Simple cast to fix a sparse warning.
Fixes: commit 5453205cd0 ("i40e/i40evf: Enable support for
SKB_GSO_UDP_TUNNEL_CSUM")
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch enables bulk Tx clean for skbs. In order to enable it we need
to pass the napi_budget value as that is used to determine if we are truly
running in NAPI mode or if we are simply calling the routine from netpoll
with a budget of 0. In order to avoid adding too many more variables I
thought it best to pass the VSI directly in a fashion similar to what we do
on igb and ixgbe with the q_vector.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
In the polling routines for i40e and i40evf we were using bitwise operators
to avoid the side effects of the logical operators, specifically the fact
that if the first case is true with "||" we skip the second case, or if it
is false with "&&" we skip the second case. This fixes an earlier patch
that converted the bitwise operators over to the logical operators and
instead replaces the entire thing with just an if statement since it should
be more readable what we are trying to do this way.
Fixes: 1a36d7fadd ("i40e/i40evf: use logical operators, not bitwise")
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
From what I can tell the practical limitation on the size of the Tx data
buffer is the fact that the Tx descriptor is limited to 14 bits. As such
we cannot use 16K as is typically used on the other Intel drivers. However
artificially limiting ourselves to 8K can be expensive as this means that
we will consume up to 10 descriptors (1 context, 1 for header, and 9 for
payload, non-8K aligned) in a single send.
I propose that we can reduce this by increasing the maximum data for a 4K
aligned block to 12K. We can reduce the descriptors used for a 32K aligned
block by 1 by increasing the size like this. In addition we still have the
4K - 1 of space that is still unused. We can use this as a bit of extra
padding when dealing with data that is not aligned to 4K.
By aligning the descriptors after the first to 4K we can improve the
efficiency of PCIe accesses as we can avoid using byte enables and can fetch
full TLP transactions after the first fetch of the buffer. This helps to
improve PCIe efficiency. Below is the results of testing before and after
with this patch:
Recv Send Send Utilization Service Demand
Socket Socket Message Elapsed Send Recv Send Recv
Size Size Size Time Throughput local remote local remote
bytes bytes bytes secs. 10^6bits/s % S % U us/KB us/KB
Before:
87380 16384 16384 10.00 33682.24 20.27 -1.00 0.592 -1.00
After:
87380 16384 16384 10.00 34204.08 20.54 -1.00 0.590 -1.00
So the net result of this patch is that we have a small gain in throughput
due to a reduction in overhead for putting together the frame.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
On all of the other Intel drivers we place checksum close to TSO as they
have a significant amount in common and it can help to reduce the decision
tree for how to handle the frame as the first check in TSO is to see if
checksumming is offloaded, and if it is not we can skip _BOTH_ TSO and Tx
checksum offload based on a single check.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch is meant to rewrite the logic for how we determine if we can
transmit the frame or if it needs to be linearized.
The previous code for this function was using a mix of division and modulus
division as a part of computing if we need to take the slow path. Instead
I have replaced this by simply working with a sliding window which will
tell us if the frame would be capable of causing a single packet to span
several descriptors.
The logic for the scan is fairly simple. If any given group of 6 fragments
is less than gso_size - 1 then it is possible for us to have one byte
coming out of the first fragment, 6 fragments, and one or more bytes coming
out of the last fragment. This gives us a total of 8 fragments
which exceeds what we can allow so we send such frames to be linearized.
Arguably the use of modulus might be more exact as the approach I propose
may generate some false positives. However the likelihood of us taking much
of a hit for those false positives is fairly low, and I would rather not
add more overhead in the case where we are receiving a frame composed of 4K
pages.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
In an upcoming patch I would like to have access to the descriptor count
used for the data portion of the frame. For this reason I am splitting up
the descriptor count function from the function that stops the ring.
Also in order to try and reduce unnecessary duplication of code I am moving
the slow-path portions of the code out of being inline calls so that we can
just jump to them and process them instead of having to build them into
each function that calls them.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The XL722 has support for providing the outer UDP tunnel checksum on
transmits. Make use of this feature to support segmenting UDP tunnels with
outer checksums enabled.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This is mostly a minor clean-up for the Rx checksum path in order to avoid
some of the unnecessary conditional checks that were being applied.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add exception handling to the Tx checksum path so that we can handle cases
of TSO where the frame is bad, or Tx checksum where we didn't recognize a
protocol
Drop I40E_TX_FLAGS_CSUM as it is unused, move the CHECKSUM_PARTIAL check
into the function itself so that we can decrease indent.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch defers writing to the Tx descriptor bits until we know we have
successfully completed a given operation. So for example we defer updating
the tunnelling portion of the context descriptor until we have fully
identified the type.
The advantage to this approach is that we can assemble values as we go
instead of having to try and kludge everything together all at once. As a
result we can significantly clean up the tunneling configuration for
instance as we can just do a pointer walk and do the math for the distance
between each set of points.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch adds support for IPv6 extension headers in setting up the Tx
checksum. Without this patch extension headers would cause IPv6 traffic to
fail as the transport protocol could not be identified.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch fixes two issues. First was the fact that iphdr(skb)->protocl
was being used to test for the outer transport protocol. This completely
breaks IPv6 support. Second was the fact that we cleared the flag for v4
going to v6, but we didn't take care of txflags going the other way. As
such we would have the v6 flag still set even if the inner header was v4.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The Tx checksum path was maintaining a set of 3 pointers and two lengths in
order to prepare the packet for being checksummed. The thing is we only
really needed 2 pointers, and the lengths that were being maintained can
easily be computed.
As such we can replace the IPv4 and IPv6 header pointers with one single
union that represents both, or a generic pointer to the start of the
network header. For the L4 headers we can do the same with TCP and a
generic pointer to the start of the transport header. The length of the
TCP header is obtained by simply multiplying doff by 4, and the network
header length can be obtained by subtracting the network header pointer
from the transport header pointer.
While I was at it I renamed l4_hdr to l4_proto to make it a bit more clear
and less likely to be confused with l4.hdr which is the transport header
pointer.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch goes through and pulls all of the spots where we were updating
either the TCP or IP checksums in the TSO and checksum path into the TSO
function. The general idea here is that we should only be updating the
header after we verify we have completed a skb_cow_head check to verify the
head is writable.
One other advantage to doing this is that it makes things much more
obvious. For example, in the case of IPv6 there was one spot where the
offset of the IPv4 header checksum was being updated which is obviously
incorrect.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch makes it so that the L4 header offsets and such can be ignored
when dealing with the L3 checksum and length update. This is done making
use of two things.
First we can just use the offset from the L4 header to the start of the
packet to determine the L4 offset, and from that we can then make use of
the data offset to determine the full length of the headers.
As far as adjusting the checksum to remove the length we can simply add the
inverse of the length instead of having to recompute the entire
pseudo-header without the length. In the case of an IPv6 header this
should be significantly cheaper since we can make use of a value we already
needed instead of having to read the source and destination address out of
the packet.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Instead of casing u32 values to u64 it makes more sense to just start out
with u64 values in the first place. This way we don't need to create a
mess with all of the casts needed to populate a 64b value.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The i40e and i40evf drivers contained code for inserting an outer checksum
on UDP tunnels. The issue however is that the upper levels of the stack
never requested such an offload and it results in possible errors.
In addition the same logic was being applied to the Rx side where it was
attempting to validate the outer checksum, but the logic there was
incorrect in that it was testing for the resultant sum to be equal to the
header checksum instead of being equal to 0.
Since this code is so massively flawed, and doing things that we didn't ask
for it to do I am just dropping it, and will bring it back later to use as
an offload for SKB_GSO_UDP_TUNNEL_CSUM which can make use of such a
feature.
As far as the Rx feature I am dropping it completely since it would need to
be massively expanded and applied to IPv4 and IPv6 checksums for all parts,
not just the one that supports Tx checksum offload for the outer.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
In the case where we have a page fully used by receive data, we need to
release the page fully to the stack. Instead of calling get_page (which
increments the page count) followed by free_page (which decrements the
page count), just donate our reference to the stack. Although this
donation is not tax deductible, it does allow us to avoid two very
expensive atomic operations that reverse each other.
Change-ID: If70739792d5748995fc175ec92ac2171ed4ad8fc
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch adds a workaround for cases where we might have
interrupts that got lost but WB happened.
If that happens without this patch we will see a tx_timeout.
To work around it, this patch goes ahead and reschedules NAPI
in that situation, if NAPI is not already scheduled.
We also add a counter in ethtool to keep track of when
we detect a case of tx_lost_interrupt.
Note: napi_reschedule() can be safely called from process/service_task
context and is done in other drivers as well without an issue.
Change-ID: I00f98f1ce3774524d9421227652bef20fcbd0d20
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Mr. Spock would certainly raise an eyebrow to see us using bitwise
operators, when we should clearly be relying on logic. Fascinating.
Change-ID: Ie338010c016f93e9faa2002c07c90b15134b7477
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Refactor the packet split Rx code to properly use half-pages for
receives. The previous code was doing way more mapping and unmapping
than it needed to, and wasn't properly using half-pages.
Increment the page use count each time we give a half-page to an skb,
knowing that the stack will probably process and release the page before
we need it again. Only free and reallocate pages if the count shows that
both half-pages are in use. Add counters to track reallocations and page
reuse.
Change-ID: I534b299196036b64be82b4861a0a4036310a8f22
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The i40e and i40evf drivers now cleanly handle allocation
failures and can avoid kernel log spew from the memory allocator
when allocations fail, so set __GFP_NOWARN on Rx buffer alloc.
Change-ID: Ic9e1b83c495e2a3ef6b069ba7fb6e52ce134cd23
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This is the "Don't Give Up" patch. Previously the
driver could fail an allocation, and then possibly stall
a queue forever, by never coming back to continue receiving
or allocating buffers.
With this patch, the driver will keep polling trying to allocate
receive buffers until it succeeds. This should keep all receive
queues running even in the face of memory pressure.
Also update copyright year in file header.
Change-ID: I2b103d1ce95b9831288a7222c3343ffa1988b81b
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
While re-enabling interrupts the driver would clear all pending
causes. This meant that if an interrupt was generated while the driver
was cleaning or polling with interrupts disabled, then that interrupt
was lost. This could cause a queue to become dead, especially for
receive. Refactored the enable_icr0 function in order to allow
it to be decided by the caller whether the CLEARPBA (clear pending
events) bit will be set while re-enabling the interrupt.
Also update copyright year in file headers.
Change-ID: Ic1db100a05e13c98919057696db147a258ca365a
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Now that the Force-WriteBack functionality in X710/XL710 devices
has been moved out of the clean routine and into the service task,
we need to make sure WriteBack-On-ITR is separated out since it
is still called from clean.
In the X722 devices, Force-WriteBack implies WriteBack-On-ITR but
without the interrupt, which put the driver into a missed
interrupt scenario and a potential tx-timeout report.
With this patch, we break the two functions out, and call the
appropriate ones at the right place. This will avoid creating missed
interrupt like scenarios for X722 devices.
Also update copyright year in file headers.
Change-ID: Iacbde39f95f332f82be8736864675052c3583a40
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don't bother trying to set up a TSO if the skb->ip_summed is not
set to CHECKSUM_PARTIAL.
Change-ID: I6495b3568e404907a2965b48cf3e2effa7c9ab55
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Driver was using an offset based off a DMA handle while mapping and
unmapping using sync_single_range_for[cpu|device], where it should
be using DMA handle (returned from alloc_coherent) and the offset of the
memory to be sync'd.
Change-ID: I208256565b1595ff0e9171ab852de06b997917c6
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: Nelson, Shannon <shannon.nelson@intel.com>
Reviewed-by: Williams, Mitch A <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We were not doing write-back on interrupt throttle for Legacy case in X722.
This patch fixes that, so we do WB_ON_ITR for Legacy as well. Plus the issue
that we should still be setting NO_ITR if we are touching the DYN_CTLN register
since we do not want to change ITR setting here.
Change-ID: I5db8491ee1544118a389db839cecc93e1bbc480e
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add write-back on interrupt throttle rate timer expiration support
for the i40evf driver, when running on X722 devices.
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
If the driver calls skb_set_hash even with a zero hash, that
indicates to the stack that the hash calculation is offloaded
in hardware. So the Stack doesn't do a SW hash which is required
for load balancing if the user decides to turn of rx-hashing
on our device.
This patch fixes the path so that we do not call skb_set_hash
if the feature is disabled.
Change-ID: Ic4debfa4ff91b5a72e447348a75768ed7a2d3e1b
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
1) remove duplicate include of tcp.h
2) put an ampersand at the end of a line instead of the beginning
3) remove a useless dev_info
4) match declaration of function to the implementation
5) repair incorrect comment
6) correct whitespace
7) remove unused define
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We shouldn't be using a bitwise operator here; it's not a bitwise
operation. Use a logical operator instead. Why doesn't c have a
logical-or-and-assign operator?
Change-ID: Id84f3ca884910bed7073c84b1e16a102e958d0de
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch fixes the memory leak which would be seen otherwise when user
programs flow-director filter using ethtool (sideband filter programming).
When ethtool is used to program flow directory filter, 'raw_buf' gets
allocated and it is supposed to be freed as part of queue cleanup. But
check of 'tx_buffer->skb' was preventing it from being freed.
Change-ID: Ief4f0a1a32a653180498bf6e987c1b4342ab8923
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch contains following changes:
- detection and recovery logic (issue SW interrupt) has been moved to
service_task from timeout function.
- added some more debug info from tx_timeout.
Logic to detect and recover TX queue hung is now two step process:
- service_task detects TX queue hung and sets a bit(hung_detected) if
it was not set.
- if bit was set (means this is back-back hung condition detected),
issue SW interrupt and clear the bit.
- napi_poll clears the bit unconditionally since it cleans TX/RX queues.
Change-ID: Ieed03a48927c845a988b3ff375090bf37caeb903
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Instead of awkwardly keeping a fixed array of pointers in the adapter
struct and then allocating ring structs individually, just keep a single
pointer and allocate a single blob for the arrays. This simplifies code,
shrinks the adapter structure, and future-proofs the driver by not
limiting the number of rings we can handle.
Change-ID: I31334ff911a6474954232cfe4bc98ccca3c769ff
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Issue a prefetch for data early in the transmit path.
This should not be generally needed for Tx traffic, but
it helps immensely for pktgen workloads and should help
for forwarding workloads as well.
Change-ID: Iefee870c20599e0c4240e1d8637e4f16b625f83a
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch fixes the issue of forcing WB too often causing us to not
benefit from NAPI.
Without this patch we were forcing WB/arming interrupt too often taking
away the benefits of NAPI and causing a performance impact.
With this patch we disable force WB in the clean routine for X710
and XL710 adapters. X722 adapters do not enable interrupt to force
a WB and benefit from WB_ON_ITR and hence force WB is left enabled
for those adapters.
For XL710 and X710 adapters if we have less than 4 packets pending
a software Interrupt triggered from service task will force a WB.
This patch also changes the conditions for setting RS bit as described
in code comments. This optimizes when the HW does a tail bump amd when
it does a WB. It also optimizes when we do a wmb.
Change-ID: Id831e1ae7d3e2ec3f52cd0917b41ce1d22d75d9d
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
When in NAPI with interrupts disabled, the HW needs to be forced to do a
write back on TX if the number of descriptors pending are less than a
cache line.
This stat helps keep track of how many times we get into this situation.
Change-ID: I76c1bcc7ebccd6bffcc5aa33bfe05f2fa1c9a984
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Code was moved into a separate function some time ago.
Change-ID: Icabbe71ce05cf5d716d3e1152cdd9cd41d11bcb5
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We would like to automatically provide busy polling support
to all NAPI drivers, without them having to implement anything.
skb_mark_napi_id() can be called from napi_gro_receive() and
napi_get_frags().
Few drivers are still calling skb_mark_napi_id() because
they use netif_receive_skb(). They should eventually call
napi_gro_receive() instead. I will leave this to drivers
maintainers.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The adaptive ITR (interrupt throttle rate) algorithm was adjusting
the hardware's interrupt rate too frequently. This caused a lot
of variation in the interrupt rate for fairly constant workloads.
Change the code to have a counter and adjust only once every N
number of interrupts.
Change-ID: I0460f1f86571037484eca5aca36ac4d889cb8389
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The dynamic algorithm, while now working, doesn't have good
performance in 40G mode.
One part of this patch addresses the high CPU utilization of some small
streaming workloads that the driver should reduce CPU in.
It also changes the minimum ITR that the dynamic algorithm
will settle on, causing our minimum latency to go from 12us
to about 14us, when using adaptive mode.
It also changes the BULK interrupt rate to allow maximum throughput
on a 40Gb connection with a single thread of transmit, clamping
interrupt rate to 8000 for TX makes single thread traffic go too
slow.
The new ULTRA bulk setting is introduced and is used
when the Rx packet rate on this queue exceeds 40000 packets per
second. This value of 40000 was chosen because the automatic tuning
of minimum ITR=20us means that a single queue can't quite achieve
that many packets per second from a round-robin test.
Change-ID: Icce8faa128688ca5fd2c4229bdd9726877a92ea2
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The driver was using a value expressed in 2us increments
for the divisor to figure out our bytes/usec values.
Fix the usecs variable to contain a value in microseconds.
Change-ID: I5c20493103c295d6f201947bb908add7040b7c41
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change moves a multi-line register setting into a function
which simplifies reading the flow of the enable function.
This also fixes a bug where the enable function was enabling
the interrupt twice while trying to update the two interrupt
throttle rate thresholds for Rx and Tx.
Change-ID: Ie308f9d0d48540204590cb9d7a5a7b1196f959bb
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
As per Eric Dumazet's previous patches:
(see commit (24d2e4a507) - tg3: use napi_complete_done())
Quoting verbatim:
Using napi_complete_done() instead of napi_complete() allows
us to use /sys/class/net/ethX/gro_flush_timeout
GRO layer can aggregate more packets if the flush is delayed a bit,
without having to set too big coalescing parameters that impact
latencies.
</end quote>
Tested
configuration: low latency via ethtool -C ethx adaptive-rx off
rx-usecs 10 adaptive-tx off tx-usecs 15
workload: streaming rx using netperf TCP_MAERTS
igb:
MIGRATED TCP MAERTS TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.0.0.1 () port 0 AF_INET : demo
...
Interim result: 941.48 10^6bits/s over 1.000 seconds ending at 1440193171.589
Alignment Offset Bytes Bytes Recvs Bytes Sends
Local Remote Local Remote Xfered Per Per
Recv Send Recv Send Recv (avg) Send (avg)
8 8 0 0 1176930056 1475.36 797726 16384.00 71905
MIGRATED TCP MAERTS TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.0.0.1 () port 0 AF_INET : demo
...
Interim result: 941.49 10^6bits/s over 0.997 seconds ending at 1440193142.763
Alignment Offset Bytes Bytes Recvs Bytes Sends
Local Remote Local Remote Xfered Per Per
Recv Send Recv Send Recv (avg) Send (avg)
8 8 0 0 1175182320 50476.00 23282 16384.00 71816
i40e:
Hard to test because the traffic is incoming so fast (24Gb/s) that GRO
always receives 87kB, even at the highest interrupt rate.
Other drivers were only compile tested.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The code in i40e and i40evf is using an "IN_NETPOLL" flag that has never
added any value due to the fact that the Rx clean-up is handled in NAPI.
As such the flag was set, the queue was scheduled via NAPI, and then polled
from the netpoll controller and if any Rx packets were processed the were
processed in the wrong context.
In addition the flag itself just added an unneeded conditional to the
hot-path so it can safely be dropped and save us a few instructions.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The polling routine for i40e was rounding up the budget for Rx cleanup to
1. This is incorrect as the netpoll poll call is expecting no Rx to be
processed as the budget passed was 0.
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add missings spaces after declarations, remove another __func__ use,
remove uncessary braces, remove unneeded breaks, and useless returns,
and generally fix up some code.
Change-ID: Ie715d6b64976c50e1c21531685fe0a2bd38c4244
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Keep track of how many times we ask the stack to linearize the
skb because the HW cannot handle skbs with more than 8 frags per
segment/single packet.
Change-ID: If455452060963a769bbe6112cba952e79e944b52
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Function i40e_clean_rx_irq() tries to reuse memory pages allocated
from the nearest node. To better support memoryless node, use
numa_mem_id() instead of numa_node_id() to get the nearest node with
memory.
This change should only affect performance.
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The arm writeback (arm_wb) code is used for kicking the Tx ring to
make sure any pending work is completed even if interrupts are
disabled. It was running when it didn't need to, and not clearing
the ring->arm_wb state after it was set. This caused Tx hangs
to still occur occasionally when there really was no hang.
Fix this by resetting the variable right after it was used.
Change-ID: I7bf75d552ba9c4bd203d40615213861a24bb5594
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch modifies the driver timeout logic by issuing a writeback
request via a software interrupt to the hardware the first time the
driver detects a hang. The driver was too aggressive in resetting a hung
queue, so back that off by removing logic to down the netdevice after
too many hangs, and move the function to the service task.
Change-ID: Ife100b9d124cd08cbdb81ab659008c1b9abbedea
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Use CTLN1 instead of CTLN for the VF relative register space.
Change-ID: Iefba63faf0307af55fec8dbb64f26059f7d91318
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
X722 supports offloading of outer UDP TX and RX checksum for tunneled
packets. This patch exposes the support and leaves it enabled by
default.
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
X722 fixes an issue from X710 where TX descriptor WB would not happen if
the interrupts were disabled. In order for the write backs to happen a
bit needs to be set in the dynamic interrupt control register called
WB_ON_ITR. With this feature, the SW driver need not arm SW interrupts to
work around the issue in X710.
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Use macros for abstracting (1 << foo) to BIT(foo)
and (1ULL << foo64) to BIT_ULL(foo64) in order to match
better with kernel requirements.
NOTE: the adminq_cmd.h file was not modified on purpose because
of the dependency upon firmware for that file.
Change-ID: I73ee2e48c880d671948aad19bd53ca6b2ac558fc
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch changes the switch statement for dynamic interrupt throttling
and adds a default case. With this patch, we check the latency setting
instead of the current ITR settings and the included refactor improves
performance.
Without this patch, the ITR setting would never change dynamically, and
there was no default.
Change-ID: Idb5a8a14c7109ec47c90f6e94bd43baa17d7ee37
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add a prefetch for the next Tx descriptor to be used when we know
there are more coming.
Change-ID: Ibb9acab11d508eec2db7da795df74debc16eeacb
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Down was requesting queue disables, but then exited immediately
without waiting for the queues to actually disable. This could
allow any function called after i40evf_down to run immediately,
including i40evf_up, and causes a memory leak.
Removing the whole reinit_locked function is the best way
to go about this, and allows for the driver to handle the
state changes by requesting reset from the periodic timer.
Also, add a couple WARN_ONs in slow path to help us recognize
if we re-introduce this issue or missed any cases.
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The driver doesn't use the time_stamp member to determine if there is a
tx_hang any more. There really isn't any point to the variable at all
so just remove it. It was left over from a previous tx_hang design.
Change-ID: I4c814827e1bcb46e45118fe37acdcfa814fb62a0
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Inlining these functions gives us about 15% more 64 byte packets per
second when using pktgen. 13.3 million to 15 million with a single
queue.
Also fix the function names in i40evf to i40evf not i40e while we are
touching the function header.
Change-ID: I3294ae9b085cf438672b6db5f9af122490ead9d0
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Eric added support for skb->xmit_more in i40e, this ports that into
i40evf as well.
Support skb->xmit_more in i40evf is straightforward; we need to move
around i40e_maybe_stop_tx() call to correctly test netif_xmit_stopped()
before taking the decision to not kick the NIC.
Change-ID: Idddda6a2e4a7ab335631c91ced51f55b25eb8468
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
There's no need for a counter so remove the TODO comment.
Change-ID: I3321dda04934c4f5fda9b279ab666192bda44214
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Without this, RSS would have done inner header load balancing. Now we can
get the benefits of ATR for tunneled packets to better align TX and RX
queues with the right core/interrupt.
Change-ID: I07d0e0a192faf28fdd33b2f04c32b2a82ff97ddd
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
None of those drivers uses last_rx for its own needs.
See 4dc89133f4 ("net: add a comment on
netdev->last_rx") for reference.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Zhangfei Gao <zhangfei.gao@linaro.org>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Wingman Kwok <w-kwok2@ti.com>
Cc: Murali Karicheri <m-karicheri2@ti.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Update i40e and i40evf to use dma_rmb. This should improve performance by
decreasing the barrier overhead on strong ordered architectures.
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Acked-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If transmit VLAN HW offloads are disabled then the network stack sends up
an skb with the protocol set to 8021q. In that case to get the correct
checksum offloads we have to reset the skb protocol to the encapsulated
ethertype.
Change-ID: I903d78533de09b1c5d3ec695ee1990dd0fa5dd0d
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
If the skb allocation fails we should not continue using the skb
pointer. Breaking out at the point of failure means that at the next
RX interrupt the driver will try the allocation again.
Change-ID: Iefaad69856ced7418bfd92afe55322676341f82e
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Inner protocol being UDP should not stop us from verifying Outer UDP
checksum correctness.
If the Outer protocol is not UDP (NVGRE) we should not be doing a UDP
checksum check. If the packet has zero checksum, skip checksum check.
Change-ID: Ie7f153feb276a59f66a54a0938901b2c0a8100fa
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Fix a bug introduced in the force writeback code, where the interrupt
rate was set to 0 (maximum) by accident.
The driver must correctly set the NOITR fields to avoid ITR update
as a side effect of triggering the software interrupt.
Change-ID: I290851ae04ef3811c43aab5ee33242029f26c1a3
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Performance can be improved a bit by imitating ixgbe and using
prefetch to get us the next Tx descriptor.
Change-ID: Ice7ffd4cd0ce87c35295059bdb7972a7f53723aa
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Use l4_tunnel type generically to keep code flow simple.
Change-ID: Ic52287e3b1ca4204e6b6e13431890c1a6ae9c422
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
A sparse complaint in i40e_debug_aq in a funky buffer write goes away by
straightening out the code out to something less convoluted.
Also fix some other sparse warnings while we are at it, making some
functions static and using NULL instead of 0.
Change-ID: I93907534fe1f1f675830774b3d14ecf1c6ffc9a0
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Conflicts:
drivers/net/ethernet/rocker/rocker.c
The rocker commit was two overlapping changes, one to rename
the ->vport member to ->pport, and another making the bitmask
expression use '1ULL' instead of plain '1'.
Signed-off-by: David S. Miller <davem@davemloft.net>
We were missing a few packet types for VXLAN offload. This patch fixes
that.
Change-ID: I4b23aa0b08e40ed49d0df6c49a5ed9f2009b44ce
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We were checking the outer Protocol flags and deciding the flow for
inner header. This patch fixes that.
This fixes the Tx checksum offload for TCP/IPv6 over vxlan.
Change-ID: I837aaea921d34f71b24c2bc32aaadea5001ddf78
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The driver was having some issues with false Tx hang detection. This
makes the driver a little more direct with the checks for progress
forward by directly checking the head write back address and tail register
when determining progress. This avoids Tx hangs where the software
gets behind, because we are directly checking hardware state when
determining hang state.
Change-ID: I774f0e861c9e8ab5ccb213634100fe15440ae24a
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The hardware has some limitations the driver needs to adhere to,
that we found in extended testing.
1) no more than 8 descriptors per packet on the wire
2) no header can span more than 3 descriptors
If one of these events occurs, the hardware will generate an internal
error and freeze the Tx queue.
This patch linearizes the skb to avoid these situations.
Change-ID: I37dab7d3966e14895a9663ec4d0aaa8eb0d9e115
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Split the receive hot path code into two, one for packet split and one
for single buffer. This improves receive performance since we only need
to check if the ring is in packet split mode once per NAPI poll time,
not several times per packet. The single buffer code is further improved
by the removal of a bunch of code and several variables that are not
needed. On a receive-oriented test this can improve single-threaded
throughput.
Also refactor the packet split receive path to use a fixed buffer for
headers, like ixgbe does. This vastly reduces the number of DMA mappings
and unmappings we need to do, allowing for much better performance in
the presence of an IOMMU.
Lastly, correct packet split descriptor types now that we are actually
using them.
Change-ID: I3a194a93af3d2c31e77ff17644ac7376da6f3e4b
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch fixes indentation issue and error found in argument
reported by static analysis. Without this patch, sparse and other
static analysis errors will be found.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds a call to u64_stats_init to Rx setup.
This done in order to avoid lockdep errors with seqcount on newer kernels.
Change-ID: Ia8ba8f0bcbd1c0e926f97d70aeee4ce4fd055e93
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch forces Tx descriptor writebacks on ITR by kicking
off the SWINT interrupt when we notice that there are non-cache-aligned
Tx descriptors waiting in the ring while interrupts are disabled
under NAPI.
Change-ID: dd6d9675629bf266c7515ad7a201394618c35444
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The same macros are used for rx as well. So rename it.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
Set skb->csum_level instead of skb->encapsulation when indicating
CHECKSUM_UNNECESSARY for an encapsulated checksum.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We are seeing situations where the driver sees a hang with less than 4
desc pending, if the driver chooses to ignore it the queue progresses
forward and the stack never experiences a real hang.
With this patch we will log a stat when this situation happens
"tx_sluggish" will increment and we can see some more details
at a higher debug level. Other than that we will ignore this
particular case of Tx hang.
Change-ID: I7d1d1666d990e2b12f4f6bed0d17d22e1b6410d5
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This device claims TSO and checksum support for vlans. It also
allows a user to control vlan acceleration offloading. As such,
it is possible to turn off vlan acceleration and configure a vlan
which will continue to support TSO and hw checksums.
In such situation the packet passed down the the device will contain
a vlan header and skb->protocol will be set to ETH_P_8021Q.
The device assumes that skb->protocol contains network protocol
value and uses that value to set up TSO and checksum information.
This results in corrupted frames sent on the wire.
This patch extract the protocol value correctly and corrects TSO
and checksums for non-accelerated traffic.
Fix this by using vlan_get_protocol() helper.
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Bruce Allan <bruce.w.allan@intel.com>
CC: Carolyn Wyborny <carolyn.wyborny@intel.com>
CC: Don Skidmore <donald.c.skidmore@intel.com>
CC: Greg Rose <gregory.v.rose@intel.com>
CC: Alex Duyck <alexander.h.duyck@intel.com>
CC: John Ronciak <john.ronciak@intel.com>
CC: Mitch Williams <mitch.a.williams@intel.com>
CC: Linux NICS <linux.nics@intel.com>
CC: e1000-devel@lists.sourceforge.net
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The hardware design requires that the driver avoid indicating
checksum offload success on some ipv6 frames with extension
headers.
The code needs to just check for the IPV6EXADD bit and if
it is set punt the checksum to the stack. I don't know why
the code was checking TCP on inner protocol, as that code
doesn't make any sense to me but seems wrong, so remove it.
Change-ID: I10d3aacdbb1819fb60b4b0eb80e6cc67ef2c9599
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-By: Jim Young <jamesx.m.young@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The HW still needs to consume it and freeing it in the function
that created it would mean we will be racing with the HW. The
i40e_clean_tx_ring() routine will free up the buffer attached once
the HW has consumed it. The clean_fdir_tx_irq function had to be fixed
to handle the freeing correctly.
Cases where we program more than one filter per flow (Ipv4), the
code had to be changed to allocate dummy buffer multiple times
since it will be freed by the clean routine. This also fixes an issue
where the filter program routine was not checking if there were
descriptors available for programming a filter.
Change-ID: Idf72028fd873221934e319d021ef65a1e51acaf7
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <jamesx.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Driver needs to initialize all members of context descriptor. Stale
data is possible otherwise.
Change-ID: Idc6b53af45583509da42d5ec0824cbaf78aee64f
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The driver was not marking packets with bad checksums
correctly, especially IPv6 packets with a bad checksum.
To do this correctly we need a define that may be set by
hardware in rare cases.
Change-ID: I1a997b72b491ded27a78ac3bce1197b2d2611130
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The TSO logic in the transmit path had some assumptions that
have been broken now that the kernel can send as much as 32kB
in a single skb->frag[.] entry, even on a system with 4kB pages.
This fixes the assumptions and allows the kernel to operate
as efficiently as possible with both SENDFILE and SEND.
In addition, the hardware limit of data contained in a descriptor is
changed to the next power of two below where it currently is in
order to align to a power of two value, preventing a single byte
of data in a descriptor.
Change-ID: I6af1f0b87c1458e10644dbd47541591075a52651
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Adding the appropriate GNU General Public License header and
update copyright year to 2014.
Change-ID: I769dd2d37d70350afd0c8727ae2859c0fd340361
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We don't need context descriptors for every packet, only tso
or timesync. This fixes a bug in the driver where it would
always add a context even if all the passed in values
to the context descriptor function were 0/default values.
Change-ID: I0101d2b893380707b5c2de61aab3e16d4310e9a1
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The hardware supports a feature to avoid updating the descriptor
ring by marking each descriptor with a DD bit, and instead
writes a memory location with an update to where the driver
should clean up to. Enable this feature.
Change-ID: I5da4e0681f0b581a6401c950a81808792267fe57
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We need to make sure that we stay away from the cache line
where the DD bit (done) may be getting written back for
the transmit ring since the hardware may be writing the
whole cache line for a partial update.
Change-ID: Id0b6dfc01f654def6a2a021af185803be1915d7e
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Jesse Brandeburg