The ISP1760 has some timing requirements where it has to delay a short
period after a write to a register has started.  However, this delay is
from the time the write hits the USB chip (the ISP1760), not from the
time where the processor started processing the write.  So on a quick
enough processor, it is sometimes possible for the write to not hit the
device before we start delaying, and we then violate the part's timing
requirements, so things stop working.

To avoid all this, insert a write barrier after the register write and
before the timing delay/register read so we can guarantee we only start
counting time after the write has hit the device.
Signed-off-by: NMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The probe method used by i2c_new_probed_device() may not be suitable
for all cases. Let the caller provide its own, optional probe
function.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Acked-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Fake "address-of" expressions that evaluate to NULL generally confuse
readers and can provoke compiler warnings.  This patch (as1412)
removes three such fake expressions, using "#ifdef"s in their place.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Tell the USB core that we can do DMA directly (instead of needing it to
memory-map the buffers for PIO).  If the xHCI host supports 64-bit addresses,
set the DMA mask accordingly.  Otherwise indicate the host can handle 32-bit DMA
addresses.

This improves performance because the USB core doesn't have to spend time
remapping buffers in high memory into the 32-bit address range.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

To tell the host controller that there are transfers on the endpoint
rings, we need to ring the endpoint doorbell.  This is a PCI MMIO write,
which can be delayed until another register read is queued.

The previous code would flush the doorbell write by reading the doorbell
register after the write.  This may take time, and it's not necessary to
force the host controller to know about the transfers right away.  Don't
flush the doorbell register writes.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The interrupter register set includes a register that says whether interrupts
are pending for each event ring (the IP bit).  Each MSI-X vector will get its
own interrupter set with separate IP bits.  The status register includes an
"Event Interrupt (EINT)" bit that is set when an IP bit is set in any of the
interrupters.

When PCI interrupts are used, the EINT bit exactly mirrors the IP bit in the
single interrupter set, and it is a waste of time to check both registers when
trying to figure out if the xHC interrupted or another device on the shared IRQ
line interrupted.  Only check the IP bit to reduce register reads.

The IP bit is automatically cleared by the xHC when MSI or MSI-X is enabled.  It
doesn't make sense to read that register to check for shared interrupts (since
MSI and MSI-X aren't shared).  It also doesn't make sense to write to that
register to clear the IP bit, since it is cleared by the hardware.

We can tell whether MSI or MSI-X is enabled by looking at the irq number in
hcd->irq.  If it's -1, we know MSI or MSI-X is enabled.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Now that the event handler functions no longer use xhci_set_hc_event_deq()
to update the event ring dequeue pointer, that function is not used by
anything in xhci-ring.c.  Move that function into xhci-mem.c and make it
static.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The xHCI specification suggests that writing the hardware event ring dequeue
pointer register too often can be an expensive operation for the xHCI hardware
to manage.  It suggests minimizing the number of writes to that register.

Originally, the driver wrote the event ring dequeue pointer after each
event was processed.  Depending on how the event ring moderation register
is set up and how fast the transfers are completing, there may be several
events processed for each interrupt.  This patch makes the hardware event
ring dequeue pointer be written only once per interrupt.

Make the transfer event handler and port status event handler only write
the software event ring dequeue pointer.  Move the updating of the
hardware event ring dequeue pointer into the interrupt function.  Move the
contents of xhci_set_hc_event_deq() into the interrupt handler.  The
interrupt handler must clear the event handler busy flag, so it might as
well also write the dequeue pointer to the same register.  This eliminates
two 32-bit PCI reads and two 32-bit PCI writes.
Reported-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

xhci_handle_event() is now only called from within xhci-ring.c, so make it
static.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Remove a duplicate register read of the interrupt pending register from
xhci_irq().  Also, remove waiting on the posted write of that register.
The host will see it eventually.  It will probably read the register
itself before deciding whether to interrupt the system again, forcing the
posted write to complete.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

When we move xhci_work() into xhci_irq(), we don't need to read the operational
register status field twice.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Most of the work for interrupt handling is done in xhci-ring.c, so it makes
sense to move the functions that are first called when an interrupt happens
(xhci_irq() or xhci_msi_irq()) into xhci-ring.c, so that the compiler can better
optimize them.

Shorten some lines to make it pass checkpatch.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

I've been using perf to measure the top symbols while transferring 1GB of data
on a USB 3.0 drive with dd.  This is using the raw disk with /dev/sdb, with a
block size of 1K.

During performance testing, the top symbol was xhci_triad_to_transfer_ring(), a
function that should return immediately if streams are not enabled for an
endpoint.  It turned out that the functions to find the endpoint ring was
defined in xhci-mem.c and used in xhci-ring.c and xhci-hcd.c.  I moved a copy of
xhci_triad_to_transfer_ring() and xhci_urb_to_transfer_ring() into xhci-ring.c
and declared them static.  I also made a static version of
xhci_urb_to_transfer_ring() in xhci.c.

This improved throughput on a 1GB read of the raw disk with dd from
186MB/s to 195MB/s, and perf reported sampling the xhci_triad_to_transfer_ring()
0.06% of the time, rather than 9.26% of the time.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The patch below on gregkh tree only creates 'lpm' file under
ehci->debug_dir, but not removes it when unloading module,

	 USB: EHCI: EHCI 1.1 addendum: preparation

which can make loading of ehci-hcd module failed after unloading it.

This patch replaces debugfs_remove with debugfs_remove_recursive
to remove ehci debugfs dir and files. It does fix the bug above,
and may simplify the removing procedure.

Also, remove the debug_registers, debug_async and debug_periodic
field from ehci_hcd struct since they are useless now.
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This adds USB 2.0 support to ssb ohci driver.
This patch was used in OpenWRT for a long time now.

CC: Steve Brown <sbrown@cortland.com>
Signed-off-by: NHauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

imx21_hc_reset() uses schedule_timeout() without setting state to
STATE_(UN)INTERRUPTIBLE. As it is called in cycle without checking of
pending signals, use schedule_timeout_uninterruptible().
Signed-off-by: NKulikov Vasiliy <segooon@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If kzalloc() fails exit with -ENOMEM.
Signed-off-by: NKulikov Vasiliy <segooon@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Enable isochronous urb enqueue.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Isochronous endpoint needs a bigger size of transfer ring. Isochronous URB
consists of multiple packets, each packet needs a isoc td to carry, and
there will be multiple trbs inserted to the ring at one time. One segment
is too small for isochronous endpoints, and it will result in
room_on_ring() check failure and the URB is failed to enqueue.

Allocate bigger ring for isochronous endpoint. 8 segments should be enough.
This will be replaced with dynamic ring expansion in the future.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch implements isochronous urb enqueue and interrupt handler part.

When an isochronous urb is passed to xHCI driver, first check the transfer
ring to guarantee there is enough room for the whole urb. Then update the
start_frame and interval field of the urb. Always assume URB_ISO_ASAP
is set, and never use urb->start_frame as input.

The number of isoc TDs is equal to urb->number_of_packets. One isoc TD is
consumed every Interval. Each isoc TD consists of an Isoch TRB chained to
zero or more Normal TRBs.

Call prepare_transfer for each TD to do initialization; then calculate the
number of TRBs needed for each TD. If the data required by an isoc TD is
physically contiguous (not crosses a page boundary), then only one isoc TRB
is needed; otherwise one or more additional normal TRB shall be chained to
the isoc TRB by the host.

Set TRB_IOC to the last TRB of each isoc TD. Do not ring endpoint doorbell
to start xHC procession until all the TDs are inserted to the endpoint
transer ring.

In irq handler, update urb status and actual_length, increase
urb_priv->td_cnt. When all the TDs are completed(td_cnt is equal to
urb_priv->length), giveback the urb to usbcore.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Add urb_priv data structure to xHCI driver. This structure allows multiple
xhci TDs to be linked to one urb, which is essential for isochronous
transfer. For non-isochronous urb, only one TD is needed for one urb;
for isochronous urb, the TD number for the urb is equal to
urb->number_of_packets.

The length field of urb_priv indicates the number of TDs in the urb.
The td_cnt field indicates the number of TDs already processed by xHC.
When td_cnt matches length, the urb can be given back to usbcore.

When an urb is dequeued or cancelled, add all the unprocessed TDs to the
endpoint's cancelled_td_list. When process a cancelled TD, increase
td_cnt field. When td_cnt equals urb_priv->length, giveback the
cancelled urb.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch adds mechanism to process Missed Service Error Event.
Sometimes the xHC is unable to process the isoc TDs in time, it will
generate Missed Service Error Event. In this case some TDs on the ring are
not processed and missed. When encounter a Missed Servce Error Event, set
the skip flag of the ep, and process the missed TDs until reach the next
processed TD, then clear the skip flag.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch adds new cases to trb_comp_code switch, and moves
the switch judgment ahead of fetching td.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Remove redundant print messages in the interrupt context.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch moves the bulk and interrupt td processing part in
handle_tx_event() into a separate function process_bulk_intr_td().
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch moves the ctrl td processing part in handle_tx_event()
into a separate function process_ctrl_td().
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch moves the td universal processing part in handle_tx_event()
into a separate function finish_td().

if finish_td() returns 1, it indicates the urb can be given back.
Signed-off-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1409) removes some dead code from the ehci-hcd
scheduler.  Thanks to the previous patch in this series, stream->depth
is no longer used.  And stream->start and stream->rescheduled
apparently have not been used for quite a while, except in some
statistics-reporting code that never gets invoked.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: David Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1408) rearranges the scheduling code in ehci-hcd, partly
to improve its structure, but mainly to change the way it works.
Whether or not a transfer exceeds the hardware schedule length will
now be determined by looking at the last frame the transfer would use,
instead of the first available frame following the end of the transfer.

The benefit of this change is that it allows the driver to accept
valid URBs which would otherwise be rejected.  For example, suppose
the schedule length is 1024 frames, the endpoint period is 256 frames,
and a four-packet URB is submitted.  The four transfers would occupy
slots that are 0, 256, 512, and 768 frames past the current frame
(plus an extra slop factor).  These don't exceed the 1024-frame limit,
so the URB should be accepted.  But the current code notices that the
next available slot would be 1024 frames (plus slop) in the future,
which is beyond the limit, and so the URB is rejected unnecessarily.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: David Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1407) fixes a bug in ehci-hcd's isochronous scheduler.
All its calculations should be done in terms of microframes, but for
full-speed devices, sched->span is stored in frames.  It needs to be
converted.

This fix is liable to expose problems in other drivers.  The old code
would accept URBs that should not have been accepted, so drivers have
had no reason to avoid submitting URBs that exceeded the maximum
schedule length.  In an attempt to partially compensate for this, the
patch also adjusts the schedule length from a minimum of 256 frames up
to a minimum of 512 frames.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: David Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1406) adds a micro-optimization to ehci-hcd's scheduling
code.  Instead of computing remainders with respect to the schedule
length, use bitwise-and (which is quicker).  We know that the schedule
length will always be a power of two, but the compiler doesn't have
this information.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: David Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1405) fixes a small bug in ehci-hcd's isochronous
scheduler.  Not all EHCI controllers are PCI, and the code shouldn't
assume that they are.  Instead, introduce a special flag for
controllers which need to delay iso scheduling for full-speed devices
beyond the scheduling threshold.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: Sarah Sharp <sarah.a.sharp@linux.intel.com>
CC: David Brownell <david-b@pacbell.net>
CC: stable <stable@kernel.org>
Acked-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>

Enable MSI/MSI-X supporting in xhci driver.

Provide the mechanism to fall back using MSI and Legacy IRQs
if MSI-X IRQs register failed.
Signed-off-by: NDong Nguyen <Dong.Nguyen@amd.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>,
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

CONFIG_ARCH_KARO doesn't exist in Kconfig and is never defined anywhere
else, therefore removing all references for it from the source code.
Signed-off-by: NChristian Dietrich <qy03fugy@stud.informatik.uni-erlangen.de>
Acked-by: NRyan Mallon <ryan@bluewatersys.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Currently devices don't get detected automatically if the ehci
module is inserted 2nd time onward. We need to disconnect and
reconnect the device for it to get detected and enumerated.

Resetting the USB PHY using PHY reset comamnd over ULPI fixes
this issue. Tested on OMAP3EVM.
Signed-off-by: NAjay Kumar Gupta <ajay.gupta@ti.com>
Acked-by: NFelipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If we use the HCD_LOCAL_MEM flag and dma_declare_coherent_memory() to
enforce the host controller's local memory utilization we also need to
disable native scatter-gather support, otherwise hcd_alloc_coherent() in
map_urb_for_dma() is called with urb->transfer_buffer == NULL, that
triggers a NULL pointer dereference.

We can also consider to add a WARN_ON() and return an error code to
better catch this problem in the future.

At the moment no driver seems to hit this bug, so I should
consider this a low-priority fix.
Signed-off-by: NAndrea Righi <arighi@develer.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Change sizeof(x) / sizeof(*x) to ARRAY_SIZE(x).
Signed-off-by: NKulikov Vasiliy <segooon@gmail.com>
Cc: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1396) adds code to uhci-hcd to support the
vendor-specific wakeup settings found in Intel's ICHx hardware.  A
couple of unnecessary memory barriers are removed.  And the root hub
isn't put back into the "suspended" state if power was lost during a
system sleep -- there's not much point in doing so because the root hub
will be resumed shortly.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1394) adds code to ehci-hcd, ohci-hcd, and uhci-hcd for
automatically resuming the root hub when the controller is resumed, if
the root hub has a wakeup request pending on some port.

During resume from system sleep this doesn't matter, because the root
hubs will naturally be resumed along with every other device in the
system.  However it _will_ matter for runtime PM: If the controller is
suspended and a remote wakeup request is received then the controller
will autoresume, but we need to ensure that the root hub also
autoresumes.  Otherwise the wakeup request would be ignored, the
controller would go back to sleep, and the cycle would repeat a large
number of times (I saw this happen before the patch was written).
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1385) adds a "do_wakeup" parameter to the pci_suspend
method used by PCI-based host controller drivers.  ehci-hcd in
particular needs to know whether or not to enable wakeup when
suspending a controller.  Although that information is currently
available through device_may_wakeup(), when support is added for
runtime suspend this will no longer be true.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>