Some good news: I got the FullDuplexSerial now working together with my CPU core and can still load the exerciser. Now I'll write some debug code that traces everything and I think the famous last bug will then have to show its ugly grimace.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Well, that's some output the head of io.spin writes to the terminal, not yet Z80.
No, it has to be page0.bin. This requires the bison and flex packages on both, Linux and CYGWIN. The file page0.z80 is actually a Z80 assembler source and is assembled by az80, which is generated using bison, flex and gcc.
HTH,
Juergen
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Toby Seckshund said...
Which would be the lesser of two evils, Windoze and Cygwin or dual boot with "XXXX" Linux?
If you want to try Linux, I would suggest Ubuntu 9.10, because it has a simple install that can re-partition your drive, and install the boot loader to choose which OS to boot. If you have a router for your internet connection that does DHCP, you won't even have to configure that. If you have a cable modem that does PPPoE (i.e. DSL modem), you need to run a setup tool after install (it was the case for me). You can even run e.g. Windows XP in a sandbox on Unbuntu. I do this with Sun's VirtualBox - a PC emulator which is really fine and for free.
CYGWIN has the advantage that you do not change your hard disk layout, and the disadvantage of being a little awkward to install, except when you decide to install everything, but that eats up a lot of hard disk space.
For the long term I would suggest giving Ubunto a try for everyone, because it's really as simple as one can imagine.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Toby Seckshund said...
I don't suppose that the +2 part is getting bits, bytes,words, ..... wrong again
Hmm.. no, I don't think so. It's one of the most complicated Z80 instructions, though. Doubly prefixed with FD and CB, followed by the signed 8bit displacement to add to IY, and finally the 4th byte containing the opcode. I looked over the code that constructs this PASM sequence over and over and don't see a reason.
I think it pokes in the wild because on the second run the first message reads "Thhs is pm80..." instead of "This is pm80...".
It may be that I missed a WZ or WC or both somewhere and are relying on undefined flags.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Yes I think I shall try another Dual boot, the only reason that there isnt one at the present is that I got into the habit of switching the PC on and going to make some coffee, or something. This always left it in the last installed, which was Ubuntu. Has the serial problems been sorted out yet ??? I might seach out the old 8.xx disk.
At work, one guy said that there is no need of any antivirus stuff, unless there are win PCs on the tail of it, is that true ???
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Toby Seckshund said...
Yes I think I shall try another Dual boot, the only reason that there isnt one at the present is that I got into the habit of switching the PC on and going to make some coffee, or something. This always left it in the last installed, which was Ubuntu. Has the serial problems been sorted out yet ??? I might seach out the old 8.xx disk.
At work, one guy said that there is no need of any antivirus stuff, unless there are win PCs on the tail of it, is that true ???
I have never seen a Linux virus. There are rumors that they shall exist, and also that root-kits should exist, while I haven't heard of anyone suffering from it. I have no virus scanner installed and no firewall either. The open ports on my box are a Webserver, CVS daemon and SSH daemon. There are even regular attempts to hack the sshd of my box, while people can try to hack the "root" account until they die, because there is none The webserver is also attacked every couple of hours, but the one I installed isn't vulnerable to these attacks.
If you let the Linux box act as a gateway for other PCs on a local network, you would better install a firewall, yes. Or trust the Windoze port blockers for that matter.
You can configure the boot menu to have a default of booting into Windows, and you can configure the time it waits for a user's choice.
I would not install an old version, because 9.10 seems very stable and recognized all my hardware without a problem.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Toby: If you have a Windows installed that you know and love. I might suggest an experiment. You see I got very tired with dual booting Windows and Linux back in '96. My solution was to dump Windows permanently but now there is a better way for you.
Install the virtualization software "VirtualBox" from Sun. It's free and excellent. www.virtualbox.org/
Now you can use that to make a virtual machine in which to install Linux, Ubuntu for example. After that you can run Linux in a Window in Windows.
This is a lot of fun by itself but to be useful you have to be able to get to the USB ports and such from the Linux machine. For example to be able to run the Linux version of BST to program Props or be able to use "dd" to write sectors to an SD card.
I have done this kind of "pass through" from Windows inside Linux, we have to confirm it is also possible/easy the other way around, should be OK.
The nice thing about running Windows in a virtual machine is that you can take a snap shot of the rare moments that it actually works and then when it screws itself just revert to the working image[noparse]:)[/noparse]
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.
I had a cd off of a mag about 2 years ago that was a SUN virtual thing but never got around to using it.
Crude but works is to switch the +5V and +12V to two HDs, both set as masters, I have to use one PC that should run WinME for legacy reasons but I Have a switched drive with XP on it as well (but not from Monday, when the brown stuff skids off the circular motion atmospheric movement actuator)
I will try another HD to start with, I have loads of 40GB ones here. That way if the video driver doesn't lke the twin WXGAs ....
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
I'm going to have to give up sleeping. Too much happens while I'm asleep!
Re sd cards, I've got 512mb standard size Legend ones that won't work with the propeller drivers. They work fine on my camera, so something is not quite right with the propeller code?
1gb san disk ones work perfectly, and the ones I'm getting are micro sd and they come with an adaptor to make them standard sized SD. Around $8 and I've been getting 8 at a time (friends and family keep 'borrowing' them. eg - "what are you doing on the computer, oh is that a pile of micro sd cards, I really need one of those for my new camera/phone etc")
I'm using the 8 8mb images, plus the boot.dsk image, and I also add a tiny standard text file MYNAME.TXT with the name of the board. There are advantages and disadvantages to using standard FAT files. I'm sure code can be optimised.
Right now I've got 456 longs free so that should be plenty for LMM for the extra opcodes.
@ Toby, Windows ME? The one operating system from Microsoft that can't do networking? I installed it once, then went 'up' to windows 95 to get the network talking. Toby, have you got boards working at the moment? Specifically re the dracblade, I've got some older designs that about one in 50 bootups won't bootup, and the sd card works fine in other boards. The more recent boards are rock solid (I've got three running the wireless network and they have not been turned off for 10 days now). The version 5 dracblade has shorter paths and the chips are closer and I am suspecting that the single NOP in the ram driver code is 'just' long enough. This version 5 also has much shorter paths for the VGA pins and the blue background is a lot more solid with less wavy lines. So layout is important, but it is hard to say what the 'right' layout is except without trial and error.
The great thing about CP/M is you can write code in small programs rather than having one huge program that takes ages to compile. So I've got one small program in sbasic that does the network login and peer to peer links. I'm now writing another one to handle packet routing and, just for fun, I'm having a go at writing that in C. I got sidetracked yesterday getting my vb.net IDE working for BDS C but it now can compile on the SIMH before downloading which is *much* faster and it also flags errors. See the attached screenshot - with the simh dos window overlaid on the IDE. There are tabs along the top for each programming language, and the one that is missing is BBC basic because for that, all the Z80 opcodes are needed. Hence I'm very interested in where the LMM idea is going.
Long, long ago, in a far off land a young student had a passion for motorcycles and stereo.
The motorcycle was from Honda but being poor the stereo amplifier was hand made from components from the junk box in a roughly hewn plywood box. Being poor the idea came "This amplifier has a hefty power supply that would also make a good battery charger for motorcycle batteries. All it needs is a big switch to put the rectified power out through the speaker terminals".
And so it was done and was a great success. Until one day at a drunken party someone threw the switch into "battery charger mode" whilst the amp was connected to some nice big expensive speakers......
In short, this sounds like a very dodgy practice.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.
pullmoll & Toby: I am not offended if Toby doesn't use my (Blade#2) board... Well actually it isn't my board anyway, it's just a clone. LOL
BTW: I published the circuit so it could be copied and I helped Drac do a latched version although I must confess, I didn't expect him to do 5 revs *grin*
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I found two more stupid bugs of mine.
1) The prefix DD/FD CB handler was patching code in the cog that was not yet there, because it ran before the code snippet was copied.
2) I had the dispatch table entries for RES and SET bit swapped in the table for DD/FD CB opcodes, thus "SET 0,(IY)" would actually do a "RES 0,(IY)".
Updated version 0.2.7 is at the front page.
To elaborate on what I did recently: the debug code now dumps the cog RAM instructions to be executed right before the jmp #ORIGIN, so I can peek at what the compiler translated for the section of Z80 opcodes listed before. Even with this dump I still didn't find the reason why the Z80 exerciser is terminating after the first test. I'll spend the rest of the day with trying to follow coredumps of the cog RAM.
The funny part is that the exerciser loop looks as if it really passed over all entries beginning at $01ae up to $024e. The value that comes off the stack after the first test is $0253, and there's a double 00 which is taken as the end mark of the table. The value should be $01b0, because $1ae was pushed to the stack earlier and the end of the test function just pops it and goes to the next table entry by incrementing twice. Somewhere in between the stack must have been corrupted.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
This via Ubuntu 9.10. It doesn't want to play nicely with the second monitor
Heater Back when I was a kid the mains electricity was 5A and 15A roundpins. There were some two pin plugs that pinched the cables under the tops of what was just a split pin. The cable for these 2 pin 5A sorts was just glorified bell wire with a single, clear insulation layer. One day the wire going to a lamp was confused with the cable going to a loudspeaker. As this speaker was about a foot or two away from the sockets, the application of 50Hz at +50dBm, the bad smell wasn't entirely due to burning voice coils.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Toby Seckshund said...
This via Ubuntu 9.10. It doesn't want to play nicely with the second monitor
You could try to get help in the forums, but searching I didn't find anyone asking for a dual WXGA setup. What cards are these? Did you try the hardware detection in the system menu?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
It's a NVidia fx5200 feeding two LG 1440x900 monitors, one on DVI (main) and 15pin Analog (second).
I did have the second monitor fed though the Extron VGA switcher, whilst the Ubuntu was going in. It should have still seen it as the switcher is supposed to pass the data lines.
I can always manage on the one monitor until I get braver about delving into the Linux.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Toby Seckshund said...
It's a NVidia fx5200 feeding two LG 1440x900 monitors, one on DVI (main) and 15pin Analog (second).
I did have the second monitor fed though the Extron VGA switcher, whilst the Ubuntu was going in. It should have still seen it as the switcher is supposed to pass the data lines.
I can always manage on the one monitor until I get braver about delving into the Linux.
Did you let the hardware checker look for the NVidia card? NVidia has a closed source video driver for Linux which is more capable than the one based upon reverse engineering. If you install that one, perhaps it recognizes the 2nd monitor.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Yes I saw some mention about that. I went to their site and selected it and a load of script came up. At this point SHE reminded me that I was supposed to be getting on with a bunch of letters which were on the WinXP HD. So its back onto Windoze for a while. This computer only has 768MB of ram (HaHaHa only 768 ......) so I do not know how the VirtualBox Heater suggested will work out.
Anyway, all this is off topic, Sorry ....
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Toby Seckshund said...
Yes I saw some mention about that. I went to their site and selected it and a load of script came up. At this point SHE reminded me that I was supposed to be getting on with a bunch of letters which were on the WinXP HD. So its back onto Windoze for a while. This computer only has 768MB of ram (HaHaHa only 768 ......) so I do not know how the VirtualBox Heater suggested will work out.
Anyway, all this is off topic, Sorry ....
I don't mind if we get off optic. There's nearly 900 replies to go to
No, actually I spent another day searching for the bug.
My signature more and more becomes a prophecy of my destiny.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
When compiling a CALL or RST instruction into memory, the return value that is pushed to the stack has to be the PC at compile time, while my code currently pushes the PC at runtime, which most of the times is several instructions behind the point where a CALL should return to.
Damn, that one was hard. Now I'm going to fix it.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
As I know 8080/8085/Z80 Stack return adres is always INSTRUCTION that flow CALL/RST address.
Program.
xx
xxxxxx
xxxx
XX(xxxx) > RST(CALL) As RST can be Software and Hardware CALL to Subroutine.
................. In Hardware type it Need always be clear with Runing instruction and Save PC next after it
XX > ADDRESS on Stack to return to
Else it will not work correctly
Regards
Regards
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Nothing is impossible, there are only different degrees of difficulty. For every stupid question there is at least one intelligent answer. Don't guess - ask instead. If you don't ask you won't know. If your gonna construct something, make it·as simple as·possible yet as versatile as posible.
As I know 8080/8085/Z80 Stack return adres is always INSTRUCTION that flow CALL/RST address.
Program.
xx
xxxxxx
xxxx
XX(xxxx) > RST(CALL) As RST can be Software and Hardware CALL to Subroutine.
................. In Hardware type it Need always be clear with Runing instruction and Save PC next after it
XX > ADDRESS on Stack to return to
Else it will not work correctly
Regards
Regards
Yes, of course it's the instruction that follows the CALL or RST, but you have to understand that in my implementation the PC variable is already at some point behind the CALL when the CALL is later executed in the cog. Thus I cannot push the current value of PC, but have to push the value of PC as it was when the CALL instruction was compiled. This then, automagically, is the address of the opcode following the CALL.
HTH,
Juergen
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Post Edited (pullmoll) : 3/14/2010 11:50:47 PM GMT
If posible can You give me any example of that. And why them are behind correct PC counter in Yours emulation.
Regards
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Nothing is impossible, there are only different degrees of difficulty. For every stupid question there is at least one intelligent answer. Don't guess - ask instead. If you don't ask you won't know. If your gonna construct something, make it·as simple as·possible yet as versatile as posible.
If posible can You give me any example of that. And why them are behind correct PC counter in Yours emulation.
Regards
Did you read the first page of this thread? Then you should understand why that was a bug.
Imagine this Z80 compiler as a level 1 cache. It looks ahead the next couple of instructions and translates them to Propeller's PASM opcodes, meanwhile incrementing PC over the code. Then it executes the compiled code and at the end compiles the next block of Z80 code. It's like a ping-pong between compiler stage and execution stage.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Now I start understand that it is not native emulator.
That exactly flow PC code.
And why You ned PC at compile time not at time rela PC counter point to.
As PC steps not fould execution time.
Thanks
Regards
Christoffer J
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Nothing is impossible, there are only different degrees of difficulty. For every stupid question there is at least one intelligent answer. Don't guess - ask instead. If you don't ask you won't know. If your gonna construct something, make it·as simple as·possible yet as versatile as posible.
Comments
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Hello world!
6c58
7044
from "a blade2". Is that good ????
PS I had to change page0.bin to page0.z80
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Well, that's some output the head of io.spin writes to the terminal, not yet Z80.
No, it has to be page0.bin. This requires the bison and flex packages on both, Linux and CYGWIN. The file page0.z80 is actually a Z80 assembler source and is assembled by az80, which is generated using bison, flex and gcc.
HTH,
Juergen
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
At least it made me wake up the "Blade2" board, DracBlade had made it get a bit neglected (Sorry Clusso, Sorry, .... )
Right who else to offend ?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
I assumed you have CYGWIN installed, no? Well, I can attach my current page0.bin and demo.bin files...
I don't believe anyone will be offended by you using or not using his board... I'd be interested to hear if it runs at all on another design.
Right now hunting for the bug in "SET 0,(IY+2)", which doesn't set bit 0 at IY + 2, but seems to poke in the wild.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
I'll look at CYGWIN.
I don't suppose that the +2 part is getting bits, bytes,words, ..... wrong again
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Post Edited (Toby Seckshund) : 3/13/2010 8:46:11 PM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
CYGWIN has the advantage that you do not change your hard disk layout, and the disadvantage of being a little awkward to install, except when you decide to install everything, but that eats up a lot of hard disk space.
For the long term I would suggest giving Ubunto a try for everyone, because it's really as simple as one can imagine.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Hmm.. no, I don't think so. It's one of the most complicated Z80 instructions, though. Doubly prefixed with FD and CB, followed by the signed 8bit displacement to add to IY, and finally the 4th byte containing the opcode. I looked over the code that constructs this PASM sequence over and over and don't see a reason.
I think it pokes in the wild because on the second run the first message reads "Thhs is pm80..." instead of "This is pm80...".
It may be that I missed a WZ or WC or both somewhere and are relying on undefined flags.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
At work, one guy said that there is no need of any antivirus stuff, unless there are win PCs on the tail of it, is that true ???
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
I have never seen a Linux virus. There are rumors that they shall exist, and also that root-kits should exist, while I haven't heard of anyone suffering from it. I have no virus scanner installed and no firewall either. The open ports on my box are a Webserver, CVS daemon and SSH daemon. There are even regular attempts to hack the sshd of my box, while people can try to hack the "root" account until they die, because there is none
If you let the Linux box act as a gateway for other PCs on a local network, you would better install a firewall, yes. Or trust the Windoze port blockers for that matter.
You can configure the boot menu to have a default of booting into Windows, and you can configure the time it waits for a user's choice.
I would not install an old version, because 9.10 seems very stable and recognized all my hardware without a problem.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Post Edited (pullmoll) : 3/13/2010 9:35:10 PM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Install the virtualization software "VirtualBox" from Sun. It's free and excellent. www.virtualbox.org/
Now you can use that to make a virtual machine in which to install Linux, Ubuntu for example. After that you can run Linux in a Window in Windows.
This is a lot of fun by itself but to be useful you have to be able to get to the USB ports and such from the Linux machine. For example to be able to run the Linux version of BST to program Props or be able to use "dd" to write sectors to an SD card.
I have done this kind of "pass through" from Windows inside Linux, we have to confirm it is also possible/easy the other way around, should be OK.
The nice thing about running Windows in a virtual machine is that you can take a snap shot of the rare moments that it actually works and then when it screws itself just revert to the working image[noparse]:)[/noparse]
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.
I had a cd off of a mag about 2 years ago that was a SUN virtual thing but never got around to using it.
Crude but works is to switch the +5V and +12V to two HDs, both set as masters, I have to use one PC that should run WinME
I will try another HD to start with, I have loads of 40GB ones here. That way if the video driver doesn't lke the twin WXGAs ....
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Re sd cards, I've got 512mb standard size Legend ones that won't work with the propeller drivers. They work fine on my camera, so something is not quite right with the propeller code?
1gb san disk ones work perfectly, and the ones I'm getting are micro sd and they come with an adaptor to make them standard sized SD. Around $8 and I've been getting 8 at a time (friends and family keep 'borrowing' them. eg - "what are you doing on the computer, oh is that a pile of micro sd cards, I really need one of those for my new camera/phone etc")
I'm using the 8 8mb images, plus the boot.dsk image, and I also add a tiny standard text file MYNAME.TXT with the name of the board. There are advantages and disadvantages to using standard FAT files. I'm sure code can be optimised.
Right now I've got 456 longs free so that should be plenty for LMM for the extra opcodes.
@ Toby, Windows ME? The one operating system from Microsoft that can't do networking? I installed it once, then went 'up' to windows 95 to get the network talking. Toby, have you got boards working at the moment? Specifically re the dracblade, I've got some older designs that about one in 50 bootups won't bootup, and the sd card works fine in other boards. The more recent boards are rock solid (I've got three running the wireless network and they have not been turned off for 10 days now). The version 5 dracblade has shorter paths and the chips are closer and I am suspecting that the single NOP in the ram driver code is 'just' long enough. This version 5 also has much shorter paths for the VGA pins and the blue background is a lot more solid with less wavy lines. So layout is important, but it is hard to say what the 'right' layout is except without trial and error.
The great thing about CP/M is you can write code in small programs rather than having one huge program that takes ages to compile. So I've got one small program in sbasic that does the network login and peer to peer links. I'm now writing another one to handle packet routing and, just for fun, I'm having a go at writing that in C. I got sidetracked yesterday getting my vb.net IDE working for BDS C but it now can compile on the SIMH before downloading which is *much* faster and it also flags errors. See the attached screenshot - with the simh dos window overlaid on the IDE. There are tabs along the top for each programming language, and the one that is missing is BBC basic because for that, all the Z80 opcodes are needed. Hence I'm very interested in where the LMM idea is going.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.smarthome.viviti.com/propeller
Let me tell you a story....
Long, long ago, in a far off land a young student had a passion for motorcycles and stereo.
The motorcycle was from Honda but being poor the stereo amplifier was hand made from components from the junk box in a roughly hewn plywood box. Being poor the idea came "This amplifier has a hefty power supply that would also make a good battery charger for motorcycle batteries. All it needs is a big switch to put the rectified power out through the speaker terminals".
And so it was done and was a great success. Until one day at a drunken party someone threw the switch into "battery charger mode" whilst the amp was connected to some nice big expensive speakers......
In short, this sounds like a very dodgy practice.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.
BTW: I published the circuit so it could be copied and I helped Drac do a latched version although I must confess, I didn't expect him to do 5 revs *grin*
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
1) The prefix DD/FD CB handler was patching code in the cog that was not yet there, because it ran before the code snippet was copied.
2) I had the dispatch table entries for RES and SET bit swapped in the table for DD/FD CB opcodes, thus "SET 0,(IY)" would actually do a "RES 0,(IY)".
Updated version 0.2.7 is at the front page.
To elaborate on what I did recently: the debug code now dumps the cog RAM instructions to be executed right before the jmp #ORIGIN, so I can peek at what the compiler translated for the section of Z80 opcodes listed before. Even with this dump I still didn't find the reason why the Z80 exerciser is terminating after the first test. I'll spend the rest of the day with trying to follow coredumps of the cog RAM.
The funny part is that the exerciser loop looks as if it really passed over all entries beginning at $01ae up to $024e. The value that comes off the stack after the first test is $0253, and there's a double 00 which is taken as the end mark of the table. The value should be $01b0, because $1ae was pushed to the stack earlier and the end of the test function just pops it and goes to the next table entry by incrementing twice. Somewhere in between the stack must have been corrupted.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Post Edited (pullmoll) : 3/14/2010 9:19:32 AM GMT
Heater Back when I was a kid the mains electricity was 5A and 15A roundpins. There were some two pin plugs that pinched the cables under the tops of what was just a split pin. The cable for these 2 pin 5A sorts was just glorified bell wire with a single, clear insulation layer. One day the wire going to a lamp was confused with the cable going to a loudspeaker. As this speaker was about a foot or two away from the sockets, the application of 50Hz at +50dBm, the bad smell wasn't entirely due to burning voice coils.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
You could try to get help in the forums, but searching I didn't find anyone asking for a dual WXGA setup. What cards are these? Did you try the hardware detection in the system menu?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
I did have the second monitor fed though the Extron VGA switcher, whilst the Ubuntu was going in. It should have still seen it as the switcher is supposed to pass the data lines.
I can always manage on the one monitor until I get braver about delving into the Linux.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
Did you let the hardware checker look for the NVidia card? NVidia has a closed source video driver for Linux which is more capable than the one based upon reverse engineering. If you install that one, perhaps it recognizes the 2nd monitor.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Anyway, all this is off topic, Sorry ....
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point
I don't mind if we get off optic. There's nearly 900 replies to go to
No, actually I spent another day searching for the bug.
My signature more and more becomes a prophecy of my destiny.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
When compiling a CALL or RST instruction into memory, the return value that is pushed to the stack has to be the PC at compile time, while my code currently pushes the PC at runtime, which most of the times is several instructions behind the point where a CALL should return to.
Damn, that one was hard. Now I'm going to fix it.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
I not understand You.
As I know 8080/8085/Z80 Stack return adres is always INSTRUCTION that flow CALL/RST address.
Program.
xx
xxxxxx
xxxx
XX(xxxx) > RST(CALL) As RST can be Software and Hardware CALL to Subroutine.
................. In Hardware type it Need always be clear with Runing instruction and Save PC next after it
XX > ADDRESS on Stack to return to
Else it will not work correctly
Regards
Regards
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nothing is impossible, there are only different degrees of difficulty.
For every stupid question there is at least one intelligent answer.
Don't guess - ask instead.
If you don't ask you won't know.
If your gonna construct something, make it·as simple as·possible yet as versatile as posible.
Sapieha
Yes, of course it's the instruction that follows the CALL or RST, but you have to understand that in my implementation the PC variable is already at some point behind the CALL when the CALL is later executed in the cog. Thus I cannot push the current value of PC, but have to push the value of PC as it was when the CALL instruction was compiled. This then, automagically, is the address of the opcode following the CALL.
HTH,
Juergen
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Post Edited (pullmoll) : 3/14/2010 11:50:47 PM GMT
Sorry but I still not understand it correctly.
If posible can You give me any example of that. And why them are behind correct PC counter in Yours emulation.
Regards
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nothing is impossible, there are only different degrees of difficulty.
For every stupid question there is at least one intelligent answer.
Don't guess - ask instead.
If you don't ask you won't know.
If your gonna construct something, make it·as simple as·possible yet as versatile as posible.
Sapieha
Did you read the first page of this thread? Then you should understand why that was a bug.
Imagine this Z80 compiler as a level 1 cache. It looks ahead the next couple of instructions and translates them to Propeller's PASM opcodes, meanwhile incrementing PC over the code. Then it executes the compiled code and at the end compiles the next block of Z80 code. It's like a ping-pong between compiler stage and execution stage.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
He died at the console of hunger and thirst.
Next day he was buried. Face down, nine edge first.
Now I start understand that it is not native emulator.
That exactly flow PC code.
And why You ned PC at compile time not at time rela PC counter point to.
As PC steps not fould execution time.
Thanks
Regards
Christoffer J
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nothing is impossible, there are only different degrees of difficulty.
For every stupid question there is at least one intelligent answer.
Don't guess - ask instead.
If you don't ask you won't know.
If your gonna construct something, make it·as simple as·possible yet as versatile as posible.
Sapieha