Kernel 4.4.109 added pp->link, pp->duplex and pp->speed setters to
mvneta_port_disable() which the mvneta patchset failed to patch out after
rebasing, leading to the following build error:
CC drivers/net/ethernet/marvell/mvneta.o
drivers/net/ethernet/marvell/mvneta.c: In function 'mvneta_port_disable':
drivers/net/ethernet/marvell/mvneta.c:1199:4: error: 'struct mvneta_port' has no member named 'link'
pp->link = 0;
^
drivers/net/ethernet/marvell/mvneta.c:1200:4: error: 'struct mvneta_port' has no member named 'duplex'
pp->duplex = -1;
^
drivers/net/ethernet/marvell/mvneta.c:1201:4: error: 'struct mvneta_port' has no member named 'speed'
pp->speed = 0;
^
Fix the issue by rebasing 134-net-mvneta-convert-to-phylink.patch to remove
these struct member accesses as well.
Fixes: 7f5a040359 ("kernel: update kernel 4.4 to version 4.4.110")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
The hardware queue scheduling is apparently configured with fixed
priorities, which creates a nasty fairness issue where traffic from one
CPU can starve traffic from all other CPUs.
Work around this issue by forcing all tx packets to go through one CPU,
until this issue is fixed properly.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
