| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 1 | By Thomas.Lange@corelatus.se 2004-Oct-05 |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 2 | ---------------------------------------- |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 3 | DbAu1xx0 are development boards from AMD containing |
| 4 | an Alchemy AU1xx0 series cpu with mips32 core. |
| 5 | Existing cpu:s are Au1000, Au1100, Au1500 and Au1550 |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 6 | |
| 7 | Limitations & comments |
| 8 | ---------------------- |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 9 | Support was originally big endian only. |
| 10 | I have not tested, but several u-boot users report working |
| 11 | configurations in little endian mode. |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 12 | |
| 13 | I named the board dbau1x00, to allow |
| 14 | support for all three development boards |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 15 | ( dbau1000, dbau1100 and dbau1500 ). |
| wdenk | 8b74bf3 | 2004-10-11 23:10:30 +0000 | [diff] [blame^] | 16 | Now there is a new board called dbau1550 also, which |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 17 | should be supported RSN. |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 18 | |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 19 | I only have a dbau1000, so my testing is limited |
| 20 | to this board. |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 21 | |
| 22 | The board has two different flash banks, that can |
| 23 | be selected via dip switch. This makes it possible |
| 24 | to test new bootloaders without thrashing the YAMON |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 25 | boot loader delivered with board. |
| 26 | |
| 27 | NOTE! When you switch between the two boot flashes, the |
| wdenk | 8b74bf3 | 2004-10-11 23:10:30 +0000 | [diff] [blame^] | 28 | base addresses will be swapped. |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 29 | Have this in mind when you compile u-boot. TEXT_BASE has |
| wdenk | 8b74bf3 | 2004-10-11 23:10:30 +0000 | [diff] [blame^] | 30 | to match the address where u-boot is located when you |
| 31 | actually launch. |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 32 | |
| 33 | Ethernet only supported for mac0. |
| 34 | |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 35 | PCMCIA only supported for slot 0, only 3.3V. |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 36 | |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 37 | PCMCIA IDE tested with Sandisk Compact Flash and |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 38 | IBM microdrive. |
| 39 | |
| 40 | ################################### |
| 41 | ######## NOTE!!!!!! ######### |
| 42 | ################################### |
| 43 | If you partition a disk on another system (e.g. laptop), |
| 44 | all bytes will be swapped on 16bit level when using |
| wdenk | 4cfaf55 | 2004-10-11 23:03:10 +0000 | [diff] [blame] | 45 | PCMCIA and running cpu in big endian mode!!!! |
| wdenk | 5da627a | 2003-10-09 20:09:04 +0000 | [diff] [blame] | 46 | |
| 47 | This is probably due to an error in Au1000 chip. |
| 48 | |
| 49 | Solution: |
| 50 | |
| 51 | a) Boot via network and partition disk directly from |
| 52 | dbau1x00. The endian will then be correct. |
| 53 | |
| 54 | b) Partition disk on "laptop" and fill it with all files |
| 55 | you need. Then write a simple program that endian swaps |
| 56 | whole disk, |
| 57 | |
| 58 | Example: |
| 59 | Original "laptop" byte order: |
| 60 | B0 B1 B2 B3 B4 B5 B6 B7 B8 B9... |
| 61 | |
| 62 | Dbau1000 byte order will then be: |
| 63 | B1 B0 B3 B2 B5 B4 B7 B6 B9 B8... |