On 15 February 2015 at 19:34, David Holland <[email protected]> wrote: > sched_getparam() is hardly a main-line function. I could easily > imagine running compat binaries for years and never happening to use > one that calls it. So I think your reasoning is suspect.
Thats the zero test coverage issue. For Linux emulation we have some ok test coverage (out of tree for now) using rump. Rump currently has no freebsd syscall module, if it did could run some tests very easily. There is also the try to run a whole distribution tests, including running test suites. If someone wants to do this happy to help. > > The problem is that these bugs affect the i386 users, most of which don't > > use compat_freebsd. This compatibility layer comes at the cost of security, > > and is - as I/you said - very incomplete. > > By this same reasoning we should not have any binary compat code at > all. The only thing you've cited that's different about the freebsd > compat code is that it's out of date -- the proper response to that is > to update it. It's not like freebsd's syscall ABI is secret or > anything. Someone has to want to do it. The motivation has generally been to run a few programs and has stopped there. SmartOS has been doing full Linux emulation for commercial reasons recently, and FreeBSD seems to have revived their efforts, again with Linux, but there are generally more interesting and fun things to work on in NetBSD. You end up having to implement all the features of the other OS to implement it. I would be happy if it was moved out of generic. Justin
