[RFT] Please test patch set intel/i945: Only write CID, PN and TCID once

[Paul Menzel]

Dear coreboot folks,

Mono pushed the patch

intel/i945: Only write CID, PN and TCID once

for review [1], which needs testing to make sure nothing breaks.

Could you please test it on Lenovo T60, Lenovo X60, the Kontron
986LCD-M/mITX and other Intel 945 based devices?

Please attach output of `inteltool` (under `util/inteltool`) without and
with the patch *and* please also attach the coreboot log with patch set
5393 [2] applied.

That would be very helpful.

Thanks,

Paul

[1] http://review.coreboot.org/5387/
[2] http://review.coreboot.org/5393/

[ron minnich]

So do I understand it correctly, that this patch is for hardware that has
been working, in some cases for 5 years; that it has
been written after reading a public doc; and that it was not developed in
response to a known bug or issue?

Code that does not conform to a public doc is NOT a bug.

A word to the wise: in all too many cases, the public docs and the private
docs disagree. In some cases, the private docs directly
contradict the public docs. And, in still other cases, the hardware
contradicts the private docs. That's what makes this so much fun.

It's the implementation of the hardware, not the documentation, that
matters.

Changes made in response to a reading of public docs, that are not known to
resolve an actual observed problem, are
a very risky thing to do, especially with old chipsets. They should be
rejected outright unless they are confirmed to fix a problem.

We've had people push this kind of change in the past and break boards.

So tread carefully.

ron