well, I've got this sun ultra 40, nice box, sun-designed mobo, lots of clever stuff.
And, for some reason, they did not put on a serial port.
Thus proving that it is possible to be too clever. Oh well, now the fun begins. I guess it will get worse from here on out.
Please, mobo vendors, just put on the serial port. USB is not an option.
ron
...and yet they somehow forgot to mention "Needs a serial port" on their "Perspectives" page ( http://www.sun.com/desktop/workstation/ultra40/perspectives.jsp ).
On 7/20/06, Ronald G Minnich rminnich@lanl.gov wrote:
well, I've got this sun ultra 40, nice box, sun-designed mobo, lots of clever stuff.
And, for some reason, they did not put on a serial port.
Thus proving that it is possible to be too clever. Oh well, now the fun begins. I guess it will get worse from here on out.
Please, mobo vendors, just put on the serial port. USB is not an option.
ron
-- linuxbios mailing list linuxbios@linuxbios.org http://www.openbios.org/mailman/listinfo/linuxbios
Ronald G Minnich wrote:
well, I've got this sun ultra 40, nice box, sun-designed mobo, lots of clever stuff.
And, for some reason, they did not put on a serial port.
Thus proving that it is possible to be too clever. Oh well, now the fun begins. I guess it will get worse from here on out.
Please, mobo vendors, just put on the serial port. USB is not an option.
ron
The legacy of PC serial, parallel, IDE, floppy and PCI is fading away quickly. They were fine in their day, but slow by todays standards. They also used lots of board space since they were so wide.
As painful as it may be to let go. We just need to get USB support up for LinuxBIOS debug.
-Bari
On 7/20/06, Bari Ari bari@onelabs.com wrote:
As painful as it may be to let go. We just need to get USB support up for LinuxBIOS debug.
ok, where do I start? I know you discussed this earlier but ... where do I start pulling on the string to make usb debug work? I guess we have to solve this. I'm just disappointed to have to be solving it just now :-)
I still think if they'd just put a serial header on these things, it would sure help. There's always a superio and it's really only 3 little wires.
ron
Another comment on serial that I recalled last night. On any pc-compatible machine, from power on reset, a simple outb('0', 0x3f8) will get you output. Having to put a whole usb stack in there and get it working, for the same effect -- UGH!
:-)
ron
* ron minnich rminnich@gmail.com [060721 17:07]:
Another comment on serial that I recalled last night. On any pc-compatible machine, from power on reset, a simple outb('0', 0x3f8) will get you output. Having to put a whole usb stack in there and get it working, for the same effect -- UGH!
Which is not really true in all cases. The SuperIO chip might very well need initialization.
What scares me more than the USB stack (because I believe this is doable in less than 100 lines for this very purpose of just using the _debug_ mode of EHCI) is that we need to see all PCI devices. This is pretty much the same problem that keeps us from just plugging in a 10$/€ PCI Serial card: Slot's behind the bridge and we might not see it early enough.
Stefan Reinauer wrote:
- ron minnich rminnich@gmail.com [060721 17:07]:
Another comment on serial that I recalled last night. On any pc-compatible machine, from power on reset, a simple outb('0', 0x3f8) will get you output. Having to put a whole usb stack in there and get it working, for the same effect -- UGH!
Which is not really true in all cases. The SuperIO chip might very well need initialization.
very true ...
What scares me more than the USB stack (because I believe this is doable in less than 100 lines for this very purpose of just using the _debug_ mode of EHCI)
that's encouraging :-)
I guess we just gotta write it.
ron
* Ronald G Minnich rminnich@lanl.gov [060721 18:08]:
What scares me more than the USB stack (because I believe this is doable in less than 100 lines for this very purpose of just using the _debug_ mode of EHCI)
that's encouraging :-)
I guess we just gotta write it.
Ugly is: The only cable that works we know of is not available anymore. So if we start to solder something we dont really know which part is wrong, the code or the cable.. But some people on this list might know an answer to this one.
Stefan
Stefan Reinauer wrote:
- Ronald G Minnich rminnich@lanl.gov [060721 18:08]:
What scares me more than the USB stack (because I believe this is doable in less than 100 lines for this very purpose of just using the _debug_ mode of EHCI)
that's encouraging :-)
I guess we just gotta write it.
Ugly is: The only cable that works we know of is not available anymore. So if we start to solder something we dont really know which part is wrong, the code or the cable.. But some people on this list might know an answer to this one.
Stefan
I can test on OLPC. We're booting into ash out of linuxbios into a minimal environment. We can build a linux with no usb and test simple programs that drive it.
maybe.
ron
-
Bari Ari -
David Hendricks -
ron minnich -
Ronald G Minnich -
Stefan Reinauer