with little shame update smartest1613

discussion of beeb/electron applications, languages, utils and educational s/w
duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Fri Mar 27, 2015 4:10 pm

it is a very slow thing to copy :)

but in the latest n.smart 27-03 it looks like it's working ??


DutchAcorn wrote:
duikkie wrote::the error with backup i don't understand.

backup mmb file ?? from 0 to 1 ??
or is it the discdrive unit ?


What I mean is:

Code: Select all

*card
*dform 450
*drive 0
*din 411
*drive 1
*din 450
*enable
*backup 0 1


I'll have another go with the latest build.

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Fri Mar 27, 2015 7:43 pm

duikkie wrote:it is a very slow thing to copy :)

but in the latest n.smart 27-03 it looks like it's working ??



I am trying to do some tests with lots of data movement.

Backup now seems to work but the copied image is still corrupted. Example attached. Have a look at the actual pictures in the copied disk image.
Attachments
AMX Art corrupt.zip
(79.36 KiB) Downloaded 43 times
Paul

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Sun Mar 29, 2015 8:30 pm

Just tried this last version of the ROM in a BBC B fitted with Econet.

Combined with ANFS 4.18
    There are known issues with Econet and the MMC, but with the SMART ROM combined with ANFS 4.18 the BBC won't even start-up. It just displays "BBC Computer 32K" and hangs. (This is the same behaviour as with the Super MMC without ANFS or DNFS fitted.)
    If ANFS is put in a higher ROM number than the SMART ROM it does not even display "BBC Computer 32K", just a flashing mode 7 cursor.

Combined with DNFS
    With a DNFS ROM in a higer priority slot it does start-up with the SMART 27-3 version.
    If the DNFS ROM is in a lower priority slot than SMART 2703 the BBC hangs after "BBC Computer 32K".
    If the SMART2703 is then replaced by the SuperMMC ROM the BBC comes to life again. (the *CARD command
does not work, as expected).
Paul

User avatar
tricky
Posts: 1916
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: working on new MMC rom YOUR missing it ?

Postby tricky » Sun Mar 29, 2015 9:13 pm

Sorry I'm missing all the fun, I have a couple of new micro SD cards on order and should have my sideways RAM beeb "retrieved" soon.

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Mar 30, 2015 7:36 am

can you make a dump of the &d00 to &e00 if anfs is fitted
allso the DNFS rom at &d00

i think there is the trouble, because i clean the hole &d00 to &d70 area with zero

is econet working with a machine with only tape version ??
&40 at &d00 the rest blank ??

still working on the *backup and *copy 0 1 thing
it is not using the right fvars% when using *backup and *copy ??
because fvars% is at &bf to &c6 is is difficult to debug
so i wrote a version to store &b6 to &d80 and can see that
the file param's are not what they are on *info
the error is to big ......
why the file param's with backup and copy are wrong is a ?
normaly something like this 00 19 23 80 AB 02 CC 02 : >> $.file 1900 8023 2AB sector 002. when load is used this is good
if *copy is used i get 00 19 AB 80 23 03 < 25 80> strange ??
only begin location is right 23 and AB switched ? length +1 ? sector strange ??

(was ik er maar nooit aan begonnen :) , all ways troubles in paradise


DutchAcorn wrote:Just tried this last version of the ROM in a BBC B fitted with Econet.

Combined with ANFS 4.18
    There are known issues with Econet and the MMC, but with the SMART ROM combined with ANFS 4.18 the BBC won't even start-up. It just displays "BBC Computer 32K" and hangs. (This is the same behaviour as with the Super MMC without ANFS or DNFS fitted.)
    If ANFS is put in a higher ROM number than the SMART ROM it does not even display "BBC Computer 32K", just a flashing mode 7 cursor.

Combined with DNFS
    With a DNFS ROM in a higer priority slot it does start-up with the SMART 27-3 version.
    If the DNFS ROM is in a lower priority slot than SMART 2703 the BBC hangs after "BBC Computer 32K".
    If the SMART2703 is then replaced by the SuperMMC ROM the BBC comes to life again. (the *CARD command
does not work, as expected).

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Mon Mar 30, 2015 9:01 am

duikkie wrote:can you make a dump of the &d00 to &e00 if anfs is fitted
allso the DNFS rom at &d00

i think there is the trouble, because i clean the hole &d00 to &d70 area with zero

is econet working with a machine with only tape version ??
&40 at &d00 the rest blank ??


The BBC is a tape-only machine without a disc interface fitted. Probably from a former school network. There is no way for me to check if the econet actually works; I do not have a network set up. It boots up with Econet station 1 (no clock), so that seems fine.

Getting stuff to and from the BBC is a bit of a challenge without a disc interface or MMC, perhaps a good time to build that UPURS cable... :wink:

I'll have a look at dumping the memory addresses later this week...
Paul

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Mar 30, 2015 1:12 pm

i read a lot about memory use for beeb
the problems are big
example
&a0..&a7= nmi workspace but allso used extensively by the disc and network filing system. so when allso used by SD card , there are troubles
i think i will reclocate the hole &a0..&a7 to &70..&77 space for user
some pointers to the new &d21 (was swap aera )
&a8.. &af are for utilities area. i did store there the skipsector and clustex% bytes. so thats no good
&b0-- &bf is filing system scratch space
&c0-- &cf is filing system dedicated workspace

&c0 -- &cf is used for active filing system , so if you switch the valeus are gone ?
all by all it is not easy

in some books &d00.. &0d5f is nmi code area
in an other part of the book
it is &d00--&d9e it is important for my beeb....mmb at &d70
at &d9f--&def are expanded vector set


so the next smarttest rom , is with new where to put the value bytes :)

i can use &70-&8f but i think some one will use that ( exmon??)
&90-&9f is econet , not a good option
&f8, &f9 are free by os1.2

zero page is to small :)

illigal playing with end stack page one ( &100--&1ff ) ???
killing page 10 (&A00--&AFF) TAPE AND RS423 ( PROBLEMS WITH your up rom ??)
killing page 11 ( softkey area ) ???

every thing you do has effect :(



DutchAcorn wrote:
duikkie wrote:can you make a dump of the &d00 to &e00 if anfs is fitted
allso the DNFS rom at &d00

i think there is the trouble, because i clean the hole &d00 to &d70 area with zero

is econet working with a machine with only tape version ??
&40 at &d00 the rest blank ??


The BBC is a tape-only machine without a disc interface fitted. Probably from a former school network. There is no way for me to check if the econet actually works; I do not have a network set up. It boots up with Econet station 1 (no clock), so that seems fine.

Getting stuff to and from the BBC is a bit of a challenge without a disc interface or MMC, perhaps a good time to build that UPURS cable... :wink:

I'll have a look at dumping the memory addresses later this week...

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Mar 30, 2015 3:26 pm

now there is smartest3003 :)

have changed the a0..&af block
not totaly empty a0..a2 is datptr% pionter
and ptr% that was a8..a9 pointer to a3 .. a4

the bytes that was at a2=sec%,seccount%,skipsec%,bytelastsec% are moved to &d21 (was *dswap area)

i don't know about econet yet
(a0..a4 still used)

but i hope the N.smart rom from 3003 will have no errors [-o<

i did backup and copy a hole din , backup takes a lot of time and you see nothing till >

try it with your backup sd card !!!
Last edited by duikkie on Fri Apr 03, 2015 7:38 am, edited 1 time in total.

User avatar
daveejhitchins
Posts: 3690
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham

Re: working on new MMC rom YOUR missing it ?

Postby daveejhitchins » Mon Mar 30, 2015 4:14 pm

OK . . . Laugh out Loud . . . But, isn't this a filing system? And, as such it has it's own workspace? If it isn't a FS couldn't you make it one?? I'm probably missing something, aren't I :oops:

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Mar 30, 2015 5:02 pm

no the mmc/sd card is not a filing system. it is a userport system.
the most other filing systems have there own chip to communigate with
what directly is linked to adress lines and datalines.
the 6522 do not have this page like cass=&a00, disc=&d00. econet &90..9f ??

i am not totaly sure but i thing the NMI claim for MMC/SD card is not the effect , the maker had in mind. claim it ???
the mmc/sd claims NMI workspace , but has no NMI ??
it stores a lot of bytes and pionters but thats all.

but you have to store somethings somewhere, else you don't know where you are in the card memory banks :)


daveejhitchins wrote:OK . . . Laugh out Loud . . . But, isn't this a filing system? And, as such it has it's own workspace? If it isn't a FS couldn't you make it one?? I'm probably missing something, aren't I :oops:

Dave H :D

User avatar
sweh
Posts: 1847
Joined: Sat Mar 10, 2012 12:05 pm
Location: New York, New York
Contact:

Re: working on new MMC rom YOUR missing it ?

Postby sweh » Mon Mar 30, 2015 6:49 pm

duikkie wrote:no the mmc/sd card is not a filing system. it is a userport system.

In Beeb terminology it _is_ a filesystem. It just happens to use the userport for communicating to the target device, rather than a dedicated chip.
Rgds
Stephen

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Mar 30, 2015 6:58 pm

you can call everything a filesystem then . a userport device where data is comming from is not a filesystem . else ram with battery is allso a filesystem.
sideway ram no filesystem. eeprom not a filesystem.

the main thing is it has no interrupt so why claiming the nmi workspace ?


sweh wrote:
duikkie wrote:no the mmc/sd card is not a filing system. it is a userport system.

In Beeb terminology it _is_ a filesystem. It just happens to use the userport for communicating to the target device, rather than a dedicated chip.

User avatar
sweh
Posts: 1847
Joined: Sat Mar 10, 2012 12:05 pm
Location: New York, New York
Contact:

Re: working on new MMC rom YOUR missing it ?

Postby sweh » Mon Mar 30, 2015 7:18 pm

duikkie wrote:you can call everything a filesystem then . a userport device where data is comming from is not a filesystem . else ram with battery is allso a filesystem.
sideway ram no filesystem. eeprom not a filesystem.

A filesystem, in Beeb terminology, has a set of requirements (eg intercepting various vectors, providing various services). The DFS ROM that you're hacking on is providing most of those services, so you don't have to do it again.

So, for example, let's take UPURS. The original routines that Martin wrote are "block transfer" type routines (eg download ROM image and load into SWR; download SSD image and write to disk). That's not a filesystem. But then he wrote the UPCFS routines; that implemented a subset of a filesystem. Then I wrote UPURSFS based on his (and JGH) work. That _is_ a filesystem. Same physical cable, but the software determines if a filesystem is implemented, or not.

Martin Mather's MMC code, SuperMMC, TurboMMC... they all implement filesystems.

As it happens there is also a ROM filesystem built into the Beeb; you can burn an EPROM in a specific format, then type *ROM and then do things like "*.", LOAD, etc. Again, a subset of the filesystem (readonly) but a filesystem.

To conclude; it's not the hardware that makes a filesystem, but the software.
the main thing is it has no interrupt so why claiming the nmi workspace ?

Any code could claim the NMI space, as long as it does it properly (correct calls). If you're not actually implementing an NMI routine then ensure &D00 holds the RTI instruction.

Note that other code could steal the NMI workspace from underneath you, though! So don't use it for persistent storage.
Rgds
Stephen

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Thu Apr 02, 2015 9:08 am

USE AT YOUR OWN RISK !!!
BACKUP FIRST YOUR BEEB.MMB FILE

USE an other SD or SDHC CARD !!!!



update and pre annoucement :)

the n.smart is your rom for spi sd card use

it can now
the command beeb for more then only beeb.mmb

hide commands like *dget and *dput

on your sd card you put the file named INOUTMEM
(take a 16k rom , rename it in INOUTMEM on your sdcard PC !!!)

do

*dmmc
*beeb ( or *din0)
*dget
(it will load inoutmem at &2000 ...&6000)

to use the sd/mmc card again for bbc
*dmmc
*beeb or *beebxxxx
*save <filename> 2000 6000 8000 8000 (if it is a rom)

to transfer memory &2000 .. &6000 to pc
YOU MUST HAVE file on SD card named INOUTMEM

*dmmc
*beeb
*load <filename> 2000
*dput

take your sd card out , place it in pc sdholder
look at INOUTMEM

you can transfer any thing length=&4000

because *dget and *dput use INOUTMEM you must put BEEB.MMB back with the command *beeb

so after transfer
*dmmc
*beeb
Last edited by duikkie on Fri Apr 03, 2015 7:38 am, edited 1 time in total.

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new SMART rom : if you DARE

Postby duikkie » Thu Apr 02, 2015 3:20 pm

for tomorrow

bug fixed if you use *dget/*dput and your sd card stays in the bbc computer
i restore values

so

*dmmc
*beeb
*dget
*din xxx
*save <filename> 2000 6000 8000 8000

get on with your life with your selected beebxxxx.mmb file
Attachments
SMARTEST0304.zip
(66.59 KiB) Downloaded 42 times

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new SMART rom : if you DARE

Postby duikkie » Mon Apr 06, 2015 9:49 am

BUG FIXING

i noticed that the SMART SPI tekst was gone
something to do with startopt%

it is back :)

also use S with BREAK to init SMART SPI
not M with break :)

the Z BREAK did never work on my machine

i changed the SMART SPI in begin PAGE=&1900
and not &1B00 ( why was that good for ??)

i hope little by little i fixed al the bugs ??
Attachments
SMARTEST0604.zip
(63.16 KiB) Downloaded 34 times

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Mon Apr 06, 2015 3:32 pm

duikkie wrote:can you make a dump of the &d00 to &e00 if anfs is fitted
allso the DNFS rom at &d00

i think there is the trouble, because i clean the hole &d00 to &d70 area with zero

is econet working with a machine with only tape version ??
&40 at &d00 the rest blank ??


Ok, not sure if this is what you mean but I installed Beebmon on the Econet machine and this is what I get at the locations &d00-&e00. DNFS is in the higher priority ROM.

I used the 0604 version. As expected the it hangs after the *card command.

dump1.jpg

dump2.jpg

dump3.jpg
Paul

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Apr 06, 2015 5:23 pm

is this if you install only the DNFS rom in bank 13 ? (19 page at & dfd.) else i do not see the DNFS rom from df0.. dff ?? little strange ??

i will convert the code , NMI from &d00 :)

does the 0604 version work on other machines ?

DutchAcorn wrote:
duikkie wrote:can you make a dump of the &d00 to &e00 if anfs is fitted
allso the DNFS rom at &d00

i think there is the trouble, because i clean the hole &d00 to &d70 area with zero

is econet working with a machine with only tape version ??
&40 at &d00 the rest blank ??


Ok, not sure if this is what you mean but I installed Beebmon on the Econet machine and this is what I get at the locations &d00-&e00. DNFS is in the higher priority ROM.

I used the 0604 version. As expected the it hangs after the *card command.

dump1.jpg

dump2.jpg

dump3.jpg

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new SMART rom : if you DARE

Postby duikkie » Mon Apr 06, 2015 5:41 pm

code at &d00
BIT &fe18
PHA
TYA
PHA
LDA #&0F : rem bank 15
STA &fE30 : select bank 15 :: dnfs rom in bank 15 ??
JMP &96BF : jmp adress in rom
STY &0D0D
STA &0D0C
LDA &F4
STA &FE30 : give back to previous rom
PLA
TAY : give back Y
PLA
BIT &FE20 ???
RTI : end interupt

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Mon Apr 06, 2015 6:32 pm

duikkie wrote:is this if you install only the DNFS rom in bank 13 ? (19 page at & dfd.) else i do not see the DNFS rom from df0.. dff ?? little strange ??


DNFS in 15
BASIC in 14
SPI in 13
Beebmon in 12

duikkie wrote:does the 0604 version work on other machines ?


It works, but the retest on the *backup failed on my Master 128. I used a FAT32 formatted SDHC card with two beeb.MMC files. I tried to copy the AMX Art images several times but the result is unpredictable. From just corrupted pictures to a corrupted catalogue. An example is attached.
Attachments
0604 AMX Art Copied.zip
(72.81 KiB) Downloaded 35 times
Paul

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Apr 06, 2015 6:48 pm

does *copy *.* works on master ?
always the *backup , het zou verboden moeten worden ;)

DutchAcorn wrote:
duikkie wrote:is this if you install only the DNFS rom in bank 13 ? (19 page at & dfd.) else i do not see the DNFS rom from df0.. dff ?? little strange ??


DNFS in 15
BASIC in 14
SPI in 13
Beebmon in 12

duikkie wrote:does the 0604 version work on other machines ?


It works, but the retest on the *backup failed on my Master 128. I used a FAT32 formatted SDHC card with two beeb.MMC files. I tried to copy the AMX Art images several times but the result is unpredictable. From just corrupted pictures to a corrupted catalogue. An example is attached.

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Apr 06, 2015 7:04 pm

something wrong with file ? can't load it in mmb image ??

DutchAcorn wrote:
duikkie wrote:is this if you install only the DNFS rom in bank 13 ? (19 page at & dfd.) else i do not see the DNFS rom from df0.. dff ?? little strange ??


DNFS in 15
BASIC in 14
SPI in 13
Beebmon in 12

duikkie wrote:does the 0604 version work on other machines ?


It works, but the retest on the *backup failed on my Master 128. I used a FAT32 formatted SDHC card with two beeb.MMC files. I tried to copy the AMX Art images several times but the result is unpredictable. From just corrupted pictures to a corrupted catalogue. An example is attached.

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Apr 06, 2015 7:09 pm

nothing to see from &df0 from dnfs rom , i think the rom do not use the same area as dfs 0.9 , which version dnfs or put the rom in ssd format
i will discramble the start up from dnfs , which area's it use and what it is with &fe18 and &fe20.


DutchAcorn wrote:
duikkie wrote:is this if you install only the DNFS rom in bank 13 ? (19 page at & dfd.) else i do not see the DNFS rom from df0.. dff ?? little strange ??


DNFS in 15
BASIC in 14
SPI in 13
Beebmon in 12

duikkie wrote:does the 0604 version work on other machines ?


It works, but the retest on the *backup failed on my Master 128. I used a FAT32 formatted SDHC card with two beeb.MMC files. I tried to copy the AMX Art images several times but the result is unpredictable. From just corrupted pictures to a corrupted catalogue. An example is attached.

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Mon Apr 06, 2015 8:05 pm

duikkie wrote:something wrong with file ? can't load it in mmb image ??


I guess it's too corrupted to load it... :(
Paul

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Mon Apr 06, 2015 8:50 pm

oke i will look at the *backup routine , i don't know what is wrong yet . but maybe i can find it ;)

is *copy *.* allso in error ? , or is it only *backup ?

DutchAcorn wrote:
duikkie wrote:something wrong with file ? can't load it in mmb image ??


I guess it's too corrupted to load it... :(

User avatar
DutchAcorn
Posts: 1631
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: working on new MMC rom YOUR missing it ?

Postby DutchAcorn » Mon Apr 06, 2015 9:00 pm

duikkie wrote:oke i will look at the *backup routine , i don't know what is wrong yet . but maybe i can find it ;)

is *copy *.* allso in error ? , or is it only *backup ?

DutchAcorn wrote:
duikkie wrote:something wrong with file ? can't load it in mmb image ??


I guess it's too corrupted to load it... :(


Just tried *copy. That gives corrupted files. Tried it on both SDHC FAT32 and SD FAT: same result....
Paul

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new SMART rom : if you DARE

Postby duikkie » Tue Apr 07, 2015 4:18 am

oke then i will look at *copy and *backup :(

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new SMART rom : if you DARE

Postby duikkie » Tue Apr 07, 2015 11:39 am

a day goes bye , and error has gone ?

hoping the *copy and *backup fixed

changed in files
j.3-drv line 9955 STY skipsec%

lines 10632 LDA #0
10634 STA skipsec %

in j.8-cow
3132 LDA #0
3234 STA skipsec%

reset to no skipsector if sector starts odd previously
Attachments
SMARTEST0704.zip
(61.68 KiB) Downloaded 40 times

User avatar
jgharston
Posts: 2756
Joined: Thu Sep 24, 2009 11:22 am
Location: Whitby/Sheffield

Re: working on new MMC rom YOUR missing it ?

Postby jgharston » Tue Apr 07, 2015 2:22 pm

duikkie wrote:i read a lot about memory use for beeb
&a0..&a7= nmi workspace but allso used extensively by the disc and network filing system. so when allso used by SD card , there are troubles
If you want to use the NMI space you must claim the NMI space with service call 12 and you must stop using it when you receive service call 12 from somebody else. You can't just go trampling on a shared resource without first stating that you want to use it.

duikkie wrote:i think i will reclocate the hole &a0..&a7 to &70..&77 space for user
&70-&77 is not allocated to the user, it is allocated to the user WHEN CALLING FROM BASIC. You MUST NOT use it at any other time.

duikkie wrote:&a8.. &af are for utilities area. i did store there the skipsector and clustex% bytes. so thats no good
&A8-&AF are the area specifically set aside for user commands. While you are running, you own &A8-&AF, as soon as you return to the caller you cease to own &A8-&AF and anybody else can trample on it.

duikkie wrote:&b0-- &bf is filing system scratch space
&c0-- &cf is filing system dedicated workspace
If you are the current filing system then you can use &B0-&CF to your heart's content. If you are not the current filing system, KEEP OFF.

Do you want to keep some persistant data? Then claim some private workspace - that's what it is for. If you only need one byte, then use your ROM's workspace byte. If you are doing something that only needs a couple of bytes and it's something that is physically impossible to use with another particular item, you can sometimes use that other item's workspace. For example, it is physically impossible to use a BBC shadow RAM system in a BBC Master, so they can use each other's workspace bytes.

duikkie wrote:no the mmc/sd card is not a filing system. it is a userport system.
Are you interfecpting FILEV? Does it allow you to do LOAD "file"? Then it's a filing system.

Code: Select all

$ bbcbasic
PDP11 BBC BASIC IV Version 0.25
(C) Copyright J.G.Harston 1989,2005-2015
>_

duikkie
Posts: 2689
Joined: Fri Feb 07, 2014 3:28 pm

Re: working on new MMC rom YOUR missing it ?

Postby duikkie » Tue Apr 07, 2015 2:45 pm

the trouble is the MMC or my SPI rom needs 4 bytes in page 0 , because &a0..&a1=datptr% , is pointer to adress used as lda (datptr%),Y=like 1900+y
i moved errptr% from &a8..&a9 to a2..a3.
the rest i moved to &d21 in claimed NMI space. i think datptr% is important to keep. no sure for how long.

problems are with DNFS rom not sure about the version number , and MMC/SPI rom. the NMI from DNFS looks at &fe18 and &FE20 ?? i see no docu on that in my books , from dfs,adv user guide. must be something for net ?? ( econet)

maybe datptr% can hide at &f8+&f9 not used by os 1.2 ??

but i don't have a econet system , so why DNFS is hanging with mmc/SPI rom i only hope i read it somewhere :)

jgharston wrote:
duikkie wrote:i read a lot about memory use for beeb
&a0..&a7= nmi workspace but allso used extensively by the disc and network filing system. so when allso used by SD card , there are troubles
If you want to use the NMI space you must claim the NMI space with service call 12 and you must stop using it when you receive service call 12 from somebody else. You can't just go trampling on a shared resource without first stating that you want to use it.

duikkie wrote:i think i will reclocate the hole &a0..&a7 to &70..&77 space for user
&70-&77 is not allocated to the user, it is allocated to the user WHEN CALLING FROM BASIC. You MUST NOT use it at any other time.

duikkie wrote:&a8.. &af are for utilities area. i did store there the skipsector and clustex% bytes. so thats no good
&A8-&AF are the area specifically set aside for user commands. While you are running, you own &A8-&AF, as soon as you return to the caller you cease to own &A8-&AF and anybody else can trample on it.

duikkie wrote:&b0-- &bf is filing system scratch space
&c0-- &cf is filing system dedicated workspace
If you are the current filing system then you can use &B0-&CF to your heart's content. If you are not the current filing system, KEEP OFF.

Do you want to keep some persistant data? Then claim some private workspace - that's what it is for. If you only need one byte, then use your ROM's workspace byte. If you are doing something that only needs a couple of bytes and it's something that is physically impossible to use with another particular item, you can sometimes use that other item's workspace. For example, it is physically impossible to use a BBC shadow RAM system in a BBC Master, so they can use each other's workspace bytes.

duikkie wrote:no the mmc/sd card is not a filing system. it is a userport system.
Are you interfecpting FILEV? Does it allow you to do LOAD "file"? Then it's a filing system.


Return to “software: other”

Who is online

Users browsing this forum: No registered users and 2 guests