The ready message from the target could be processed
before the host HW init has completed. In this case,
htc_process_target_rdy() would assume the target has timed
out, when it hasn't. Fix this by checking if the target
has sent the ready message properly.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The HTC state has to be setup before initializing
the target because the ready message could possibly
come before the control endpoints in HTC have been
identified.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The URBs have to be allocated before uploading
the firmware to the target. This is needed to process
the target ready message properly.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Check the mode in channel switch ie for either 0 or 1 on transmission.
A channel switch mode set to 1 means that the STA in a BSS to which the
frame containing the element is addressed shall transmit no further
frames within the BSS until the scheduled channel switch.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This adds support for offloading the channel switch
operation to devices that support such, typically
by having specific firmware API for it. The reasons
for this could be that the firmware provides better
timing or that regulatory enforcement done by the
device requires special handling of CSAs.
In order to allow drivers to specify the timing to
the device, the new channel_switch callback will
pass through the received frame's mactime, where
available.
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When we process a frame, we currently just match it
to the work struct by the MAC addresses, and not by
the work type. This means that we can end up doing
the work for an association request item when (for
whatever reason) we receive another frame type, for
example a probe response. Processing the wrong type
of frame will lead to completely invalid data being
processed, and will lead to various problems like
thinking the association was successful even if the
AP never sent an assocation response.
Fix this by making each processing function check
that it is invoked for the right work struct type
only and continue processing otherwise (and drop
frames that we didn't expect).
This bug was uncovered during the debugging for
https://bugzilla.kernel.org/show_bug.cgi?id=15862
but doesn't seem to be the cause for any of the
various problems reported there.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The generic gpiolib provides a debugfs interface to GPIOs which
provides identical (but nicer looking) data as the OMAP specific one.
This patch completely drops the OMAP specific interface
(/debug/omap_gpio) in favor of using the generic one (/debug/gpio.)
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
Ensure GPIO debounce clocks are disabled when idle. Otherwise,
clocks will prevent PER powerdomain from entering retention.
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
Currently, the GPIO 'prepare' hook is only called when going to
off-mode, while the function is called 'prepare_for_retention.' This
patch renames the function to 'prepare_for_idle' and calls it for any
powersate != PWRDM_POWER_ON passing in the powerstate.
The hook itself is then responsible for doing various preparation
based on the powerstate.
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
According to the GPIO 'Wakeup and Interrupt' section of the TRM[1],
wake-up requests can only be generated on edge transitions.
Also for OMAP3, only edge GPIOs may lose interrupts when PER enters
RET/OFF state, this is addressed by gpio prepare|resume idle functions
[1] Section 25.5.3.1 OMAP34xx_ES3.1_TRM_V_Q
Signed-off-by: Chunqiu Wang <cqwang@motorola.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
Off mode is now using the omap2 retention fix code for scanning GPIOs
during off-mode transitions. All the *non_wakeup_gpios variables
are now used for off-mode transition tracking on OMAP3. This patch fixes
cases where GPIO state changes are missed during off-mode.
Signed-off-by: Tero Kristo <tero.kristo@nokia.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
We can remove this wakeup dependency since now, when
GPIO2-6 are enabled for IO-pad wakeup, PER domain is gauranteed
to be awake or be woken up to service.
The previous dependency did not handle all corner cases. Since there
was no sleep dependency between CORE and PER domains, if PER enters
RET and CORE is ON, PER will not be active for GPIO handling.
Signed-off-by: Mike Chan <mike@android.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
IO events can also come from GPIO modules, which reside in the PER domain.
It is possible for the PER to enter RET while CORE is still in ON.
If GPIO 2-6 are enabled for IO-pad wakeups, the PER domain will not
wakeup in this case, unless we enable it.
Signed-off-by: Mike Chan <mike@android.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
This patch enables the wakeup capabilities of ads7846 touchscreen driver.
ads7846 driver can now wakeup the system from suspend on OMAP3430 EVM
and SDP boards.
The earlier approach of enabling wakeup on the touchscreen GPIO pin during
board level mux init is removed. Instead the wakeup flag in
ads7846_platform_data is enabled. Based on the flag, the ads7846 driver
will do an enable_irq_wake which will eventually call into the OMAP GPIO
layer and will enable the wakeup capability on the GPIO pin.
Signed-off-by: Ranjith Lohithakshan <ranjithl@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
OMAP3530 TRM section 7.4.4.4.2 requires OFFOUTENABLE to be set (active low)
if wakeup capabilities are enabled on a pad. During OFF mode testing
on OMAP3530 EVM, it was observed that the device was not residing in
the OFF state. The device enters into the OFF state and immediately exits
from that state as if an IO wakeup event has occured. The issue was traced
down to the pad configuration of wkaeup enabled pad's.
Signed-off-by: Ranjith Lohithakshan <ranjithl@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
While handling PRCM IRQs, mask out interrupts that are not enabled in
PRM_IRQENABLE_MPU. If these are not masked out, non-enabled
interrupts are caught, a WARN() is printed due to no 'handler' and the
events are cleared. In addition to being noisy, this can also
interfere with independent polling of this register by SR/VP code.
This was noticed using SmartReflex transitions which cause the VPx_*
interrupts to be handled since they are set in PRM_IRQSTATUS_MPU even
but not enabled in PRM_IRQENABLE_MPU.
Acked-by: Mike Turquette <mturquette@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
We shouldn't free things here because we free them later.
The call tree looks like this:
iser_connect() ==> initiating the connection establishment
and later
iser_cma_handler() => iser_route_handler() => iser_create_ib_conn_res()
if we fail here, eventually iser_conn_release() is called, resulting
in a double free.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The iser connection teardown flow isn't over until the underlying
Connection Manager (e.g the IB CM) delivers a disconnected or timeout
event through the RDMA-CM. When the remote (target) side isn't
reachable, e.g when some HW e.g port/hca/switch isn't functioning or
taken down administratively, the CM timeout flow is used and the event
may be generated only after relatively long time -- on the order of
tens of seconds.
The current iser code exposes this possibly long delay to higher
layers, specifically to the iscsid daemon and iscsi kernel stack. As a
result, the iscsi stack doesn't respond well: this low-level CM delay
is added to the fail-over time under HA schemes such as the one
provided by DM multipath through the multipathd(8) service.
This patch enhances the reference counting scheme on iser's IB
connections so that the disconnect flow initiated by iscsid from user
space (ep_disconnect) doesn't wait for the CM to deliver the
disconnect/timeout event. (The connection teardown isn't done from
iser's view point until the event is delivered)
The iser ib (rdma) connection object is destroyed when its reference
count reaches zero. When this happens on the RDMA-CM callback
context, extra care is taken so that the RDMA-CM does the actual
destroying of the associated ID, since doing it in the callback is
prohibited.
The reference count of iser ib connection normally reaches three,
where the <ref, deref> relations are
1. conn <init, terminate>
2. conn <bind, stop/destroy>
3. cma id <create, disconnect/error/timeout callbacks>
With this patch, multipath fail-over time is about 30 seconds, while
without this patch, multipath fail-over time is about 130 seconds.
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The iscsi connection object life cycle includes binding and unbinding
(conn_stop) to/from the iscsi transport connection object. Since
iscsi connection objects are recycled, at the time the transport
connection (e.g iser's IB connection) is released, it is not valid to
touch the iscsi connection tied to the transport back-pointer since it
may already point to a different transport connection.
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add handler to handle events such as port up and down. This is useful
when testing high-availability schemes such as multi-pathing.
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This fixes the following warning,
arch/arm/mach-msm/smd_debug.c:240: warning: initialization from incompatible pointer type
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
Without a debug uart selected you get this failure,
arch/arm/mach-msm/board-trout.c:85: error: 'MSM_DEBUG_UART_PHYS' undeclared here (not in a function)
arch/arm/mach-msm/board-trout.c:86: error: 'MSM_DEBUG_UART_BASE' undeclared here (not in a function)
This just removes these lines in that case.
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
This removes the Kconfig menu option. SMD can still be selected
but it's done inside the Kconfig file and not via the menu.
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
When booting up we need to wait for the modem processor to
partially boot. This is because the modem processor does
resource allocation for us. If we don't wait the modem won't
honor our requests and we end up crashing or in an unknown
state. This change just formalizes the waiting process.
Signed-off-by: Daniel Walker <c_dwalke@quicinc.com>
This irq handler isn't used in all cases, so add the proper ifdef. This
eliminates a compiler warning due to the function not getting used.
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
This just removed some unneeded predefines. One needed a whole
function moved down further. The others could just be deleted.
No functional changes.
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
This moves the msm_a2m_int() function into the header, and
does a small macro clean up to be more inline with Linux
norms. No functional changes.
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
"unsigned" translates to "unsigned int", but this value holds an
address. We always want to use unsigned long for addresses since
it will change size to fit the machine.
This just convert the one address holder to unsigned long.
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
This modifies SMD to use either the package v3 or package v4,
but not both. The current code tries to allocate as v4 on all
system which can produce a scary looking error message on boot up,
smem_find(16, 40): wrong size 16424
smd_alloc_channel() cid=02 size=08192 'SMD_RPCCALL'
With this error the code then falls back on the package v3 allocation
method. This method is inefficient because it causes a slow down
on some systems even when the allocation method can be determined
at compile time. It also causes a kernel size increase that effects
all system and is not needed.
This change corrects the allocation to use one method or the other
and not both.
Signed-off-by: Daniel Walker <c_dwalke@quicinc.com>
Forcing the alignment prevents gcc from generating byte reads for word
member variables. Lack of this caused issues when the app processor
modified struct members and the modem saw a partial word write.
Signed-off-by: Dima Zavin <dima@android.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
Some smd clients may write from multiple threads, in which case it's
not safe to call smd_write without holding a lock. smd_write_atomic()
provides the same functionality as smd_write() but obtains the smd
lock first.
Signed-off-by: Brian Swetland <swetland@google.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
Don't mark a channel as allocated if we failed to allocate it
(perhaps the modem updated one table but not the other, etc)
Signed-off-by: Brian Swetland <swetland@google.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
When we read data out of the sender's fifo, we need to advance the sender's
tail pointer, not the receiver's.
Signed-off-by: Haley Teng <Haley_Teng@htc.com>
Acked-by: Dima Zavin <dima@android.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
- QSD8250 has a DSP that speaks SMD, in addition to the modem
- handle a separate list of modem vs dsp channels
- install dsp smd irq handler as necessary
Signed-off-by: Brian Swetland <swetland@google.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
- pull debug code into smd_debug.c
- move necessary structures and defines into smd_private.h
- fix some comment formatting, etc
Signed-off-by: Brian Swetland <swetland@google.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
The new protocol require writing to two state fields, and reading
several fields.
Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
- support both v2 and v1 style smd channels
- support both v2 and v1 smsm shared state
- update smsm state defines and smem item enum
- prep work for dealing with smd to qdsp6
- simplify some smem access to minimize use of smem_alloc() at runtime
Signed-off-by: Brian Swetland <swetland@google.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
This code provides the low level interface to the "shared memory
state machine" (smsm), and the virtual serial channels (smd), used
to communicate with the baseband processor. Higher level transports
(rpc, ethernet, AT command channel, etc) ride on top of this.
Signed-off-by: Brian Swetland <swetland@google.com>
Also, convert all SDCC IRQ resources to be named. No longer pass status_irq
in the platform_data
Signed-off-by: Dima Zavin <dima@android.com>
Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
