Mirror of the lttng-dev mailing list
 help / color / mirror / Atom feed
From: simon.marchi@ericsson.com (Simon Marchi)
Subject: [lttng-dev] [PATCH lttng-modules] Add support for i2c tracepoints
Date: Tue, 4 Oct 2016 15:48:34 -0400	[thread overview]
Message-ID: <f09f33cd-54d1-a8e9-2fbb-46a62a77ff89@ericsson.com> (raw)
In-Reply-To: <1386914197.45790.1475602673694.JavaMail.zimbra@efficios.com>

On 16-10-04 01:37 PM, Mathieu Desnoyers wrote:
> Perhaps we could rename the option to "trace_sensitive_payload" ?

Of all those:

- trace_sensitive_data
- trace_sensitive_payload
- allow_sensitive_payload_extraction
- extract_sensitive_payload

I prefer "extract_sensitive_payload".  But it's your call, just tell me which
one you prefer.

> I think we should keep it at the probe level, like you do in your prototype.
> 
>>
>>
>> From 69aa1c3d73b5b0dbf031b90c6e38d4b2a0ad8699 Mon Sep 17 00:00:00 2001
>> From: Simon Marchi <simon.marchi@ericsson.com>
>> Date: Tue, 4 Oct 2016 11:38:19 -0400
>> Subject: [PATCH] Provide knob to control tracing of i2c buffer contents
>>
>> ---
>> instrumentation/events/lttng-module/i2c.h | 6 ++++--
>> probes/lttng-probe-i2c.c                  | 7 +++++++
>> 2 files changed, 11 insertions(+), 2 deletions(-)
>>
>> diff --git a/instrumentation/events/lttng-module/i2c.h
>> b/instrumentation/events/lttng-module/i2c.h
>> index 68ce80d..1ede276 100644
>> --- a/instrumentation/events/lttng-module/i2c.h
>> +++ b/instrumentation/events/lttng-module/i2c.h
>> @@ -26,7 +26,8 @@ LTTNG_TRACEPOINT_EVENT(i2c_write,
>> 		ctf_integer(__u16, addr, msg->addr)
>> 		ctf_integer(__u16, flags, msg->flags)
>> 		ctf_integer(__u16, len, msg->len)
>> -		ctf_sequence_hex(__u8, buf, msg->buf, __u16, msg->len)
>> +		ctf_sequence_hex(__u8, buf, trace_sensitive_data ? msg->buf : NULL,
>> +				 __u16, trace_sensitive_data ? msg->len : 0)
> 
> Since "trace_sensitive_data" can be changed at runtime concurrently
> with tracing, we should load it with LOAD_ONCE(), keep it in a local
> variable (TRACEPOINT_EVENT_CODE, TP_locvar and TP_code), and use that
> in as argument. Otherwise the compiler can perform two fetch of that
> variable, and its value may change in between, thus leading to a
> NULL pointer dereference (OOPS).

Ahh good catch.  I'll look up those things you mentioned, but do you
have a usage pattern in mind that we could use here?  Where would the
local variable be?  I know of gcc's statement expression, we could use
that, but do you have something else in mind?

>> 	)
>> )
>>
>> @@ -63,7 +64,8 @@ LTTNG_TRACEPOINT_EVENT(i2c_reply,
>> 		ctf_integer(__u16, addr, msg->addr)
>> 		ctf_integer(__u16, flags, msg->flags)
>> 		ctf_integer(__u16, len, msg->len)
>> -		ctf_sequence_hex(__u8, buf, msg->buf, __u16, msg->len)
>> +		ctf_sequence_hex(__u8, buf, trace_sensitive_data ? msg->buf : NULL,
>> +				 __u16, trace_sensitive_data ? msg->len : 0)
>> 	)
>> )
>>
> 
> Same here.
> 
> Can you also fold it with your i2c patch ?

I'll send a v2 of the original patch once it's clear how to do the load once thingy.

Thanks!

Simon



  parent reply	other threads:[~2016-10-04 19:48 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-10-03 18:21 Simon Marchi
2016-10-03 22:21 ` Mathieu Desnoyers
2016-10-04 17:08   ` Simon Marchi
2016-10-04 17:37     ` Mathieu Desnoyers
2016-10-04 19:19       ` Mathieu Desnoyers
2016-10-04 19:48       ` Simon Marchi [this message]
2016-10-04 21:21         ` Mathieu Desnoyers

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f09f33cd-54d1-a8e9-2fbb-46a62a77ff89@ericsson.com \
    --to=simon.marchi@ericsson.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox