPORT in your board configuration file.
583 If your ULPI phy needs a different reference clock than the
584 standard 24 MHz then you have to define CFG_ULPI_REF_CLK to
585 the appropriate value in Hz.
587 - MMC Support:
588 CONFIG_SH_MMCIF
589 Support for Renesas on-chip MMCIF controller
591 CONFIG_SH_MMCIF_ADDR
592 Define the base address of MMCIF registers
594 CONFIG_SH_MMCIF_CLK
595 Define the clock frequency for MMCIF
597 - USB Device Firmware Update (DFU) class support:
598 CONFIG_DFU_OVER_USB
599 This enables the USB portion of the DFU USB class
601 CONFIG_DFU_NAND
602 This enables support for exposing NAND devices via DFU.
604 CONFIG_DFU_RAM
605 This enables support for exposing RAM via DFU.
606 Note: DFU spec refer to non-volatile memory usage, but
607 allow usages beyond the scope of spec - here RAM usage,
608 one that would help mostly the developer.
610 CONFIG_SYS_DFU_DATA_BUF_SIZE
611 Dfu transfer uses a buffer before writing data to the
612 raw storage device. Make the size (in bytes) of this buffer
613 configurable. The size of this buffer is also configurable
614 through the "dfu_bufsiz" environment variable.
616 CONFIG_SYS_DFU_MAX_FILE_SIZE
617 When updating files rather than the raw storage device,
618 we use a static buffer to copy the file into and then write
619 the buffer once we've been given the whole file. Define
620 this to the maximum filesize (in bytes) for the buffer.
621 Default is 4 MiB if undefined.
623 DFU_DEFAULT_POLL_TIMEOUT
624 Poll timeout [ms], is the timeout a device can send to the
625 host. The host must wait for this timeout before sending
626 a subsequent DFU_GET_STATUS request to the device.
628 DFU_MANIFEST_POLL_TIMEOUT
629 Poll timeout [ms], which the device sends to the host when
630 entering dfuMANIFEST state. Host waits this timeout, before
631 sending again an USB request to the device.
633 - Keyboard Support:
634 See Kconfig help for available keyboard drivers.
636 - MII/PHY support:
637 CONFIG_PHY_CLOCK_FREQ (ppc4xx)
639 The clock frequency of the MII bus
641 CONFIG_PHY_CMD_DELAY (ppc4xx)
643 Some PHY like Intel LXT971A need extra delay after
644 command issued before MII status register can be read
646 - BOOTP Recovery Mode:
647 CONFIG_BOOTP_RANDOM_DELAY
649 If you have many targets in a network that try to
650 boot using BOOTP, you may want to avoid that all
651 systems send out BOOTP requests at precisely the same
652 moment (which would happen for instance at recovery
653 from a power failure, when all systems will try to
654 boot, thus flooding the BOOTP server. Defining
655 CONFIG_BOOTP_RANDOM_DELAY causes a random delay to be
656 inserted before sending out BOOTP requests. The
657 following delays are inserted then:
659 1st BOOTP request: delay 0 ... 1 sec
660 2nd BOOTP request: delay 0 ... 2 sec
661 3rd BOOTP request: delay 0 ... 4 sec
662 4th and following
663 BOOTP requests: delay 0 ... 8 sec
665 CFG_BOOTP_ID_CACHE_SIZE
667 BOOTP packets are uniquely identified using a 32-bit ID. The
668 server will copy the ID from client requests to responses and
669 U-Boot will use this to determine if it is the destination of
670 an incoming response. Some servers will check that addresses
671 aren't in use before handing them out (usually using an ARP
672 ping) and therefore take up to a few hundred milliseconds to
673 respond. Network congestion may also influence the time it
674 takes for a response to make it back to the client. If that
675 time is too long, U-Boot will retransmit requests. In order
676 to allow earlier responses to still be accepted after these
677 retransmissions, U-Boot's BOOTP client keeps a small cache of
678 IDs. The CFG_BOOTP_ID_CACHE_SIZE controls the size of this
679 cache. The default is to keep IDs for up to four outstanding
680 requests. Increasing this will allow U-Boot to accept offers
681 from a BOOTP client in networks with unusually high latency.
683 - DHCP Advanced Options:
685 - Link-local IP address negotiation:
686 Negotiate with other link-local clients on the local network
687 for an address that doesn't require explicit configuration.
688 This is especially useful if a DHCP server cannot be guaranteed
689 to exist in all environments that the device must operate.
691 See doc/README.link-local for more information.
693 - MAC address from environment variables
695 FDT_SEQ_MACADDR_FROM_ENV
697 Fix-up device tree with MAC addresses fetched sequentially from
698 environment variables. This config work on assumption that
699 non-usable ethernet node of device-tree are either not present
700 or their status has been marked as "disabled".
702 - CDP Options:
703 CONFIG_CDP_DEVICE_ID
705 The device id used in CDP trigger frames.
707 CONFIG_CDP_DEVICE_ID_PREFIX
709 A two character string which is prefixed to the MAC address
710 of the device.
712 CONFIG_CDP_PORT_ID
714 A printf format string which contains the ascii name of
715 the port. Normally is set to "eth%d" which sets
716 eth0 for the first Ethernet, eth1 for the second etc.
718 CONFIG_CDP_CAPABILITIES
720 A 32bit integer which indicates the device capabilities;
721 0x00000010 for a normal host which does not forwards.
723 CONFIG_CDP_VERSION
725 An ascii string containing the version of the software.
727 CONFIG_CDP_PLATFORM
729 An ascii string containing the name of the platform.
731 CONFIG_CDP_TRIGGER
733 A 32bit integer sent on the trigger.
735 CONFIG_CDP_POWER_CONSUMPTION
737 A 16bit integer containing the power consumption of the
738 device in .1 of milliwatts.
740 CONFIG_CDP_APPLIANCE_VLAN_TYPE
742 A byte containing the id of the VLAN.
744 - Status LED: CONFIG_LED_STATUS
746 Several configurations allow to display the current
747 status using a LED. For instance, the LED will blink
748 fast while running U-Boot code, stop blinking as
749 soon as a reply to a BOOTP request was received, and
750 start blinking slow once the Linux kernel is running
751 (supported by a status LED driver in the Linux
752 kernel). Defining CONFIG_LED_STATUS enables this
753 feature in U-Boot.
755 Additional options:
757 CONFIG_LED_STATUS_GPIO
758 The status LED can be connected to a GPIO pin.
759 In such cases, the gpio_led driver can be used as a
760 status LED backend implementation. Define CONFIG_LED_STATUS_GPIO
761 to include the gpio_led driver in the U-Boot binary.
763 CFG_GPIO_LED_INVERTED_TABLE
764 Some GPIO connected LEDs may have inverted polarity in which
765 case the GPIO high value corresponds to LED off state and
766 GPIO low value corresponds to LED on state.
767 In such cases CFG_GPIO_LED_INVERTED_TABLE may be defined
768 with a list of GPIO LEDs that have inverted polarity.
770 - I2C Support:
771 CFG_SYS_NUM_I2C_BUSES
772 Hold the number of i2c buses you want to use.
774 CFG_SYS_I2C_DIRECT_BUS
775 define this, if you don't use i2c muxes on your hardware.
776 if CFG_SYS_I2C_MAX_HOPS is not defined or == 0 you can
777 omit this define.
779 CFG_SYS_I2C_MAX_HOPS
780 define how many muxes are maximal consecutively connected
781 on one i2c bus. If you not use i2c muxes, omit this
782 define.
784 CFG_SYS_I2C_BUSES
785 hold a list of buses you want to use, only used if
786 CFG_SYS_I2C_DIRECT_BUS is not defined, for example
787 a board with CFG_SYS_I2C_MAX_HOPS = 1 and
788 CFG_SYS_NUM_I2C_BUSES = 9:
790 CFG_SYS_I2C_BUSES {{0, {I2C_NULL_HOP}}, \
791 {0, {{I2C_MUX_PCA9547, 0x70, 1}}}, \
792 {0, {{I2C_MUX_PCA9547, 0x70, 2}}}, \
793 {0, {{I2C_MUX_PCA9547, 0x70, 3}}}, \
794 {0, {{I2C_MUX_PCA9547, 0x70, 4}}}, \
795 {0, {{I2C_MUX_PCA9547, 0x70, 5}}}, \
796 {1, {I2C_NULL_HOP}}, \
797 {1, {{I2C_MUX_PCA9544, 0x72, 1}}}, \
798 {1, {{I2C_MUX_PCA9544, 0x72, 2}}}, \
799 }
801 which defines
802 bus 0 on adapter 0 without a mux
803 bus 1 on adapter 0 with a PCA9547 on address 0x70 port 1
804 bus 2 on adapter 0 with a PCA9547 on address 0x70 port 2
805 bus 3 on adapter 0 with a PCA9547 on address 0x70 port 3
806 bus 4 on adapter 0 with a PCA9547 on address 0x70 port 4
807 bus 5 on adapter 0 with a PCA9547 on address 0x70 port 5
808 bus 6 on adapter 1 without a mux
809 bus 7 on adapter 1 with a PCA9544 on address 0x72 port 1
810 bus 8 on adapter 1 with a PCA9544 on address 0x72 port 2
812 If you do not have i2c muxes on your board, omit this define.
814 - Legacy I2C Support:
815 If you use the software i2c interface (CONFIG_SYS_I2C_SOFT)
816 then the following macros need to be defined (examples are
817 from include/configs/lwmon.h):
819 I2C_INIT
821 (Optional). Any commands necessary to enable the I2C
822 controller or configure ports.
824 eg: #define I2C_INIT (immr->im_cpm.cp_pbdir |= PB_SCL)
826 I2C_ACTIVE
828 The code necessary to make the I2C data line active
829 (driven). If the data line is open collector, this
830 define can be null.
832 eg: #define I2C_ACTIVE (immr->im_cpm.cp_pbdir |= PB_SDA)
834 I2C_TRISTATE
836 The code necessary to make the I2C data line tri-stated
837 (inactive). If the data line is open collector, this
838 define can be null.
840 eg: #define I2C_TRISTATE (immr->im_cpm.cp_pbdir &= ~PB_SDA)
842 I2C_READ
844 Code that returns true if the I2C data line is high,
845 false if it is low.
847 eg: #define I2C_READ ((immr->im_cpm.cp_pbdat & PB_SDA) != 0)
849 I2C_SDA(bit)
851 If <bit> is true, sets the I2C data line high. If it
852 is false, it clears it (low).
854 eg: #define I2C_SDA(bit) \
855 if(bit) immr->im_cpm.cp_pbdat |= PB_SDA; \
856 else immr->im_cpm.cp_pbdat &= ~PB_SDA
858 I2C_SCL(bit)
860 If <bit> is true, sets the I2C clock line high. If it
861 is false, it clears it (low).
863 eg: #define I2C_SCL(bit) \
864 if(bit) immr->im_cpm.cp_pbdat |= PB_SCL; \
865 else immr->im_cpm.cp_pbdat &= ~PB_SCL
867 I2C_DELAY
869 This delay is invoked four times per clock cycle so this
870 controls the rate of data transfer. The data rate thus
871 is 1 / (I2C_DELAY * 4). Often defined to be something
872 like:
874 #define I2C_DELAY udelay(2)
876 CONFIG_SOFT_I2C_GPIO_SCL / CONFIG_SOFT_I2C_GPIO_SDA
878 If your arch supports the generic GPIO framework (asm/gpio.h),
879 then you may alternatively define the two GPIOs that are to be
880 used as SCL / SDA. Any of the previous I2C_xxx macros will
881 have GPIO-based defaults assigned to them as appropriate.
883 You should define these to the GPIO value as given directly to
884 the generic GPIO functions.
886 CFG_I2C_MULTI_BUS
888 This option allows the use of multiple I2C buses, each of which
889 must have a controller. At any point in time, only one bus is
890 active. To switch to a different bus, use the 'i2c dev' command.
891 Note that bus numbering is zero-based.
893 CFG_SYS_I2C_NOPROBES
895 This option specifies a list of I2C devices that will be skipped
896 when the 'i2c probe' command is issued.
898 e.g.
899 #define CFG_SYS_I2C_NOPROBES {0x50,0x68}
901 will skip addresses 0x50 and 0x68 on a board with one I2C bus
903 CFG_SYS_RTC_BUS_NUM
905 If defined, then this indicates the I2C bus number for the RTC.
906 If not defined, then U-Boot assumes that RTC is on I2C bus 0.
908 CONFIG_SOFT_I2C_READ_REPEATED_START
910 defining this will force the i2c_read() function in
911 the soft_i2c driver to perform an I2C repeated start
912 between writing the address pointer and reading the
913 data. If this define is omitted the default behaviour
914 of doing a stop-start sequence will be used. Most I2C
915 devices can use either method, but some require one or
916 the other.
918 - SPI Support: CONFIG_SPI
920 Enables SPI driver (so far only tested with
921 SPI EEPROM, also an instance works with Crystal A/D and
922 D/As on the SACSng board)
924 CFG_SYS_SPI_MXC_WAIT
925 Timeout for waiting until spi transfer completed.
926 default: (CONFIG_SYS_HZ/100) /* 10 ms */
928 - FPGA Support: CONFIG_FPGA
930 Enables FPGA subsystem.
932 CONFIG_FPGA_<vendor>
934 Enables support for specific chip vendors.
935 (ALTERA, XILINX)
937 CONFIG_FPGA_<family>
939 Enables support for FPGA family.
940 (SPARTAN2, SPARTAN3, VIRTEX2, CYCLONE2, ACEX1K, ACEX)
942 CONFIG_SYS_FPGA_CHECK_BUSY
944 Enable checks on FPGA configuration interface busy
945 status by the configuration function. This option
946 will require a board or device specific function to
947 be written.
949 CFG_FPGA_DELAY
951 If defined, a function that provides delays in the FPGA
952 configuration driver.
954 CFG_SYS_FPGA_CHECK_ERROR
956 Check for configuration errors during FPGA bitfile
957 loading. For example, abort during Virtex II
958 configuration if the INIT_B line goes low (which
959 indicated a CRC error).
961 CFG_SYS_FPGA_WAIT_INIT
963 Maximum time to wait for the INIT_B line to de-assert
964 after PROB_B has been de-asserted during a Virtex II
965 FPGA configuration sequence. The default time is 500
966 ms.
968 CFG_SYS_FPGA_WAIT_BUSY
970 Maximum time to wait for BUSY to de-assert during
971 Virtex II FPGA configuration. The default is 5 ms.
973 CFG_SYS_FPGA_WAIT_CONFIG
975 Time to wait after FPGA configuration. The default is
976 200 ms.
978 - Vendor Parameter Protection:
980 U-Boot considers the values of the environment
981 variables "serial#" (Board Serial Number) and
982 "ethaddr" (Ethernet Address) to be parameters that
983 are set once by the board vendor / manufacturer, and
984 protects these variables from casual modification by
985 the user. Once set, these variables are read-only,
986 and write or delete attempts are rejected. You can
987 change this behaviour:
989 If CONFIG_ENV_OVERWRITE is #defined in your config
990 file, the write protection for vendor parameters is
991 completely disabled. Anybody can change or delete
992 these parameters.
994 The same can be accomplished in a more flexible way
995 for any variable by configuring the type of access
996 to allow for those variables in the ".flags" variable
997 or define CFG_ENV_FLAGS_LIST_STATIC.
999 - Protected RAM:
1000 CFG_PRAM
1002 Define this variable to enable the reservation of
1003 "protected RAM", i. e. RAM which is not overwritten
1004 by U-Boot. Define CFG_PRAM to hold the number of
1005 kB you want to reserve for pRAM. You can overwrite
1006 this default value by defining an environment
1007 variable "pram" to the number of kB you want to
1008 reserve. Note that the board info structure will
1009 still show the full amount of RAM. If pRAM is
1010 reserved, a new environment variable "mem" will
1011 automatically be defined to hold the amount of
1012 remaining RAM in a form that can be passed as boot
1013 argument to Linux, for instance like that:
1015 setenv bootargs ... mem=\${mem}
1016 saveenv
1018 This way you can tell Linux not to use this memory,
1019 either, which results in a memory region that will
1020 not be affected by reboots.
1022 *WARNING* If your board configuration uses automatic
1023 detection of the RAM size, you must make sure that
1024 this memory test is non-destructive. So far, the
1025 following board configurations are known to be
1026 "pRAM-clean":
1028 IVMS8, IVML24, SPD8xx,
1029 HERMES, IP860, RPXlite, LWMON,
1030 FLAGADM
1032 - Error Recovery:
1033 Note:
1035 In the current implementation, the local variables
1036 space and global environment variables space are
1037 separated. Local variables are those you define by
1038 simply typing `name=value'. To access a local
1039 variable later on, you have write `$name' or
1040 `${name}'; to execute the contents of a variable
1041 directly type `$name' at the command prompt.
1043 Global environment variables are those you use
1044 setenv/printenv to work with. To run a command stored
1045 in such a variable, you need to use the run command,
1046 and you must not use the '$' sign to access them.
1048 To store commands and special characters in a
1049 variable, please use double quotation marks
1050 surrounding the whole text of the variable, instead
1051 of the backslashes before semicolons and special
1052 symbols.
1054 - Default Environment:
1055 CFG_EXTRA_ENV_SETTINGS
1057 Define this to contain any number of null terminated
1058 strings (variable = value pairs) that will be part of
1059 the default environment compiled into the boot image.
1061 For example, place something like this in your
1062 board's config file:
1064 #define CFG_EXTRA_ENV_SETTINGS \
1065 "myvar1=value1\0" \
1066 "myvar2=value2\0"
1068 Warning: This method is based on knowledge about the
1069 internal format how the environment is stored by the
1070 U-Boot code. This is NOT an official, exported
1071 interface! Although it is unlikely that this format
1072 will change soon, there is no guarantee either.
1073 You better know what you are doing here.
1075 Note: overly (ab)use of the default environment is
1076 discouraged. Make sure to check other ways to preset
1077 the environment like the "source" command or the
1078 boot command first.
1080 CONFIG_DELAY_ENVIRONMENT
1082 Normally the environment is loaded when the board is
1083 initialised so that it is available to U-Boot. This inhibits
1084 that so that the environment is not available until
1085 explicitly loaded later by U-Boot code. With CONFIG_OF_CONTROL
1086 this is instead controlled by the value of
1087 /config/load-environment.
1089 - Automatic software updates via TFTP server
1090 CONFIG_UPDATE_TFTP
1091 CONFIG_UPDATE_TFTP_CNT_MAX
1092 CONFIG_UPDATE_TFTP_MSEC_MAX
1094 These options enable and control the auto-update feature;
1095 for a more detailed description refer to doc/README.update.
1097 - MTD Support (mtdparts command, UBI support)
1098 CONFIG_MTD_UBI_WL_THRESHOLD
1099 This parameter defines the maximum difference between the highest
1100 erase counter value and the lowest erase counter value of eraseblocks
1101 of UBI devices. When this threshold is exceeded, UBI starts performing
1102 wear leveling by means of moving data from eraseblock with low erase
1103 counter to eraseblocks with high erase counter.
1105 The default value should be OK for SLC NAND flashes, NOR flashes and
1106 other flashes which have eraseblock life-cycle 100000 or more.
1107 However, in case of MLC NAND flashes which typically have eraseblock
1108 life-cycle less than 10000, the threshold should be lessened (e.g.,
1109 to 128 or 256, although it does not have to be power of 2).
1111 default: 4096
1113 CONFIG_MTD_UBI_BEB_LIMIT
1114 This option specifies the maximum bad physical eraseblocks UBI
1115 expects on the MTD device (per 1024 eraseblocks). If the
1116 underlying flash does not admit of bad eraseblocks (e.g. NOR
1117 flash), this value is ignored.
1119 NAND datasheets often specify the minimum and maximum NVM
1120 (Number of Valid Blocks) for the flashes' endurance lifetime.
1121 The maximum expected bad eraseblocks per 1024 eraseblocks
1122 then can be calculated as "1024 * (1 - MinNVB / MaxNVB)",
1123 which gives 20 for most NANDs (MaxNVB is basically the total
1124 count of eraseblocks on the chip).
1126 To put it differently, if this value is 20, UBI will try to
1127 reserve about 1.9% of physical eraseblocks for bad blocks
1128 handling. And that will be 1.9% of eraseblocks on the entire
1129 NAND chip, not just the MTD partition UBI attaches. This means
1130 that if you have, say, a NAND flash chip admits maximum 40 bad
1131 eraseblocks, and it is split on two MTD partitions of the same
1132 size, UBI will reserve 40 eraseblocks when attaching a
1133 partition.
1135 default: 20
1137 CONFIG_MTD_UBI_FASTMAP
1138 Fastmap is a mechanism which allows attaching an UBI device
1139 in nearly constant time. Instead of scanning the whole MTD device it
1140 only has to locate a checkpoint (called fastmap) on the device.
1141 The on-flash fastmap contains all information needed to attach
1142 the device. Using fastmap makes only sense on large devices where
1143 attaching by scanning takes long. UBI will not automatically install
1144 a fastmap on old images, but you can set the UBI parameter
1145 CONFIG_MTD_UBI_FASTMAP_AUTOCONVERT to 1 if you want so. Please note
1146 that fastmap-enabled images are still usable with UBI implementations
1147 without fastmap support. On typical flash devices the whole fastmap
1148 fits into one PEB. UBI will reserve PEBs to hold two fastmaps.
1150 CONFIG_MTD_UBI_FASTMAP_AUTOCONVERT
1151 Set this parameter to enable fastmap automatically on images
1152 without a fastmap.
1153 default: 0
1155 CONFIG_MTD_UBI_FM_DEBUG
1156 Enable UBI fastmap debug
1157 default: 0
1159 - SPL framework
1160 CONFIG_SPL
1161 Enable building of SPL globally.
1163 CONFIG_SPL_PANIC_ON_RAW_IMAGE
1164 When defined, SPL will panic() if the image it has
1165 loaded does not have a signature.
1166 Defining this is useful when code which loads images
1167 in SPL cannot guarantee that absolutely all read errors
1168 will be caught.
1169 An example is the LPC32XX MLC NAND driver, which will
1170 consider that a completely unreadable NAND block is bad,
1171 and thus should be skipped silently.
1173 CONFIG_SPL_DISPLAY_PRINT
1174 For ARM, enable an optional function to print more information
1175 about the running system.
1177 CONFIG_SPL_MPC83XX_WAIT_FOR_NAND
1178 Set this for NAND SPL on PPC mpc83xx targets, so that
1179 start.S waits for the rest of the SPL to load before
1180 continuing (the hardware starts execution after just
1181 loading the first page rather than the full 4K).
1183 CONFIG_SPL_UBI
1184 Support for a lightweight UBI (fastmap) scanner and
1185 loader
1187 CONFIG_SYS_NAND_5_ADDR_CYCLE, CONFIG_SYS_NAND_PAGE_SIZE,
1188 CONFIG_SYS_NAND_OOBSIZE, CONFIG_SYS_NAND_BLOCK_SIZE,
1189 CONFIG_SYS_NAND_BAD_BLOCK_POS, CFG_SYS_NAND_ECCPOS,
1190 CFG_SYS_NAND_ECCSIZE, CFG_SYS_NAND_ECCBYTES
1191 Defines the size and behavior of the NAND that SPL uses
1192 to read U-Boot
1194 CFG_SYS_NAND_U_BOOT_DST
1195 Location in memory to load U-Boot to
1197 CFG_SYS_NAND_U_BOOT_SIZE
1198 Size of image to load
1200 CFG_SYS_NAND_U_BOOT_START
1201 Entry point in loaded image to jump to
1203 CONFIG_SPL_RAM_DEVICE
1204 Support for running image already present in ram, in SPL binary
1206 CONFIG_SPL_FIT_PRINT
1207 Printing information about a FIT image adds quite a bit of
1208 code to SPL. So this is normally disabled in SPL. Use this
1209 option to re-enable it. This will affect the output of the
1210 bootm command when booting a FIT image.
1212 - Interrupt support (PPC):
1214 There are common interrupt_init() and timer_interrupt()
1215 for all PPC archs. interrupt_init() calls interrupt_init_cpu()
1216 for CPU specific initialization. interrupt_init_cpu()
1217 should set decrementer_count to appropriate value. If
1218 CPU resets decrementer automatically after interrupt
1219 (ppc4xx) it should set decrementer_count to zero.
1220 timer_interrupt() calls timer_interrupt_cpu() for CPU
1221 specific handling. If board has watchdog / status_led
1222 / other_activity_monitor it works automatically from
1223 general timer_interrupt().
1226 Board initialization settings:
1227 ------------------------------
1229 During Initialization u-boot calls a number of board specific functions
1230 to allow the preparation of board specific prerequisites, e.g. pin setup
1231 before drivers are initialized. To enable these callbacks the
1232 following configuration macros have to be defined. Currently this is
1233 architecture specific, so please check arch/your_architecture/lib/board.c
1234 typically in board_init_f() and board_init_r().
1236 - CONFIG_BOARD_EARLY_INIT_F: Call board_early_init_f()
1237 - CONFIG_BOARD_EARLY_INIT_R: Call board_early_init_r()
1238 - CONFIG_BOARD_LATE_INIT: Call board_late_init()
1240 Configuration Settings:
1241 -----------------------
1243 - CONFIG_SYS_LONGHELP: Defined when you want long help messages included;
1244 undefine this when you're short of memory.
1246 - CFG_SYS_HELP_CMD_WIDTH: Defined when you want to override the default
1247 width of the commands listed in the 'help' command output.
1249 - CONFIG_SYS_PROMPT: This is what U-Boot prints on the console to
1250 prompt for user input.
1252 - CFG_SYS_BAUDRATE_TABLE:
1253 List of legal baudrate settings for this board.
1255 - CFG_SYS_MEM_RESERVE_SECURE
1256 Only implemented for ARMv8 for now.
1257 If defined, the size of CFG_SYS_MEM_RESERVE_SECURE memory
1258 is substracted from total RAM and won't be reported to OS.
1259 This memory can be used as secure memory. A variable
1260 gd->arch.secure_ram is used to track the location. In systems
1261 the RAM base is not zero, or RAM is divided into banks,
1262 this variable needs to be recalcuated to get the address.
1264 - CFG_SYS_SDRAM_BASE:
1265 Physical start address of SDRAM. _Must_ be 0 here.
1267 - CFG_SYS_FLASH_BASE:
1268 Physical start address of Flash memory.
1270 - CONFIG_SYS_MALLOC_LEN:
1271 Size of DRAM reserved for malloc() use.
1273 - CFG_SYS_BOOTMAPSZ:
1274 Maximum size of memory mapped by the startup code of
1275 the Linux kernel; all data that must be processed by
1276 the Linux kernel (bd_info, boot arguments, FDT blob if
1277 used) must be put below this limit, unless "bootm_low"
1278 environment variable is defined and non-zero. In such case
1279 all data for the Linux kernel must be between "bootm_low"
1280 and "bootm_low" + CFG_SYS_BOOTMAPSZ. The environment
1281 variable "bootm_mapsize" will override the value of
1282 CFG_SYS_BOOTMAPSZ. If CFG_SYS_BOOTMAPSZ is undefined,
1283 then the value in "bootm_size" will be used instead.
1285 - CONFIG_SYS_BOOT_GET_CMDLINE:
1286 Enables allocating and saving kernel cmdline in space between
1287 "bootm_low" and "bootm_low" + BOOTMAPSZ.
1289 - CONFIG_SYS_BOOT_GET_KBD:
1290 Enables allocating and saving a kernel copy of the bd_info in
1291 space between "bootm_low" and "bootm_low" + BOOTMAPSZ.
1293 - CONFIG_SYS_FLASH_PROTECTION
1294 If defined, hardware flash sectors protection is used
1295 instead of U-Boot software protection.
1297 - CONFIG_SYS_FLASH_CFI:
1298 Define if the flash driver uses extra elements in the
1299 common flash structure for storing flash geometry.
1301 - CONFIG_FLASH_CFI_DRIVER
1302 This option also enables the building of the cfi_flash driver
1303 in the drivers directory
1305 - CONFIG_FLASH_CFI_MTD
1306 This option enables the building of the cfi_mtd driver
1307 in the drivers directory. The driver exports CFI flash
1308 to the MTD layer.
1310 - CONFIG_SYS_FLASH_USE_BUFFER_WRITE
1311 Use buffered writes to flash.
1313 - CONFIG_ENV_FLAGS_LIST_DEFAULT
1314 - CFG_ENV_FLAGS_LIST_STATIC
1315 Enable validation of the values given to environment variables when
1316 calling env set. Variables can be restricted to only decimal,
1317 hexadecimal, or boolean. If CONFIG_CMD_NET is also defined,
1318 the variables can also be restricted to IP address or MAC address.
1320 The format of the list is:
1321 type_attribute = [s|d|x|b|i|m]
1322 access_attribute = [a|r|o|c]
1323 attributes = type_attribute[access_attribute]
1324 entry = variable_name[:attributes]
1325 list = entry[,list]
1327 The type attributes are:
1328 s - String (default)
1329 d - Decimal
1330 x - Hexadecimal
1331 b - Boolean ([1yYtT|0nNfF])
1332 i - IP address
1333 m - MAC address
1335 The access attributes are:
1336 a - Any (default)
1337 r - Read-only
1338 o - Write-once
1339 c - Change-default
1341 - CONFIG_ENV_FLAGS_LIST_DEFAULT
1342 Define this to a list (string) to define the ".flags"
1343 environment variable in the default or embedded environment.
1345 - CFG_ENV_FLAGS_LIST_STATIC
1346 Define this to a list (string) to define validation that
1347 should be done if an entry is not found in the ".flags"
1348 environment variable. To override a setting in the static
1349 list, simply add an entry for the same variable name to the
1350 ".flags" variable.
1352 If CONFIG_REGEX is defined, the variable_name above is evaluated as a
1353 regular expression. This allows multiple variables to define the same
1354 flags without explicitly listing them for each variable.
1356 The following definitions that deal with the placement and management
1357 of environment data (variable area); in general, we support the
1358 following configurations:
1360 BE CAREFUL! The first access to the environment happens quite early
1361 in U-Boot initialization (when we try to get the setting of for the
1362 console baudrate). You *MUST* have mapped your NVRAM area then, or
1363 U-Boot will hang.
1365 Please note that even with NVRAM we still use a copy of the
1366 environment in RAM: we could work on NVRAM directly, but we want to
1367 keep settings there always unmodified except somebody uses "saveenv"
1368 to save the current settings.
1370 BE CAREFUL! For some special cases, the local device can not use
1371 "saveenv" command. For example, the local device will get the
1372 environment stored in a remote NOR flash by SRIO or PCIE link,
1373 but it can not erase, write this NOR flash by SRIO or PCIE interface.
1375 - CONFIG_NAND_ENV_DST
1377 Defines address in RAM to which the nand_spl code should copy the
1378 environment. If redundant environment is used, it will be copied to
1379 CONFIG_NAND_ENV_DST + CONFIG_ENV_SIZE.
1381 Please note that the environment is read-only until the monitor
1382 has been relocated to RAM and a RAM copy of the environment has been
1383 created; also, when using EEPROM you will have to use env_get_f()
1384 until then to read environment variables.
1386 The environment is protected by a CRC32 checksum. Before the monitor
1387 is relocated into RAM, as a result of a bad CRC you will be working
1388 with the compiled-in default environment - *silently*!!! [This is
1389 necessary, because the first environment variable we need is the
1390 "baudrate" setting for the console - if we have a bad CRC, we don't
1391 have any device yet where we could complain.]
1393 Note: once the monitor has been relocated, then it will complain if
1394 the default environment is used; a new CRC is computed as soon as you
1395 use the "saveenv" command to store a valid environment.
1397 - CONFIG_SYS_FAULT_MII_ADDR:
1398 MII address of the PHY to check for the Ethernet link state.
1400 - CONFIG_DISPLAY_BOARDINFO
1401 Display information about the board that U-Boot is running on
1402 when U-Boot starts up. The board function checkboard() is called
1403 to do this.
1405 - CONFIG_DISPLAY_BOARDINFO_LATE
1406 Similar to the previous option, but display this information
1407 later, once stdio is running and output goes to the LCD, if
1408 present.
1410 Low Level (hardware related) configuration options:
1411 ---------------------------------------------------
1413 - CONFIG_SYS_CACHELINE_SIZE:
1414 Cache Line Size of the CPU.
1416 - CONFIG_SYS_CCSRBAR_DEFAULT:
1417 Default (power-on reset) physical address of CCSR on Freescale
1418 PowerPC SOCs.
1420 - CFG_SYS_CCSRBAR:
1421 Virtual address of CCSR. On a 32-bit build, this is typically
1422 the same value as CONFIG_SYS_CCSRBAR_DEFAULT.
1424 - CFG_SYS_CCSRBAR_PHYS:
1425 Physical address of CCSR. CCSR can be relocated to a new
1426 physical address, if desired. In this case, this macro should
1427 be set to that address. Otherwise, it should be set to the
1428 same value as CONFIG_SYS_CCSRBAR_DEFAULT. For example, CCSR
1429 is typically relocated on 36-bit builds. It is recommended
1430 that this macro be defined via the _HIGH and _LOW macros:
1432 #define CFG_SYS_CCSRBAR_PHYS ((CFG_SYS_CCSRBAR_PHYS_HIGH
1433 * 1ull) << 32 | CFG_SYS_CCSRBAR_PHYS_LOW)
1435 - CFG_SYS_CCSRBAR_PHYS_HIGH:
1436 Bits 33-36 of CFG_SYS_CCSRBAR_PHYS. This value is typically
1437 either 0 (32-bit build) or 0xF (36-bit build). This macro is
1438 used in assembly code, so it must not contain typecasts or
1439 integer size suffixes (e.g. "ULL").
1441 - CFG_SYS_CCSRBAR_PHYS_LOW:
1442 Lower 32-bits of CFG_SYS_CCSRBAR_PHYS. This macro is
1443 used in assembly code, so it must not contain typecasts or
1444 integer size suffixes (e.g. "ULL").
1446 - CONFIG_SYS_IMMR: Physical address of the Internal Memory.
1447 DO NOT CHANGE unless you know exactly what you're
1448 doing! (11-4) [MPC8xx systems only]
1450 - CFG_SYS_INIT_RAM_ADDR:
1452 Start address of memory area that can be used for
1453 initial data and stack; please note that this must be
1454 writable memory that is working WITHOUT special
1455 initialization, i. e. you CANNOT use normal RAM which
1456 will become available only after programming the
1457 memory controller and running certain initialization
1458 sequences.
1460 U-Boot uses the following memory types:
1461 - MPC8xx: IMMR (internal memory of the CPU)
1463 - CONFIG_SYS_SCCR: System Clock and reset Control Register (15-27)
1465 - CONFIG_SYS_OR_TIMING_SDRAM:
1466 SDRAM timing
1468 - CONFIG_SYS_SRIOn_MEM_VIRT:
1469 Virtual Address of SRIO port 'n' memory region
1471 - CONFIG_SYS_SRIOn_MEM_PHYxS:
1472 Physical Address of SRIO port 'n' memory region
1474 - CONFIG_SYS_SRIOn_MEM_SIZE:
1475 Size of SRIO port 'n' memory region
1477 - CONFIG_SYS_NAND_BUSWIDTH_16BIT
1478 Defined to tell the NAND controller that the NAND chip is using
1479 a 16 bit bus.
1480 Not all NAND drivers use this symbol.
1481 Example of drivers that use it:
1482 - drivers/mtd/nand/raw/ndfc.c
1483 - drivers/mtd/nand/raw/mxc_nand.c
1485 - CONFIG_SYS_NDFC_EBC0_CFG
1486 Sets the EBC0_CFG register for the NDFC. If not defined
1487 a default value will be used.
1489 - CONFIG_SYS_SPD_BUS_NUM
1490 If SPD EEPROM is on an I2C bus other than the first
1491 one, specify here. Note that the value must resolve
1492 to something your driver can deal with.
1494 - CONFIG_FSL_DDR_INTERACTIVE
1495 Enable interactive DDR debugging. See doc/README.fsl-ddr.
1497 - CONFIG_FSL_DDR_SYNC_REFRESH
1498 Enable sync of refresh for multiple controllers.
1500 - CONFIG_FSL_DDR_BIST
1501 Enable built-in memory test for Freescale DDR controllers.
1503 - CONFIG_RMII
1504 Enable RMII mode for all FECs.
1505 Note that this is a global option, we can't
1506 have one FEC in standard MII mode and another in RMII mode.
1508 - CONFIG_CRC32_VERIFY
1509 Add a verify option to the crc32 command.
1510 The syntax is:
1512 => crc32 -v <address> <count> <crc32>
1514 Where address/count indicate a memory area
1515 and crc32 is the correct crc32 which the
1516 area should have.
1518 - CONFIG_LOOPW
1519 Add the "loopw" memory command. This only takes effect if
1520 the memory commands are activated globally (CONFIG_CMD_MEMORY).
1522 - CONFIG_CMD_MX_CYCLIC
1523 Add the "mdc" and "mwc" memory commands. These are cyclic
1524 "md/mw" commands.
1525 Examples:
1527 => mdc.b 10 4 500
1528 This command will print 4 bytes (10,11,12,13) each 500 ms.
1530 => mwc.l 100 12345678 10
1531 This command will write 12345678 to address 100 all 10 ms.
1533 This only takes effect if the memory commands are activated
1534 globally (CONFIG_CMD_MEMORY).
1536 - CONFIG_SPL_BUILD
1537 Set when the currently running compilation is for an artifact
1538 that will end up in one of the 'xPL' builds, i.e. SPL, TPL or
1539 VPL. Code that needs phase-specific behaviour can check this,
1540 or (where possible) use spl_phase() instead.
1542 Note that CONFIG_SPL_BUILD *is* always defined when either
1543 of CONFIG_TPL_BUILD / CONFIG_VPL_BUILD is defined. This can be
1544 counter-intuitive and should perhaps be changed.
1546 - CONFIG_TPL_BUILD
1547 Set when the currently running compilation is for an artifact
1548 that will end up in the TPL build (as opposed to SPL, VPL or
1549 U-Boot proper). Code that needs phase-specific behaviour can
1550 check this, or (where possible) use spl_phase() instead.
1552 - CONFIG_VPL_BUILD
1553 Set when the currently running compilation is for an artifact
1554 that will end up in the VPL build (as opposed to the SPL, TPL
1555 or U-Boot proper). Code that needs phase-specific behaviour can
1556 check this, or (where possible) use spl_phase() instead.
1558 - CONFIG_ARCH_MAP_SYSMEM
1559 Generally U-Boot (and in particular the md command) uses
1560 effective address. It is therefore not necessary to regard
1561 U-Boot address as virtual addresses that need to be translated
1562 to physical addresses. However, sandbox requires this, since
1563 it maintains its own little RAM buffer which contains all
1564 addressable memory. This option causes some memory accesses
1565 to be mapped through map_sysmem() / unmap_sysmem().
1567 - CONFIG_X86_RESET_VECTOR
1568 If defined, the x86 reset vector code is included. This is not
1569 needed when U-Boot is running from Coreboot.
1571 Freescale QE/FMAN Firmware Support:
1572 -----------------------------------
1574 The Freescale QUICCEngine (QE) and Frame Manager (FMAN) both support the
1575 loading of "firmware", which is encoded in the QE firmware binary format.
1576 This firmware often needs to be loaded during U-Boot booting, so macros
1577 are used to identify the storage device (NOR flash, SPI, etc) and the address
1578 within that device.
1580 - CONFIG_SYS_FMAN_FW_ADDR
1581 The address in the storage device where the FMAN microcode is located. The
1582 meaning of this address depends on which CONFIG_SYS_QE_FMAN_FW_IN_xxx macro
1583 is also specified.
1585 - CONFIG_SYS_QE_FW_ADDR
1586 The address in the storage device where the QE microcode is located. The
1587 meaning of this address depends on which CONFIG_SYS_QE_FMAN_FW_IN_xxx macro
1588 is also specified.
1590 - CONFIG_SYS_QE_FMAN_FW_LENGTH
1591 The maximum possible size of the firmware. The firmware binary format
1592 has a field that specifies the actual size of the firmware, but it
1593 might not be possible to read any part of the firmware unless some
1594 local storage is allocated to hold the entire firmware first.
1596 - CONFIG_SYS_QE_FMAN_FW_IN_NOR
1597 Specifies that QE/FMAN firmware is located in NOR flash, mapped as
1598 normal addressable memory via the LBC. CONFIG_SYS_FMAN_FW_ADDR is the
1599 virtual address in NOR flash.
1601 - CONFIG_SYS_QE_FMAN_FW_IN_NAND
1602 Specifies that QE/FMAN firmware is located in NAND flash.
1603 CONFIG_SYS_FMAN_FW_ADDR is the offset within NAND flash.
1605 - CONFIG_SYS_QE_FMAN_FW_IN_MMC
1606 Specifies that QE/FMAN firmware is located on the primary SD/MMC
1607 device. CONFIG_SYS_FMAN_FW_ADDR is the byte offset on that device.
1609 - CONFIG_SYS_QE_FMAN_FW_IN_REMOTE
1610 Specifies that QE/FMAN firmware is located in the remote (master)
1611 memory space. CONFIG_SYS_FMAN_FW_ADDR is a virtual address which
1612 can be mapped from slave TLB->slave LAW->slave SRIO or PCIE outbound
1613 window->master inbound window->master LAW->the ucode address in
1614 master's memory space.
1616 Freescale Layerscape Management Complex Firmware Support:
1617 ---------------------------------------------------------
1618 The Freescale Layerscape Management Complex (MC) supports the loading of
1619 "firmware".
1620 This firmware often needs to be loaded during U-Boot booting, so macros
1621 are used to identify the storage device (NOR flash, SPI, etc) and the address
1622 within that device.
1624 - CONFIG_FSL_MC_ENET
1625 Enable the MC driver for Layerscape SoCs.
1627 Freescale Layerscape Debug Server Support:
1628 -------------------------------------------
1629 The Freescale Layerscape Debug Server Support supports the loading of
1630 "Debug Server firmware" and triggering SP boot-rom.
1631 This firmware often needs to be loaded during U-Boot booting.
1633 - CONFIG_SYS_MC_RSV_MEM_ALIGN
1634 Define alignment of reserved memory MC requires
1637 Building the Software:
1638 ======================
1640 Building U-Boot has been tested in several native build environments
1641 and in many different cross environments. Of course we cannot support
1642 all possibly existing versions of cross development tools in all
1643 (potentially obsolete) versions. In case of tool chain problems we
1644 recommend to use the ELDK (see https://www.denx.de/wiki/DULG/ELDK)
1645 which is extensively used to build and test U-Boot.
1647 If you are not using a native environment, it is assumed that you
1648 have GNU cross compiling tools available in your path. In this case,
1649 you must set the environment variable CROSS_COMPILE in your shell.
1650 Note that no changes to the Makefile or any other source files are
1651 necessary. For example using the ELDK on a 4xx CPU, please enter:
1653 $ CROSS_COMPILE=ppc_4xx-
1654 $ export CROSS_COMPILE
1656 U-Boot is intended to be simple to build. After installing the
1657 sources you must configure U-Boot for one specific board type. This
1658 is done by typing:
1660 make NAME_defconfig
1662 where "NAME_defconfig" is the name of one of the existing configu-
1663 rations; see configs/*_defconfig for supported names.
1665 Note: for some boards special configuration names may exist; check if
1666 additional information is available from the board vendor; for
1667 instance, the TQM823L systems are available without (standard)
1668 or with LCD support. You can select such additional "features"
1669 when choosing the configuration, i. e.
1671 make TQM823L_defconfig
1672 - will configure for a plain TQM823L, i. e. no LCD support
1674 make TQM823L_LCD_defconfig
1675 - will configure for a TQM823L with U-Boot console on LCD
1677 etc.
1680 Finally, type "make all", and you should get some working U-Boot
1681 images ready for download to / installation on your system:
1683 - "u-boot.bin" is a raw binary image
1684 - "u-boot" is an image in ELF binary format
1685 - "u-boot.srec" is in Motorola S-Record format
1687 By default the build is performed locally and the objects are saved
1688 in the source directory. One of the two methods can be used to change
1689 this behavior and build U-Boot to some external directory:
1691 1. Add O= to the make command line invocations:
1693 make O=/tmp/build distclean
1694 make O=/tmp/build NAME_defconfig
1695 make O=/tmp/build all
1697 2. Set environment variable KBUILD_OUTPUT to point to the desired location:
1699 export KBUILD_OUTPUT=/tmp/build
1700 make distclean
1701 make NAME_defconfig
1702 make all
1704 Note that the command line "O=" setting overrides the KBUILD_OUTPUT environment
1705 variable.
1707 User specific CPPFLAGS, AFLAGS and CFLAGS can be passed to the compiler by
1708 setting the according environment variables KCPPFLAGS, KAFLAGS and KCFLAGS.
1709 For example to treat all compiler warnings as errors:
1711 make KCFLAGS=-Werror
1713 Please be aware that the Makefiles assume you are using GNU make, so
1714 for instance on NetBSD you might need to use "gmake" instead of
1715 native "make".
1718 If the system board that you have is not listed, then you will need
1719 to port U-Boot to your hardware platform. To do this, follow these
1720 steps:
1722 1. Create a new directory to hold your board specific code. Add any
1723 files you need. In your board directory, you will need at least
1724 the "Makefile" and a "<board>.c".
1725 2. Create a new configuration file "include/configs/<board>.h" for
1726 your board.
1727 3. If you're porting U-Boot to a new CPU, then also create a new
1728 directory to hold your CPU specific code. Add any files you need.
1729 4. Run "make <board>_defconfig" with your new name.
1730 5. Type "make", and you should get a working "u-boot.srec" file
1731 to be installed on your target system.
1732 6. Debug and solve any problems that might arise.
1733 [Of course, this last step is much harder than it sounds.]
1736 Testing of U-Boot Modifications, Ports to New Hardware, etc.:
1737 ==============================================================
1739 If you have modified U-Boot sources (for instance added a new board
1740 or support for new devices, a new CPU, etc.) you are expected to
1741 provide feedback to the other developers. The feedback normally takes
1742 the form of a "patch", i.e. a context diff against a certain (latest
1743 official or latest in the git repository) version of U-Boot sources.
1745 But before you submit such a patch, please verify that your modifi-
1746 cation did not break existing code. At least make sure that *ALL* of
1747 the supported boards compile WITHOUT ANY compiler warnings. To do so,
1748 just run the buildman script (tools/buildman/buildman), which will
1749 configure and build U-Boot for ALL supported system. Be warned, this
1750 will take a while. Please see the buildman README, or run 'buildman -H'
1751 for documentation.
1754 See also "U-Boot Porting Guide" below.
1757 Monitor Commands - Overview:
1758 ============================
1760 go - start application at address 'addr'
1761 run - run commands in an environment variable
1762 bootm - boot application image from memory
1763 bootp - boot image via network using BootP/TFTP protocol
1764 bootz - boot zImage from memory
1765 tftpboot- boot image via network using TFTP protocol
1766 and env variables "ipaddr" and "serverip"
1767 (and eventually "gatewayip")
1768 tftpput - upload a file via network using TFTP protocol
1769 rarpboot- boot image via network using RARP/TFTP protocol
1770 diskboot- boot from IDE devicebootd - boot default, i.e., run 'bootcmd'
1771 loads - load S-Record file over serial line
1772 loadb - load binary file over serial line (kermit mode)
1773 loadm - load binary blob from source address to destination address
1774 md - memory display
1775 mm - memory modify (auto-incrementing)
1776 nm - memory modify (constant address)
1777 mw - memory write (fill)
1778 ms - memory search
1779 cp - memory copy
1780 cmp - memory compare
1781 crc32 - checksum calculation
1782 i2c - I2C sub-system
1783 sspi - SPI utility commands
1784 base - print or set address offset
1785 printenv- print environment variables
1786 pwm - control pwm channels
1787 seama - load SEAMA NAND image
1788 setenv - set environment variables
1789 saveenv - save environment variables to persistent storage
1790 protect - enable or disable FLASH write protection
1791 erase - erase FLASH memory
1792 flinfo - print FLASH memory information
1793 nand - NAND memory operations (see doc/README.nand)
1794 bdinfo - print Board Info structure
1795 iminfo - print header information for application image
1796 coninfo - print console devices and informations
1797 ide - IDE sub-system
1798 loop - infinite loop on address range
1799 loopw - infinite write loop on address range
1800 mtest - simple RAM test
1801 icache - enable or disable instruction cache
1802 dcache - enable or disable data cache
1803 reset - Perform RESET of the CPU
1804 echo - echo args to console
1805 version - print monitor version
1806 help - print online help
1807 ? - alias for 'help'
1810 Monitor Commands - Detailed Description:
1811 ========================================
1813 TODO.
1815 For now: just type "help <command>".
1818 Note for Redundant Ethernet Interfaces:
1819 =======================================
1821 Some boards come with redundant Ethernet interfaces; U-Boot supports
1822 such configurations and is capable of automatic selection of a
1823 "working" interface when needed. MAC assignment works as follows:
1825 Network interfaces are numbered eth0, eth1, eth2, ... Corresponding
1826 MAC addresses can be stored in the environment as "ethaddr" (=>eth0),
1827 "eth1addr" (=>eth1), "eth2addr", ...
1829 If the network interface stores some valid MAC address (for instance
1830 in SROM), this is used as default address if there is NO correspon-
1831 ding setting in the environment; if the corresponding environment
1832 variable is set, this overrides the settings in the card; that means:
1834 o If the SROM has a valid MAC address, and there is no address in the
1835 environment, the SROM's address is used.
1837 o If there is no valid address in the SROM, and a definition in the
1838 environment exists, then the value from the environment variable is
1839 used.
1841 o If both the SROM and the environment contain a MAC address, and
1842 both addresses are the same, this MAC address is used.
1844 o If both the SROM and the environment contain a MAC address, and the
1845 addresses differ, the value from the environment is used and a
1846 warning is printed.
1848 o If neither SROM nor the environment contain a MAC address, an error
1849 is raised. If CONFIG_NET_RANDOM_ETHADDR is defined, then in this case
1850 a random, locally-assigned MAC is used.
1852 If Ethernet drivers implement the 'write_hwaddr' function, valid MAC addresses
1853 will be programmed into hardware as part of the initialization process. This
1854 may be skipped by setting the appropriate 'ethmacskip' environment variable.
1855 The naming convention is as follows:
1856 "ethmacskip" (=>eth0), "eth1macskip" (=>eth1) etc.
1858 Image Formats:
1859 ==============
1861 U-Boot is capable of booting (and performing other auxiliary operations on)
1862 images in two formats:
1864 New uImage format (FIT)
1865 -----------------------
1867 Flexible and powerful format based on Flattened Image Tree -- FIT (similar
1868 to Flattened Device Tree). It allows the use of images with multiple
1869 components (several kernels, ramdisks, etc.), with contents protected by
1870 SHA1, MD5 or CRC32. More details are found in the doc/uImage.FIT directory.
1873 Old uImage format
1874 -----------------
1876 Old image format is based on binary files which can be basically anything,
1877 preceded by a special header; see the definitions in include/image.h for
1878 details; basically, the header defines the following image properties:
1880 * Target Operating System (Provisions for OpenBSD, NetBSD, FreeBSD,
1881 4.4BSD, Linux, SVR4, Esix, Solaris, Irix, SCO, Dell, NCR, VxWorks,
1882 LynxOS, pSOS, QNX, RTEMS, INTEGRITY;
1883 Currently supported: Linux, NetBSD, VxWorks, QNX, RTEMS, INTEGRITY).
1884 * Target CPU Architecture (Provisions for Alpha, ARM, Intel x86,
1885 IA64, MIPS, Nios II, PowerPC, IBM S390, SuperH, Sparc, Sparc 64 Bit;
1886 Currently supported: ARM, Intel x86, MIPS, Nios II, PowerPC).
1887 * Compression Type (uncompressed, gzip, bzip2)
1888 * Load Address
1889 * Entry Point
1890 * Image Name
1891 * Image Timestamp
1893 The header is marked by a special Magic Number, and both the header
1894 and the data portions of the image are secured against corruption by
1895 CRC32 checksums.
1898 Linux Support:
1899 ==============
1901 Although U-Boot should support any OS or standalone application
1902 easily, the main focus has always been on Linux during the design of
1903 U-Boot.
1905 U-Boot includes many features that so far have been part of some
1906 special "boot loader" code within the Linux kernel. Also, any
1907 "initrd" images to be used are no longer part of one big Linux image;
1908 instead, kernel and "initrd" are separate images. This implementation
1909 serves several purposes:
1911 - the same features can be used for other OS or standalone
1912 applications (for instance: using compressed images to reduce the
1913 Flash memory footprint)
1915 - it becomes much easier to port new Linux kernel versions because
1916 lots of low-level, hardware dependent stuff are done by U-Boot
1918 - the same Linux kernel image can now be used with different "initrd"
1919 images; of course this also means that different kernel images can
1920 be run with the same "initrd". This makes testing easier (you don't
1921 have to build a new "zImage.initrd" Linux image when you just
1922 change a file in your "initrd"). Also, a field-upgrade of the
1923 software is easier now.
1926 Linux HOWTO:
1927 ============
1929 Porting Linux to U-Boot based systems:
1930 ---------------------------------------
1932 U-Boot cannot save you from doing all the necessary modifications to
1933 configure the Linux device drivers for use with your target hardware
1934 (no, we don't intend to provide a full virtual machine interface to
1935 Linux :-).
1937 But now you can ignore ALL boot loader code (in arch/powerpc/mbxboot).
1939 Just make sure your machine specific header file (for instance
1940 include/asm-ppc/tqm8xx.h) includes the same definition of the Board
1941 Information structure as we define in include/asm-<arch>/u-boot.h,
1942 and make sure that your definition of IMAP_ADDR uses the same value
1943 as your U-Boot configuration in CONFIG_SYS_IMMR.
1945 Note that U-Boot now has a driver model, a unified model for drivers.
1946 If you are adding a new driver, plumb it into driver model. If there
1947 is no uclass available, you are encouraged to create one. See
1948 doc/driver-model.
1951 Configuring the Linux kernel:
1952 -----------------------------
1954 No specific requirements for U-Boot. Make sure you have some root
1955 device (initial ramdisk, NFS) for your target system.
1958 Building a Linux Image:
1959 -----------------------
1961 With U-Boot, "normal" build targets like "zImage" or "bzImage" are
1962 not used. If you use recent kernel source, a new build target
1963 "uImage" will exist which automatically builds an image usable by
1964 U-Boot. Most older kernels also have support for a "pImage" target,
1965 which was introduced for our predecessor project PPCBoot and uses a
1966 100% compatible format.
1968 Example:
1970 make TQM850L_defconfig
1971 make oldconfig
1972 make dep
1973 make uImage
1975 The "uImage" build target uses a special tool (in 'tools/mkimage') to
1976 encapsulate a compressed Linux kernel image with header information,
1977 CRC32 checksum etc. for use with U-Boot. This is what we are doing:
1979 * build a standard "vmlinux" kernel image (in ELF binary format):
1981 * convert the kernel into a raw binary image:
1983 ${CROSS_COMPILE}-objcopy -O binary \
1984 -R .note -R .comment \
1985 -S vmlinux linux.bin
1987 * compress the binary image:
1989 gzip -9 linux.bin
1991 * package compressed binary image for U-Boot:
1993 mkimage -A ppc -O linux -T kernel -C gzip \
1994 -a 0 -e 0 -n "Linux Kernel Image" \
1995 -d linux.bin.gz uImage
1998 The "mkimage" tool can also be used to create ramdisk images for use
1999 with U-Boot, either separated from the Linux kernel image, or
2000 combined into one file. "mkimage" encapsulates the images with a 64
2001 byte header containing information about target architecture,
2002 operating system, image type, compression method, entry points, time
2003 stamp, CRC32 checksums, etc.
2005 "mkimage" can be called in two ways: to verify existing images and
2006 print the header information, or to build new images.
2008 In the first form (with "-l" option) mkimage lists the information
2009 contained in the header of an existing U-Boot image; this includes
2010 checksum verification:
2012 tools/mkimage -l image
2013 -l ==> list image header information
2015 The second form (with "-d" option) is used to build a U-Boot image
2016 from a "data file" which is used as image payload:
2018 tools/mkimage -A arch -O os -T type -C comp -a addr -e ep \
2019 -n name -d data_file image
2020 -A ==> set architecture to 'arch'
2021 -O ==> set operating system to 'os'
2022 -T ==> set image type to 'type'
2023 -C ==> set compression type 'comp'
2024 -a ==> set load address to 'addr' (hex)
2025 -e ==> set entry point to 'ep' (hex)
2026 -n ==> set image name to 'name'
2027 -d ==> use image data from 'datafile'
2029 Right now, all Linux kernels for PowerPC systems use the same load
2030 address (0x00000000), but the entry point address depends on the
2031 kernel version:
2033 - 2.2.x kernels have the entry point at 0x0000000C,
2034 - 2.3.x and later kernels have the entry point at 0x00000000.
2036 So a typical call to build a U-Boot image would read:
2038 -> tools/mkimage -n '2.4.4 kernel for TQM850L' \
2039 > -A ppc -O linux -T kernel -C gzip -a 0 -e 0 \
2040 > -d /opt/elsk/ppc_8xx/usr/src/linux-2.4.4/arch/powerpc/coffboot/vmlinux.gz \
2041 > examples/uImage.TQM850L
2042 Image Name: 2.4.4 kernel for TQM850L
2043 Created: Wed Jul 19 02:34:59 2000
2044 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2045 Data Size: 335725 Bytes = 327.86 kB = 0.32 MB
2046 Load Address: 0x00000000
2047 Entry Point: 0x00000000
2049 To verify the contents of the image (or check for corruption):
2051 -> tools/mkimage -l examples/uImage.TQM850L
2052 Image Name: 2.4.4 kernel for TQM850L
2053 Created: Wed Jul 19 02:34:59 2000
2054 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2055 Data Size: 335725 Bytes = 327.86 kB = 0.32 MB
2056 Load Address: 0x00000000
2057 Entry Point: 0x00000000
2059 NOTE: for embedded systems where boot time is critical you can trade
2060 speed for memory and install an UNCOMPRESSED image instead: this
2061 needs more space in Flash, but boots much faster since it does not
2062 need to be uncompressed:
2064 -> gunzip /opt/elsk/ppc_8xx/usr/src/linux-2.4.4/arch/powerpc/coffboot/vmlinux.gz
2065 -> tools/mkimage -n '2.4.4 kernel for TQM850L' \
2066 > -A ppc -O linux -T kernel -C none -a 0 -e 0 \
2067 > -d /opt/elsk/ppc_8xx/usr/src/linux-2.4.4/arch/powerpc/coffboot/vmlinux \
2068 > examples/uImage.TQM850L-uncompressed
2069 Image Name: 2.4.4 kernel for TQM850L
2070 Created: Wed Jul 19 02:34:59 2000
2071 Image Type: PowerPC Linux Kernel Image (uncompressed)
2072 Data Size: 792160 Bytes = 773.59 kB = 0.76 MB
2073 Load Address: 0x00000000
2074 Entry Point: 0x00000000
2077 Similar you can build U-Boot images from a 'ramdisk.image.gz' file
2078 when your kernel is intended to use an initial ramdisk:
2080 -> tools/mkimage -n 'Simple Ramdisk Image' \
2081 > -A ppc -O linux -T ramdisk -C gzip \
2082 > -d /LinuxPPC/images/SIMPLE-ramdisk.image.gz examples/simple-initrd
2083 Image Name: Simple Ramdisk Image
2084 Created: Wed Jan 12 14:01:50 2000
2085 Image Type: PowerPC Linux RAMDisk Image (gzip compressed)
2086 Data Size: 566530 Bytes = 553.25 kB = 0.54 MB
2087 Load Address: 0x00000000
2088 Entry Point: 0x00000000
2090 The "dumpimage" tool can be used to disassemble or list the contents of images
2091 built by mkimage. See dumpimage's help output (-h) for details.
2093 Installing a Linux Image:
2094 -------------------------
2096 To downloading a U-Boot image over the serial (console) interface,
2097 you must convert the image to S-Record format:
2099 objcopy -I binary -O srec examples/image examples/image.srec
2101 The 'objcopy' does not understand the information in the U-Boot
2102 image header, so the resulting S-Record file will be relative to
2103 address 0x00000000. To load it to a given address, you need to
2104 specify the target address as 'offset' parameter with the 'loads'
2105 command.
2107 Example: install the image to address 0x40100000 (which on the
2108 TQM8xxL is in the first Flash bank):
2110 => erase 40100000 401FFFFF
2112 .......... done
2113 Erased 8 sectors
2115 => loads 40100000
2116 ## Ready for S-Record download ...
2117 ~>examples/image.srec
2118 1 2 3 4 5 6 7 8 9 10 11 12 13 ...
2119 ...
2120 15989 15990 15991 15992
2121 [file transfer complete]
2122 [connected]
2123 ## Start Addr = 0x00000000
2126 You can check the success of the download using the 'iminfo' command;
2127 this includes a checksum verification so you can be sure no data
2128 corruption happened:
2130 => imi 40100000
2132 ## Checking Image at 40100000 ...
2133 Image Name: 2.2.13 for initrd on TQM850L
2134 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2135 Data Size: 335725 Bytes = 327 kB = 0 MB
2136 Load Address: 00000000
2137 Entry Point: 0000000c
2138 Verifying Checksum ... OK
2141 Boot Linux:
2142 -----------
2144 The "bootm" command is used to boot an application that is stored in
2145 memory (RAM or Flash). In case of a Linux kernel image, the contents
2146 of the "bootargs" environment variable is passed to the kernel as
2147 parameters. You can check and modify this variable using the
2148 "printenv" and "setenv" commands:
2151 => printenv bootargs
2152 bootargs=root=/dev/ram
2154 => setenv bootargs root=/dev/nfs rw nfsroot=10.0.0.2:/LinuxPPC nfsaddrs=10.0.0.99:10.0.0.2
2156 => printenv bootargs
2157 bootargs=root=/dev/nfs rw nfsroot=10.0.0.2:/LinuxPPC nfsaddrs=10.0.0.99:10.0.0.2
2159 => bootm 40020000
2160 ## Booting Linux kernel at 40020000 ...
2161 Image Name: 2.2.13 for NFS on TQM850L
2162 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2163 Data Size: 381681 Bytes = 372 kB = 0 MB
2164 Load Address: 00000000
2165 Entry Point: 0000000c
2166 Verifying Checksum ... OK
2167 Uncompressing Kernel Image ... OK
2168 Linux version 2.2.13 (wd@denx.local.net) (gcc version 2.95.2 19991024 (release)) #1 Wed Jul 19 02:35:17 MEST 2000
2169 Boot arguments: root=/dev/nfs rw nfsroot=10.0.0.2:/LinuxPPC nfsaddrs=10.0.0.99:10.0.0.2
2170 time_init: decrementer frequency = 187500000/60
2171 Calibrating delay loop... 49.77 BogoMIPS
2172 Memory: 15208k available (700k kernel code, 444k data, 32k init) [c0000000,c1000000]
2173 ...
2175 If you want to boot a Linux kernel with initial RAM disk, you pass
2176 the memory addresses of both the kernel and the initrd image (PPBCOOT
2177 format!) to the "bootm" command:
2179 => imi 40100000 40200000
2181 ## Checking Image at 40100000 ...
2182 Image Name: 2.2.13 for initrd on TQM850L
2183 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2184 Data Size: 335725 Bytes = 327 kB = 0 MB
2185 Load Address: 00000000
2186 Entry Point: 0000000c
2187 Verifying Checksum ... OK
2189 ## Checking Image at 40200000 ...
2190 Image Name: Simple Ramdisk Image
2191 Image Type: PowerPC Linux RAMDisk Image (gzip compressed)
2192 Data Size: 566530 Bytes = 553 kB = 0 MB
2193 Load Address: 00000000
2194 Entry Point: 00000000
2195 Verifying Checksum ... OK
2197 => bootm 40100000 40200000
2198 ## Booting Linux kernel at 40100000 ...
2199 Image Name: 2.2.13 for initrd on TQM850L
2200 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2201 Data Size: 335725 Bytes = 327 kB = 0 MB
2202 Load Address: 00000000
2203 Entry Point: 0000000c
2204 Verifying Checksum ... OK
2205 Uncompressing Kernel Image ... OK
2206 ## Loading RAMDisk Image at 40200000 ...
2207 Image Name: Simple Ramdisk Image
2208 Image Type: PowerPC Linux RAMDisk Image (gzip compressed)
2209 Data Size: 566530 Bytes = 553 kB = 0 MB
2210 Load Address: 00000000
2211 Entry Point: 00000000
2212 Verifying Checksum ... OK
2213 Loading Ramdisk ... OK
2214 Linux version 2.2.13 (wd@denx.local.net) (gcc version 2.95.2 19991024 (release)) #1 Wed Jul 19 02:32:08 MEST 2000
2215 Boot arguments: root=/dev/ram
2216 time_init: decrementer frequency = 187500000/60
2217 Calibrating delay loop... 49.77 BogoMIPS
2218 ...
2219 RAMDISK: Compressed image found at block 0
2220 VFS: Mounted root (ext2 filesystem).
2222 bash#
2224 Boot Linux and pass a flat device tree:
2225 -----------
2227 First, U-Boot must be compiled with the appropriate defines. See the section
2228 titled "Linux Kernel Interface" above for a more in depth explanation. The
2229 following is an example of how to start a kernel and pass an updated
2230 flat device tree:
2232 => print oftaddr
2233 oftaddr=0x300000
2234 => print oft
2235 oft=oftrees/mpc8540ads.dtb
2236 => tftp $oftaddr $oft
2237 Speed: 1000, full duplex
2238 Using TSEC0 device
2239 TFTP from server 192.168.1.1; our IP address is 192.168.1.101
2240 Filename 'oftrees/mpc8540ads.dtb'.
2241 Load address: 0x300000
2242 Loading: #
2243 done
2244 Bytes transferred = 4106 (100a hex)
2245 => tftp $loadaddr $bootfile
2246 Speed: 1000, full duplex
2247 Using TSEC0 device
2248 TFTP from server 192.168.1.1; our IP address is 192.168.1.2
2249 Filename 'uImage'.
2250 Load address: 0x200000
2251 Loading:############
2252 done
2253 Bytes transferred = 1029407 (fb51f hex)
2254 => print loadaddr
2255 loadaddr=200000
2256 => print oftaddr
2257 oftaddr=0x300000
2258 => bootm $loadaddr - $oftaddr
2259 ## Booting image at 00200000 ...
2260 Image Name: Linux-2.6.17-dirty
2261 Image Type: PowerPC Linux Kernel Image (gzip compressed)
2262 Data Size: 1029343 Bytes = 1005.2 kB
2263 Load Address: 00000000
2264 Entry Point: 00000000
2265 Verifying Checksum ... OK
2266 Uncompressing Kernel Image ... OK
2267 Booting using flat device tree at 0x300000
2268 Using MPC85xx ADS machine description
2269 Memory CAM mapping: CAM0=256Mb, CAM1=256Mb, CAM2=0Mb residual: 0Mb
2270 [snip]
2273 More About U-Boot Image Types:
2274 ------------------------------
2276 U-Boot supports the following image types:
2278 "Standalone Programs" are directly runnable in the environment
2279 provided by U-Boot; it is expected that (if they behave
2280 well) you can continue to work in U-Boot after return from
2281 the Standalone Program.
2282 "OS Kernel Images" are usually images of some Embedded OS which
2283 will take over control completely. Usually these programs
2284 will install their own set of exception handlers, device
2285 drivers, set up the MMU, etc. - this means, that you cannot
2286 expect to re-enter U-Boot except by resetting the CPU.
2287 "RAMDisk Images" are more or less just data blocks, and their
2288 parameters (address, size) are passed to an OS kernel that is
2289 being started.
2290 "Multi-File Images" contain several images, typically an OS
2291 (Linux) kernel image and one or more data images like
2292 RAMDisks. This construct is useful for instance when you want
2293 to boot over the network using BOOTP etc., where the boot
2294 server provides just a single image file, but you want to get
2295 for instance an OS kernel and a RAMDisk image.
2297 "Multi-File Images" start with a list of image sizes, each
2298 image size (in bytes) specified by an "uint32_t" in network
2299 byte order. This list is terminated by an "(uint32_t)0".
2300 Immediately after the terminating 0 follow the images, one by
2301 one, all aligned on "uint32_t" boundaries (size rounded up to
2302 a multiple of 4 bytes).
2304 "Firmware Images" are binary images containing firmware (like
2305 U-Boot or FPGA images) which usually will be programmed to
2306 flash memory.
2308 "Script files" are command sequences that will be executed by
2309 U-Boot's command interpreter; this feature is especially
2310 useful when you configure U-Boot to use a real shell (hush)
2311 as command interpreter.
2313 Booting the Linux zImage:
2314 -------------------------
2316 On some platforms, it's possible to boot Linux zImage. This is done
2317 using the "bootz" command. The syntax of "bootz" command is the same
2318 as the syntax of "bootm" command.
2320 Note, defining the CONFIG_SUPPORT_RAW_INITRD allows user to supply
2321 kernel with raw initrd images. The syntax is slightly different, the
2322 address of the initrd must be augmented by it's size, in the following
2323 format: "<initrd addres>:<initrd size>".
2326 Standalone HOWTO:
2327 =================
2329 One of the features of U-Boot is that you can dynamically load and
2330 run "standalone" applications, which can use some resources of
2331 U-Boot like console I/O functions or interrupt services.
2333 Two simple examples are included with the sources:
2335 "Hello World" Demo:
2336 -------------------
2338 'examples/hello_world.c' contains a small "Hello World" Demo
2339 application; it is automatically compiled when you build U-Boot.
2340 It's configured to run at address 0x00040004, so you can play with it
2341 like that:
2343 => loads
2344 ## Ready for S-Record download ...
2345 ~>examples/hello_world.srec
2346 1 2 3 4 5 6 7 8 9 10 11 ...
2347 [file transfer complete]
2348 [connected]
2349 ## Start Addr = 0x00040004
2351 => go 40004 Hello World! This is a test.
2352 ## Starting application at 0x00040004 ...
2353 Hello World
2354 argc = 7
2355 argv[0] = "40004"
2356 argv[1] = "Hello"
2357 argv[2] = "World!"
2358 argv[3] = "This"
2359 argv[4] = "is"
2360 argv[5] = "a"
2361 argv[6] = "test."
2362 argv[7] = "<NULL>"
2363 Hit any key to exit ...
2365 ## Application terminated, rc = 0x0
2367 Another example, which demonstrates how to register a CPM interrupt
2368 handler with the U-Boot code, can be found in 'examples/timer.c'.
2369 Here, a CPM timer is set up to generate an interrupt every second.
2370 The interrupt service routine is trivial, just printing a '.'
2371 character, but this is just a demo program. The application can be
2372 controlled by the following keys:
2374 ? - print current values og the CPM Timer registers
2375 b - enable interrupts and start timer
2376 e - stop timer and disable interrupts
2377 q - quit application
2379 => loads
2380 ## Ready for S-Record download ...
2381 ~>examples/timer.srec
2382 1 2 3 4 5 6 7 8 9 10 11 ...
2383 [file transfer complete]
2384 [connected]
2385 ## Start Addr = 0x00040004
2387 => go 40004
2388 ## Starting application at 0x00040004 ...
2389 TIMERS=0xfff00980
2390 Using timer 1
2391 tgcr @ 0xfff00980, tmr @ 0xfff00990, trr @ 0xfff00994, tcr @ 0xfff00998, tcn @ 0xfff0099c, ter @ 0xfff009b0
2393 Hit 'b':
2394 [q, b, e, ?] Set interval 1000000 us
2395 Enabling timer
2396 Hit '?':
2397 [q, b, e, ?] ........
2398 tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0xef6, ter=0x0
2399 Hit '?':
2400 [q, b, e, ?] .
2401 tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0x2ad4, ter=0x0
2402 Hit '?':
2403 [q, b, e, ?] .
2404 tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0x1efc, ter=0x0
2405 Hit '?':
2406 [q, b, e, ?] .
2407 tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0x169d, ter=0x0
2408 Hit 'e':
2409 [q, b, e, ?] ...Stopping timer
2410 Hit 'q':
2411 [q, b, e, ?] ## Application terminated, rc = 0x0
2414 Implementation Internals:
2415 =========================
2417 The following is not intended to be a complete description of every
2418 implementation detail. However, it should help to understand the
2419 inner workings of U-Boot and make it easier to port it to custom
2420 hardware.
2423 Initial Stack, Global Data:
2424 ---------------------------
2426 The implementation of U-Boot is complicated by the fact that U-Boot
2427 starts running out of ROM (flash memory), usually without access to
2428 system RAM (because the memory controller is not initialized yet).
2429 This means that we don't have writable Data or BSS segments, and BSS
2430 is not initialized as zero. To be able to get a C environment working
2431 at all, we have to allocate at least a minimal stack. Implementation
2432 options for this are defined and restricted by the CPU used: Some CPU
2433 models provide on-chip memory (like the IMMR area on MPC8xx and
2434 MPC826x processors), on others (parts of) the data cache can be
2435 locked as (mis-) used as memory, etc.
2437 Chris Hallinan posted a good summary of these issues to the
2438 U-Boot mailing list:
2440 Subject: RE: [U-Boot-Users] RE: More On Memory Bank x (nothingness)?
2441 From: "Chris Hallinan" <clh@net1plus.com>
2442 Date: Mon, 10 Feb 2003 16:43:46 -0500 (22:43 MET)
2443 ...
2445 Correct me if I'm wrong, folks, but the way I understand it
2446 is this: Using DCACHE as initial RAM for Stack, etc, does not
2447 require any physical RAM backing up the cache. The cleverness
2448 is that the cache is being used as a temporary supply of
2449 necessary storage before the SDRAM controller is setup. It's
2450 beyond the scope of this list to explain the details, but you
2451 can see how this works by studying the cache architecture and
2452 operation in the architecture and processor-specific manuals.
2454 OCM is On Chip Memory, which I believe the 405GP has 4K. It
2455 is another option for the system designer to use as an
2456 initial stack/RAM area prior to SDRAM being available. Either
2457 option should work for you. Using CS 4 should be fine if your
2458 board designers haven't used it for something that would
2459 cause you grief during the initial boot! It is frequently not
2460 used.
2462 CFG_SYS_INIT_RAM_ADDR should be somewhere that won't interfere
2463 with your processor/board/system design. The default value
2464 you will find in any recent u-boot distribution in
2465 walnut.h should work for you. I'd set it to a value larger
2466 than your SDRAM module. If you have a 64MB SDRAM module, set
2467 it above 400_0000. Just make sure your board has no resources
2468 that are supposed to respond to that address! That code in
2469 start.S has been around a while and should work as is when
2470 you get the config right.
2472 -Chris Hallinan
2473 DS4.COM, Inc.
2475 It is essential to remember this, since it has some impact on the C
2476 code for the initialization procedures:
2478 * Initialized global data (data segment) is read-only. Do not attempt
2479 to write it.
2481 * Do not use any uninitialized global data (or implicitly initialized
2482 as zero data - BSS segment) at all - this is undefined, initiali-
2483 zation is performed later (when relocating to RAM).
2485 * Stack space is very limited. Avoid big data buffers or things like
2486 that.
2488 Having only the stack as writable memory limits means we cannot use
2489 normal global data to share information between the code. But it
2490 turned out that the implementation of U-Boot can be greatly
2491 simplified by making a global data structure (gd_t) available to all
2492 functions. We could pass a pointer to this data as argument to _all_
2493 functions, but this would bloat the code. Instead we use a feature of
2494 the GCC compiler (Global Register Variables) to share the data: we
2495 place a pointer (gd) to the global data into a register which we
2496 reserve for this purpose.
2498 When choosing a register for such a purpose we are restricted by the
2499 relevant (E)ABI specifications for the current architecture, and by
2500 GCC's implementation.
2502 For PowerPC, the following registers have specific use:
2503 R1: stack pointer
2504 R2: reserved for system use
2505 R3-R4: parameter passing and return values
2506 R5-R10: parameter passing
2507 R13: small data area pointer
2508 R30: GOT pointer
2509 R31: frame pointer
2511 (U-Boot also uses R12 as internal GOT pointer. r12
2512 is a volatile register so r12 needs to be reset when
2513 going back and forth between asm and C)
2515 ==> U-Boot will use R2 to hold a pointer to the global data
2517 Note: on PPC, we could use a static initializer (since the
2518 address of the global data structure is known at compile time),
2519 but it turned out that reserving a register results in somewhat
2520 smaller code - although the code savings are not that big (on
2521 average for all boards 752 bytes for the whole U-Boot image,
2522 624 text + 127 data).
2524 On ARM, the following registers are used:
2526 R0: function argument word/integer result
2527 R1-R3: function argument word
2528 R9: platform specific
2529 R10: stack limit (used only if stack checking is enabled)
2530 R11: argument (frame) pointer
2531 R12: temporary workspace
2532 R13: stack pointer
2533 R14: link register
2534 R15: program counter
2536 ==> U-Boot will use R9 to hold a pointer to the global data
2538 Note: on ARM, only R_ARM_RELATIVE relocations are supported.
2540 On Nios II, the ABI is documented here:
2541 https://www.altera.com/literature/hb/nios2/n2cpu_nii51016.pdf
2543 ==> U-Boot will use gp to hold a pointer to the global data
2545 Note: on Nios II, we give "-G0" option to gcc and don't use gp
2546 to access small data sections, so gp is free.
2548 On RISC-V, the following registers are used:
2550 x0: hard-wired zero (zero)
2551 x1: return address (ra)
2552 x2: stack pointer (sp)
2553 x3: global pointer (gp)
2554 x4: thread pointer (tp)
2555 x5: link register (t0)
2556 x8: frame pointer (fp)
2557 x10-x11: arguments/return values (a0-1)
2558 x12-x17: arguments (a2-7)
2559 x28-31: temporaries (t3-6)
2560 pc: program counter (pc)
2562 ==> U-Boot will use gp to hold a pointer to the global data
2564 Memory Management:
2565 ------------------
2567 U-Boot runs in system state and uses physical addresses, i.e. the
2568 MMU is not used either for address mapping nor for memory protection.
2570 The available memory is mapped to fixed addresses using the memory
2571 controller. In this process, a contiguous block is formed for each
2572 memory type (Flash, SDRAM, SRAM), even when it consists of several
2573 physical memory banks.
2575 U-Boot is installed in the first 128 kB of the first Flash bank (on
2576 TQM8xxL modules this is the range 0x40000000 ... 0x4001FFFF). After
2577 booting and sizing and initializing DRAM, the code relocates itself
2578 to the upper end of DRAM. Immediately below the U-Boot code some
2579 memory is reserved for use by malloc() [see CONFIG_SYS_MALLOC_LEN
2580 configuration setting]. Below that, a structure with global Board
2581 Info data is placed, followed by the stack (growing downward).
2583 Additionally, some exception handler code is copied to the low 8 kB
2584 of DRAM (0x00000000 ... 0x00001FFF).
2586 So a typical memory configuration with 16 MB of DRAM could look like
2587 this:
2589 0x0000 0000 Exception Vector code
2590 :
2591 0x0000 1FFF
2592 0x0000 2000 Free for Application Use
2593 :
2594 :
2596 :
2597 :
2598 0x00FB FF20 Monitor Stack (Growing downward)
2599 0x00FB FFAC Board Info Data and permanent copy of global data
2600 0x00FC 0000 Malloc Arena
2601 :
2602 0x00FD FFFF
2603 0x00FE 0000 RAM Copy of Monitor Code
2604 ... eventually: LCD or video framebuffer
2605 ... eventually: pRAM (Protected RAM - unchanged by reset)
2606 0x00FF FFFF [End of RAM]
2609 System Initialization:
2610 ----------------------
2612 In the reset configuration, U-Boot starts at the reset entry point
2613 (on most PowerPC systems at address 0x00000100). Because of the reset
2614 configuration for CS0# this is a mirror of the on board Flash memory.
2615 To be able to re-map memory U-Boot then jumps to its link address.
2616 To be able to implement the initialization code in C, a (small!)
2617 initial stack is set up in the internal Dual Ported RAM (in case CPUs
2618 which provide such a feature like), or in a locked part of the data
2619 cache. After that, U-Boot initializes the CPU core, the caches and
2620 the SIU.
2622 Next, all (potentially) available memory banks are mapped using a
2623 preliminary mapping. For example, we put them on 512 MB boundaries
2624 (multiples of 0x20000000: SDRAM on 0x00000000 and 0x20000000, Flash
2625 on 0x40000000 and 0x60000000, SRAM on 0x80000000). Then UPM A is
2626 programmed for SDRAM access. Using the temporary configuration, a
2627 simple memory test is run that determines the size of the SDRAM
2628 banks.
2630 When there is more than one SDRAM bank, and the banks are of
2631 different size, the largest is mapped first. For equal size, the first
2632 bank (CS2#) is mapped first. The first mapping is always for address
2633 0x00000000, with any additional banks following immediately to create
2634 contiguous memory starting from 0.
2636 Then, the monitor installs itself at the upper end of the SDRAM area
2637 and allocates memory for use by malloc() and for the global Board
2638 Info data; also, the exception vector code is copied to the low RAM
2639 pages, and the final stack is set up.
2641 Only after this relocation will you have a "normal" C environment;
2642 until that you are restricted in several ways, mostly because you are
2643 running from ROM, and because the code will have to be relocated to a
2644 new address in RAM.
2647 Contributing
2648 ============
2650 The U-Boot projects depends on contributions from the user community.
2651 If you want to participate, please, have a look at the 'General'
2652 section of https://docs.u-boot.org/en/latest/develop/index.html
2653 where we describe coding standards and the patch submission process.