XDP RxQ info contains frag_size, which depends on the MTU. This makes the
old way of registering RxQ info before calculating new buffer sizes
invalid. Currently, it leads to frag_size being outdated, making it
sometimes impossible to grow tailroom in a mbuf packet. E.g. fragments are
actually 3K+, but frag size is still as if MTU was 1500.
Always register new XDP RxQ info after reconfiguring memory pools.
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Link: https://patch.msgid.link/20260305111253.2317394-4-larysa.zaremba@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
The ice Tx/Rx hotpath has a few statistics counters for tracking unexpected
events. These values are stored as u64 but are not accumulated using the
u64_stats API. This could result in load/tear stores on some architectures.
Even some 64-bit architectures could have issues since the fields are not
read or written using ACCESS_ONCE or READ_ONCE.
A following change is going to refactor the stats accumulator code to use
the u64_stats API for all of these stats, and to use u64_stats_read and
u64_stats_inc properly to prevent load/store tears on all architectures.
Using u64_stats_inc and the syncp pointer is slightly verbose and would be
duplicated in a number of places in the Tx and Rx hot path. Add accessor
macros for the cases where only a single stat value is touched at once. To
keep lines short, also shorten the stats names and convert ice_txq_stats
and ice_rxq_stats to struct_group.
This will ease the transition to properly using the u64_stats API in the
following change.
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The __ice_update_sample and __ice_get_ethtool_stats functions directly
accesses the pkts and bytes counters from the ring stats. A following
change is going to update the fields to be u64_stats_t type, and will need
to be accessed appropriately. This will ensure that the accesses do not
cause load/store tearing.
Add helper functions similar to the ones used for updating the stats
values, and use them. This ensures use of the syncp pointer on 32-bit
architectures. Once the fields are updated to u64_stats_t, it will then
properly avoid tears on all architectures.
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Add second page_pool for header buffers to each Rx queue and ability
to toggle the header split on/off using Ethtool (default to off to
match the current behaviour).
Unlike idpf, all HW backed up by ice doesn't require any W/As and
correctly splits all types of packets as configured: after L4 headers
for TCP/UDP/SCTP, after L3 headers for other IPv4/IPv6 frames, after
the Ethernet header otherwise (in case of tunneling, same as above,
but after innermost headers).
This doesn't affect the XSk path as there are no benefits of having
it there.
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Tested-by: Alexander Nowlin <alexander.nowlin@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
This patch completes the transition of the ice driver to use the Page Pool
and libeth APIs, following the same direction as commit 5fa4caff59
("iavf: switch to Page Pool"). With the legacy page splitting and recycling
logic already removed, the driver is now in a clean state to adopt the
modern memory model.
The Page Pool integration simplifies buffer management by offloading
DMA mapping and recycling to the core infrastructure. This eliminates
the need for driver-specific handling of headroom, buffer sizing, and
page order. The libeth helper is used for CPU-side processing, while
DMA-for-device is handled by the Page Pool core.
Additionally, this patch extends the conversion to cover XDP support.
The driver now uses libeth_xdp helpers for Rx buffer processing,
and optimizes XDP_TX by skipping per-frame DMA mapping. Instead, all
buffers are mapped as bi-directional up front, leveraging Page Pool's
lifecycle management. This significantly reduces overhead in virtualized
environments.
Performance observations:
- In typical scenarios (netperf, XDP_PASS, XDP_DROP), performance remains
on par with the previous implementation.
- In XDP_TX mode:
* With IOMMU enabled, performance improves dramatically - over 5x
increase - due to reduced DMA mapping overhead and better memory reuse.
* With IOMMU disabled, performance remains comparable to the previous
implementation, with no significant changes observed.
- In XDP_DROP mode:
* For small MTUs, (where multiple buffers can be allocated on a single
memory page), a performance drop of approximately 20% is observed.
According to 'perf top' analysis, the bottleneck is caused by atomic
reference counter increments in the Page Pool.
* For normal MTUs, (where only one buffer can be allocated within a
single memory page), performance remains comparable to baseline
levels.
This change is also a step toward a more modular and unified XDP
implementation across Intel Ethernet drivers, aligning with ongoing
efforts to consolidate and streamline feature support.
Suggested-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Suggested-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Michal Kubiak <michal.kubiak@intel.com>
Tested-by: Alexander Nowlin <alexander.nowlin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
As part of the transition toward Page Pool integration, remove the
legacy page splitting and recycling logic from the ice driver. This
mirrors the approach taken in commit 920d86f3c5 ("iavf: drop page
splitting and recycling").
The previous model attempted to reuse partially consumed pages by
splitting them and tracking their usage across descriptors. While
this was once a memory optimization, it introduced significant
complexity and overhead in the Rx path, including:
- Manual refcount management and page reuse heuristics;
- Per-descriptor buffer shuffling, which could involve moving dozens
of `ice_rx_buf` structures per NAPI cycle;
- Increased branching and cache pressure in the hotpath.
This change simplifies the Rx logic by always allocating fresh pages
and letting the networking stack handle their lifecycle. Although this
may temporarily reduce performance (up to ~98% in some XDP cases), it
greatly improves maintainability and paves the way for Page Pool,
which will restore and exceed previous performance levels.
The `ice_rx_buf` array is retained for now to minimize diffstat and
ease future replacement with a shared buffer abstraction.
Co-developed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Michal Kubiak <michal.kubiak@intel.com>
Tested-by: Alexander Nowlin <alexander.nowlin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The commit 53844673d5 ("iavf: kill 'legacy-rx' for good") removed
the legacy Rx path in the iavf driver. This change applies the same
rationale to the ice driver.
The legacy Rx path relied on manual skb allocation and header copying,
which has become increasingly inefficient and difficult to maintain.
With the stabilization of build_skb() and the growing adoption of
features like XDP, page_pool, and multi-buffer support, the legacy
approach is no longer viable.
Key drawbacks of the legacy path included:
- Higher memory pressure due to direct page allocations and splitting;
- Redundant memcpy() operations for packet headers;
- CPU overhead from eth_get_headlen() and Flow Dissector usage;
- Compatibility issues with XDP, which imposes strict headroom and
tailroom requirements.
The ice driver, like iavf, does not benefit from the minimal headroom
savings that legacy Rx once offered, as it already splits pages into
fixed halves. Removing this path simplifies the Rx logic, eliminates
unnecessary branches in the hotpath, and prepares the driver for
upcoming enhancements.
In addition to removing the legacy Rx path, this change also eliminates
the custom construct_skb() functions from both the standard and
zero-copy (ZC) Rx paths. These are replaced with the build_skb()
and standardized xdp_build_skb_from_zc() helpers, aligning the driver
with the modern XDP infrastructure and reducing code duplication.
This cleanup also reduces code complexity and improves maintainability
as we move toward a more unified and modern Rx model across drivers.
Co-developed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Michal Kubiak <michal.kubiak@intel.com>
Tested-by: Alexander Nowlin <alexander.nowlin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
E830 supports Earliest TxTime First (ETF) hardware offload, which is
configured via the ETF Qdisc on a per-queue basis (see tc-etf(8)). ETF
introduces a new Tx flow mechanism that utilizes a timestamp ring
(tstamp_ring) alongside the standard Tx ring. This timestamp ring is
used to indicate when hardware will transmit a packet. Tx Time is
supported on the first 2048 Tx queues of the device, and the NVM image
limits the maximum number of Tx queues to 2048 for the device.
The allocation and initialization of the timestamp ring occur when the
feature is enabled on a specific Tx queue via tc-etf. The requested Tx
Time queue index cannot be greater than the number of Tx queues
(vsi->num_txq).
To support ETF, the following flags and bitmap are introduced:
- ICE_F_TXTIME: Device feature flag set for E830 NICs, indicating ETF
support.
- txtime_txqs: PF-level bitmap set when ETF is enabled and cleared
when disabled for a specific Tx queue. It is used by
ice_is_txtime_ena() to check if ETF is allocated and configured on
any Tx queue, which is checked during Tx ring allocation.
- ICE_TX_FLAGS_TXTIME: Per Tx ring flag set when ETF is allocated and
configured for a specific Tx queue. It determines ETF status during
packet transmission and is checked by ice_is_txtime_ena() to verify
if ETF is enabled on any Tx queue.
Due to a hardware issue that can result in a malicious driver detection
event, additional timestamp descriptors are required when wrapping
around the timestamp ring. Up to 64 additional timestamp descriptors
are reserved, reducing the available Tx descriptors.
To accommodate this, ICE_MAX_NUM_DESC_BY_MAC is introduced, defining:
- E830: Maximum Tx descriptor count of 8096 (8K - 32 - 64 for timestamp
fetch descriptors).
- E810 and E82X: Maximum Tx descriptor count of 8160 (8K - 32).
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Co-developed-by: Alice Michael <alice.michael@intel.com>
Signed-off-by: Alice Michael <alice.michael@intel.com>
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Cross-merge networking fixes after downstream PR (net-6.17-rc7).
No conflicts.
Adjacent changes:
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h
9536fbe10c ("net/mlx5e: Add PSP steering in local NIC RX")
7601a0a462 ("net/mlx5e: Add a miss level for ipsec crypto offload")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The ice_put_rx_mbuf() function handles calling ice_put_rx_buf() for each
buffer in the current frame. This function was introduced as part of
handling multi-buffer XDP support in the ice driver.
It works by iterating over the buffers from first_desc up to 1 plus the
total number of fragments in the frame, cached from before the XDP program
was executed.
If the hardware posts a descriptor with a size of 0, the logic used in
ice_put_rx_mbuf() breaks. Such descriptors get skipped and don't get added
as fragments in ice_add_xdp_frag. Since the buffer isn't counted as a
fragment, we do not iterate over it in ice_put_rx_mbuf(), and thus we don't
call ice_put_rx_buf().
Because we don't call ice_put_rx_buf(), we don't attempt to re-use the
page or free it. This leaves a stale page in the ring, as we don't
increment next_to_alloc.
The ice_reuse_rx_page() assumes that the next_to_alloc has been incremented
properly, and that it always points to a buffer with a NULL page. Since
this function doesn't check, it will happily recycle a page over the top
of the next_to_alloc buffer, losing track of the old page.
Note that this leak only occurs for multi-buffer frames. The
ice_put_rx_mbuf() function always handles at least one buffer, so a
single-buffer frame will always get handled correctly. It is not clear
precisely why the hardware hands us descriptors with a size of 0 sometimes,
but it happens somewhat regularly with "jumbo frames" used by 9K MTU.
To fix ice_put_rx_mbuf(), we need to make sure to call ice_put_rx_buf() on
all buffers between first_desc and next_to_clean. Borrow the logic of a
similar function in i40e used for this same purpose. Use the same logic
also in ice_get_pgcnts().
Instead of iterating over just the number of fragments, use a loop which
iterates until the current index reaches to the next_to_clean element just
past the current frame. Unlike i40e, the ice_put_rx_mbuf() function does
call ice_put_rx_buf() on the last buffer of the frame indicating the end of
packet.
For non-linear (multi-buffer) frames, we need to take care when adjusting
the pagecnt_bias. An XDP program might release fragments from the tail of
the frame, in which case that fragment page is already released. Only
update the pagecnt_bias for the first descriptor and fragments still
remaining post-XDP program. Take care to only access the shared info for
fragmented buffers, as this avoids a significant cache miss.
The xdp_xmit value only needs to be updated if an XDP program is run, and
only once per packet. Drop the xdp_xmit pointer argument from
ice_put_rx_mbuf(). Instead, set xdp_xmit in the ice_clean_rx_irq() function
directly. This avoids needing to pass the argument and avoids an extra
bit-wise OR for each buffer in the frame.
Move the increment of the ntc local variable to ensure its updated *before*
all calls to ice_get_pgcnts() or ice_put_rx_mbuf(), as the loop logic
requires the index of the element just after the current frame.
Now that we use an index pointer in the ring to identify the packet, we no
longer need to track or cache the number of fragments in the rx_ring.
Cc: Christoph Petrausch <christoph.petrausch@deepl.com>
Cc: Jesper Dangaard Brouer <hawk@kernel.org>
Reported-by: Jaroslav Pulchart <jaroslav.pulchart@gooddata.com>
Closes: https://lore.kernel.org/netdev/CAK8fFZ4hY6GUJNENz3wY9jaYLZXGfpr7dnZxzGMYoE44caRbgw@mail.gmail.com/
Fixes: 743bbd93cf ("ice: put Rx buffers after being done with current frame")
Tested-by: Michal Kubiak <michal.kubiak@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Tested-by: Priya Singh <priyax.singh@intel.com>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
xdp_update_skb_shared_info() needs to update skb state which
was maintained in xdp_buff / frame. Pass full flags into it,
instead of breaking it out bit by bit. We will need to add
a bit for unreadable frags (even tho XDP doesn't support
those the driver paths may be common), at which point almost
all call sites would become:
xdp_update_skb_shared_info(skb, num_frags,
sinfo->xdp_frags_size,
MY_PAGE_SIZE * num_frags,
xdp_buff_is_frag_pfmemalloc(xdp),
xdp_buff_is_frag_unreadable(xdp));
Keep a helper for accessing the flags, in case we need to
transform them somehow in the future (e.g. to cover up xdp_buff
vs xdp_frame differences).
While we are touching call callers - rename the helper to
xdp_update_skb_frags_info(), previous name may have implied that
it's shinfo that's updated. We are updating flags in struct sk_buff
based on frags that got attched.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://patch.msgid.link/20250905221539.2930285-2-kuba@kernel.org
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, the driver increments `alloc_page_failed` when buffer allocation fails
in `ice_clean_rx_irq()`. However, this counter is intended for page allocation
failures, not buffer allocation issues.
This patch corrects the counter by incrementing `alloc_buf_failed` instead,
ensuring accurate statistics reporting for buffer allocation failures.
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Reported-by: Jacob Keller <jacob.e.keller@intel.com>
Suggested-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Michal Kubiak <michal.kubiak@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Tested-by: Priya Singh <priyax.singh@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The "ice" driver implementation uses the control VSI to handle
the flow director configuration for PFs and VFs.
Unfortunately, although a separate VSI type was created to handle flow
director queues, the Rx queue handler was shared between the flow
director and a standard NAPI Rx handler.
Such a design approach was not very flexible. First, it mixed hotpath
and slowpath code, blocking their further optimization. It also created
a huge overkill for the flow director command processing, which is
descriptor-based only, so there is no need to allocate Rx data buffers.
For the above reasons, implement a separate Rx handler for the control
VSI. Also, remove from the NAPI handler the code dedicated to
configuring the flow director rules on VFs.
Do not allocate Rx data buffers to the flow director queues because
their processing is descriptor-based only.
Finally, allow Rx data queues to be allocated only for VSIs that have
netdev assigned to them.
This handler splitting approach is the first step in converting the
driver to use the Page Pool (which can only be used for data queues).
Test hints:
1. Create a VF for any PF managed by the ice driver.
2. In a loop, add and delete flow director rules for the VF, e.g.:
for i in {1..128}; do
q=$(( i % 16 ))
ethtool -N ens802f0v0 flow-type tcp4 dst-port "$i" action "$q"
done
for i in {0..127}; do
ethtool -N ens802f0v0 delete "$i"
done
Suggested-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Suggested-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Michal Kubiak <michal.kubiak@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
tc clsact qdisc allows us to add offloaded egress rules with commands such
as the following one:
tc filter add dev <ifname> egress protocol lldp flower skip_sw action drop
Support the egress rule drop action when added to PF, with a few caveats:
* in switchdev mode, all PF traffic has to go uplink with an exception for
LLDP that can be delegated to a single VSI at a time
* in legacy mode, we cannot delegate LLDP functionality to another VSI, so
such packets from PF should not be blocked.
Also, simplify the rule direction logic, it was previously derived from
actions, but actually can be inherited from the tc block (and flipped in
case of port representors).
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
E830 supports raw receive and generic transmit checksum offloads.
Raw receive checksum support is provided by hardware calculating the
checksum over the whole packet, regardless of type. The calculated
checksum is provided to driver in the Rx flex descriptor. Then the driver
assigns the checksum to skb->csum and sets skb->ip_summed to
CHECKSUM_COMPLETE.
Generic transmit checksum support is provided by hardware calculating the
checksum given two offsets: the start offset to begin checksum calculation,
and the offset to insert the calculated checksum in the packet. Support is
advertised to the stack using NETIF_F_HW_CSUM feature.
E830 has the following limitations when both generic transmit checksum
offload and TCP Segmentation Offload (TSO) are enabled:
1. Inner packet header modification is not supported. This restriction
includes the inability to alter TCP flags, such as the push flag. As a
result, this limitation can impact the receiver's ability to coalesce
packets, potentially degrading network throughput.
2. The Maximum Segment Size (MSS) is limited to 1023 bytes, which prevents
support of Maximum Transmission Unit (MTU) greater than 1063 bytes.
Therefore NETIF_F_HW_CSUM and NETIF_F_ALL_TSO features are mutually
exclusive. NETIF_F_HW_CSUM hardware feature support is indicated but is not
enabled by default. Instead, IP checksums and NETIF_F_ALL_TSO are the
defaults. Enforcement of mutual exclusivity of NETIF_F_HW_CSUM and
NETIF_F_ALL_TSO is done in ice_set_features(). Mutual exclusivity
of IP checksums and NETIF_F_HW_CSUM is handled by netdev_fix_features().
When NETIF_F_HW_CSUM is requested the provided skb->csum_start and
skb->csum_offset are passed to hardware in the Tx context descriptor
generic checksum (GCS) parameters. Hardware calculates the 1's complement
from skb->csum_start to the end of the packet, and inserts the result in
the packet at skb->csum_offset.
Co-developed-by: Alice Michael <alice.michael@intel.com>
Signed-off-by: Alice Michael <alice.michael@intel.com>
Co-developed-by: Eric Joyner <eric.joyner@intel.com>
Signed-off-by: Eric Joyner <eric.joyner@intel.com>
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://patch.msgid.link/20250310174502.3708121-2-anthony.l.nguyen@intel.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Ever since removing switchdev control VSI and using PF for port
representor Tx/Rx, switchdev slow-path has been working improperly after
failover in SR-IOV LAG. LAG assumes that the first uplink to be added to
the aggregate will own VFs and have switchdev configured. After
failing-over to the other uplink, representors are still configured to
Tx through the uplink they are set up on, which fails because that
uplink is now down.
On failover, update all PRs on primary uplink to use the currently
active uplink for Tx. Call netif_keep_dst(), as the secondary uplink
might not be in switchdev mode. Also make sure to call
ice_eswitch_set_target_vsi() if uplink is in LAG.
On the Rx path, representors are already working properly, because
default Tx from VFs is set to PF owning the eswitch. After failover the
same PF is receiving traffic from VFs, even though link is down.
Fixes: defd52455a ("ice: do Tx through PF netdev in slow-path")
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Tested-by: Sujai Buvaneswaran <sujai.buvaneswaran@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Idea behind having ice_rx_buf::act was to simplify and speed up the Rx
data path by walking through buffers that were representing cleaned HW
Rx descriptors. Since it caused us a major headache recently and we
rolled back to old approach that 'puts' Rx buffers right after running
XDP prog/creating skb, this is useless now and should be removed.
Get rid of ice_rx_buf::act and related logic. We still need to take care
of a corner case where XDP program releases a particular fragment.
Make ice_run_xdp() to return its result and use it within
ice_put_rx_mbuf().
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
If we store the pgcnt on few fragments while being in the middle of
gathering the whole frame and we stumbled upon DD bit not being set, we
terminate the NAPI Rx processing loop and come back later on. Then on
next NAPI execution we work on previously stored pgcnt.
Imagine that second half of page was used actively by networking stack
and by the time we came back, stack is not busy with this page anymore
and decremented the refcnt. The page reuse algorithm in this case should
be good to reuse the page but given the old refcnt it will not do so and
attempt to release the page via page_frag_cache_drain() with
pagecnt_bias used as an arg. This in turn will result in negative refcnt
on struct page, which was initially observed by Xu Du.
Therefore, move the page count storage from ice_get_rx_buf() to a place
where we are sure that whole frame has been collected, but before
calling XDP program as it internally can also change the page count of
fragments belonging to xdp_buff.
Fixes: ac07533911 ("ice: Store page count inside ice_rx_buf")
Reported-and-tested-by: Xu Du <xudu@redhat.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Co-developed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Introduce a new helper ice_put_rx_mbuf() that will go through gathered
frags from current frame and will call ice_put_rx_buf() on them. Current
logic that was supposed to simplify and optimize the driver where we go
through a batch of all buffers processed in current NAPI instance turned
out to be broken for jumbo frames and very heavy load that was coming
from both multi-thread iperf and nginx/wrk pair between server and
client. The delay introduced by approach that we are dropping is simply
too big and we need to take the decision regarding page
recycling/releasing as quick as we can.
While at it, address an error path of ice_add_xdp_frag() - we were
missing buffer putting from day 1 there.
As a nice side effect we get rid of annoying and repetitive three-liner:
xdp->data = NULL;
rx_ring->first_desc = ntc;
rx_ring->nr_frags = 0;
by embedding it within introduced routine.
Fixes: 1dc1a7e7f4 ("ice: Centrallize Rx buffer recycling")
Reported-and-tested-by: Xu Du <xudu@redhat.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Co-developed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
net_dim() is currently passed a struct dim_sample argument by value.
struct dim_sample is 24 bytes. Since this is greater 16 bytes, x86-64
passes it on the stack. All callers have already initialized dim_sample
on the stack, so passing it by value requires pushing a duplicated copy
to the stack. Either witing to the stack and immediately reading it, or
perhaps dereferencing addresses relative to the stack pointer in a chain
of push instructions, seems to perform quite poorly.
In a heavy TCP workload, mlx5e_handle_rx_dim() consumes 3% of CPU time,
94% of which is attributed to the first push instruction to copy
dim_sample on the stack for the call to net_dim():
// Call ktime_get()
0.26 |4ead2: call 4ead7 <mlx5e_handle_rx_dim+0x47>
// Pass the address of struct dim in %rdi
|4ead7: lea 0x3d0(%rbx),%rdi
// Set dim_sample.pkt_ctr
|4eade: mov %r13d,0x8(%rsp)
// Set dim_sample.byte_ctr
|4eae3: mov %r12d,0xc(%rsp)
// Set dim_sample.event_ctr
0.15 |4eae8: mov %bp,0x10(%rsp)
// Duplicate dim_sample on the stack
94.16 |4eaed: push 0x10(%rsp)
2.79 |4eaf1: push 0x10(%rsp)
0.07 |4eaf5: push %rax
// Call net_dim()
0.21 |4eaf6: call 4eafb <mlx5e_handle_rx_dim+0x6b>
To allow the caller to reuse the struct dim_sample already on the stack,
pass the struct dim_sample by reference to net_dim().
Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Arthur Kiyanovski <akiyano@amazon.com>
Reviewed-by: Louis Peens <louis.peens@corigine.com>
Link: https://patch.msgid.link/20241031002326.3426181-2-csander@purestorage.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When working on multi-buffer packet on arch that has PAGE_SIZE >= 8192,
truesize is calculated and stored in xdp_buff::frame_sz per each
processed Rx buffer. This means that frame_sz will contain the truesize
based on last received buffer, but commit 1dc1a7e7f4 ("ice:
Centrallize Rx buffer recycling") assumed this value will be constant
for each buffer, which breaks the page recycling scheme and mess up the
way we update the page::page_offset.
To fix this, let us work on constant truesize when PAGE_SIZE >= 8192
instead of basing this on size of a packet read from Rx descriptor. This
way we can simplify the code and avoid calculating truesize per each
received frame and on top of that when using
xdp_update_skb_shared_info(), current formula for truesize update will
be valid.
This means ice_rx_frame_truesize() can be removed altogether.
Furthermore, first call to it within ice_clean_rx_irq() for 4k PAGE_SIZE
was redundant as xdp_buff::frame_sz is initialized via xdp_init_buff()
in ice_vsi_cfg_rxq(). This should have been removed at the point where
xdp_buff struct started to be a member of ice_rx_ring and it was no
longer a stack based variable.
There are two fixes tags as my understanding is that the first one
exposed us to broken truesize and page_offset handling and then second
introduced broken skb_shared_info update in ice_{construct,build}_skb().
Reported-and-tested-by: Luiz Capitulino <luizcap@redhat.com>
Closes: https://lore.kernel.org/netdev/8f9e2a5c-fd30-4206-9311-946a06d031bb@redhat.com/
Fixes: 1dc1a7e7f4 ("ice: Centrallize Rx buffer recycling")
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
For bigger PAGE_SIZE archs, ice driver works on 3k Rx buffers.
Therefore, ICE_LAST_OFFSET should take into account ICE_RXBUF_3072, not
ICE_RXBUF_2048.
Fixes: 7237f5b0db ("ice: introduce legacy Rx flag")
Suggested-by: Luiz Capitulino <luizcap@redhat.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Architectures that have PAGE_SIZE >= 8192 such as arm64 should act the
same as x86 currently, meaning reuse of a page should only take place
when no one else is busy with it.
Do two things independently of underlying PAGE_SIZE:
- store the page count under ice_rx_buf::pgcnt
- then act upon its value vs ice_rx_buf::pagecnt_bias when making the
decision regarding page reuse
Fixes: 2b245cb294 ("ice: Implement transmit and NAPI support")
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
xsk_buff_pool pointers that ice ring structs hold are updated via
ndo_bpf that is executed in process context while it can be read by
remote CPU at the same time within NAPI poll. Use synchronize_net()
after pointer update and {READ,WRITE}_ONCE() when working with mentioned
pointer.
Fixes: 2d4238f556 ("ice: Add support for AF_XDP")
Reviewed-by: Shannon Nelson <shannon.nelson@amd.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
__napi_alloc_skb() is napi_alloc_skb() with the added flexibility
of choosing gfp_mask. This is a NAPI function, so GFP_ATOMIC is
implied. The only practical choice the caller has is whether to
set __GFP_NOWARN. But that's a false choice, too, allocation failures
in atomic context will happen, and printing warnings in logs,
effectively for a packet drop, is both too much and very likely
non-actionable.
This leads me to a conclusion that most uses of napi_alloc_skb()
are simply misguided, and should use __GFP_NOWARN in the first
place. We also have a "standard" way of reporting allocation
failures via the queue stat API (qstats::rx-alloc-fail).
The direct motivation for this patch is that one of the drivers
used at Meta calls napi_alloc_skb() (so prior to this patch without
__GFP_NOWARN), and the resulting OOM warning is the top networking
warning in our fleet.
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20240327040213.3153864-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
xdp_rxq_info struct can be registered by drivers via two functions -
xdp_rxq_info_reg() and __xdp_rxq_info_reg(). The latter one allows
drivers that support XDP multi-buffer to set up xdp_rxq_info::frag_size
which in turn will make it possible to grow the packet via
bpf_xdp_adjust_tail() BPF helper.
Currently, ice registers xdp_rxq_info in two spots:
1) ice_setup_rx_ring() // via xdp_rxq_info_reg(), BUG
2) ice_vsi_cfg_rxq() // via __xdp_rxq_info_reg(), OK
Cited commit under fixes tag took care of setting up frag_size and
updated registration scheme in 2) but it did not help as
1) is called before 2) and as shown above it uses old registration
function. This means that 2) sees that xdp_rxq_info is already
registered and never calls __xdp_rxq_info_reg() which leaves us with
xdp_rxq_info::frag_size being set to 0.
To fix this misbehavior, simply remove xdp_rxq_info_reg() call from
ice_setup_rx_ring().
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20240124191602.566724-7-maciej.fijalkowski@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Fix an OOM panic in XDP_DRV mode when a XDP program shrinks a
multi-buffer packet by 4k bytes and then redirects it to an AF_XDP
socket.
Since support for handling multi-buffer frames was added to XDP, usage
of bpf_xdp_adjust_tail() helper within XDP program can free the page
that given fragment occupies and in turn decrease the fragment count
within skb_shared_info that is embedded in xdp_buff struct. In current
ice driver codebase, it can become problematic when page recycling logic
decides not to reuse the page. In such case, __page_frag_cache_drain()
is used with ice_rx_buf::pagecnt_bias that was not adjusted after
refcount of page was changed by XDP prog which in turn does not drain
the refcount to 0 and page is never freed.
To address this, let us store the count of frags before the XDP program
was executed on Rx ring struct. This will be used to compare with
current frag count from skb_shared_info embedded in xdp_buff. A smaller
value in the latter indicates that XDP prog freed frag(s). Then, for
given delta decrement pagecnt_bias for XDP_DROP verdict.
While at it, let us also handle the EOP frag within
ice_set_rx_bufs_act() to make our life easier, so all of the adjustments
needed to be applied against freed frags are performed in the single
place.
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20240124191602.566724-5-maciej.fijalkowski@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tony Nguyen says:
====================
intel: use bitfield operations
Jesse Brandeburg says:
After repeatedly getting review comments on new patches, and sporadic
patches to fix parts of our drivers, we should just convert the Intel code
to use FIELD_PREP() and FIELD_GET(). It's then "common" in the code and
hopefully future change-sets will see the context and do-the-right-thing.
This conversion was done with a coccinelle script which is mentioned in the
commit messages. Generally there were only a couple conversions that were
"undone" after the automatic changes because they tried to convert a
non-contiguous mask.
Patch 1 is required at the beginning of this series to fix a "forever"
issue in the e1000e driver that fails the compilation test after conversion
because the shift / mask was out of range.
The second patch just adds all the new #includes in one go.
The patch titled: "ice: fix pre-shifted bit usage" is needed to allow the
use of the FIELD_* macros and fix up the unexpected "shifts included"
defines found while creating this series.
The rest are the conversion to use FIELD_PREP()/FIELD_GET(), and the
occasional leXX_{get,set,encode}_bits() call, as suggested by Alex.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactor ice driver to use FIELD_PREP(), which reduces lines of code
and adds clarity of intent.
This code was generated by the following coccinelle/spatch script and
then manually repaired.
Several places I changed to OR into a single variable with |= instead of
using a multi-line statement with trailing OR operators, as it
(subjectively) makes the code clearer.
A local variable vmvf_and_timeout was created and used to avoid multiple
logical ORs being __le16 converted, which shortened some lines and makes
the code cleaner.
Also clean up a couple of places where conversions were made to have the
code read more clearly/consistently.
@prep2@
constant shift,mask;
type T;
expression a;
@@
-(((T)(a) << shift) & mask)
+FIELD_PREP(mask, a)
@prep@
constant shift,mask;
type T;
expression a;
@@
-((T)((a) << shift) & mask)
+FIELD_PREP(mask, a)
Cc: Julia Lawall <Julia.Lawall@inria.fr>
CC: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
In order to use XDP hints via kfuncs we need to put
RX descriptor and miscellaneous data next to xdp_buff.
Same as in hints implementations in other drivers, we achieve
this through putting xdp_buff into a child structure.
Currently, xdp_buff is stored in the ring structure,
so replace it with union that includes child structure.
This way enough memory is available while existing XDP code
remains isolated from hints.
Minimum size of the new child structure (ice_xdp_buff) is exactly
64 bytes (single cache line). To place it at the start of a cache line,
move 'next' field from CL1 to CL4, as it isn't used often. This still
leaves 192 bits available in CL3 for packet context extensions.
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20231205210847.28460-5-larysa.zaremba@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Currently, rx_ptype variable is used only as an argument
to ice_process_skb_fields() and is computed
just before the function call.
Therefore, there is no reason to pass this value as an argument.
Instead, remove this argument and compute the value directly inside
ice_process_skb_fields() function.
Also, separate its calculation into a short function, so the code
can later be reused in .xmo_() callbacks.
Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Link: https://lore.kernel.org/r/20231205210847.28460-4-larysa.zaremba@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Before performing a Tx timestamp in ice_stamp(), the driver checks a ptp_tx
ring variable to see if timestamping is enabled on that ring. This value is
set for all rings whenever userspace configures Tx timestamping.
Ostensibly this was done to avoid wasting cycles checking other fields when
timestamping has not been enabled. However, for Tx timestamps we already
get an individual per-SKB flag indicating whether userspace wants to
request a timestamp on that packet. We do not gain much by also having
a separate flag to check for whether timestamping was enabled.
In fact, the driver currently fails to restore the field after a PF reset.
Because of this, if a PF reset occurs, timestamps will be disabled.
Since this flag doesn't add value in the hotpath, remove it and always
provide a timestamp if the SKB flag has been set.
A following change will fix the reset path to properly restore user
timestamping configuration completely.
This went unnoticed for some time because one of the most common
applications using Tx timestamps, ptp4l, will reconfigure the socket as
part of its fault recovery logic.
Fixes: ea9b847cda ("ice: enable transmit timestamps for E810 devices")
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
As not all ICE_TX_FLAGS_* fit in current 16-bit limited
tx_flags field that was introduced in the Fixes commit,
VLAN-related information would be discarded completely.
As such, creating a vlan and trying to run ping through
would result in no traffic passing.
Fix that by refactoring tx_flags variable into flags only and
a separate variable that holds VLAN ID. As there is some space left,
type variable can fit between those two. Pahole reports no size
change to ice_tx_buf struct.
Fixes: aa1d3faf71 ("ice: Robustify cleaning/completing XDP Tx buffers")
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
make modules W=1 returns:
.../ice/ice_txrx_lib.c:448: warning: Function parameter or member 'first_idx' not described in 'ice_finalize_xdp_rx'
.../ice/ice_txrx.c:948: warning: Function parameter or member 'ntc' not described in 'ice_get_rx_buf'
.../ice/ice_txrx.c:1038: warning: Excess function parameter 'rx_buf' description in 'ice_construct_skb'
Fix these warnings by adding and deleting the deviant arguments.
Fixes: 2fba7dc515 ("ice: Add support for XDP multi-buffer on Rx side")
Fixes: d7956d81f1 ("ice: Pull out next_to_clean bump out of ice_put_rx_buf()")
CC: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: Piotr Raczynski <piotr.raczynski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Daniel Borkmann says:
====================
pull-request: bpf-next 2023-02-17
We've added 64 non-merge commits during the last 7 day(s) which contain
a total of 158 files changed, 4190 insertions(+), 988 deletions(-).
The main changes are:
1) Add a rbtree data structure following the "next-gen data structure"
precedent set by recently-added linked-list, that is, by using
kfunc + kptr instead of adding a new BPF map type, from Dave Marchevsky.
2) Add a new benchmark for hashmap lookups to BPF selftests,
from Anton Protopopov.
3) Fix bpf_fib_lookup to only return valid neighbors and add an option
to skip the neigh table lookup, from Martin KaFai Lau.
4) Add cgroup.memory=nobpf kernel parameter option to disable BPF memory
accouting for container environments, from Yafang Shao.
5) Batch of ice multi-buffer and driver performance fixes,
from Alexander Lobakin.
6) Fix a bug in determining whether global subprog's argument is
PTR_TO_CTX, which is based on type names which breaks kprobe progs,
from Andrii Nakryiko.
7) Prep work for future -mcpu=v4 LLVM option which includes usage of
BPF_ST insn. Thus improve BPF_ST-related value tracking in verifier,
from Eduard Zingerman.
8) More prep work for later building selftests with Memory Sanitizer
in order to detect usages of undefined memory, from Ilya Leoshkevich.
9) Fix xsk sockets to check IFF_UP earlier to avoid a NULL pointer
dereference via sendmsg(), from Maciej Fijalkowski.
10) Implement BPF trampoline for RV64 JIT compiler, from Pu Lehui.
11) Fix BPF memory allocator in combination with BPF hashtab where it could
corrupt special fields e.g. used in bpf_spin_lock, from Hou Tao.
12) Fix LoongArch BPF JIT to always use 4 instructions for function
address so that instruction sequences don't change between passes,
from Hengqi Chen.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (64 commits)
selftests/bpf: Add bpf_fib_lookup test
bpf: Add BPF_FIB_LOOKUP_SKIP_NEIGH for bpf_fib_lookup
riscv, bpf: Add bpf trampoline support for RV64
riscv, bpf: Add bpf_arch_text_poke support for RV64
riscv, bpf: Factor out emit_call for kernel and bpf context
riscv: Extend patch_text for multiple instructions
Revert "bpf, test_run: fix &xdp_frame misplacement for LIVE_FRAMES"
selftests/bpf: Add global subprog context passing tests
selftests/bpf: Convert test_global_funcs test to test_loader framework
bpf: Fix global subprog context argument resolution logic
LoongArch, bpf: Use 4 instructions for function address in JIT
bpf: bpf_fib_lookup should not return neigh in NUD_FAILED state
bpf: Disable bh in bpf_test_run for xdp and tc prog
xsk: check IFF_UP earlier in Tx path
Fix typos in selftest/bpf files
selftests/bpf: Use bpf_{btf,link,map,prog}_get_info_by_fd()
samples/bpf: Use bpf_{btf,link,map,prog}_get_info_by_fd()
bpftool: Use bpf_{btf,link,map,prog}_get_info_by_fd()
libbpf: Use bpf_{btf,link,map,prog}_get_info_by_fd()
libbpf: Introduce bpf_{btf,link,map,prog}_get_info_by_fd()
...
====================
Link: https://lore.kernel.org/r/20230217221737.31122-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Enable sending BIG TCP packets on IPv6 in the ice driver using generic
ipv6_hopopt_jumbo_remove helper for stripping HBH header.
Tested:
netperf -t TCP_RR -H 2001:db8:0:f101::1 -- -r80000,80000 -O MIN_LATENCY,P90_LATENCY,P99_LATENCY,TRANSACTION_RATE
Tested on two different setups. In both cases, the following settings were
applied after loading the changed driver:
ip link set dev enp175s0f1np1 gso_max_size 130000
ip link set dev enp175s0f1np1 gro_max_size 130000
ip link set dev enp175s0f1np1 mtu 9000
First setup:
Before:
Minimum 90th 99th Transaction
Latency Percentile Percentile Rate
Microseconds Latency Latency Tran/s
Microseconds Microseconds
134 279 410 3961.584
After:
Minimum 90th 99th Transaction
Latency Percentile Percentile Rate
Microseconds Latency Latency Tran/s
Microseconds Microseconds
135 178 216 6093.404
The other setup:
Before:
Minimum 90th 99th Transaction
Latency Percentile Percentile Rate
Microseconds Latency Latency Tran/s
Microseconds Microseconds
218 414 478 2944.765
After:
Minimum 90th 99th Transaction
Latency Percentile Percentile Rate
Microseconds Latency Latency Tran/s
Microseconds Microseconds
146 238 266 4700.596
Signed-off-by: Pawel Chmielewski <pawel.chmielewski@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
After the recent mbuf changes, ice_xmit_xdp_ring() became a 3-liner.
It makes no sense to keep it global in a different file than its caller.
Move it just next to the sole call site and mark static. Also, it
doesn't need a full xdp_convert_frame_to_buff(). Save several cycles
and fill only the fields used by __ice_xmit_xdp_ring() later on.
Finally, since it doesn't modify @xdpf anyhow, mark the argument const
to save some more (whole -11 bytes of .text! :D).
Thanks to 1 jump less and less calcs as well, this yields as many as
6.7 Mpps per queue. `xdp.data_hard_start = xdpf` is fully intentional
again (see xdp_convert_buff_to_frame()) and just works when there are
no source device's driver issues.
Signed-off-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/bpf/20230210170618.1973430-7-alexandr.lobakin@intel.com
As already mentioned, freeing any &xdp_frame via page_frag_free() is
wrong, as it assumes the frame is backed by either an order-0 page or
a page with no "patrons" behind them, while in fact frames backed by
Page Pool can be redirected to a device, which's driver doesn't use it.
Keep storing a pointer to the raw buffer and then freeing it
unconditionally via page_frag_free() for %XDP_TX frames, but introduce
a separate type in the enum for frames coming through .ndo_xdp_xmit(),
and free them via xdp_return_frame_bulk(). Note that saving xdpf as
xdp_buff->data_hard_start is intentional and is always true when
everything is configured properly.
After this change, %XDP_REDIRECT from a Page Pool based driver to ice
becomes zero-alloc as it should be and horrendous 3.3 Mpps / queue
turn into 6.6, hehe.
Let it go with no "Fixes:" tag as it spans across good 5+ commits and
can't be trivially backported.
Signed-off-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/bpf/20230210170618.1973430-6-alexandr.lobakin@intel.com
When queueing frames from a Page Pool for redirecting to a device backed
by the ice driver, `perf top` shows heavy load on page_alloc() and
page_frag_free(), despite that on a properly working system it must be
fully or at least almost zero-alloc. The problem is in fact a bit deeper
and raises from how ice cleans up completed Tx buffers.
The story so far: when cleaning/freeing the resources related to
a particular completed Tx frame (skbs, DMA mappings etc.), ice uses some
heuristics only without setting any type explicitly (except for dummy
Flow Director packets, which are marked via ice_tx_buf::tx_flags).
This kinda works, but only up to some point. For example, currently ice
assumes that each frame coming to __ice_xmit_xdp_ring(), is backed by
either plain order-0 page or plain page frag, while it may also be
backed by Page Pool or any other possible memory models introduced in
future. This means any &xdp_frame must be freed properly via
xdp_return_frame() family with no assumptions.
In order to do that, the whole heuristics must be replaced with setting
the Tx buffer/frame type explicitly, just how it's always been done via
an enum. Let us reuse 16 bits from ::tx_flags -- 1 bit-and instr won't
hurt much -- especially given that sometimes there was a check for
%ICE_TX_FLAGS_DUMMY_PKT, which is now turned from a flag to an enum
member. The rest of the changes is straightforward and most of it is
just a conversion to rely now on the type set in &ice_tx_buf rather than
to some secondary properties.
For now, no functional changes intended, the change only prepares the
ground for starting freeing XDP frames properly next step. And it must
be done atomically/synchronously to not break stuff.
Signed-off-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/bpf/20230210170618.1973430-5-alexandr.lobakin@intel.com
The tagged commit started sending %XDP_TX frames from XSk Rx ring
directly without converting it to an &xdp_frame. However, when XSk is
enabled on a queue pair, it has its separate Tx cleaning functions, so
neither ice_clean_xdp_irq() nor ice_unmap_and_free_tx_buf() ever happens
there.
Remove impossible branches in order to reduce the diffstat of the
upcoming change.
Fixes: a24b4c6e9a ("ice: xsk: Do not convert to buff to frame for XDP_TX")
Suggested-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/bpf/20230210170618.1973430-4-alexandr.lobakin@intel.com
