Re: possible trace_printk() bug in v5.19-rc1

[Chuck Lever III]

> On Jun 27, 2022, at 1:11 PM, Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> 
> On Mon, 27 Jun 2022 17:08:03 +0000
> Chuck Lever III <chuck.lever@xxxxxxxxxx> wrote:
> 
>> Updated from your copr repo. The problem persists.
> 
> The symbol lookup problem still exists. What about the get_sockaddr() not
> processing?

"trace-cmd report" no longer produces the get_sockaddr warning messages,
but tracepoints that use __get_sockaddr() still FAIL TO PARSE on my
system:

            nfsd-1167  [005]   117.853235: nfsd_cb_probe:        [FAILED TO PARSE] state=0x1 cl_boot=1656349219 cl_id=3054917767 addr=ARRAY[02, 00, 00, 00, c0, a8, 02, 43, 00, 00, 00, 00, 00, 00, 00, 00]
   kworker/u24:2-985   [003]   117.853368: nfsd_cb_setup:        [FAILED TO PARSE] cl_boot=1656349219 cl_id=3054917767 authflavor=0x1 addr=ARRAY[02, 00, 00, 00, c0, a8, 02, 43, 00, 00, 00, 00, 00, 00, 00, 00] netid=rdma
   kworker/u24:2-985   [003]   117.853370: nfsd_cb_state:        [FAILED TO PARSE] state=0x0 cl_boot=1656349219 cl_id=3054917767 addr=ARRAY[02, 00, 00, 00, c0, a8, 02, 43, 00, 00, 00, 00, 00, 00, 00, 00]


> Would you be able to send me a trace.dat file that has this issue?

Sending under separate cover.


--
Chuck Lever