https://bugzilla.novell.com/show_bug.cgi?id=229210#c57
Tejun Heo changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|FIXED |
--- Comment #57 from Tejun Heo 2007-10-08 19:57:55 MST ---
That's a serious regression and shouldn't be happening. Reopening. libata now
has a logic which detects spindown commands issued from userland and skips
kernel side spin down if it already happened but, for safety's sake, flush
cache is never skipped. This sometimes leads to double spin down depending on
how the disk drive behaves when FLUSH CACHE is issued while the disk is spun
down with empty cache.
1. userland flushes cache
2. userland issues spin down (libata records the last command was spin down
request)
3. kernel shutdown initiates
4. kernel flushes cache (some drives spin back up here to flush the already
empty cache)
5. kernel skips spin down and warns
Not many drives spins back up as in #4 and in any case, there should be no data
loss. I'll look deeper into what's going on.
Thanks.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.