From: Lu Baolu <baolu.lu-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
To: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>
Cc: kevin.tian-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org,
mika.westerberg-VuQAYsv1563Yd54FQh9/CA@public.gmane.org,
ashok.raj-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org,
Konrad Rzeszutek Wilk
<konrad.wilk-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>,
alan.cox-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org,
Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org,
linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
pengfei.xu-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org,
jacob.jun.pan-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org,
David Woodhouse <dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
Subject: Re: [PATCH v3 02/10] swiotlb: Factor out slot allocation and free
Date: Thu, 25 Apr 2019 10:07:19 +0800 [thread overview]
Message-ID: <a189444b-15c9-8069-901d-8cdf9af7fc3c@linux.intel.com> (raw)
In-Reply-To: <20190424144532.GA21480-jcswGhMUV9g@public.gmane.org>
Hi,
On 4/24/19 10:45 PM, Christoph Hellwig wrote:
> On Tue, Apr 23, 2019 at 03:32:16PM +0800, Lu Baolu wrote:
>> When we add the bounce buffer between IOVA and physical buffer, the
>> bounced buffer must starts from the same offset in a page, otherwise,
>> IOMMU can't work here.
>
> Why? Even with the odd hardware descriptors typical in Intel land that
> only allow offsets in the first page, not the following ones, having
> a zero offset where we previously had one should be fine.
>
This is not VT-d specific. It's just how generic IOMMU works.
Normally, IOMMU works in paging mode. So if a driver issues DMA with
IOVA 0xAAAA0123, IOMMU can remap it with a physical address 0xBBBB0123.
But we should never expect IOMMU to remap 0xAAAA0123 with physical
address of 0xBBBB0000. That's the reason why I said that IOMMU will not
work there.
swiotlb System
IOVA bounce page Memory
.---------. .---------. .---------.
| | | | | |
| | | | | |
buffer_start .---------. .---------.buffer_start .---------.
| |----->| | | |
| | | |************>| |
| | | | swiotlb | |
IOMMU Page '---------' '---------' mapping '---------'
Boundary | | | |
| | | |
| | | |
| |----------------------------->| |
| | IOMMU mapping | |
| | | |
IOMMU Page .---------. .---------.
Boundary | | | |
| | | |
| |----------------------------->| |
| | IOMMU mapping | |
| | | |
| | | |
IOMMU Page .---------. .---------. .---------.
Boundary | | | | | |
| | | |************>| |
| |----->| | swiotlb | |
buffer_end '---------' '---------' mapping '---------'
| | | | | |
| | | | | |
'---------' '---------' '---------'
This is the whole view of iommu bounce page. I expect the buffer_start
returned by swiotlb_tbl_map_single() starts from the same page_offset as
the buffer_start in IOVA.
Best regards,
Lu Baolu
WARNING: multiple messages have this Message-ID (diff)
From: Lu Baolu <baolu.lu@linux.intel.com>
To: Christoph Hellwig <hch@lst.de>
Cc: kevin.tian@intel.com, mika.westerberg@linux.intel.com,
ashok.raj@intel.com,
Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
alan.cox@intel.com, Robin Murphy <robin.murphy@arm.com>,
iommu@lists.linux-foundation.org, linux-kernel@vger.kernel.org,
pengfei.xu@intel.com, jacob.jun.pan@intel.com,
David Woodhouse <dwmw2@infradead.org>
Subject: Re: [PATCH v3 02/10] swiotlb: Factor out slot allocation and free
Date: Thu, 25 Apr 2019 10:07:19 +0800 [thread overview]
Message-ID: <a189444b-15c9-8069-901d-8cdf9af7fc3c@linux.intel.com> (raw)
Message-ID: <20190425020719.RxYsaErZzi85ZABLnzF8uHqsuFedDT6bDS8iEw8s1vw@z> (raw)
In-Reply-To: <20190424144532.GA21480@lst.de>
Hi,
On 4/24/19 10:45 PM, Christoph Hellwig wrote:
> On Tue, Apr 23, 2019 at 03:32:16PM +0800, Lu Baolu wrote:
>> When we add the bounce buffer between IOVA and physical buffer, the
>> bounced buffer must starts from the same offset in a page, otherwise,
>> IOMMU can't work here.
>
> Why? Even with the odd hardware descriptors typical in Intel land that
> only allow offsets in the first page, not the following ones, having
> a zero offset where we previously had one should be fine.
>
This is not VT-d specific. It's just how generic IOMMU works.
Normally, IOMMU works in paging mode. So if a driver issues DMA with
IOVA 0xAAAA0123, IOMMU can remap it with a physical address 0xBBBB0123.
But we should never expect IOMMU to remap 0xAAAA0123 with physical
address of 0xBBBB0000. That's the reason why I said that IOMMU will not
work there.
swiotlb System
IOVA bounce page Memory
.---------. .---------. .---------.
| | | | | |
| | | | | |
buffer_start .---------. .---------.buffer_start .---------.
| |----->| | | |
| | | |************>| |
| | | | swiotlb | |
IOMMU Page '---------' '---------' mapping '---------'
Boundary | | | |
| | | |
| | | |
| |----------------------------->| |
| | IOMMU mapping | |
| | | |
IOMMU Page .---------. .---------.
Boundary | | | |
| | | |
| |----------------------------->| |
| | IOMMU mapping | |
| | | |
| | | |
IOMMU Page .---------. .---------. .---------.
Boundary | | | | | |
| | | |************>| |
| |----->| | swiotlb | |
buffer_end '---------' '---------' mapping '---------'
| | | | | |
| | | | | |
'---------' '---------' '---------'
This is the whole view of iommu bounce page. I expect the buffer_start
returned by swiotlb_tbl_map_single() starts from the same page_offset as
the buffer_start in IOVA.
Best regards,
Lu Baolu
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu
next prev parent reply other threads:[~2019-04-25 2:07 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-04-21 1:17 [PATCH v3 00/10] iommu: Bounce page for untrusted devices Lu Baolu
2019-04-21 1:17 ` Lu Baolu
[not found] ` <20190421011719.14909-1-baolu.lu-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
2019-04-21 1:17 ` [PATCH v3 01/10] iommu: Add helper to get minimal page size of domain Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-29 10:55 ` Robin Murphy
2019-04-29 10:55 ` Robin Murphy
2019-04-30 0:40 ` Lu Baolu
2019-04-30 0:40 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 02/10] swiotlb: Factor out slot allocation and free Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-22 16:45 ` Christoph Hellwig
2019-04-22 16:45 ` Christoph Hellwig
2019-04-23 1:58 ` Lu Baolu
2019-04-23 1:58 ` Lu Baolu
2019-04-23 6:12 ` Christoph Hellwig
2019-04-23 6:12 ` Christoph Hellwig
2019-04-23 7:32 ` Lu Baolu
2019-04-23 7:32 ` Lu Baolu
2019-04-24 14:45 ` Christoph Hellwig
2019-04-24 14:45 ` Christoph Hellwig
[not found] ` <20190424144532.GA21480-jcswGhMUV9g@public.gmane.org>
2019-04-25 2:07 ` Lu Baolu [this message]
2019-04-25 2:07 ` Lu Baolu
2019-04-26 15:04 ` Christoph Hellwig
2019-04-26 15:04 ` Christoph Hellwig
2019-04-29 5:10 ` Lu Baolu
2019-04-29 5:10 ` Lu Baolu
2019-04-29 11:06 ` Robin Murphy
2019-04-29 11:06 ` Robin Murphy
2019-04-29 11:44 ` Christoph Hellwig
2019-04-29 11:44 ` Christoph Hellwig
2019-05-06 1:54 ` Lu Baolu
2019-05-13 7:05 ` Christoph Hellwig
2019-05-16 1:53 ` Lu Baolu
2019-04-30 2:02 ` Lu Baolu
2019-04-30 2:02 ` Lu Baolu
2019-04-30 9:53 ` Robin Murphy
2019-04-30 9:53 ` Robin Murphy
2019-05-02 1:47 ` Lu Baolu
2019-05-02 1:47 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 03/10] swiotlb: Limit tlb address range inside slot pool Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 04/10] swiotlb: Extend swiotlb to support page bounce Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 05/10] iommu: Add bounce page APIs Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 06/10] iommu/vt-d: Add trace events for domain map/unmap Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 07/10] iommu/vt-d: Keep swiotlb on if bounce page is necessary Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-22 16:47 ` Christoph Hellwig
2019-04-22 16:47 ` Christoph Hellwig
2019-04-23 2:00 ` Lu Baolu
2019-04-23 2:00 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 08/10] iommu/vt-d: Check whether device requires bounce buffer Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-22 16:47 ` Christoph Hellwig
2019-04-22 16:47 ` Christoph Hellwig
[not found] ` <20190422164755.GC31181-jcswGhMUV9g@public.gmane.org>
2019-04-23 2:03 ` Lu Baolu
2019-04-23 2:03 ` Lu Baolu
2019-04-23 6:08 ` Christoph Hellwig
2019-04-23 6:08 ` Christoph Hellwig
2019-04-23 7:35 ` Lu Baolu
2019-04-23 7:35 ` Lu Baolu
[not found] ` <e11489cf-87cc-a956-7c6c-ffdd26a4e3ec-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
2019-04-24 18:27 ` Konrad Rzeszutek Wilk
2019-04-24 18:27 ` Konrad Rzeszutek Wilk
2019-04-21 1:17 ` [PATCH v3 09/10] iommu/vt-d: Add dma sync ops for untrusted devices Lu Baolu
2019-04-21 1:17 ` Lu Baolu
2019-04-21 1:17 ` [PATCH v3 10/10] iommu/vt-d: Use bounce buffer " Lu Baolu
2019-04-21 1:17 ` Lu Baolu
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=a189444b-15c9-8069-901d-8cdf9af7fc3c@linux.intel.com \
--to=baolu.lu-vuqaysv1563yd54fqh9/ca@public.gmane.org \
--cc=alan.cox-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
--cc=ashok.raj-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
--cc=dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org \
--cc=hch-jcswGhMUV9g@public.gmane.org \
--cc=iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org \
--cc=jacob.jun.pan-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
--cc=kevin.tian-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
--cc=konrad.wilk-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org \
--cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=mika.westerberg-VuQAYsv1563Yd54FQh9/CA@public.gmane.org \
--cc=pengfei.xu-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
--cc=robin.murphy-5wv7dgnIgG8@public.gmane.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox