On Tue, Apr 20, 2021 at 3:48 PM Guenter Roeck [off-list ref] wrote:

On 4/20/21 2:43 AM, Eric Dumazet wrote:

quoted

Unfortunately that doesn't fix the problem for me. With this patch applied
on top of next-20210419, I still get the same crash as before:

udhcpc: sending discover^M
Unable to handle kernel paging request at virtual address 0000000000000004^M
udhcpc(169): Oops -1^M
pc = [<0000000000000004>]  ra = [<fffffc0000b8c5b8>]  ps = 0000    Not tainted^M
pc is at 0x4^M
ra is at napi_gro_receive+0x68/0x150^M
v0 = 0000000000000000  t0 = 0000000000000008  t1 = 0000000000000000^M
t2 = 0000000000000000  t3 = 000000000000000e  t4 = 0000000000000038^M
t5 = 000000000000ffff  t6 = fffffc00002f298a  t7 = fffffc0002c78000^M
s0 = fffffc00010b3ca0  s1 = 0000000000000000  s2 = fffffc00011267e0^M
s3 = 0000000000000000  s4 = fffffc00025f2008  s5 = fffffc00002f2940^M
s6 = fffffc00025f2040^M
a0 = fffffc00025f2008  a1 = fffffc00002f2940  a2 = fffffc0002ca000c^M
a3 = fffffc00000250d0  a4 = 0000000effff0008  a5 = 0000000000000000^M
t8 = fffffc00010b3c80  t9 = fffffc0002ca04cc  t10= 0000000000000000^M
t11= 00000000000004c0  pv = fffffc0000b8bc40  at = 0000000000000000^M
gp = fffffc00010f9fb8  sp = 00000000df74db09^M
Disabling lock debugging due to kernel taint^M
Trace:^M
[<fffffc0000b8c5b8>] napi_gro_receive+0x68/0x150^M
[<fffffc00009b409c>] receive_buf+0x50c/0x1b80^M
[<fffffc00009b58b8>] virtnet_poll+0x1a8/0x5b0^M
[<fffffc00009b58ec>] virtnet_poll+0x1dc/0x5b0^M
[<fffffc0000b8d17c>] __napi_poll+0x4c/0x270^M
[<fffffc0000b8d670>] net_rx_action+0x130/0x2c0^M
[<fffffc0000bd6cb0>] sch_direct_xmit+0x170/0x360^M
[<fffffc0000bd7000>] __qdisc_run+0x160/0x6c0^M
[<fffffc0000337b64>] do_softirq+0xa4/0xd0^M
[<fffffc0000337ca4>] __local_bh_enable_ip+0x114/0x120^M
[<fffffc0000b89554>] __dev_queue_xmit+0x484/0xa60^M
[<fffffc0000cd072c>] packet_sendmsg+0xe7c/0x1ba0^M
[<fffffc0000b53338>] __sys_sendto+0xf8/0x170^M
[<fffffc0000cfec18>] _raw_spin_unlock+0x18/0x30^M
[<fffffc0000a9bf7c>] ehci_irq+0x2cc/0x5c0^M
[<fffffc0000a71334>] usb_hcd_irq+0x34/0x50^M
[<fffffc0000b521bc>] move_addr_to_kernel+0x3c/0x60^M
[<fffffc0000b532e4>] __sys_sendto+0xa4/0x170^M
[<fffffc0000b533d4>] sys_sendto+0x24/0x40^M
[<fffffc0000cfea38>] _raw_spin_lock+0x18/0x30^M
[<fffffc0000cfec18>] _raw_spin_unlock+0x18/0x30^M
[<fffffc0000325298>] clipper_enable_irq+0x98/0x100^M
[<fffffc0000cfec18>] _raw_spin_unlock+0x18/0x30^M
[<fffffc0000311514>] entSys+0xa4/0xc0^M

OK, it would be nice if you could get line number from this stack trace.

(scripts/decode_stacktrace.sh is your friend)