http://defect.opensolaris.org/bz/show_bug.cgi?id=5546
Bill Holler <bill.holler at sun.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P4 |P2
Severity|minor |normal
Status|NEW |ACCEPTED
--- Comment #1 from Bill Holler <bill.holler at sun.com> 2008-12-02 11:24:54
---
Raising priority; must be fixed before putback. Interrupt storm regresses
performance. From code inspection it is likely his bug will occur on
all x86 platforms which have an HPET and processor(s) which do not
support deep C-states?
The HPET's general interrupt status register contains 0 when read from
the interrupt service routine.
Currently pcplusmp adds the HPET interrupt early in boot. Later when
the cpudrv determines it can support C-states, the cpudrv plumbs the
interrupt through the IOAPIC. We did not want to add a callback into
pcplusmp to add the HPET's interrupt from the cpudrv. Perhaps we need
to bite the bullet and either add an interface in pcplusmp to add the
HPET's interrupt, or we need to add an ACPI nexus which can be used with
the existing ddi interfaces to add the HPET interrupt.
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
