From: Jason Gunthorpe via iommu <iommu@lists.linux-foundation.org>
To: Jacob Pan <jacob.jun.pan@linux.intel.com>
Cc: "Tian, Kevin" <kevin.tian@intel.com>,
Tony Luck <tony.luck@intel.com>,
Dave Jiang <dave.jiang@intel.com>,
Raj Ashok <ashok.raj@intel.com>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
"Kumar, Sanjay K" <sanjay.k.kumar@intel.com>,
LKML <linux-kernel@vger.kernel.org>,
Christoph Hellwig <hch@infradead.org>,
iommu@lists.linux-foundation.org,
Jacob Pan <jacob.jun.pan@intel.com>,
Barry Song <21cnbao@gmail.com>,
Dan Williams <dan.j.williams@intel.com>,
Jean-Philippe Brucker <jean-philippe@linaro.com>,
"Zanussi, Tom" <tom.zanussi@intel.com>
Subject: Re: [PATCH 3/4] iommu/vt-d: Support PASID DMA for in-kernel usage
Date: Fri, 10 Dec 2021 13:48:48 -0400 [thread overview]
Message-ID: <20211210174848.GH6385@nvidia.com> (raw)
In-Reply-To: <20211210095025.38af67ce@jacob-builder>
On Fri, Dec 10, 2021 at 09:50:25AM -0800, Jacob Pan wrote:
> > Tying pasid to an iommu_domain is not a good idea. An iommu_domain
> > represents an I/O address translation table. It could be attached to a
> > device or a PASID on the device.
>
> I don;t think we can avoid storing PASID at domain level or the group's
> default domain. IOTLB flush is per domain. Default domain of DMA type
> is already tying to PASID0, right?
No, it is just wrong.
If the HW requires a list of everything that is connected to the
iommu_domain then it's private iommu_domain should have that list.
But it is a *list* not a single PASID.
If one device has 10 PASID's pointing to this domain you must flush
them all if that is what the HW requires.
Jason
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
WARNING: multiple messages have this Message-ID (diff)
From: Jason Gunthorpe <jgg@nvidia.com>
To: Jacob Pan <jacob.jun.pan@linux.intel.com>
Cc: Lu Baolu <baolu.lu@linux.intel.com>,
iommu@lists.linux-foundation.org,
LKML <linux-kernel@vger.kernel.org>,
Joerg Roedel <joro@8bytes.org>,
Christoph Hellwig <hch@infradead.org>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Jean-Philippe Brucker <jean-philippe@linaro.com>,
Jacob Pan <jacob.jun.pan@intel.com>,
Raj Ashok <ashok.raj@intel.com>,
"Kumar, Sanjay K" <sanjay.k.kumar@intel.com>,
Dave Jiang <dave.jiang@intel.com>,
Tony Luck <tony.luck@intel.com>, Yi Liu <yi.l.liu@intel.com>,
"Tian, Kevin" <kevin.tian@intel.com>,
Barry Song <21cnbao@gmail.com>,
"Zanussi, Tom" <tom.zanussi@intel.com>,
Dan Williams <dan.j.williams@intel.com>
Subject: Re: [PATCH 3/4] iommu/vt-d: Support PASID DMA for in-kernel usage
Date: Fri, 10 Dec 2021 13:48:48 -0400 [thread overview]
Message-ID: <20211210174848.GH6385@nvidia.com> (raw)
In-Reply-To: <20211210095025.38af67ce@jacob-builder>
On Fri, Dec 10, 2021 at 09:50:25AM -0800, Jacob Pan wrote:
> > Tying pasid to an iommu_domain is not a good idea. An iommu_domain
> > represents an I/O address translation table. It could be attached to a
> > device or a PASID on the device.
>
> I don;t think we can avoid storing PASID at domain level or the group's
> default domain. IOTLB flush is per domain. Default domain of DMA type
> is already tying to PASID0, right?
No, it is just wrong.
If the HW requires a list of everything that is connected to the
iommu_domain then it's private iommu_domain should have that list.
But it is a *list* not a single PASID.
If one device has 10 PASID's pointing to this domain you must flush
them all if that is what the HW requires.
Jason
next prev parent reply other threads:[~2021-12-10 17:48 UTC|newest]
Thread overview: 93+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-07 13:47 [PATCH 0/4] Enable PASID for DMA API users Jacob Pan
2021-12-07 13:47 ` Jacob Pan
2021-12-07 13:47 ` [PATCH 1/4] ioasid: Reserve a global PASID for in-kernel DMA Jacob Pan
2021-12-07 13:47 ` Jacob Pan
2021-12-09 11:03 ` Jean-Philippe Brucker
2021-12-09 11:03 ` Jean-Philippe Brucker
2021-12-09 18:14 ` Jacob Pan
2021-12-09 18:14 ` Jacob Pan
2021-12-10 9:06 ` Jean-Philippe Brucker
2021-12-10 9:06 ` Jean-Philippe Brucker
2021-12-10 12:31 ` Jason Gunthorpe via iommu
2021-12-10 12:31 ` Jason Gunthorpe
2021-12-10 18:05 ` Jacob Pan
2021-12-10 18:05 ` Jacob Pan
2021-12-11 8:39 ` Tian, Kevin
2021-12-11 8:39 ` Tian, Kevin
2021-12-12 23:34 ` Jason Gunthorpe via iommu
2021-12-12 23:34 ` Jason Gunthorpe
2021-12-07 13:47 ` [PATCH 2/4] iommu: Add PASID support for DMA mapping API users Jacob Pan
2021-12-07 13:47 ` Jacob Pan
2021-12-08 2:31 ` Lu Baolu
2021-12-08 2:31 ` Lu Baolu
2021-12-08 18:49 ` Jacob Pan
2021-12-08 18:49 ` Jacob Pan
2021-12-09 1:56 ` Tian, Kevin
2021-12-09 1:56 ` Tian, Kevin
2021-12-09 2:21 ` Lu Baolu
2021-12-09 2:21 ` Lu Baolu
2021-12-09 16:32 ` Jacob Pan
2021-12-09 16:32 ` Jacob Pan
2021-12-09 16:57 ` Raj, Ashok
2021-12-09 16:57 ` Raj, Ashok
2021-12-09 17:34 ` Jacob Pan
2021-12-09 17:34 ` Jacob Pan
2021-12-07 13:47 ` [PATCH 3/4] iommu/vt-d: Support PASID DMA for in-kernel usage Jacob Pan
2021-12-07 13:47 ` Jacob Pan
2021-12-08 13:22 ` Jason Gunthorpe via iommu
2021-12-08 13:22 ` Jason Gunthorpe
2021-12-08 19:16 ` Jacob Pan
2021-12-08 19:16 ` Jacob Pan
2021-12-09 2:32 ` Lu Baolu
2021-12-09 2:32 ` Lu Baolu
2021-12-09 23:21 ` Jacob Pan
2021-12-09 23:21 ` Jacob Pan
2021-12-09 23:41 ` Jason Gunthorpe via iommu
2021-12-09 23:41 ` Jason Gunthorpe
2021-12-10 6:46 ` Lu Baolu
2021-12-10 6:46 ` Lu Baolu
2021-12-10 17:50 ` Jacob Pan
2021-12-10 17:50 ` Jacob Pan
2021-12-10 17:48 ` Jason Gunthorpe via iommu [this message]
2021-12-10 17:48 ` Jason Gunthorpe
2021-12-10 18:18 ` Jacob Pan
2021-12-10 18:18 ` Jacob Pan
2021-12-10 18:53 ` Jason Gunthorpe via iommu
2021-12-10 18:53 ` Jason Gunthorpe
2021-12-07 13:47 ` [PATCH 4/4] dmaengine: idxd: Use DMA API for in-kernel DMA with PASID Jacob Pan
2021-12-07 13:47 ` Jacob Pan
2021-12-07 23:27 ` Dave Jiang
2021-12-07 23:27 ` Dave Jiang
2021-12-08 4:56 ` Vinod Koul
2021-12-08 4:56 ` Vinod Koul
2021-12-08 17:36 ` Jacob Pan
2021-12-08 17:36 ` Jacob Pan
2021-12-08 13:13 ` Jason Gunthorpe via iommu
2021-12-08 13:13 ` Jason Gunthorpe
2021-12-08 15:35 ` Dave Jiang
2021-12-08 15:35 ` Dave Jiang
2021-12-08 17:51 ` Jason Gunthorpe via iommu
2021-12-08 17:51 ` Jason Gunthorpe
2021-12-09 1:48 ` Tian, Kevin
2021-12-09 1:48 ` Tian, Kevin
2021-12-09 19:18 ` Jacob Pan
2021-12-09 19:18 ` Jacob Pan
2021-12-08 19:55 ` Jacob Pan
2021-12-08 19:55 ` Jacob Pan
2021-12-08 20:30 ` Jason Gunthorpe via iommu
2021-12-08 20:30 ` Jason Gunthorpe
2021-12-08 21:59 ` Jacob Pan
2021-12-08 21:59 ` Jacob Pan
2021-12-08 23:39 ` Jason Gunthorpe via iommu
2021-12-08 23:39 ` Jason Gunthorpe
2021-12-09 0:12 ` Dave Jiang
2021-12-09 0:12 ` Dave Jiang
2021-12-09 2:06 ` Tian, Kevin
2021-12-09 2:06 ` Tian, Kevin
2021-12-08 18:37 ` kernel test robot
2021-12-08 18:37 ` kernel test robot
2021-12-08 18:37 ` kernel test robot
2021-12-08 13:10 ` [PATCH 0/4] Enable PASID for DMA API users Jason Gunthorpe via iommu
2021-12-08 13:10 ` Jason Gunthorpe
2021-12-08 18:15 ` Jacob Pan
2021-12-08 18:15 ` Jacob Pan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211210174848.GH6385@nvidia.com \
--to=iommu@lists.linux-foundation.org \
--cc=21cnbao@gmail.com \
--cc=ashok.raj@intel.com \
--cc=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=gregkh@linuxfoundation.org \
--cc=hch@infradead.org \
--cc=jacob.jun.pan@intel.com \
--cc=jacob.jun.pan@linux.intel.com \
--cc=jean-philippe@linaro.com \
--cc=jgg@nvidia.com \
--cc=kevin.tian@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=sanjay.k.kumar@intel.com \
--cc=tom.zanussi@intel.com \
--cc=tony.luck@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.