write data command

[deiKruve]

Hallo, can I zero a block of RAM with the 'wd' command? I am short of memory so I would dearly like to see if I overwrite my tail with the stack.

Cheers j.

[dcwbrown]

It's not as easy as a purpose built fill command, but you can provide up to 16 bytes at a time to wd, so you can zero e.g. the 16 bytes at 0x90 with the command:

wd90 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

How much do you need to fill, and is 16 bytes at a time workable for you?

(If not, I should consider adding a set of fill commands.)

-- Dave.

[deiKruve]

Hi, my 2 cents is that 'wd' must stay as it is.

fil

Would be excellent, but I can live without it, for the moment. If I fill -say- 3 x 16 bytes that should tell me whether I have enough stackspace left (out of 256 bytes).

I have been looking at the programming issue a bit closer: I never need to write more than twice to program all locations properly. Then I have increased the bulk capacitor on the supply to 10uF ceramic and I put a big screencable-type ferrite around the connections to the fdti. And I did look at your code. It all made no difference at all. In fact I have never noticed any mangled data on the connection. So I guess it is the chip that does not program quite as good at 3.3 volt supply.

Cheers,

j.

On Fri, Sep 15, 2017 at 1:58 PM, David C W Brown notifications@github.com wrote:

It's not as easy as a purpose built fill command, but you can provide up to 16 bytes at a time to wd, so you can zero e.g. the 16 bytes at 0x90 with the command:

wd90 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

How much do you need to fill, and is 16 bytes at a time workable for you?

(If not, I should consider adding a set of fill commands.)

-- Dave.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/dcwbrown/dwire-debug/issues/26#issuecomment-329761436, or mute the thread https://github.com/notifications/unsubscribe-auth/AEaMoMA0Pto0lozqWwSRX7yFdhISs_szks5simZagaJpZM4PX4gg .

[deiKruve]

hallo,

1. I put up a cable for avrdude / dwire mixed use on github.

2. I noticed just now that programming a virgin attiny441 with dwire went faultless in 1 cycle. But the second time over I had to program twice again as before. Could this issue have to do with the way you erase the locations to be reprogrammed? Or does it get erased perhaps but not reprogrammed the first time around?

Cheers, enjoy wat's left of the day.

j.

On Fri, Sep 15, 2017 at 5:11 PM, Jan de Kruyf jan.de.kruyf@gmail.com wrote:

Hi, my 2 cents is that 'wd' must stay as it is.

fil

Would be excellent, but I can live without it, for the moment. If I fill -say- 3 x 16 bytes that should tell me whether I have enough stackspace left (out of 256 bytes).

I have been looking at the programming issue a bit closer: I never need to write more than twice to program all locations properly. Then I have increased the bulk capacitor on the supply to 10uF ceramic and I put a big screencable-type ferrite around the connections to the fdti. And I did look at your code. It all made no difference at all. In fact I have never noticed any mangled data on the connection. So I guess it is the chip that does not program quite as good at 3.3 volt supply.

Cheers,

j.

On Fri, Sep 15, 2017 at 1:58 PM, David C W Brown <notifications@github.com

wrote:

It's not as easy as a purpose built fill command, but you can provide up to 16 bytes at a time to wd, so you can zero e.g. the 16 bytes at 0x90 with the command:

wd90 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

How much do you need to fill, and is 16 bytes at a time workable for you?

(If not, I should consider adding a set of fill commands.)

-- Dave.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/dcwbrown/dwire-debug/issues/26#issuecomment-329761436, or mute the thread https://github.com/notifications/unsubscribe-auth/AEaMoMA0Pto0lozqWwSRX7yFdhISs_szks5simZagaJpZM4PX4gg .