- 20 May, 2022 28 commits
-
-
Geliang Tang authored
Add the multiple subflows test case for MP_FAIL, to test the MP_FAIL reset case. Use the test_linkfail value to make 1024KB test files. Invoke reset_with_fail() to use 'iptables' and 'tc action pedit' rules to produce the bit flips to trigger the checksum failures on ns2eth2. Add delays on ns2eth1 to make sure more data can translate on ns2eth2. The check_invert flag is enabled in reset_with_fail(), so this test prints out the inverted bytes, instead of the file mismatch errors. Invoke pedit_action_pkts() to get the numbers of the packets edited by the tc pedit actions, and print this numbers to the output. Co-developed-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
Geliang Tang <geliang.tang@suse.com> Signed-off-by:
Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Mat Martineau authored
The MPTCP socket's conn_list (list of subflows) requires the socket lock to access. The MP_FAIL timeout code added such an access, where it would check the list of subflows both in timer context and (later) in workqueue context where the socket lock is held. Rather than check the list twice, remove the check in the timeout handler and only depend on the check in the workqueue. Also remove the MPTCP_FAIL_NO_RESPONSE flag, since mptcp_mp_fail_no_response() has insignificant overhead and can be checked on each worker run. Fixes: 49fa1919 ("mptcp: reset subflow when MP_FAIL doesn't respond") Reported-by:
-
Mat Martineau authored
MP_FAIL timeout (waiting for a peer to respond to an MP_FAIL with another MP_FAIL) is implemented using the MPTCP socket's sk_timer. That timer is also used at MPTCP socket close, so it's important to not have the two timer users interfere with each other. At MPTCP socket close, all subflows are orphaned before sk_timer is manipulated. By checking the SOCK_DEAD flag on the subflows, each subflow can determine if the timer is safe to alter without acquiring any MPTCP-level lock. This replaces code that was using the mptcp_data_lock and MPTCP-level socket state checks that did not correctly protect the timer. Fixes: 49fa1919 ("mptcp: reset subflow when MP_FAIL doesn't respond") Reviewed-by:
-
Paolo Abeni authored
The mentioned helper requires the msk socket lock, and the current callers don't own it nor can't acquire it, so the access is racy. All the current callers are really checking for infinite mapping fallback, and the latter condition is explicitly tracked by the relevant msk variable: we can safely remove the caller usage - and the caller itself. The issue is present since MP_FAIL implementation, but the fix only applies since the infinite fallback support, ence the somewhat unexpected fixes tag. Fixes: 0530020a ("mptcp: track and update contiguous data status") Acked-and-tested-by:
-
Yuchung Cheng authored
This patch improves TCP PRR loss recovery behavior for a corner case. Previously during PRR conservation-bound mode, it strictly sends the amount equals to the amount newly acked or s/acked. The patch changes s.t. PRR may send additional amount that was banked previously (e.g. application-limited) in the conservation-bound mode, similar to the slow-start mode. This unifies and simplifies the algorithm further and may improve the recovery latency. This change still follow the general packet conservation design principle and always keep inflight/cwnd below the slow start threshold set by the congestion control module. PRR is described in RFC 6937. We'll include this change in the latest revision rfc6937-bis as well. Reported-by:
Neal Cardwell <ncardwell@google.com> Signed-off-by:
Yuchung Cheng <ycheng@google.com> Signed-off-by:
Eric Dumazet <edumazet@google.com> Link: https://lore.kernel.org/r/20220519003410.2531936-1-ycheng@google.comSigned-off-by:
-
Jakub Kicinski authored
The current title of our section of the documentation is Linux Networking Documentation. Since we're describing a section of Linux Documentation repeating those two words seems redundant. Link: https://lore.kernel.org/r/20220518234346.2088436-1-kuba@kernel.orgSigned-off-by:
-
Jakub Kicinski authored
GCC 12 seems upset that we check ipa_irq against array bound but then proceed, anyway: drivers/net/ipa/ipa_interrupt.c: In function ‘ipa_interrupt_add’: drivers/net/ipa/ipa_interrupt.c:196:27: warning: array subscript 30 is above array bounds of ‘void (*[30])(struct ipa *, enum ipa_irq_id)’ [-Warray-bounds] 196 | interrupt->handler[ipa_irq] = handler; | ~~~~~~~~~~~~~~~~~~^~~~~~~~~ drivers/net/ipa/ipa_interrupt.c:42:27: note: while referencing ‘handler’ 42 | ipa_irq_handler_t handler[IPA_IRQ_COUNT]; | ^~~~~~~ Reviewed-by:Alex Elder <elder@linaro.org> Link: https://lore.kernel.org/r/20220519004417.2109886-1-kuba@kernel.orgSigned-off-by:
-
Jakub Kicinski authored
GCC 12 warns: drivers/net/wwan/iosm/iosm_ipc_protocol_ops.c: In function ‘ipc_protocol_dl_td_process’: drivers/net/wwan/iosm/iosm_ipc_protocol_ops.c:406:13: warning: the comparison will always evaluate as ‘true’ for the address of ‘cb’ will never be NULL [-Waddress] 406 | if (!IPC_CB(skb)) { | ^ Indeed the check seems entirely pointless. Hopefully the other validation checks will catch if the cb is bad, but it can't be NULL. Reviewed-by:M Chetan Kumar <m.chetan.kumar@intel.com> Link: https://lore.kernel.org/r/20220519004342.2109832-1-kuba@kernel.orgSigned-off-by:
-
Jakub Kicinski authored
Martin Blumenstingl says: ==================== lantiq_gswip: Two small fixes While updating the Lantiq target in OpenWrt to Linux 5.15 I came across an FDB related error message. While that still needs to be solved I found two other small issues on the way. This series fixes the two minor issues found while revisiting the FDB code in the lantiq_gswip driver: - The first patch fixes the start index used in gswip_port_fdb() to find the entry with the matching bridge. The updated logic is now consistent with the rest of the driver. - The second patch fixes a typo in a dev_err() message. [0] https://lore.kernel.org/netdev/20220517194015.1081632-1-martin.blumenstingl@googlemail.com/ ==================== Link: https://lore.kernel.org/r/20220518220051.1520023-1-martin.blumenstingl@googlemail.comSigned-off-by:
-
Martin Blumenstingl authored
gswip_port_fdb_dump() reads the MAC bridge entries. The error message should say "failed to read mac bridge entry". While here, also add the index to the error print so humans can get to the cause of the problem easier. Acked-by:
Hauke Mehrtens <hauke@hauke-m.de> Reviewed-by:
Vladimir Oltean <olteanv@gmail.com> Signed-off-by:
Martin Blumenstingl <martin.blumenstingl@googlemail.com> Signed-off-by:
-
Martin Blumenstingl authored
The first N entries in priv->vlans are reserved for managing ports which are not part of a bridge. Use priv->hw_info->max_ports to consistently access per-bridge entries at index 7. Starting at priv->hw_info->cpu_port (6) is harmless in this case because priv->vlan[6].bridge is always NULL so the comparison result is always false (which results in this entry being skipped). Acked-by:
-
Ricardo Martinez authored
t7xx_request_irq() error: uninitialized symbol 'ret'. t7xx_core_hk_handler() error: potentially dereferencing uninitialized 'event'. If the condition to enter the loop that waits for the handshake event is false on the first iteration then the uninitialized 'event' will be dereferenced, fix this by initializing 'event' to NULL. t7xx_port_proxy_recv_skb() warn: variable dereferenced before check 'skb'. No need to check skb at t7xx_port_proxy_recv_skb() since we know it is always called with a valid skb by t7xx_cldma_gpd_rx_from_q(). Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Ricardo Martinez <ricardo.martinez@linux.intel.com> Link: https://lore.kernel.org/r/20220518195529.126246-1-ricardo.martinez@linux.intel.comSigned-off-by:
-
Jakub Kicinski authored
Russell King says: ==================== mtk_eth_soc phylink updates This series ultimately updates mtk_eth_soc to use phylink_pcs, with some fixes along the way. Previous attempts to update this driver (which is now marked as legacy) have failed due to lack of testing. I am hoping that this time will be different; Marek can test RGMII modes, but not SGMII. So all that we know is that this patch series probably doesn't break RGMII. 1) remove unused mac_mode and sgmii flags members from structures. 2) remove unnecessary interpretation of speed when configuring 1000 and 2500 Base-X 3) move configuration of SGMII duplex setting from mac_config() to link_up() 4) only pass in interface mode to mtk_sgmii_setup_mode_force() 5) move decision about which mtk_sgmii_setup_mode_*() function to call into mtk_sgmii.c 6) add a fixme comment for RGMII explaning why the call to mtk_gmac0_rgmii_adjust() is completely wrong - this needs to be addressed by someone who has the hardware and can test an appropriate fix. This fixme means that the driver still can't become non-legacy. 7) move gmac setup from mac_config() to mac_finish() - this preserves the order that we write to the hardware when we eventually convert to phylink_pcs() 8) move configuration of syscfg0 in SGMII/802.3z mode to mac_finish() for the same reasons as (7). 9) convert mtk_sgmii.c code structure and the mtk_sgmii structure to suit conversion to phylink_pcs 10) finally convert to phylink_pcs As there has been no feedback from mtk_eth_soc maintainers to my RFC on April 6th, not my reminder on April 11th, so it's now time to merge this anyway. Mediatek code seems to be submitted to the kernel and then the maintainers scarper... ==================== Link: https://lore.kernel.org/r/YoUIX+BN/ZbyXzTT@shell.armlinux.org.ukTested-by:
Marek Behún <kabel@kernel.org> Signed-off-by:
-
Russell King (Oracle) authored
Partially convert mtk_eth_soc to phylink_pcs, moving the configuration, link up and AN restart over. However, it seems mac_pcs_get_state() doesn't actually get the state from the PCS, so we can't convert that over without a better understanding of the hardware. Signed-off-by:
Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Signed-off-by:
-
Russell King (Oracle) authored
Provide a mtk_pcs structure which encapsulates everything that the PCS functions need (the regmap and ana_rgc3 offset), and use this in the PCS functions. Provide shim functions to convert from the existing "mtk_sgmii_*" interface to the converted PCS functions. Signed-off-by:
-
Russell King (Oracle) authored
The SGMIISYS configuration is performed while ETHSYS_SYSCFG0 is in a disabled state. In order to preserve this when we switch to phylink_pcs we need to move the restoration of this register to the mac_finish() callback. Signed-off-by:
-
Russell King (Oracle) authored
Move the setting of the MTK_MAC_MCR register from the end of mac_config into the phylink mac_finish() method, to keep it as the very last write that is done during configuration. Signed-off-by:
-
Russell King (Oracle) authored
Add a fixme comment for the last remaining incorrect usage of state->speed in the mac_config() method, which is strangely in a code path which is only run when the PHY interface mode changes. This means if we are in RGMII mode, changes in state->speed will not cause the INTF_MODE, TRGMII_RCK_CTRL and TRGMII_TCK_CTRL registers to be set according to the speed, nor will the TRGPLL clock be set to the correct value. Signed-off-by:
-
Russell King (Oracle) authored
Provide mtk_sgmii_config() to wrap up the decisions about which SGMII configuration will be called. Signed-off-by:
-
Russell King (Oracle) authored
Now that mtk_sgmii_setup_mode_force() only uses the interface mode from the phylink state, pass just the interface mode into this function. Signed-off-by:
-
Russell King authored
Phylink does not guarantee that state->duplex will be set correctly in the mac_config() call, so it's a bug that the driver makes use of it. Move the 802.3z PCS duplex configuration to mac_link_up(). Signed-off-by:
-
Russell King (Oracle) authored
Phylink does not guarantee that state->speed will be set correctly in the mac_config() call, so it's a bug that the driver makes use of it. Moreover, it is making use of it in a function that is only ever called for 1000BASE-X and 2500BASE-X which operate at a fixed speed which happens to be the same setting irrespective of the interface mode. We can simply remove the switch statement and just set the SGMII interface speed. Signed-off-by:
-
Russell King (Oracle) authored
The PCS speed setting is a two bit field, but it is defined as two separate bits. Add a bitfield mask for the speed definitions, an use the FIELD_PREP() macro to define each PCS speed. Signed-off-by:
-
Russell King (Oracle) authored
The "flags" member of struct mtk_sgmii appears to be unused, as are the MTK_SGMII_PHYSPEED_* and MTK_HAS_FLAGS() macros. Remove them. Signed-off-by:
-
Russell King (Oracle) authored
mac->mode is only ever written to in one location, and is thus superflous. Remove it. Signed-off-by:
-
Jakub Kicinski authored
Artem points out that skb may try to take over the skb and queue it to its own list. Unlink the skb before calling out. Fixes: b1a2c178 ("tls: rx: clear ctx->recv_pkt earlier") Reported-by:
Artem Savkov <asavkov@redhat.com> Tested-by:
-
Dan Carpenter authored
The netif_receive_skb() function frees "skb" so store skb->len before it is freed. Fixes: fd3040b9 ("net: ethernet: Add driver for Sunplus SP7021") Signed-off-by:
-
David Ober authored
Lenovo Thunderbolt 4 Dock, and other Lenovo USB Docks are using the original Realtek USB ethernet Vendor and Product IDs If the Network device is Realtek verify that it is on a Lenovo USB hub before enabling the passthru feature This also adds in the device IDs for the Lenovo USB Dongle and one other USB-C dock V2 fix formating of code V3 remove Generic define for Device ID 0x8153 and change it to use value V4 rearrange defines and case statement to put them in better order v5 create helper function to do the testing work as suggested Signed-off-by:
David Ober <dober6023@gmail.com> Link: https://lore.kernel.org/r/20220517180539.25839-1-dober6023@gmail.comSigned-off-by:
-
- 19 May, 2022 12 commits
-
-
Jakub Kicinski authored
Merge tag 'linux-can-next-for-5.19-20220519' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next Marc Kleine-Budde says: ==================== pull-request: can-next 2022-05-19 Oliver Hartkopp contributes a patch for the ISO-TP CAN protocol to update the validation of address information during bind. The next patch is by Jakub Kicinski and converts the CAN network drivers from netif_napi_add() to the netif_napi_add_weight() function. Another patch by Oliver Hartkopp removes obsolete CAN specific LED support. Vincent Mailhol's patch for the mcp251xfd driver fixes a -Wunaligned-access warning by clang-14. * tag 'linux-can-next-for-5.19-20220519' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next: can: mcp251xfd: silence clang's -Wunaligned-access warning can: can-dev: remove obsolete CAN LED support can: can-dev: move to netif_napi_add_weight() can: isotp: isotp_bind(): do not validate unused address information ==================== Link: https://lore.kernel.org/r/20220519202308.1435903-1-mkl@pengutronix.deSigned-off-by:
-
Vincent Mailhol authored
clang emits a -Wunaligned-access warning on union mcp251xfd_tx_ojb_load_buf. The reason is that field hw_tx_obj (not declared as packed) is being packed right after a 16 bits field inside a packed struct: | union mcp251xfd_tx_obj_load_buf { | struct __packed { | struct mcp251xfd_buf_cmd cmd; | /* ^ 16 bits fields */ | struct mcp251xfd_hw_tx_obj_raw hw_tx_obj; | /* ^ not declared as packed */ | } nocrc; | struct __packed { | struct mcp251xfd_buf_cmd_crc cmd; | struct mcp251xfd_hw_tx_obj_raw hw_tx_obj; | __be16 crc; | } crc; | } ____cacheline_aligned; Starting from LLVM 14, having an unpacked struct nested in a packed struct triggers a warning. c.f. [1]. This is a false positive because the field is always being accessed with the relevant put_unaligned_*() function. Adding __packed to the structure declaration silences the warning. [1] https://github.com/llvm/llvm-project/issues/55520 Link: https://lore.kernel.org/all/20220518114357.55452-1-mailhol.vincent@wanadoo.frSigned-off-by:Vincent Mailhol <mailhol.vincent@wanadoo.fr> Reported-by:
kernel test robot <lkp@intel.com> Tested-by: Nathan Chancellor <nathan@kernel.org> # build Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
Oliver Hartkopp authored
Since commit 30f3b421 ("can: mark led trigger as broken") the CAN specific LED support was disabled and marked as BROKEN. As the common LED support with CONFIG_LEDS_TRIGGER_NETDEV should do this work now the code can be removed as preparation for a CAN netdevice Kconfig rework. Link: https://lore.kernel.org/all/20220518154527.29046-1-socketcan@hartkopp.netSuggested-by:
Oliver Hartkopp <socketcan@hartkopp.net> [mkl: remove led.h from MAINTAINERS] Signed-off-by:
-
Jakub Kicinski authored
We want to remove the weight argument from the basic version of the netif_napi_add() call. Move all the callers in drivers/net/can that pass a custom weight (i.e. not NAPI_POLL_WEIGHT or 64) to the netif_napi_add_weight() API. Link: https://lore.kernel.org/all/20220517002345.1812104-1-kuba@kernel.orgSigned-off-by:
-
Oliver Hartkopp authored
With commit 2aa39889 ("can: isotp: isotp_bind(): return -EINVAL on incorrect CAN ID formatting") the bind() syscall returns -EINVAL when the given CAN ID needed to be sanitized. But in the case of an unconfirmed broadcast mode the rx CAN ID is not needed and may be uninitialized from the caller - which is ok. This patch makes sure the result of an inproper CAN ID format is only provided when the address information is needed. Link: https://lore.kernel.org/all/20220517145653.2556-1-socketcan@hartkopp.netSigned-off-by:
-
Jakub Kicinski authored
Merge tag 'wireless-next-2022-05-19' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next Kalle Valo says: ==================== wireless-next patches for v5.19 Second set of patches for v5.19 and most likely the last one. rtw89 got support for 8852ce devices and mt76 now supports Wireless Ethernet Dispatch. Major changes: cfg80211/mac80211 - support disabling EHT mode rtw89 - add support for Realtek 8852ce devices mt76 - Wireless Ethernet Dispatch support for flow offload - non-standard VHT MCS10-11 support - mt7921 AP mode support - mt7921 ipv6 NS offload support ath11k - enable keepalive during WoWLAN suspend - implement remain-on-channel support * tag 'wireless-next-2022-05-19' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (135 commits) iwlwifi: mei: fix potential NULL-ptr deref iwlwifi: mei: clear the sap data header before sending iwlwifi: mvm: remove vif_count iwlwifi: mvm: always tell the firmware to accept MCAST frames in BSS iwlwifi: mvm: add OTP info in case of init failure iwlwifi: mvm: fix assert 1F04 upon reconfig iwlwifi: fw: init SAR GEO table only if data is present iwlwifi: mvm: clean up authorized condition iwlwifi: mvm: use NULL instead of ERR_PTR when parsing wowlan status iwlwifi: pcie: simplify MSI-X cause mapping rtw89: pci: only mask out INT indicator register for disable interrupt v1 rtw89: convert rtw89_band to nl80211_band precisely rtw89: 8852c: update txpwr tables to HALRF_027_00_052 rtw89: cfo: check mac_id to avoid out-of-bounds rtw89: 8852c: set TX antenna path rtw89: add ieee80211::sta_rc_update ops wireless: Fix Makefile to be in alphabetical order mac80211: refactor freeing the next_beacon cfg80211: fix kernel-doc for cfg80211_beacon_data mac80211: minstrel_ht: support ieee80211_rate_status ... ==================== Link: https://lore.kernel.org/r/20220519153334.8D051C385AA@smtp.kernel.orgSigned-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski authored
drivers/net/ethernet/mellanox/mlx5/core/main.c b3388697 ("net/mlx5: Initialize flow steering during driver probe") 40379a00 ("net/mlx5_fpga: Drop INNOVA TLS support") f2b41b32 ("net/mlx5: Remove ipsec_ops function table") https://lore.kernel.org/all/20220519040345.6yrjromcdistu7vh@sx1/ 16d42d31 ("net/mlx5: Drain fw_reset when removing device") 8324a02c ("net/mlx5: Add exit route when waiting for FW") https://lore.kernel.org/all/20220519114119.060ce014@canb.auug.org.au/ tools/testing/selftests/net/mptcp/mptcp_join.sh e274f715 ("selftests: mptcp: add subflow limits test-cases") b6e074e1 ("selftests: mptcp: add infinite map testcase") 5ac1d2d6 ("selftests: mptcp: Add tests for userspace PM type") https://lore.kernel.org/all/20220516111918.366d747f@canb.auug.org.au/ net/mptcp/options.c ba2c89e0 ("mptcp: fix checksum byte order") 1e39e5a3 ("mptcp: infinite mapping sending") ea66758c ("tcp: allow MPTCP to update the announced window") https://lore.kernel.org/all/20220519115146.751c3a37@canb.auug.org.au/ net/mptcp/pm.c 95d68651 ("mptcp: fix subflow accounting on close") 4d25247d ("mptcp: bypass in-kernel PM restrictions for non-kernel PMs") https://lore.kernel.org/all/20220516111435.72f35dca@canb.auug.org.au/ net/mptcp/subflow.c ae66fb2b ("mptcp: Do TCP fallback on early DSS checksum failure") 0348c690 ("mptcp: add the fallback check") f8d4bcac ("mptcp: infinite mapping receiving") https://lore.kernel.org/all/20220519115837.380bb8d4@canb.auug.org.au/Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netLinus Torvalds authored
Pull networking fixes from Paolo Abeni: "Including fixes from can, xfrm and netfilter subtrees. Notably this reverts a recent TCP/DCCP netns-related change to address a possible UaF. Current release - regressions: - tcp: revert "tcp/dccp: get rid of inet_twsk_purge()" - xfrm: set dst dev to blackhole_netdev instead of loopback_dev in ifdown Previous releases - regressions: - netfilter: flowtable: fix TCP flow teardown - can: revert "can: m_can: pci: use custom bit timings for Elkhart Lake" - xfrm: check encryption module availability consistency - eth: vmxnet3: fix possible use-after-free bugs in vmxnet3_rq_alloc_rx_buf() - eth: mlx5: initialize flow steering during driver probe - eth: ice: fix crash when writing timestamp on RX rings Previous releases - always broken: - mptcp: fix checksum byte order - eth: lan966x: fix assignment of the MAC address - eth: mlx5: remove HW-GRO from reported features - eth: ftgmac100: disable hardware checksum on AST2600" * tag 'net-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (50 commits) net: bridge: Clear offload_fwd_mark when passing frame up bridge interface. ptp: ocp: change sysfs attr group handling selftests: forwarding: fix missing backslash netfilter: nf_tables: disable expression reduction infra netfilter: flowtable: move dst_check to packet path netfilter: flowtable: fix TCP flow teardown net: ftgmac100: Disable hardware checksum on AST2600 igb: skip phy status check where unavailable nfc: pn533: Fix buggy cleanup order mptcp: Do TCP fallback on early DSS checksum failure mptcp: fix checksum byte order net: af_key: check encryption module availability consistency net: af_key: add check for pfkey_broadcast in function pfkey_process net/mlx5: Drain fw_reset when removing device net/mlx5e: CT: Fix setting flow_source for smfs ct tuples net/mlx5e: CT: Fix support for GRE tuples net/mlx5e: Remove HW-GRO from reported features net/mlx5e: Properly block HW GRO when XDP is enabled net/mlx5e: Properly block LRO when XDP is enabled net/mlx5e: Block rx-gro-hw feature in switchdev mode ... -
Boris Pismenny authored
TLS device offload copies sendfile data to a bounce buffer before transmitting. It allows to maintain the valid MAC on TLS records when the file contents change and a part of TLS record has to be retransmitted on TCP level. In many common use cases (like serving static files over HTTPS) the file contents are not changed on the fly. In many use cases breaking the connection is totally acceptable if the file is changed during transmission, because it would be received corrupted in any case. This commit allows to optimize performance for such use cases to providing a new optional mode of TLS sendfile(), in which the extra copy is skipped. Removing this copy improves performance significantly, as TLS and TCP sendfile perform the same operations, and the only overhead is TLS header/trailer insertion. The new mode can only be enabled with the new socket option named TLS_TX_ZEROCOPY_SENDFILE on per-socket basis. It preserves backwards compatibility with existing applications that rely on the copying behavior. The new mode is safe, meaning that unsolicited modifications of the file being sent can't break integrity of the kernel. The worst thing that can happen is sending a corrupted TLS record, which is in any case not forbidden when using regular TCP sockets. Sockets other than TLS device offload are not affected by the new socket option. The actual status of zerocopy sendfile can be queried with sock_diag. Performance numbers in a single-core test with 24 HTTPS streams on nginx, under 100% CPU load: * non-zerocopy: 33.6 Gbit/s * zerocopy: 79.92 Gbit/s CPU: Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz Signed-off-by:
Boris Pismenny <borisp@nvidia.com> Signed-off-by:
Tariq Toukan <tariqt@nvidia.com> Signed-off-by:
Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by:
-
Yinjun Zhang authored
Current implementation of ct merging doesn't support the case that the fields mangling in pre_ct rules are matched in post_ct rules. This change is to support merging when mangling mac address, ip address, tos, ttl and l4 port. VLAN and MPLS mangling is not involved yet. Signed-off-by:
Yinjun Zhang <yinjun.zhang@corigine.com> Signed-off-by:
Louis Peens <louis.peens@corigine.com> Signed-off-by:
Simon Horman <simon.horman@corigine.com> Link: https://lore.kernel.org/r/20220518075055.130649-1-simon.horman@corigine.comSigned-off-by:
-
Michael Trimarchi authored
Make ndo_open less sensitive to memory pressure. Signed-off-by:
Michael Trimarchi <michael@amarulasolutions.com> Reviewed-by:
-
Andrew Lunn authored
It is possible to stack bridges on top of each other. Consider the following which makes use of an Ethernet switch: br1 / \ / \ / \ br0.11 wlan0 | br0 / | \ p1 p2 p3 br0 is offloaded to the switch. Above br0 is a vlan interface, for vlan 11. This vlan interface is then a slave of br1. br1 also has a wireless interface as a slave. This setup trunks wireless lan traffic over the copper network inside a VLAN. A frame received on p1 which is passed up to the bridge has the skb->offload_fwd_mark flag set to true, indicating that the switch has dealt with forwarding the frame out ports p2 and p3 as needed. This flag instructs the software bridge it does not need to pass the frame back down again. However, the flag is not getting reset when the frame is passed upwards. As a result br1 sees the flag, wrongly interprets it, and fails to forward the frame to wlan0. When passing a frame upwards, clear the flag. This is the Rx equivalent of br_switchdev_frame_unmark() in br_dev_xmit(). Fixes: f1c2eddf ("bridge: switchdev: Use an helper to clear forward mark") Signed-off-by:Andrew Lunn <andrew@lunn.ch> Reviewed-by:
Ido Schimmel <idosch@nvidia.com> Tested-by:
Nikolay Aleksandrov <razor@blackwall.org> Link: https://lore.kernel.org/r/20220518005840.771575-1-andrew@lunn.chSigned-off-by:
-
