[Bug 212139] r298900 introduced a fatal failure case for >2TB disk size reporting bugs

bugzilla-noreply at freebsd.org bugzilla-noreply at freebsd.org
Thu Aug 25 20:53:28 UTC 2016


https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=212139

--- Comment #4 from Peter Wemm <peter at FreeBSD.org> ---
FWIW; the thought experiment patch is wrong - it does not implement what I
thought  it did.

I still think it is probably the best option though.  The root of the problem
is the read-ahead and we cannot trust size reporting.  The old code did not do
read-ahead. Preventing read-ahead from crossing the end of disk, or an alias of
the end of disk should be more than sufficient to restore old behavior without
adding new hard failure cases.

-- 
You are receiving this mail because:
You are the assignee for the bug.


More information about the freebsd-bugs mailing list