On Sat, Jan 10, 2026 at 9:42 AM John Baldwin <[email protected]> wrote:
> On 1/9/26 21:28, Konstantin Belousov wrote: > > On Fri, Jan 09, 2026 at 04:35:22PM +0000, Olivier Certner wrote: > >> The branch main has been updated by olce: > >> > >> URL: > https://cgit.FreeBSD.org/src/commit/?id=67599eef01f5417715f673ec81fc1467917c7a10 > >> > >> commit 67599eef01f5417715f673ec81fc1467917c7a10 > >> Author: Olivier Certner <[email protected]> > >> AuthorDate: 2026-01-09 16:23:59 +0000 > >> Commit: Olivier Certner <[email protected]> > >> CommitDate: 2026-01-09 16:34:56 +0000 > >> > >> sys/x86/NOTES: Add vt_efifb > >> > >> Contrary to what is stated in commit f224591746bd ("Add ASMC_DEBUG > make > >> option"), the various NOTES files should list all available > options. > > No, the f224591746bd commit is correct. GENERIC + LINT should cover all > > options and reverse of them, as much as possible. In other words, LINT > > should be complementary to GENERIC, for most options. > > > > At least this is how bde@ described the intent of LINT, and I think > this is > > a right thing to follow. > > Hmmm, I've always understood it that options that change behavior (like, > say > HZ) should have non-default values in LINT to test the setting of > non-default > options, but that LINT itself should include "all the things" in terms of > coverage, so 'device' statements and 'options' that add in new code like > filesystems should always be in LINT. > That's what I always try to do as well... In addition to documenting what often is an under documented part of the tree. Warner
