From: Stanislaw Gruszka <sgruszka@redhat.com> To: Robin Murphy <robin.murphy@arm.com> Cc: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>, Samuel Sieb <samuel@sieb.net>, linux-wireless <linux-wireless@vger.kernel.org>, linux-kernel@vger.kernel.org, iommu@lists.linux-foundation.org, Rosen Penev <rosenp@gmail.com>, Alexander Duyck <alexander.h.duyck@linux.intel.com> Subject: Re: MT76x2U crashes XHCI driver on AMD Ryzen system Date: Tue, 19 Feb 2019 12:08:52 +0100 [thread overview] Message-ID: <20190219110851.GA23223@redhat.com> (raw) In-Reply-To: <2cc5674a-a3a0-d8fe-65f5-4357da9b85d3@arm.com> On Mon, Feb 18, 2019 at 05:01:59PM +0000, Robin Murphy wrote: > On 18/02/2019 14:37, Stanislaw Gruszka wrote: > [...] > >Another issue is that dma_map_sg() & dma_map_page() may require some > >constraints. I'm not sure about that and I want to clarify that with > >CCed mm maintainers. I think DMA drivers may expect sg->offset < PAGE_SIZE > >for both dma_map_sg() and dma_map_page(). Additionally dma_map_page() > >maight expect that offset & length specify buffer within one page. > > Luckily, this came up a while back[1] and we seemed to reach a > consensus that sg->offset >= PAGE_SIZE for dma_map_sg() was weird > but valid. IIRC it was only the Intel IOMMU code which failed to > handle that case appropriately (and which I fixed) - the AMD IOMMU > code always looked like it should be OK, but I'm not sure I've ever > seen definitive test results (and I don't have hardware to do so > myself). Funny that we have problems on AMD IOMMU and not with Intel IOMMU. > For dma_map_page(), length >= PAGE_SIZE should be perfectly valid > and handled correctly. The offset >= PAGE_SIZE case is a bit harder > to justify, but at the same time has less scope for the DMA API > backend to get it wrong, so either way is likely to be OK in > practice (in particular the AMD IOMMU code looks like it won't have > a problem, since its map_page() implementation converts page and > offset to a plain physical address before doing anything else). Thanks for clarify this. So my patch which do: - page = virt_to_head_page(data); + page = virt_to_page(data); offset = data - page_address(page); sg_set_page(&urb->sg[i], page, sglen, offset); should not be necessary as IOMMU driver do exactly the same internally. Are there any alignment requirement for offset for dma_map_{page,sg} ? It will work with let say sg->offset=113 or we have make sure it is aligned to some boundary. If so, what boundary ? Stanislaw
WARNING: multiple messages have this Message-ID (diff)
From: Stanislaw Gruszka <sgruszka-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> To: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org> Cc: Lorenzo Bianconi <lorenzo.bianconi-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>, Samuel Sieb <samuel-yL8jFUf1A1M@public.gmane.org>, linux-wireless <linux-wireless-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, Rosen Penev <rosenp-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>, Alexander Duyck <alexander.h.duyck-VuQAYsv1563Yd54FQh9/CA@public.gmane.org> Subject: Re: MT76x2U crashes XHCI driver on AMD Ryzen system Date: Tue, 19 Feb 2019 12:08:52 +0100 [thread overview] Message-ID: <20190219110851.GA23223@redhat.com> (raw) In-Reply-To: <2cc5674a-a3a0-d8fe-65f5-4357da9b85d3-5wv7dgnIgG8@public.gmane.org> On Mon, Feb 18, 2019 at 05:01:59PM +0000, Robin Murphy wrote: > On 18/02/2019 14:37, Stanislaw Gruszka wrote: > [...] > >Another issue is that dma_map_sg() & dma_map_page() may require some > >constraints. I'm not sure about that and I want to clarify that with > >CCed mm maintainers. I think DMA drivers may expect sg->offset < PAGE_SIZE > >for both dma_map_sg() and dma_map_page(). Additionally dma_map_page() > >maight expect that offset & length specify buffer within one page. > > Luckily, this came up a while back[1] and we seemed to reach a > consensus that sg->offset >= PAGE_SIZE for dma_map_sg() was weird > but valid. IIRC it was only the Intel IOMMU code which failed to > handle that case appropriately (and which I fixed) - the AMD IOMMU > code always looked like it should be OK, but I'm not sure I've ever > seen definitive test results (and I don't have hardware to do so > myself). Funny that we have problems on AMD IOMMU and not with Intel IOMMU. > For dma_map_page(), length >= PAGE_SIZE should be perfectly valid > and handled correctly. The offset >= PAGE_SIZE case is a bit harder > to justify, but at the same time has less scope for the DMA API > backend to get it wrong, so either way is likely to be OK in > practice (in particular the AMD IOMMU code looks like it won't have > a problem, since its map_page() implementation converts page and > offset to a plain physical address before doing anything else). Thanks for clarify this. So my patch which do: - page = virt_to_head_page(data); + page = virt_to_page(data); offset = data - page_address(page); sg_set_page(&urb->sg[i], page, sglen, offset); should not be necessary as IOMMU driver do exactly the same internally. Are there any alignment requirement for offset for dma_map_{page,sg} ? It will work with let say sg->offset=113 or we have make sure it is aligned to some boundary. If so, what boundary ? Stanislaw
next prev parent reply other threads:[~2019-02-19 11:09 UTC|newest] Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-01-10 22:55 MT76x2U crashes XHCI driver on AMD Ryzen system Rosen Penev 2019-01-11 17:29 ` Lorenzo Bianconi 2019-01-11 19:01 ` Rosen Penev 2019-01-13 13:33 ` Lorenzo Bianconi [not found] ` <1547404075.1582.0@smtp.gmail.com> 2019-01-13 19:00 ` Lorenzo Bianconi 2019-01-14 2:20 ` Rosen Penev 2019-01-14 3:13 ` Samuel Sieb 2019-01-14 9:18 ` Lorenzo Bianconi 2019-01-14 9:22 ` Tom Psyborg 2019-01-14 20:06 ` Rosen Penev 2019-01-15 9:04 ` Lorenzo Bianconi 2019-02-18 14:37 ` Stanislaw Gruszka 2019-02-18 14:37 ` Stanislaw Gruszka 2019-02-18 15:15 ` Lorenzo Bianconi 2019-02-18 15:15 ` Lorenzo Bianconi 2019-02-18 17:01 ` Robin Murphy 2019-02-19 11:08 ` Stanislaw Gruszka [this message] 2019-02-19 11:08 ` Stanislaw Gruszka 2019-02-26 10:05 ` Joerg Roedel 2019-02-26 10:05 ` Joerg Roedel 2019-02-26 10:34 ` Stanislaw Gruszka 2019-02-26 10:44 ` Joerg Roedel 2019-02-26 11:24 ` Stanislaw Gruszka 2019-02-28 9:04 ` Stanislaw Gruszka 2019-02-28 10:42 ` Stanislaw Gruszka 2019-02-28 12:19 ` Stanislaw Gruszka 2019-02-28 13:40 ` Joerg Roedel 2019-03-04 7:10 ` Stanislaw Gruszka 2019-03-04 7:20 ` Rosen Penev 2019-03-11 8:43 ` Stanislaw Gruszka 2019-03-11 8:43 ` Stanislaw Gruszka 2019-03-11 8:43 ` Stanislaw Gruszka 2019-03-11 9:03 ` [PATCH] iommu/amd: fix sg->dma_address for sg->offset bigger than PAGE_SIZE Stanislaw Gruszka 2019-03-11 9:03 ` Stanislaw Gruszka 2019-03-11 9:03 ` Stanislaw Gruszka 2019-03-11 15:47 ` Alexander Duyck 2019-03-11 15:47 ` Alexander Duyck 2019-03-11 15:47 ` Alexander Duyck 2019-03-12 7:08 ` Stanislaw Gruszka 2019-03-12 7:08 ` Stanislaw Gruszka 2019-03-12 15:18 ` Alexander Duyck 2019-03-12 15:18 ` Alexander Duyck 2019-03-12 15:18 ` Alexander Duyck 2019-03-13 9:03 ` [PATCH v2] " Stanislaw Gruszka 2019-03-13 9:03 ` Stanislaw Gruszka 2019-03-13 9:03 ` Stanislaw Gruszka 2019-03-18 10:17 ` Joerg Roedel 2019-03-18 10:17 ` Joerg Roedel [not found] ` <20190325003822.2AB72213F2@mail.kernel.org> 2019-03-25 11:19 ` Stanislaw Gruszka 2019-03-12 7:13 ` MT76x2U crashes XHCI driver on AMD Ryzen system Stanislaw Gruszka 2019-03-12 7:13 ` Stanislaw Gruszka 2019-03-12 7:13 ` Stanislaw Gruszka
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=20190219110851.GA23223@redhat.com \ --to=sgruszka@redhat.com \ --cc=alexander.h.duyck@linux.intel.com \ --cc=iommu@lists.linux-foundation.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-wireless@vger.kernel.org \ --cc=lorenzo.bianconi@redhat.com \ --cc=robin.murphy@arm.com \ --cc=rosenp@gmail.com \ --cc=samuel@sieb.net \ /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: linkBe 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.