[v4,01/17] PCI: dwc: Use the DMA-API to get the MSI address

Message ID 20171103134722.5532-2-niklas.cassel@axis.com
State Needs Review / ACK
Headers show
Series
  • dwc MSI fixes, ARTPEC-6 EP mode support, ARTPEC-7 SoC support
Related show

Commit Message

Niklas Cassel Nov. 3, 2017, 1:47 p.m.
Use the DMA-API to get the MSI address. This address will be written to
our PCI config space and to the register which determines which AXI
address the DWC IP will spoof for incoming MSI irqs.

Since it is a PCIe endpoint device, rather than the CPU, that is supposed
to write to the MSI address, the proper way to get the MSI address is by
using the DMA API, not by using virt_to_phys().

Using virt_to_phys() might work on some systems, but using the DMA API
should work on all systems.

This is essentially the same thing as allocating a buffer in a driver
to which the endpoint will write to. To do this, we use the DMA API.

Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
---
 drivers/pci/dwc/pcie-designware-host.c | 15 ++++++++++++---
 drivers/pci/dwc/pcie-designware.h      |  3 ++-
 2 files changed, 14 insertions(+), 4 deletions(-)

Comments

Bjorn Helgaas Nov. 8, 2017, 12:56 a.m. | #1
On Fri, Nov 03, 2017 at 02:47:05PM +0100, Niklas Cassel wrote:
> Use the DMA-API to get the MSI address. This address will be written to
> our PCI config space and to the register which determines which AXI
> address the DWC IP will spoof for incoming MSI irqs.
> 
> Since it is a PCIe endpoint device, rather than the CPU, that is supposed
> to write to the MSI address, the proper way to get the MSI address is by
> using the DMA API, not by using virt_to_phys().
> 
> Using virt_to_phys() might work on some systems, but using the DMA API
> should work on all systems.
> 
> This is essentially the same thing as allocating a buffer in a driver
> to which the endpoint will write to. To do this, we use the DMA API.
> 
> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>

I'm expecting Jingoo and/or Joao to chime in and ack the
DesignWare-related patches.  I think all the others are in
pretty good shape.

> ---
>  drivers/pci/dwc/pcie-designware-host.c | 15 ++++++++++++---
>  drivers/pci/dwc/pcie-designware.h      |  3 ++-
>  2 files changed, 14 insertions(+), 4 deletions(-)
> ...
Joao Pinto Nov. 8, 2017, 12:45 p.m. | #2
Hello to all,

Às 12:56 AM de 11/8/2017, Bjorn Helgaas escreveu:
> On Fri, Nov 03, 2017 at 02:47:05PM +0100, Niklas Cassel wrote:
>> Use the DMA-API to get the MSI address. This address will be written to
>> our PCI config space and to the register which determines which AXI
>> address the DWC IP will spoof for incoming MSI irqs.
>>
>> Since it is a PCIe endpoint device, rather than the CPU, that is supposed
>> to write to the MSI address, the proper way to get the MSI address is by
>> using the DMA API, not by using virt_to_phys().
>>
>> Using virt_to_phys() might work on some systems, but using the DMA API
>> should work on all systems.
>>
>> This is essentially the same thing as allocating a buffer in a driver
>> to which the endpoint will write to. To do this, we use the DMA API.
>>
>> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> 
> I'm expecting Jingoo and/or Joao to chime in and ack the
> DesignWare-related patches.  I think all the others are in
> pretty good shape.
> 
>> ---
>>  drivers/pci/dwc/pcie-designware-host.c | 15 ++++++++++++---
>>  drivers/pci/dwc/pcie-designware.h      |  3 ++-
>>  2 files changed, 14 insertions(+), 4 deletions(-)
>> ...

Let me test this patch-set in my setup first!
I will give feedback until Friday.

Thanks,
Joao
Bjorn Helgaas Nov. 8, 2017, 5:27 p.m. | #3
On Wed, Nov 08, 2017 at 12:45:49PM +0000, Joao Pinto wrote:
> Hello to all,
> 
> Às 12:56 AM de 11/8/2017, Bjorn Helgaas escreveu:
> > On Fri, Nov 03, 2017 at 02:47:05PM +0100, Niklas Cassel wrote:
> >> Use the DMA-API to get the MSI address. This address will be written to
> >> our PCI config space and to the register which determines which AXI
> >> address the DWC IP will spoof for incoming MSI irqs.
> >>
> >> Since it is a PCIe endpoint device, rather than the CPU, that is supposed
> >> to write to the MSI address, the proper way to get the MSI address is by
> >> using the DMA API, not by using virt_to_phys().
> >>
> >> Using virt_to_phys() might work on some systems, but using the DMA API
> >> should work on all systems.
> >>
> >> This is essentially the same thing as allocating a buffer in a driver
> >> to which the endpoint will write to. To do this, we use the DMA API.
> >>
> >> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
> > 
> > I'm expecting Jingoo and/or Joao to chime in and ack the
> > DesignWare-related patches.  I think all the others are in
> > pretty good shape.
> > 
> >> ---
> >>  drivers/pci/dwc/pcie-designware-host.c | 15 ++++++++++++---
> >>  drivers/pci/dwc/pcie-designware.h      |  3 ++-
> >>  2 files changed, 14 insertions(+), 4 deletions(-)
> >> ...
> 
> Let me test this patch-set in my setup first!
> I will give feedback until Friday.

OK, thanks.  This may push this series out until v4.16 because the
v4.15 merge window will probably open Sunday, and I don't like to add
code during the window and merge it immediately.

Bjorn
Niklas Cassel Nov. 9, 2017, 9:20 a.m. | #4
On 11/08/2017 06:27 PM, Bjorn Helgaas wrote:
> On Wed, Nov 08, 2017 at 12:45:49PM +0000, Joao Pinto wrote:
>> Hello to all,
>>
>> Às 12:56 AM de 11/8/2017, Bjorn Helgaas escreveu:
>>> On Fri, Nov 03, 2017 at 02:47:05PM +0100, Niklas Cassel wrote:
>>>> Use the DMA-API to get the MSI address. This address will be written to
>>>> our PCI config space and to the register which determines which AXI
>>>> address the DWC IP will spoof for incoming MSI irqs.
>>>>
>>>> Since it is a PCIe endpoint device, rather than the CPU, that is supposed
>>>> to write to the MSI address, the proper way to get the MSI address is by
>>>> using the DMA API, not by using virt_to_phys().
>>>>
>>>> Using virt_to_phys() might work on some systems, but using the DMA API
>>>> should work on all systems.
>>>>
>>>> This is essentially the same thing as allocating a buffer in a driver
>>>> to which the endpoint will write to. To do this, we use the DMA API.
>>>>
>>>> Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
>>>
>>> I'm expecting Jingoo and/or Joao to chime in and ack the
>>> DesignWare-related patches.  I think all the others are in
>>> pretty good shape.
>>>
>>>> ---
>>>>  drivers/pci/dwc/pcie-designware-host.c | 15 ++++++++++++---
>>>>  drivers/pci/dwc/pcie-designware.h      |  3 ++-
>>>>  2 files changed, 14 insertions(+), 4 deletions(-)
>>>> ...
>>
>> Let me test this patch-set in my setup first!
>> I will give feedback until Friday.
> 
> OK, thanks.  This may push this series out until v4.16 because the
> v4.15 merge window will probably open Sunday, and I don't like to add
> code during the window and merge it immediately.
> 

I suspected as much. The general rule seems to be that
a patch has to be in linux-next for at least 2 weeks
before sending it to Linus.

No worries, there is not really a rush.

Regards,
Niklas

Patch

diff --git a/drivers/pci/dwc/pcie-designware-host.c b/drivers/pci/dwc/pcie-designware-host.c
index 81e2157a7cfb..33b52fe98a01 100644
--- a/drivers/pci/dwc/pcie-designware-host.c
+++ b/drivers/pci/dwc/pcie-designware-host.c
@@ -83,10 +83,19 @@  irqreturn_t dw_handle_msi_irq(struct pcie_port *pp)
 
 void dw_pcie_msi_init(struct pcie_port *pp)
 {
+	struct dw_pcie *pci = to_dw_pcie_from_pp(pp);
+	struct device *dev = pci->dev;
+	struct page *page;
 	u64 msi_target;
 
-	pp->msi_data = __get_free_pages(GFP_KERNEL, 0);
-	msi_target = virt_to_phys((void *)pp->msi_data);
+	page = alloc_page(GFP_KERNEL | GFP_DMA32);
+	pp->msi_data = dma_map_page(dev, page, 0, PAGE_SIZE, DMA_FROM_DEVICE);
+	if (dma_mapping_error(dev, pp->msi_data)) {
+		dev_err(dev, "failed to map MSI data\n");
+		__free_page(page);
+		return;
+	}
+	msi_target = (u64)pp->msi_data;
 
 	/* program the msi_data */
 	dw_pcie_wr_own_conf(pp, PCIE_MSI_ADDR_LO, 4,
@@ -187,7 +196,7 @@  static void dw_msi_setup_msg(struct pcie_port *pp, unsigned int irq, u32 pos)
 	if (pp->ops->get_msi_addr)
 		msi_target = pp->ops->get_msi_addr(pp);
 	else
-		msi_target = virt_to_phys((void *)pp->msi_data);
+		msi_target = (u64)pp->msi_data;
 
 	msg.address_lo = (u32)(msi_target & 0xffffffff);
 	msg.address_hi = (u32)(msi_target >> 32 & 0xffffffff);
diff --git a/drivers/pci/dwc/pcie-designware.h b/drivers/pci/dwc/pcie-designware.h
index e5d9d77b778e..ecdede68522a 100644
--- a/drivers/pci/dwc/pcie-designware.h
+++ b/drivers/pci/dwc/pcie-designware.h
@@ -14,6 +14,7 @@ 
 #ifndef _PCIE_DESIGNWARE_H
 #define _PCIE_DESIGNWARE_H
 
+#include <linux/dma-mapping.h>
 #include <linux/irq.h>
 #include <linux/msi.h>
 #include <linux/pci.h>
@@ -168,7 +169,7 @@  struct pcie_port {
 	const struct dw_pcie_host_ops *ops;
 	int			msi_irq;
 	struct irq_domain	*irq_domain;
-	unsigned long		msi_data;
+	dma_addr_t		msi_data;
 	DECLARE_BITMAP(msi_irq_in_use, MAX_MSI_IRQS);
 };