| Age | Commit message (Collapse) | Author |
|---|
|
Remove AF family obfuscation by creating symmetric structs for RX and
TX IPsec flow steering chains. This simplifies to us low level IPsec
FS creation logic without need to dig into multiple levels of structs.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
Instead of performing redefinition of XFRM core defines to same
values but with MLX5_* prefix, cache the input values as is by making
sure that the proper storage objects are used.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
As a preparation for future extension of IPsec hardware object to allow
configuration of packet offload mode, extend the XFRM validator to check
replay window values.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
Add needed capabilities check to determine if device supports IPsec
packet offload mode.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
Add all needed bits to support IPsec packet offload mode.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
There is no need in hiding returned ASO WQE type by providing void*,
use the real type instead. Do it together with zeroing that memory,
so ASO WQE will be ready to use immediately.
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
Currently lan966x, doesn't allow to run PTP over interfaces that are
part of the bridge. The reason is when the lan966x was receiving a
PTP frame (regardless if L2/IPv4/IPv6) the HW it would flood this
frame.
Now that it is possible to add VCAP rules to the HW, such to trap these
frames to the CPU, it is possible to run PTP also over interfaces that
are part of the bridge.
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Add the function vcap_rule_get_key_u32 which allows to get the value and
the mask of a key that exist on the rule. If the key doesn't exist,
it would return error.
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Add the function vcap_mod_rule which allows to update an existing rule
in the vcap. It is required for the rule to exist in the vcap to be able
to modify it.
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Add function vcap_get_rule which returns a rule based on the internal
rule id.
The entire functionality of reading and decoding the rule from the VCAP
was inside vcap_api_debugfs file. So move the entire implementation in
vcap_api as this is used also by vcap_get_rule.
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The skb is delivered to napi_gro_receive() which may free it, after
calling this, dereferencing skb may trigger use-after-free.
Fixes: 57c5bc9ad7d7 ("net: hisilicon: add hix5hd2 mac driver")
Signed-off-by: Liu Jian <liujian56@huawei.com>
Link: https://lore.kernel.org/r/20221203094240.1240211-2-liujian56@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The skb is delivered to napi_gro_receive() which may free it, after
calling this, dereferencing skb may trigger use-after-free.
Fixes: 542ae60af24f ("net: hisilicon: Add Fast Ethernet MAC driver")
Signed-off-by: Liu Jian <liujian56@huawei.com>
Link: https://lore.kernel.org/r/20221203094240.1240211-1-liujian56@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The nicvf_probe() won't destroy workqueue when register_netdev()
failed. Add destroy_workqueue err handle case to fix this issue.
Fixes: 2ecbe4f4a027 ("net: thunderx: replace global nicvf_rx_mode_wq work queue for all VFs to private for each of them.")
Signed-off-by: Yongqiang Liu <liuyongqiang13@huawei.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Link: https://lore.kernel.org/r/20221203094125.602812-1-liuyongqiang13@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The skb is delivered to napi_gro_receive() which may free it, after calling this,
dereferencing skb may trigger use-after-free.
Fixes: 1c59eb678cbd ("ravb: Fillup ravb_rx_gbeth() stub")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20221203092941.10880-1-yuehaibing@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The mchp_sparx5_probe() won't destroy workqueue created by
create_singlethread_workqueue() in sparx5_start() when later
inits failed. Add destroy_workqueue in the cleanup_ports case,
also add it in mchp_sparx5_remove()
Fixes: b37a1bae742f ("net: sparx5: add mactable support")
Signed-off-by: Qiheng Lin <linqiheng@huawei.com>
Link: https://lore.kernel.org/r/20221203070259.19560-1-linqiheng@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
After calling napi_complete_done(), the NAPIF_STATE_SCHED bit may be
cleared, and another CPU can start napi thread and access per-CQ variable,
cq->work_done. If the other thread (for example, from busy_poll) sets
it to a value >= budget, this thread will continue to run when it should
stop, and cause memory corruption and panic.
To fix this issue, save the per-CQ work_done variable in a local variable
before napi_complete_done(), so it won't be corrupted by a possible
concurrent thread after napi_complete_done().
Also, add a flag bit to advertise to the NIC firmware: the NAPI work_done
variable race is fixed, so the driver is able to reliably support features
like busy_poll.
Cc: stable@vger.kernel.org
Fixes: e1b5683ff62e ("net: mana: Move NAPI from EQ to CQ")
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Link: https://lore.kernel.org/r/1670010190-28595-1-git-send-email-haiyangz@microsoft.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In dt-binding snps,dwmac.yaml, some properties under "snps,axi-config"
node are named without "axi_" prefix, but the driver expects the
prefix. Since the dt-binding has been there for a long time, we'd
better make driver match the binding for compatibility.
Fixes: afea03656add ("stmmac: rework DMA bus setting and introduce new platform AXI structure")
Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Link: https://lore.kernel.org/r/20221202161739.2203-1-jszhang@kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Since Wireless Ethernet Dispatcher is now available for mt7986 in mt76,
enable hw flow support for MT7986 SoC.
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/fdcaacd827938e6a8c4aa1ac2c13e46d2c08c821.1670072898.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Return -EOPNOTSUPP, when user requests l4_4_bytes for raw IP4 or
IP6 flow director filters. Flow director does not support filtering
on l4 bytes for PCTYPEs used by IP4 and IP6 filters.
Without this patch, user could create filters with l4_4_bytes fields,
which did not do any filtering on L4, but only on L3 fields.
Fixes: 36777d9fa24c ("i40e: check current configured input set when adding ntuple filters")
Signed-off-by: Przemyslaw Patynowski <przemyslawx.patynowski@intel.com>
Signed-off-by: Kamil Maziarz <kamil.maziarz@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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 spawning max VFs on a PF, some VFs were not getting resources and
their MAC addresses were 0. This was caused by PF sleeping before flushing
HW registers which caused VIRTCHNL_VFR_VFACTIVE to not be set in time for
VF.
Fix by adding a sleep after hw flush.
Fixes: e4b433f4a741 ("i40e: reset all VFs in parallel when rebuilding PF")
Signed-off-by: Sylwester Dziedziuch <sylwesterx.dziedziuch@intel.com>
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
|
|
During tx rings configuration default XPS queue config is set and
__I40E_TX_XPS_INIT_DONE is locked. __I40E_TX_XPS_INIT_DONE state is
cleared and set again with default mapping only during queues build,
it means after first setup or reset with queues rebuild. (i.e.
ethtool -L <interface> combined <number>) After other resets (i.e.
ethtool -t <interface>) XPS_INIT_DONE is not cleared and those default
maps cannot be set again. It results in cleared xps_cpus mapping
until queues are not rebuild or mapping is not set by user.
Add clearing __I40E_TX_XPS_INIT_DONE state during reset to let
the driver set xps_cpus to defaults again after it was cleared.
Fixes: 6f853d4f8e93 ("i40e: allow XPS with QoS enabled")
Signed-off-by: Michal Jaron <michalx.jaron@intel.com>
Signed-off-by: Kamil Maziarz <kamil.maziarz@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
|
|
The pp->indir[0] value comes from the user. It is passed to:
if (cpu_online(pp->rxq_def))
inside the mvneta_percpu_elect() function. It needs bounds checkeding
to ensure that it is not beyond the end of the cpu bitmap.
Fixes: cad5d847a093 ("net: mvneta: Fix the CPU choice in mvneta_percpu_elect")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Rewrite nfp_net_set_rx_mode() to implement interface to delivery
mc address and operations to firmware by using general mailbox
for filtering multicast packets.
The operations include add mc address and delete mc address.
And the limitation of mc addresses number is 1024 for each net
device.
User triggers adding mc address by using command below:
ip maddress add <mc address> dev <interface name>
User triggers deleting mc address by using command below:
ip maddress del <mc address> dev <interface name>
Signed-off-by: Diana Wang <na.wang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
consume_skb on transmitted, kfree_skb on dropped, do not free on
TX_BUSY.
Previously the xmit function could return -EBUSY without freeing, which
supposedly is interpreted as a drop. And was using kfree on successfully
transmitted packets.
sparx5_fdma_xmit and sparx5_inject returns error code, where -EBUSY
indicates TX_BUSY and any other error code indicates dropped.
Fixes: f3cad2611a77 ("net: sparx5: add hostmode with phylink support")
Signed-off-by: Casper Andersson <casper.casan@gmail.com>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In otx2_init_tc(), if rhashtable_init() failed, it does not free
tc->tc_entries_bitmap which is allocated in otx2_tc_alloc_ent_bitmap().
Fixes: 2e2a8126ffac ("octeontx2-pf: Unify flow management variables")
Signed-off-by: Ziyang Xuan <william.xuanziyang@huawei.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
A loop for reading MISTAT register continues while regmap_read() fails
and (mistat & BUSY), but if regmap_read() fails a value of mistat is
undefined.
The patch proposes to check for BUSY flag only when regmap_read()
succeed. Compile test only.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: d70e53262f5c ("net: Microchip encx24j600 driver")
Signed-off-by: Valentina Goncharenko <goncharenko.vp@ispras.ru>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In functions regmap_encx24j600_phy_reg_read() and
regmap_encx24j600_phy_reg_write() in the conditions of the waiting
cycles for filling the variable 'ret' it is necessary to add parentheses
to prevent wrong assignment due to logical operations precedence.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: d70e53262f5c ("net: Microchip encx24j600 driver")
Signed-off-by: Valentina Goncharenko <goncharenko.vp@ispras.ru>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Add support for the Multi-Gigabit Ethernet (MGBE/XPCS) IP found on
NVIDIA Tegra234 SoCs.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Bhadram Varka <vbhadram@nvidia.com>
Co-developed-by: Revanth Kumar Uppala <ruppala@nvidia.com>
Signed-off-by: Revanth Kumar Uppala <ruppala@nvidia.com>
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The Tegra MGBE ethernet controller requires that the SERDES link is
powered-up after the PHY link is up, otherwise the link fails to
become ready following a resume from suspend. Add a variable to indicate
that the SERDES link must be powered-up after the PHY link.
Signed-off-by: Revanth Kumar Uppala <ruppala@nvidia.com>
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Allow users to configure xfrm states with packet offload mode.
The packet mode must be requested both for policy and state, and
such requires us to do not implement fallback.
We explicitly return an error if requested packet mode can't
be configured.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
When there's only one buffer to dma and its length is 4096, then
only one data descriptor is needed to carry it according to current
descriptor definition. So the descriptor type should be `simple`
instead of `gather`, the latter requires more than one descriptor,
otherwise it'll be dropped by application firmware.
Fixes: c10d12e3dce8 ("nfp: add support for NFDK data path")
Fixes: d9d950490a0a ("nfp: nfdk: implement xdp tx path for NFDK")
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Reviewed-by: Richard Donkin <richard.donkin@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20221202134646.311108-1-simon.horman@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
There are reports about r8169 not reaching full line speed on certain
systems (e.g. SBC's) with a 2.5Gbps link.
There was a time when hardware interrupt coalescing was enabled per
default, but this was changed due to ASPM-related issues on few systems.
So let's use software interrupt coalescing instead and enable it
using new function netdev_sw_irq_coalesce_default_on().
Even with these conservative settings interrupt load on my 1Gbps test
system reduced significantly.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In order to fix the following sleep while atomic bug always alloc pages
with GFP_ATOMIC in mtk_wed_wo_queue_refill since page_frag_alloc runs in
spin_lock critical section.
[ 9.049719] Hardware name: MediaTek MT7986a RFB (DT)
[ 9.054665] Call trace:
[ 9.057096] dump_backtrace+0x0/0x154
[ 9.060751] show_stack+0x14/0x1c
[ 9.064052] dump_stack_lvl+0x64/0x7c
[ 9.067702] dump_stack+0x14/0x2c
[ 9.071001] ___might_sleep+0xec/0x120
[ 9.074736] __might_sleep+0x4c/0x9c
[ 9.078296] __alloc_pages+0x184/0x2e4
[ 9.082030] page_frag_alloc_align+0x98/0x1ac
[ 9.086369] mtk_wed_wo_queue_refill+0x134/0x234
[ 9.090974] mtk_wed_wo_init+0x174/0x2c0
[ 9.094881] mtk_wed_attach+0x7c8/0x7e0
[ 9.098701] mt7915_mmio_wed_init+0x1f0/0x3a0 [mt7915e]
[ 9.103940] mt7915_pci_probe+0xec/0x3bc [mt7915e]
[ 9.108727] pci_device_probe+0xac/0x13c
[ 9.112638] really_probe.part.0+0x98/0x2f4
[ 9.116807] __driver_probe_device+0x94/0x13c
[ 9.121147] driver_probe_device+0x40/0x114
[ 9.125314] __driver_attach+0x7c/0x180
[ 9.129133] bus_for_each_dev+0x5c/0x90
[ 9.132953] driver_attach+0x20/0x2c
[ 9.136513] bus_add_driver+0x104/0x1fc
[ 9.140333] driver_register+0x74/0x120
[ 9.144153] __pci_register_driver+0x40/0x50
[ 9.148407] mt7915_init+0x5c/0x1000 [mt7915e]
[ 9.152848] do_one_initcall+0x40/0x25c
[ 9.156669] do_init_module+0x44/0x230
[ 9.160403] load_module+0x1f30/0x2750
[ 9.164135] __do_sys_init_module+0x150/0x200
[ 9.168475] __arm64_sys_init_module+0x18/0x20
[ 9.172901] invoke_syscall.constprop.0+0x4c/0xe0
[ 9.177589] do_el0_svc+0x48/0xe0
[ 9.180889] el0_svc+0x14/0x50
[ 9.183929] el0t_64_sync_handler+0x9c/0x120
[ 9.188183] el0t_64_sync+0x158/0x15c
Fixes: 799684448e3e ("net: ethernet: mtk_wed: introduce wed wo support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Link: https://lore.kernel.org/r/67ca94bdd3d9eaeb86e52b3050fbca0bcf7bb02f.1669908312.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Refill RX queue in batches of descriptors to improve performance. Refill
is allowed to fail as long as a minimum number of descriptors is active.
Thus, a limited number of failed RX buffer allocations is now allowed
for normal operation. Previously every failed allocation resulted in a
dropped frame.
If the minimum number of active descriptors is reached, then RX buffers
are still reused and frames are dropped. This ensures that the RX queue
never runs empty and always continues to operate.
Prework for future XDP support.
Signed-off-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Without interrupt throttling, iperf server mode generates a CPU load of
100% (A53 1.2GHz). Also the throughput suffers with less than 900Mbit/s
on a 1Gbit/s link. The reason is a high interrupt load with interrupts
every ~20us.
Reduce interrupt load by throttling of interrupts. Interrupt delay
default is 64us. For iperf server mode the CPU load is significantly
reduced to ~20% and the throughput reaches the maximum of 941MBit/s.
Interrupts are generated every ~140us.
RX and TX coalesce can be configured with ethtool. RX coalesce has
priority over TX coalesce if the same interrupt is used.
Signed-off-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Allow user space to read number of TX and RX queue. This is useful for
device dependent qdisc configurations like TAPRIO with hardware offload.
Also ethtool::get_per_queue_coalesce / set_per_queue_coalesce requires
that interface.
Signed-off-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2022-11-30 (e1000e, igb)
This series contains updates to e1000e and igb drivers.
Akihiko Odaki fixes calculation for checking whether space for next
frame exists for e1000e and properly sets MSI-X vector to fix failing
ethtool interrupt test for igb.
* '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
igb: Allocate MSI-X vector when testing
e1000e: Fix TX dispatch condition
====================
Link: https://lore.kernel.org/r/20221130194228.3257787-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
I must have missed that these stats are only exposed
via the unstructured ethtool -S when they got merged.
Plumb in the structured form.
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20221130013108.90062-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The open code is defined as a helper function(tp_to_dev) on r8169_main.c,
which the open code is &tp->pci_dev->dev. The helper function was added
in commit 1e1205b7d3e9 ("r8169: add helper tp_to_dev"). And then later,
commit f1e911d5d0df ("r8169: add basic phylib support") added
r8169_phylink_handler function but it didn't use the helper function.
Thus, tp_to_dev() replaces the open code. This patch doesn't change logic.
Signed-off-by: Juhee Kang <claudiajkang@gmail.com>
Reviewed-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/20221129161244.5356-1-claudiajkang@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
After the introduction of a private mac_lock that serializes access to
priv->mac (and port_priv->mac in the switch), the only remaining purpose
of rtnl_lock() is to satisfy the locking requirements of
phylink_fwnode_phy_connect() and phylink_disconnect_phy().
But the functions these live in, dpaa2_mac_connect() and
dpaa2_mac_disconnect(), have contradictory locking requirements.
While phylink_fwnode_phy_connect() wants rtnl_lock() to be held,
phylink_create() wants it to not be held.
Move the rtnl_lock() from top-level (in the dpaa2-eth and dpaa2-switch
drivers) to only surround the phylink calls that require it, in the
dpaa2-mac library code.
This is possible because dpaa2_mac_connect() and dpaa2_mac_disconnect()
run unlocked, and there isn't any danger of an AB/BA deadlock between
the rtnl_mutex and other private locks.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The dpaa2-switch driver uses a DPMAC in the same way as the dpaa2-eth
driver, so we need to duplicate the locking solution established by the
previous change to the switch driver as well.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The dpaa2 architecture permits dynamic connections between objects on
the fsl-mc bus, specifically between a DPNI object (represented by a
struct net_device) and a DPMAC object (represented by a struct phylink).
The DPNI driver is notified when those connections are created/broken
through the dpni_irq0_handler_thread() method. To ensure that ethtool
operations, as well as netdev up/down operations serialize with the
connection/disconnection of the DPNI with a DPMAC,
dpni_irq0_handler_thread() takes the rtnl_lock() to block those other
operations from taking place.
There is code called by dpaa2_mac_connect() which wants to acquire the
rtnl_mutex once again, see phylink_create() -> phylink_register_sfp() ->
sfp_bus_add_upstream() -> rtnl_lock(). So the strategy doesn't quite
work out, even though it's fairly simple.
Create a different strategy, where all code paths in the dpaa2-eth
driver access priv->mac only while they are holding priv->mac_lock.
The phylink instance is not created or connected to the PHY under the
priv->mac_lock, but only assigned to priv->mac then. This will eliminate
the reliance on the rtnl_mutex.
Add lockdep annotations and put comments where holding the lock is not
necessary, and priv->mac can be dereferenced freely.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
dpaa2_eth_connect_mac() is called both from dpaa2_eth_probe() and from
dpni_irq0_handler_thread().
It could happen that the DPNI gets connected to a DPMAC on the fsl-mc
bus exactly during probe, as soon as the "endpoint change" interrupt is
requested in dpaa2_eth_setup_irqs(). This will cause the
dpni_irq0_handler_thread() to register a phylink instance for that DPMAC.
Then, the probing function will also try to register a phylink instance
for the same DPMAC, operation which should fail (and this will fail the
probing of the driver).
Reorder dpaa2_eth_setup_irqs() and dpaa2_eth_connect_mac(), such that
dpni_irq0_handler_thread() never races with the DPMAC-related portion of
the probing path.
Also reorder dpaa2_eth_disconnect_mac() to be in the mirror position of
dpaa2_eth_connect_mac() in the teardown path.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The helper function will gain a lockdep annotation in a future patch.
Make sure to benefit from it.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
DPNIs and DPSW objects can connect and disconnect at runtime from DPMAC
objects on the same fsl-mc bus. The DPMAC object also holds "ethtool -S"
unstructured counters. Those counters are only shown for the entity
owning the netdev (DPNI, DPSW) if it's connected to a DPMAC.
The ethtool stringset code path is split into multiple callbacks, but
currently, connecting and disconnecting the DPMAC takes the rtnl_lock().
This blocks the entire ethtool code path from running, see
ethnl_default_doit() -> rtnl_lock() -> ops->prepare_data() ->
strset_prepare_data().
This is going to be a problem if we are going to no longer require
rtnl_lock() when connecting/disconnecting the DPMAC, because the DPMAC
could appear between ops->get_sset_count() and ops->get_strings().
If it appears out of the blue, we will provide a stringset into an array
that was dimensioned thinking the DPMAC wouldn't be there => array
accessed out of bounds.
There isn't really a good way to work around that, and I don't want to
put too much pressure on the ethtool framework by playing locking games.
Just make the DPMAC counters be always available. They'll be zeroes if
the DPNI or DPSW isn't connected to a DPMAC.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The dpaa2-switch has the exact same locking requirements when connected
to a DPMAC, so it needs port_priv->mac to always point either to NULL,
or to a DPMAC with a fully initialized phylink instance.
Make the same preparatory change in the dpaa2-switch driver as in the
dpaa2-eth one.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
There are 2 requirements for correct code:
- Any time the driver accesses the priv->mac pointer at runtime, it
either holds NULL to indicate a DPNI-DPNI connection (or unconnected
DPNI), or a struct dpaa2_mac whose phylink instance was fully
initialized (created and connected to the PHY). No changes are made to
priv->mac while it is being used. Currently, rtnl_lock() watches over
the call to dpaa2_eth_connect_mac(), so it serves the purpose of
serializing this with all readers of priv->mac.
- dpaa2_mac_connect() should run unlocked, because inside it are 2
phylink calls with incompatible locking requirements: phylink_create()
requires that the rtnl_mutex isn't held, and phylink_fwnode_phy_connect()
requires that the rtnl_mutex is held. The only way to solve those
contradictory requirements is to let dpaa2_mac_connect() take
rtnl_lock() when it needs to.
To solve both requirements, we need to identify the writer side of the
priv->mac pointer, which can be wrapped in a mutex private to the driver
in a future patch. The dpaa2_mac_connect() cannot be part of the writer
side critical section, because of an AB/BA deadlock with rtnl_lock().
So the strategy needs to be that where we prepare the DPMAC by calling
dpaa2_mac_connect(), and only make priv->mac point to it once it's fully
prepared. This ensures that the writer side critical section has the
absolute minimum surface it can.
The reverse strategy is adopted in the dpaa2_eth_disconnect_mac() code
path. This makes sure that priv->mac is NULL when we start tearing down
the DPMAC that we disconnected from, and concurrent code will simply not
see it.
No locking changes in this patch (concurrent code is still blocked by
the rtnl_mutex).
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
dpaa2_mac_disconnect() will only be called with a NULL mac->phylink if
dpaa2_mac_connect() failed, or was never called.
The callers are these:
dpaa2_eth_disconnect_mac():
if (dpaa2_eth_is_type_phy(priv))
dpaa2_mac_disconnect(priv->mac);
dpaa2_switch_port_disconnect_mac():
if (dpaa2_switch_port_is_type_phy(port_priv))
dpaa2_mac_disconnect(port_priv->mac);
priv->mac can be NULL, but in that case, dpaa2_eth_is_type_phy() returns
false, and dpaa2_mac_disconnect() is never called. Similar for
dpaa2-switch.
When priv->mac is non-NULL, it means that dpaa2_mac_connect() returned
zero (success), and therefore, priv->mac->phylink is also a valid
pointer.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The phylink handling is intended to be hidden inside the dpaa2_mac
object. Move the phylink_start() call into dpaa2_mac_start(), and
phylink_stop() into dpaa2_mac_stop().
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
