Hi Marc and everyone, third version part 2, functionally I had to move from spin_lock to spin_lock_irqsave because of an interrupt that was calling start_xmit, see attached stack. This is tested on tcan455x but I don't have the integrated hardware myself so any testing is appreciated. The series implements many small and bigger throughput improvements and adds rx/tx coalescing at the end. Best, Markus Changes in v3: - Remove parenthesis in error messages - Use memcpy_and_pad for buffer copy in 'can: m_can: Write transmit header and data in one transaction'. - Replace spin_lock with spin_lock_irqsave. I got a report of a interrupt that was calling start_xmit just after the netqueue was woken up before the locked region was exited. spin_lock_irqsave should fix this. I attached the full stack at the end of the mail if someone wants to know. - Rebased to v6.3-rc1. - Removed tcan4x5x patches from this series. Changes in v2: - Rebased on v6.2-rc5 - Fixed missing/broken accounting for non peripheral m_can devices. part 1: v1 - https://lore.kernel.org/lkml/20221116205308.2996556-1-msp@xxxxxxxxxxxx v2 - https://lore.kernel.org/lkml/20221206115728.1056014-1-msp@xxxxxxxxxxxx part 2: v1 - https://lore.kernel.org/lkml/20221221152537.751564-1-msp@xxxxxxxxxxxx v2 - https://lore.kernel.org/lkml/20230125195059.630377-1-msp@xxxxxxxxxxxx stack of calling start_xmit within locked region: [ 308.170171] dump_backtrace+0x0/0x1a0 [ 308.173841] show_stack+0x18/0x70 [ 308.177158] sched_show_task+0x154/0x180 [ 308.181084] dump_cpu_task+0x44/0x54 [ 308.184664] rcu_dump_cpu_stacks+0xe8/0x12c [ 308.188846] rcu_sched_clock_irq+0x9f4/0xd10 [ 308.193118] update_process_times+0x9c/0xec [ 308.197304] tick_sched_handle+0x34/0x60 [ 308.201231] tick_sched_timer+0x4c/0xa4 [ 308.205071] __hrtimer_run_queues+0x138/0x1e0 [ 308.209429] hrtimer_interrupt+0xe8/0x244 [ 308.213440] arch_timer_handler_phys+0x38/0x50 [ 308.217890] handle_percpu_devid_irq+0x84/0x130 [ 308.222422] handle_domain_irq+0x60/0x90 [ 308.226347] gic_handle_irq+0x54/0x130 [ 308.230099] do_interrupt_handler+0x34/0x60 [ 308.234286] el1_interrupt+0x30/0x80 [ 308.237861] el1h_64_irq_handler+0x18/0x24 [ 308.241958] el1h_64_irq+0x78/0x7c [ 308.245360] queued_spin_lock_slowpath+0xf4/0x390 [ 308.250067] m_can_start_tx+0x20/0xb0 [m_can] [ 308.254431] m_can_start_xmit+0xd8/0x230 [m_can] [ 308.259054] dev_hard_start_xmit+0xd4/0x15c [ 308.263241] sch_direct_xmit+0xe8/0x370 [ 308.267080] __qdisc_run+0x118/0x650 [ 308.270660] net_tx_action+0x118/0x230 [ 308.274409] _stext+0x124/0x2a0 [ 308.277549] __irq_exit_rcu+0xe4/0x100 [ 308.281302] irq_exit+0x10/0x20 [ 308.284444] handle_domain_irq+0x64/0x90 [ 308.288367] gic_handle_irq+0x54/0x130 [ 308.292119] call_on_irq_stack+0x2c/0x54 [ 308.296043] do_interrupt_handler+0x54/0x60 [ 308.300228] el1_interrupt+0x30/0x80 [ 308.303804] el1h_64_irq_handler+0x18/0x24 [ 308.307901] el1h_64_irq+0x78/0x7c [ 308.311303] __netif_schedule+0x78/0xa0 [ 308.315138] netif_tx_wake_queue+0x50/0x7c [ 308.319237] m_can_isr+0x474/0x1710 [m_can] [ 308.323425] irq_thread_fn+0x2c/0x9c [ 308.327005] irq_thread+0x178/0x2c0 [ 308.330497] kthread+0x150/0x160 [ 308.333727] ret_from_fork+0x10/0x20 Markus Schneider-Pargmann (16): can: m_can: Remove repeated check for is_peripheral can: m_can: Always acknowledge all interrupts can: m_can: Remove double interrupt enable can: m_can: Disable unused interrupts can: m_can: Keep interrupts enabled during peripheral read can: m_can: Write transmit header and data in one transaction can: m_can: Implement receive coalescing can: m_can: Implement transmit coalescing can: m_can: Add rx coalescing ethtool support can: m_can: Add tx coalescing ethtool support can: m_can: Cache tx putidx can: m_can: Use the workqueue as queue can: m_can: Introduce a tx_fifo_in_flight counter can: m_can: Use tx_fifo_in_flight for netif_queue control can: m_can: Implement BQL can: m_can: Implement transmit submission coalescing drivers/net/can/m_can/m_can.c | 503 ++++++++++++++++++++++++++-------- drivers/net/can/m_can/m_can.h | 35 ++- 2 files changed, 415 insertions(+), 123 deletions(-) -- 2.39.2
