Hi Joerg,
I see your point(s). First, to the second point, I
should (and will) add a check to the AMD Family/Model coupled to this change.
Also, I will append the patch(es) to include the perf IOMMU PMU so that this
change is put into context.
Thanks for your time and advise.
Best Regards,
Steve
-----Original Message-----
From: Joerg Roedel [mailto:[email protected]]
Sent: Thursday, December 06, 2012 4:33 AM
To: Kinney, Steven
Cc: Hiroshi DOYU; Stephen Warren; Jiri Kosina; Kukjin Kim;
[email protected]; [email protected]
Subject: Re: [PATCH 1/1] iommuv2/amd: Enable Performance Counters On Family 15h
Models 10h-1Fh
On Tue, Dec 04, 2012 at 04:19:06PM -0600, Steven Kinney wrote:
> drivers/iommu/Kconfig | 10 +++++
> drivers/iommu/amd_iommu_init.c | 85
> +++++++++++++++++++++++++++++++++++++++
> drivers/iommu/amd_iommu_types.h | 13 +++++-
> 3 files changed, 107 insertions(+), 1 deletion(-)
So first of all, this patch alone doesn't make sense. You should submit it in a
patch-set that enabled IOMMUv2 performance counters for the end-user. This
means together with the PMU implementation and tools support.
Further I don't see how this patch is specific to Family 15h Models 10h-1Fh as
described in the subject-line.
Any comments on these two points?
Cheers,
Joerg
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu