From f4cdf7ca9a1fdcca413157df19753f388a5a224e Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 19 Aug 2026 10:09:22 -0700 Subject: Merge tag 'media/v7.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media Pull media updates from Mauro Carvalho Chehab: - v4l2-core: added ISP statistics support and per-block validation - v4l2-core: Allow unknown HDR10 white point and luminance - New camera sensors: Sony IMX678 and IMX471m, Himax HM1092 IR sensor - New codec: Milos: VPU v2.0 codec support - isp driver: gained support for Dreamchip RPPX1 ISP framework - vsp1 driver: gained support for RZ/T2H and RZ/N2H - Novalake driver: gained CVS support for new NVL hardware - dvb-core: fix feed leak on failed DMX_ADD_PID - several driver fixes, cleanups and minor improvements * tag 'media/v7.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (308 commits) media: ipu-bridge: check all DMI entries when overriding sensor rotation media: v4l2-async: avoid deleting unlinked ASC entry on link error media: rzg2l-cru: Align bytesperline to hardware DMA stride requirement media: intel/ipu6: fix async notifier cleanup leak on parse error media: staging/ipu7: fix async notifier UAF on probe error path media: amd: isp4: fix self-deadlock in isp4sd_pwron_and_init() error path media: amd: isp4: release partial allocations in isp4if_alloc_fw_gpumem() media: rcar-isp: Fix VSPX reference leaks media: rcar-isp: Release ISPCORE resources media: i2c: imx415: Release runtime PM reference on VBLANK error media: i2c: imx415: Return test pattern write errors media: renesas: vsp1: Declare index variables in for loop statement media: renesas: vsp1: Make reset control optional to support platforms without a reset line media: dt-bindings: media: renesas,vsp1: Document RZ/T2H and RZ/N2H SoCs media: dt-bindings: media: renesas,fcp: Document RZ/T2H and RZ/N2H SoCs media: nxp: imx8-isi: Add additional 32-bit RGB format support media: nxp: imx8-isi: Add 16-bit raw Bayer format support media: nxp: imx8-isi: Implement per-stream reference counting for multiplexed streams media: nxp: imx8-isi: Use BIT_ULL() for 64-bit stream masks media: nxp: imx8-isi: Correct color map between V4L2 and ISI ... --- include/net/rps.h | 197 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 197 insertions(+) create mode 100644 include/net/rps.h (limited to 'include/net/rps.h') diff --git a/include/net/rps.h b/include/net/rps.h new file mode 100644 index 000000000..e33c6a2fa --- /dev/null +++ b/include/net/rps.h @@ -0,0 +1,197 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +#ifndef _NET_RPS_H +#define _NET_RPS_H + +#include +#include +#include +#include + +#ifdef CONFIG_RPS +#include + +extern struct static_key_false rps_needed; +extern struct static_key_false rfs_needed; + +/* + * This structure holds an RPS map which can be of variable length. The + * map is an array of CPUs. + */ +struct rps_map { + unsigned int len; + struct rcu_head rcu; + u16 cpus[]; +}; +#define RPS_MAP_SIZE(_num) (sizeof(struct rps_map) + ((_num) * sizeof(u16))) + +/* + * The rps_dev_flow structure contains the mapping of a flow to a CPU, the + * tail pointer for that CPU's input queue at the time of last enqueue, a + * hardware filter index, and the hash of the flow if aRFS is enabled. + */ +struct rps_dev_flow { + u16 cpu; + u16 filter; + unsigned int last_qtail; +#ifdef CONFIG_RFS_ACCEL + u32 hash; +#endif +}; +#define RPS_NO_FILTER 0xffff + +/* + * The rps_sock_flow_table contains mappings of flows to the last CPU + * on which they were processed by the application (set in recvmsg). + * Each entry is a 32bit value. Upper part is the high-order bits + * of flow hash, lower part is CPU number. + * rps_cpu_mask is used to partition the space, depending on number of + * possible CPUs : rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1 + * For example, if 64 CPUs are possible, rps_cpu_mask = 0x3f, + * meaning we use 32-6=26 bits for the hash. + */ +struct rps_sock_flow_table { + u32 ent; +}; + +#define RPS_NO_CPU 0xffff + +static inline void rps_record_sock_flow(rps_tag_ptr tag_ptr, u32 hash) +{ + unsigned int index = hash & rps_tag_to_mask(tag_ptr); + u32 val = hash & ~net_hotdata.rps_cpu_mask; + struct rps_sock_flow_table *table; + + /* We only give a hint, preemption can change CPU under us */ + val |= raw_smp_processor_id(); + + table = rps_tag_to_table(tag_ptr); + /* The following WRITE_ONCE() is paired with the READ_ONCE() + * here, and another one in get_rps_cpu(). + */ + if (READ_ONCE(table[index].ent) != val) + WRITE_ONCE(table[index].ent, val); +} + +static inline void _sock_rps_record_flow_hash(__u32 hash) +{ + rps_tag_ptr tag_ptr; + + if (!hash) + return; + rcu_read_lock(); + tag_ptr = READ_ONCE(net_hotdata.rps_sock_flow_table); + if (tag_ptr) + rps_record_sock_flow(tag_ptr, hash); + rcu_read_unlock(); +} + +static inline void _sock_rps_record_flow(const struct sock *sk) +{ + /* Reading sk->sk_rxhash might incur an expensive cache line + * miss. + * + * TCP_ESTABLISHED does cover almost all states where RFS + * might be useful, and is cheaper [1] than testing : + * IPv4: inet_sk(sk)->inet_daddr + * IPv6: ipv6_addr_any(&sk->sk_v6_daddr) + * OR an additional socket flag + * [1] : sk_state and sk_prot are in the same cache line. + */ + if (sk->sk_state == TCP_ESTABLISHED) { + /* This READ_ONCE() is paired with the WRITE_ONCE() + * from sock_rps_save_rxhash() and sock_rps_reset_rxhash(). + */ + _sock_rps_record_flow_hash(READ_ONCE(sk->sk_rxhash)); + } +} + +static inline void _sock_rps_delete_flow(const struct sock *sk) +{ + struct rps_sock_flow_table *table; + rps_tag_ptr tag_ptr; + u32 hash, index; + + hash = READ_ONCE(sk->sk_rxhash); + if (!hash) + return; + + rcu_read_lock(); + tag_ptr = READ_ONCE(net_hotdata.rps_sock_flow_table); + if (tag_ptr) { + index = hash & rps_tag_to_mask(tag_ptr); + table = rps_tag_to_table(tag_ptr); + if (READ_ONCE(table[index].ent) != RPS_NO_CPU) + WRITE_ONCE(table[index].ent, RPS_NO_CPU); + } + rcu_read_unlock(); +} +#endif /* CONFIG_RPS */ + +static inline bool rfs_is_needed(void) +{ +#ifdef CONFIG_RPS + return static_branch_unlikely(&rfs_needed); +#else + return false; +#endif +} + +static inline void sock_rps_record_flow_hash(__u32 hash) +{ +#ifdef CONFIG_RPS + if (!rfs_is_needed()) + return; + + _sock_rps_record_flow_hash(hash); +#endif +} + +static inline void sock_rps_record_flow(const struct sock *sk) +{ +#ifdef CONFIG_RPS + if (!rfs_is_needed()) + return; + + _sock_rps_record_flow(sk); +#endif +} + +static inline void sock_rps_delete_flow(const struct sock *sk) +{ +#ifdef CONFIG_RPS + if (!rfs_is_needed()) + return; + + _sock_rps_delete_flow(sk); +#endif +} + +static inline u32 rps_input_queue_tail_incr(struct softnet_data *sd) +{ +#ifdef CONFIG_RPS + return ++sd->input_queue_tail; +#else + return 0; +#endif +} + +static inline void rps_input_queue_tail_save(u32 *dest, u32 tail) +{ +#ifdef CONFIG_RPS + WRITE_ONCE(*dest, tail); +#endif +} + +static inline void rps_input_queue_head_add(struct softnet_data *sd, int val) +{ +#ifdef CONFIG_RPS + WRITE_ONCE(sd->input_queue_head, sd->input_queue_head + val); +#endif +} + +static inline void rps_input_queue_head_incr(struct softnet_data *sd) +{ + rps_input_queue_head_add(sd, 1); +} + +#endif /* _NET_RPS_H */ -- cgit v1.3.1