Hello, I've uploaded the state-so-far of my monitor to: http://www.axio.ms/projects/e2/e2-monitor-0.22.tgz (the README from the tar is also at http://www.axio.ms/projects/e2/e2-monitor-0.22-README ) What it can do: o Provides basic monitor with commands such as peek/poke/go o XMODEM upload/download RAM over serial (though this is currently v.picky about terminal s/w used) o Read NAND flash to RAM o Write/erase NAND flash from RAM data o CRC32 chunks of memory o Initialise LCD a bit/provide simple string/hex printing to it o Boot from PBL What it can't yet do: o Search the rest of NAND for executable client programs, and boot them o Access SMC cards (I have been unable to coax PBL into accessing them, but it looks like much of the support is there in PBL. (Not necessarily complete/functional of course)) It's pretty fast at downloading data and writing it to flash, or just executing code, so I find it's a better development environment than EXP'ing stuff down with pblq-alikes. I've usually started up the E2, interrupted the splash screen over JTAG and downloaded the ELF (over jtag); now that it boots from flash, it'll accept a new version of itself via XMODEM and run that happily. (Whew, my Windows box is now unnecessary!) I don't know if anyone's interested in trying it out or using it; I'd be grateful for any comments, suggestions or code contributions! It's not really there yet - well, XMODEM works for me but probably won't for anyone else =] That's next on my list. -Matt ps: Here are the supported commands, from its help text: Help (commands are case-sensitive): h/? - This help dump <m|fd> <addr> <len> - Hexdump RAM or flash crc <addr> <len> - CRC32 memory get <addr> <len> - XMODEM download from RAM put <addr> - XMODEM upload to RAM flashwr <flash offset> <len> <ram address> - Write RAM data into flash device flashrd <flash offset> <len> <ram address> - Copy flash data into RAM listdevs - List flash devices go <addr> - Jump to address r{b|w} <addr> - Read byte/word w{b|w} <addr> <data> - Write byte/word fill <addr> <len> <byte> - Fill memory lcd <string> - Write string to LCD console [ fd above is a flash device ID, 0 = internal, 1 = external etc., as listed in listdevs ] [ All numbers/addresses/offsets are hexadecimal - zero padding is unnecessary ] pps: Any better names..?
Hi I have been lurking here for a little while and finally bought an E2 from ebay about a month ago (followed by a much cheaper one from a car boot sale!). I was very happy when I got pblq to work with my level shifter but unfortunately lost my splash screen (no big deal) when I tried out the Hello World example. Tonight I thought I'd try the monitor program because I liked the idea of browsing the NAND 'live' rather than downloading. Unfortunately, on downloading my monitor.pbl using pblq it looks as if my pbl has been trashed i.e. the E2 boots straight into the monitor's LCD debug console without the flashing lights of the pbl. Incidentally, I used "pblq writeflash monitor.pbl 0". So: is this a one-way trip for this E2 i.e. Monitor-only - no pbl possible, or is there some other path? (I mean via the ext port rather than JTAG) Another mistake was that my 115200 is a little flaky and I meant to see if I could drop the monitor terminal speed a little before loading it. I guess its all good experience! NB: still one untouched E2 to play with! Adam. --- Matt Evans <matt@axio.ms> wrote:
Hello,
I've uploaded the state-so-far of my monitor to:
http://www.axio.ms/projects/e2/e2-monitor-0.22.tgz
(the README from the tar is also at
http://www.axio.ms/projects/e2/e2-monitor-0.22-README
)
_______________________________________________ e3-hacking mailing list e3-hacking@earth.li
http://www.earth.li/cgi-bin/mailman/listinfo/e3-hacking
___________________________________________________________ How much free photo storage do you get? Store your holiday snaps for FREE with Yahoo! Photos http://uk.photos.yahoo.com
Evening Adam, On 25 May 2005, at 19:00, Adam Coventry wrote:
I have been lurking here for a little while and finally bought an E2 from ebay about a month ago (followed by a much cheaper one from a car boot sale!).
I was very happy when I got pblq to work with my level shifter but unfortunately lost my splash screen (no big deal) when I tried out the Hello World example.
Hello welcome!
Unfortunately, on downloading my monitor.pbl using pblq it looks as if my pbl has been trashed i.e. the E2 boots straight into the monitor's LCD debug console without the flashing lights of the pbl. Incidentally, I used "pblq writeflash monitor.pbl 0". So: is this a one-way trip for this E2 i.e. Monitor-only - no pbl possible, or is there some other path?
:-) That's what's supposed to happen. (It sounds like you see the monitor message on the LCD, right? And you get serial port output?) I think there's some confusion about what PBL is. There are two flash chips; one is a NOR flash chip of 64K in which PBL lives. The other is an 8MB NAND flash chip that PBL then accesses and loads the main OS from (including the splash screen). The splash screen is the first chunk in the NAND flash, offset zero. The rest of the OS appears to live +1MB into the NAND, and the splash screen 'chains' onto the rest. i.e. the boot sequence is: 1. CPU starts at phys addr. 0 where NOR flash is mapped, & executes PBL. 2. PBL trawls the NAND flash and finds a valid bootable block at 0. (This is the splashscreen (or helloWorld, or monitor)) 3. The block, the splashscreen, is loaded to RAM and executed and then either it or PBL loads the next block in flash (the machine OS). So what you have done is replaced the splash screen with the monitor code; the machine boots and runs the monitor straight off instead of the splash screen and the rest of the OS. I'm not sure if I understand your question correctly - are you wondering whether it will 'dual-boot', i.e. if the monitor can then boot the main Amstrad OS? No is the short answer - though it ought to be possible. When you say "no pbl possible" what do you mean? PBL is still there - it's what boots the monitor from NAND. If you want the splash screen back you could reflash the first 2 blocks (16K) of NAND with the blocks you backed up with pblq before reflashing ;-)
Another mistake was that my 115200 is a little flaky and I meant to see if I could drop the monitor terminal speed a little before loading it.
Ahhh yep; well, you can download a replacement monitor into the monitor you've already got running (if your serial is reliable enough to do so). Edit uart.S:uart_init, there's a comment for some other baud rate divisors. You can "put 100000" to XMODEM the resulting (blessed) 'monitor.pbl' into RAM, and then "flashwr 0 4000 100000" to reflash over itself. Glad to hear you got it working thus far, though! I've a newer version of the monitor I'll upload tonight. -Matt
OK - panic over. I had assumed the rapid light flashing was driven by an active pbl and that the absence of those lights meant that I had somehow trashed pbl. I assume what actually happens is that the lights flash as pbl scans the NAND for headers. Since the monitor was loaded into address 0 and did not chain other blocks there was no searching to do and therefore no reassuring light show. I _did_ manage to interrupt my (failed?) monitor program with pblq and reloaded a 57600 version (to overcome my flaky serial). Now I am presented with a proper monitor prompt, and all is well! Thanks for the great work you guys are doing - I don't mind hanging on to your coat tails! Adam. --- Matt Evans <matt@axio.ms> wrote: [deleted - I don't know if etiquette demands I leave the original text, but I worry about bandwidth ;-)]
e3-hacking mailing list e3-hacking@earth.li
http://www.earth.li/cgi-bin/mailman/listinfo/e3-hacking
Adam Coventry BSc. MCP ascoventry@yahoo.co.uk ___________________________________________________________ Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail http://uk.messenger.yahoo.com
Hi Adam, On 27 May 2005, at 08:02, Adam Coventry wrote:
I had assumed the rapid light flashing was driven by an active pbl and that the absence of those lights meant that I had somehow trashed pbl. I assume what actually happens is that the lights flash as pbl scans the NAND for headers. Since the monitor was loaded into address 0 and did not chain other blocks there was no searching to do and therefore no reassuring light show.
Ah yes, I think that's what happens. (Or, it's flashing when it's loading the first splashscreen thing - which is much larger. Maybe it currently does flash but only once, really quickly, so you don't see it or something... my E2 is out of the case so I don't have LEDs wired up.)
I _did_ manage to interrupt my (failed?) monitor program with pblq and reloaded a 57600 version (to overcome my flaky serial). Now I am presented with a proper monitor prompt, and all is well!
Good stuff! I forgot to upload the other night, but there's a new version of the monitor that you may find more useful: http://www.axio.ms/projects/e2/e2-monitor-0.24-README ..contained in.. http://www.axio.ms/projects/e2/e2-monitor-0.24.tgz This version boots other stuff from flash (so is actually useful (he says - my biggest test program so far splats a big image on the screen...)). The XMODEM stuff should be slightly more resilient/well-behaved but still isn't perfect. Cheers, Matt
participants (2)
-
Adam Coventry -
Matt Evans