On Fri, Dec 10, 2004 at 12:27:32AM -0600, Jonathan Berry wrote: > On Thu, 9 Dec 2004 18:28:12 -0800, Nifty Hat Mitch wrote: > > Those with problems powering off cleanly need > > to be very clear... > Hey Tom, > > Glad to hear that work is being done on this. Where is the best place > to post info concerning this? Bugzilla? Well bugzilla is the common good answer but you need to read the existing bug to make sure that you are adding new info and not adding noise to the issue. I see that there is a lot of confusion in most of the bug postings. For example people are confusing acpi and apic. It makes sense to 'reherse' bugzilla additions on a list like this. Then after some comments and reflection post to bugzill. > Is there FC specific code > that is changing here, or is there work being done upstream (or both)? It seems that the upstream and current folks are staying linked up on this. Upstream is where much of the change is going on. I this case 'acpi' folks at Intel have been doing some of the interesting work. The best site on ACPI is Intel's. Give it a look. Look fro the reference code for the interpreter and tools. http://www.intel.com/technology/iapc/acpi/ Intel has standards documents, examples and more ... > As this seems to only affect some hardware, perhaps some info about > your hardware setup (motherboard and cpu) should be included too. > How are you comparing the sources? Just downloading the .src.rpm's > and extracting them? I'd like to look at the code that has changed to > possibly see what is going on. Not that I'd probably understand it, > but it would be interesting : ). In my case I had an old and new kernel build area handy. That made it easy for me to do a diff of the two. If you are new to diffs try: vim -d version1 version2 You can download the old and new src rpm, and install them. You can also use cvs tools to compare files. This is one of the best reasons to attempt to build your own kernel. You will find that the README files are very good. One thing that I have noticed is the way reset and interrupts are handled. In one case (apic) power down is full and a reset does not restart the box. In the other (acpi) reset works but external interrupts do not wakeup the box as they once did (lan wakeup). I have not figured out how to do is track the interrupt masks/vectoring and the use of halt instructions through the kernel+acpi interpreter dance. Interrupts do involve the mother board chip sets... the acpi interpreter does get involved in vectoring interrupts. It is low level and so most easy debug tricks do not play. Later, Tom -- T o m M i t c h e l l spam unwanted email. SPAM, good eats, and a trademark of Hormel Foods.
