Re: [Etherboot-developers] PXEification testing
Brought to you by:
marty_connor,
stefanhajnoczi
From: Timothy L. <tl...@ro...> - 2004-05-30 01:26:51
|
On Sat, 2004-05-29 at 21:35, Marty Connor wrote: > At this rate, it'll take quite a while to get a release out, and I'm > not sure it's worth it, since there are a lot of other good fixes that > will benefit people as it stands. Marty I would say lets doe a release without the pxefication No one is beating down the door for drivers that work with the undi driver. > I'm seriously thinking for now of doing the quick and dirty thing that > Michael mentions in README.pixify: [snip] > This way, we get the drivers done, we get out 5.3.8, and we can start > testing. > Then we can update drivers the right way. for 5.3.9+. > > Which brings me to the question of testing. Adding it would add most of the support stuff that should be added anyway. > Has the function: > > nontrivial_irq_debug ( undi.irq ); > > been committed to CVS? I see it mentioned in an email message, but > can't find it in the source. Michael or Timothy, could commit it? I just checked it in. You will need to add a call at the beginning of undi_poll(), as nontrivial_irq_debug ( undi.irq ); > Also, is this still the simplest testing method: > On May 10, 2004, at 8:05 PM, Michael Brown wrote: > Probably the easiest way to test out a pixified driver is to boot > that > driver (e.g. sundance) from a floppy disk, then have that download > undi.zpxe with -DPXELOADER_KEEP_ALL. This won't test the UNDI > loader > portion of the code, but that part isn't driver-specific anyway. > > (I assume this means to compile an undi.zpxe image with > -DPXELOADER_KEEP_ALL defined). Yes, it is a lot easier than the alternative of burning a rom, testing, making changes ... Tim |