From: Will Deacon <will.deacon@arm.com>
To: Nicolas Boichat <drinkcat@chromium.org>
Cc: Robin Murphy <robin.murphy@arm.com>,
Christoph Lameter <cl@linux.com>,
Vlastimil Babka <vbabka@suse.cz>, Michal Hocko <mhocko@suse.com>,
Matthias Brugger <matthias.bgg@gmail.com>,
hch@infradead.org, Matthew Wilcox <willy@infradead.org>,
Joerg Roedel <joro@8bytes.org>, Pekka Enberg <penberg@kernel.org>,
David Rientjes <rientjes@google.com>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Andrew Morton <akpm@linux-foundation.org>,
Mel Gorman <mgorman@techsingularity.net>,
Levin Alexander <Alexander.Levin@microsoft.com>,
Huaisheng Ye <yehs1@lenovo.com>,
Mike Rapoport <rppt@linux.vnet.ibm.com>,
linux-arm Mailing List <linux-arm-kernel@lists.infradead.org>,
iommu@lists.linux-foundation.org,
lkml <linux-kernel@vger.kernel.org>,
linux-mm@kvack.org, Yong Wu <yong.wu@mediatek.com>,
Tomasz Figa <tfiga@google.com>,
yingjoe.chen@mediatek.com, Hsin-Yi Wang <hsinyi@chromium.org>,
Daniel Kurtz <djkurtz@chromium.org>
Subject: Re: [PATCH v2 0/3] iommu/io-pgtable-arm-v7s: Use DMA32 zone for page tables
Date: Tue, 4 Dec 2018 16:28:44 +0000 [thread overview]
Message-ID: <20181204162844.GA8169@arm.com> (raw)
In-Reply-To: <CANMq1KDxmRcWhtaJbrLHqx6yPGkNaK7WNYYf+iFjH1e8XdrwRg@mail.gmail.com>
On Tue, Dec 04, 2018 at 05:37:13PM +0800, Nicolas Boichat wrote:
> On Sun, Nov 11, 2018 at 5:04 PM Nicolas Boichat <drinkcat@chromium.org> wrote:
> >
> > This is a follow-up to the discussion in [1], to make sure that the page
> > tables allocated by iommu/io-pgtable-arm-v7s are contained within 32-bit
> > physical address space.
> >
> > [1] https://lists.linuxfoundation.org/pipermail/iommu/2018-November/030876.html
>
> Hi everyone,
>
> Let's try to summarize here.
>
> First, we confirmed that this is a regression, and IOMMU errors happen
> on 4.19 and linux-next/master on MT8173 (elm, Acer Chromebook R13).
> The issue most likely starts from ad67f5a6545f ("arm64: replace
> ZONE_DMA with ZONE_DMA32"), i.e. 4.15, and presumably breaks a number
> of Mediatek platforms (and maybe others?).
>
> We have a few options here:
> 1. This series [2], that adds support for GFP_DMA32 slab caches,
> _without_ adding kmalloc caches (since there are no users of
> kmalloc(..., GFP_DMA32)). I think I've addressed all the comments on
> the 3 patches, and AFAICT this solution works fine.
> 2. genalloc. That works, but unless we preallocate 4MB for L2 tables
> (which is wasteful as we usually only need a handful of L2 tables),
> we'll need changes in the core (use GFP_ATOMIC) to allow allocating on
> demand, and as it stands we'd have no way to shrink the allocation.
> 3. page_frag [3]. That works fine, and the code is quite simple. One
> drawback is that fragments in partially freed pages cannot be reused
> (from limited experiments, I see that IOMMU L2 tables are rarely
> freed, so it's unlikely a whole page would get freed). But given the
> low number of L2 tables, maybe we can live with that.
>
> I think 2 is out. Any preference between 1 and 3? I think 1 makes
> better use of the memory, so that'd be my preference. But I'm probably
> missing something.
FWIW, I'm open to any solution at this point, since I'd like to see this
regression fixed. (1) does sound better longer-term, but (3) looks pretty
much ready to do afaict.
Will
WARNING: multiple messages have this Message-ID (diff)
From: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
To: Nicolas Boichat <drinkcat-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
Cc: Michal Hocko <mhocko-IBi9RG/b67k@public.gmane.org>,
Daniel Kurtz <djkurtz-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>,
Levin Alexander
<Alexander.Levin-0li6OtcxBFHby3iVrkZq2A@public.gmane.org>,
linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org,
Christoph Lameter <cl-vYTEC60ixJUAvxtiuMwx3w@public.gmane.org>,
Huaisheng Ye <yehs1-6jq1YtArVR3QT0dZR+AlfA@public.gmane.org>,
Matthew Wilcox <willy-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>,
linux-arm Mailing List
<linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org>,
David Rientjes <rientjes-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>,
yingjoe.chen-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org,
Vlastimil Babka <vbabka-AlSwsSmVLrQ@public.gmane.org>,
Tomasz Figa <tfiga-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>,
Mike Rapoport
<rppt-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org>,
Hsin-Yi Wang <hsinyi-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>,
Matthias Brugger
<matthias.bgg-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Joonsoo Kim <iamjoonsoo.kim-Hm3cg6mZ9cc@public.gmane.org>,
Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>,
lkml <linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>,
Pekka Enberg <penberg-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org,
Andrew Morton
<akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>,
Mel Gorman
<mgorman-3eNAlZScCAx27rWaFMvyedHuzzzSOjJt@public.gmane.org>
Subject: Re: [PATCH v2 0/3] iommu/io-pgtable-arm-v7s: Use DMA32 zone for page tables
Date: Tue, 4 Dec 2018 16:28:44 +0000 [thread overview]
Message-ID: <20181204162844.GA8169@arm.com> (raw)
In-Reply-To: <CANMq1KDxmRcWhtaJbrLHqx6yPGkNaK7WNYYf+iFjH1e8XdrwRg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
On Tue, Dec 04, 2018 at 05:37:13PM +0800, Nicolas Boichat wrote:
> On Sun, Nov 11, 2018 at 5:04 PM Nicolas Boichat <drinkcat-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org> wrote:
> >
> > This is a follow-up to the discussion in [1], to make sure that the page
> > tables allocated by iommu/io-pgtable-arm-v7s are contained within 32-bit
> > physical address space.
> >
> > [1] https://lists.linuxfoundation.org/pipermail/iommu/2018-November/030876.html
>
> Hi everyone,
>
> Let's try to summarize here.
>
> First, we confirmed that this is a regression, and IOMMU errors happen
> on 4.19 and linux-next/master on MT8173 (elm, Acer Chromebook R13).
> The issue most likely starts from ad67f5a6545f ("arm64: replace
> ZONE_DMA with ZONE_DMA32"), i.e. 4.15, and presumably breaks a number
> of Mediatek platforms (and maybe others?).
>
> We have a few options here:
> 1. This series [2], that adds support for GFP_DMA32 slab caches,
> _without_ adding kmalloc caches (since there are no users of
> kmalloc(..., GFP_DMA32)). I think I've addressed all the comments on
> the 3 patches, and AFAICT this solution works fine.
> 2. genalloc. That works, but unless we preallocate 4MB for L2 tables
> (which is wasteful as we usually only need a handful of L2 tables),
> we'll need changes in the core (use GFP_ATOMIC) to allow allocating on
> demand, and as it stands we'd have no way to shrink the allocation.
> 3. page_frag [3]. That works fine, and the code is quite simple. One
> drawback is that fragments in partially freed pages cannot be reused
> (from limited experiments, I see that IOMMU L2 tables are rarely
> freed, so it's unlikely a whole page would get freed). But given the
> low number of L2 tables, maybe we can live with that.
>
> I think 2 is out. Any preference between 1 and 3? I think 1 makes
> better use of the memory, so that'd be my preference. But I'm probably
> missing something.
FWIW, I'm open to any solution at this point, since I'd like to see this
regression fixed. (1) does sound better longer-term, but (3) looks pretty
much ready to do afaict.
Will
WARNING: multiple messages have this Message-ID (diff)
From: Will Deacon <will.deacon@arm.com>
To: Nicolas Boichat <drinkcat@chromium.org>
Cc: Michal Hocko <mhocko@suse.com>,
Daniel Kurtz <djkurtz@chromium.org>,
Levin Alexander <Alexander.Levin@microsoft.com>,
linux-mm@kvack.org, Christoph Lameter <cl@linux.com>,
Huaisheng Ye <yehs1@lenovo.com>, Joerg Roedel <joro@8bytes.org>,
Matthew Wilcox <willy@infradead.org>,
hch@infradead.org,
linux-arm Mailing List <linux-arm-kernel@lists.infradead.org>,
David Rientjes <rientjes@google.com>,
yingjoe.chen@mediatek.com, Vlastimil Babka <vbabka@suse.cz>,
Tomasz Figa <tfiga@google.com>,
Mike Rapoport <rppt@linux.vnet.ibm.com>,
Hsin-Yi Wang <hsinyi@chromium.org>,
Matthias Brugger <matthias.bgg@gmail.com>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Yong Wu <yong.wu@mediatek.com>,
Robin Murphy <robin.murphy@arm.com>,
lkml <linux-kernel@vger.kernel.org>,
Pekka Enberg <penberg@kernel.org>,
iommu@lists.linux-foundation.org,
Andrew Morton <akpm@linux-foundation.org>,
Mel Gorman <mgorman@techsingularity.net>
Subject: Re: [PATCH v2 0/3] iommu/io-pgtable-arm-v7s: Use DMA32 zone for page tables
Date: Tue, 4 Dec 2018 16:28:44 +0000 [thread overview]
Message-ID: <20181204162844.GA8169@arm.com> (raw)
In-Reply-To: <CANMq1KDxmRcWhtaJbrLHqx6yPGkNaK7WNYYf+iFjH1e8XdrwRg@mail.gmail.com>
On Tue, Dec 04, 2018 at 05:37:13PM +0800, Nicolas Boichat wrote:
> On Sun, Nov 11, 2018 at 5:04 PM Nicolas Boichat <drinkcat@chromium.org> wrote:
> >
> > This is a follow-up to the discussion in [1], to make sure that the page
> > tables allocated by iommu/io-pgtable-arm-v7s are contained within 32-bit
> > physical address space.
> >
> > [1] https://lists.linuxfoundation.org/pipermail/iommu/2018-November/030876.html
>
> Hi everyone,
>
> Let's try to summarize here.
>
> First, we confirmed that this is a regression, and IOMMU errors happen
> on 4.19 and linux-next/master on MT8173 (elm, Acer Chromebook R13).
> The issue most likely starts from ad67f5a6545f ("arm64: replace
> ZONE_DMA with ZONE_DMA32"), i.e. 4.15, and presumably breaks a number
> of Mediatek platforms (and maybe others?).
>
> We have a few options here:
> 1. This series [2], that adds support for GFP_DMA32 slab caches,
> _without_ adding kmalloc caches (since there are no users of
> kmalloc(..., GFP_DMA32)). I think I've addressed all the comments on
> the 3 patches, and AFAICT this solution works fine.
> 2. genalloc. That works, but unless we preallocate 4MB for L2 tables
> (which is wasteful as we usually only need a handful of L2 tables),
> we'll need changes in the core (use GFP_ATOMIC) to allow allocating on
> demand, and as it stands we'd have no way to shrink the allocation.
> 3. page_frag [3]. That works fine, and the code is quite simple. One
> drawback is that fragments in partially freed pages cannot be reused
> (from limited experiments, I see that IOMMU L2 tables are rarely
> freed, so it's unlikely a whole page would get freed). But given the
> low number of L2 tables, maybe we can live with that.
>
> I think 2 is out. Any preference between 1 and 3? I think 1 makes
> better use of the memory, so that'd be my preference. But I'm probably
> missing something.
FWIW, I'm open to any solution at this point, since I'd like to see this
regression fixed. (1) does sound better longer-term, but (3) looks pretty
much ready to do afaict.
Will
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2018-12-04 16:28 UTC|newest]
Thread overview: 96+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-11 9:03 [PATCH v2 0/3] iommu/io-pgtable-arm-v7s: Use DMA32 zone for page tables Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-11 9:03 ` [PATCH v2 1/3] mm: slab/slub: Add check_slab_flags function to check for valid flags Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-11 9:03 ` [PATCH v2 2/3] mm: Add support for SLAB_CACHE_DMA32 Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-21 18:32 ` Christopher Lameter
2018-11-21 18:32 ` Christopher Lameter
2018-11-22 0:52 ` Nicolas Boichat
2018-11-22 0:52 ` Nicolas Boichat
2018-11-22 0:52 ` Nicolas Boichat
2018-11-11 9:03 ` [PATCH v2 3/3] iommu/io-pgtable-arm-v7s: Request DMA32 memory, and improve debugging Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-11 9:03 ` Nicolas Boichat
2018-11-21 16:46 ` Will Deacon
2018-11-21 16:46 ` Will Deacon
2018-11-21 17:38 ` Christopher Lameter
2018-11-21 17:38 ` Christopher Lameter
2018-11-21 17:43 ` Robin Murphy
2018-11-21 17:43 ` Robin Murphy
2018-11-21 18:18 ` Christopher Lameter
2018-11-21 18:18 ` Christopher Lameter
2018-11-21 18:02 ` Michal Hocko
2018-11-21 18:02 ` Michal Hocko
2018-11-22 1:20 ` Nicolas Boichat
2018-11-22 1:20 ` Nicolas Boichat
2018-11-22 1:20 ` Nicolas Boichat
2018-11-23 12:15 ` Vlastimil Babka
2018-11-23 12:15 ` Vlastimil Babka
2018-11-21 18:20 ` [PATCH v2 0/3] iommu/io-pgtable-arm-v7s: Use DMA32 zone for page tables Christopher Lameter
2018-11-21 18:20 ` Christopher Lameter
2018-11-21 21:38 ` Matthew Wilcox
2018-11-21 21:38 ` Matthew Wilcox
2018-11-21 21:38 ` Matthew Wilcox
2018-11-21 22:26 ` Robin Murphy
2018-11-21 22:26 ` Robin Murphy
2018-11-21 22:26 ` Robin Murphy
2018-11-22 1:05 ` Nicolas Boichat
2018-11-22 1:05 ` Nicolas Boichat
2018-11-22 1:05 ` Nicolas Boichat
2018-11-22 2:35 ` Matthew Wilcox
2018-11-22 2:35 ` Matthew Wilcox
2018-11-22 2:35 ` Matthew Wilcox
2018-11-22 5:56 ` Nicolas Boichat
2018-11-22 5:56 ` Nicolas Boichat
2018-11-22 5:56 ` Nicolas Boichat
2018-11-22 8:26 ` Christoph Hellwig
2018-11-22 8:26 ` Christoph Hellwig
2018-11-22 8:26 ` Christoph Hellwig
2018-11-22 15:16 ` Matthew Wilcox
2018-11-22 15:16 ` Matthew Wilcox
2018-11-22 15:16 ` Matthew Wilcox
2018-11-22 15:19 ` Christoph Hellwig
2018-11-22 15:19 ` Christoph Hellwig
2018-11-22 15:19 ` Christoph Hellwig
2018-11-22 8:23 ` Christoph Hellwig
2018-11-22 8:23 ` Christoph Hellwig
2018-11-22 8:23 ` Christoph Hellwig
2018-11-23 3:04 ` Nicolas Boichat
2018-11-23 3:04 ` Nicolas Boichat
2018-11-23 3:04 ` Nicolas Boichat
2018-11-23 5:37 ` Nicolas Boichat
2018-11-23 5:37 ` Nicolas Boichat
2018-11-23 5:37 ` Nicolas Boichat
2018-11-23 12:23 ` Vlastimil Babka
2018-11-23 12:23 ` Vlastimil Babka
2018-11-23 12:23 ` Vlastimil Babka
2018-11-23 12:30 ` Michal Hocko
2018-11-23 12:30 ` Michal Hocko
2018-11-23 12:30 ` Michal Hocko
2018-11-26 8:02 ` Christoph Hellwig
2018-11-26 8:02 ` Christoph Hellwig
2018-11-26 8:02 ` Christoph Hellwig
2018-11-28 8:55 ` Nicolas Boichat
2018-11-28 8:55 ` Nicolas Boichat
2018-11-28 8:55 ` Nicolas Boichat
2018-12-04 9:37 ` Nicolas Boichat
2018-12-04 9:37 ` Nicolas Boichat
2018-12-04 14:35 ` Vlastimil Babka
2018-12-04 14:35 ` Vlastimil Babka
2018-12-04 14:35 ` Vlastimil Babka
2018-12-05 2:04 ` Nicolas Boichat
2018-12-05 2:04 ` Nicolas Boichat
2018-12-05 2:04 ` Nicolas Boichat
2018-12-05 5:51 ` Nicolas Boichat
2018-12-05 5:51 ` Nicolas Boichat
2018-12-05 5:51 ` Nicolas Boichat
2018-12-05 14:41 ` Will Deacon
2018-12-05 14:41 ` Will Deacon
2018-12-05 14:41 ` Will Deacon
2018-12-04 16:28 ` Will Deacon [this message]
2018-12-04 16:28 ` Will Deacon
2018-12-04 16:28 ` Will Deacon
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=20181204162844.GA8169@arm.com \
--to=will.deacon@arm.com \
--cc=Alexander.Levin@microsoft.com \
--cc=akpm@linux-foundation.org \
--cc=cl@linux.com \
--cc=djkurtz@chromium.org \
--cc=drinkcat@chromium.org \
--cc=hch@infradead.org \
--cc=hsinyi@chromium.org \
--cc=iamjoonsoo.kim@lge.com \
--cc=iommu@lists.linux-foundation.org \
--cc=joro@8bytes.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=matthias.bgg@gmail.com \
--cc=mgorman@techsingularity.net \
--cc=mhocko@suse.com \
--cc=penberg@kernel.org \
--cc=rientjes@google.com \
--cc=robin.murphy@arm.com \
--cc=rppt@linux.vnet.ibm.com \
--cc=tfiga@google.com \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
--cc=yehs1@lenovo.com \
--cc=yingjoe.chen@mediatek.com \
--cc=yong.wu@mediatek.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.