On Wed, Aug 16, 2023 at 4:01 PM Johannes Berg <johannes@xxxxxxxxxxxxxxxx> wrote: > > On Wed, 2023-08-16 at 15:39 +0200, Aleksandr Nogikh wrote: > > On Wed, Aug 16, 2023 at 10:09 AM Johannes Berg > > <johannes@xxxxxxxxxxxxxxxx> wrote: > > > > > > Hi, > > > > > > Yeah, I know this is old ... still happening though, so I've been > > > looking at it a bit. > > > > > > > The issue was bisected to: > > > > > > > > commit 7dfd8ac327301f302b03072066c66eb32578e940 > > > > Author: Ramon Fontes <ramonreisfontes@xxxxxxxxx> > > > > Date: Thu Oct 10 18:13:07 2019 +0000 > > > > > > > > mac80211_hwsim: add support for OCB > > > > > > I'm not sure that make sense, FWIW. There isn't even any OCB? The syz > > > script just contains a channel switch command. > > > > > > > WARNING: CPU: 1 PID: 6900 at net/mac80211/tx.c:4875 __ieee80211_beacon_get+0xb59/0x1aa0 net/mac80211/tx.c:4875 > > > > > > I also can't reproduce this though. > > > > Did you try to run the kernel attached in the assets? > > https://github.com/google/syzkaller/blob/master/docs/syzbot_assets.md#run-a-c-reproducer > > > > I've just followed the instructions from there and the C repro did > > crash the kernel in ~20 seconds: > > > > [ 56.809692][ C1] ------------[ cut here ]------------ > > [ 56.810656][ C1] WARNING: CPU: 1 PID: 5358 at > > net/mac80211/tx.c:5011 __ieee80211_beacon_get+0x1495/0x16e0 > > > > To be fair, I didn't, I figured the reproducer was simple enough to just > have a go at it with my own test infra. > > Is there an easy way to rebuild the kernel for it? How does an easy way differ from a difficult one in this case? :) I've just built the v6.5-rc6 kernel with the https://syzkaller.appspot.com/text?tag=KernelConfig&x=3e670757e16affb config and run the C repro mentioned above from root. It crashed the kernel: [ 78.057333][ C0] ------------[ cut here ]------------ [ 78.058289][ C0] WARNING: CPU: 0 PID: 5377 at net/mac80211/tx.c:5011 __ieee80211_beacon_get+0x1495/0x16e0 -- Aleksandr > > johannes
