Add waiting for receiving Ethernet gadget state on the Windows host
side before dropping pullup, but keep it for debug.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Port USB gadget RNDIS protocol support from linux-2.6.26
(.27 gadget stack actually has composite drivers).
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Disconnecting USB gadget with pending interrupt may cause its wrong
handling in the next time when interface will be started again
(especially actual for RNDIS). This interrupt may force the gadget
to queue unexpected response before setup stage.
Despite the fact that such interrupt handled after dropped pullup
also may add pending response, this will not bring to any issues due to
usb_ep_disable (which clears the queue) called on gadget unregistering.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
This adds support for using USB Ethernet dongles in host mode. This is just
the framework - drivers will come later. A new config option called
CONFIG_USB_HOST_ETHER can be defined in board config files to switch this
on.
The was originally written by NVIDIA and was cleaned up for release by the
Chromium authors.
Signed-off-by: Simon Glass <sjg@chromium.org>
Changed both to use a common timeout for URB submission, since they were using
different values and EHCI's was too short.
Also fixed EHCI to actually check if urb submission succeeded, rather than
silently continuing into the weeds.
Change-Id: I7f71499ffaa05187d8e5618db2419e1606007b82
Signed-off-by: Simon Glass <sjg@chromium.org>
Calling eth_bind at usb_eth_init time causes renaming of the network
device from 'usb_ether' to 'usb0'. Fixing this to keep the first name.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Since the ether may not be the only one usb gadget would be used
in the uboot, it is neccessary to do the register each time the
eth begin to work to make usb gadget driver less confussed when
we want to use two different usb gadget at the same time.
Usb gadget driver could simple ignore the register operation, if
it find the driver has been registered already.
Signed-off-by: Lei Wen <leiwen@marvell.com>
Calling usb_dev_init() from within the EHCI host driver is wrong.
The EHCI host driver should have no dependency/interconnection to the
USB device driver.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Remy Bohmer <linux@bohmer.net>
This patch fixes a problem noticed on lwmon5 (PPC440EPx) using the
common EHCI driver, when "usb reset" is issued multiple times.
Upon the 2nd (and further) "usb reset" command, the command fails
with the following messages:
=> usb reset
(Re)start USB...
USB: Register 1111 NbrPorts 1
USB EHCI 1.00
scanning bus for devices... 5 USB Device(s) found
scanning bus for storage devices... 2 Storage Device(s) found
=> usb reset
(Re)start USB...
USB: EHCI fail to reset
Error, couldn't init Lowlevel part
This patch fixes this problem. Now "usb reset" can be called multiple
times.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Remy Bohmer <linux@bohmer.net>
This watchdog reset call is needed here, otherwise the lwmon5 board
(PPC440EPx based) will reset upon the "usb reset" command.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Remy Bohmer <linux@bohmer.net>
Checking the status field of the qTD token in the current code
do not take into acount cases where endpoint stall (halted) bit
is set together with XactErr status bit. As a result clearing
stall on an endpoint won't be done if this status bit was also
set. Check for halted bit and report USB_ST_STALLED status
if the host controller also indicates endpoit stall condition.
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Before this commit, weak symbols were not overridden by non-weak symbols
found in archive libraries when linking with recent versions of
binutils. As stated in the System V ABI, "the link editor does not
extract archive members to resolve undefined weak symbols".
This commit changes all Makefiles to use partial linking (ld -r) instead
of creating library archives, which forces all symbols to participate in
linking, allowing non-weak symbols to override weak symbols as intended.
This approach is also used by Linux, from which the gmake function
cmd_link_o_target (defined in config.mk and used in all Makefiles) is
inspired.
The name of each former library archive is preserved except for
extensions which change from ".a" to ".o". This commit updates
references accordingly where needed, in particular in some linker
scripts.
This commit reveals board configurations that exclude some features but
include source files that depend these disabled features in the build,
resulting in undefined symbols. Known such cases include:
- disabling CMD_NET but not CMD_NFS;
- enabling CONFIG_OF_LIBFDT but not CONFIG_QE.
Signed-off-by: Sebastien Carlier <sebastien.carlier@gmail.com>
This patch solves a problem with USB hanging under higher load on a
i.MX31 board. It falls into class of typical USB problems and fixes:
if you don't understand the real cause, add a delay somewhere.
The problem appeared after introduction of ELF relocation, which
results in smaller code, which appears to run faster (probably because
it fits better in the cache); turning off the instruction cache,
adding debug printf()s and increasing the delay have all been found to
make the problem go away.
Moving the original "udelay(1)" up in the code to it's new place made
the problem appear much less frequently. Increasing the delay to 2
microseconds then made the code run reliably in all (hour-long) tests.
To be on the safe side, we set it to 5 microseconds here.
Signed-off-by: Heiko schocher <hs@denx.de>
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Remy Bohmer <linux@bohmer.net>
Cc: Stefano Babic <sbabic@denx.de>
Commit 3ed1607 "USB: sync Queue Element Transfer Descriptor against
EHCI spec" added an "__attribute__ ((aligned (32)))" to the
declaration of struct qTD, as used for example in the Linux kernel as
well.
However, it turns out that this attribute causes errors in "usb start"
(like "ERROR: NOT USB_CONFIG_DESC 7b" and similar). Drop the attribute
again.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Dan Lykowski <lykowdk@gmail.com>
Cc: Remy Bohmer <linux@bohmer.net>
Cc: Stefano Babic <sbabic@denx.de>
It is useful to know the EHCI-PCI hccr, hcor and hc_lenght to make sure it was
successfully registered, and at the correct location.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Appendix B "EHCI 64-Bit Data Structures" of the "Enhanced Host
Controller Interface Specification for Universal Serial Bus" (Rev.
1.0, March 12, 2002) defines additional fields which were missing in
U-Boot's struct qTD; as these are also present in recent versions of
struct ehci_qtd in the Linux kernel, we add them here, too.
This fixes some nasty memory corruption problems.
Reported-by: Dan Lykowski <lykowdk@gmail.com>
See http://permalink.gmane.org/gmane.comp.boot-loaders.u-boot/76942
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Remy Bohmer <linux@bohmer.net>
Cc: Dan Lykowski <lykowdk@gmail.com>
Cc: Stefano Babic <sbabic@denx.de>
Tested-by: Stefano Babic <sbabic@denx.de>
Sometimes a usb tree is not popolated after a system reset.
It seems a delay is required after setting the USB ports
for the MX.31 before resetting the ehci controller.
Signed-off-by: Stefano Babic <sbabic@denx.de>
This is a change similar to what is already in the Linux driver. We
should only program the CLRDATATOG bit when the current mode indicates
that it is needed.
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Signed-off-by: Cliff Cai <cliff.cai@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
After gadget reinitializaton (after tftp has been done once)
packet_received may become equal to 1 due to nuking OUT_EP
while disabling it in eth_reset_config.
rx_submit called from usb_eth_init queues rx_req first time.
But the first call of usb_eth_recv from NetLoop queues rx_req
again due to packet_received = 1.
The following flow shows the path of functions calls when
this happens:
net/net.c:NetLoop
|
+-net/eth.c:eth_init
| ether.c:usb_eth_init
| |
| +-udc_driver:usb_gadget_handle_interrupts
| | udc_driver:...
| | ether.c:eth_setup
| | ether.c:eth_set_config
| | ether.c:eth_reset_config
| | udc_driver:usb_ep_disable
| | udc_driver:nuke
| | ether.c:rx_complete
| | ether.c: packet_received = 1;
| |
| +-ether.c:rx_submit
| udc_driver:usb_ep_queue --- The first time when rx_req is queued
|
+-net/eth.c:eth_rx
ether.c:usb_eth_recv
|
+-udc_driver:usb_gadget_handle_interrupts
| udc_driver:... --- no interrupts, returning
+-ether.c: if (packet_received) { ...
ether.c:rx_submit
udc_driver:usb_ep_queue --- The second time!
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Remove and fix needless and destructive operations with tx/rx_req.
1) 'req' in rx_complete is always not NULL and always equals to rx_req
2) Free allocated tx_req if rx_req allocation has failed
3) Do not zero out tx/rx_req in usb_eth_init, leave this for
eth_reset_config which will be called at the next use of usb0
4) Additional check in usb_eth_recv is not required
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
This fixes TIMEOUT with my Kingston 32GB USB3.0 flashdrive, which I experienced
on my PXA270 (USB 1.1) Vpac270 board.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Fixes most of checkpatch warnings and errors in USB gadget stack.
The most frequently encountered problems are:
1) "(foo*)", "foo * bar", "foo* bar"
2) C99 // comments
3) No spaces before/after/around '?', ':', '=', '==', ',', '&', '('
4) Spaces before '['
5) Spaces between function names and '('
6) Block braces in wrong places
7) Spaces before tabs
8) Macros with complex values not enclosed in parenthesis
9) Multiline comments start just after /*
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
The patch removes an endless loop in the usb_eth_send
if the tx_complete is not called before going
in the loop. The driver interrupt routine is called
allowing the driver to check if the TX is completed.
Signed-off-by: Stefano Babic <sbabic@denx.de>
The buffer for the status request must be word aligned
because it is accessed with 32 bit pointer in the
eth_status_complete function.
Signed-off-by: Stefano Babic <sbabic@denx.de>
The patch is based on commit bb9496c6f7e853e5d4edd5397c9d45f1968d623c
(done by Julia Lawall <julia@diku.dk>) from linux-2.6.git.
Since num is unsigned, it would seem better to use simple_strtoul that
simple_strtol.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Fix potential oops on rare error path.
The patch is based on commit e7b13ec9235b9fded90f826ceeb8c34548631351
(done by David Brownell <david-b@pacbell.net>) from linux-2.6.git.
Description of the issue taken from linux kernel bugzilla:
(https://bugzilla.kernel.org/show_bug.cgi?id=9594)
The potential error can be tracked down as follows:
(1) line 807: let the second conjunct on the "if" statment be false
meaning "dev->status_ep" is null. This means the "if" evaluates
to false.
follow thru the code until...
(2) line 808: usb_ep_disable(dev->status_ep) passes in a null argument,
however "usb_ep_disable" cannot handle that:
(from include/linux/usb/gadget.h)
191 static inline int
192 usb_ep_disable (struct usb_ep *ep)
193 {
194 return ep->ops->disable (ep);
195 }
--
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Fix possible oops on stat_req->buf initialization and fix ep0 and
status_ep confusion (last one is just intended for stat_req keeping).
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Signed-off-by: Stefano Babic <sbabic@denx.de>
Fix in_ep and out_ep confusion (rx_req was allocated from out_ep, not
from in_ep) and add lost dev->req freeing.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Add lost 'qmult' definition for High Speed devices and make it
configurable through CONFIG_USB_ETH_QMULT.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Prohibit simultaneous usage of both old and new gadget stacks and
allow UDC drivers to be dependent on CONFIG_USB_ETHER.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
The MUSB_TXCSR_MODE register setting isn't supported on Blackfin musb
parts, so stub it out to 0. This matches Linux behavior.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The Blackfin implementation of musb has a TXCOUNT register that needs to
be programmed when transmitting data.
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Signed-off-by: Cliff Cai <cliff.cai@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
AM35x supports only 32bit read operations so we need to have
workaround for 8bit and 16bit read operations.
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
