Date: Fri, 21 Nov 2014 08:14:45 -0800 From: Kirk McKusick <mckusick@mckusick.com> To: Konstantin Belousov <kostikbel@gmail.com> Cc: FreeBSD Filesystems <freebsd-fs@freebsd.org> Subject: Re: RFC: patch to make d_fileno 64bits Message-ID: <201411211614.sALGEjRb082219@chez.mckusick.com> In-Reply-To: <20141121155754.GN17068@kib.kiev.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
> Date: Fri, 21 Nov 2014 17:57:54 +0200 > From: Konstantin Belousov <kostikbel@gmail.com> > To: Rick Macklem <rmacklem@uoguelph.ca> > Subject: Re: RFC: patch to make d_fileno 64bits > > I do not think we need to have in-kernel compatibility shims. > The work, big but relatively trivial, is to convert filesystems to > use the new ino_t, even if the on-disk structures still use 32bit > inode number. > > Really problematic part of this change is the usermode ABI breakage. > The struct dirent is only the start of the whole issue. ino_t is > embedded into more structures which are part of the contract, e.g. > struct stat. We have to provide new syscalls which accept or return > the affected structures. > > And then, there are libraries which embed ino_t into their ABI. > Immediate example is fts(3) in libc. Look at the FTSENT.fts_ino. Even > after the base system is fixed by properly providing the compat shims > and symbol versions for the affected libraries, we get the same problem > with the binaries not from base. > > Summary of the issue with ino_t is that it is not too hard to fix the > kernel, comparing with the ABI issues which must be solved in usermode. You are quite right that this is a big and complex process. It was first attempted as a Google Summer of Code project which was later (in August 2011) integrated in projects/ino64. The hurdle for getting it in was too high and it has since languished. We discussed the need to get this done at the MeetBSD developer summit and I agreed to take a fresh look to see if we could pull it off in time for FreeBSD-11. I have started looking at resurrecting the work done in projects/ino64 and will work with Rick to come up with a proposal. As you note, getting a kernel working with backward compatibility is straight-forward. If you have ideas on how to handle the usermode ABI issues, they would be most appreciated. Kirk McKusick
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201411211614.sALGEjRb082219>