All of lore.kernel.org
 help / color / mirror / Atom feed
* [U-Boot] [PATCH 0/1] dm: usb: Fix divide by 0 errors in host drivers caused by not setting maxpsize
@ 2015-04-29  9:40 Hans de Goede
  2015-04-29  9:40 ` [U-Boot] [PATCH] " Hans de Goede
  0 siblings, 1 reply; 3+ messages in thread
From: Hans de Goede @ 2015-04-29  9:40 UTC (permalink / raw)
  To: u-boot

Hi Simon,

While working on moving the sunxi ehci code over to the driver model I've
found this bug in the usb uclass which causes a div by zero error.

I've spend a couple of hours yesterday evening debugging this and I'm happy
to report that I've a fix now. This fix has not seen as much testing as
I would have liked, but I believe it is correct and want to share it now
anyways to avoid people doing double work.

While thinking more about this I've come up with what I believe is a better
way to deal with this all together. I'll send that out right after this one,
note that the intend is that only one of the 2 patches gets applied, preferably
the latter as I believe that one is the best.

Regards,

Hans

p.s.

I've found 2 other issues with the dm usb code:

1) usb-1 devices / hubs behind a usb-2 (root)hub do not seem to work, this is
a familiar issue which I fixed in the pre dm usb code a while back. I'll
investigate this one myself.

2) "usb reset" is buggered, it either hangs or reboots the system. I would
appreciate if you could take a look (assuming you can reproduce it).

^ permalink raw reply	[flat|nested] 3+ messages in thread

* [U-Boot] [PATCH] dm: usb: Fix divide by 0 errors in host drivers caused by not setting maxpsize
  2015-04-29  9:40 [U-Boot] [PATCH 0/1] dm: usb: Fix divide by 0 errors in host drivers caused by not setting maxpsize Hans de Goede
@ 2015-04-29  9:40 ` Hans de Goede
  2015-05-01  1:54   ` Simon Glass
  0 siblings, 1 reply; 3+ messages in thread
From: Hans de Goede @ 2015-04-29  9:40 UTC (permalink / raw)
  To: u-boot

usb_child_pre_probe() initializes the final struct usb_device by redoingthe
usb_select_config() done by usb_scan_device() but not the usb_prepare_device()
call, this leads to a call into the hcd to get the device descriptors without
ep0's maxpacketsize being set, which leads to a device by 0 error in the
ehci driver.

This commit fixes this by passing the maxpacketsize through the
usb_device_platdata and setting it in usb_child_pre_probe().

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/usb/host/usb-uclass.c | 4 ++++
 include/usb.h                 | 3 +++
 2 files changed, 7 insertions(+)

diff --git a/drivers/usb/host/usb-uclass.c b/drivers/usb/host/usb-uclass.c
index 714bc0e..2af8611 100644
--- a/drivers/usb/host/usb-uclass.c
+++ b/drivers/usb/host/usb-uclass.c
@@ -539,6 +539,7 @@ int usb_scan_device(struct udevice *parent, int port,
 	plat->speed = udev->speed;
 	plat->slot_id = udev->slot_id;
 	plat->portnr = port;
+	plat->maxpacketsize = udev->maxpacketsize;
 	debug("** device '%s': stashing slot_id=%d\n", dev->name,
 	      plat->slot_id);
 	priv->next_addr++;
@@ -613,6 +614,9 @@ int usb_child_pre_probe(struct udevice *dev)
 	udev->slot_id = plat->slot_id;
 	udev->portnr = plat->portnr;
 	udev->speed = plat->speed;
+	udev->maxpacketsize = plat->maxpacketsize;
+	udev->epmaxpacketin[0] = 1 << (udev->maxpacketsize + 3);
+	udev->epmaxpacketout[0] = 1 << (udev->maxpacketsize + 3);
 	debug("** device '%s': getting slot_id=%d\n", dev->name, plat->slot_id);
 
 	ret = usb_select_config(udev);
diff --git a/include/usb.h b/include/usb.h
index 1984e8f..a5eb7fe 100644
--- a/include/usb.h
+++ b/include/usb.h
@@ -576,6 +576,7 @@ struct usb_platdata {
  * @slot_id:	USB3 slot ID, which is separate from the device address
  * @portnr:	Port number of this device on its parent hub, numbered from 1
  *		(0 mean this device is the root hub)
+ * @maxpacketsize: EP0 maxpacketsize
  * @strings:	List of descriptor strings (for sandbox emulation purposes)
  * @desc_list:	List of descriptors (for sandbox emulation purposes)
  */
@@ -585,6 +586,8 @@ struct usb_dev_platdata {
 	int devnum;
 	int slot_id;
 	int portnr;	/* Hub port number, 1..n */
+	int maxpacketsize; /* Maximum packet size; one of: PACKET_SIZE_* */
+
 #ifdef CONFIG_SANDBOX
 	struct usb_string *strings;
 	/* NULL-terminated list of descriptor pointers */
-- 
2.3.6

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* [U-Boot] [PATCH] dm: usb: Fix divide by 0 errors in host drivers caused by not setting maxpsize
  2015-04-29  9:40 ` [U-Boot] [PATCH] " Hans de Goede
@ 2015-05-01  1:54   ` Simon Glass
  0 siblings, 0 replies; 3+ messages in thread
From: Simon Glass @ 2015-05-01  1:54 UTC (permalink / raw)
  To: u-boot

On 29 April 2015 at 03:40, Hans de Goede <hdegoede@redhat.com> wrote:
> usb_child_pre_probe() initializes the final struct usb_device by redoingthe
> usb_select_config() done by usb_scan_device() but not the usb_prepare_device()
> call, this leads to a call into the hcd to get the device descriptors without
> ep0's maxpacketsize being set, which leads to a device by 0 error in the
> ehci driver.
>
> This commit fixes this by passing the maxpacketsize through the
> usb_device_platdata and setting it in usb_child_pre_probe().
>
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> ---
>  drivers/usb/host/usb-uclass.c | 4 ++++
>  include/usb.h                 | 3 +++
>  2 files changed, 7 insertions(+)

Acked-by: Simon Glass <sjg@chromium.org>

>
> diff --git a/drivers/usb/host/usb-uclass.c b/drivers/usb/host/usb-uclass.c
> index 714bc0e..2af8611 100644
> --- a/drivers/usb/host/usb-uclass.c
> +++ b/drivers/usb/host/usb-uclass.c
> @@ -539,6 +539,7 @@ int usb_scan_device(struct udevice *parent, int port,
>         plat->speed = udev->speed;
>         plat->slot_id = udev->slot_id;
>         plat->portnr = port;
> +       plat->maxpacketsize = udev->maxpacketsize;
>         debug("** device '%s': stashing slot_id=%d\n", dev->name,
>               plat->slot_id);
>         priv->next_addr++;
> @@ -613,6 +614,9 @@ int usb_child_pre_probe(struct udevice *dev)
>         udev->slot_id = plat->slot_id;
>         udev->portnr = plat->portnr;
>         udev->speed = plat->speed;
> +       udev->maxpacketsize = plat->maxpacketsize;
> +       udev->epmaxpacketin[0] = 1 << (udev->maxpacketsize + 3);
> +       udev->epmaxpacketout[0] = 1 << (udev->maxpacketsize + 3);
>         debug("** device '%s': getting slot_id=%d\n", dev->name, plat->slot_id);
>
>         ret = usb_select_config(udev);
> diff --git a/include/usb.h b/include/usb.h
> index 1984e8f..a5eb7fe 100644
> --- a/include/usb.h
> +++ b/include/usb.h
> @@ -576,6 +576,7 @@ struct usb_platdata {
>   * @slot_id:   USB3 slot ID, which is separate from the device address
>   * @portnr:    Port number of this device on its parent hub, numbered from 1
>   *             (0 mean this device is the root hub)
> + * @maxpacketsize: EP0 maxpacketsize
>   * @strings:   List of descriptor strings (for sandbox emulation purposes)
>   * @desc_list: List of descriptors (for sandbox emulation purposes)
>   */
> @@ -585,6 +586,8 @@ struct usb_dev_platdata {
>         int devnum;
>         int slot_id;
>         int portnr;     /* Hub port number, 1..n */
> +       int maxpacketsize; /* Maximum packet size; one of: PACKET_SIZE_* */
> +
>  #ifdef CONFIG_SANDBOX
>         struct usb_string *strings;
>         /* NULL-terminated list of descriptor pointers */
> --
> 2.3.6
>

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2015-05-01  1:54 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-04-29  9:40 [U-Boot] [PATCH 0/1] dm: usb: Fix divide by 0 errors in host drivers caused by not setting maxpsize Hans de Goede
2015-04-29  9:40 ` [U-Boot] [PATCH] " Hans de Goede
2015-05-01  1:54   ` Simon Glass

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.