From 37dbe302158c7c567835ed64c118236dfc0425b2 Mon Sep 17 00:00:00 2001 From: Rosen Penev Date: Tue, 28 Mar 2023 15:29:19 -0700 Subject: [PATCH] ramips: mt7621: remove set-affinity script From https://github.com/openwrt/openwrt/pull/12280#issuecomment-1489279860 On Ethernet and WLAN, NAPI is threaded for all queues. This means that the processing work is not stuck on the CPU that fired the IRQ. Under heavy load, IRQs get disabled anyway, so it should not matter at all which CPUs the IRQs fire on. Basic testing indicates this to be true. There's no speedup or slowdown. Signed-off-by: Rosen Penev --- .../base-files/etc/init.d/set-irq-affinity | 19 ------------------- 1 file changed, 19 deletions(-) delete mode 100755 target/linux/ramips/mt7621/base-files/etc/init.d/set-irq-affinity diff --git a/target/linux/ramips/mt7621/base-files/etc/init.d/set-irq-affinity b/target/linux/ramips/mt7621/base-files/etc/init.d/set-irq-affinity deleted file mode 100755 index c118d928a7..0000000000 --- a/target/linux/ramips/mt7621/base-files/etc/init.d/set-irq-affinity +++ /dev/null @@ -1,19 +0,0 @@ -#!/bin/sh /etc/rc.common - -START=99 - -start() { - if grep -q 'processor.*: 2' /proc/cpuinfo; then - mask=4 - elif grep -q 'processor.*: 1' /proc/cpuinfo; then - mask=2 - else - return - fi - - for irq in $(grep "mt76..e" /proc/interrupts | cut -d: -f1 | sed 's, *,,') - do - echo "$mask" > "/proc/irq/$irq/smp_affinity" - [ $mask = 4 ] && mask=8 - done -}