Teleconference information for software status meeting tonight.
smithbone at gmail.com
Tue Oct 10 18:59:19 EDT 2006
> can't make the meeting, please at least update the status on action
> items. Let's update our bugs even if the status is "no progress this
I'm not going to be able to make the meeting tonight. I _might_ be
able to catch the tail end depending on how long it goes..
> - follow up on trac 113. jg has memory that full reset is possible.
I seem to remember some one else mentioning that as well.
> olpcflash verify fails with "SST" flash part
Bricked one board so far 3 to go. :)
The sst part is more different than I orginally realized.
I can now erase sectors. (thus how I created the brick)
There is no page write command on this part. Instead you put it into
auto inc mode and you send it words at a time it programs and then
increments the address. I can still do this a "page" at a time but
the mechanics inside the write routine are totally different.
So I have to have 2 seperate write routines. I've got the new write
routine done and the framework to select it in place but while testing
early this morning I found that:
1) The ID data does not seem to pop out fo the chip in the order the
datasheet says it should and
2) The spansion part dosen't support the JDEC ID command. So I need
to rework the part ID code to use the alternate RDID command which is
sort of mfg specific.
Then finish up the testing of the new SST write method.
I'm very interested in the progress Mitch makes on the SPI recovery
tool for BTest boards.
> PCI Pin A needs to be routed to a real interrupt
I'm looking at this. Tsylla said that interrupt routeing is not in
the LinuxBIOS code but it looks like its there to me. Ron also seems
to think whats there should work if I poke in the right value.
I was planning on looking at this deeper later on tonight.
> e820 map incorrectly excludes most of range 0xa0000-100000
> a-f segments cause errors in memtest, and fixed region configs are
> possibly not set right.
No change from last week. I've got the code in my private tree I've
just not gotten around to testing it.
Richard A. Smith
More information about the Devel