I think the operative word in that sentence was "reasonable".
In the case where the changes to the interface are reasonable, and the driver
change is necessary, then modifying the affected files will save time for the
maintainers - they don't each need to immediately go and investigate the
implications of the changes, because it's been done for them by the person who
understands the 'offending' interface completely.
I've don't recall anyone ever complaining about a change which they thought
was reasonable. The point here is that the change wasn't necessary, and made
the code more difficult to maintain. The maintainer is quite justified in
complaining, in that case.
In the case of a 'reasonable' change, then I'd much rather let the person who
changed the interface go and update the drivers accordingly. I've just taken
about a month to work out that I needed four one-line patches in arcnet.c to
use (un)register_netdevice instead of (un)register_netdev, because of new
locking semantics. Believe me, I wouldn't have objected if the person
responsible for that change had done it for me. Neither would the users who
were without arcnet for a month while they were waiting for me to finish my
exams.
---- ---- ----
David Woodhouse, Robinson College, CB3 9AN, England. (+44) 0976 658355
Dave@imladris.demon.co.uk http://www.imladris.demon.co.uk
finger pgp@dwmw2.robinson.cam.ac.uk for PGP key.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu