From: Mitch Williams Date: Wed, 4 Jun 2014 08:45:19 +0000 (+0000) Subject: i40evf: invite vector 0 to the interrupt party X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=164ec1bfa13d34834a72b2ace3ecb521234f444f;p=GitHub%2FLineageOS%2FG12%2Fandroid_kernel_amlogic_linux-4.9.git i40evf: invite vector 0 to the interrupt party The i40evf_irq_enable and i40evf_fire_sw_interrupt functions were unfairly discriminating against MSI-X vector 0, just because it doesn't handle traffic. That doesn't mean it's not essential to the operation of the driver. This change allows the watchdog to fire vector 0 via software, which makes the driver tolerant of dropped interrupts on that vector. Buck up, vector 0! You can be part of our gang! Change-ID: I37131d955018a6b3e711e1732d21428acd0d767e Signed-off-by: Mitch Williams Tested-by: Jim Young Signed-off-by: Jeff Kirsher --- diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c index 5f29e58bf5ea..6186149a279e 100644 --- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c +++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c @@ -260,6 +260,12 @@ static void i40evf_fire_sw_int(struct i40evf_adapter *adapter, int i; uint32_t dyn_ctl; + if (mask & 1) { + dyn_ctl = rd32(hw, I40E_VFINT_DYN_CTL01); + dyn_ctl |= I40E_VFINT_DYN_CTLN_SWINT_TRIG_MASK | + I40E_VFINT_DYN_CTLN_CLEARPBA_MASK; + wr32(hw, I40E_VFINT_DYN_CTL01, dyn_ctl); + } for (i = 1; i < adapter->num_msix_vectors; i++) { if (mask & (1 << i)) { dyn_ctl = rd32(hw, I40E_VFINT_DYN_CTLN1(i - 1)); @@ -278,6 +284,7 @@ void i40evf_irq_enable(struct i40evf_adapter *adapter, bool flush) { struct i40e_hw *hw = &adapter->hw; + i40evf_misc_irq_enable(adapter); i40evf_irq_enable_queues(adapter, ~0); if (flush)