[PATCH] usb: gadget: aspeed: set port_dev dma mask

[PATCH] usb: gadget: aspeed: set port_dev dma mask

[rentao.bupt]

From: Tao Ren <rentao.bupt@gmail.com>

Set aspeed-usb vhub port_dev's dma mask to pass the dma_mask test in
"dma_map_page_attrs" function, and the dma_mask test was added in
'commit f959dcd6ddfd ("dma-direct: Fix potential NULL pointer
dereference")'.

Below is the backtrace without the patch:
[<80106550>] show_stack+0x20/0x24
[<80106868>] dump_stack+0x28/0x30
[<80823540>] __warn+0xfc/0x110
[<8011ac30>] warn_slowpath_fmt+0xb0/0xc0
[<8011ad44>] dma_map_page_attrs+0x24c/0x314
[<8016a27c>] usb_gadget_map_request_by_dev+0x100/0x1e4
[<805cedd8>] usb_gadget_map_request+0x1c/0x20
[<805cefbc>] ast_vhub_epn_queue+0xa0/0x1d8
[<7f02f710>] usb_ep_queue+0x48/0xc4
[<805cd3e8>] ecm_do_notify+0xf8/0x248
[<7f145920>] ecm_set_alt+0xc8/0x1d0
[<7f145c34>] composite_setup+0x680/0x1d30
[<7f00deb8>] ast_vhub_ep0_handle_setup+0xa4/0x1bc
[<7f02ee94>] ast_vhub_dev_irq+0x58/0x84
[<7f0309e0>] ast_vhub_irq+0xb0/0x1c8
[<7f02e118>] __handle_irq_event_percpu+0x50/0x19c
[<8015e5bc>] handle_irq_event_percpu+0x38/0x8c
[<8015e758>] handle_irq_event+0x38/0x4c

Signed-off-by: Tao Ren <rentao.bupt@gmail.com>
---
 drivers/usb/gadget/udc/aspeed-vhub/dev.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/usb/gadget/udc/aspeed-vhub/dev.c b/drivers/usb/gadget/udc/aspeed-vhub/dev.c
index d268306a7bfe..9eb3904a6ff9 100644
--- a/drivers/usb/gadget/udc/aspeed-vhub/dev.c
+++ b/drivers/usb/gadget/udc/aspeed-vhub/dev.c
@@ -569,6 +569,7 @@ int ast_vhub_init_dev(struct ast_vhub *vhub, unsigned int idx)
 	device_initialize(d->port_dev);
 	d->port_dev->release = ast_vhub_dev_release;
 	d->port_dev->parent = parent;
+	d->port_dev->dma_mask = parent->dma_mask;
 	dev_set_name(d->port_dev, "%s:p%d", dev_name(parent), idx + 1);
 	rc = device_add(d->port_dev);
 	if (rc)
-- 
2.17.1

Re: [PATCH] usb: gadget: aspeed: set port_dev dma mask

[Robin Murphy]

On 2021-03-26 07:02, rentao.bupt@gmail.com wrote:
> From: Tao Ren <rentao.bupt@gmail.com>
> 
> Set aspeed-usb vhub port_dev's dma mask to pass the dma_mask test in
> "dma_map_page_attrs" function, and the dma_mask test was added in
> 'commit f959dcd6ddfd ("dma-direct: Fix potential NULL pointer
> dereference")'.
> 
> Below is the backtrace without the patch:
> [<80106550>] show_stack+0x20/0x24
> [<80106868>] dump_stack+0x28/0x30
> [<80823540>] __warn+0xfc/0x110
> [<8011ac30>] warn_slowpath_fmt+0xb0/0xc0
> [<8011ad44>] dma_map_page_attrs+0x24c/0x314
> [<8016a27c>] usb_gadget_map_request_by_dev+0x100/0x1e4
> [<805cedd8>] usb_gadget_map_request+0x1c/0x20
> [<805cefbc>] ast_vhub_epn_queue+0xa0/0x1d8
> [<7f02f710>] usb_ep_queue+0x48/0xc4
> [<805cd3e8>] ecm_do_notify+0xf8/0x248
> [<7f145920>] ecm_set_alt+0xc8/0x1d0
> [<7f145c34>] composite_setup+0x680/0x1d30
> [<7f00deb8>] ast_vhub_ep0_handle_setup+0xa4/0x1bc
> [<7f02ee94>] ast_vhub_dev_irq+0x58/0x84
> [<7f0309e0>] ast_vhub_irq+0xb0/0x1c8
> [<7f02e118>] __handle_irq_event_percpu+0x50/0x19c
> [<8015e5bc>] handle_irq_event_percpu+0x38/0x8c
> [<8015e758>] handle_irq_event+0x38/0x4c
> 
> Signed-off-by: Tao Ren <rentao.bupt@gmail.com>
> ---
>   drivers/usb/gadget/udc/aspeed-vhub/dev.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/usb/gadget/udc/aspeed-vhub/dev.c b/drivers/usb/gadget/udc/aspeed-vhub/dev.c
> index d268306a7bfe..9eb3904a6ff9 100644
> --- a/drivers/usb/gadget/udc/aspeed-vhub/dev.c
> +++ b/drivers/usb/gadget/udc/aspeed-vhub/dev.c
> @@ -569,6 +569,7 @@ int ast_vhub_init_dev(struct ast_vhub *vhub, unsigned int idx)
>   	device_initialize(d->port_dev);
>   	d->port_dev->release = ast_vhub_dev_release;
>   	d->port_dev->parent = parent;
> +	d->port_dev->dma_mask = parent->dma_mask;

This might happen to work out, but is far from correct. Just wait until 
you try it on a platform where the USB controller is behind an IOMMU...

It looks like something is more fundamentally wrong here - the device 
passed to DMA API calls must be the actual hardware device performing 
the DMA, which in USB-land I believe means the controller's sysdev.

Robin.

>   	dev_set_name(d->port_dev, "%s:p%d", dev_name(parent), idx + 1);
>   	rc = device_add(d->port_dev);
>   	if (rc)
> 

Re: [PATCH] usb: gadget: aspeed: set port_dev dma mask

[Christoph Hellwig]

On Fri, Mar 26, 2021 at 12:03:03PM +0000, Robin Murphy wrote:
> This might happen to work out, but is far from correct. Just wait until you 
> try it on a platform where the USB controller is behind an IOMMU...
>
> It looks like something is more fundamentally wrong here - the device 
> passed to DMA API calls must be the actual hardware device performing the 
> DMA, which in USB-land I believe means the controller's sysdev.

The shiny new usb_intf_get_dma_device API provides the device to use.

Re: [PATCH] usb: gadget: aspeed: set port_dev dma mask

[Tao Ren]

On Fri, Mar 26, 2021 at 01:05:26PM +0100, Christoph Hellwig wrote:
> On Fri, Mar 26, 2021 at 12:03:03PM +0000, Robin Murphy wrote:
> > This might happen to work out, but is far from correct. Just wait until you 
> > try it on a platform where the USB controller is behind an IOMMU...
> >
> > It looks like something is more fundamentally wrong here - the device 
> > passed to DMA API calls must be the actual hardware device performing the 
> > DMA, which in USB-land I believe means the controller's sysdev.
> 
> The shiny new usb_intf_get_dma_device API provides the device to use.

Thanks Robin and Christoph for the feedback.

If I understand correctly, usb_intf_get_dma_device API is mainly for usb
host drivers? I just found usb_gadget_map_request_by_dev API: does it
make sense to replace usb_gadget_map_request with
usb_gadget_map_request_by_dev so we can pass the actual DMA-capable
hardware device (aspeed-vhub platform device) to the API?


Cheers,

Tao

Re: [PATCH] usb: gadget: aspeed: set port_dev dma mask

[Christoph Hellwig]

On Sat, Mar 27, 2021 at 03:17:59PM -0700, Tao Ren wrote:
> On Fri, Mar 26, 2021 at 01:05:26PM +0100, Christoph Hellwig wrote:
> > On Fri, Mar 26, 2021 at 12:03:03PM +0000, Robin Murphy wrote:
> > > This might happen to work out, but is far from correct. Just wait until you 
> > > try it on a platform where the USB controller is behind an IOMMU...
> > >
> > > It looks like something is more fundamentally wrong here - the device 
> > > passed to DMA API calls must be the actual hardware device performing the 
> > > DMA, which in USB-land I believe means the controller's sysdev.
> > 
> > The shiny new usb_intf_get_dma_device API provides the device to use.
> 
> Thanks Robin and Christoph for the feedback.
> 
> If I understand correctly, usb_intf_get_dma_device API is mainly for usb
> host drivers? I just found usb_gadget_map_request_by_dev API: does it
> make sense to replace usb_gadget_map_request with
> usb_gadget_map_request_by_dev so we can pass the actual DMA-capable
> hardware device (aspeed-vhub platform device) to the API?

Oh, right you're dealing with a gadget side driver.  Not sure about
the API there, I'll let the relevant maintainers chime in.

Re: [PATCH] usb: gadget: aspeed: set port_dev dma mask

[Tao Ren]

On Mon, Mar 29, 2021 at 08:17:35AM +0200, Christoph Hellwig wrote:
> On Sat, Mar 27, 2021 at 03:17:59PM -0700, Tao Ren wrote:
> > On Fri, Mar 26, 2021 at 01:05:26PM +0100, Christoph Hellwig wrote:
> > > On Fri, Mar 26, 2021 at 12:03:03PM +0000, Robin Murphy wrote:
> > > > This might happen to work out, but is far from correct. Just wait until you 
> > > > try it on a platform where the USB controller is behind an IOMMU...
> > > >
> > > > It looks like something is more fundamentally wrong here - the device 
> > > > passed to DMA API calls must be the actual hardware device performing the 
> > > > DMA, which in USB-land I believe means the controller's sysdev.
> > > 
> > > The shiny new usb_intf_get_dma_device API provides the device to use.
> > 
> > Thanks Robin and Christoph for the feedback.
> > 
> > If I understand correctly, usb_intf_get_dma_device API is mainly for usb
> > host drivers? I just found usb_gadget_map_request_by_dev API: does it
> > make sense to replace usb_gadget_map_request with
> > usb_gadget_map_request_by_dev so we can pass the actual DMA-capable
> > hardware device (aspeed-vhub platform device) to the API?
> 
> Oh, right you're dealing with a gadget side driver.  Not sure about
> the API there, I'll let the relevant maintainers chime in.

Given this is not the right path, I will drop the patch and work out a
new fix soon (by calling usb_gadget_map_request_by_dev, and with
modified subject).


Cheers,

Tao