| yluo_ | hello all, i am new to openrisc. After days of trying, I am able to boot Linux on a DE2 (not DE2-115), but the boot procedure stopped at "Unpacking initramfs". Any hints will be greatly appreciated. | 18:37 |
|---|---|---|
| yluo_ | btw, I followed the nice tutorial from http://kevinmehall.net/openrisc/guide/ | 18:38 |
| Erant | yluo_: Are you sure there's ramdisk in your kernel image? | 18:46 |
| Erant | Now mind you, you're further along than I am, there's something off about my attempts. | 18:47 |
| yluo_ | in fact I am not sure about whether ramdisk is in my kernel image. how can I check it? | 18:55 |
| Erant | Well, for one, how large is your image? | 18:57 |
| yluo_ | 4759072 bytes | 19:00 |
| Erant | Sounds about right. What are the bootargs? (I'm trying to debug here from a general knowledge of how this is supposed to work, not specific OpenRISC knowledge) | 19:03 |
| yluo_ | thanks. this is what I have for bootarg --- console=uart,mmio,0x90000000,115200 | 19:14 |
| yluo_ | i agree that this is not really an OpenRISC issue, more on the linux side. | 19:14 |
| yluo_ | I only have 8MB sdram on the DE2 board. could it be the issue? | 19:15 |
| Erant | Euh, yes. What target did you compile for? | 19:16 |
| yluo_ | here is something from the boot messages: Memory: 4208k/8192k available (2053k kernel code, 3984k reserved, 300k data, 1416k init, 0k highmem) | 19:16 |
| yluo_ | target is or32-linux | 19:17 |
| Erant | Though, really, it'd panic if it ran out of memory. | 19:17 |
| Erant | I meant more, what platform did you synthesize for? | 19:17 |
| stekern | naw, I think it just hangs there | 19:18 |
| stekern | at least it does that in or1ksim the times I've had a too large image compared to the memory size defined there | 19:18 |
| yluo_ | Erant, were you asking the FPGA part number? it is Cyclone II EP2C35F672C6 on DE2 | 19:20 |
| yluo_ | does it mean that with only 8MB SDRAM, DE2 is not able to boot a linux? | 19:21 |
| Erant | stekern: Would it block trying to wire down the memory? | 19:22 |
| Erant | yluo_: No, you just don't want to go the initramfs route. Do you have storage on the board? Like a SPI flash? | 19:22 |
| yluo_ | I have 4Mbyte Flash Memory and SD card socket | 19:23 |
| Erant | You can try booting off an SD card instead of using a ramdisk. | 19:24 |
| yluo_ | that is a good idea. do you have any pointers that can help me start? | 19:25 |
| Erant | Google would be a good a place as any, really. | 19:26 |
| yluo_ | haha, yes google. Is u-boot helpful? | 19:26 |
| Erant | You'll be wanting a bootloader, sure, you'll be modifying the boot-args a lot. | 19:27 |
| Erant | But u-boot itself can't do anything about where Linux gets its root filesystem. | 19:28 |
| Erant | You'll have to figure out where the SD blockdevice is (assuming there's a driver for it in your kernel), supply the 'root=' bootarg, tell Linux that there's no initial ramdisk through noinitrd, maybe hint at the root filesystem by telling it through rootfs=, etc. etc. | 19:30 |
| yluo_ | so I can have the root fs in either on-board flash memory or SD card, right? | 19:31 |
| yluo_ | I saw orpsoc defines a SD controller, is that what I need to look into first? | 19:31 |
| Erant | Sure. as long as there's a blockdevice that represents it. | 19:34 |
| yluo_ | something like /dev/sda , right? I prefer SD card over the on-board flash memory, as it is easier to update the root file system. Am I understanding this correctly? | 19:36 |
| Erant | Yeah. And it can be many different things. /dev/mmc<blah> would probably be more common. | 19:37 |
| Erant | /dev/mmcblk0pX | 19:38 |
| Erant | Again, google is your friend. | 19:38 |
| yluo_ | thanks for the direction. Since it becomes more of a linux problem than OpenRISC, I will not flood this channel any more. I just want to say thank you for your help, and thanks the OpenRISC community for such wonderful platform/tools. | 19:41 |
| Erant | I actually had an orpsoc related question. bootrom.S doesn't seem to do any init of the DRAM block, but then loads the flash image into RAM. | 20:07 |
| Erant | Are the DRAM blocks self-configuring? | 20:07 |
| stekern | Erant: yes, all memory controllers currently present are self-configuring | 21:02 |
| Erant | Fair enough. That explains that at least. Was the openRISC modeled after any existing architecture? | 21:02 |
| stekern | not really, but it keeps much reassemblence with DLX and MIPS | 21:03 |
| Erant | Fair enough, I can write MIPS. I'm having weird issues where I can't seem to get any output from the core. Going to have to go write some assembly to figure it out (or, ya know, wire up a JTAG dongle) | 21:04 |
| Erant | Or ChipScope. | 21:05 |
| stekern | openrisc assembly is pretty straight forward, if you know mips you'll pretty much be all set already | 21:07 |
| Erant | What about delay slots? | 21:08 |
| yluo_ | i have question related to SD controller. I got this error when running quartus II analysis & synthesis: Can't infer register for "new_bw" at sd_bd.v(132). Any hints? | 23:08 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!