More on the whole interruptless networking

[Sean Fagan]

I've made a series of diffs.  I've only tested them in vmware fusion for now, so that's if_lem that gets involved.  I need to track down a system with igb and try that; I'm much less sure of those changes than I am the if_em / if_lem changes.

Note that, for various reasons, I used git for my source control.  One of those reasons was "I really need to get some experience with git, so this seems like a good excuse."  So pardon me if I produced them incorrectly.

I'm attaching two of the sets for perusal; the others are over on one of my machines.

First up:  netdump-10-diffs.txt, attached to this message.

This gets the old netdump branch compiling in 10.0.

Second, also attached to this message, kernel-netdump-diffs.txt, which then takes that, and splits out some of the code, so that I could use it in the real purpose for doing all of this.

Third, which is at http://earth.kithrup.com/~sef/netdump/kernel-kdp-diffs.txt, is an in-progress port (hence all the output it creates) of the KDP protocol Apple uses for kernel debugging over ethernet.  It's not entirely functional yet, but it's hard to tell how much of that is this side, and how much is the gdb side.

Which leads me to fourth, at http://earth.kithrup.com/~sef/netdump/gdb-diffs.txt, which are changes to gdb (also in-progress, hence it can be very verbose) to attach to the KDP code in the third patch.

Despite my splitting them out like that, I haven't tried doing them separately.  See earlier comment about git for why that is; if anyone really wants, I can poke around and try to get the diffs from each stage.  Obviously, however, they really are all fairly interconnected.

Sean.