Re: Regarding the ULTRIX and OSF1 compats
Jason Thorpe <[email protected]>
| Newsgroups | gmane.os.netbsd.devel.kernel,gmane.os.netbsd.ports.pmax,gmane.os.netbsd.ports.alpha |
|---|---|
| Message-ID | <[email protected]> |
> On Mar 16, 2019, at 9:09 AM, Maxime Villard <[email protected]> wrote: > Anyway, people, let's wrap it up. The status right now is that I've disabled > COMPAT_OSF1 on Alpha, and I'm not sure if there is an improper dependency of > COMPAT_LINUX on COMPAT_OSF1, because I see calls like osf1_sys_wait4 in > sys/compat/linux/arch/alpha/, but there is no dependency in the modules. > > Right now I'm waiting to see if the build cluster fails, to know for sure. > If there is indeed a dependency, we'll have to untangle it before retiring > COMPAT_OSF1. I don't think it is really complicated, the functions used are > just wrappers, that we can easily move into Alpha's COMPAT_LINUX. Ah, yes, Linux... When Linux was brought up on Alpha, they tried to make the system call ABI compatible with OSF/1's ... that's just the way Linux was done back in those days... I don't know if it's still the case, and I don't honestly know how much utility COMPAT_LINUX has on Alpha, either. Linux is a "high-touch" emulation ... stuff that should be platform-independent "leaks" out into arch-dependent code in Linux, so making it work is a lot more than just doing the exec glue and signal trampolines -- sometimes you have to have completely alternate implementations of some system calls. -- thorpej