Re: general protection fault in dev_map_hash_update_elem

Hillf Danton [off-list ref] writes:

quoted

syzbot has found a reproducer for the following crash on Sat, 07 Sep 2019 18:59:06 -0700

HEAD commit:    a2c11b03 kcm: use BPF_PROG_RUN
git tree:       bpf-next
console output: https://syzkaller.appspot.com/x/log.txt?x=13d46ec1600000
kernel config:  https://syzkaller.appspot.com/x/.config?x=cf0c85d15c20ade3
dashboard link: https://syzkaller.appspot.com/bug?extid=4e7a85b1432052e8d6f8
compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1220b2d1600000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=1360b26e600000

general protection fault: 0000 [#1] PREEMPT SMP KASAN
CPU: 1 PID: 10210 Comm: syz-executor910 Not tainted 5.3.0-rc7+ #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS  
Google 01/01/2011
RIP: 0010:__write_once_size include/linux/compiler.h:226 [inline]
RIP: 0010:__hlist_del include/linux/list.h:762 [inline]
RIP: 0010:hlist_del_rcu include/linux/rculist.h:455 [inline]
RIP: 0010:__dev_map_hash_update_elem kernel/bpf/devmap.c:668 [inline]
RIP: 0010:dev_map_hash_update_elem+0x3c8/0x6e0 kernel/bpf/devmap.c:691

Fix commit 6f9d451ab1a3 ("xdp: Add devmap_hash map type for looking
up devices by hashed index")

While this minimal patch does fix the bug (as Jesper already noted), I
prefer to rework the logic instead of just repeating the lookup; a patch
is on its way :)

-Toke