human-readable dmesg

[Bill Hacker]

Sascha Wildner wrote:
Matthew Dillon schrieb:
    I like the idea of putting it in a kernel module so we can have
    the messages if we want them.  I'm just trying to avoid source
    duplication.
OK, but still we have to be careful with removing/changing pcidevs file
entries. They are not consistent across the different BSDs.
Sascha

And what *is* cross-BSD consistent, these days?

Seriously - given that devices are one area still 100% driven by 'outside 
world', eg: other-than OS developers [1], might it not be fertile ground for a 
jointly-maintained pan-BSD device dataset?

Could save some labor and improve rate and accuracy of update - even where what 
to *DO* with the information still goes by separate roads.

Bill

[1] Possible exception of Apple getting into custom silicon & firmware, as Sun 
may soon cease to be silicon-involved, IBM & HP have long-since become 
'specifiers' at best, commodity kit-builders more often - as with Dell.

What little 'bespoke' hardware remaons is of diminishing relevance, or even 
visibility, to the *BSD's - or even non-fee Linux.