Date: Wed, 27 Oct 2010 23:55:23 +0200 From: Attilio Rao <attilio@freebsd.org> To: John Baldwin <jhb@freebsd.org> Cc: freebsd-arch@freebsd.org Subject: Re: [PATCH] Headers for the x86 subtree Message-ID: <AANLkTinuQHVycJ4nu2yazVSX5NOHumvPNJods9_3F6rv@mail.gmail.com> In-Reply-To: <201010271749.54619.jhb@freebsd.org> References: <AANLkTiktoYyxmE8nyGeoc4_ov35fR7iN83444MfhYg-e@mail.gmail.com> <201010271649.47880.jhb@freebsd.org> <AANLkTikwT9%2BwQycEHTGapw7j9TeuiCaPet6UF4=MPMSg@mail.gmail.com> <201010271749.54619.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
2010/10/27 John Baldwin <jhb@freebsd.org>: > On Wednesday, October 27, 2010 5:25:20 pm Attilio Rao wrote: >> 2010/10/27 John Baldwin <jhb@freebsd.org>: >> > On Wednesday, October 27, 2010 4:00:15 pm Attilio Rao wrote: >> >> 2010/10/27 John Baldwin <jhb@freebsd.org>: >> >> > On Wednesday, October 27, 2010 10:56:06 am Attilio Rao wrote: >> >> >> This patch should convert a (simple and 100% shared between amd64 = and >> >> >> i386 header) under the x86 sub-tree. Please note that in this patc= h I >> >> >> "svn cp" the file from sys/amd64/include/mptable.h into >> >> >> sys/x86/include/mptable.h: >> >> >> http://www.freebsd.org/~attilio/headers-x86.diff >> >> >> >> >> >> This is someway a POC, that I really want to get in. The idea is >> >> >> simple and someway follows the pc98 case (even if not entirely): t= he >> >> >> files under machine/include/* became just mere stubs for x86/inclu= de/* >> >> >> contents and redirect there. >> >> >> This won't particulary help reducing the number of available files= , >> >> >> but generally removing verbatim and would also be the way to go fo= r >> >> >> handling MFCs. >> >> >> If you find this is the right way I'll commit the fix and start mo= ving >> >> >> other files as time permits. >> >> > >> >> > No, we want to do this differently because we also want this to wor= k in >> >> > userland. =C2=A0(e.g. I'd like to outright move mca.h to x86/includ= e and then use >> >> > '#include <x86/mca.h>' in both kernel and userland for it). =C2=A0W= e'd need some >> >> > special glue to setup an 'x86' symlink during a kernel build that p= oints to >> >> > @/x86/include as we do now to setup an 'i386' link for pc98 kernels= . >> >> > >> >> > We'd also need to install the x86 headers into /usr/include during = an >> >> > installworld. =C2=A0Warner has some more pointers on this I think. >> >> >> >> So you probabilly are suggesting to go w/ the "pc98 approach". >> >> I'm fine with it, I'll try to look for how it works and implement as = well. >> > >> > Thanks. =C2=A0I think it is fine to use '#include <x86/foo.h>' in code= directly >> > with this approach as well. =C2=A0I only think we should provide wrapp= ers in >> > /usr/include/machine if compatibility is needed. =C2=A0mca.h and mptab= le.h >> > shouldn't need compatibility for example, but specialreg.h might. >> >> I'd expect to potentially mirror any header, otherwise on which >> criterias we can say some are needed and other not? > > If it is widely used in userland. :) =C2=A0mptable.h is probably only use= d by the > mptable utility which is easily fixed. =C2=A0It will not be of interest t= o any 3rd > party software. =C2=A0mca.h may eventually be used by third party softwar= e, but it > is very new, so we can probably move it now without causing any harm. > > Kernel-only headers such as intr_machdep.h certainly do not need any comp= at > wrappers. > > We can always add a compat header in machine if we find it is needed, but= I'd > like to minimize the amount of compat cruft as much as possible. I'd say in general I agree, so let's add them just on-demand is possibly fi= ne. Thanks, Attilio --=20 Peace can only be achieved by understanding - A. Einstein
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTinuQHVycJ4nu2yazVSX5NOHumvPNJods9_3F6rv>