add pci map direction in atl1c_buffer flags, it is used when call pci_unmap
apis.
Signed-off-by: Jie Yang <jie.yang@atheros.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Unified firmware image may not contain MN type of firmware.
Driver should fall back to NOMN firmware type instead
of going to flash.
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
o netif_running() check for enabling interrupt at end of napi poll is
not enough to cover firmwar recovery. Instead test __NX_DEV_UP bit.
o Avoid re-entry into to netxen_nic_down() with __NX_DEV_UP bit check.
Acked-by: Dhananjay Phadke <dhananjay.phadke@qlogic.com>
Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
o To prevent race conditions with other reset events.
During suspend/resume and firmware recovery, acquire rtnl_lock,
while changing interface state.
Acked-by: Dhananjay Phadke <dhananjay.phadke@qlogic.com>
Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Various additions and improvements to the Gigaset driver's README
file, and added comments to its userspace visible include file.
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
When built with debugging support, the Gigaset driver enabled some
debugging messages by default. Change the default to "all off".
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the locking scheme on the fec_mpc52xx driver. This device can
receive IRQs from three sources; the FEC itself, the tx DMA, and the
rx DMA. Mutual exclusion was handled by taking a spin_lock() in the
critical regions, but because the handlers are run with IRQs enabled,
spin_lock() is insufficient and the driver can end up interrupting
a critical region anyway from another IRQ.
Asier Llano discovered that this occurs when an error IRQ is raised
in the middle of handling rx irqs which resulted in an sk_buff memory
leak.
In addition, locking is spotty at best in the driver and inspection
revealed quite a few places with insufficient locking.
This patch is based on Asier's initial work, but reworks a number of
things so that locks are held for as short a time as possible, so
that spin_lock_irqsave() is used everywhere, and so the locks are
dropped when calling into the network stack (because the lock only
protects the hardware interface; not the network stack).
Boot tested on a lite5200 with an NFS root. Has not been performance
tested.
Signed-off-by: Asier Llano <a.llano@ziv.es>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use a more effective rss hash by default (src + dst, rather than just
src).
Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
in routed mode, we don't have a hardware address so netdev_ops doesnt
need to validate our hardware address via .ndo_validate_addr
Reported-by: Manuel Fuentes <mfuentes@agenciaefe.com>
Signed-off-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>
due to reference counting sk_wmem_alloc now has a value of 1 when all
the outstanding data has been sent.
Signed-off-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>
fix oops when initializing lane interfaces. lec should probably be
changed to use alloc_netdev() instead.
Signed-off-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adds kerneldoc for inet_twsk_unhash() & inet_twsk_bind_unhash().
With help from Randy Dunlap.
Suggested-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we find a timewait connection in __inet_hash_connect() and reuse
it for a new connection request, we have a race window, releasing bind
list lock and reacquiring it in __inet_twsk_kill() to remove timewait
socket from list.
Another thread might find the timewait socket we already chose, leading to
list corruption and crashes.
Fix is to remove timewait socket from bind list before releasing the bind lock.
Note: This problem happens if sysctl_tcp_tw_reuse is set.
Reported-by: kapil dakhane <kdakhane@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
First patch changes __inet_hash_nolisten() and __inet6_hash()
to get a timewait parameter to be able to unhash it from ehash
at same time the new socket is inserted in hash.
This makes sure timewait socket wont be found by a concurrent
writer in __inet_check_established()
Reported-by: kapil dakhane <kdakhane@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Acked-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Here is an updated version, because ixgbe_get_ethtool_stats()
needs to call dev_get_stats() or "ethtool -S" wont give
correct tx_bytes/tx_packets values.
Several cpus can update netdev->stats.tx_bytes & netdev->stats.tx_packets
in parallel. In this case, TX stats are under estimated and false sharing
takes place.
After a pktgen session sending exactly 200000000 packets :
# ifconfig fiber0 | grep TX
TX packets:198501982 errors:0 dropped:0 overruns:0 carrier:0
Multi queue devices should instead use txq->tx_bytes & txq->tx_packets
in their xmit() method (appropriate txq lock already held by caller, no
cache line miss), or use appropriate locking.
After patch, same pktgen session gives :
# ifconfig fiber0 | grep TX
TX packets:200000000 errors:0 dropped:0 overruns:0 carrier:0
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A workaround added for all ESB2 devices (adds a delay for all MDIC accesses
which resolves an issue with the MDIC ready bit being set prematurely) is
applicable only to devices in which the MAC-PHY interconnect is not
operating in a certain mode with in-band MDIO. Check the control register
for the operating mode and enable the workaround accordingly.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The GG82563_REG() macro should not be used to determine the offset provided
to the e1000e_[read|write]_kmrn_reg() functions since the first argument to
the macro is already implied and gets masked off anyway in the functions.
The resultant register reads/writes with this patch are functionally the
same as before.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Bit 7 in the CTRL_REG register is actually the Software Definable Pin 3,
not the Software Definable Pin 7.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
GCC even warns about it, as reported by Andrew Morton:
net/ipv4/tcp.c: In function 'do_tcp_getsockopt':
net/ipv4/tcp.c:2544: warning: comparison is always false due to limited range of data type
Signed-off-by: David S. Miller <davem@davemloft.net>
Entries in the i2c table aren't always ordered
by id. This allows us to remove some quirks
that are no longer needed.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
entries in the ss table aren't always ordered
by id.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
- Don't add dac load detection property to DVI-D
- Make sure i2c info is valid before adding DP aux chan bus
- Don't create scaling_mode_property twice
- fix typo that prevented coherent and load detection from working
- add coherent prop to DP (for dp->dvi adapters)
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
* 'timers-for-linus-ntp' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
ntp: Provide compability defines (You say MOD_NANO, I say ADJ_NANO)
* 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
genirq: do not execute DEBUG_SHIRQ when irq setup failed
Sometimes we will use a crtc for integerated LVDS, which is different with
that assigned by BIOS. If we want to get flicker-free transitions,
then we could read out the current state for it and set our current state
accordingly.
But it is true that if we aren't reading current state out, we do need
to turn everything off before modesetting. Otherwise the clocks can get very
angry and we get things worse than a flicker at boot.
In fact we also do the similar thing in UMS mode. We will disable all the
possible outputs/crtcs for the first modesetting.
So we disable all the possible outputs/crtcs before entering the KMS mode.
Before we configure connector/encoder/crtc, the function of
drm_helper_disable_unused_function can disable all the possible outputs/crtcs.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Rafal Milecki <zajec5@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
* 'timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
timers, init: Limit the number of per cpu calibration bootup messages
posix-cpu-timers: optimize and document timer_create callback
clockevents: Add missing include to pacify sparse
x86: vmiclock: Fix printk format
x86: Fix printk format due to variable type change
sparc: fix printk for change of variable type
clocksource/events: Fix fallout of generic code changes
nohz: Allow 32-bit machines to sleep for more than 2.15 seconds
nohz: Track last do_timer() cpu
nohz: Prevent clocksource wrapping during idle
nohz: Type cast printk argument
mips: Use generic mult/shift factor calculation for clocks
clocksource: Provide a generic mult/shift factor calculation
clockevents: Use u32 for mult and shift factors
nohz: Introduce arch_needs_cpu
nohz: Reuse ktime in sub-functions of tick_check_idle.
time: Remove xtime_cache
time: Implement logarithmic time accumulation
* 'timers-for-linus-hpet' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: hpet: Make WARN_ON understandable
x86: arch specific support for remapping HPET MSIs
intr-remap: generic support for remapping HPET MSIs
x86, hpet: Simplify the HPET code
x86, hpet: Disable per-cpu hpet timer if ARAT is supported
Because of OMAP off-mode, powerdomain can go off when I2C is idle.
Save enough state, and do a re-init for each transfer.
Additional save/restore state added by Jagadeesh Bhaskar Pakaravoor
(SYSC_REG) and Aaro Koskinen (wakeup sources.)
Also, The OMAP3430 TRM states:
"During active mode (I2Ci.I2C_CON[15] I2C_EN bit is set to 1), make no
changes to the I2Ci.I2C_SCLL and I2Ci.I2C_SCLH registers. Changes may
result in unpredictable behavior."
Hence, the I2C_EN bit should be clearer when modifying these
registers. Please note that clearing the entire I2C_CON register to
disable the I2C module is safe, because the I2C_CON register is
re-configured for each transfer.
Signed-off-by: Jouni Hogander <jouni.hogander@nokia.com>
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Cc: Jagadeesh Bhaskar Pakaravoor <j-pakaravoor@ti.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Jon Hunter <jon-hunter@ti.com>
Cc: Hu Tao <taohu@motorola.com>
Cc: Xiaolong Chen <A21785@motorola.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
In the case of no-ACKs, we don't want to see dev_err() messages in the
console, because some utilities like i2c-tools are capable of printing
decorated console output. This patch will ease such situations.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
* ABRT_MASTER_DIS: Fix a typo.
* i2c_dw_handle_tx_abort: Return an appropriate error number
depending on abort_source.
* i2c_dw_xfer: Add a missing abort_source initialization.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Suppose TX_ABRT occurs in the middle of processing i2c_msg msgs[], and
a STOP condition has already been generated on the bus. In this case,
subsequent i2c_dw_xfer_msg() might initiate a new and unnecessary I2C
transaction, which we'd have to avoid.
Furthermore, anytime TX_ABRT is set, the contents of tx/rx buffers are
flushed, so we don't have to process RX_FULL and TX_EMPTY.
Disable interrupts, and skip them.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Current error handling procedures are not good in two respects:
* Forgot to mark dev->cmd_complete as "completed" on errors
Once an I2C transaction is initiated, wait_for_completion_
interruptible_timeout() waits for dev->cmd_complete to be completed.
We have to take care of it whenever an error is detected, otherwise
we will have a needless HZ timeout.
* Forgot to disable interrupts
In the previous patch, interrupt mask operations have been changed.
We don't disable interrupts at the end of the interrupt handler any
more, and try to keep RX_FULL (and TX_EMPTY if required) enabled
during the transaction so that we can send longer data than the size
of Tx/Rx FIFO.
If an error is detected, we need to disable interrupts before
quitting current transaction.
We can work around above points using dev->msg_err effectively.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Currently we disable TX_EMPTY interrupt when buf_len is zero, but this
is wrong. (buf_len == 0) means that all transmit data in the current
i2c_msg message has been sent out, but that doesn't necessarily mean
all i2c_msg messages have been processed.
TX_EMPTY interrupt is used as the driving force of DW I2C transactions,
so we need to keep it enabled as long as i2c_msg messages are available.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Currently we process the first i2c_dw_xfer_msg() in i2c_dw_xfer(),
but in this case there is a possibility to be interrupted by certain
interrupts. As described before in this patchset, we need to keep
providing new transmit data within a given time period, otherwise Tx
FIFO underrun takes place and STOP condition will be generated on the
bus, even if we have more bytes to be written.
In order to exclude all such possibilities, change TX_EMPTY interrupt
usage as below:
* DW_IC_INTR_DEFAULT_MASK: Define a default interrupt mask set, and
put TX_EMPTY there.
* i2c_dw_xfer_init: Enable DW_IC_INTR_DEFAULT_MASK prior to initiating
a new I2C transaction. The first TX_EMPTY will be triggered shortly.
With the help of it, we can make the first call to i2c_dw_xfer_msg()
in the interrupt handler.
* i2c_dw_xfer_msg: Fixup intr_mask operation accordingly. Make sure
that TX_EMPTY operations need to be reversed.
* request_irq: Set IRQF_DISABLED so that we could load transmit data
into Tx FIFO without being distracted by other interrupts.
* Remove i2c_dw_xfer_msg() in i2c_dw_xfer().
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
I2c_dw_xfer_msg() also has the same target address inconsistency check,
and furthermore it checks across all i2c_msg messages, while
i2c_dw_read() walks through i2c_msg messages only with_ I2C_M_RD flag.
That is, target address check in i2c_dw_read() is redundant and useless.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Set proper I2C_FUNC_SMBUS_* bits so that the driver could be used with
some utilities requiring SMBus functionalities, such as i2c-tools.
Note that DW I2C core doesn't support I2C_FUNC_SMBUS_QUICK, as it's not
capable of zero-length data transactions.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
As the driver and hardware always process the given data in parallel,
then it would be better to initialize tx_limit, rx_limit and rx_valid
variables just prior to being used.
This will help us to send / receive as much data as possible.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
While we have a local variable "buf_len" for dev->tx_buf_len, we don't
have such local variable for dev->tx_buf pointer. While "buf_len" is
restored at first then updated when we start processing a new i2c_msg
(determined by STATUS_WRITE_IN_PROGRESS flag), ->tx_buf is different.
Such inconsistency makes the code slightly hard to follow.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
We have some steps at the top of i2c_dw_xfer_msg() to set up a slave
address and enable DW I2C core. And it's executed only when we don't
have STATUS_WRITE_IN_PROGRESS.
But we need to make sure that STATUS_WRITE_IN_PROGRESS only indicates
that we have a pending i2c_msg to process. In other words, even if
STATUS_WRITE_IN_PROGRESS is not set, that doesn't mean we're at initial
state in the I2C transaction.
Since i2c_dw_xfer_msg() will be invoked again and again during a
transaction, those init steps have a possibility to be re-processed
needlessly. For example, this issue easily takes place when processing
a combined transaction with a certain condition (the number of tx bytes
in the first i2c_msg, equals to the Tx FIFO depth).
Consequently we should not use STATUS_WRITE_IN_PROGRESS to determine
where we're at in an I2C transaction. It would be better to separate
those initialization steps from i2c_dw_xfer_msg().
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Enable RX_FULL interrupt mask by default, and hook it in the interrupt
handler. If requested amount of rx data (defined by IC_RX_TL) is not
available, we don't have to process i2c_dw_read().
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
As a hardware feature, DW I2C core generates a STOP condition whenever
the Tx FIFO becomes empty (strictly speaking, whenever the last byte in
the Tx FIFO is sent out), even if we have more bytes to be written.
In other words, we must never make "Tx FIFO underrun" happen during
a transaction, except for the last byte. For the safety's sake, we'd
make TX_EMPTY interrupt get triggered every time one byte is processed.
The Rx FIFO threshold needs to be set as well.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Symptom:
--------
When we're going to send/receive the longer size of data than the Tx
FIFO length, the I2C transaction will be divided into several separated
transactions, limited by the Tx FIFO length.
Details:
--------
As a hardware feature, DW I2C core generates a STOP condition whenever
the Tx FIFO becomes empty (strictly speaking, whenever the last byte in
the Tx FIFO is sent out), even if we have more bytes to be written.
Then, once a new transmit data is written to the Tx FIFO, DW I2C core
will initiate a new transaction, which leads to another START condition.
This explains how the transaction in question goes, and implies that
current tasklet-based dw_i2c_pump_msg() strategy couldn't meet the
timing constraint required for avoiding Tx FIFO underrun.
To avoid this scenario, we must keep providing new transmit data within
a given time period. In case of Fast-mode + 32-byte Tx FIFO, for
instance, it takes about 22.5[us] to process single byte, and 720[us] in
total.
This patch removes the existing tasklet-based "pump" system, and move
its jobs into the interrupt handler.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
In case a work-in-progress i2c_msg has more bytes to be written, we
need to set STATUS_WRITE_IN_PROGRESS and exit from the msg_write_idx-
searching loop. Otherwise, we will overtake the current msg_write_idx
without waiting for its transmission to be processed.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
* Calculate with accurate conditional expressions from DW manuals.
* Round ic_clk by adding 0.5 as it's important at high ic_clk rate.
* Take into account "tHD;STA" issue for _HCNT calculation.
* Take into account "tf" for _LCNT calculation.
* Add "cond" and "offset" fot further correction requirements.
For _HCNT calculation, there's one issue needs to be carefully
considered; DesignWare I2C core doesn't seem to have solid strategy
to meet the tHD;STA timing spec. If you configure _HCNT based on the
tHIGH timing spec, it easily results in violation of the tHD;STA spec.
After many trials, we came to the conclusion that the tHD;STA period
is proportional to (_HCNT + 3). For the safety's sake, this should be
selected by default.
As for _LCNT calculation, DW I2C core has one characteristic behavior;
he starts counting the SCL CNTs for the LOW period of the SCL clock
(tLOW) as soon as it pulls the SCL line. At that time, he doesn't take
into account the fall time of SCL signal (tf), IOW, he starts counting
CNTs without confirming the SCL input voltage has dropped to below VIL.
This characteristics becomes a problem on some platforms where tf is
considerably long, and results in violation of the tLOW timing spec.
To make the driver configurable as much as possible for various cases,
we'd have separated arguments "tf" and "offset", and for safety default
values should be 0.3 us and 0, respectively.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
