From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Michael Ellerman <michael@ellerman.id.au>
Cc: linuxppc-dev@ozlabs.org, cbe-oss-dev@ozlabs.org
Subject: Re: [PATCH 4/4] Avoid DMA exception when using axon_msi with IOMMU
Date: Tue, 05 Feb 2008 11:21:25 +1100 [thread overview]
Message-ID: <1202170885.7079.30.camel@pasglop> (raw)
In-Reply-To: <f1b51505f34145d7b096fef167c14c4409f0baf6.1201240265.git.michael@ellerman.id.au>
On Fri, 2008-01-25 at 16:59 +1100, Michael Ellerman wrote:
> There's a brown-paper-bag bug in axon_msi, we pass the address of our
> FIFO directly to the hardware, without DMA mapping it. This leads to
> DMA exceptions if you enable MSI & the IOMMU.
>
> The fix is to correctly DMA map the fifo, dma_alloc_coherent() does
> what we want - and we need to track the virt & phys addresses.
>
> Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> ---
> arch/powerpc/platforms/cell/axon_msi.c | 21 ++++++++++-----------
> 1 files changed, 10 insertions(+), 11 deletions(-)
>
> diff --git a/arch/powerpc/platforms/cell/axon_msi.c b/arch/powerpc/platforms/cell/axon_msi.c
> index ea3dc8c..b9a97c4 100644
> --- a/arch/powerpc/platforms/cell/axon_msi.c
> +++ b/arch/powerpc/platforms/cell/axon_msi.c
> @@ -65,7 +65,8 @@
>
> struct axon_msic {
> struct irq_host *irq_host;
> - __le32 *fifo;
> + __le32 *fifo_virt;
> + dma_addr_t fifo_phys;
> dcr_host_t dcr_host;
> u32 read_offset;
> };
> @@ -91,7 +92,7 @@ static void axon_msi_cascade(unsigned int irq, struct irq_desc *desc)
>
> while (msic->read_offset != write_offset) {
> idx = msic->read_offset / sizeof(__le32);
> - msi = le32_to_cpu(msic->fifo[idx]);
> + msi = le32_to_cpu(msic->fifo_virt[idx]);
> msi &= 0xFFFF;
>
> pr_debug("axon_msi: woff %x roff %x msi %x\n",
> @@ -306,7 +307,6 @@ static int axon_msi_shutdown(struct of_device *device)
> static int axon_msi_probe(struct of_device *device,
> const struct of_device_id *device_id)
> {
> - struct page *page;
> struct device_node *dn = device->node;
> struct axon_msic *msic;
> unsigned int virq;
> @@ -338,16 +338,14 @@ static int axon_msi_probe(struct of_device *device,
> goto out_free_msic;
> }
>
> - page = alloc_pages_node(of_node_to_nid(dn), GFP_KERNEL,
> - get_order(MSIC_FIFO_SIZE_BYTES));
> - if (!page) {
> + msic->fifo_virt = dma_alloc_coherent(&device->dev, MSIC_FIFO_SIZE_BYTES,
> + &msic->fifo_phys, GFP_KERNEL);
> + if (!msic->fifo_virt) {
> printk(KERN_ERR "axon_msi: couldn't allocate fifo for %s\n",
> dn->full_name);
> goto out_free_msic;
> }
>
> - msic->fifo = page_address(page);
> -
> msic->irq_host = irq_alloc_host(of_node_get(dn), IRQ_HOST_MAP_NOMAP,
> NR_IRQS, &msic_host_ops, 0);
> if (!msic->irq_host) {
> @@ -370,9 +368,9 @@ static int axon_msi_probe(struct of_device *device,
> pr_debug("axon_msi: irq 0x%x setup for axon_msi\n", virq);
>
> /* Enable the MSIC hardware */
> - msic_dcr_write(msic, MSIC_BASE_ADDR_HI_REG, (u64)msic->fifo >> 32);
> + msic_dcr_write(msic, MSIC_BASE_ADDR_HI_REG, msic->fifo_phys >> 32);
> msic_dcr_write(msic, MSIC_BASE_ADDR_LO_REG,
> - (u64)msic->fifo & 0xFFFFFFFF);
> + msic->fifo_phys & 0xFFFFFFFF);
> msic_dcr_write(msic, MSIC_CTRL_REG,
> MSIC_CTRL_IRQ_ENABLE | MSIC_CTRL_ENABLE |
> MSIC_CTRL_FIFO_SIZE);
> @@ -390,7 +388,8 @@ static int axon_msi_probe(struct of_device *device,
> out_free_host:
> kfree(msic->irq_host);
> out_free_fifo:
> - __free_pages(virt_to_page(msic->fifo), get_order(MSIC_FIFO_SIZE_BYTES));
> + dma_free_coherent(&device->dev, MSIC_FIFO_SIZE_BYTES, msic->fifo_virt,
> + msic->fifo_phys);
> out_free_msic:
> kfree(msic);
> out:
next prev parent reply other threads:[~2008-02-05 0:21 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-01-25 5:59 [PATCH 1/4] Search for and publish cell OF platform devices earlier Michael Ellerman
2008-01-25 5:59 ` [PATCH 2/4] Create and hook up of_platform_device_shutdown Michael Ellerman
2008-02-05 0:19 ` Benjamin Herrenschmidt
2008-01-25 5:59 ` [PATCH 4/4] Avoid DMA exception when using axon_msi with IOMMU Michael Ellerman
2008-02-05 0:21 ` Benjamin Herrenschmidt [this message]
2008-01-25 5:59 ` [PATCH 3/4] Convert axon_msi to an of_platform driver Michael Ellerman
2008-02-05 0:20 ` Benjamin Herrenschmidt
2008-02-05 0:19 ` [PATCH 1/4] Search for and publish cell OF platform devices earlier Benjamin Herrenschmidt
-- strict thread matches above, loose matches on Subject: below --
2008-01-22 11:04 Michael Ellerman
2008-01-22 11:04 ` [PATCH 4/4] Avoid DMA exception when using axon_msi with IOMMU Michael Ellerman
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=1202170885.7079.30.camel@pasglop \
--to=benh@kernel.crashing.org \
--cc=cbe-oss-dev@ozlabs.org \
--cc=linuxppc-dev@ozlabs.org \
--cc=michael@ellerman.id.au \
/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).