avrdudes / avrdude

AVRDUDE is a utility to program AVR microcontrollers
GNU General Public License v2.0
738 stars 136 forks source link

AT89S51 and AT89S52 support for avrdude #1304

Closed orbanp closed 1 year ago

orbanp commented 1 year ago

Hello Everyone,

I am new here, so please advise if this is the wrong place to ask the following question.

I am trying to program an AT89S52 chip. I am using avrdude 6.3 and an ebay programmer which is a version of the usbasp programmer. The OS is Kubuntu 20.04 LTS. I am getting the following error message:

$ avrdude -p usbasp -c 5082 -U flash:r:"test.bin":r
avrdude: error at /etc/avrdude.conf:336: programmer type must be written as "id_type"
avrdude: error reading system wide configuration file "/etc/avrdude.conf"

I updated the configuration file that has an entry for the AT89S52 chip, the programmer description is the same that came with the installation. The suggestion that the programmer type should be written in the configuration file with double quotes did not make any difference.

Previously I installed avrdude 7.0 and got the same error. avrdude 6.3 came from the Software Center of the OS.

Any help would be appreciated! Thanks, Peter

MCUdude commented 1 year ago

Hi!

First of all, Avrdude doesn't support programming of AT89 chips out of the box, but with a custom avrdude.conf file it might work.

First, can you share the part of your avrdude.conf where the AT89S52 chip is defined?

orbanp commented 1 year ago

Thanks for the reply!

The error message complains about the programmer type!

I did try the original installation as well, without the updated config file, with another programmer, and with another device that were in the original config file, and got the same error message about the programmer type.

Enclosing the config file section about the AT89S52.

Thanks, Peter

#------------------------------------------------------------
# AT89S52
#------------------------------------------------------------
part
    id               = "8052";
    desc             = "AT89S52";
    signature        = 0x1E 0x52 0x06;
    chip_erase_delay = 500000;
    pgm_enable       = "1 0 1 0  1 1 0 0    0 1 0 1  0 0 1 1",
                       "x x x x  x x x x    x x x x  x x x x";

    chip_erase       = "1 0 1 0  1 1 0 0    1 0 0 x  x x x x",
                       "x x x x  x x x x    x x x x  x x x x";

    timeout      = 200;
    stabdelay      = 100;
    cmdexedelay      = 25;
    synchloops      = 32;
    bytedelay      = 0;
    pollindex      = 3;
    pollvalue      = 0x53;
    predelay      = 1;
    postdelay      = 1;
    pollmethod      = 0;

    memory "flash"
        size            = 8192;
        paged           = no;
        min_write_delay = 4000;
        max_write_delay = 9000;
        readback_p1     = 0xff;
        readback_p2     = 0xff;
        read            = "  0   0   1   0    0   0   0   0",
                          "  x   x   x a12  a11 a10  a9  a8",
                          " a7  a6  a5  a4   a3  a2  a1  a0",
                          "  o   o   o   o    o   o   o   o";

        write           = "  0   1   0   0    0   0   0   0",
                          "  x   x   x a12  a11 a10  a9  a8",
                          " a7  a6  a5  a4   a3  a2  a1  a0",
                          "  i   i   i   i    i   i   i   i";
   mode      = 0x21;
   delay      = 12;
      ;

    memory "signature"
        size            = 3;
        read            = "0  0  1  0   1  0  0  0   x  x  x  0   0  0 a1 a0",
                          "0  0  0  0   0  0  0  0   o  o  o  o   o  o  o  o";
      ;
  ; 
MCUdude commented 1 year ago

It might be something else that's wrong with your avrdude.conf. I'm using the latest 7.1 version, and I'm able to get Avrdude to run even though there are a few warnings. Try building Avrddue yourself, and add the AT89S52 target to the avrdude.conf file that was created by the build.

$ ./avrdude -cusbasp -p8052
avrdude warning: a1 would normally be expected to be a9 [/Users/hans/Downloads/avrdude/src/avrdude.conf:16243]
avrdude warning: a0 would normally be expected to be a8 [/Users/hans/Downloads/avrdude/src/avrdude.conf:16243]
avrdude warning: mcuid -1 for AT89S52 is out of range [0..2039], use a free number >= 372 [/Users/hans/Downloads/avrdude/src/avrdude.conf:16245]

avrdude error: program enable: target does not answer (0x01)
avrdude error: initialization failed, rc=-1
        - double check the connections and try again
        - use -B to set lower ISP clock frequency, e.g. -B 125kHz
        - use -F to override this check

avrdude done.  Thank you.
mcuee commented 1 year ago

Since this is not supported by avrdude, I will close this issue.

The command line mentioned in the first post is totally wrong. It should be avrdude -cusbasp -p8052 ....

llinjupt commented 1 year ago

Since this is not supported by avrdude, I will close this issue.

The command line mentioned in the first post is totally wrong. It should be avrdude -cusbasp -p8052 ....

Even though avrdude doesn't support AT89S series, for coherence, the 7.1 should behaviour the same as 7.0 which works well.

$ ./avrdude.exe -P COM180 -c avrisp -p 8052 -U flash:r:flash.bin:r

avrdude.exe: AVR device initialized and ready to accept instructions

Reading | ################################################## | 100% 0.00s

avrdude.exe: Device signature = 0x1e5206 (probably 89s52)
avrdude.exe: reading flash memory:

Reading | ################################################## | 100% 3.51s

avrdude.exe: writing output file "flash.bin"

avrdude.exe done.  Thank you.

It seems that 7.1 couldn't analyse the old avrdude.conf well, and throwed out the error:

$ ./avrdude.exe -P COM180 -c avrisp -p 8052 -U flash:r:flash.bin:r
avrdude warning: a1 would normally be expected to be a9 [F:\avrdude7.1\win64\avrdude.conf:49]
avrdude warning: a0 would normally be expected to be a8 [F:\avrdude7.1\win64\avrdude.conf:49]
avrdude warning: mcuid -1 for AT89S52 is out of range [0..2039], use a free number >= 372 [F:\avrdude7.1\win64\avrdude.conf:51]

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0xffffff (probably .xmega) (retrying)
avrdude: device signature = 0xffffff (probably .xmega) (retrying)
avrdude: device signature = 0xffffff (probably .xmega)
avrdude error: Yikes!  Invalid device signature.
avrdude error: expected signature for AT89S52 is 1E 52 06
        Double check connections and try again, or use -F to override
        this check.

avrdude done.  Thank you.

Anyway, for convenience I strongly recommand to keep the coherence with 7.0.

llinjupt commented 1 year ago

BTW, I have tried to update the avrdude.conf as below to fit 7.1:

part
    id               = "89s52";
    desc             = "AT89S52";
    stk500_devcode   = 0xE1;
    mcuid            = 372;
    signature        = 0x1e 0x52 0x06;
    chip_erase_delay = 500000;
    reset            = dedicated;
    pgm_enable       = "1010.1100--0101.0011--xxxx.xxxx--xxxx.xxxx";
    chip_erase       = "1010.1100--100x.xxxx--xxxx.xxxx--xxxx.xxxx";

    memory "flash"
        size            = 8192 ;             # bytes
        read            = "0010.0000--xxxa.aaaa--aaaa.aaaa--oooo.oooo";
        write           = "0100.0000--xxxa.aaaa--aaaa.aaaa--iiii.iiii";
        ;
    memory "signature"
        size            = 3;
        read            = "0010.1000--0000.00aa--0000.0000--oooo.oooo";
        ;
  ;

part parent "89s52"
    id = "89s51";
    desc = "AT89S51";
    signature = 0x1e 0x51 0x06;
    mcuid     = 373;
    memory "flash"
        size            = 4096 ;             # bytes
        read            = "0010.0000--xxxx.aaaa--aaaa.aaaa--oooo.oooo";
        write           = "0100.0000--xxxx.aaaa--aaaa.aaaa--iiii.iiii";
      ;
;

And I got another error. Since I know nothing about the a1 bit, any advice?

$ ./avrdude.exe -P COM180 -c avrisp -p AT89S52 -U flash:r:flash.bin:r
avrdude warning: AT89S52's signature read misses a necessary address bit a1 [F:\avrdude7.1\win64\avrdude.conf:20]

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e1e1e
avrdude error: expected signature for AT89S52 is 1E 52 06
        double check chip or use -F to override this check

avrdude done.  Thank you.
mcuee commented 1 year ago

Anyway, for convenience I strongly recommand to keep the coherence with 7.0.

No, the warnings will be there if you use old-style avrdude.conf file. And to me it is not worth the effort to support AT89S52 in avrdude.

mcuee commented 1 year ago

That being said, maybe @MCUdude or @stefanrueger will be able to see if they can help you here or not.

mcuee commented 1 year ago

Hmm, it seems to be that PlatformIO seems to support AT89S51/AT89S52 using avrdude. So maybe there are valid use case here. I will leave this to other admins to decide. https://docs.platformio.org/en/latest/boards/intel_mcs51/AT89S51.html https://docs.platformio.org/en/latest/boards/intel_mcs51/AT89S52.html

Therefore I have re-opened the issue and changed the issue titile and the issue label to "enhancement".

stefanrueger commented 1 year ago

Try the following entries [edit: added prog_modes]

#------------------------------------------------------------
# AT89S52
#------------------------------------------------------------

part
    desc                   = "AT89S52";
    id                     = "8052";
    prog_modes             = PM_ISP | PM_HVPP;
    mcuid                  = 373;
    chip_erase_delay       = 500000;
    signature              = 0x1e 0x52 0x06;
    timeout                = 200;
    stabdelay              = 100;
    cmdexedelay            = 25;
    synchloops             = 32;
    pollindex              = 3;
    pollvalue              = 0x53;
    predelay               = 1;
    postdelay              = 1;
    chip_erase             = "1010.1100--100x.xxxx--xxxx.xxxx--xxxx.xxxx";
    pgm_enable             = "1010.1100--0101.0011--xxxx.xxxx--xxxx.xxxx";

    memory "flash"
        size               = 8192;
        min_write_delay    = 4000;
        max_write_delay    = 9000;
        readback           = 0xff 0xff;
        mode               = 0x21;
        delay              = 12;
        read               = "0010.0000--xxxa.aaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxa.aaaa--aaaa.aaaa--iiii.iiii";
    ;

    memory "signature"
        size               = 3;
        read               = "0 0 1 0 1 0 0 0  x x x 0 0 0 a1 a0  0 0 0 0 0 0 0 0  o o o o o o o o";
    ;
;

#------------------------------------------------------------
# AT89S51
#------------------------------------------------------------

part parent "AT89S52"
    desc                   = "AT89S51";
    id                     = "89s51";
    mcuid                  = 372;
    signature              = 0x1e 0x51 0x06;

    memory "flash"
        size               = 4096;
        read               = "0010.0000--xxxx.aaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxx.aaaa--aaaa.aaaa--iiii.iiii";
    ;
;

One problem with this part is the odd documentation for the read-signature command (see below); it claims to utilise A7..A12, but that cannot be because signatures have 3 bytes, so only A0 and A1 should be relevant. I suspect the documentation is wrong, and that A0 and A1 need to be set. I looks like this part has A0 and A1 not in their standard position (end of Byte 3) but instead at end of Byte 2. All other parts known to AVRDUDE have the address bits in byte2 and byte3 in a systematic way, which has allowed a new shorter syntax for the command. So the read signature command for AT90S51/52 must use the old-style notation.

Fun fact: It used to be that contributors could not cope with the old-style notation to the effect that some 25% of the avrdude.conf entries had bits in the wrong position. This is why the warning message has been introduced. It can be ignored in this case (assuming A0/A1 have to be at the unsystematic position for these parts).

at89s52-read-signature

stefanrueger commented 1 year ago

@orbanp: The avrdude command mixed up the -p part and -c programmer flags. It should be -c usbasp -p 5082, not the other way round.

I am getting the following error message:

$ avrdude -p usbasp -c 5082 -U flash:r:"test.bin":r
avrdude: error at /etc/avrdude.conf:336: programmer type must be written as "id_type"
avrdude: error reading system wide configuration file "/etc/avrdude.conf"

Please try again and report back.

llinjupt commented 1 year ago

@stefanrueger, THX. I have tried your new avrdude.conf, but unluckly it didn't work:

$ ./avrdude.exe -P COM180 -c avrisp -p AT89S52 -U flash:r:flash.bin:r
avrdude warning: a1 would normally be expected to be a9 [F:\BiTForest\avrdude7.1\win64\avrdude.conf:36]
avrdude warning: a0 would normally be expected to be a8 [F:\BiTForest\avrdude7.1\win64\avrdude.conf:36]

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0xffffff (probably .xmega) (retrying)
avrdude: device signature = 0xffffff (probably .xmega) (retrying)
avrdude: device signature = 0xffffff (probably .xmega)
avrdude error: Yikes!  Invalid device signature.
avrdude error: expected signature for AT89S52 is 1E 52 06
        Double check connections and try again, or use -F to override
        this check.

avrdude done.  Thank you.

BTW, in the same environment, V7.0 with this old avrdude.conf works well:

part
    id               = "89s52";
    desc             = "AT89S52";
    stk500_devcode   = 0xE1;
    signature        = 0x1e 0x52 0x06;
    chip_erase_delay = 500000;
    reset            = dedicated;
    pgm_enable       = "1 0 1 0  1 1 0 0    0 1 0 1  0 0 1 1",
                       "x x x x  x x x x    x x x x  x x x x";

    chip_erase       = "1 0 1 0  1 1 0 0    1 0 0 x  x x x x",
                       "x x x x  x x x x    x x x x  x x x x";

    memory "flash"
        size            = 8192 ;             # bytes
        read            = "0 0 1 0 0 0 0 0  x x x a12 a11 a10 a9 a8",
                          "a7 a6 a5 a4 a3 a2 a1 a0 o o o o o o o o";
        write           = "0 1 0 0 0 0 0 0 x x x a12 a11 a10 a9 a8",
                          "a7 a6 a5 a4 a3 a2 a1 a0 i i i i i i i i";
        ;
    memory "signature"
        size            = 3;
        read            = "0 0 1 0 1 0 0 0  0 0 0 0 0 0 a1 a0",
                          "0 0 0 0 0 0 0 0  o o o o o o  o  o";
        ;
  ;
part parent "89s52"
    id = "89s51";
    desc = "AT89S51";
    signature = 0x1e 0x51 0x06;
    memory "flash"
        size            = 4096 ;             # bytes
        read            = "0 0 1 0 0 0 0 0  x x x x a11 a10 a9 a8",
                          "a7 a6 a5 a4 a3 a2 a1 a0 o o o o o o o o";
        write           = "0 1 0 0 0 0 0 0 x x x x a11 a10 a9 a8",
                          "a7 a6 a5 a4 a3 a2 a1 a0 i i i i i i i i";
      ;
;
stefanrueger commented 1 year ago

@llinjupt Thanks for sharing a working 7.0 entry. My guess was based on @orbanp's entry.

It might have been the missing stk500_devcode entry. Try in v7.1 the following (based on your entry)

#------------------------------------------------------------
# AT89S52
#------------------------------------------------------------

part
    desc                   = "AT89S52";
    id                     = "89s52";
    prog_modes             = PM_ISP | PM_HVPP;
    mcuid                  = 373;
    stk500_devcode         = 0xe1;
    chip_erase_delay       = 500000;
    signature              = 0x1e 0x52 0x06;
    chip_erase             = "1010.1100--100x.xxxx--xxxx.xxxx--xxxx.xxxx";
    pgm_enable             = "1010.1100--0101.0011--xxxx.xxxx--xxxx.xxxx";

    memory "flash"
        size               = 8192;
        read               = "0010.0000--xxxa.aaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxa.aaaa--aaaa.aaaa--iiii.iiii";
    ;

    memory "signature"
        size               = 3;
        read               = "0 0 1 0 1 0 0 0  0 0 0 0 0 0 a1 a0  0 0 0 0 0 0 0 0  o o o o o o o o";
    ;
;

#------------------------------------------------------------
# AT89S51
#------------------------------------------------------------

part parent "89s52"
    desc                   = "AT89S51";
    id                     = "89s51";
    mcuid                  = 372;
    signature              = 0x1e 0x51 0x06;

    memory "flash"
        size               = 4096;
        read               = "0010.0000--xxxx.aaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxx.aaaa--aaaa.aaaa--iiii.iiii";
    ;
;
llinjupt commented 1 year ago

@stefanrueger Good news it works and bad news with warninings:

$ ./avrdude.exe -P COM180 -c avrisp -p AT89S52 -U flash:r:flash.bin:r
avrdude warning: a1 would normally be expected to be a9 [F:\BiTForest\avrdude7.1\win64\avrdude.conf:24]
avrdude warning: a0 would normally be expected to be a8 [F:\BiTForest\vrdude7.1\win64\avrdude.conf:24]

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e5206 (probably 89s52)
avrdude: reading flash memory ...

Reading | ################################################## | 100% 3.54s

avrdude: writing output file flash.bin

avrdude done.  Thank you.
stefanrueger commented 1 year ago

This is the first time the project has come across an AVR where the address bits do not conform to their usual positions. When we create a PR to support AT89S51/52, we can try to suppress the warning for these (odd) parts.

Before I do a PR, please also try to write flash and verify. If you have the AT89S51, please also try that part.

BTW, in order to migrate a part entry from an older avrdude version, you can normally put that into avrdude.conf, add new parameters needed for the new version (here, prog_modes and mcuid, though neither would have been needed I guess) and then run the new avrdude with

$ avrdude -pAT89S5[12]/s

This outputs migrated, canonical entries suitable for the new version.

I also noted that the flash r/w commands have one address bit too many: it's word addresses, not byte addresses for flash. This should be better [edit - also updated stk500_devcode for the AT89S51]:

#------------------------------------------------------------
# AT89S52
#------------------------------------------------------------

part
    desc                   = "AT89S52";
    id                     = "89s52";
    prog_modes             = PM_ISP | PM_HVPP;
    mcuid                  = 373;
    stk500_devcode         = 0xe1;
    chip_erase_delay       = 500000;
    signature              = 0x1e 0x52 0x06;
    chip_erase             = "1010.1100--100x.xxxx--xxxx.xxxx--xxxx.xxxx";
    pgm_enable             = "1010.1100--0101.0011--xxxx.xxxx--xxxx.xxxx";

    memory "flash"
        size               = 8192;
        read               = "0010.0000--xxxx.aaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxx.aaaa--aaaa.aaaa--iiii.iiii";
    ;

    memory "signature"
        size               = 3;
        read               = "0 0 1 0 1 0 0 0  0 0 0 0 0 0 a1 a0  0 0 0 0 0 0 0 0  o o o o o o o o";
    ;
;

#------------------------------------------------------------
# AT89S51
#------------------------------------------------------------

part parent "89s52"
    desc                   = "AT89S51";
    id                     = "89s51";
    mcuid                  = 372;
    stk500_devcode         = 0xe0;
    signature              = 0x1e 0x51 0x06;

    memory "flash"
        size               = 4096;
        read               = "0010.0000--xxxx.xaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxx.xaaa--aaaa.aaaa--iiii.iiii";
    ;
;
stefanrueger commented 1 year ago

@llinjupt ... and I suspect that the AT89S51 needs a different stk500_devcode = 0xe0; (not the inherited 0xe1): https://github.com/avrdudes/avrdude/blob/4019edf60d34e2dcfdafe6d00c33f58db8615781/src/avrdude.conf.in#L344-L345

llinjupt commented 1 year ago

@llinjupt ... and I suspect that the AT89S51 needs a different stk500_devcode = 0xe0; (not the inherited 0xe1):

https://github.com/avrdudes/avrdude/blob/4019edf60d34e2dcfdafe6d00c33f58db8615781/src/avrdude.conf.in#L344-L345

@stefanrueger , Yes, you are right that AT89S51 has its own devcode. And I have checked how it is used. Actually, in avrisp, it's only used to detemine the rst polarity, and that's why the old avrdude.conf is accidentally ok for AT89S51:

rst_active_high = (param.devicecode >= 0xe0);

Reading back the flash and fill it up to the full falsh size: truncate.exe -s 8192 flash.bin

Wrting failed with your lastest avrdude.conf and check the datasheet there is really an A12 which seems can be ignored:

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e5206 (probably 89s52)
avrdude: Note: flash memory has been specified, an erase cycle will be performed.
         To disable this feature, specify the -D option.
avrdude: erasing chip
avrdude: reading input file flash.bin for flash
         with 8192 bytes in 1 section within [0, 0x1fff]
avrdude: writing 8192 bytes flash ...

Writing | #################################################- | 99% 7.02s

 ***failed;
 ***failed;
 ***failed;
 ***failed;
 ***failed;
avrdude: 8192 bytes of flash written
avrdude: verifying flash memory against flash.bin

Reading | ################################################## | 100% 3.50s

avrdude warning: verification mismatch
        device 0x00 != input 0x11 at addr 0x0040 (error)
avrdude error: verification mismatch

avrdude done.  Thank you.

There is no AT89S51 on hand, so I have ordered some chips already. Maybe two or three days the feedback will be ready.

Another subtle problem is that I have noticed the lock bits in the datasheet as well as the picture you pasted above. For reading there are 3 bits LB3/2/1 and for writing there are 2 bits B1/B2. Isn't it a bit confusing? Would you like to config it too for both 51/52 in avrdude.conf? And I could verify at the same time.

stefanrueger commented 1 year ago

Wrting failed

Thanks for testing. I suspect for writing to work the flash memory should specify the correct page_size > 1.

Maybe try, 2, 4, 8, 16, ...

It is a pity there does not seem to be an .atdf file (or .xml file as they were called in the early days) containing the necessary write delay etc parameters that need to go into the .conf file. The data sheet is very terse, too.

In order for HVPP programming to work, other parameters would need to be specified, too, in avrdude.conf. Maybe try setting the other parameters like in @orbanp 's entry? Or copying parameters from "similar" other parts in avrdude.conf for this entry? You could try a few settings.

I suspect AVRDUDE will be able to read the lock bits but won't be able to write them: The .conf file can set pretty much any SPI command (the old notation has a very versatile syntax!), but apparently the footnote in the data sheet says that one needs several writes of different SPI-commands to effect certain lock levels... That procedure isn't in the code base (and is probably not worth our while implementing!)

stefanrueger commented 1 year ago

check the datasheet there is really an A12

Ooops that's right. And the data sheet says "byte mode" for flash r/w. So, page size 1 would appear correct for flash. And one would need a12 in the flash read/write commands (as it's a byte address not a word address).

So, this makes me wonder. Has your .conf file also worked for writing in AVRDUDE v7.0 or any earlier AVRDUDE? If so let us know.

stefanrueger commented 1 year ago

Here another one to try with some sensible other values set

#------------------------------------------------------------
# AT89S51
#------------------------------------------------------------

part
    desc                   = "AT89S51";
    id                     = "89S51";
    prog_modes             = PM_ISP | PM_HVPP;
    mcuid                  = 372;
    stk500_devcode         = 0xe0;
    chip_erase_delay       = 250000;
    signature              = 0x1e 0x51 0x06;
    timeout                = 200;
    stabdelay              = 100;
    cmdexedelay            = 25;
    synchloops             = 32;
    pollindex              = 4;
    pollvalue              = 0x69;
    predelay               = 1;
    postdelay              = 1;
    chiperasepulsewidth    = 15;
    programfusepulsewidth  = 2;
    programlockpolltimeout = 1;
    chip_erase             = "1010.1100--100x.xxxx--xxxx.xxxx--xxxx.xxxx";
    pgm_enable             = "1010.1100--0101.0011--xxxx.xxxx--xxxx.xxxx";

    memory "flash"
        size               = 4096;
        min_write_delay    = 4500;
        max_write_delay    = 4500;
        readback           = 0xff 0xff;
        mode               = 0x02;
        delay              = 10;
        blocksize          = 256;
        read               = "0010.0000--xxxa.aaaa--aaaa.aaaa--oooo.oooo";
        write              = "0100.0000--xxxa.aaaa--aaaa.aaaa--iiii.iiii";
        # Page mode is available in the device, but not implemented here.
    ;

    memory "signature"
        size               = 3;
        read               = "0 0 1 0 1 0 0 0  x x x x x x a1 a0  x x x x x x x 0  o o o o o o o o";
    ;
    # Lock bits are available in the device but not implemented here
;

#------------------------------------------------------------
# AT89S52
#------------------------------------------------------------

part parent "89S51"
    desc                   = "AT89S52";
    id                     = "89S52";
    mcuid                  = 373;
    stk500_devcode         = 0xe1;
    signature              = 0x1e 0x52 0x06;

    memory "flash"
        size               = 8192;
        # Page mode is available in the device but not implemented here
    ;
  # Lock bits are available in the device but not implemented here
;
llinjupt commented 1 year ago

@stefanrueger

That procedure isn't in the code base (and is probably not worth our while implementing!)

I see. If the code base don't suppport, it's reasonable and acceptable.

Has your .conf file also worked for writing in AVRDUDE v7.0 or any earlier AVRDUDE? If so let us know.

Yes. I have tried 6.3 and it worked well.

With your newest avrdude.conf, both reading and writing are passed for 89S52. And as soon as 89S51 is arrived I will give feedback.

$ ./avrdude.exe -c avrisp -P COM245 -p AT89S52 -U flash:w:AT89S52.bin:r
avrdude warning: a1 would normally be expected to be a9 [F:\BiTForest\ATMEL\TEST_BINs\avrdude.conf:42]
avrdude warning: a0 would normally be expected to be a8 [F:\BiTForest\ATMEL\TEST_BINs\avrdude.conf:42]

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e5206 (probably 89S52)
avrdude: Note: flash memory has been specified, an erase cycle will be performed.
         To disable this feature, specify the -D option.
avrdude: erasing chip
avrdude: reading input file AT89S52.bin for flash
         with 8192 bytes in 1 section within [0, 0x1fff]
avrdude: writing 8192 bytes flash ...

Writing | ################################################## | 100% 24.57s

avrdude: 8192 bytes of flash written
avrdude: verifying flash memory against AT89S52.bin

Reading | ################################################## | 100% 8.19s

avrdude: 8192 bytes of flash verified

avrdude done.  Thank you.
llinjupt commented 1 year ago

And the data sheet says "byte mode" for flash r/w. So, page size 1 would appear correct for flash. And one would need a12 in the flash read/write commands (as it's a byte address not a word address).

Yes, the address space is exactly the same as the flash size. And I have noticed AT89S51/2 all support page mode read/write. Is it possible to support it? If so, it would be great for speed consideration.

orbanp commented 1 year ago

Hello Everyone,

I appreciate the continuing help and discussions!

To reiterate, I have avrdude version 3.6 on my machine, ubuntu 20.04 with KDE frontend.

I have two hardware interfaces for programming. One of them is a pair of usbasp dongles. The usbasp interfces do work, I did upgrade the firmware in them, programming one with the other. It was a couple of years ago when I obtained them on ebay, not sure about the firmware version.

The other interface is a simplified version of the STK200 dongle, a cable from the parallel port of the PC, with four simple wires for the reset, in and out data, and clock signals, plus ground. The cable works. I could read/write (program) the AT89S52 flash with it using an old program under W10, the IspPgm30a, a program that has been long abandoned and has some issues.

Looking at the signals with a scope, using the usbasp interface, that hardware does not handle the reset signal. The AT89S52 has an active high reset, and it needs to be set to high to initiate ISP programming and the usbasp interface does not do that.

"My STK200" dongle shows some life when trying to program. It initiates contact with the processor, however the signature and data returned is invalid. I defined a new programmer in the config file, same features as that of the STK200, with the reset signal inverted. Using that hardware description the reset signal is raised high and I can see the clock signal and the input data signal changing, however the data is not what I would expect when forcing the flash read operation. There is a correct program in the processor, it runs as it is supposed to, but the data received is a simple counter, the values increasing to 00x to FFx. The processor is running at 12MHz, the ISP programming max speed could be 1/16-the of that clock speed, that is 750kHz. The scope shows that the ISP clock speed is 500kHz.

Again, here is the processor description in the config file that I have been using with avrdude 6.3:

#------------------------------------------------------------
# AT89S52
#------------------------------------------------------------
part
    id               = "8052";
    desc             = "AT89S52";
    signature        = 0x1E 0x52 0x06;
    chip_erase_delay = 500000;
    pgm_enable       = "1 0 1 0  1 1 0 0    0 1 0 1  0 0 1 1",
                       "x x x x  x x x x    x x x x  x x x x";

    chip_erase       = "1 0 1 0  1 1 0 0    1 0 0 x  x x x x",
                       "x x x x  x x x x    x x x x  x x x x";

    timeout      = 200;
    stabdelay      = 100;
    cmdexedelay      = 25;
    synchloops      = 32;
    bytedelay      = 0;
    pollindex      = 3;
    pollvalue      = 0x53;
    predelay      = 1;
    postdelay      = 1;
    pollmethod      = 0;

    memory "flash"
        size            = 8192;
        paged           = no;
        min_write_delay = 4000;
        max_write_delay = 9000;
        readback_p1     = 0xff;
        readback_p2     = 0xff;
        read            = "  0   0   1   0    0   0   0   0",
                          "  x   x   x a12  a11 a10  a9  a8",
                          " a7  a6  a5  a4   a3  a2  a1  a0",
                          "  o   o   o   o    o   o   o   o";

        write           = "  0   1   0   0    0   0   0   0",
                          "  x   x   x a12  a11 a10  a9  a8",
                          " a7  a6  a5  a4   a3  a2  a1  a0",
                          "  i   i   i   i    i   i   i   i";
   mode      = 0x21;
   delay      = 12;
      ;

    memory "signature"
        size            = 3;
        read            = "0  0  1  0   1  0  0  0   x  x  x  0   0  0 a1 a0",
                          "0  0  0  0   0  0  0  0   o  o  o  o   o  o  o  o";
      ;
  ; 

Also enclose the STK200 description with the inverted reset:

# Parallel cable programmer with inverted RESET signal
# It is suitable, among others, for the AT89S52

programmer
  id    = "stk200_I_R";
  desc  = "STK200 with inverted Reset";
  type  = "par";
  connection_type = parallel;
  buff  = 4, 5;
  sck   = 6;
  mosi  = 7;
  reset = ~9;
  miso  = 10;
;

Thanks, Peter

stefanrueger commented 1 year ago

@orbanp I suspect adding the correct stk500_devcode will make a difference (0xe0 for 89s50 and 0xe1 for 89s51); also try out PR #1310.

@llinjupt Thanks for confirming r/w now works for AT89S52. I've created PR #1310 so you can try that out (also later with the AT89S51). Note I added lock read; please check (though you probably won't see anything more interesting than 0xff, I suspect!)

AT89S51/2 all support page mode read/write. Is it possible to support it?

No, page mode is nonstandard so unless we receive a working PR we won't be able to support that. Same with lock write. I guess(!) you could add memories lockmode0, lockmode1, lockmode2 and lockmode3 each with a different lock write command as detailed in the data sheet and then write the lock modes in sequence -U lockmode0:w:0xff:m -U lockmode1:w:0xff:m -U lockmode2:w:0xff:m ... as many times as the lock bit(s) you want to set as alluded to in Note 1 on p27 (but no warranties if my interpretation of the data sheet bricks your device :)

llinjupt commented 1 year ago

@stefanrueger

With your patch, AT89S51/2 are OK to read and write.

$ ./avrdude.exe -P COM245 -c avrisp -p at89s51 -U flash:w:AT89S51.bin:r

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e5106 (probably 89S51)
avrdude: Note: flash memory has been specified, an erase cycle will be performed.
         To disable this feature, specify the -D option.
avrdude: erasing chip
avrdude: reading input file AT89S51.bin for flash
         with 4096 bytes in 1 section within [0, 0xfff]
avrdude: writing 4096 bytes flash ...

Writing | ################################################## | 100% 12.29s

avrdude: 4096 bytes of flash written
avrdude: verifying flash memory against AT89S51.bin

Reading | ################################################## | 100% 4.10s

avrdude: 4096 bytes of flash verified

avrdude done.  Thank you.

as many times as the lock bit(s) you want to set as alluded to in Note 1 on p27 (but no warranties if my interpretation of the data sheet bricks your device :)

Yes, your interpretation is correct. I have write and read back lock bits and the value is just like the datasheet said. BTW, these lock bits' fucntion is very different from morden MCUs and I don't think there are many people who are interested in it. I just put the write cmd here for reference.

    memory "lock"
        size               = 1;
        read               = "0010.0100--xxxx.xxxx--xxxx.xxxx--xxxo.ooxx";
        write              = "1010.1100--1110.00ii--xxxx.xxxx--xxxx.xxxx";
        # Nonstandard lock write is available but not implemented
    ;
orbanp commented 1 year ago

Hello Everyone,

I am glad to hear that at least for some of you the AT89S52 programming works with the avrisp. It still does not work for me with the usbasp programmer!

I installed avrdude 7.1 with the LINUXSPI and PARPORT options added. I can read/write the ATmega8A on the usbasp dongle, I updated the firmware to the latest version.

Added to the config file stefanrueger's latest S51/S52 part descriptions. (Also added the STK200 type programmers, that is why some of the warnings.) When trying to read the AT89S52 I am getting the following errors:

$ sudo avrdude -p 89s52 -c usbasp -U flash:r:flash.hex -B 125kHz -F
avrdude warning: mosi is deprecated, will be removed in v8.0, use sdo [/usr/local/etc/avrdude.conf:457]
avrdude warning: miso is deprecated, will be removed in v8.0, use sdi [/usr/local/etc/avrdude.conf:459]
avrdude warning: a1 would normally be expected to be a9 [/usr/local/etc/avrdude.conf:2441]
avrdude warning: a0 would normally be expected to be a8 [/usr/local/etc/avrdude.conf:2441]

avrdude: set SCK frequency to 93750 Hz
avrdude error: program enable: target does not answer (0x01)
avrdude error: initialization failed, rc=-1
        - double check the connections and try again
        - use -B to set lower ISP clock frequency, e.g. -B 125kHz
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x000000 (retrying)
avrdude: device signature = 0x000000 (retrying)
avrdude: device signature = 0x000000
avrdude error: Yikes!  Invalid device signature.
avrdude warning: expected signature for AT89S52 is 1E 52 06

avrdude done.  Thank you.

Interesting that first the device fails to initialize, but after a while it does, but the programmer still fails to get the correct signature. Also, -F does not seem to have any effect.

Any clue what is wrong when trying to use the usbasp for programming the S52?

Thanks, Peter

llinjupt commented 1 year ago

@orbanp,

  1. Check if there is stk500_devcode for AT89S51/2 in avrdude.conf. Trying PR #1310 is a better choice.
  2. AFAIK, usbasp has a special firmware to support AT8xxx,did you update the right FW? Could you check if the rst is pulled high with a scope while do programming.
  3. If still not work, with -vvvv and paste the message.
orbanp commented 1 year ago

Thanks for the help!

Yes, there is the stk500_devcode for AT89S51/2 in avrdude.conf:

stk500_devcode = 0xe1;

I am not sure what do you mean by "PR #1310 is a better choice"? That link brings me back to this same discussion!

For updating the firmware in the the usbasp dongles I used "usbasp.atmega8.2011-05-28.hex" file from Thomas Fischl's web site. The usbasp dongles I have have ATmega8A processors.

Checking with a scope the reset line is not pulled high when trying to program with the usbasp, as you suggested.

This is the reply with the -vvvv option, hopefully it gives some clue as to what is amiss:

$ sudo avrdude -p 89s52 -c usbasp -U flash:r:flash.hex -F -vvvv -B 125kHz

avrdude: Version 7.1-20230304 (4019edf)
         Copyright the AVRDUDE authors;
         see https://github.com/avrdudes/avrdude/blob/main/AUTHORS

         System wide configuration file is /usr/local/etc/avrdude.conf
avrdude yywarning() [config.c:242] warning: mosi is deprecated, will be removed in v8.0, use sdo [/usr/local/etc/avrdude.conf:457]
avrdude yywarning() [config.c:242] warning: miso is deprecated, will be removed in v8.0, use sdi [/usr/local/etc/avrdude.conf:459]
avrdude yywarning() [config.c:242] warning: a1 would normally be expected to be a9 [/usr/local/etc/avrdude.conf:2441]
avrdude yywarning() [config.c:242] warning: a0 would normally be expected to be a8 [/usr/local/etc/avrdude.conf:2441]
         User configuration file is /root/.avrduderc
         User configuration file does not exist or is not a regular file, skipping

         Using Port                    : usb
         Using Programmer              : usbasp
         Setting bit clk period        : 8.0
avrdude: usbasp_open("usb")
avrdude: seen device from vendor >www.fischl.de<
avrdude: seen product >USBasp<
         AVR Part                      : AT89S52
         Chip Erase delay              : 250000 us
         RESET disposition             : dedicated
         RETRY pulse                   : SCK
         Serial program mode           : yes
         Parallel program mode         : yes
         Timeout                       : 200
         StabDelay                     : 100
         CmdexeDelay                   : 25
         SyncLoops                     : 32
         PollIndex                     : 4
         PollValue                     : 0x69
         Memory Detail                 :

                                           Block Poll               Page                       Polled
           Memory Type Alias    Mode Delay Size  Indx Paged  Size   Size #Pages MinW  MaxW   ReadBack
           ----------- -------- ---- ----- ----- ---- ------ ------ ---- ------ ----- ----- ---------
           flash                   2    10   256    0 no       8192    1      0  4500  4500 0xff 0xff
                                           Block Poll               Page                       Polled
           Memory Type Alias    Mode Delay Size  Indx Paged  Size   Size #Pages MinW  MaxW   ReadBack
           ----------- -------- ---- ----- ----- ---- ------ ------ ---- ------ ----- ----- ---------
           signature               0     0     0    0 no          3    1      0     0     0 0x00 0x00

         Programmer Type : usbasp
         Description     : USBasp, http://www.fischl.de/usbasp/

avrdude: usbasp_initialize()
avrdude: usbasp_transmit("USBASP_FUNC_GETCAPABILITIES", 0x00, 0x00, 0x00, 0x00)
         <= [01] [00] [00] [00] 
avrdude: usbasp_spi_set_sck_period(8e-06)
avrdude: try to set SCK period to 8e-06 s (= 125000 Hz)
avrdude: connected USBasp is not cabable of 3 MHz SCK
avrdude: set SCK frequency to 93750 Hz
avrdude: usbasp_transmit("USBASP_FUNC_SETISPSCK", 0x08, 0x00, 0x00, 0x00)
         <= [00] 
avrdude: usbasp_transmit("USBASP_FUNC_CONNECT", 0x00, 0x00, 0x00, 0x00)
avrdude: usbasp_program_enable()
avrdude: usbasp_transmit("USBASP_FUNC_ENABLEPROG", 0x00, 0x00, 0x00, 0x00)
         <= [01] 
avrdude usbasp_spi_program_enable() [usbasp.c:726] error: program enable: target does not answer (0x01)
avrdude main() [main.c:1246] error: initialization failed, rc=-1
        - double check the connections and try again
        - use -B to set lower ISP clock frequency, e.g. -B 125kHz
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x000000 (retrying)
avrdude: device signature = 0x000000 (retrying)
avrdude: device signature = 0x000000
avrdude main() [main.c:1349] error: Yikes!  Invalid device signature.
avrdude main() [main.c:1364] warning: expected signature for AT89S52 is 1E 52 06
avrdude: usbasp_close()
avrdude: usbasp_transmit("USBASP_FUNC_DISCONNECT", 0x00, 0x00, 0x00, 0x00)

avrdude done.  Thank you.

Thanks, Peter

llinjupt commented 1 year ago

For updating the firmware in the the usbasp dongles I used "usbasp.atmega8.2011-05-28.hex" file from Thomas Fischl's web site. The usbasp dongles I have have ATmega8A processors.

No. AFAIK, these firmwares don't support AT89 series. Google or search in https://www.avrfreaks.net/forum may help you find one working firmware hacked for AT89 series. If the souce is avaliable, search AT89/89S51/2 as keywords to see if if is supported. AT89 series need a high-level reset signal which is not the same as other parts.

mcuee commented 1 year ago

I am not sure what do you mean by "PR https://github.com/avrdudes/avrdude/pull/1310 is a better choice"? That link brings me back to this same discussion!

@orbanp

I do not have the AT89S51/52 chip to test but you are not using PR #1310. You are using git main (https://github.com/avrdudes/avrdude/commit/4019edf60d34e2dcfdafe6d00c33f58db8615781) PR #1310 has one more commit https://github.com/stefanrueger/avrdude/commit/6c8f7e10fd4e6d3ee9184e9e7375efcf64e17277 than git main.

You can build PR #1310 using the following command.

git clone https://github.com/stefanrueger/avrdude.git avrdude_pr1310
cd avrdude_pr1310
git checkout at89s5x
./build.sh
mcuee commented 1 year ago

@llinjupt

Just wondering which usbasp firmware are you using to program the AT89S51/52. Which USBasp variant are you using?

From the following thread -- looks like only the Chinese clone usbasp has the firmware to support AT89S51/52. https://www.avrfreaks.net/s/topic/a5C3l000000UV6aEAG/t133362

There are usually two types of USBasp clones. 1) using ATmega8A, often called USBASP, this will usually came with some modified usbasp firmware which can work with avrdude but often had some warnings. I am not so sure whether they support AT89S51/52 or not.

2) using ATmeag88A, which comes with USB HID FW by default, often called USBISP. This USB HID FW is supposed to support Atmel AVR and AT89S51/52 using the ProgISP Windows utility. But this USB HID FW is not supported by by avrdude but you can flash alternative firmware to get it working with avrdude (eg: https://github.com/nerdralph/usbasp)

Both variants are said to be supported by Chinese ProgISP utility (Windows only) to program the AVR chips. But I am not so sure about ATmel AT89 8051 chips.

ProgISP supports many Atmel 8051 chips but the software was developed by a Chinese company for their own 8051/AVR programmers. So I am not so sure which usbasp firmware can support AT89 chips. https://github.com/gen-so/PROGISP-V1.72/tree/main/eng

Are you using the first variant (ATmega8A) with the Chinese mod firmware?

What does that -P COM245 means in your command?

$ ./avrdude.exe -P COM245 -c avrisp -p at89s51 -U flash:w:AT89S51.bin:r
mcuee commented 1 year ago

For updating the firmware in the the usbasp dongles I used "usbasp.atmega8.2011-05-28.hex" file from Thomas Fischl's web site. The usbasp dongles I have have ATmega8A processors.

No. AFAIK, these firmwares don't support AT89 series. Google or search in https://www.avrfreaks.net/forum may help you find one working firmware hacked for AT89 series. If the souce is avaliable, search AT89/89S51/2 as keywords to see if if is supported. AT89 series need a high-level reset signal which is not the same as other parts.

@llinjupt Indeed you are right. Thanks for the info. Google found this one, which has the firmware. https://labprojectsbd.com/2020/04/19/make-your-own-8051-mcu-programmer-with-usbasp/

@orbanp You may want to give the alternative Chinese firmware a try to see if that works for you or not. Please use PR #1310 for your test. Thanks.

mcuee commented 1 year ago

@stefanrueger

I am thinking that we need to have a new programmer type called usbasp_8051 to differentiate the above two types of FW (Chinese FW which supports 8051, original USBasp firmware or improvement firmware which does not support 8051).

As far as I know, the Chinese FW is not open as I can not find the source code.

BTW, I also can not find the USB HID FW source codes for the USBISP (ATmega88A) but that is anyway not supported by avrdude.

llinjupt commented 1 year ago

@llinjupt Indeed you are right. Thanks for the info. Google found this one, which has the firmware. https://labprojectsbd.com/2020/04/19/make-your-own-8051-mcu-programmer-with-usbasp/

I don't remember when I started fiddling with USBASP just like the Chinese clone one. And it didnt' seem verfy powerful and useful to me due to no labels of pins, no open and things like this. So I gave up and never touch it. I recommand avrisp and avrispmkii which is faster, open, easy to be hacking.

BTW I am working on a project named iLINK Infinite (means my or your own programmer with infite abilities if you like hacking) which is still on going, trying to support most popular MCUs, including STM32/STM8/AVR(ISP,UPDI,TPI,PDI)/PIC8/32/MSP430xx, even programming SPI EEPROMs and etc. Moreover, it's tiny just like the STLINKv2 Clone in an aluminum case, while it's also an open platform for people to diy they own firmwares and upload to it without taking the risk of losing the original firmware. Maybe to your surprise, the ultimate aim is to make it even cheaper than a cup of Starbucks and support most open soueced PC programs like avrdude. Guess it's going to be released in the second quarter of this year. If you are interested, I will let you know that time.

mcuee commented 1 year ago

@llinjupt

Nice one with your iLINK Infinite. I am interested for sure -- I am interested in PIC18/32, AVR and STM32 myself.

BTW, for usbasp, it is actually very popular for avrdude users (for Microchip/Atmel AVR chips, not for Microchip/Atmel 8051 chips which are not officially supported) due to good support from avrdude and available good alternative FW (eg: https://github.com/nerdralph/usbasp and https://github.com/dioannidis/usbasp).

There are alternative FW/SW for more features for USBasp.

Another thing, last time Simon Qian developed Versaloon with similar objective like your iLINK Infinite. It is supported by OpenOCD to program ARM chips (not only STM32). It has many targets (https://github.com/versaloon/versaloon/tree/master/src/target), including AVR, AVR32, MSP430, STM8 and STM32. It was popular for a while.

Versaloon https://github.com/versaloon/versaloon (official but not updated any more) https://github.com/lamfe/versaloon-git (fork) https://github.com/zoobab/versaloon (fork with more updates)

orbanp commented 1 year ago

Hello Everyone,

Some progress.

Installed avrdude_pr1310

Updated the firmware in the usbasp dongle. The dongle version I have is a Chinese, eBay purchased one, marked "USBASP V2.0 LC Technology". It uses the ATmega 8A processor running at 12MHz. First I tried the Chinese firmware, I did not talk to the AT89S52, neither to the ATmega8A in the other dongle.

After that I used the firmware found here on github: https://github.com/mamut-tme/UsbAsp-flash_89S52/blob/master/README.md This combination of avrdude and firmware talks to the AT89S52, gets the signature correctly, however the data read back from the chip is garbled, it does not match what is in the processor. Similarly, when writing the flash the signature is correct, but the correct data transfer fails, garbled data is written out. Enclosing the printout with the -vvvv qualifier:

$ sudo avrdude -p 89s52 -c usbasp -U flash:w:code.hex -vvvv

avrdude: Version 7.1-20230306 (6c8f7e1)
         Copyright the AVRDUDE authors;
         see https://github.com/avrdudes/avrdude/blob/main/AUTHORS

         System wide configuration file is /usr/local/etc/avrdude.conf
         User configuration file is /root/.avrduderc
         User configuration file does not exist or is not a regular file, skipping

         Using Port                    : usb
         Using Programmer              : usbasp
avrdude: usbasp_open("usb")
avrdude: seen device from vendor >www.fischl.de<
avrdude: seen product >USBasp<
avrdude: input file code.hex auto detected as Intel Hex
         AVR Part                      : AT89S52
         Chip Erase delay              : 250000 us
         RESET disposition             : dedicated
         RETRY pulse                   : SCK
         Serial program mode           : yes
         Parallel program mode         : yes
         Timeout                       : 200
         StabDelay                     : 100
         CmdexeDelay                   : 25
         SyncLoops                     : 32
         PollIndex                     : 4
         PollValue                     : 0x69
         Memory Detail                 :

                                           Block Poll               Page                       Polled
           Memory Type Alias    Mode Delay Size  Indx Paged  Size   Size #Pages MinW  MaxW   ReadBack
           ----------- -------- ---- ----- ----- ---- ------ ------ ---- ------ ----- ----- ---------
           flash                   2    10   256    0 no       8192    1      0  4500  4500 0xff 0xff
                                           Block Poll               Page                       Polled
           Memory Type Alias    Mode Delay Size  Indx Paged  Size   Size #Pages MinW  MaxW   ReadBack
           ----------- -------- ---- ----- ----- ---- ------ ------ ---- ------ ----- ----- ---------
           lock                    0     0     0    0 no          1    1      0     0     0 0x00 0x00
                                           Block Poll               Page                       Polled
           Memory Type Alias    Mode Delay Size  Indx Paged  Size   Size #Pages MinW  MaxW   ReadBack
           ----------- -------- ---- ----- ----- ---- ------ ------ ---- ------ ----- ----- ---------
           signature               0     0     0    0 no          3    1      0     0     0 0x00 0x00

         Programmer Type : usbasp
         Description     : USBasp, http://www.fischl.de/usbasp/

avrdude: usbasp_initialize()
avrdude: usbasp_transmit("USBASP_FUNC_GETCAPABILITIES", 0x00, 0x00, 0x00, 0x00)
         <= [01] [00] [01] [01] 
avrdude: usbasp_spi_set_sck_period(0)
avrdude: auto set sck period (because given equals null)
avrdude: usbasp_transmit("USBASP_FUNC_SETISPSCK", 0x00, 0x00, 0x00, 0x00)
         <= [00] 
avrdude: usbasp_transmit("USBASP_FUNC_CONNECT", 0x00, 0x00, 0x00, 0x00)
avrdude: usbasp_program_enable()
avrdude: usbasp_transmit("USBASP_FUNC_ENABLEPROG", 0x00, 0x00, 0x00, 0x00)
         <= [00] 
avrdude: AVR device initialized and ready to accept instructions

Reading |                                                    | 0% 0.00 s avrdude: usbasp_spi_cmd(0x28, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x28, 0x00, 0x00, 0x00)
         <= [00] [28] [00] [1e] 
avrdude: usbasp_spi_cmd() => 0x00, 0x28, 0x00, 0x1e
avrdude: usbasp_spi_cmd(0x28, 0x01, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x28, 0x01, 0x00, 0x00)
         <= [00] [28] [01] [52] 
avrdude: usbasp_spi_cmd() => 0x00, 0x28, 0x01, 0x52
Reading | #################                                  | 33% 0.00 s avrdude: usbasp_spi_cmd(0x28, 0x02, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x28, 0x02, 0x00, 0x00)
         <= [00] [28] [02] [06] 
avrdude: usbasp_spi_cmd() => 0x00, 0x28, 0x02, 0x06
Reading | ################################################## | 100% 0.00 s 

avrdude: device signature = 0x1e5206 (probably 89S52)
avrdude: Note: flash memory has been specified, an erase cycle will be performed.
         To disable this feature, specify the -D option.
avrdude: erasing chip
avrdude: usbasp_chip_erase()
avrdude: usbasp_spi_cmd(0xac, 0x80, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0xac, 0x80, 0x00, 0x00)
         <= [00] [ac] [00] [00] 
avrdude: usbasp_spi_cmd() => 0x00, 0xac, 0x00, 0x00
avrdude: usbasp_initialize()
avrdude: usbasp_transmit("USBASP_FUNC_GETCAPABILITIES", 0x00, 0x00, 0x00, 0x00)
         <= [01] [00] [01] [01] 
avrdude: usbasp_spi_set_sck_period(0)
avrdude: auto set sck period (because given equals null)
avrdude: usbasp_transmit("USBASP_FUNC_SETISPSCK", 0x00, 0x00, 0x00, 0x00)
         <= [00] 
avrdude: usbasp_transmit("USBASP_FUNC_CONNECT", 0x00, 0x00, 0x00, 0x00)
avrdude: usbasp_program_enable()
avrdude: usbasp_transmit("USBASP_FUNC_ENABLEPROG", 0x00, 0x00, 0x00, 0x00)
         <= [00] 
avrdude: reading input file code.hex for flash
         with 5144 bytes in 1 section within [0, 0x1fff]
avrdude: writing 5144 bytes flash ...

Writing |                                                    | 0% 0.00 s avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x40, 0x00, 0x00, 0x02) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x40, 0x00, 0x00, 0x02)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff] 
avrdude: usbasp_spi_cmd() => 0xff, 0xff, 0xff, 0xff
avrdude: usbasp_spi_cmd(0x20, 0x00, 0x00, 0x00) ...
avrdude: usbasp_transmit("USBASP_FUNC_TRANSMIT", 0x20, 0x00, 0x00, 0x00)
         <= [ff] [ff] [ff] [ff]

Thanks, Peter

stefanrueger commented 1 year ago

So it seems with PR #1310 avrdude supports AT89S51/2 with the avrisp programmer. So, the suggested avrdude.conf entries work in principle, ie, the SPI commands do what they should

Unclear to me why usbasp deosn't work. Has that ever worked in previous versions of avrdude? If so, which .conf entry?

mcuee commented 1 year ago

Unclear to me why usbasp deosn't work. Has that ever worked in previous versions of avrdude? If so, which .conf entry?

@stefanrueger There are different reports and different FW involved. I am actually thinking that we may not be able to support usbasp very effectively for AT89S51/AT89S52.

1) Chinese FW versions are not open but is supposed to work with ProgISP which supports many 8051 chips and memory chips along with AVR chips. There are some reports that modified avrdude works with this type of FW.

2) This repo claims support of AT89S51 and AT89S52 with patched version avrudude 6.3. However it does not seem to have a proper .conf entry for these two chips. It seems to contain some changes to usbasp.c and usbasp.h. https://github.com/mamut-tme/UsbAsp-flash_89S52/tree/master

@orbanp Have you tried the patched version of avrdude 6.3 mentioned in the above repo?

stefanrueger commented 1 year ago

When PR #1310 gets merged this issue will be closed b/c AT89S51/2 is then supported by -c avrisp.

@orbanp once you know more about reasons for your usbasp programmer not working for these parts, please open a dedicated issue for that, ideally with a suggested solution that you might find in the pointers that @mcuee has found above.

We rely on community contributions, particularly for non-standard parts.

orbanp commented 1 year ago

Hello Everyone,

Just curious, where do I find the patched version of avrdude 6.3? The link that mentioned it could not find the page.

I will see if I can put together a simple logic analyzer to test the programming interface with an arduino uno board that I have. I downloaded PulseView, sigrok, and an arduino library that implements logic analyzer protocols. Also found an "ISP decoder" for PulseView to display the communication in a more readable format. Hopefully I can make the above work.

Another question. Whenever I sign into github it always asks for and sends a verification code. It is a bit tiresome. Is this normal, or just my browser is configured too "strictly"? If that is the case, specifically, what do I need to allow in my browser to get in without verification?

Thanks, Peter

stefanrueger commented 1 year ago

browser is configured too "strictly"

Try allowing github to store cookies so that it can remember you.

mcuee commented 1 year ago

Just curious, where do I find the patched version of avrdude 6.3? The link that mentioned it could not find the page.

Have you tried to build from the following source? https://github.com/mamut-tme/UsbAsp-flash_89S52/tree/master/avrdude-6.3_patch-src

orbanp commented 1 year ago

When trying to compile the patched avrdude 6.3,  ./configure complains about the lack of the "libhid" library.Unfortunately the sw repositories for ubuntu do not have that library.I did find a libhid library on github from Branson Fosdick, but it is not clear how to make it available for avrdude 6.3.Also looked at the avrdude.conf file for the patched 6.3 version, it does not even has a part description for the AT89Sxx devices. Thanks, Peter

mcuee commented 1 year ago

I see, Sorry I an not help you much here as I do not have the AT89S51/52 target myself. You may want to ask in other places.

mcuee commented 1 year ago

@orbanp

But I have ordered AT89S52 chip and it should arive roughly in a month . I have USBASP/USBISP so I can test this once it is here.

llinjupt commented 1 year ago

@orbanp @mcuee After glancing over those souce codes of USBasp firmware which claimed to support AT89S5x, I realize that none of them could support avrdude natively.

Spending some time doing hacking, I found it's easy to make the official USBasp version to support avrdude:

$ ./avrdude.exe -P usb -c usbasp -p AT89S52 -U flash:w:AT89S52.bin:r

avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e5206 (probably 89S52)
avrdude: Note: flash memory has been specified, an erase cycle will be performed.
         To disable this feature, specify the -D option.
avrdude: erasing chip
avrdude: reading input file AT89S52.bin for flash
         with 8192 bytes in 1 section within [0, 0x1fff]
avrdude: writing 8192 bytes flash ...

Writing | ################################################## | 100% 9.76s

avrdude: 8192 bytes of flash written
avrdude: verifying flash memory against AT89S52.bin

Reading | ################################################## | 100% 3.19s

avrdude: 8192 bytes of flash verified

avrdude done.  Thank you.

Keep in mind that you only need to do 2 things:

  1. Rst polarity of AT89S5x is opposite of AVR.
  2. probe AT89S5x, the fourth byte will get 0x69

Since I didn't hack it on a AVR MCU, so I paste alike hacked code below for you to refer:

uchar ispEnterProgrammingMode() {
    uchar check;
    uchar count = 32;

       ISP_OUT |= (1 << ISP_RST); 
    while (count--) {
        ispTransmit(0xAC);
        ispTransmit(0x53);
        ispTransmit(0);
        check = ispTransmit(0);

        if (check == 0x69) {
            return 0;
        }

        spiHWdisable();

        /* pulse RST */
        ispDelay();
        ISP_OUT &= ~(1 << ISP_RST); /* RST low */
        ispDelay();
        ISP_OUT |= (1 << ISP_RST); /* RST high */
                ispDelay();

        if (ispTransmit == ispTransmit_hw) {
            spiHWenable();
        }
    }

    return 1; /* error: device dosn't answer */
}
mcuee commented 1 year ago

@llinjupt

Thanks a lot for looking at usbasp firmware. This is much better for avrdude project. It is always better to support open firmware and not some unknown firmware.

I will try your modification once I get the AT89S52 chip.

@dioannidis Just wonder if you are willing to add the support in your usbasp firmware. You need to update the following code as per the suggestion from @llinjupt https://github.com/dioannidis/usbasp/blob/master/firmware/isp.c

dioannidis commented 1 year ago

Hi,

@mcuee , @llinjupt

Just wonder if you are willing to add the support in your usbasp firmware. You need to update the following code as per the suggestion from @llinjupt https://github.com/dioannidis/usbasp/blob/master/firmware/isp.c

OK I'll order those parts and after testing I'll add it

mcuee commented 1 year ago

Keep in mind that you only need to do 2 things:

Rst polarity of AT89S5x is opposite of AVR. probe AT89S5x, the fourth byte will get 0x69

@llinjupt and @stefanrueger

In this case, it is possible to just change avrdude (to change usbasp.c and avrdude.conf.in, create a variant of usbasp, say, usbasp_at89s) without changing the usbasp open source firmware?

I have also re-opened this issue to see if there is a way to support these two AT89S chips with other programmers, at least for a few more popular ones. I am guessing STK500v2 based programmers may already work. I am not so sure about USBtinyISP.

Edit: adding a reference here: reset pin polarity handling can be done by HW as well. https://www.avrfreaks.net/s/topic/a5C3l000000UKhJEAW/t093355?page=2

llinjupt commented 1 year ago

@mcuee IMHO, if the programmer doesn't support setting reset polarity, that's impossible to do so. And there are more other strange chips like S2051/AT89S8252/AT89S8253 that behave pretty differently, see below probe function which comes from one variant of USBasp.

uchar ispEnterProgrammingMode() {
    uchar check, check89 = 0;               //.kbv
    uchar count = 4;
    int i;
    // this is crap.   We need to do a proper RESET pulse before the start sequence.
    // we never want to insert false SCK pulses.
    for (i = 0; i < 256; i++) AT89buf[i] = 0xFF;
    while (count--) {
        reset_active_high = 0;          // all AVRs are active-low
        SS_inp();
        AT89mode = 0;
        spimode_1 = 0;
        ispReset();             // .kbv put into ISP mode for another go
        ispTransmit(0xAC);
        ispTransmit(0x53);
        check = ispTransmit(0xAA);
        check89 = ispTransmit(0x33);        // NOP on a S8252
        if (check == 0x53 && check89 == 0xAA) { // AVR
            return 0;
        }
        reset_active_high = 1;          // AT89 are active-high
        SS_out();
        if (((PINC & (1<<PC2)) == 0) || (eeprom_read_byte(&AT89reset_low) == 0x55))
            reset_active_high = 0;      // SLOW jumper overrules
        AT89mode = 0;
        spimode_1 = 0;
        ispReset();             // .kbv put into ISP mode for another go
        SS_low();
        ispTransmit(0xAC);
        ispTransmit(0x53);
        check = ispTransmit(0xAA);
        check89 = ispTransmit(0x33);
        if (check == 0x53 && check89 == 0xAA) {         // AVR or S2051
            if (reset_active_high == 0) {
                SS_inp();
                return 0;       // AVR
            }
            AT89mode = 1;           // S2051
            AT89page_size = 32;
            return 0;
        }
        if (check89 == 0x69) {
            AT89mode = 4;           // AT89S52
            AT89page_size = 256;
            return 0;
        }
        check = ispTransmit(0x55);
        check89 = ispTransmit(0);
        if (check == 0x33 && check89 == 0x55) { // AT89S8252
            AT89mode = 2;
            AT89page_size = 0;
            return 0;
        }

        spimode_1 = 1;              // only 8253 uses SPI mode #1
        ispReset();             // .kbv put into ISP mode for another go
        SS_low();
        ispTransmit(0xAC);
        ispTransmit(0x53);
        check = ispTransmit(0xAA);
        check89 = ispTransmit(0);
        if (check == 0x53 && check89 == 0xAA) { // AT89S8253
            AT89mode = 5;
            AT89page_size = 64;
            return 0;
        }

        spimode_1 = 0;
        ispReset();             // .kbv put into ISP mode for another go
        SS_low();
        ispTransmit(0xAA);
        ispTransmit(0xAC);
        ispTransmit(0x53);
        check = ispTransmit(0);
        check89 = ispTransmit(0);
        SS_high();
        if (check == 0x53) {
            AT89mode = 3;           // AT89LP4052
            AT89page_size = 32;
            return 0;
        }

        spimode_1 = 0;
        ispReset();             // .kbv put into ISP mode for another go
        SS_low();
        ispTransmit(0xAA);
        ispTransmit(0x55);
        ispTransmit(0xAC);
        ispTransmit(0x53);
        check = ispTransmit(0);
        check89 = ispTransmit(0);
        SS_high();
        if (check == 0x53) {
            AT89mode = 6;           // AT89LP51RD2, AT89LP6440 etc
            AT89page_size = 64;
            return 0;
        }
    }

    SS_high();
    SS_inp();
    reset_active_high = 0; // all AVRs are active-low
    return 1; /* error: device dosn't answer */
}