stmmac: fix stmmac_resume removing not yet used shutdown flag
The commit to convert to use the dev_pm_ops struct
introduces a bug. The shutdown flag is not yet used
because the hibernation on memory is done by using
the freeze callback.
Thanks to Vlad for having reported it.
Breno Leitao [Thu, 25 Nov 2010 07:53:55 +0000 (07:53 +0000)]
cxgb3: Removing unused return variable
Currently the ret variable is not used for anything other than
receive the value of the t3_adapter_error(), which will always be 0,
because the reset parameter is 0.
Jozsef Kadlecsik [Thu, 25 Nov 2010 03:15:07 +0000 (03:15 +0000)]
ipv6: Prepare the tree for un-inlined jhash.
jhash is widely used in the kernel and because the functions
are inlined, the cost in size is significant. Also, the new jhash
functions are slightly larger than the previous ones so better un-inline.
As a preparation step, the calls to the internal macros are replaced
with the plain jhash function calls.
Jonas Bonn [Thu, 25 Nov 2010 02:30:30 +0000 (02:30 +0000)]
ethoc: rework mdio read/write
MDIO read and write were checking whether a timeout had expired to determine
whether to recheck the result of the MDIO operation. Under heavy CPU usage,
however, it was possible for the timeout to expire before the routine got
around to be able to check a second time even, thus erroneousy returning an
-EBUSY.
This patch changes the the MDIO IO routines to try up to five times to complete
the operation before giving up, thus lessening the dependency on CPU load.
This resolves a problem whereby a ping flood would keep the CPU so busy that
the above problem would manifest itself; the MDIO command to check link status
would fail and the interface would erroneously be shut down.
Jonas Bonn [Thu, 25 Nov 2010 02:30:29 +0000 (02:30 +0000)]
ethoc: rework interrupt handling
The old interrupt handling was incorrect in that it did not account for the
fact that the interrupt source bits get set irregardless of whether or not
their corresponding mask is set. This patch fixes that by masking off the
source bits for masked interrupts.
Furthermore, the handling of transmission events is moved to the NAPI polling
handler alongside the reception handler, thus preventing a whole bunch of
interrupts during heavy traffic.
Jonas Bonn [Thu, 25 Nov 2010 02:30:28 +0000 (02:30 +0000)]
ethoc: Double check pending RX packet
An interrupt may occur between checking bd.stat and clearing the
interrupt source register which would result in the packet going totally
unnoticed as the interrupt will be missed. Double check bd.stat after
clearing the interrupt source register to guard against such an
occurrence.
Adam Edvardsson [Thu, 25 Nov 2010 02:30:27 +0000 (02:30 +0000)]
ethoc: enable interrupts after napi_complete
Occasionally, it seems that some race is causing the interrupts to not be
reenabled otherwise with the end result that networking just stops working.
Enabling interrupts after calling napi_complete is more in line with what
other drivers do.
Jonas Bonn [Thu, 25 Nov 2010 02:30:25 +0000 (02:30 +0000)]
ethoc: Add device tree configuration
This patch adds the ability to describe ethernet devices via a flattened
device tree. As device tree remains an optional feature, these bits all
need to be guarded by CONFIG_OF ifdefs.
MAC address is settable via the device tree parameter "local-mac-address";
however, the selection of the phy id is limited to probing, for now.
Michael Chan [Wed, 24 Nov 2010 13:48:55 +0000 (13:48 +0000)]
bnx2: Remove config access to non-standard registers
In KVM passthrough mode, the driver may not have config access to
non-standard registers. The BNX2_PCICFG_MISC_CONFIG config register
access to setup mailbox swapping can be done using MMIO.
Eddie Wai [Wed, 24 Nov 2010 13:48:54 +0000 (13:48 +0000)]
bnx2: Fix reset bug on 5709
The 5709 chip requires the BNX2_MISC_NEW_CORE_CTL_DMA_ENABLE bit to be
cleared and polling for pending DMAs to complete before chip reset.
Without this step, we've seen NMIs during repeated resets of the chip.
Rajesh Borundia [Tue, 23 Nov 2010 03:08:27 +0000 (03:08 +0000)]
netxen: avoid using reset_devices as it may become obsolete
In kdump environment do not depend on reset_devices
parameter to reset the device as the parameter may become obsolete.
Instead use an adapter specific mechanism to determine if the device
needs a reset.
Driver maintains a count of number of pci functions probed
and decrements the count when remove handler of that pci function
is called. If the first probe, probe of function 0,
detects the count as non zero then reset the device.
Rajesh Borundia [Tue, 23 Nov 2010 01:25:21 +0000 (01:25 +0000)]
qlcnic: avoid using reset_devices as it may become obsolete.
In kdump environment do not depend upon reset_devices parameter
to reset the pci function as this parameter may become obsolete.
Instead use an adapter specific mechanism to determine if the pci
function needs to be reset.
Per function refcount is maintained in driver, which is set in probe
and reset in remove handler of adapter. If the probe detects the count
as non zero then reset the function.
Thomas Graf [Mon, 22 Nov 2010 01:31:54 +0000 (01:31 +0000)]
rtnl: make link af-specific updates atomic
As David pointed out correctly, updates to af-specific attributes
are currently not atomic. If multiple changes are requested and
one of them fails, previous updates may have been applied already
leaving the link behind in a undefined state.
This patch splits the function parse_link_af() into two functions
validate_link_af() and set_link_at(). validate_link_af() is placed
to validate_linkmsg() check for errors as early as possible before
any changes to the link have been made. set_link_af() is called to
commit the changes later.
This method is not fail proof, while it is currently sufficient
to make set_link_af() inerrable and thus 100% atomic, the
validation function method will not be able to detect all error
scenarios in the future, there will likely always be errors
depending on states which are f.e. not protected by rtnl_mutex
and thus may change between validation and setting.
Also, instead of silently ignoring unknown address families and
config blocks for address families which did not register a set
function the errors EAFNOSUPPORT respectively EOPNOSUPPORT are
returned to avoid comitting 4 out of 5 update requests without
notifying the user.
Sathya Perla [Sun, 21 Nov 2010 23:25:50 +0000 (23:25 +0000)]
be2net: adding support for Lancer family of CNAs
Key changes are:
- EQ ids are not assigned consecutively in Lancer. So, fix mapping of MSIx
vector to EQ-id.
- BAR mapping and some req locations different for Lancer.
- TCP,UDP,IP checksum fields must be compulsorily set in TX wrb for TSO in
Lancer.
- CEV_IST reg not present in Lancer; so, peek into event queue to check for
new entries
- cq_create and mcc_create cmd interface is different for Lancer; handle
accordingly
Categorize AR7010 & AR9287 devices based on driver_info
of usb_device_id, instead of PIDs. This avoids per-device cases
and minimize code changes for new device addition.
ath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)
AR9287 devices (PCI/USB) use different eeprom start location
to read nvram. New devices might endup with same devid. So use
driver_info to set offset, instead of devid. driver_info is
valid for HTC devices alone which is filled in usb_device_id.
The driver_info stores the device category information which
is used to load appropriate device firmware, select firmware offset
and eeprom starting location. The driver_info is accessed across
ath9k_htc and ath9k_hw. Hence placed under common structure.
Johannes Berg [Wed, 24 Nov 2010 07:10:06 +0000 (08:10 +0100)]
mac80211: implement packet loss notification
For drivers that have accurate TX status reporting
we can report the number of consecutive lost packets
to userspace using the new cfg80211 CQM event. The
threshold is fixed right now, this may need to be
improved in the future.
Johannes Berg [Wed, 24 Nov 2010 07:10:05 +0000 (08:10 +0100)]
cfg80211: allow using CQM event to notify packet loss
This adds the ability for drivers to use CQM events
to notify about packet loss for specific stations
(which could be the AP for the managed mode case).
Since the threshold might be determined by the
driver (it isn't passed in right now) it will be
passed out of the driver to userspace in the event.
Felix Fietkau [Tue, 23 Nov 2010 02:10:32 +0000 (03:10 +0100)]
mac80211: reduce the number of retries for nullfunc probing
Since nullfunc frames are transmitted as unicast frames, they're more
reliable than the broadcast probe requests, so we need fewer retries
to figure out whether the AP is really gone.
Felix Fietkau [Tue, 23 Nov 2010 02:10:31 +0000 (03:10 +0100)]
mac80211: use nullfunc instead of probe request for connection monitoring
nullfunc frames are better for connection monitoring, because probe requests
are answered even if the AP has already dropped the connection, whereas
nullfunc frames from an unassociated station will trigger a disassoc/deauth
frame from the AP (WLAN_REASON_CLASS3_FRAME_FROM_NONASSOC_STA), which allows
the station to reconnect immediately instead of waiting until it attempts to
transmit the next unicast frame.
This only works on hardware with reliable tx ACK reporting, any other hardware
needs to fall back to the probe request method.
- store the multicast rate as an index instead of the rate value
(reduces cpu overhead in a hotpath)
- validate the rate values (must match a bitrate in at least one sband)
Felix Fietkau [Tue, 23 Nov 2010 19:28:03 +0000 (20:28 +0100)]
mac80211: probe the AP when resuming
Check the connection by probing the AP (either using nullfunc or a
probe request). If nullfunc probing is supported and the assoc is no
longer valid, the AP will send a disassoc/deauth immediately.
Felix Fietkau [Fri, 19 Nov 2010 21:55:38 +0000 (22:55 +0100)]
mac80211: calculate beacon loss time accurately
Instead of using a fixed 2 second timeout, calculate beacon loss interval
from the advertised beacon interval and a frame count. With this beacon
loss happens after N (default 7) consecutive frames are missed which
for a typical setup (100TU beacon interval) is ~700ms (or ~1/3 previous).
This change deserves a bit more explanation. You might include something like
"These tables came from reverse engineering the 5.10.56.46 version of the
Broadcom driver. Trace comparisons between b43 and the current Broadcom driver
(5.10.120.0) show byte reversals for the PHY register writes."
Tom Herbert [Sun, 21 Nov 2010 13:17:27 +0000 (13:17 +0000)]
xps: Transmit Packet Steering
This patch implements transmit packet steering (XPS) for multiqueue
devices. XPS selects a transmit queue during packet transmission based
on configuration. This is done by mapping the CPU transmitting the
packet to a queue. This is the transmit side analogue to RPS-- where
RPS is selecting a CPU based on receive queue, XPS selects a queue
based on the CPU (previously there was an XPS patch from Eric
Dumazet, but that might more appropriately be called transmit completion
steering).
Each transmit queue can be associated with a number of CPUs which will
use the queue to send packets. This is configured as a CPU mask on a
per queue basis in:
/sys/class/net/eth<n>/queues/tx-<n>/xps_cpus
The mappings are stored per device in an inverted data structure that
maps CPUs to queues. In the netdevice structure this is an array of
num_possible_cpu structures where each structure holds and array of
queue_indexes for queues which that CPU can use.
The benefits of XPS are improved locality in the per queue data
structures. Also, transmit completions are more likely to be done
nearer to the sending thread, so this should promote locality back
to the socket on free (e.g. UDP). The benefits of XPS are dependent on
cache hierarchy, application load, and other factors. XPS would
nominally be configured so that a queue would only be shared by CPUs
which are sharing a cache, the degenerative configuration woud be that
each CPU has it's own queue.
Below are some benchmark results which show the potential benfit of
this patch. The netperf test has 500 instances of netperf TCP_RR test
with 1 byte req. and resp.
bnx2x on 16 core AMD
XPS (16 queues, 1 TX queue per CPU) 1234K at 100% CPU
No XPS (16 queues) 996K at 100% CPU
Tom Herbert [Sun, 21 Nov 2010 13:17:29 +0000 (13:17 +0000)]
xps: Improvements in TX queue selection
In dev_pick_tx, don't do work in calculating queue
index or setting
the index in the sock unless the device has more than one queue. This
allows the sock to be set only with a queue index of a multi-queue
device which is desirable if device are stacked like in a tunnel.
We also allow the mapping of a socket to queue to be changed. To
maintain in order packet transmission a flag (ooo_okay) has been
added to the sk_buff structure. If a transport layer sets this flag
on a packet, the transmit queue can be changed for the socket.
Presumably, the transport would set this if there was no possbility
of creating OOO packets (for instance, there are no packets in flight
for the socket). This patch includes the modification in TCP output
for setting this flag.
stmmac: add init/exit callback in plat_stmmacenet_data struct
This patch adds in the plat_stmmacenet_data
the init and exit callbacks that can be used
for invoking specific platform functions.
For example, on ST targets, these call the
PAD manager functions to set PIO lines and
syscfg registers.
The patch removes the stmmac_claim_resource
only used on STM Kernels as well.
bnx2x: Do interrupt mode initialization and NAPIs adding before register_netdev()
Move the interrupt mode configuration and NAPIs adding before a
register_netdev() call to prevent netdev->open() from running
before these functions are done.
Eric Dumazet [Wed, 24 Nov 2010 03:45:10 +0000 (03:45 +0000)]
bnx2x: Disable local BHes to prevent a dead-lock situation
According to Eric's suggestion:
Disable local BHes to prevent a dead-lock situation between sch_direct_xmit()
(Soft_IRQ context) and bnx2x_tx_int (called by bnx2x_run_loopback() - syscall
context), as both are taking a netif_tx_lock().
Matt Carlson [Wed, 24 Nov 2010 08:31:52 +0000 (08:31 +0000)]
tg3: use dma_alloc_coherent() instead of pci_alloc_consistent()
Using dma_alloc_coherent() permits to use GFP_KERNEL allocations instead
of GFP_ATOMIC ones. Its better when a machine is out of memory, because
this allows driver to sleep to get its memory and succeed its init,
especially when allocating high order pages.
Matt Carlson [Wed, 24 Nov 2010 08:31:48 +0000 (08:31 +0000)]
tg3: Fix 5719 internal FIFO overflow problem
Under load, there an internal FIFO can overflow on the 5719. The fix is
to scale back the PCIe maximum read request size based on the current
link speed and width.
Matt Carlson [Wed, 24 Nov 2010 08:31:47 +0000 (08:31 +0000)]
tg3: Always turn on APE features in mac_mode reg
The APE needs certain bits in the mac_mode register to be enabled for
traffic to flow correctly. This patch changes the code to always enable
these bits in the presence of the APE.
Matt Carlson [Wed, 24 Nov 2010 08:31:46 +0000 (08:31 +0000)]
tg3: Assign correct tx margin for 5719
Commit d309a46e42542223946d3a9e4e239fdc945cb53e, entitled
"tg3: 5719: Prevent tx data corruption", was supposed to contain the tx
margin adjustment but it looks like it somehow was omitted. This patch
fixes the problem.
Matt Carlson [Wed, 24 Nov 2010 08:31:45 +0000 (08:31 +0000)]
tg3: Apply 10Mbps fix to all 57765 revisions
Commit a977dbe8445b8a81d6127c4aa9112a2c29a1a008, entitled
"tg3: Reduce 57765 core clock when link at 10Mbps" needs to be applied
to all revisions of the 57765 asic rev, not just the A0 revision.
Eric Paris [Tue, 23 Nov 2010 06:28:08 +0000 (06:28 +0000)]
SELinux: indicate fatal error in compat netfilter code
The SELinux ip postroute code indicates when policy rejected a packet and
passes the error back up the stack. The compat code does not. This patch
sends the same kind of error back up the stack in the compat code.
Eric Paris [Tue, 23 Nov 2010 06:28:02 +0000 (06:28 +0000)]
SELinux: Only return netlink error when we know the return is fatal
Some of the SELinux netlink code returns a fatal error when the error might
actually be transient. This patch just silently drops packets on
potentially transient errors but continues to return a permanant error
indicator when the denial was because of policy.
Felix Fietkau [Mon, 22 Nov 2010 14:39:39 +0000 (15:39 +0100)]
ath9k_hw: fix A-MPDU key search issues on AR9003
Under load, a large number of frames can produce decryption errors, even when
no key cache update is being done. Performing a key search for every single
frame in an A-MPDU improves reliability.
projects / linux.git / log