Return-Path: Received: from mail-qg0-f45.google.com ([209.85.192.45]:36068 "EHLO mail-qg0-f45.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755582AbbHZKbI (ORCPT ); Wed, 26 Aug 2015 06:31:08 -0400 Received: by qgeb6 with SMTP id b6so123233408qge.3 for ; Wed, 26 Aug 2015 03:31:07 -0700 (PDT) Date: Wed, 26 Aug 2015 06:31:03 -0400 From: Jeff Layton To: Pratyush Anand Cc: bfields@fieldses.org, rostedt@goodmis.org, linux-nfs@vger.kernel.org, Ingo Molnar , "J. Bruce Fields" , Jeff Layton , linux-kernel@vger.kernel.org (open list), Trond Myklebust Subject: Re: [PATCH 1/2] net: sunrpc: fix tracepoint Warning: unknown op '->' Message-ID: <20150826063103.6f036582@tlielax.poochiereds.net> In-Reply-To: References: MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-nfs-owner@vger.kernel.org List-ID: On Tue, 25 Aug 2015 11:34:19 +0530 Pratyush Anand wrote: > `perf stat -e sunrpc:svc_xprt_do_enqueue true` results in > > Warning: unknown op '->' > Warning: [sunrpc:svc_xprt_do_enqueue] unknown op '->' > > Similar warning for svc_handle_xprt as well. > > Actually TP_printk() should never dereference an address saved in the ring > buffer that points somewhere in the kernel. There's no guarantee that that > object still exists (with the exception of static strings). > > Therefore change all the arguments for TP_printk(), so that it references > values existing in the ring buffer only. > > While doing that, also fix another possible bug when argument xprt could be > NULL and TP_fast_assign() tries to access it's elements. > > Signed-off-by: Pratyush Anand > Cc: Steven Rostedt > --- > include/trace/events/sunrpc.h | 21 ++++++++++++++++----- > 1 file changed, 16 insertions(+), 5 deletions(-) > > diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h > index fd1a02cb3c82..937e482d80e9 100644 > --- a/include/trace/events/sunrpc.h > +++ b/include/trace/events/sunrpc.h > @@ -529,18 +529,24 @@ TRACE_EVENT(svc_xprt_do_enqueue, > > TP_STRUCT__entry( > __field(struct svc_xprt *, xprt) > + __field(struct sockaddr *, addr) > __field(struct svc_rqst *, rqst) > + __field(int, pid) > + __field(unsigned long, flags) > ), > > TP_fast_assign( > __entry->xprt = xprt; > + __entry->addr = > + xprt ? (struct sockaddr *)&xprt->xpt_remote : NULL; I don't get it. It's not safe to save a pointer to xprt and dereference that in a tracepoint (and I understand why that is), but it is safe to save a pointer to a structure embedded inside of xprt? Shouldn't you be saving a copy of the entire sockaddr struct instead? Ahh, ok -- I think I see the confusion. %pIScp does not print the address of the sockaddr, but instead dereferences the pointer and prints it as a formatted address string. See pointer() in lib/vsprintf.c. You do want to save off a copy of the structure instead. > __entry->rqst = rqst; > + __entry->pid = rqst? rqst->rq_task->pid : 0; > + __entry->flags = xprt ? xprt->xpt_flags : 0; > ), > > TP_printk("xprt=0x%p addr=%pIScp pid=%d flags=%s", > __entry->xprt, > - (struct sockaddr *)&__entry->xprt->xpt_remote, > - __entry->rqst ? __entry->rqst->rq_task->pid : 0, > - show_svc_xprt_flags(__entry->xprt->xpt_flags)) > + __entry->addr, __entry->pid, > + show_svc_xprt_flags(__entry->flags)) > ); > > TRACE_EVENT(svc_xprt_dequeue, > @@ -589,16 +595,21 @@ TRACE_EVENT(svc_handle_xprt, > TP_STRUCT__entry( > __field(struct svc_xprt *, xprt) > __field(int, len) > + __field(struct sockaddr *, addr) > + __field(unsigned long, flags) > ), > > TP_fast_assign( > __entry->xprt = xprt; > + __entry->addr = > + xprt ? (struct sockaddr > *)&xprt->xpt_remote : NULL; __entry->len = len; > + __entry->flags = xprt ? xprt->xpt_flags : 0; > ), > > TP_printk("xprt=0x%p addr=%pIScp len=%d flags=%s", > __entry->xprt, > - (struct sockaddr *)&__entry->xprt->xpt_remote, > __entry->len, > - show_svc_xprt_flags(__entry->xprt->xpt_flags)) > + __entry->addr, __entry->len, > + show_svc_xprt_flags(__entry->flags)) > ); > #endif /* _TRACE_SUNRPC_H */ > -- Jeff Layton