linux-toradex.git/drivers/usb/class/cdc-wdm.c, branch v4.7-rc4 Linux kernel for Apalis and Colibri modules cdc-wdm: error returns need to be translated 2015-03-26T09:51:57+00:00 Oliver Neukum oneukum@suse.de 2015-03-23T13:34:43+00:00 85e8a0b9a3565c8185068b6b340cc8c6dd4411f4 One more case of error codes not correctly being correctly returned to user space. Signed-off-by: Olive Neukum <oneukum@suse.com>0 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
One more case of error codes not correctly being
correctly returned to user space.

Signed-off-by: Olive Neukum <oneukum@suse.com>0
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cdc-wdm: fix endianness bug in debug statements 2015-03-26T09:51:56+00:00 Oliver Neukum oneukum@suse.de 2015-03-20T13:29:34+00:00 323ece54e0761198946ecd0c2091f1d2bfdfcb64 Values directly from descriptors given in debug statements must be converted to native endianness. Signed-off-by: Oliver Neukum <oneukum@suse.de> CC: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
Values directly from descriptors given in debug statements
must be converted to native endianness.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
CC: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cdc-wdm: unify error handling in write 2015-03-26T09:51:56+00:00 Oliver Neukum oneukum@suse.de 2015-03-20T13:29:18+00:00 28965e17ee7a9591c241b831fee050d2391688c6 This makes sure the error handling path is the same for all error conditions, thus reducing code duplication. Signed-off-by: Oliver Neukum <oneukum@suse.de>0 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
This makes sure the error handling path is the same for
all error conditions, thus reducing code duplication.

Signed-off-by: Oliver Neukum <oneukum@suse.de>0
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cdc-wdm: return correct error codes 2015-03-26T09:50:52+00:00 Oliver Neukum oneukum@suse.de 2015-03-20T13:28:56+00:00 76cb03e7d5d7ba49175784dce961696da66c44cc Lieing to user space is wrong. The real reason for a failure to write should be returned to user space. Signed-off-by: Oliver Neukum <oneukum@suse.de>0 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
Lieing to user space is wrong. The real reason for a failure
to write should be returned to user space.

Signed-off-by: Oliver Neukum <oneukum@suse.de>0
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
usb: cdc-wdm: resp_count can be 0 even if WDM_READ is set 2014-01-13T04:13:28+00:00 Bjørn Mork bjorn@mork.no 2014-01-12T20:48:53+00:00 f563926fed982f26b391ca42493f55f2447f1b0a Do not decrement resp_count if it's already 0. We set resp_count to 0 when the device is closed. The next open and read will try to clear the WDM_READ flag if there was leftover data in the read buffer. This fix is necessary to prevent resubmitting the read URB in a tight loop because resp_count becomes negative. The bug can easily be triggered from userspace by not reading all data in the read buffer, and then closing and reopening the chardev. Fixes: 8dd5cd5395b9 ("usb: cdc-wdm: avoid hanging on zero length reads") Cc: <stable@vger.kernel.org> # 3.13 Signed-off-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
Do not decrement resp_count if it's already 0.

We set resp_count to 0 when the device is closed.  The next open and
read will try to clear the WDM_READ flag if there was leftover data
in the read buffer. This fix is necessary to prevent resubmitting
the read URB in a tight loop because resp_count becomes negative.

The bug can easily be triggered from userspace by not reading all
data in the read buffer, and then closing and reopening the chardev.

Fixes: 8dd5cd5395b9 ("usb: cdc-wdm: avoid hanging on zero length reads")
Cc: <stable@vger.kernel.org> # 3.13
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Merge 3.13-rc5 into usb-next 2013-12-24T18:18:03+00:00 Greg Kroah-Hartman gregkh@linuxfoundation.org 2013-12-24T18:18:03+00:00 99f14bd4d1b2a1b1b6cd508e08efdbc5e3919198 This resolves the merge issue with drivers/usb/host/ohci-at91.c Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
This resolves the merge issue with drivers/usb/host/ohci-at91.c

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
usb: cdc-wdm: avoid hanging on zero length reads 2013-12-20T20:06:46+00:00 Bjørn Mork bjorn@mork.no 2013-12-20T13:07:24+00:00 8dd5cd5395b90070d98149d0a94e5981a74cd2ec commit 73e06865ead1 ("USB: cdc-wdm: support back-to-back USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications") implemented queued response handling. This added a new requirement: The read urb must be resubmitted every time we clear the WDM_READ flag if the response counter indicates that the device is waiting for a read. Fix by factoring out the code handling the WMD_READ clearing and possible urb submission, calling it everywhere we clear the flag. Without this fix, the driver ends up in a state where the read urb is inactive, but the response counter is positive after a zero length read. This prevents the read urb from ever being submitted again and the driver appears to be hanging. Fixes: 73e06865ead1 ("USB: cdc-wdm: support back-to-back USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications") Cc: Greg Suarez <gsuarez@smithmicro.com> Signed-off-by: Bjørn Mork <bjorn@mork.no> Cc: stable <stable@vger.kernel.org> # 3.13 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 73e06865ead1 ("USB: cdc-wdm: support back-to-back
USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications") implemented
queued response handling. This added a new requirement: The read
urb must be resubmitted every time we clear the WDM_READ flag if
the response counter indicates that the device is waiting for a
read.

Fix by factoring out the code handling the WMD_READ clearing and
possible urb submission, calling it everywhere we clear the flag.

Without this fix, the driver ends up in a state where the read urb
is inactive, but the response counter is positive after a zero
length read.  This prevents the read urb from ever being submitted
again and the driver appears to be hanging.

Fixes: 73e06865ead1 ("USB: cdc-wdm: support back-to-back USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications")
Cc: Greg Suarez <gsuarez@smithmicro.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Cc: stable <stable@vger.kernel.org> # 3.13
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
usb: cdc-wdm: manage_power should always set needs_remote_wakeup 2013-12-09T21:14:11+00:00 Bjørn Mork bjorn@mork.no 2013-11-29T19:17:45+00:00 4144bc861ed7934d56f16d2acd808d44af0fcc90 Cc: stable <stable@vger.kernel.org> Reported-by: Oliver Neukum <oneukum@suse.de> Signed-off-by: Bjørn Mork <bjorn@mork.no> Acked-by: Oliver Neukum <oneukum@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
Cc: stable <stable@vger.kernel.org>
Reported-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Acked-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
usb: cdc-wdm: ignore speed change notifications 2013-10-30T00:02:41+00:00 Bjørn Mork bjorn@mork.no 2013-10-29T08:52:57+00:00 9983d6dc4e8a8a1ce599cb426c0ec9384219b476 The only notification supported by the Device Management class is Response Available. But this driver is also used as a subdriver of other CDC classes, allowing notifications like Speed Change and Network Connection. This results in log messages which are only confusing to an end user: [66255.801874] cdc_mbim 1-3:1.5: unknown notification 42 received: index 5 len 8 These drivers use cdc-wdm as a subdriver to allow access to an embedded management protocol, and all management is expected to use this protocol. There is therefore no need to handle any of these optional CDC notifications. Instead we can let the cdc-wdm driver recognize them and log a debug level message instead of an error. Reported-by: Rob Gardner <robmatic@gmail.com> Signed-off-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
The only notification supported by the Device Management class is
Response Available. But this driver is also used as a subdriver of
other CDC classes, allowing notifications like Speed Change and
Network Connection. This results in log messages which are only
confusing to an end user:

 [66255.801874] cdc_mbim 1-3:1.5: unknown notification 42 received: index 5 len 8

These drivers use cdc-wdm as a subdriver to allow access to an
embedded management protocol, and all management is expected to
use this protocol. There is therefore no need to handle any of
these optional CDC notifications. Instead we can let the cdc-wdm
driver recognize them and log a debug level message instead of an
error.

Reported-by: Rob Gardner <robmatic@gmail.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
USB: cdc-wdm: support back-to-back USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications 2013-10-30T00:02:41+00:00 Greg Suarez gpsuarez2512@gmail.com 2013-10-29T17:29:10+00:00 73e06865ead1bec8d1c179e1c647dc77adde9116 Some MBIM devices send back-to-back USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications when sending a message over multiple fragments or when there are unsolicited messages available. Count up the number of USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications received and decrement the count and submit the urb for the next response each time userspace completes a read the response. Signed-off-by: Greg Suarez <gsuarez@smithmicro.com> Acked-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
Some MBIM devices send back-to-back USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications
when sending a message over multiple fragments or when there are unsolicited
messages available.

Count up the number of USB_CDC_NOTIFY_RESPONSE_AVAILABLE notifications received
and decrement the count and submit the urb for the next response each time userspace
completes a read the response.

Signed-off-by: Greg Suarez <gsuarez@smithmicro.com>
Acked-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>