Since highmem PIO URB handling was introduced in:
8e8a551 usb: musb: host: Handle highmem in PIO mode
when a URB is being handled it may happen that the static use_sg flag
was set by a previous URB with buffer in highmem. This leads to error
in handling the present URB.
Fix this by making the use_sg flag URB specific.
Cc: stable <stable@vger.kernel.org> # 3.7+
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Virupax Sadashivpetimath <virupax.sadashivpetimath@stericsson.com>
Signed-off-by: Fabio Baltieri <fabio.baltieri@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
MUSB controller cannot work in DMA mode with misaligned buffers,
switching in PIO mode.
HCD core has hooks that allow to override the default DMA
mapping and unmapping routines for host controllers that have
special DMA requirements, such as alignment constraints.
It is observed that work in PIO mode is slow and it's better
to align buffers properly before passing them to MUSB
This increased throughput 80->120 MBits/s over musb@omap4 with
USB Gigabit Ethernet adapter attached.
Some ideas are taken from ehci-tegra.c
Signed-off-by: Ruslan Bilovol <ruslan.bilovol@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
all other functions under drivers/usb/ start
with usb_, let's do the same thing.
This patch is in preparation for moving otg_state_string
to usb-common.c and deleting otg.c completely.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Whenever an URB is programmed for transfer, the TXFIFO
is flushed. This results in valid packets of the
previous transfer to get flushed when double buffering
is enabled (The MUSB_TXCSR_FIFONOTEMPTY bit in TXCSR
is set indicating that a packet in the FIFO is yet to be sent)
For ex:- In Host mode Audio, noise is heard in the headset
when double buffering is enabled on the ISO endpoint.
The fifo flush is removed for double buffering case.
The fifo is now flushed only in cases of error or when
aborting a transfer.
Also, In Host MSC case, data toggle errors are seen when double
buffering is enabled on the bulk endpoint. Whenever an URB is
programmed for transfer, the data toggle is set manually
resulting in data toggle errors on the bus. Leave the data
toggle handling upto the hardware in the double buffering case.
Signed-off-by: supriya karanth <supriya.karanth@stericsson.com>
Signed-off-by: Praveena NADAHALLY <praveen.nadahally@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The "mult" factor is not updated properly for the can_bulk_split()
case.
The AUTOSET bit in the TXCSR is not being set if the "mult"
factor is greater than 1 for the High Bandwidth ISO case.
But the "mult" factor is also greater than 1 in case of Full speed
bulk transfers with the packet splitting in TXMAXP register
enabled with can_bulk_split().
Without the AUTOSET the DMA transfers will not progress in mode1
[ balbi@ti.com : fix braces placement ]
Signed-off-by: supriya karanth <supriya.karanth@stericsson.com>
Signed-off-by: Praveena NADAHALLY <praveen.nadahally@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This is part of the workaround for AM35x advisory Advisory 1.1.20.
The advisory says that the IPSS bridge can't handle 8 & 16 bit read
access. An 16bit read access to MUSB_INTRTXE results in an 32bit read
access which also reads INTRRX and therefore may lose interrupts.
This patch uses a shadow register of MUSB_INTRTXE so we only perform
write access to it.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This resolves the merge problems with:
drivers/usb/dwc3/gadget.c
drivers/usb/musb/tusb6010.c
that had been seen in linux-next.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
when using musb_urb_enqueue to submit three urbs to the same endpoint, when
hep->hcpriv is NULL, qh will be allocated when the first urb is completed.
When the IRQ completes the next two urbs, qh->hep->hcpriv will be set to NULL.
Now the second urb get musb->lock and executes musb_schedule(), but
next_urb(qh) is NULL, so musb_start_urb will Oops.
[ balbi@ti.com : practically rewrote commit log so it makes sense ]
Signed-off-by: mayuzheng <myz147@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
In case of USB bulk transfer, when himem page
is received, the usb_sg_init function sets the
urb transfer buffer to NULL. When such URB
transfer is handled, kernel crashes in PIO mode.
Handle this by mapping the highmem buffer in PIO mode.
Signed-off-by: Virupax Sadashivpetimath <virupax.sadashivpetimath@stericsson.com>
Signed-off-by: Praveena NADAHALLY <praveen.nadahally@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Fixes endpoint starvation issue when more than one bulk QH is
multiplexed on the reserved bulk TX endpoint.
This patch sets the NAK timeout interval for such QHs, and when
a timeout triggers the next QH will be scheduled.
This scheme doesn't work for devices which are connected to a
high to full speed tree (transaction translator) as there is
no NAK timeout interrupt from the musb controller from such
devices.
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This is to reduce the overhead of dma programming for zero byte
transmit as same can be done using pio mode.
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Just two patches here:
First we have a fix to disable DMA in case DMA channel
request fails.
Second, there's a fix for situations where the user
kills musb (by rmmod or any other means) while a
transfer is on the fly. In such cases, we could be
led into a NULL pointer dereference due to endpoint
being disabled and endpoint descriptor being NULL.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)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=4uPD
-----END PGP SIGNATURE-----
Merge tag 'musb-for-v3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb into usb-next
usb: musb: patches for v3.6 merge window
Just two patches here:
First we have a fix to disable DMA in case DMA channel
request fails.
Second, there's a fix for situations where the user
kills musb (by rmmod or any other means) while a
transfer is on the fly. In such cases, we could be
led into a NULL pointer dereference due to endpoint
being disabled and endpoint descriptor being NULL.
Currently DMA channels are allocated and they remain allocated
even if there is no active data transfer. Added channel_release()
whenever there is no pending request.
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Currently in case of MUSB DMA channel request failure we are not
clearing MUSB_RXCSR_DMAENAB, MUSB_RXCSR_H_AUTOREQ and
MUSB_RXCSR_AUTOCLEAR bits of MUSB RXCSR of MUSB DMA. Which is
causing failure in receipt of data packets in next transfer.
Fix is to disable the MUSB DMA mode and related bits incase of
DMA channel request fails
Signed-off-by: Mantesh Sarashetti <mantesh@ti.com>
Signed-off-by: Ruslan Bilovol <ruslan.bilovol@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Control transfers with data expected from device to host will use usb_rcvctrlpipe()
for urb->pipe so for such urbs 'is_in' will be set causing control urb to fall
into the first "if" condition in musb_cleanup_urb().
Fixed by adding logic to check for non control endpoints.
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
TXMAXP register is not getting programmed correctly for a full speed device
as can_bulk_split() have been removed by
"0662481: usb: musb: disable double buffering when it's broken" patch.
Adding back the case for can_bulk_split() to fix the reset message seen with
a full speed stick.
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Now ${LINUX}/drivers/usb/* can use usb_endpoint_maxp(desc) to get maximum packet size
instead of le16_to_cpu(desc->wMaxPacketSize).
This patch fix it up
Cc: Armin Fuerst <fuerst@in.tum.de>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Johannes Erdfelt <johannes@erdfelt.com>
Cc: Vojtech Pavlik <vojtech@suse.cz>
Cc: Oliver Neukum <oliver@neukum.name>
Cc: David Kubicek <dave@awk.cz>
Cc: Johan Hovold <jhovold@gmail.com>
Cc: Brad Hards <bhards@bigpond.net.au>
Acked-by: Felipe Balbi <balbi@ti.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Thomas Dahlmann <dahlmann.thomas@arcor.de>
Cc: David Brownell <david-b@pacbell.net>
Cc: David Lopo <dlopo@chipidea.mips.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Michal Nazarewicz <m.nazarewicz@samsung.com>
Cc: Xie Xiaobo <X.Xie@freescale.com>
Cc: Li Yang <leoli@freescale.com>
Cc: Jiang Bo <tanya.jiang@freescale.com>
Cc: Yuan-hsin Chen <yhchen@faraday-tech.com>
Cc: Darius Augulis <augulis.darius@gmail.com>
Cc: Xiaochen Shen <xiaochen.shen@intel.com>
Cc: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Cc: OKI SEMICONDUCTOR, <toshiharu-linux@dsn.okisemi.com>
Cc: Robert Jarzmik <robert.jarzmik@free.fr>
Cc: Ben Dooks <ben@simtec.co.uk>
Cc: Thomas Abraham <thomas.ab@samsung.com>
Cc: Herbert Pötzl <herbert@13thfloor.at>
Cc: Arnaud Patard <arnaud.patard@rtp-net.org>
Cc: Roman Weissgaerber <weissg@vienna.at>
Acked-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: Tony Olech <tony.olech@elandigitalsystems.com>
Cc: Florian Floe Echtler <echtler@fs.tum.de>
Cc: Christian Lucht <lucht@codemercs.com>
Cc: Juergen Stuber <starblue@sourceforge.net>
Cc: Georges Toth <g.toth@e-biz.lu>
Cc: Bill Ryder <bryder@sgi.com>
Cc: Kuba Ober <kuba@mareimbrium.org>
Cc: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Variable d is a struct usb_iso_packet_descriptor. The status filed is usually
negative when an error happens.
Signed-off-by: Márton Németh <nm127@freemail.hu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
musb_ep_program() contains obviously duplicate check for 'dma_channel' in its
IN/receive path -- removing it allows to save one level of indentation. While
at it, improve the comment style...
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
We have a generic way of enabling/disabling
different debug messages on a driver called
DYNAMIC_PRINTK. Anyone interested in enabling
just part of the debug messages, please read
the documentation under:
Documentation/dynamic-debug-howto.txt
for information on how to use that great
infrastructure.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Commit 3dacdf11 "usb: factor out state_string() on otg drivers"
broke building musb drivers since there is already another
otg_state_string() function in musb drivers, but with different
prototype. Fix musb drivers to use common otg_state_string(), too.
Also provide a nop for otg_state_string() if CONFIG_USB_OTG_UTILS
is not defined.
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This is needed to resolve some merge conflicts that were found
in the USB host controller patches, and reported by Stephen Rothwell.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The convention is to prefix symbols exported from the USB HCD core with
"usb_hcd". This change makes unmap_urb_setup_for_dma() and
unmap_urb_for_dma() consistent with that.
Signed-off-by: Robert Morell <rmorell@nvidia.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
We know that blackfin doesn't support double
buffering feature as of today. So we add a
flag set by musb_platform_init() to forcefully
disable that feature.
Such flag is created and marked as deprecated
to force us to find a solution for the missing
double buffering support on blackfin.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Add support for MUSB Host DMA transfers greater than max
channel length, so that such transfers won't be truncated.
Signed-off-by: Anil Shetty <anil@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The USB stack maps the buffer for DMA if the controller supports DMA.
MUSB controller can perform DMA as well as PIO transfers.
The buffer needs to be unmapped before CPU can perform
PIO data transfers.
Export unmap_urb_for_dma() so that drivers can perform
the DMA unmapping in a sane way.
Signed-off-by: Maulik Mankad <x0082077@ti.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch fixes the issue which was observed while transfering
a large file ( > 20MB) over USB (OMAP MUSB controller acts as USB host)
to an attached USB thumb drive.
It was found that CDB field of CBW packet was set to 0x0. This was
due to missing a barrier before DMA engine starts transfer.
This buffer is allocated using dma_alloc_coherent which gives
non-cacheble but bufferable memory and hence needed a write
memory barrier to flush the write buffer.
More info on this thread is here:
http://www.spinics.net/lists/linux-omap/msg33987.html
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: Maulik Mankad <x0082077@ti.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
We assign "urb->hcpriv = qh;" a few lines down. I'm pretty sure we
want it "urb->hcpriv" to be NULL not a freed value.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The recent commit "usb: musb: Fix for isochronous IN transfer" (f82a689fa)
seems to have been against an older kernel version. It uses the old style
naming of variables. Unfortunately, this breaks building for most MUSB
users out there since "bDesiredMode" has been renamed to "desired_mode".
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Acked-by: Anand Gadiyar <gadiyar@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Trying to use double buffer modes in RTL versions <2.0 may result in
infinite hangs or data corruption. So avoid them with older versions.
Signed-off-by: Cliff Cai <cliff.cai@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch implements the work around for a Mentor controller related
bug where it's observed a BULK Tx toggle error on the bus when a
BULK IO gets scheduled on an endpoint that was earlier used for
handling ISOC transaction and needed to start on 1 toggle. When such
a situation arises even if the TXCSR toggle bits are programmed
correctly by the musb driver the data gets transmitted with 0 toggle
which leads to toggle error on the bus and the BULK transaction fails.
In case of MSC write, the device gets reset by the Host.
This Mentor bug is observed on almost all Mentor versions (1.3, 1.5,
1.8). Confirmed on DM644x, DM355, DM365, OMAPL13x platforms.
Signed-off-by: Swaminathan S <swami.iyer@ti.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Fix the following sparse warning:
drivers/usb/musb/musb_host.c:1642:9: warning: symbol 'status' shadows an earlier one
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Isochronous Tx DMA is getting programmed but never getting started
for CPPI and TUSB DMAs and thus Isochronous Tx doesn't work.
Fixing it by starting DMAs using musb_h_tx_dma_start().
Signed-off-by: Swaminathan S <swami.iyer@ti.com>
Signed-off-by: Babu Ravi <ravibabu@ti.com>
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Commit c9cd06b3d6 (musb_host: refactor
URB giveback) included due to my overlook the change incorrect in the
context of the current kernel -- undo it.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Get rid of some obnoxious and inappropriate messaging, mostly on
DaVinci, when usbcore tries to autosuspend a root hub if just a
mini/micro-A connector is connected. Symptom: endless stream of
messages reading like:
musb_bus_suspend 2221: trying to suspend as a_wait_vrise is_active=1
Improve that musb bus suspend primitive a bit. Take advantage of
this call to update the OTG state machine if appropriate, moving
the device out of the A_WAIT_VRISE state. There's basically no
timer for that state transition just now, except with tusb6010;
that can make trouble.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Tested on OMAP3 host side with Creative (Live! Cam Optia) USB camera
which uses high bandwidth isochronous IN endpoints. FIFO mode 4 is
updated to provide the needed 4K endpoint buffer without breaking
the g_nokia composite gadget configuration. (This is the only
gadget driver known to use enough endpoints to notice the change.)
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Let the otg_transceiver in MUSB be managed by an external driver;
don't assume it's integrated. OMAP3 chips need it to be external,
and there may be ways to interact with the transceiver which add
functionality to the system.
Platform init code is responsible for setting up the transeciver,
probably using the NOP transceiver for integrated transceivers.
External ones will use whatever the board init code provided,
such as twl4030 or something more hands-off.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
As musb_advance_schedule() is now the only remaning
caller of musb_giveback() (and the only valid context
of such call), just fold the latter into the former
and then rename __musb_giveback() into musb_giveback().
This is a net minor shrink.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The argument for the 'is_in' parameter of musb_cleanup_urb()
is always extracted from an URB that's passed to the function.
So that parameter is superfluous; remove it.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The existance of the scheduling list shouldn't matter in
determining whether there's currectly an URB executing on a
hardware endpoint. What should actually matter is the 'in_qh'
or 'out_qh' fields of the 'struct musb_hw_ep' -- those are
set in musb_start_urb() and cleared in musb_giveback() when
the endpoint's URB list drains. Hence we should be able to
replace the big *switch* statements in musb_urb_dequeue()
and musb_h_disable() with mere musb_ep_get_qh() calls...
While at it, do some more changes:
- add 'is_in' variable to musb_urb_dequeue();
- remove the unnecessary 'epnum' variable from musb_h_disable();
- fix the comment style in the vicinity.
This is a minor shrink of source and object code.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Factor out the often used code to get/set the active 'qh'
pointer for the hardware endpoint. Change the way the case
of a shared FIFO is handled by setting *both* 'in_qh' and
'out_qh' fields of 'struct musb_hw_ep'. That seems more
consistent and makes getting to the current 'qh' easy when
the code knows the direction beforehand.
While at it, turn some assignments into intializers and
fix declaration style in the vicinity.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Refactor musb_save_toggle() as follows:
- replace 'struct musb_hw_ep *ep' parameter by 'struct
musb_qh *qh' to avoid re-calculating this value
- move usb_settogle() call out of the *if* operator.
This is a net minor shrink of source and object code.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Multi-frame isochronous TX URBs transfers in DMA mode never
complete with CPPI DMA because musb_host_tx() doesn't restart
DMA on the second frame, only emitting a debug message.
With Inventra DMA they complete, but in PIO mode. To fix:
- Factor out programming of the DMA transfer from
musb_ep_program() into musb_tx_dma_program();
- Reorder the code at the end of musb_host_tx() to
facilitate the fallback to PIO iff DMA fails;
- Handle the buffer offset consistently for both
PIO and DMA modes;
- Add an argument to musb_ep_program() for the same
reason (it only worked correctly with non-zero
offset of the first frame in PIO mode);
- Set the completed isochronous frame descriptor's
'actual_length' and 'status' fields correctly in
DMA mode.
Also, since CPPI reportedly doesn't like sending isochronous
packets in the RNDIS mode, change the criterion for this
mode to be used only for multi-packet transfers. (There's
no need for that mode in the single-packet case anyway.)
[ dbrownell@users.sourceforge.net: split comment paragraph
into bullet list, shrink patch delta, style tweaks ]
Signed-off-by: Pavel Kiryukhin <pkiryukhin@ru.mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The MUSB code clears TXCSR_DMAMODE incorrectly in several
places, either asserting that TXCSR_DMAENAB is clear (when
sometimes it isn't) or clearing both bits together. Recent
versions of the programmer's guide require DMAENAB to be
cleared first, although some older ones didn't.
Fix this and while at it:
- In musb_gadget::txstate(), stop clearing the AUTOSET
and DMAMODE bits for the CPPI case since they never
get set anyway (the former bit is reserved on DaVinci);
but do clear the DMAENAB bit on the DMA error path.
- In musb_host::musb_ep_program(), remove the duplicate
DMA controller specific code code clearing the TXCSR
previous state, add the code to clear TXCSR DMA bits
on the Inventra DMA error path, to replace such code
(executed late) on the PIO path.
- In musbhsdma::dma_channel_abort()/dma_controller_irq(),
add/use the 'offset' variable to avoid MUSB_EP_OFFSET()
invocations on every RXCSR/TXCSR access.
[dbrownell@users.sourceforge.net: don't introduce CamelCase,
shrink diff]
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
We really want to use DMA mode 1 for all multi-packet transfers;
that's one IRQ on DMA completion, instead of one per packet.
There is an important issue with such transfers, especially on
the host side: when such transfers end with a full-size packet,
we must defer musb_dma_completion() calls until the FIFO empties.
Else we report URB completions too soon, and may clobber data in
the FIFO fifo when writing the next packet (losing data).
The Inventra DMA support uses DMA mode 1, but it ignores that
issue. The CPPI DMA support uses mode 0, but doesn't handle
its TXPKTRDY interrupts quite right either; it can get stale
"packet ready" interrupts, and report transfer completion too
early using slightly different code paths, also losing data.
So I'm solving it in a generic way -- by adding a sort of the
"interrupt filter" into musb_host_tx(), catching these cases
where a DMA completion IRQ doesn't suffice and removing some
needlessly controller-specific logic. When a TXDMA interrupt
happens and DMA request mode 1 is active, that filter resets
to mode 0 and defers URB completion processing until TXPKTRDY,
unless the FIFO is already empty. Related filtering logic in
Inventra and CPPI code gets removed.
Since it should be competely safe now to use the DMA request
mode 1 for host side transfers with the CPPI DMA controller,
set it in musb_h_tx_dma_start() ... now renamed (and shared).
[ dbrownell@users.sourceforge.net: don't introduce more
CamElCase; use more concise explanations ]
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The MUSB host side can't share generic TX FIFO flush logic
with EP0; the EP0 TX status register bits are different
from those for other entpoints.
Resolve this issue by providing a new EP0-specific routine
to flush and reset the FIFO, which pays careful attention to
restrictions listed in the latest programmer's guide. This
gets rid of an open issue whereby the usbtest control write
test (#14) failed.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Someone noted that the enqueue path used an unlocked access
for usb_host_endpoint->hcpriv ... fix that, by being safe
and always accessing it under spinlock protection.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Fixes endpoint starvation issue when more than one bulk QH is
multiplexed on the reserved bulk RX endpoint, which is normal
for cases like serial and ethernet adapters.
This patch sets the NAK timeout interval for such QHs, and when
a timeout triggers the next QH will be scheduled. (This resembles
the bulk scheduling done in hardware by EHCI, OHCI, and UHCI.)
This scheme doesn't work for devices which are connected to a
high to full speed tree (transaction translator) as there is
no NAK timeout interrupt from the musb controller from such
devices.
Tested with PIO, Inventra DMA, CPPI DMA.
[ dbrownell@users.sourceforge.net: fold in start_urb() update;
clarify only for bulk RX; don't accidentally clear WZC bits ]
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The current MUSB host code doesn't make use of all the available
FIFOs in for periodic transfers since it wrongly assumes the RX
and TX sides of any given hw_ep always share one FIFO.
Change: use 'in_qh' and 'out_qh' fields of the 'struct musb_hw_ep'
to check the endpoint's business; get rid of the now-unused 'periodic'
array in the 'struct musb'. Also optimize a loop induction variable
in the endpoint lookup code.
(Based on a previous patch from Ajay Kumar Gupta <ajay.gupta@ti.com>)
[ dbrownell@users.sourceforge.net: clarify description and origin
of this fix; whitespace ]
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Remove wrongly applied upper limit on the interrupt transfer
interval for low speed devices (not much of an error per se,
according to USB specs).
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
