Aw: RE: [OpenAFS] compile fails kernel version 4.4.0-1-default

Benjamin Kaduk kaduk@MIT.EDU
Mon, 7 Mar 2016 01:42:22 -0500 (EST)

On Thu, 3 Mar 2016, Michael Dressel wrote:

> Hi,
> it is me who reported the issue on archlinux. Let me know if I can help
> with reproducing the issue or anything else.
> For curiosity, could anyone explain to me what change in the 4.4 kernel
> created the issue for openafs? Or rather what was the intention of the
> change in 4.4?

I am given to understand that the proximal trigger is linux commit,
which addds a path wherein -ERESTARTSYS can be returned from within VFS
library code.  (Maybe there are other such paths, but we maybe just didn't
notice before?)  This particular function, splice_from_pipe_next(), ends
up getting called from the low-level afs_linux_storeproc() routine.  There
are many call paths in the cache manager that end up at this function,
most of which are not prepared to properly handle an ERESTARTSYS return.
Since this status can be returned after some data has already been
written, the correct behavior upon receiving it is far from clear ... a
path towards a client free of this vector for data corruption may involve
avoiding the dependence on splice_from_pipe_next() in preference to
adopting all call sites to handle the ERSTARTSYS case.