- 28 Dec, 2021 9 commits
-
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx. ixgbe driver runs Tx completion polling cycle right before the Rx one and uses napi_consume_skb() to feed the cache with skbuff_heads of completed entries, so it's never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx. Signed-off-by:
Alexander Lobakin <alexandr.lobakin@intel.com> Reviewed-by:
Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Tested-by:
Gurucharan G <gurucharanx.g@intel.com> Signed-off-by:
Tony Nguyen <anthony.l.nguyen@intel.com>
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx. igc driver runs Tx completion polling cycle right before the Rx one and uses napi_consume_skb() to feed the cache with skbuff_heads of completed entries, so it's never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx. Signed-off-by:
Nechama Kraus <nechamax.kraus@linux.intel.com> Signed-off-by:
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx. igb driver runs Tx completion polling cycle right before the Rx one and uses napi_consume_skb() to feed the cache with skbuff_heads of completed entries, so it's never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx. Signed-off-by:
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx. ice driver runs Tx completion polling cycle right before the Rx one and uses napi_consume_skb() to feed the cache with skbuff_heads of completed entries, so it's never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx. Signed-off-by:
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx. iavf driver runs Tx completion polling cycle right before the Rx one and uses napi_consume_skb() to feed the cache with skbuff_heads of completed entries, so it's never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx. Signed-off-by:
Konrad Jankowski <konrad0.jankowski@intel.com> Signed-off-by:
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx. i40e driver runs Tx completion polling cycle right before the Rx one and uses napi_consume_skb() to feed the cache with skbuff_heads of completed entries, so it's never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx. Signed-off-by:
-
Alexander Lobakin authored
napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save some cycles on freeing/allocating skbuff_heads on every new Rx or completed Tx element. e1000 driver runs Tx completion polling cycle right before the Rx one. Now that e1000 uses napi_consume_skb() to put skbuff_heads of completed entries into the cache, it will never empty and always warm at that moment. Switch to the napi_build_skb() to relax mm pressure on heavy Rx and increase throughput. Signed-off-by:
Tony Brelinski <tony.brelinski@intel.com> Signed-off-by:
-
Alexander Lobakin authored
In order to take the best from per-cpu NAPI skbuff_head caches and CPU cycles, let's switch from dev_kfree_skb_any(), which passes skb back to the mm layer, to napi_consume_skb(), which feeds those caches on non-zero budget instead (falls back to the former on 0). Do the replacement in e1000_unmap_and_free_tx_resource(). There are 4 call sites of this function throughout the driver: * e1000_clean_tx_ring(). Slowpath, process context, cleans the whole Tx ring on ifdown. Use budget of 0 here; * e1000_tx_map(). Hotpath, net Tx softirq, unmaps the buffers in case of error. Use 0 as well; * e1000_clean_tx_irq(). Hotpath, NAPI Tx completion polling cycle. As the driver doesn't count completed Tx entries towards the NAPI budget, just use the poll budget of 64 to utilize caches. Apart from being a preparation for switching to napi_build_skb(), this is useful on its own as well, as napi_consume_skb() flushes skb caches by batches of 32 instead of one-at-a-time. Signed-off-by:
-
Colin Ian King authored
Variable expectlen is being assigned a value that is never read, the assignment occurs before a return statement. The assignment is redundant and can be removed. Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
- 27 Dec, 2021 19 commits
-
-
Wen Zhiwei authored
Delete the initialization of three static variables because it is meaningless. Signed-off-by:
Wen Zhiwei <wenzhiwei@kylinos.cn> Signed-off-by:
-
Remi Pommarel authored
In compat mode SIOC{G,S}IFBR ioctls were only supporting BRCTL_GET_VERSION returning an artificially version to spur userland tool to use SIOCDEVPRIVATE instead. But some userland tools ignore that and use SIOC{G,S}IFBR unconditionally as seen with busybox's brctl. Example of non working 32-bit brctl with CONFIG_COMPAT=y: $ brctl show brctl: SIOCGIFBR: Invalid argument Example of fixed 32-bit brctl with CONFIG_COMPAT=y: $ brctl show bridge name bridge id STP enabled interfaces br0 Signed-off-by:Remi Pommarel <repk@triplefau.lt> Co-developed-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
Nikolay Aleksandrov <nikolay@nvidia.com> Signed-off-by:
-
Lad Prabhakar authored
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining. In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq() for DT users only. While at it propagate error code in case request_irq() fails instead of returning -EBUSY. Signed-off-by:
Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com> Signed-off-by:
-
Lad Prabhakar authored
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining. In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq(). Signed-off-by:
-
Lad Prabhakar authored
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining. In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq(). Signed-off-by:
-
Lad Prabhakar authored
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining. In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq(). While doing so return error pointer from read_dts_node() as platform_get_irq() may return -EPROBE_DEFER. Signed-off-by:
-
Lad Prabhakar authored
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining. In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq(). Signed-off-by:
-
Lad Prabhakar authored
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining. In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq(). Signed-off-by:
-
Kai-Heng Feng authored
We are seeing spurious wakeup caused by Intel 7560 WWAN on AMD laptops. This prevent those laptops to stay in s2idle state. >From what I can understand, the intention of ipc_pcie_suspend() is to put the device to D3cold, and ipc_pcie_suspend_s2idle() is to keep the device at D0. However, the device can still be put to D3hot/D3cold by PCI core. So explicitly let PCI core know this device should stay at D0, to solve the spurious wakeup. Signed-off-by:
Kai-Heng Feng <kai.heng.feng@canonical.com> Signed-off-by:
-
Kai-Heng Feng authored
pci_pm_suspend_noirq() and pci_pm_resume_noirq() already handle power transition for system-wide suspend and resume, so it's not necessary to do it in the driver. Signed-off-by:
-
Horatiu Vultur authored
The blamed commit changed the vlan used by the host ports to be 4095 instead of 0. Because of this change the following issues are seen: - when the port is probed first it was adding an entry in the MAC table with the wrong vlan (port->pvid which is default 0) and not HOST_PVID - when the port is removed from a bridge, it was using the wrong vlan to add entries in the MAC table. It was using the old PVID and not the HOST_PVID This patch fixes this two issues by using the HOST_PVID instead of port->pvid. Fixes: 6d2c186a ("net: lan966x: Add vlan support.") Signed-off-by:
Horatiu Vultur <horatiu.vultur@microchip.com> Signed-off-by:
-
David S. Miller authored
Michael Chan says: ==================== bnxt_en: Update for net-next This series includes some added error logging for firmware error reports, DIM interrupt sampling for the latest 5750X chips, CQE interrupt coalescing mode support, and to use RX page frag buffers for better software GRO performance. ==================== Signed-off-by: -
Jakub Kicinski authored
If NETIF_F_GRO_HW is disabled, the existing driver code uses kmalloc'ed data for RX buffers. This causes inefficient SW GRO performance because the GRO data is merged using the less efficient frag_list. Use netdev_alloc_frag() and friends instead so that GRO data can be merged into skb_shinfo(skb)->frags for better performance. [Use skb_free_frag() - Vikas Gupta] Signed-off-by:
Jakub Kicinski <kuba@kernel.org> Reviewed-by:
Pavan Chebbi <pavan.chebbi@broadcom.com> Reviewed-by:
Andy Gospodarek <gospo@broadcom.com> Signed-off-by:
Vikas Gupta <vikas.gupta@broadcom.com> Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Signed-off-by:
-
Edwin Peer authored
The xdp_do_flush_map function has been replaced with the more general xdp_do_flush(). Signed-off-by:
Edwin Peer <edwin.peer@broadcom.com> Signed-off-by:
-
Michael Chan authored
Support showing and setting the CQE mode in ethtool. Reviewed-by:
-
Michael Chan authored
CQE coalescing mode is the same as the timer reset coalescing mode on Broadcom devices. Currently this mode is always enabled if it is supported by the device. Restructure the code slightly to support dynamically changing this mode. Add a flags field to struct bnxt_coal. Initially, the CQE flag will be set for the RX and TX side if the device supports it. We need to move bnxt_init_dflt_coal() to set up default coalescing until the capability is determined. Reviewed-by:
-
Andy Gospodarek authored
5750X (P5) chips handle receiving packets on the NQ rather than the main completion queue so we need to get and set stats from the correct spots for dynamic interrupt moderation. Signed-off-by:
-
Michael Chan authored
Log the unrecognized error report type value as well. Reviewed-by:
-
Somnath Kotur authored
FW has been modified to send a new async event when it detects a pause storm. Register for this new event and log it upon receipt. Reviewed-by:
Somnath Kotur <somnath.kotur@broadcom.com> Signed-off-by:
-
- 24 Dec, 2021 4 commits
-
-
Christophe JAILLET authored
Use 'bitmap_zalloc()' to simplify code, improve the semantic and reduce some open-coded arithmetic in allocator arguments. Also change the corresponding 'kfree()' into 'bitmap_free()' to keep consistency. Signed-off-by:
Christophe JAILLET <christophe.jaillet@wanadoo.fr> Link: https://lore.kernel.org/r/f9541b085ec68e573004e1be200c11c9c901181a.1640295165.git.christophe.jaillet@wanadoo.frSigned-off-by:
-
Horatiu Vultur authored
Add config_init for LAN8814. This function is required for the following reasons: - we need to make sure that the PHY is reset, - disable ANEG with QSGMII PCS Host side - swap the MDI-X A,B transmit so that there will not be any link flip-flaps when the PHY gets a link. Signed-off-by:
-
Randy Dunlap authored
Use the possessive "its" instead of the contraction of "it is" ("it's") in user messages. Signed-off-by:Randy Dunlap <rdunlap@infradead.org> Cc: Krzysztof Halasa <khc@pm.waw.pl> Cc: "David S. Miller" <davem@davemloft.net> Cc: Jakub Kicinski <kuba@kernel.org> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski authored
include/net/sock.h commit 8f905c0e ("inet: fully convert sk->sk_rx_dst to RCU rules") commit 43f51df4 ("net: move early demux fields close to sk_refcnt") https://lore.kernel.org/all/20211222141641.0caa0ab3@canb.auug.org.au/Signed-off-by:
-
- 23 Dec, 2021 8 commits
-
-
Nobuhiro Iwamatsu authored
ETHER_CLK_SEL_FREQ_SEL_2P5M is not 0 bit of the register. This is a value, which is 0. Fix from BIT(0) to 0. Reported-by:
Yuji Ishikawa <yuji2.ishikawa@toshiba.co.jp> Fixes: b38dd98f ("net: stmmac: Add Toshiba Visconti SoCs glue driver") Signed-off-by:
Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp> Link: https://lore.kernel.org/r/20211223073633.101306-1-nobuhiro1.iwamatsu@toshiba.co.jpSigned-off-by:
-
Jakub Kicinski authored
Hayes Wang says: ==================== r8152: fix bugs Patch #1 fix the issue of force speed mode for RTL8156. Patch #2 fix the issue of unexpected ocp_base. ==================== Link: https://lore.kernel.org/r/20211223092702.23841-386-nic_swsd@realtek.comSigned-off-by:
-
Hayes Wang authored
There are some chances that the actual base of hardware is different from the value recorded by driver, so we have to reset the variable of ocp_base to sync it. Set ocp_base to -1. Then, it would be updated and the new base would be set to the hardware next time. Signed-off-by:
Hayes Wang <hayeswang@realtek.com> Signed-off-by:
-
Hayes Wang authored
It needs to set mdio force mode. Otherwise, link off always occurs when setting force speed. Fixes: 195aae32 ("r8152: support new chips") Signed-off-by:
-
Remi Pommarel authored
Commit 561d8352 ("bridge: use ndo_siocdevprivate") changed the source and destination arguments of copy_{to,from}_user in bridge's old_deviceless() from args[1] to uarg breaking SIOC{G,S}IFBR ioctls. Commit cbd7ad29 ("net: bridge: fix ioctl old_deviceless bridge argument") fixed only BRCTL_{ADD,DEL}_BRIDGES commands leaving BRCTL_GET_BRIDGES one untouched. The fixes BRCTL_GET_BRIDGES as well and has been tested with busybox's brctl. Example of broken brctl: $ brctl show bridge name bridge id STP enabled interfaces brctl: can't get bridge name for index 0: No such device or address Example of fixed brctl: $ brctl show bridge name bridge id STP enabled interfaces br0 8000.000000000000 no Fixes: 561d8352 ("bridge: use ndo_siocdevprivate") Signed-off-by:
-
Xiaoliang Yang authored
Convert the u32 variable to type u64 in a context where expression of type u64 is required to avoid potential overflow. Fixes: e9e37200 ("net: stmmac: ptp: update tas basetime after ptp adjust") Signed-off-by:
Xiaoliang Yang <xiaoliang.yang_1@nxp.com> Link: https://lore.kernel.org/r/20211223073928.37371-1-xiaoliang.yang_1@nxp.comSigned-off-by:
-
Xiaoliang Yang authored
For Ocelot switches, the CPU injected frames have an injection header where it can specify the QoS class of the packet and the DSA tag, now it uses the SKB priority to set that. If a traffic class to priority mapping is configured on the netdevice (with mqprio for example ...), it won't be considered for CPU injected headers. This patch make the QoS class aligned to the priority to traffic class mapping if it exists. Fixes: 8dce89aa ("net: dsa: ocelot: add tagger for Ocelot/Felix switches") Signed-off-by:
Marouen Ghodhbane <marouen.ghodhbane@nxp.com> Link: https://lore.kernel.org/r/20211223072211.33130-1-xiaoliang.yang_1@nxp.comSigned-off-by:
-
Paolo Abeni authored
After commit d3256efd ("veth: allow enabling NAPI even without XDP"), if GRO is enabled on a veth device and TSO is disabled on the peer device, TCP skbs will go through the NAPI callback. If there is no XDP program attached, the veth code does not perform any share check, and shared/cloned skbs could enter the GRO engine. Ignat reported a BUG triggered later-on due to the above condition: [ 53.970529][ C1] kernel BUG at net/core/skbuff.c:3574! [ 53.981755][ C1] invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI [ 53.982634][ C1] CPU: 1 PID: 19 Comm: ksoftirqd/1 Not tainted 5.16.0-rc5+ #25 [ 53.982634][ C1] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015 [ 53.982634][ C1] RIP: 0010:skb_shift+0x13ef/0x23b0 [ 53.982634][ C1] Code: ea 03 0f b6 04 02 48 89 fa 83 e2 07 38 d0 7f 08 84 c0 0f 85 41 0c 00 00 41 80 7f 02 00 4d 8d b5 d0 00 00 00 0f 85 74 f5 ff ff <0f> 0b 4d 8d 77 20 be 04 00 00 00 4c 89 44 24 78 4c 89 f7 4c 89 8c [ 53.982634][ C1] RSP: 0018:ffff8881008f7008 EFLAGS: 00010246 [ 53.982634][ C1] RAX: 0000000000000000 RBX: ffff8881180b4c80 RCX: 0000000000000000 [ 53.982634][ C1] RDX: 0000000000000002 RSI: ffff8881180b4d3c RDI: ffff88810bc9cac2 [ 53.982634][ C1] RBP: ffff8881008f70b8 R08: ffff8881180b4cf4 R09: ffff8881180b4cf0 [ 53.982634][ C1] R10: ffffed1022999e5c R11: 0000000000000002 R12: 0000000000000590 [ 53.982634][ C1] R13: ffff88810f940c80 R14: ffff88810f940d50 R15: ffff88810bc9cac0 [ 53.982634][ C1] FS: 0000000000000000(0000) GS:ffff888235880000(0000) knlGS:0000000000000000 [ 53.982634][ C1] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 53.982634][ C1] CR2: 00007ff5f9b86680 CR3: 0000000108ce8004 CR4: 0000000000170ee0 [ 53.982634][ C1] Call Trace: [ 53.982634][ C1] <TASK> [ 53.982634][ C1] tcp_sacktag_walk+0xaba/0x18e0 [ 53.982634][ C1] tcp_sacktag_write_queue+0xe7b/0x3460 [ 53.982634][ C1] tcp_ack+0x2666/0x54b0 [ 53.982634][ C1] tcp_rcv_established+0x4d9/0x20f0 [ 53.982634][ C1] tcp_v4_do_rcv+0x551/0x810 [ 53.982634][ C1] tcp_v4_rcv+0x22ed/0x2ed0 [ 53.982634][ C1] ip_protocol_deliver_rcu+0x96/0xaf0 [ 53.982634][ C1] ip_local_deliver_finish+0x1e0/0x2f0 [ 53.982634][ C1] ip_sublist_rcv_finish+0x211/0x440 [ 53.982634][ C1] ip_list_rcv_finish.constprop.0+0x424/0x660 [ 53.982634][ C1] ip_list_rcv+0x2c8/0x410 [ 53.982634][ C1] __netif_receive_skb_list_core+0x65c/0x910 [ 53.982634][ C1] netif_receive_skb_list_internal+0x5f9/0xcb0 [ 53.982634][ C1] napi_complete_done+0x188/0x6e0 [ 53.982634][ C1] gro_cell_poll+0x10c/0x1d0 [ 53.982634][ C1] __napi_poll+0xa1/0x530 [ 53.982634][ C1] net_rx_action+0x567/0x1270 [ 53.982634][ C1] __do_softirq+0x28a/0x9ba [ 53.982634][ C1] run_ksoftirqd+0x32/0x60 [ 53.982634][ C1] smpboot_thread_fn+0x559/0x8c0 [ 53.982634][ C1] kthread+0x3b9/0x490 [ 53.982634][ C1] ret_from_fork+0x22/0x30 [ 53.982634][ C1] </TASK> Address the issue by skipping the GRO stage for shared or cloned skbs. To reduce the chance of OoO, try to unclone the skbs before giving up. v1 -> v2: - use avoid skb_copy and fallback to netif_receive_skb - Eric Reported-by:
Ignat Korchagin <ignat@cloudflare.com> Fixes: d3256efd ("veth: allow enabling NAPI even without XDP") Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Tested-by:
Eric Dumazet <edumazet@google.com> Link: https://lore.kernel.org/r/b5f61c5602aab01bac8d711d8d1bfab0a4817db7.1640197544.git.pabeni@redhat.comSigned-off-by:
-
