New board ID moving into production

[wad]

Due to minor board changes (but still, they are changes), the motherboard ID will be incremented.

This change will be delayed until the necessary software changes have been made, tested, and communicated to Quanta. Quanta would like to make the change ASAP.

The new ID (C3) will be (5/8) / 3VPCU =2.06V. R535 will stay as 10K. R534 will change to 6.04K.

The motherboard changes are mainly replacement of 0 ohm resistors with traces. A resistor was also added to improve the robustness of the external microphone input.

[rsmith]

Replying to wad:

This change will be delayed until the necessary software changes have been made, tested, and communicated to Quanta. Quanta would like to make the change ASAP.

The EC modifications for this are trivial. I will make up a new test firmware tomorrow and wad is going to mod a board with the new resistor. The new ID will be C3.

The _real_ question is whats going to break when the kernel/olpc-configure sees a board ID of C3. In the past audio stopped working cause it falls back on the older wiring scheme.

Kernel folk can you please take gander at the code that deals with board ID and see if a 'C3' is in whatever case statements are there?

[dsaxena]

Posting Andres' email message here:

I can't seem to log into trac right nowww, so:

The last time this came up, we replaced all the code that depended upon
specific model versions.  The kernel currently checks for model
features with things like:

if (olpc_board_at_least(0xb3))

So, things should Just Work (heh, famous last words...)

I also just did a quick sanity check of the places we're calling olpc_board_at_least(), everything looks OK.

[gregorio]

Milestone Never Assigned deleted

[rsmith]

This is in 709 and ready for QA