Re: splice: infinite busy loop lockup bug

[Al Viro]

On Fri, Aug 07, 2020 at 01:27:27PM +0100, Al Viro wrote:
> On Fri, Aug 07, 2020 at 07:35:08PM +0900, Tetsuo Handa wrote:
> > syzbot is reporting hung task at pipe_release() [1], for for_each_bvec() from
> > iterate_bvec() from iterate_all_kinds() from iov_iter_alignment() from
> > ext4_unaligned_io() from ext4_dio_write_iter() from ext4_file_write_iter() from
> > call_write_iter() from do_iter_readv_writev() from do_iter_write() from
> > vfs_iter_write() from iter_file_splice_write() falls into infinite busy loop
> > with pipe->mutex held.
> > 
> > The reason of falling into infinite busy loop is that iter_file_splice_write()
> > for some reason generates "struct bio_vec" entry with .bv_len=0 and .bv_offset=0
> > while for_each_bvec() cannot handle .bv_len == 0.
> 
> broken in 1bdc76aea115 "iov_iter: use bvec iterator to implement iterate_bvec()",
> unless I'm misreading it...
> 
> Zero-length segments are not disallowed; it's not all that hard to filter them
> out in iter_file_splice_write(), but the intent had always been to have
> iterate_all_kinds() et.al. able to cope with those.
> 
> How are these pipe_buffers with ->len == 0 generated in that reproducer, BTW?
> There might be something else fishy going on...

FWIW, my preference would be to have for_each_bvec() advance past zero-length
segments; I'll need to go through its uses elsewhere in the tree first, though
(after I grab some sleep),