This commit adds support for Enhanced TX MPWQE feature in the regular
(SKB) data path. A MPWQE (multi-packet work queue element) can serve
multiple packets, reducing the PCI bandwidth on control traffic.
Two new stats (tx*_mpwqe_blks and tx*_mpwqe_pkts) are added. The feature
is on by default and controlled by the skb_tx_mpwqe private flag.
In a MPWQE, eseg is shared among all packets, so eseg-based offloads
(IPSEC, GENEVE, checksum) run on a separate eseg that is compared to the
eseg of the current MPWQE session to decide if the new packet can be
added to the same session.
MPWQE is not compatible with certain offloads and features, such as TLS
offload, TSO, nonlinear SKBs. If such incompatible features are in use,
the driver gracefully falls back to non-MPWQE.
This change has no performance impact in TCP single stream test and
XDP_TX single stream test.
UDP pktgen, 64-byte packets, single stream, MPWQE off:
Packet rate: 16.96 Mpps (±0.12 Mpps) -> 17.01 Mpps (±0.20 Mpps)
Instructions per packet: 421 -> 429
Cycles per packet: 156 -> 161
Instructions per cycle: 2.70 -> 2.67
UDP pktgen, 64-byte packets, single stream, MPWQE on:
Packet rate: 16.96 Mpps (±0.12 Mpps) -> 20.94 Mpps (±0.33 Mpps)
Instructions per packet: 421 -> 329
Cycles per packet: 156 -> 123
Instructions per cycle: 2.70 -> 2.67
Enabling MPWQE can reduce PCI bandwidth:
PCI Gen2, pktgen at fixed rate of 36864000 pps on 24 CPU cores:
Inbound PCI utilization with MPWQE off: 80.3%
Inbound PCI utilization with MPWQE on: 59.0%
PCI Gen3, pktgen at fixed rate of 56064000 pps on 24 CPU cores:
Inbound PCI utilization with MPWQE off: 65.4%
Inbound PCI utilization with MPWQE on: 49.3%
Enabling MPWQE can also reduce CPU load, increasing the packet rate in
case of CPU bottleneck:
PCI Gen2, pktgen at full rate on 24 CPU cores:
Packet rate with MPWQE off: 37.5 Mpps
Packet rate with MPWQE on: 49.0 Mpps
PCI Gen3, pktgen at full rate on 24 CPU cores:
Packet rate with MPWQE off: 57.0 Mpps
Packet rate with MPWQE on: 66.8 Mpps
Burst size in all pktgen tests is 32.
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz (x86_64)
NIC: Mellanox ConnectX-6 Dx
GCC 10.2.0
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
mlx5e_txwqe_complete performs some actions that can be taken to separate
functions:
1. Update the flags needed for hardware timestamping.
2. Stop the TX queue if it's full.
Take these actions into separate functions to be reused by the MPWQE
code in the following commit and to maintain clear responsibilities of
functions.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
As preparation for the upcoming TX MPWQE support for SKBs, rename struct
mlx5e_xdp_mpwqe to mlx5e_tx_mpwqe and move it above struct mlx5e_txqsq.
This structure will be reused in the regular SQ and in the regular TX
data path. Also rename mlx5e_xdp_xmit_data to mlx5e_xmit_data - it will
be used in the upcoming TX MPWQE flow.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
As preparation for the upcoming TX MPWQE for SKBs, create a function
(mlx5e_tx_mpwqe_is_full) to check whether an MPWQE session is full. This
function will be shared by MPWQE code for XDP and for SKBs. Defines are
renamed and moved to make them not XDP-specific.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
TX MPWQE support for SKBs is coming in one of the following patches, and
a single MPWQE can send multiple SKBs. This commit prepares the TX path
code to handle such cases:
1. An additional FIFO for SKBs is added, just like the FIFO for DMA
chunks.
2. struct mlx5e_tx_wqe_info will contain num_fifo_pkts. If a given WQE
contains only one packet, num_fifo_pkts will be zero, and the SKB will
be stored in mlx5e_tx_wqe_info, as usual. If num_fifo_pkts > 0, the SKB
pointer will be NULL, and the SKBs will be stored in the FIFO.
This change has no performance impact in TCP single stream test and
XDP_TX single stream test.
When compiled with a recent GCC, this change shows no visible
performance impact on UDP pktgen (burst 32) single stream test either:
Packet rate: 16.95 Mpps (±0.15 Mpps) -> 16.96 Mpps (±0.12 Mpps)
Instructions per packet: 429 -> 421
Cycles per packet: 160 -> 156
Instructions per cycle: 2.69 -> 2.70
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz (x86_64)
NIC: Mellanox ConnectX-6 Dx
GCC 10.2.0
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Before this patch, mlx5e_ktls_tx_handle_resync_dump_comp checked for
resync_dump_frag_page. It happened for all WQEs without an SKB,
including padding WQEs, and required a function call. Normally, padding
WQEs happen more often than TLS resyncs. Take this check out of the
function and put it to an inline function to save a call on all padding
WQEs.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
A constant for the number of DS in an empty WQE (i.e. a WQE without data
segments) is needed in multiple places (normal TX data path, MPWQE in
XDP), but currently we have a constant for XDP and an inline formula in
normal TX. This patch introduces a common constant.
Additionally, mlx5e_xdp_mpwqe_session_start is converted to use struct
assignment, because the code nearby is touched.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Use MLX5E_XDP_MPW_MAX_WQEBBS to reserve space for a MPWQE, because it's
actually the maximal size a MPWQE can take.
Reorganize the logic that checks when to close the MPWQE session:
1. Put all checks into a single function.
2. When inline is on, make only one comparison - if it's false, the less
strict one will also be false. The compiler probably optimized it out
anyway, but it's clearer to also reflect it in the code.
The MLX5E_XDP_INLINE_WQE_* defines are also changed to make the
calculations more correct from the logical point of view. Though
MLX5E_XDP_INLINE_WQE_MAX_DS_CNT used to be 16 and didn't change its
value, the calculation used to be DIV_ROUND_UP(max inline packet size,
MLX5_SEND_WQE_DS), and the numerator should have included sizeof(struct
mlx5_wqe_inline_seg).
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
A huge function mlx5e_sq_xmit was split into several to achieve multiple
goals:
1. Reuse the code in IPoIB.
2. Better intergrate with TLS, IPSEC, GENEVE and checksum offloads. Now
it's possible to reserve space in the WQ before running eseg-based
offloads, so:
2.1. It's not needed to copy cseg and eseg after mlx5e_fill_sq_frag_edge
anymore.
2.2. mlx5e_txqsq_get_next_pi will be used instead of the legacy
mlx5e_fill_sq_frag_edge for better code maintainability and reuse.
3. Prepare for the upcoming TX MPWQE for SKBs. It will intervene after
mlx5e_sq_calc_wqe_attr to check if it's possible to use MPWQE, and the
code flow will split into two paths: MPWQE and non-MPWQE.
Two high-level functions are provided to send packets:
* mlx5e_xmit is called by the networking stack, runs offloads and sends
the packet. In one of the following patches, MPWQE support will be added
to this flow.
* mlx5e_sq_xmit_simple is called by the TLS offload, runs only the
checksum offload and sends the packet.
This change has no performance impact in TCP single stream test and
XDP_TX single stream test.
When compiled with a recent GCC, this change shows no visible
performance impact on UDP pktgen (burst 32) single stream test either:
Packet rate: 16.86 Mpps (±0.15 Mpps) -> 16.95 Mpps (±0.15 Mpps)
Instructions per packet: 434 -> 429
Cycles per packet: 158 -> 160
Instructions per cycle: 2.75 -> 2.69
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz (x86_64)
NIC: Mellanox ConnectX-6 Dx
GCC 10.2.0
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Move mlx5e_tx_wqe_inline_mode from en/txrx.h to en_tx.c as it's only
used there.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Struct assignment guarantees that all fields of the structure are
initialized (those that are not mentioned are zeroed). It makes code
mode robust and reduces chances for unpredictable behavior when one
forgets to reset some field and it holds an old value from previous
iterations of using the structure.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
As preparation for the next patch, don't increase ihs to calculate
ds_cnt and then decrease it, but rather calculate the intermediate value
temporarily. This code has the same amount of arithmetic operations, but
now allows to split out ds_cnt calculation, which will be performed in
the next patch.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Vladimir Oltean says:
====================
Fix broken tc-flower rules for mscc_ocelot switches
All 3 switch drivers from the Ocelot family have the same bug in the
VCAP IS2 key offsets, which is that some keys are in the incorrect
order.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The IS2 IP4_TCP_UDP key offsets do not correspond to the VSC7514
datasheet. Whether they work or not is unknown to me. On VSC9959 and
VSC9953, with the same mistake and same discrepancy from the
documentation, tc-flower src_port and dst_port rules did not work, so I
am assuming the same is true here.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since these were copied from the Felix VCAP IS2 code, and only the
offsets were adjusted, the order of the bit fields is still wrong.
Fix it.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some of the IS2 IP4_TCP_UDP keys are not correct, like L4_DPORT,
L4_SPORT and other L4 keys. This prevents offloaded tc-flower rules from
matching on src_port and dst_port for TCP and UDP packets.
Signed-off-by: Xiaoliang Yang <xiaoliang.yang_1@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When calling the RCU brother of br_vlan_get_pvid(), lockdep warns:
=============================
WARNING: suspicious RCU usage
5.9.0-rc3-01631-g13c17acb8e38-dirty #814 Not tainted
-----------------------------
net/bridge/br_private.h:1054 suspicious rcu_dereference_protected() usage!
Call trace:
lockdep_rcu_suspicious+0xd4/0xf8
__br_vlan_get_pvid+0xc0/0x100
br_vlan_get_pvid_rcu+0x78/0x108
The warning is because br_vlan_get_pvid_rcu() calls nbp_vlan_group()
which calls rtnl_dereference() instead of rcu_dereference(). In turn,
rtnl_dereference() calls rcu_dereference_protected() which assumes
operation under an RCU write-side critical section, which obviously is
not the case here. So, when the incorrect primitive is used to access
the RCU-protected VLAN group pointer, READ_ONCE() is not used, which may
cause various unexpected problems.
I'm sad to say that br_vlan_get_pvid() and br_vlan_get_pvid_rcu() cannot
share the same implementation. So fix the bug by splitting the 2
functions, and making br_vlan_get_pvid_rcu() retrieve the VLAN groups
under proper locking annotations.
Fixes: 7582f5b70f ("bridge: add br_vlan_get_pvid_rcu()")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGhZs6bAKwk/OTgTpSD+KveBX+j4FAl9pQ8EACgkQSD+KveBX
+j7I2wf/cu9W3mC8sNeJaZKIbJ+H6KhgZsGbeLud5tFscjcf5IaCpR97hyeZPfEG
doNRtcsT9Pj5YJn458L/p+zTVeWOuaOGPMsV8pdP/8OlFzjJW/rGXnBrEUt0ehkS
Sa//xGD6V8+nW9Z34fwQqrrqJeZik3H9V/RkriZUTsJ/zR/otLF3fVOQFwrS9Ka2
/dl1ERFepjBWupY39PSMFS2S2BZ6LYY8G/ewgHKeexbqLykxU27P3+mFz46YPmP6
jdIMmvo+fuPqyu9Tjtg6pGjYpCWttnBBtDmeSg+ewf61qW4mSemJzfGcbZYY2XT6
CxRsm4aTJ5COTEx05JFOqIhpP5LuAA==
=Hcsv
-----END PGP SIGNATURE-----
Merge tag 'mlx5-fixes-2020-09-18' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5 fixes-2020-09-18
This series introduces some fixes to mlx5 driver.
Please pull and let me know if there is any problem.
v1->v2:
Remove missing patch from -stable list.
For -stable v5.1
('net/mlx5: Fix FTE cleanup')
For -stable v5.3
('net/mlx5e: TLS, Do not expose FPGA TLS counter if not supported')
('net/mlx5e: Enable adding peer miss rules only if merged eswitch is supported')
For -stable v5.7
('net/mlx5e: Fix memory leak of tunnel info when rule under multipath not ready')
For -stable v5.8
('net/mlx5e: Use RCU to protect rq->xdp_prog')
('net/mlx5e: Fix endianness when calculating pedit mask first bit')
('net/mlx5e: Use synchronize_rcu to sync with NAPI')
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Update maintainers for MediaTek switch driver with Landen Chao who is
familiar with MediaTek MT753x switch devices and will help maintenance
from the vendor side.
Cc: Steven Liu <steven.liu@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Landen Chao <Landen.Chao@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The spinlock only needed when accessing the channel's icosq, grab the lock
after the buf allocation in resync_post_get_progress_params() to avoid
kzalloc(GFP_KERNEL) in atomic context.
Fixes: 0419d8c9d8 ("net/mlx5e: kTLS, Add kTLS RX resync support")
Reported-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Using synchronize_rcu() is sufficient to wait until running NAPI quits.
See similar upstream fix with detailed explanation:
("net/mlx5e: Use synchronize_rcu to sync with NAPI")
This change also fixes a possible use-after-free as the NAPI
might be already released at this stage.
Fixes: 0419d8c9d8 ("net/mlx5e: kTLS, Add kTLS RX resync support")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The set of TLS TX global SW counters in mlx5e_tls_sw_stats_desc
is updated from all rings by using atomic ops.
This set of stats is used only in the FPGA TLS use case, not in
the Connect-X TLS one, where regular per-ring counters are used.
Do not expose them in the Connect-X use case, as this would cause
counter duplication. For example, tx_tls_drop_no_sync_data would
appear twice in the ethtool stats.
Fixes: d2ead1f360 ("net/mlx5e: Add kTLS TX HW offload support")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
The cited commit started to reuse function mlx5e_update_ndo_stats() for
the representors as well.
However, the function is hard-coded to work on mlx5e_nic_stats_grps only.
Due to this issue, the representors statistics were not updated in the
output of "ip -s".
Fix it to work with the correct group by extracting it from the caller's
profile.
Also, while at it and since this function became generic, move it to
en_stats.c and rename it accordingly.
Fixes: 8a236b1514 ("net/mlx5e: Convert rep stats to mlx5e_stats_grp-based infra")
Signed-off-by: Alaa Hleihel <alaa@nvidia.com>
Reviewed-by: Vlad Buslov <vladbu@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Currently the FW does not generate events for counters other than error
counters. Unlike ".get_ethtool_stats", ".ndo_get_stats64" (which ip -s
uses) might run in atomic context, while the FW interface is non atomic.
Thus, 'ip' is not allowed to issue FW commands, so it will only display
cached counters in the driver.
Add a SW counter (mcast_packets) in the driver to count rx multicast
packets. The counter also counts broadcast packets, as we consider it a
special case of multicast.
Use the counter value when calling "ip -s"/"ifconfig".
Fixes: f62b8bb8f2 ("net/mlx5: Extend mlx5_core to support ConnectX-4 Ethernet functionality")
Signed-off-by: Ron Diskin <rondi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The field mask value is provided in network byte order and has to
be converted to host byte order before calculating pedit mask
first bit.
Fixes: 88f30bbcba ("net/mlx5e: Bit sized fields rewrite support")
Signed-off-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
The cited commit creates peer miss group during switchdev mode
initialization in order to handle miss packets correctly while in VF
LAG mode. This is done regardless of FW support of such groups which
could cause rules setups failure later on.
Fix by adding FW capability check before creating peer groups/rule.
Fixes: ac004b8321 ("net/mlx5e: E-Switch, Add peer miss rules")
Signed-off-by: Maor Dickman <maord@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Raed Salem <raeds@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Add missing mapping remove call when removing ct rule,
as the mapping was allocated when ct rule was adding with ct_label.
Also there is a missing mapping remove call in error flow.
Fixes: 54b154ecfb ("net/mlx5e: CT: Map 128 bits labels to 32 bit map ID")
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
When deleting vxlan flow rule under multipath, tun_info in parse_attr is
not freed when the rule is not ready.
Fixes: ef06c9ee89 ("net/mlx5e: Allow one failure when offloading tc encap rules under multipath")
Signed-off-by: Jianbo Liu <jianbol@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
As described in the previous commit, napi_synchronize doesn't quite fit
the purpose when we just need to wait until the currently running NAPI
quits. Its implementation waits until NAPI is not running by polling and
waiting for 1ms in between. In cases where we need to deactivate one
queue (e.g., recovery flows) or where we deactivate them one-by-one
(deactivate channel flow), we may get stuck in napi_synchronize forever
if other queues keep NAPI active, causing a soft lockup. Depending on
kernel configuration (CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC), it may result
in a kernel panic.
To fix the issue, use synchronize_rcu to wait for NAPI to quit, and wrap
the whole NAPI in rcu_read_lock.
Fixes: acc6c5953a ("net/mlx5e: Split open/close channels to stages")
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Currently, the RQs are temporarily deactivated while hot-replacing the
XDP program, and napi_synchronize is used to make sure rq->xdp_prog is
not in use. However, napi_synchronize is not ideal: instead of waiting
till the end of a NAPI cycle, it polls and waits until NAPI is not
running, sleeping for 1ms between the periodic checks. Under heavy
workloads, this loop will never end, which may even lead to a kernel
panic if the kernel detects the hangup. Such workloads include XSK TX
and possibly also heavy RX (XSK or normal).
The fix is inspired by commit 326fe02d1e ("net/mlx4_en: protect
ring->xdp_prog with rcu_read_lock"). As mlx5e_xdp_handle is already
protected by rcu_read_lock, and bpf_prog_put uses call_rcu to free the
program, there is no need for additional synchronization if proper RCU
functions are used to access the pointer. This patch converts all
accesses to rq->xdp_prog to use RCU functions.
Fixes: 86994156c7 ("net/mlx5e: XDP fast RX drop bpf programs support")
Fixes: db05815b36 ("net/mlx5e: Add XSK zero-copy support")
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Currently, when an FTE is allocated, its refcount is decreased to 0
with the purpose it will not be a stand alone steering object and every
rule (destination) of the FTE would increase the refcount.
When mlx5_cleanup_fs is called while not all rules were deleted by the
steering users, it hit refcount underflow on the FTE once clean_tree
calls to tree_remove_node after the deleted rules already decreased
the refcount to 0.
FTE is no longer destroyed implicitly when the last rule (destination)
is deleted. mlx5_del_flow_rules avoids it by increasing the refcount on
the FTE and destroy it explicitly after all rules were deleted. So we
can avoid the refcount underflow by making FTE as stand alone object.
In addition need to set del_hw_func to FTE so the HW object will be
destroyed when the FTE is deleted from the cleanup_tree flow.
refcount_t: underflow; use-after-free.
WARNING: CPU: 2 PID: 15715 at lib/refcount.c:28 refcount_warn_saturate+0xd9/0xe0
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
Call Trace:
tree_put_node+0xf2/0x140 [mlx5_core]
clean_tree+0x4e/0xf0 [mlx5_core]
clean_tree+0x4e/0xf0 [mlx5_core]
clean_tree+0x4e/0xf0 [mlx5_core]
clean_tree+0x5f/0xf0 [mlx5_core]
clean_tree+0x4e/0xf0 [mlx5_core]
clean_tree+0x5f/0xf0 [mlx5_core]
mlx5_cleanup_fs+0x26/0x270 [mlx5_core]
mlx5_unload+0x2e/0xa0 [mlx5_core]
mlx5_unload_one+0x51/0x120 [mlx5_core]
mlx5_devlink_reload_down+0x51/0x90 [mlx5_core]
devlink_reload+0x39/0x120
? devlink_nl_cmd_reload+0x43/0x220
genl_rcv_msg+0x1e4/0x420
? genl_family_rcv_msg_attrs_parse+0x100/0x100
netlink_rcv_skb+0x47/0x110
genl_rcv+0x24/0x40
netlink_unicast+0x217/0x2f0
netlink_sendmsg+0x30f/0x430
sock_sendmsg+0x30/0x40
__sys_sendto+0x10e/0x140
? handle_mm_fault+0xc4/0x1f0
? do_page_fault+0x33f/0x630
__x64_sys_sendto+0x24/0x30
do_syscall_64+0x48/0x130
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Fixes: 718ce4d601 ("net/mlx5: Consolidate update FTE for all removal changes")
Fixes: bd71b08ec2 ("net/mlx5: Support multiple updates of steering rules in parallel")
Signed-off-by: Maor Gottlieb <maorg@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
BCM72113 features a 28nm integrated EPHY, add an entry to the driver for
it.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some kernels builds might inline vfs_getattr call within fstat
syscall code path, so fentry/vfs_getattr trampoline is not called.
Add security_inode_getattr to allowlist and switch the d_path test stat
trampoline to security_inode_getattr.
Keeping dentry_open and filp_close, because they are in their own
files, so unlikely to be inlined, but in case they are, adding
security_file_open.
Adding flags that indicate trampolines were called and failing
the test if any of them got missed, so it's easier to identify
the issue next time.
Fixes: e4d1af4b16 ("selftests/bpf: Add test for d_path helper")
Suggested-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200918112338.2618444-1-jolsa@kernel.org
If a bucket contains a lot of sockets, during bpf_iter traversing
a bucket, concurrent userspace bpf_map_update_elem() and
bpf program bpf_sk_storage_{get,delete}() may experience
some undesirable delays as they will compete with bpf_iter
for bucket lock.
Note that the number of buckets for bpf_sk_storage_map
is roughly the same as the number of cpus. So if there
are lots of sockets in the system, each bucket could
contain lots of sockets.
Different actual use cases may experience different delays.
Here, using selftest bpf_iter subtest bpf_sk_storage_map,
I hacked the kernel with ktime_get_mono_fast_ns()
to collect the time when a bucket was locked
during bpf_iter prog traversing that bucket. This way,
the maximum incurred delay was measured w.r.t. the
number of elements in a bucket.
# elems in each bucket delay(ns)
64 17000
256 72512
2048 875246
The potential delays will be further increased if
we have even more elemnts in a bucket. Using rcu_read_lock()
is a reasonable compromise here. It may lose some precision, e.g.,
access stale sockets, but it will not hurt performance of
bpf program or user space application which also tries
to get/delete or update map elements.
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200916224645.720172-1-yhs@fb.com
Lorenz Bauer says:
====================
Changes in v4:
- Output the desired type on BTF ID mismatch (Martin)
Changes in v3:
- Fix BTF_ID_LIST_SINGLE if BTF is disabled (Martin)
- Drop incorrect arg_btf_id in bpf_sk_storage.c (Martin)
- Check for arg_btf_id in check_func_proto (Martin)
- Drop incorrect PTR_TO_BTF_ID from fullsock_types (Martin)
- Introduce btf_seq_file_ids in bpf_trace.c to reduce duplication
Changes in v2:
- Make the series stand alone (Martin)
- Drop incorrect BTF_SET_START fix (Andrii)
- Only support a single BTF ID per argument (Martin)
- Introduce BTF_ID_LIST_SINGLE macro (Andrii)
- Skip check_ctx_reg iff register is NULL
- Change output of check_reg_type slightly, to avoid touching tests
Original cover letter:
Currently, check_func_arg has this pretty gnarly if statement that
compares the valid arg_type with the actualy reg_type. Sprinkled
in-between are checks for register_is_null, to short circuit these
tests if we're dealing with a nullable arg_type. There is also some
code for later bounds / access checking hidden away in there.
This series of patches refactors the function into something like this:
if (reg_is_null && arg_type_is_nullable)
skip type checking
do type checking, including BTF validation
do bounds / access checking
The type checking is now table driven, which makes it easy to extend
the acceptable types. Maybe more importantly, using a table makes it
easy to provide more helpful verifier output (see the last patch).
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The mapping between bpf_arg_type and bpf_reg_type is encoded in a big
hairy if statement that is hard to follow. The debug output also leaves
to be desired: if a reg_type doesn't match we only print one of the
options, instead printing all the valid ones.
Convert the if statement into a table which is then used to drive type
checking. If none of the reg_types match we print all options, e.g.:
R2 type=rdonly_buf expected=fp, pkt, pkt_meta, map_value
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-12-lmb@cloudflare.com
check_func_arg has a plethora of weird if statements with empty branches.
They work around the fact that *_OR_NULL argument types should accept a
SCALAR_VALUE register, as long as it's value is 0. These statements make
it difficult to reason about the type checking logic.
Instead, skip more detailed type checking logic iff the register is 0,
and the function expects a nullable type. This allows simplifying the type
checking itself.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-11-lmb@cloudflare.com
Move the check for PTR_TO_MAP_VALUE to check_func_arg, where all other
checking is done as well. Move the invocation of process_spin_lock away
from the register type checking, to allow a future refactoring.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-10-lmb@cloudflare.com
If we encounter a pointer to memory, we set meta->raw_mode depending
on the type of memory we point at. What isn't obvious is that this
information is only used when the next memory size argument is
encountered.
Move the assignment closer to where it's used, and add a comment that
explains what is going on.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-9-lmb@cloudflare.com
Always check context access if the register we're operating on is
PTR_TO_CTX, rather than relying on ARG_PTR_TO_CTX. This allows
simplifying the arg_type checking section of the function.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-8-lmb@cloudflare.com
Instead of dealing with reg->ref_obj_id individually for every arg type that
needs it, rely on the fact that ref_obj_id is zero if the register is not
reference tracked.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-7-lmb@cloudflare.com
Perform BTF type checks if the register we're working on contains a BTF
pointer, rather than if the argument is for a BTF pointer. This is easier
to understand, and allows removing the code from the arg_type checking
section of the function.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-6-lmb@cloudflare.com
Function prototypes using ARG_PTR_TO_BTF_ID currently use two ways to signal
which BTF IDs are acceptable. First, bpf_func_proto.btf_id is an array of
IDs, one for each argument. This array is only accessed up to the highest
numbered argument that uses ARG_PTR_TO_BTF_ID and may therefore be less than
five arguments long. It usually points at a BTF_ID_LIST. Second, check_btf_id
is a function pointer that is called by the verifier if present. It gets the
actual BTF ID of the register, and the argument number we're currently checking.
It turns out that the only user check_arg_btf_id ignores the argument, and is
simply used to check whether the BTF ID has a struct sock_common at it's start.
Replace both of these mechanisms with an explicit BTF ID for each argument
in a function proto. Thanks to btf_struct_ids_match this is very flexible:
check_arg_btf_id can be replaced by requiring struct sock_common.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-5-lmb@cloudflare.com
Add a convenience macro that allows defining a BTF ID list with
a single item. This lets us cut down on repetitive macros.
Suggested-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-4-lmb@cloudflare.com
Move the check for a NULL or zero register to check_helper_mem_access. This
makes check_stack_boundary easier to understand.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-3-lmb@cloudflare.com
bsearch doesn't modify the contents of the array, so we can take a const pointer.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200921121227.255763-2-lmb@cloudflare.com
Maxim Mikityanskiy