I have put together a webrev with fixes to address lint warnings generated thanks to MBLKL and company being lint-unclean. They are now lint clean. Additionally, I did related lint abd whitespace (cstyle) in a bunch of drivers. Some of the drivers I was even able to remove all of the LINTTAGS overrides, yielding truly lint clean source! (I could have done more of the drivers, but the effort to clean *all* of uts up is rather larger than I want to take on right now.) I also fixed the fact that one driver (amd8111s) was pointlessly including a dependency upon drv/ip. (I've verified that the link dependency is not necessary, both by manual examination of the symbol table, and by loading the generated object without the dependency on a system.)
The webrev is at http://cr.opensolaris.org/~gdamore/mblkl There is a closed version, which includes one closed file fix, at http://jurassic.sfbay/~gd78059/webrev/mblkl At this point, IMO, there is little reason for *any* new kernel code to require the use of either /*LINTED*/, or -erroff. Once I get these changes in, I'm going to ask the RTI advocates to start requiring really strong justification for use of either of these in kernel putbacks. (Its seems I'm the only person who really took the following comment in the typical module Makefiles to heart:) # # For now, disable these lint checks; maintainers should endeavor # to investigate and remove these for maximum lint coverage. # Please do not carry these forward to new Makefiles. # Now that the most annoying macros in strsun.h will be fixed, I'd really like other kernel module maintainers to clean up their LINT issues. Note that one of the other annoyances, which is dereferencing char * pointers such as mp->b_rptr, to some other data type triggers E_BAD_PTR_TYPE_ALIGN, is often easily resolved without special lint overrides by way of casting through void *. You'll see lots of evidence of that in some of my changes above. (Note that this assumes that one is certain that the alignment really is safe.) Thanks! -- Garrett _______________________________________________ opensolaris-code mailing list opensolaris-code@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
