linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Scott Wood <scottwood@freescale.com>
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Scott Wood <scottwood@freescale.com>,
	linuxppc-dev@lists.ozlabs.org,
	Shaohui Xie <Shaohui.Xie@freescale.com>
Subject: [PATCH 2/4] powerpc/64: Honor swiotlb limit in coherent allocations
Date: Fri, 8 Aug 2014 18:40:43 -0500	[thread overview]
Message-ID: <1407541245-27617-2-git-send-email-scottwood@freescale.com> (raw)
In-Reply-To: <1407541245-27617-1-git-send-email-scottwood@freescale.com>

FSL PCI cannot directly address the whole lower 4 GiB due to
conflicts with PCICSRBAR and outbound windows, and thus
max_direct_dma_addr is less than 4GiB.  Honor that limit in
dma_direct_alloc_coherent().

Note that setting the DMA mask to 31 bits is not an option, since many
PCI drivers would fail if we reject 32-bit DMA in dma_supported(), and
we have no control over the setting of coherent_dma_mask if
dma_supported() returns true.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Cc: Shaohui Xie <Shaohui.Xie@freescale.com>
---
 arch/powerpc/kernel/dma.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/kernel/dma.c b/arch/powerpc/kernel/dma.c
index dfd99ef..a7b0156 100644
--- a/arch/powerpc/kernel/dma.c
+++ b/arch/powerpc/kernel/dma.c
@@ -15,6 +15,7 @@
 #include <asm/vio.h>
 #include <asm/bug.h>
 #include <asm/machdep.h>
+#include <asm/swiotlb.h>
 
 /*
  * Generic direct DMA implementation
@@ -25,6 +26,18 @@
  * default the offset is PCI_DRAM_OFFSET.
  */
 
+static u64 __maybe_unused get_pfn_limit(struct device *dev)
+{
+	u64 pfn = (dev->coherent_dma_mask >> PAGE_SHIFT) + 1;
+	struct dev_archdata __maybe_unused *sd = &dev->archdata;
+
+#ifdef CONFIG_SWIOTLB
+	if (sd->max_direct_dma_addr && sd->dma_ops == &swiotlb_dma_ops)
+		pfn = min_t(u64, pfn, sd->max_direct_dma_addr >> PAGE_SHIFT);
+#endif
+
+	return pfn;
+}
 
 void *dma_direct_alloc_coherent(struct device *dev, size_t size,
 				dma_addr_t *dma_handle, gfp_t flag,
@@ -40,7 +53,7 @@ void *dma_direct_alloc_coherent(struct device *dev, size_t size,
 #else
 	struct page *page;
 	int node = dev_to_node(dev);
-	u64 pfn = (dev->coherent_dma_mask >> PAGE_SHIFT) + 1;
+	u64 pfn = get_pfn_limit(dev);
 	int zone;
 
 	zone = dma_pfn_limit_to_zone(pfn);
-- 
1.9.1

  reply	other threads:[~2014-08-08 23:41 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-08-08 23:40 [PATCH 1/4] powerpc: Dynamic DMA zone limits Scott Wood
2014-08-08 23:40 ` Scott Wood [this message]
2014-08-08 23:40 ` [PATCH 3/4] powerpc/64: Limit ZONE_DMA32 to 4GiB in swiotlb_detect_4g() Scott Wood
2014-08-08 23:40 ` [PATCH 4/4] powerpc/fsl-pci: Limit ZONE_DMA32 to 2GiB on 64-bit platforms Scott Wood
2014-10-13  7:14 ` [PATCH 1/4] powerpc: Dynamic DMA zone limits Anton Blanchard
2014-10-13  7:30   ` Benjamin Herrenschmidt
2014-10-13  9:00   ` Michael Ellerman
2014-10-14  7:39     ` Scott Wood
2014-10-14  7:57       ` Benjamin Herrenschmidt
2014-10-14  9:44         ` Scott Wood

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=1407541245-27617-2-git-send-email-scottwood@freescale.com \
    --to=scottwood@freescale.com \
    --cc=Shaohui.Xie@freescale.com \
    --cc=benh@kernel.crashing.org \
    --cc=linuxppc-dev@lists.ozlabs.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;
as well as URLs for NNTP newsgroup(s).