MS serial data block size

[elaw]

Phil,

James told me in a response to a post I made in the MS3 development forum that there's a hardcoded 512-byte limit in TS on the size of the data block that can be read from the Megasquirt for runtime variables (http://www.msextra.com/forums/viewtopic ... 25&t=59482).

I'm going to be doing some custom mods to the MS3 code and it would be very helpful to have that limit increased. Is that hard to do?

[LT401Vette]

I assume you are referring to the SD read command addressing for MS3?
It seems the command had a hard address in that is now in the way of the block size. I guess they didn't figure he would need that much space in the future. In TS they will need to be externalized so the command can be overridden for firmwares that need it.

It is on the todo list for the next major release which will be going beta shortly.

[elaw]

Phil,

Sorry for the delayed response... I never got a notification of your reply!

I'm not 100% clear on the terminology and maybe I'm confused, but I thought the limit was on the size of the block defined under "OutputChannels" in the .ini, which I understand defines the values that can be logged in TS?

If you look at "core.ini" in the source for MS3 v1.4.0, at line 14107 there's a comment that says "very nearly out of space. 512+ requires firmware changes". That of course implies the problem is on the MS side, but something James posted implied there was a limit in TS so maybe it's both?

[LT401Vette]

If you look at "core.ini" in the source for MS3 v1.4.0, at line 14107 there's a comment that says "very nearly out of space. 512+ requires firmware changes". That of course implies the problem is on the MS side, but something James posted implied there was a limit in TS so maybe it's both?

Yes, sort of a circular dependency.

TS doesn't care how big the Outpc read is....

The trouble is that the command to set RTC on the MS3 includes a hard address just above the outpc. It is not really a command, but a write instruction to an address. This command is compiled into TS, so James can not change the address on the firmware until TS externalizes the command in some way to change the address in both the firmware and TS's command together.

Ultimately I need to get the next major release to beta status.

[elaw]

Ah, okay. I should have known if it was something simple it would have been taken care of already!