Turn pppoe(4) back to kernel lock. We can't predict netlock state within
authormvs <mvs@openbsd.org>
Thu, 14 Jul 2022 11:03:15 +0000 (11:03 +0000)
committermvs <mvs@openbsd.org>
Thu, 14 Jul 2022 11:03:15 +0000 (11:03 +0000)
commitff335594b74dd061d0e7cd1abfb0c5c256a082f7
tree249ac19aeb99409559bc0f88d21975f5425939f8
parent4f290cc7e96086a9796d1b142eff92423ee42caf
Turn pppoe(4) back to kernel lock. We can't predict netlock state within
pppoe_start(), so we can't use it for pppoe(4) data protection. Except
input path, pppoe(4) always accessed with kernel lock held, so grab it
around pppoeintr() too.

Interfaces should not use netlock for their data protection. They should
rely on kernel lock or implement their own.

ok bluhm@ bket@
sys/net/if.c
sys/net/if_pppoe.c