HEADSUP: new linux infrastructure ports are committed

[Kostik Belousov]

On Fri, Apr 03, 2009 at 12:22:51PM +0200, Alexander Leidinger wrote:
> Quoting Kostik Belousov <kostikbel at gmail.com> (from Fri, 3 Apr 2009  
> 12:51:27 +0300):
> 
> >On Fri, Apr 03, 2009 at 08:27:42AM +0200, Alexander Leidinger wrote:
> >>Quoting "Sean C. Farley" <scf at FreeBSD.org> (from Thu, 2 Apr 2009
> >>12:33:42 -0500 (CDT)):
> >>
> >>>What cannot be MFC'd besides TLS and NPTL?  For syscalls that cannot
> >>>be MFC'd, can calls be written that help to some degree without
> >>>breaking ABI?  The wiki page[1] is incomplete on what changes are to
> >>>be MFC'd.  I can say that fstatat64() seems popular (called by rm)
> >>>in /var/log/messages.
> >>
> >>*at() can not be MFCed as it would change the native VFS ABI (Roman
> >>implemented the FreeBSD side of *at() calls to be able to emulate the
> >>linux *at() calls).
> >It can be MFCed with low-impact KBI breakage, and I have re@ approval
> 
> You need to rebuild all FS then, don't you? Would FUSE in this case  
> refuse to load when not rebuild? If not, what would be the impact of  
> not rebuilding but loading it?
No, fs modules are actually immune, mostly. What changes is the namei()
binary interface. It is needed for new ZFS too, for what it worth.

> 
> >for the MFC. What actually stopped the merge is an issue with missed
> >audit bits for dirfd descriptors. They are missed in HEAD too, so
> >8.0 needs some action to finish this.
> 
> Thanks for the info,
> Alexander.
> 
> -- 
> Scintillation is not always identification for an auric substance.
> 
> http://www.Leidinger.net    Alexander @ Leidinger.net: PGP ID = B0063FE7
> http://www.FreeBSD.org       netchild @ FreeBSD.org  : PGP ID = 72077137
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-emulation/attachments/20090403/8af09e2c/attachment.pgp