In article <46C9DB48.9EE972C@spam.comcast.net>, David J Dachtera writes:
>
> Did you have the source and the build environment to build a new (unsupported)
> VMS kernel, or did you just include kernel-mode code in a routine that could be
> LINKed to your driver code, bypassing the kernel code you were replacing?


For drivers I've either written my own routine as part of the driver,
or added code to my driver so that I could use the faulty kernel code
as was.

I have never built a VMS kernel. I've added kernel code and I've
patched kernel code (on VAXen, using PATCH on the .EXE or console
deposit commands for temporary edits, and on Alpha using kernel mode
code instead of deposit commands) to change behaviour.

Fortunately I've only had to change writeable kernel data cells on
Alpha. PATCHing kernel .EXE on VAXen to change the instruction
stream was more fun.