[PATCH v2 0/3] usb: Allow compilation on platforms where NO_DMA=y

[PATCH v2 0/3] usb: Allow compilation on platforms where NO_DMA=y

[Geert Uytterhoeven]

	Hi Greg,

Some platforms don't have DMA, but we should still be able to build USB
drivers for these platforms. They could still be used through vhci_hcd,
usbip_host, or maybe something like USB passthrough in UML from a
capable host.

This patch series allows to compile USB on platforms where NO_DMA=y.
I've been using it for quite a while to make allmodconfig/allyesconfig
build for m68k/sun3.
Vegard used this on UML to read a USB flash drive plugged into the host
machine using USB-over-IP.

Changes compared to v1:
  - Replace remaining #ifdefs by IS_ENABLED() checks,
  - Add to patch description that this actually allows using USB on UML,
  - Add Acked-by.

Thanks!

Geert Uytterhoeven (3):
  usb: core: Allow compilation on platforms where NO_DMA=y
  usb: host: Host drivers relying on DMA should depend on HAS_DMA
  usb: dwc2: USB_DWC2 should depend on HAS_DMA

 drivers/usb/core/buffer.c | 18 ++++++++++++------
 drivers/usb/core/hcd.c    | 16 ++++++++++------
 drivers/usb/dwc2/Kconfig  |  1 +
 drivers/usb/host/Kconfig  |  5 ++++-
 4 files changed, 27 insertions(+), 13 deletions(-)

-- 
1.9.1

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

[PATCH v2 1/3] usb: core: Allow compilation on platforms where NO_DMA=y

[Geert Uytterhoeven]

Some platforms don't have DMA, but we should still be able to build USB
drivers for these platforms. They could still be used through vhci_hcd,
usbip_host, or maybe something like USB passthrough in UML from a
capable host.

If NO_DMA=y:

    ERROR: "dma_pool_destroy" [drivers/usb/core/usbcore.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/core/usbcore.ko] undefined!
    ERROR: "dma_pool_free" [drivers/usb/core/usbcore.ko] undefined!
    ERROR: "dma_pool_alloc" [drivers/usb/core/usbcore.ko] undefined!
    ERROR: "dma_pool_create" [drivers/usb/core/usbcore.ko] undefined!

Add a few checks for CONFIG_HAS_DMA to fix this.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Vegard Nossum <vegard.nossum@oracle.com>
---
v2:
  - Replace remaining #ifdefs by IS_ENABLED() checks,
  - Add to patch description that this actually allows using USB on UML,
  - Add Acked-by.
---
 drivers/usb/core/buffer.c | 18 ++++++++++++------
 drivers/usb/core/hcd.c    | 16 ++++++++++------
 2 files changed, 22 insertions(+), 12 deletions(-)

diff --git a/drivers/usb/core/buffer.c b/drivers/usb/core/buffer.c
index 89f2e7765093955b..2741566ee4f25849 100644
--- a/drivers/usb/core/buffer.c
+++ b/drivers/usb/core/buffer.c
@@ -62,8 +62,9 @@ int hcd_buffer_create(struct usb_hcd *hcd)
 	char		name[16];
 	int		i, size;
 
-	if (!hcd->self.controller->dma_mask &&
-	    !(hcd->driver->flags & HCD_LOCAL_MEM))
+	if (!IS_ENABLED(CONFIG_HAS_DMA) ||
+	    (!hcd->self.controller->dma_mask &&
+	     !(hcd->driver->flags & HCD_LOCAL_MEM)))
 		return 0;
 
 	for (i = 0; i < HCD_BUFFER_POOLS; i++) {
@@ -93,6 +94,9 @@ void hcd_buffer_destroy(struct usb_hcd *hcd)
 {
 	int i;
 
+	if (!IS_ENABLED(CONFIG_HAS_DMA))
+		return;
+
 	for (i = 0; i < HCD_BUFFER_POOLS; i++) {
 		struct dma_pool *pool = hcd->pool[i];
 
@@ -119,8 +123,9 @@ void *hcd_buffer_alloc(
 	int			i;
 
 	/* some USB hosts just use PIO */
-	if (!bus->controller->dma_mask &&
-	    !(hcd->driver->flags & HCD_LOCAL_MEM)) {
+	if (!IS_ENABLED(CONFIG_HAS_DMA) ||
+	    (!bus->controller->dma_mask &&
+	     !(hcd->driver->flags & HCD_LOCAL_MEM))) {
 		*dma = ~(dma_addr_t) 0;
 		return kmalloc(size, mem_flags);
 	}
@@ -145,8 +150,9 @@ void hcd_buffer_free(
 	if (!addr)
 		return;
 
-	if (!bus->controller->dma_mask &&
-	    !(hcd->driver->flags & HCD_LOCAL_MEM)) {
+	if (!IS_ENABLED(CONFIG_HAS_DMA) ||
+	    (!bus->controller->dma_mask &&
+	     !(hcd->driver->flags & HCD_LOCAL_MEM))) {
 		kfree(addr);
 		return;
 	}
diff --git a/drivers/usb/core/hcd.c b/drivers/usb/core/hcd.c
index df0e3b92533a745f..b277977cada2b382 100644
--- a/drivers/usb/core/hcd.c
+++ b/drivers/usb/core/hcd.c
@@ -1408,7 +1408,8 @@ static void hcd_free_coherent(struct usb_bus *bus, dma_addr_t *dma_handle,
 
 void usb_hcd_unmap_urb_setup_for_dma(struct usb_hcd *hcd, struct urb *urb)
 {
-	if (urb->transfer_flags & URB_SETUP_MAP_SINGLE)
+	if (IS_ENABLED(CONFIG_HAS_DMA) &&
+	    (urb->transfer_flags & URB_SETUP_MAP_SINGLE))
 		dma_unmap_single(hcd->self.controller,
 				urb->setup_dma,
 				sizeof(struct usb_ctrlrequest),
@@ -1440,17 +1441,20 @@ void usb_hcd_unmap_urb_for_dma(struct usb_hcd *hcd, struct urb *urb)
 	usb_hcd_unmap_urb_setup_for_dma(hcd, urb);
 
 	dir = usb_urb_dir_in(urb) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
-	if (urb->transfer_flags & URB_DMA_MAP_SG)
+	if (IS_ENABLED(CONFIG_HAS_DMA) &&
+	    (urb->transfer_flags & URB_DMA_MAP_SG))
 		dma_unmap_sg(hcd->self.controller,
 				urb->sg,
 				urb->num_sgs,
 				dir);
-	else if (urb->transfer_flags & URB_DMA_MAP_PAGE)
+	else if (IS_ENABLED(CONFIG_HAS_DMA) &&
+		 (urb->transfer_flags & URB_DMA_MAP_PAGE))
 		dma_unmap_page(hcd->self.controller,
 				urb->transfer_dma,
 				urb->transfer_buffer_length,
 				dir);
-	else if (urb->transfer_flags & URB_DMA_MAP_SINGLE)
+	else if (IS_ENABLED(CONFIG_HAS_DMA) &&
+		 (urb->transfer_flags & URB_DMA_MAP_SINGLE))
 		dma_unmap_single(hcd->self.controller,
 				urb->transfer_dma,
 				urb->transfer_buffer_length,
@@ -1492,7 +1496,7 @@ int usb_hcd_map_urb_for_dma(struct usb_hcd *hcd, struct urb *urb,
 	if (usb_endpoint_xfer_control(&urb->ep->desc)) {
 		if (hcd->self.uses_pio_for_control)
 			return ret;
-		if (hcd->self.uses_dma) {
+		if (IS_ENABLED(CONFIG_HAS_DMA) && hcd->self.uses_dma) {
 			urb->setup_dma = dma_map_single(
 					hcd->self.controller,
 					urb->setup_packet,
@@ -1518,7 +1522,7 @@ int usb_hcd_map_urb_for_dma(struct usb_hcd *hcd, struct urb *urb,
 	dir = usb_urb_dir_in(urb) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
 	if (urb->transfer_buffer_length != 0
 	    && !(urb->transfer_flags & URB_NO_TRANSFER_DMA_MAP)) {
-		if (hcd->self.uses_dma) {
+		if (IS_ENABLED(CONFIG_HAS_DMA) && hcd->self.uses_dma) {
 			if (urb->num_sgs) {
 				int n;
 
-- 
1.9.1

[PATCH v2 2/3] usb: host: Host drivers relying on DMA should depend on HAS_DMA

[Geert Uytterhoeven]

If NO_DMA=y:

    ERROR: "bad_dma_ops" [drivers/usb/host/xhci-plat-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/xhci-mtk.ko] undefined!
    ERROR: "dma_pool_destroy" [drivers/usb/host/xhci-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/xhci-hcd.ko] undefined!
    ERROR: "dma_pool_free" [drivers/usb/host/xhci-hcd.ko] undefined!
    ERROR: "dma_pool_alloc" [drivers/usb/host/xhci-hcd.ko] undefined!
    ERROR: "dma_pool_create" [drivers/usb/host/xhci-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/ohci-platform.ko] undefined!
    ERROR: "dma_pool_destroy" [drivers/usb/host/ohci-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/ohci-hcd.ko] undefined!
    ERROR: "dma_pool_free" [drivers/usb/host/ohci-hcd.ko] undefined!
    ERROR: "dma_pool_alloc" [drivers/usb/host/ohci-hcd.ko] undefined!
    ERROR: "dma_pool_create" [drivers/usb/host/ohci-hcd.ko] undefined!
    ERROR: "dma_pool_create" [drivers/usb/host/fotg210-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/fotg210-hcd.ko] undefined!
    ERROR: "dma_pool_destroy" [drivers/usb/host/fotg210-hcd.ko] undefined!
    ERROR: "dma_pool_alloc" [drivers/usb/host/fotg210-hcd.ko] undefined!
    ERROR: "dma_pool_free" [drivers/usb/host/fotg210-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/ehci-platform.ko] undefined!
    ERROR: "dma_pool_destroy" [drivers/usb/host/ehci-hcd.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/host/ehci-hcd.ko] undefined!
    ERROR: "dma_pool_free" [drivers/usb/host/ehci-hcd.ko] undefined!
    ERROR: "dma_pool_alloc" [drivers/usb/host/ehci-hcd.ko] undefined!
    ERROR: "dma_pool_create" [drivers/usb/host/ehci-hcd.ko] undefined!

Add dependencies on HAS_DMA to fix this.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
---
v2:
  - Reword one-line summary.
---
 drivers/usb/host/Kconfig | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/host/Kconfig b/drivers/usb/host/Kconfig
index 1f117c360ebbba33..573789698474ad8d 100644
--- a/drivers/usb/host/Kconfig
+++ b/drivers/usb/host/Kconfig
@@ -17,6 +17,7 @@ config USB_C67X00_HCD
 
 config USB_XHCI_HCD
 	tristate "xHCI HCD (USB 3.0) support"
+	depends on HAS_DMA
 	---help---
 	  The eXtensible Host Controller Interface (xHCI) is standard for USB 3.0
 	  "SuperSpeed" host controller hardware.
@@ -70,6 +71,7 @@ endif # USB_XHCI_HCD
 
 config USB_EHCI_HCD
 	tristate "EHCI HCD (USB 2.0) support"
+	depends on HAS_DMA
 	---help---
 	  The Enhanced Host Controller Interface (EHCI) is standard for USB 2.0
 	  "high speed" (480 Mbit/sec, 60 Mbyte/sec) host controller hardware.
@@ -361,7 +363,7 @@ config USB_ISP1362_HCD
 
 config USB_FOTG210_HCD
 	tristate "FOTG210 HCD support"
-	depends on USB
+	depends on USB && HAS_DMA
 	---help---
 	  Faraday FOTG210 is an OTG controller which can be configured as
 	  an USB2.0 host. It is designed to meet USB2.0 EHCI specification
@@ -383,6 +385,7 @@ config USB_MAX3421_HCD
 
 config USB_OHCI_HCD
 	tristate "OHCI HCD (USB 1.1) support"
+	depends on HAS_DMA
 	---help---
 	  The Open Host Controller Interface (OHCI) is a standard for accessing
 	  USB 1.1 host controller hardware.  It does more in hardware than Intel's
-- 
1.9.1

[PATCH v2 3/3] usb: dwc2: USB_DWC2 should depend on HAS_DMA

[Geert Uytterhoeven]

If NO_DMA=y:

    ERROR: "usb_gadget_map_request" [drivers/usb/dwc2/dwc2.ko] undefined!
    ERROR: "usb_gadget_unmap_request" [drivers/usb/dwc2/dwc2.ko] undefined!
    ERROR: "bad_dma_ops" [drivers/usb/dwc2/dwc2.ko] undefined!

Add a dependency on HAS_DMA to fix this.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: John Youn <johnyoun@synopsys.com>
---
v2:
  - Add Acked-by.
---
 drivers/usb/dwc2/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/usb/dwc2/Kconfig b/drivers/usb/dwc2/Kconfig
index fd95ba6ec317fdac..f0decc0d69b5197d 100644
--- a/drivers/usb/dwc2/Kconfig
+++ b/drivers/usb/dwc2/Kconfig
@@ -1,5 +1,6 @@
 config USB_DWC2
 	tristate "DesignWare USB2 DRD Core Support"
+	depends on HAS_DMA
 	depends on USB || USB_GADGET
 	help
 	  Say Y here if your system has a Dual Role Hi-Speed USB
-- 
1.9.1

Re: [PATCH v2 1/3] usb: core: Allow compilation on platforms where NO_DMA=y

[Greg Kroah-Hartman]

On Tue, Feb 16, 2016 at 04:10:57PM +0100, Geert Uytterhoeven wrote:
> Some platforms don't have DMA, but we should still be able to build USB
> drivers for these platforms. They could still be used through vhci_hcd,
> usbip_host, or maybe something like USB passthrough in UML from a
> capable host.
> 
> If NO_DMA=y:
> 
>     ERROR: "dma_pool_destroy" [drivers/usb/core/usbcore.ko] undefined!
>     ERROR: "bad_dma_ops" [drivers/usb/core/usbcore.ko] undefined!
>     ERROR: "dma_pool_free" [drivers/usb/core/usbcore.ko] undefined!
>     ERROR: "dma_pool_alloc" [drivers/usb/core/usbcore.ko] undefined!
>     ERROR: "dma_pool_create" [drivers/usb/core/usbcore.ko] undefined!
> 
> Add a few checks for CONFIG_HAS_DMA to fix this.
> 
> Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Acked-by: Vegard Nossum <vegard.nossum@oracle.com>
> ---
> v2:
>   - Replace remaining #ifdefs by IS_ENABLED() checks,
>   - Add to patch description that this actually allows using USB on UML,
>   - Add Acked-by.

This patch didn't apply to my tree, can you rebase it against usb-next
of usb.git and resend?

thanks,

greg k-h

Re: [PATCH v2 1/3] usb: core: Allow compilation on platforms where NO_DMA=y

[Geert Uytterhoeven]

Hi Greg,

On Sun, Feb 21, 2016 at 5:23 AM, Greg Kroah-Hartman
<gregkh@linuxfoundation.org> wrote:
> On Tue, Feb 16, 2016 at 04:10:57PM +0100, Geert Uytterhoeven wrote:
>> Some platforms don't have DMA, but we should still be able to build USB
>> drivers for these platforms. They could still be used through vhci_hcd,
>> usbip_host, or maybe something like USB passthrough in UML from a
>> capable host.
>>
>> If NO_DMA=y:
>>
>>     ERROR: "dma_pool_destroy" [drivers/usb/core/usbcore.ko] undefined!
>>     ERROR: "bad_dma_ops" [drivers/usb/core/usbcore.ko] undefined!
>>     ERROR: "dma_pool_free" [drivers/usb/core/usbcore.ko] undefined!
>>     ERROR: "dma_pool_alloc" [drivers/usb/core/usbcore.ko] undefined!
>>     ERROR: "dma_pool_create" [drivers/usb/core/usbcore.ko] undefined!
>>
>> Add a few checks for CONFIG_HAS_DMA to fix this.
>>
>> Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
>> Acked-by: Vegard Nossum <vegard.nossum@oracle.com>
>> ---
>> v2:
>>   - Replace remaining #ifdefs by IS_ENABLED() checks,
>>   - Add to patch description that this actually allows using USB on UML,
>>   - Add Acked-by.
>
> This patch didn't apply to my tree, can you rebase it against usb-next
> of usb.git and resend?

Are you sure it's this one that didn't apply? It's already in usb-testing?

"[2/3] usb: host: Host drivers relying on DMA should depend on HAS_DMA"
doesn't apply, as your tree gained some HAS_IOMEM dependencies, so I'll
resend.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH v2 1/3] usb: core: Allow compilation on platforms where NO_DMA=y

[Greg Kroah-Hartman]

On Sun, Feb 21, 2016 at 10:46:59PM +0100, Geert Uytterhoeven wrote:
> Hi Greg,
> 
> On Sun, Feb 21, 2016 at 5:23 AM, Greg Kroah-Hartman
> <gregkh@linuxfoundation.org> wrote:
> > On Tue, Feb 16, 2016 at 04:10:57PM +0100, Geert Uytterhoeven wrote:
> >> Some platforms don't have DMA, but we should still be able to build USB
> >> drivers for these platforms. They could still be used through vhci_hcd,
> >> usbip_host, or maybe something like USB passthrough in UML from a
> >> capable host.
> >>
> >> If NO_DMA=y:
> >>
> >>     ERROR: "dma_pool_destroy" [drivers/usb/core/usbcore.ko] undefined!
> >>     ERROR: "bad_dma_ops" [drivers/usb/core/usbcore.ko] undefined!
> >>     ERROR: "dma_pool_free" [drivers/usb/core/usbcore.ko] undefined!
> >>     ERROR: "dma_pool_alloc" [drivers/usb/core/usbcore.ko] undefined!
> >>     ERROR: "dma_pool_create" [drivers/usb/core/usbcore.ko] undefined!
> >>
> >> Add a few checks for CONFIG_HAS_DMA to fix this.
> >>
> >> Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
> >> Acked-by: Vegard Nossum <vegard.nossum@oracle.com>
> >> ---
> >> v2:
> >>   - Replace remaining #ifdefs by IS_ENABLED() checks,
> >>   - Add to patch description that this actually allows using USB on UML,
> >>   - Add Acked-by.
> >
> > This patch didn't apply to my tree, can you rebase it against usb-next
> > of usb.git and resend?
> 
> Are you sure it's this one that didn't apply? It's already in usb-testing?
> 
> "[2/3] usb: host: Host drivers relying on DMA should depend on HAS_DMA"
> doesn't apply, as your tree gained some HAS_IOMEM dependencies, so I'll
> resend.

Thanks, I think you were right.

greg k-h