On Tue, 25 Jan 2022 10:01:38 -0500
Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> On Tue, 25 Jan 2022 16:21:46 +1100
> Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx> wrote:
>
> > Hi all,
> >
> > On Tue, 25 Jan 2022 16:05:05 +1100 Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx> wrote:
> > >
> > > My qemu boot test of a powerpc pseries_le_defconfig kernel produces the
> > > following trace:
> > >
> > > ------------[ cut here ]------------
> > > WARNING: CPU: 0 PID: 0 at kernel/trace/trace_events.c:417 trace_event_raw_init+0x194/0x730
> > > Modules linked in:
> > > CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.17.0-rc1 #2
> > > NIP: c0000000002bdbb4 LR: c0000000002bdcb0 CTR: c0000000002bdb70
> > >
> > > I have no idea what has caused this :-( Maybe commit
> > >
> > > 5544d5318802 ("SUNRPC: Same as SVC_RQST_ENDPOINT, but without the xid")
> >
> > Actually, reverting commits
> >
> > 6ff851d98af8 ("SUNRPC: Improve sockaddr handling in the svc_xprt_create_error trace point")
> > 5544d5318802 ("SUNRPC: Same as SVC_RQST_ENDPOINT, but without the xid")
> > e2d3613db12a ("SUNRPC: Record endpoint information in trace log")
> >
> > makes the warning go away.
> >
>
> We added a new way to save items on the ring buffer, but did not update the
> safety checks to know about them. I'll fix this shortly.
>
This should fix it:
I'll make it a real patch and start running it through my tests.
-- Steve
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 3147614c1812..f527ae807e77 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -384,6 +384,12 @@ static void test_event_printk(struct trace_event_call *call)
if (!(dereference_flags & (1ULL << arg)))
goto next_arg;
+ /* Check for __get_sockaddr */;
+ if (str_has_prefix(fmt + i, "__get_sockaddr(")) {
+ dereference_flags &= ~(1ULL << arg);
+ goto next_arg;
+ }
+
/* Find the REC-> in the argument */
c = strchr(fmt + i, ',');
r = strstr(fmt + i, "REC->");
