Re: [PATCH v4 08/28] x86/vvtd: Add MMIO handler for VVTD

From: Chao Gao <chao.gao@intel.com>
To: "Roger Pau Monné" <roger.pau@citrix.com>
Cc: Lan Tianyu <tianyu.lan@intel.com>,
	Kevin Tian <kevin.tian@intel.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Wei Liu <wei.liu2@citrix.com>,
	George Dunlap <george.dunlap@eu.citrix.com>,
	Ian Jackson <ian.jackson@eu.citrix.com>, Tim Deegan <tim@xen.org>,
	xen-devel@lists.xen.org, Jan Beulich <jbeulich@suse.com>,
	Andrew Cooper <andrew.cooper3@citrix.com>
Subject: Re: [PATCH v4 08/28] x86/vvtd: Add MMIO handler for VVTD
Date: Thu, 22 Feb 2018 14:20:12 +0800	[thread overview]
Message-ID: <20180222062011.GA105139@skl-4s-chao.sh.intel.com> (raw)
In-Reply-To: <20180209175129.fil4uetd2nxzhvfr@MacBook-Pro-de-Roger.local>

On Fri, Feb 09, 2018 at 05:51:29PM +0000, Roger Pau Monné wrote:
>On Sat, Feb 10, 2018 at 01:21:09AM +0800, Chao Gao wrote:
>> On Fri, Feb 09, 2018 at 04:39:15PM +0000, Roger Pau Monné wrote:
>> >On Fri, Nov 17, 2017 at 02:22:15PM +0800, Chao Gao wrote:
>> >> This patch adds VVTD MMIO handler to deal with MMIO access.
>> >> 
>> >> Signed-off-by: Chao Gao <chao.gao@intel.com>
>> >> Signed-off-by: Lan Tianyu <tianyu.lan@intel.com>
>> >> ---
>> >> v4:
>> >>  - only trap the register emulated in vvtd_in_range().
>> >>    i.e. replace PAGE_SIZE with the VVTD_MAX_OFFSET
>> >> ---
>> >>  xen/drivers/passthrough/vtd/vvtd.c | 55 ++++++++++++++++++++++++++++++++++++++
>> >>  1 file changed, 55 insertions(+)
>> >> 
>> >> diff --git a/xen/drivers/passthrough/vtd/vvtd.c b/xen/drivers/passthrough/vtd/vvtd.c
>> >> index 9f76ccf..d78d878 100644
>> >> --- a/xen/drivers/passthrough/vtd/vvtd.c
>> >> +++ b/xen/drivers/passthrough/vtd/vvtd.c
>> >
>> >Now that I look at this, this is the wrong folder. This should be in
>> >xen/arch/x86/hvm with the rest of the emulated devices.
>> 
>> It is a problem we discussed in previous versions. AMD puts its vIOMMU
>> (iommu_guest.c) in xen/drivers/passthrough/amd/. We are following what
>> they did. I don't have special taste on this. If no one objects to your
>> suggestion, I will move it to xen/arch/x86/hvm/. Maybe create a new
>> intel directory since it's intel-specific and won't be used by AMD.
>
>Oh, it's been quite some time since I've reviewed that, so TBH I
>didn't remember that discussion.
>
>If the AMD viommu thing is already there I guess it doesn't hurt...
>Also, have you checked whether it can be converted to use the
>infrastructure that you add here?

Not yet. It seems that we have no method to use AMD vIOMMU now.
And I notice that Wei plans to remove AMD vIOMMU.

I can convert AMD vIOMMU implementation to use this infrastructure if we
finally decide to preserve AMD vIOMMU.

Thanks
Chao

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel