Received-SPF: pass (google.com: domain of linux-kernel+bounces-40296-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Fri, 26 Jan 2024 11:44:27 -0400
From: Jason Gunthorpe <jgg@ziepe.ca>
To: Timothy Pearson <tpearson@raptorengineering.com>
Cc: Shivaprasad G Bhat <sbhat@linux.ibm.com>, iommu <iommu@lists.linux.dev>,
	linuxppc-dev <linuxppc-dev@lists.ozlabs.org>,
	linux-kernel <linux-kernel@vger.kernel.org>,
	Michael Ellerman <mpe@ellerman.id.au>, npiggin <npiggin@gmail.com>,
	christophe leroy <christophe.leroy@csgroup.eu>,
	aneesh kumar <aneesh.kumar@kernel.org>,
	naveen n rao <naveen.n.rao@linux.ibm.com>,
	jroedel <jroedel@suse.de>, aik <aik@amd.com>,
	bgray <bgray@linux.ibm.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	gbatra <gbatra@linux.vnet.ibm.com>, vaibhav <vaibhav@linux.ibm.com>
Subject: Re: [PATCH 1/2] powerpc: iommu: Bring back table group
 release_ownership() call
Message-ID: <20240126154427.GB50608@ziepe.ca>
References: <170618450592.3805.8216395093813382208.stgit@ltcd48-lp2.aus.stglab.ibm.com>
 <170618451433.3805.9015493852395837391.stgit@ltcd48-lp2.aus.stglab.ibm.com>
 <20240125155017.GW50608@ziepe.ca>
 <b825dd04-3d32-4fbd-91e3-523ddf96fc7a@linux.ibm.com>
 <20240126151701.GZ50608@ziepe.ca>
 <392247278.10124607.1706282995795.JavaMail.zimbra@raptorengineeringinc.com>
 <20240126153806.GA50608@ziepe.ca>
 <1453449220.10126352.1706283596477.JavaMail.zimbra@raptorengineeringinc.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1453449220.10126352.1706283596477.JavaMail.zimbra@raptorengineeringinc.com>

On Fri, Jan 26, 2024 at 09:39:56AM -0600, Timothy Pearson wrote:
> > Just forget about the weird KVM and SPAPR stuff, leave it under the
> > kconfig of the old code and nobody will run it. Almost nobody already
> > runs it, apparently.
> 
> We actually use QEMU/KVM/VFIO extensively at Raptor, so need the
> support and need it to be performant...

I wonder if you alone are the "almost" :)

The KVM entanglement was hairy and scary. I never did figure out what
was really going on there. Maybe you don't need all of it and can be
successful with a more typical iommu working model?

Suggest to tackle it after getting the first parts done.

Jason