Re: EINVAL from copyin/out - how?

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Wed, 13 Aug 2014 04:38:19 +1000
    From:        matthew green <mrg%eterna.com.au@localhost>
    Message-ID:  <26310.1407868699%splode.eterna.com.au@localhost>

  | these can trigger page faults, which in turn can trigger paging
  | (in this case, most likely from swap.)

Unlikely in this case, and I can easily arrange for no swap space to
validate that - the test is to boot, mount the cd (dvd), cd to it, and
tar cf - . >/dev/null

Page faults are entirely possible, but they should be zfod pages being
created, rather than anything read from swap (there's plenty of ram
available, and fortunately, on this particular dvd, the bad files are
found very early - if after that, I let tar continue, it will happily
read lots more from the dvd and happily write it to /dev/null).

  | the result of that is eventually handed back to copyout(),

It would be useful to have copyout(9) at least mention whatever error
codes are possible ...

I am going to keep hunting - when I first started looking, I was expecting
to find the EINVAL being generated by something looking at the structure of
the filesystem on the DVD, and then perhaps to initiate a "discussion" on
whether it was mkisofs or cd9660fs interpreting the spec in an unexpected
way ...   but that isn't how it turned out.

kre