Private rx_csum flags are now duplicate of netdev->features &
NETIF_F_RXCSUM. We remove those duplicates and now use the net_device_ops
ndo_set_features. This was based on the original patch submitted by
Michal Miroslaw <mirq-linux@rere.qmqm.pl>. I also removed the special
case not requiring a reset for X540 hardware. It is needed just as it is
in 82599 hardware.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Cc: Michal Miroslaw <mirq-linux@rere.qmqm.pl>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is meant to address possible race conditions from the status
and error bits on the RX descriptors being re-read by multiple functions in
the RX cleanup path. To resolve this I have added code that will pass the
staterr value to those functions.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change moves work_limit, total_packets, and total_bytes into the ring
container struct of the q_vector. The advantage of this is that it should
reduce the size of memory used in the event of multiple rings being
assigned to a single q_vector. In addition it should help to reduce the
total workload for calculating itr since now total_packets and total_bytes
will be the total work done of the interrupt instead of for the ring.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch adds support for a ring container structure to be used within
the q_vector. The basic idea is to provide a means of separating the RX
and TX rings while maintaining a common structure for their containment.
The advantage to this is that later we should be able to pass this
structure to the update_itr functions without needing to pass individual
rings.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch updates the current methods used for determining if we have
enough space to transmit a given skb. The current method is quite wasteful
as it has us go through and determine how each page is going to be broken
up. That only needs to be done if pages are larger than our maximum data
per TXD. As such I have wrapped that in a page size check.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
There is a significant amount of shared functionality between the checksum
and TSO offload configuration that is shared in regards to how they setup
the context descriptors. Since so much of the functionality is shared it
makes sense to move the shared functionality into a single function and
just call that function from the two context descriptor specific routines.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is a minor cleanup that converts the IXGBE_DESC_UNUSED macro
into a static inline function just for the case of the code being a bit
cleaner.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This code adds support for displaying the filters that were added via the
nfc interface. This is primarily to test the interface for now, but I am
also looking into the feasibility of moving all of the ntuple filter code
in ixgbe over to the nfc interface since it seems to be better implemented.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is meant to update the internal framework of ixgbe so that
perfect filters can be stored and tracked via software.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The number of TX and RX queues allocated depends on the device
type, the current features set, online CPUs, and various
compile flags.
To enable DCB with multiple queues and allow it to coexist with
all the features currently implemented it has to setup a valid
queue count. This is done at init time using the FDIR and RSS
max queue counts and allowing each TC to allocate a queue per
CPU.
DCB will now use available queues up to (8 x TCs) this is somewhat
arbitrary cap but allows DCB to use up to 64 queues. Its easy to
increase this later if that is needed.
This is prep work to enable Flow Director with DCB. After this
DCB can easily coexist with existing features and no longer
needs its own DCB feature ring.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change merges the over-temp task into the service task. As a result
all tasklets are finally combined into once single tasklet for easier
management.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change merges the ATR table reinitialization into the service task.
This is yet another opportunity to avoid any race conditions as we don't
want to be attempting to reinitialize the table during a possible reset.
In addition this change adds a counter for table reinitialization so that
it can be tracked as part of the regular statistics.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Evan Swanson <evan.swanson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is meant to further help to reduce possible configuration
collisions between the various tasklets. This change combines the device
reset with the service task. As a result it is now not possible to be
updating the link on the device while also resetting the part.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Evan Swanson <evan.swanson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch is meant to merge the functionality of the ixgbe watchdog task
into the service task. By doing this all link state functionality will be
controlled by a single task. As a result the reliability of the interface
will be improved as the likelihood of any race conditions is further
reduced.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is meant to address several race conditions with multi-speed
fiber SFP+ modules in 82599 adapters. Specifically issues have been seen
in which both the SFP configuration and the multi-speed fiber configuration
are running simultaneously which will result in the device getting into an
erroneous link down state.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Evan Swanson <evan.swanson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change moves flags and state into the same cacheline. The reason for
this change is because both are frequently read around the same time and
infrequently written. By combining them into the same cacheline this
should help to reduce memory utilization.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Evan Swanson <evan.swanson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add infrastructure in the PF driver to support macvlan in the VF driver.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Based on the original patch submitted by Stephen Hemminger.
This patch makes the following changes:
- Change ETHTOOL_ID_INACTIVE return value to 2 (blinks/sec)
- Fix restoring of IXGBE_LEDCTL
CC: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Evan Swanson <evan.swanson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Implemented ixgbe_ndo_set_vf_bw function which is being used by iproute2
tool. In addition, updated ixgbe_ndo_get_vf_config function to show the
actual rate limit to the user.
The rate limitation can be configured only when the link is up and the
link speed is 10Gb.
The rate limit value can be 0 or ranged between 11 and actual link
speed measured in Mbps. A value of '0' disables the rate limit for
this specific VF.
iproute2 usage will be 'ip link set ethX vf Y rate Z'.
After the command is made, the rate will be changed instantly.
To view the current rate limit, use 'ip link show ethX'.
The rates will be zeroed only upon driver reload or a link speed change.
This feature is being supported by 82599 and X540 devices.
Signed-off-by: Lior Levy <lior.levy@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This enables multiple {Tx|Rx} rings per traffic class while in DCB
mode. In order to get this working as expected the tc_to_tx net
device mapping is configured as well as the prio_tc_map.
skb priorities are mapped across a range of queue pairs to get
a distribution per traffic class. The maximum number of
queue pairs used while in DCB mode is capped at 64. The hardware
max is actually 128 queues but 64 is sufficient for now and
allocating more seemed a bit excessive. It is easy enough to
increase the cap later if need be.
To get the 802.1Q priority tags inserted correctly ixgbe was
previously using the skb queue_mapping field to directly set
the 802.1Q priority. This no longer works because we have removed
the 1:1 mapping between queues and traffic class. Each ring
is aligned with an 802.1Qaz traffic class so here we add an
extra field to the ring struct to identify the 802.1Q traffic
class. This uses an extra byte of the ixgbe_ring struct
fortunately there was a 2byte hole,
struct ixgbe_ring {
void * desc; /* 0 8 */
struct device * dev; /* 8 8 */
struct net_device * netdev; /* 16 8 */
union {
struct ixgbe_tx_buffer * tx_buffer_info; /* 8 */
struct ixgbe_rx_buffer * rx_buffer_info; /* 8 */
}; /* 24 8 */
long unsigned int state; /* 32 8 */
u8 atr_sample_rate; /* 40 1 */
u8 atr_count; /* 41 1 */
u16 count; /* 42 2 */
u16 rx_buf_len; /* 44 2 */
u16 next_to_use; /* 46 2 */
u16 next_to_clean; /* 48 2 */
u8 queue_index; /* 50 1 */
u8 reg_idx; /* 51 1 */
u16 work_limit; /* 52 2 */
/* XXX 2 bytes hole, try to pack */
u8 * tail; /* 56 8 */
/* --- cacheline 1 boundary (64 bytes) --- */
Now we can set the VLAN priority directly and it will be
correct. User space can indicate the 802.1Qaz priority
using the SO_PRIORITY setsocket() option and QOS layer will
steer the skb to the correct rings. Additionally using
the multiq qdisc with a queue_mapping action works as
well.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This implements dcbnl get and set capabilities ops. The
devices supported by ixgbe can be configured to run in
IEEE or CEE modes but not both.
With the DCBX set capabilities bit we add an explicit
signal that must be used to toggle between these modes.
This patch adds logic to fail the CEE command set_hw_all()
which programs the device with a CEE configuration if
the CEE caps bit is not set. Similarly, IEEE set
commands will fail if the IEEE caps bit is not set. We
allow most CEE config set commands to occur because they
do not touch the hardware until set_hw_all() is called.
The one exception to the above is the {set|get}app routines.
These must always be protected by caps bits to ensure
side effects do not corrupt the current configured mode.
By requiring the caps bit to be set correctly we can
maintain a consistent configuration in the hardware
for CEE or IEEE modes and prevent partial hardware
configurations that may occur if user space does
not send a complete IEEE or CEE configurations.
It is expected that user space will signal a DCBX mode
before programming device.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add support to the ndo_fcoe_ddp_target() to allow the Intel 82599 device to
also provide DDP offload capability when the upper FCoE protocol stack is
operating as a target.
Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Updating the copyrights for 2011 as well as make the ixgbe_copyright string
a constant.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We didn't need the prototype and it was causing namespace complaints so
I made it static.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Implements 802.1Qaz support for ixgbe driver. Additionally,
this adds IEEE_8021QAZ_TSA_{} defines to dcbnl.h this is to
avoid having to use cryptic numeric codes for the TSA type.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change fixes several issues found in ntuple filtering while I was
doing the ATR refactor.
Specifically I updated the masks to work correctly with the latest version
of ethtool, I cleaned up the exception handling and added detailed error
output when a filter is rejected, and corrected several bits that were set
incorrectly in ixgbe_type.h.
The previous version of this patch included a printk that was left over from
me fixing the filter setup. This patch does not include that printk.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This change adds a compressed input type for atr signature hash
computation. It also drops the use of the set functions when setting up
the ATR input since we can then directly setup the hash input as two dwords
that can be stored and passed as registers.
With these changes the cost of computing the has is low enough that we can
perform a hash computation on each TCP SYN flagged packet allowing us to
drop the number of flow director misses considerably in tests such as
netperf TCP_CRR.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This change cleans up the layout of the flow director data, and the
algorithm used to calculate the hash resulting in a 35x / 3500% performance
increase versus the old flow director hash computation. The overall effect
is only a 1% increase in transactions per second though due to the fact
that only 1 packet in 20 are actually hashed upon.
TCP_RR before:
Socket Size Request Resp. Elapsed Trans.
Send Recv Size Size Time Rate
bytes Bytes bytes bytes secs. per sec
16384 87380 1 1 60.00 23059.27
16384 87380
TCP_RR after:
Socket Size Request Resp. Elapsed Trans.
Send Recv Size Size Time Rate
bytes Bytes bytes bytes secs. per sec
16384 87380 1 1 60.00 23239.98
16384 87380
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When disable the Rx logic globally, we would also want to disable the per Rx
queue receive logic by per queue Rx control register RXDCTL so no more DMA is
happening from the packet buffer to the receive buffer associated with the Rx
ring, before we start unmapping Rx ring receive buffer. The hardware may take
max of 100us before the corresponding Rx queue is really disabled. Added
ixgbe_disable_rx_queue() for this purpose.
Signed-off-by: Yi Zou <yi.zou@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds support for the x540 MAC which is the next MAC in the
82598/82599 line.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Adds the new x540.c file and Aquantia 1202 PHY for X540 support.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The Tx hang logic has been known to detect false hangs when
the device is receiving pause frames or has delayed processing
for some other reason.
This patch makes the logic more robust and resolves these
known issues. The old logic checked to see if the device
was paused by querying the HW then the hang logic was
aborted if the device was currently paused. This check was
racy because the device could have been in the pause state
any time up to this check. The other operation of the
hang logic is to verify the Tx ring is still advancing
the old logic checked the EOP timestamp. This is not
sufficient to determine the ring is not advancing but
only infers that it may be moving slowly.
Here we add logic to track the number of completed Tx
descriptors and use the adapter stats to check if any
pause frames have been received since the previous Tx
hang check. This way we avoid racing with the HW
register and do not detect false hangs if the ring is
advancing slowly.
This patch is primarily the work of Jesse Brandeburg. I
clean it up some and fixed the PFC checking.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change cleans up some of the items in ixgbe_map_rings_to_vectors.
Specifically it merges the two for loops and drops the unnecessary vectors
parameter.
It also moves the vector names into the q_vectors themselves.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change cleans up the use of rsc_count and changes it to a boolean since
the actual numerical value is used nowhere in the Rx cleanup path. I am
also moving the skb count into the RSC_CB path since it is much easier to
track it there than when it is passed as a parameter to various function
calls.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change adds a set of state flags to the rings that allow them to
independently function allowing for features like RSC, packet split, and
TX hang detection to be done per ring instead of for the entire device.
This is accomplished by re-purposing the flow director reinit_state member
and making it a global state instead since a long for a single bit flag is
a bit wasteful.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This is the start of work to sort out what belongs in the rings and what
belongs in the q_vector. Items like the CPU variable for make much more
sense in the q_vector since the CPU is a per-interrupt thing rather than a
per ring thing.
I also added a back-pointer from the ring to the q_vector.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change places a netdev pointer directly into the ring structure. This
way we can avoid having to determine which netdev we are supposed to be
using and can just access the one on the ring directly.
As a result of this change further collapse of the code is possible by
dropping the adapter from ixgbe_alloc_rx_buffers, and the netdev pointer
from ixgbe_xmit_frame_ring_adv and ixgbe_maybe_stop_tx.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change moved some of the RX and TX stats into separate structures and
them placed those structures in a union in order to help reduce the size of
the ring structure.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is meant to simplify DMA map/unmap by providing a device
pointer. As a result the adapter pointer can be dropped from many of
the calls.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change drops ring->head since it is not used in any hot-path and can
easily be determined using IXGBE_[RT]DH(ring->reg_idx).
It also changes ring->tail into a true pointer so we can avoid unnecessary
pointer math to find the location of the tail.
In addition I also dropped the setting of head and tail in
ixgbe_clean_[rx|tx]_ring. The only location that should be setting the head
and tail values is ixgbe_configure_[rx|tx]_ring and that is only while the
queue is disabled.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change re-orders alloc_rx_buffers to make better use of the packet
split enabled flag. The new setup should require less branching in the
code since now we are down to fewer if statements since we either are
handling packet split or aren't.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change simplifies the work being done by the TX interrupt handler and
pushes it into the tx_map call. This allows for fewer cache misses since
the TX cleanup now accesses almost none of the skb members.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Currently the high and low water marks for PFC are being set
conservatively for jumbo frames. This means the RX buffers
are being underutilized in the default 1500 MTU. This patch
fixes this so that the water marks are set as described in
the data sheet considering the MTU size.
The equation used is,
RTT * 1.44 + MTU * 1.44 + MTU
Where RTT is the round trip time and MTU is the max frame size
in KB. To avoid floating point arithmetic FC_HIGH_WATER is
defined
((((RTT + MTU) * 144) + 99) / 100) + MTU
This changes how the hardware field fc.low_water and
fc.high_water are used. With this change they are no longer
storing the actual low water and high water markers but are
storing the required head room in the buffer. This simplifies
the logic and we do not need to account for the size of the
buffer when setting the thresholds.
Testing with iperf and 16 threads showed a slight uptick in
throughput over a single traffic class .1-.2Gbps and a reduction
in pause frames. Without the patch a 30 second run would show
~10-15 pause frames being transmitted with the patch ~2-5 are
seen. Test were run back to back with 82599.
Note RXPBSIZE is in KB and low and high water marks fields are
also in KB. However the FCRT* registers are 32B granularity and
right shifted 5 into the register,
(((rx_pbsize - water_mark) * 1024) / 32) << 5
is the most explicit conversion here we simplify
(rx_pbsize - water_mark) * 32 << 5 = (rx_pbsize - water_mark) << 10
This patch updates the PFC thresholds and legacy FC thresholds.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
"cat /proc/net/dev" uses RCU protection only.
Its quite possible we call a driver get_stats() method while device is
dismantling and freeing its data structures.
So get_stats() methods must be very careful not accessing driver private
data without appropriate locking.
In ixgbe case, we access rx_ring pointers. These pointers are freed in
ixgbe_clear_interrupt_scheme() and set to NULL, this can trigger NULL
dereference in ixgbe_get_stats64()
A possible fix is to use RCU locking in ixgbe_get_stats64() and defer
rx_ring freeing after a grace period in ixgbe_clear_interrupt_scheme()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Reported-by: Tantilov, Emil S <emil.s.tantilov@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Current ixgbe stats have following problems :
- Not 64 bit safe (on 32bit arches)
- Not safe in ixgbe_clean_rx_irq() :
All cpus dirty a common location (netdev->stats.rx_bytes &
netdev->stats.rx_packets) without proper synchronization.
This slow down a bit multiqueue operations, and possibly miss some
updates.
Fixes :
Implement ndo_get_stats64() method to provide accurate 64bit rx|tx
bytes/packets counters, using 64bit safe infrastructure.
ixgbe_get_ethtool_stats() also use this infrastructure to provide 64bit
safe counters.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make the ixgbe driver use the new vlan accleration model.
Signed-off-by: Jesse Gross <jesse@nicira.com>
CC: Peter Waskiewicz <peter.p.waskiewicz.jr@intel.com>
CC: Emil Tantilov <emil.s.tantilov@intel.com>
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the new infrastructure to balance interrupts for flow
alignment when ATR or Flow Director are enabled.
Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This change makes it so that the ethtool loopback test uses the standard
ring configuration and allocation functions. As a result the loopback test
will be much more effective at testing core driver functionality.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All of the DESC_ADV macros are currently needing the pointers to be
de-referenced before accessing the ring. Instead of having to add all of
the asterisks it is easier to just update the macro to expect a pointer to
the ring.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Bump the header size for packet split to 512 bytes since this makes the
best use of the 1k buffer that is allocated for any skb 512 bytes or
smaller.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
