This patch clears the irqstatus register with the exact same events it
has read from it. Since the read-write operation is not atomic, a new
irqstatus bit could have been set in between these operations and would
then be cleared accidentally.
Secondly, we now don't need any spin lock protection when
scheduling/completing napi poll as the isr will not execute anymore (as
we turn off all interrupts now).
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch modifies the throughput mode poll settings to reduce the
number of interrupts. This is only used by older hardware that need a
timer irq in throughput mode.
Secondly, this patch increases the default rx ring from 128 to 512. This
drastically improves bandwidth utilization for small packets sizes i.e
512 bytes.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the logic to moderate the interrupts by changing the
mode between throughput and poll. If there has been a large amount of
time without any burst of network load, the code will transition to pure
throughput mode (where each tx/rx/other will cause an interrupt). If
bursts of network load occurs, it will transition to poll based mode to
help reduce cpu utilization (it will not interrupt on each packet) while
maintaining the optimum network bandwidth utilization.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is only a subset of changes so that it is easier to see the
modifications. This patch removes the isr 'for' loop and shifts all the
logic to account for new tab spacing.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A new optimization mode called Dynamic has been added. This will be mode
where interrupt moderation logic will dynamically switch between pure
throughput mode and poll based (called 'cpu') mode.
Also, for newer chipsets, the timer irq is not needed for throughput
mode. Secondly, since we are modifying the irqmask to change between
modes, msix is not supported.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The napi poll routine has been modified to handle all interrupt events
and process them accordingly. Therefore, the ISR will now only schedule
the napi poll and disable all interrupts instead of just disabling rx
interrupt.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are two tx_done routines to handle tx completion processing. Both
these functions now take in a limit value and return the amount of tx
completions. This will be used by a future patch to determine the total
amount of work done.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch removes unnecessary overhead code. Firstly, there is no nead
to mask off unwanted interrupts as we will be checking against the
irqmask field anyways. Secondly, there has been no value in last few
years from detecting error or unknown interrupts.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch will save the irq events in the driver's context so that the
napi routine knows which interrupts have occurred. Subsequent changes
will be moving all interrupt processing into the napi poll routine.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch removes support for msix running in conjunction with napi.
There has been reported issues regarding the behaviour of irqmask and
generation of interrupts by the HW when in MSIX mode. When running napi,
the driver is constantly turning off/on the irqmask. For the time being,
I am going to disable it until I can root cause the issue.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds missing napi enable/disable calls.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Newer versions of the stats feature would not encompass all older
versions. This would result in only retreiving a subset of all available
stats in HW.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Protocols that use packet_type can be __read_mostly section for better
locality. Elminate any unnecessary initializations of NULL.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove huge board config structure from each instance, read
only necessary fields from flash.
Replace board_type with port_type (1G/10G), there's another
board_type field describing card type (SFP/XFP/CX4).
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MAX_RCV_CTX was set to 1, there's only rx context per
PCI function.
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rearrange open and close into hardware attach(), detach() and
nic up() and down(). this will be used for suspend/resume
subsequently.
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
o remove unused rx fragment handling code.
o imporove check for status descriptor ownership.
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There was a bug, which occasionally caused failure in PRAM initialization after
the cold boot.
Also incremented version number to 1.45.27.
Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adding a proper cast to the argument of PAGE_ALIGN macro so that the output
won't depend on its original type. Without this cast aligned value will be
truncated to the size of the argument type.
Reported-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com>
Tested-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sge_buff_size may not be more than 0xffff.
Reported-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com>
Tested-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This was introduced in an earlier net-next patch.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With some asic configurations xmit of frames smaller than 60 bytes may
fail.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Moving netif_napi_del() up the call chain so it will get called from all
exit points.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The completion interrupt disable routine is only called from the ISR, so
there is no need for irqsave/restore.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Chip does not do UDP checksum when fragmentation occurs.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Get rid of extraneous ql_cycle_adapter. It's only called from the one
place.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes a number of race conditions in the driver.
Up until now, "entry" pointer was initialized before acquiring the right lock.
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This fixes compilation, if the PCI core is disabled.
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
debugfs ath9k/phy#/wiphy can be used to show the current list of
virtual wiphys and to add/remove virtual wiphys. Eventually, this
interface could be replaced with a cfg80211/nl80211 command that is
passed through mac80211.
For example:
# cat /debug/ath9k/phy0/wiphy
primary: phy0
# echo add > /debug/ath9k/phy0/wiphy
# cat /debug/ath9k/phy0/wiphy
primary: phy0
secondary: phy1
# echo del=phy1 > /debug/ath9k/phy0/wiphy
# cat /debug/ath9k/phy0/wiphy
primary: phy0
In addition, following commands can be used to test pausing and
unpausing of the virtual wiphys:
pause=phy1
unpause=phy1
select=phy1
(select pauses and unpauses wiphys automatically based on channel)
schedule=500
(set wiphy scheduling interval in msec; 0 = disable; default value: 500)
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This is a very simple scheduler that goes through the wiphys and
schedules one at a time every N milliseconds (current default value:
500 ms). This is enough for initial testing, but there are number of
areas where a more complex scheduler can improve operations greatly.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Allow mac80211-controlled channel changes on an active wiphy and
especially during a scan. We need this as long as the scan is
controlled by mac80211. Moving this control into the driver could
allow some optimizations on scanning while using multiple virtual
interfaces, but for now, try to work as well as possible with the
current scan mechanism.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
It looks like channel change may fail in some cases and end up leaving
the hardware in state where it cannot transmit any frames. Add a
workaround to recover from this state if we detect that wiphy
selection is failing due to wiphys not leaving PAUSING state.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211 should not be requesting us to transmit frames on paused
wiphys since we stop the TX queues. Just in case, add debug code to
make sure we catch if this were to happen.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
For now, allow channel changes immediately and just force the other
virtual wiphys to paused state. This is needed to allow
mac80211-controlled scan to control channel changes.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Notify the AP that we may be sleeping longer to allow the AP power
save code to buffer larger number of frames for us when using virtual
wiphys.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Instead of always going through initialization/deinitialization steps,
do this only for the first/last wiphy to not break the other wiphys.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k_wiphy_select() can be used to select a virtual wiphy to be
activated. Other virtual wiphys will be paused and once that is done,
the operational channel is changed and the wiphys that are on the
selected channel will be unpaused.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Allow virtual wiphys to be paused/unpaused to allow off-channel
operations. Pause will stop all TX queues for the wiphy and move the
STA into power save mode if in managed mode. Unpause wakes up the TX
queues and notifies the AP that the STA woke up if in managed mode.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Allow RX filter to pass through all broadcast/multicast frames (i.e.,
no BSSID filtering) if virtual interfaces are used. Software filtering
will be used in this case to drop broadcast/multicast frames for
foreign BSSIDs.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The new struct ath_softc::sec_wiphy array is used to store information
about virtual wiphys and select which wiphy is used in calls to
mac80211. Each virtual wiphy will be assigned a different MAC address
based on the virtual wiphy index.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This is the initial step in allowing ath9k to register multiple
virtual radios (wiphys). The goal of virtual radios is to allow the
same radio to be shared for multiple virtual interfaces that may
operate on different channels. The mac80211 virtual interface support
is designed only for single channel operation and as such, it is not
suitable for this type of use. Anyway, it can be used on top of the
virtual radio concept, if desired (e.g., use two virtual radios to
handle two channels and then add multiple mac80211 virtual interfaces
on top of each virtual radio).
The new struct ath_wiphy is now registered as the driver data
structure for wiphy. This structure has a pointer to the shared (among
virtual wiphys of the same physical radio) struct ath_softc data. The
primary wiphy maintains the allocated memory for ath_softc. Secondary
(virtual) wiphys will only allocate the new ath_wiphy structure.
Registration of secondary wiphys is added in a separate patch.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Instead of using a hardcoded BSSID mask (mask for own addresses),
iterate through all active interfaces and determine the minimal mask
that covers all local addresses.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ayaz Abdulla
Stephen Hemminger
Dhananjay Phadke
David S. Miller
Eilon Greenstein
Ron Mercer
Christian Lamparter
Michael Buesch
Jouni Malinen