Date: Tue, 29 Jan 2013 01:45:18 +0400 From: Chagin Dmitry <chagin.dmitry@gmail.com> To: John Baldwin <jhb@freebsd.org> Cc: emulation@freebsd.org Subject: Re: [PATCH] Move MI bits out of MD linux.h Message-ID: <20130128214518.GA5802@dchagin.static.corbina.net> In-Reply-To: <201301281353.36994.jhb@freebsd.org> References: <201301280947.39690.jhb@freebsd.org> <20130128174944.GA4105@dchagin.static.corbina.net> <201301281353.36994.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--OgqxwSJOaUobr8KG Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, Jan 28, 2013 at 01:53:36PM -0500, John Baldwin wrote: > On Monday, January 28, 2013 12:49:44 pm Chagin Dmitry wrote: > > On Mon, Jan 28, 2013 at 09:47:39AM -0500, John Baldwin wrote: > > > This patch attempts to reduce duplication between i386/linux/linux.h = and=20 > > > amd64/linux32/linux.h by moving bits that are MI out into headers in= =20 > > > compat/linux. I only moved constants that were defined in MI headers= in Linux=20 > > > itself, not necessarily all constants that have the same value. Furt= her=20 > > > reduction could be possible if we wanted to make a linux-i386.h that = these two=20 > > > headers could share, but that would be a separate project. > > >=20 > > > http://www.FreeBSD.org/~jhb/patches/linux_mi.patch > > >=20 > >=20 > > Hi, John! > > Looks ok to me, furthemore I have similar patches in the local x86_64 l= inuxulator > > repo. So, commit, please. >=20 > So do you have an open PR with patches for the 64-bit linux stuff? I'd > like to look at what you have and see about getting it into the tree if > so. I'm curious about how you handled things (e.g. do you recompile > things like linux_file.c twice once as 32-bit once as native, and do you > have linux32.ko for the existing module and linux.ko for 64-bit)? >=20 Plan - switch linuxulator to the use of the native 1:1 threads first. I'll send patches to review soon. need to write a brief explanation of the new linuxulator. Second, commit vdso, which is used on x86_64 linux by default (now trap at known address used to enter syscall()). Yes, on amd64 thow modules of linuxulator: linux64.ko && linux.ko as requested by kib@. --=20 Have fun! chd --OgqxwSJOaUobr8KG Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (FreeBSD) iEYEARECAAYFAlEG8W4ACgkQ0t2Tb3OO/O2hSACgk6P9HcM09ynmii9M6RVSUHih AGUAnj+pg7gvknuG66niRyzn0ghdmfS0 =pbqe -----END PGP SIGNATURE----- --OgqxwSJOaUobr8KG--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130128214518.GA5802>