Re: git: 9fd0d9b16e93 - main - ktls: Remove the KERN_TLS option from the i386 and amd64 LINT-NOIP kernel configurations.
Date: Mon, 13 Jun 2022 14:28:59 UTC
On 6/13/22 16:22, John Baldwin wrote: > On 6/12/22 4:43 AM, Hans Petter Selasky wrote: >> On 6/11/22 22:04, Bjoern A. Zeeb wrote: >>> On Sat, 11 Jun 2022, Hans Petter Selasky wrote: >>> >>>> The branch main has been updated by hselasky: >>>> >>>> URL: >>>> https://cgit.FreeBSD.org/src/commit/?id=9fd0d9b16e93ff2a3bd375a98763dca0150dcee0 >>>> >>>> >>>> >>>> commit 9fd0d9b16e93ff2a3bd375a98763dca0150dcee0 >>>> Author: Hans Petter Selasky <hselasky@FreeBSD.org> >>>> AuthorDate: 2022-06-11 19:29:42 +0000 >>>> Commit: Hans Petter Selasky <hselasky@FreeBSD.org> >>>> CommitDate: 2022-06-11 19:31:28 +0000 >>>> >>>> ktls: Remove the KERN_TLS option from the i386 and amd64 LINT-NOIP >>>> kernel configurations. >>>> >>>> Kernel TLS depends on INET or INET6 being enabled. >>> >>> That is only a short-term solution hiding the actual problem and will >>> now propably never be fixed again like these device drivers which were >>> on that list ofr ages... >> >> +1 >> >> There is a plan to replace config at some point. I believe this is a >> limitation of sys/conf/files, that it doesn't provide a good way to >> handle conflicts. There are many ifdef KERN_TLS, so just skipping one >> file for building won't help. >> >> https://hackmd.io/w1Tf8mmVQVuZok7-LNEhgw > > What we often do though is just add in #ifdef's to make a given file a > kind of stub. In theory KTLS could work over UNIX domain sockets if > someone wanted to do the work, and that would just entail adding the > right #ifdef INET, etc. in this file. > Is this something you plan to work on, or something you want done? Feel free to revert my bandaid once a better solution is in place. --HPS