Tom Rini | 5d4dbf1 | 2013-07-08 12:15:18 -0400 | [diff] [blame] | 1 | Summary |
| 2 | ======= |
| 3 | |
| 4 | This document covers various features of the 'am335x_evm' build, and some of |
| 5 | the related build targets (am335x_evm_uartN, etc). |
| 6 | |
| 7 | Hardware |
| 8 | ======== |
| 9 | |
| 10 | The binary produced by this board supports, based on parsing of the EEPROM |
| 11 | documented in TI's reference designs: |
| 12 | - AM335x GP EVM |
| 13 | - AM335x EVM SK |
| 14 | - Beaglebone White |
| 15 | - Beaglebone Black |
Tom Rini | d7ccfc5 | 2013-08-20 08:53:50 -0400 | [diff] [blame] | 16 | |
| 17 | Customization |
| 18 | ============= |
| 19 | |
| 20 | Given that all of the above boards are reference platforms (and the |
| 21 | Beaglebone platforms are OSHA), it is likely that this platform code and |
| 22 | configuration will be used as the basis of a custom platform. It is |
| 23 | worth noting that aside from things such as NAND or MMC only being |
| 24 | required if a custom platform makes use of these blocks, the following |
| 25 | are required, depending on design: |
| 26 | |
| 27 | - GPIO is only required if DDR3 power is controlled in a way similar to |
| 28 | EVM SK |
| 29 | - SPI is only required for SPI flash, or exposing the SPI bus. |
| 30 | |
| 31 | The following blocks are required: |
| 32 | - I2C, to talk with the PMIC and ensure that we do not run afoul of |
| 33 | errata 1.0.24. |
| 34 | |
| 35 | When removing options as part of customization, |
| 36 | CONFIG_EXTRA_ENV_SETTINGS will need additional care to update for your |
| 37 | needs and to remove no longer relevant options as in some cases we |
| 38 | define additional text blocks (such as for NAND or DFU strings). Also |
| 39 | note that all of the SPL options are grouped together, rather than with |
| 40 | the IP blocks, so both areas will need their choices updated to reflect |
| 41 | the custom design. |
| 42 | |
Tom Rini | 97559b5 | 2013-07-17 12:24:30 -0400 | [diff] [blame] | 43 | NAND |
| 44 | ==== |
| 45 | |
| 46 | The AM335x GP EVM ships with a 256MiB NAND available in most profiles. In |
| 47 | this example to program the NAND we assume that an SD card has been |
| 48 | inserted with the files to write in the first SD slot and that mtdparts |
| 49 | have been configured correctly for the board. As a time saving measure we |
| 50 | load MLO into memory in one location, copy it into the three locatations |
| 51 | that the ROM checks for additional valid copies, then load U-Boot into |
| 52 | memory. We then write that whole section of memory to NAND. |
| 53 | |
| 54 | U-Boot # mmc rescan |
| 55 | U-Boot # env default -f -a |
| 56 | U-Boot # nand erase.chip |
| 57 | U-Boot # saveenv |
| 58 | U-Boot # load mmc 0 81000000 MLO |
| 59 | U-Boot # cp.b 81000000 81020000 20000 |
| 60 | U-Boot # cp.b 81000000 81040000 20000 |
| 61 | U-Boot # cp.b 81000000 81060000 20000 |
| 62 | U-Boot # load mmc 0 81080000 u-boot.img |
| 63 | U-Boot # nand write 81000000 0 260000 |
| 64 | U-Boot # load mmc 0 ${loadaddr} uImage |
| 65 | U-Boot # nand write ${loadaddr} kernel 500000 |
Tom Rini | 5d4dbf1 | 2013-07-08 12:15:18 -0400 | [diff] [blame] | 66 | |
Tom Rini | 486da22 | 2013-07-18 15:13:05 -0400 | [diff] [blame] | 67 | NOR |
| 68 | === |
| 69 | |
| 70 | The Beaglebone White can be equiped with a "memory cape" that in turn can |
| 71 | have a NOR module plugged into it. In this case it is then possible to |
| 72 | program and boot from NOR. Note that due to how U-Boot is architectured we |
| 73 | must build a specific version of U-Boot that knows we have NOR flash. This |
| 74 | build is named 'am335x_evm_nor'. Further, we have a 'am335x_evm_norboot' |
| 75 | build that will assume that the environment is on NOR rather than NAND. In |
| 76 | the following example we assume that and SD card has been populated with |
| 77 | MLO and u-boot.img from a 'am335x_evm_nor' build and also contains the |
| 78 | 'u-boot.bin' from a 'am335x_evm_norboot' build. When booting from NOR, a |
| 79 | binary must be written to the start of NOR, with no header or similar |
| 80 | prepended. In the following example we use a size of 512KiB (0x80000) |
| 81 | as that is how much space we set aside before the environment, as per |
| 82 | the config file. |
| 83 | |
| 84 | U-Boot # mmc rescan |
| 85 | U-Boot # load mmc 0 ${loadaddr} u-boot.bin |
| 86 | U-Boot # protect off 08000000 +80000 |
| 87 | U-Boot # erase 08000000 +80000 |
| 88 | U-Boot # cp.b ${loadaddr} 08000000 ${filesize} |
| 89 | |
Tom Rini | 5d4dbf1 | 2013-07-08 12:15:18 -0400 | [diff] [blame] | 90 | Falcon Mode |
| 91 | =========== |
| 92 | |
| 93 | The default build includes "Falcon Mode" (see doc/README.falcon) via NAND, |
| 94 | eMMC (or raw SD cards) and FAT SD cards. Our default behavior currently is |
| 95 | to read a 'c' on the console while in SPL at any point prior to loading the |
| 96 | OS payload (so as soon as possible) to opt to booting full U-Boot. Also |
| 97 | note that while one can program Falcon Mode "in place" great care needs to |
| 98 | be taken by the user to not 'brick' their setup. As these are all eval |
| 99 | boards with multiple boot methods, recovery should not be an issue in this |
| 100 | worst-case however. |
| 101 | |
| 102 | Falcon Mode: eMMC |
| 103 | ================= |
| 104 | |
| 105 | The recommended layout in this case is: |
| 106 | |
| 107 | MMC BLOCKS |--------------------------------| LOCATION IN BYTES |
| 108 | 0x0000 - 0x007F : MBR or GPT table : 0x000000 - 0x020000 |
| 109 | 0x0080 - 0x00FF : ARGS or FDT file : 0x010000 - 0x020000 |
| 110 | 0x0100 - 0x01FF : SPL.backup1 (first copy used) : 0x020000 - 0x040000 |
| 111 | 0x0200 - 0x02FF : SPL.backup2 (second copy used) : 0x040000 - 0x060000 |
| 112 | 0x0300 - 0x06FF : U-Boot : 0x060000 - 0x0e0000 |
| 113 | 0x0700 - 0x08FF : U-Boot Env + Redundant : 0x0e0000 - 0x120000 |
| 114 | 0x0900 - 0x28FF : Kernel : 0x120000 - 0x520000 |
| 115 | |
| 116 | Note that when we run 'spl export' it will prepare to boot the kernel. |
| 117 | This includes relocation of the uImage from where we loaded it to the entry |
| 118 | point defined in the header. As these locations overlap by default, it |
| 119 | would leave us with an image that if written to MMC will not boot, so |
| 120 | instead of using the loadaddr variable we use 0x81000000 in the following |
| 121 | example. In this example we are loading from the network, for simplicity, |
| 122 | and assume a valid partition table already exists and 'mmc dev' has already |
| 123 | been run to select the correct device. Also note that if you previously |
| 124 | had a FAT partition (such as on a Beaglebone Black) it is not enough to |
| 125 | write garbage into the area, you must delete it from the partition table |
| 126 | first. |
| 127 | |
| 128 | # Ensure we are able to talk with this mmc device |
| 129 | U-Boot # mmc rescan |
| 130 | U-Boot # tftp 81000000 am335x/MLO |
| 131 | # Write to two of the backup locations ROM uses |
| 132 | U-Boot # mmc write 81000000 100 100 |
| 133 | U-Boot # mmc write 81000000 200 100 |
| 134 | # Write U-Boot to the location set in the config |
| 135 | U-Boot # tftp 81000000 am335x/u-boot.img |
| 136 | U-Boot # mmc write 81000000 300 400 |
| 137 | # Load kernel and device tree into memory, perform export |
| 138 | U-Boot # tftp 81000000 am335x/uImage |
| 139 | U-Boot # run findfdt |
| 140 | U-Boot # tftp ${fdtaddr} am335x/${fdtfile} |
| 141 | U-Boot # run mmcargs |
| 142 | U-Boot # spl export fdt 81000000 - ${fdtaddr} |
| 143 | # Write the updated device tree to MMC |
| 144 | U-Boot # mmc write ${fdtaddr} 80 80 |
| 145 | # Write the uImage to MMC |
| 146 | U-Boot # mmc write 81000000 900 2000 |
| 147 | |
| 148 | Falcon Mode: FAT SD cards |
| 149 | ========================= |
| 150 | |
| 151 | In this case the additional file is written to the filesystem. In this |
| 152 | example we assume that the uImage and device tree to be used are already on |
| 153 | the FAT filesystem (only the uImage MUST be for this to function |
| 154 | afterwards) along with a Falcon Mode aware MLO and the FAT partition has |
| 155 | already been created and marked bootable: |
| 156 | |
| 157 | U-Boot # mmc rescan |
| 158 | # Load kernel and device tree into memory, perform export |
| 159 | U-Boot # load mmc 0:1 ${loadaddr} uImage |
| 160 | U-Boot # run findfdt |
| 161 | U-Boot # load mmc 0:1 ${fdtaddr} ${fdtfile} |
| 162 | U-Boot # run mmcargs |
| 163 | U-Boot # spl export fdt ${loadaddr} - ${fdtaddr} |
| 164 | |
| 165 | This will print a number of lines and then end with something like: |
| 166 | Using Device Tree in place at 80f80000, end 80f85928 |
| 167 | Using Device Tree in place at 80f80000, end 80f88928 |
| 168 | So then you: |
| 169 | |
| 170 | U-Boot # fatwrite mmc 0:1 0x80f80000 args 8928 |
| 171 | |
| 172 | Falcon Mode: NAND |
| 173 | ================= |
| 174 | |
| 175 | In this case the additional data is written to another partition of the |
| 176 | NAND. In this example we assume that the uImage and device tree to be are |
| 177 | already located on the NAND somewhere (such as fileystem or mtd partition) |
| 178 | along with a Falcon Mode aware MLO written to the correct locations for |
| 179 | booting and mtdparts have been configured correctly for the board: |
| 180 | |
| 181 | U-Boot # nand read ${loadaddr} kernel |
| 182 | U-Boot # load nand rootfs ${fdtaddr} /boot/am335x-evm.dtb |
| 183 | U-Boot # run nandargs |
| 184 | U-Boot # spl export fdt ${loadaddr} - ${fdtaddr} |
| 185 | U-Boot # nand erase.part u-boot-spl-os |
| 186 | U-Boot # nand write ${fdtaddr} u-boot-spl-os |