Re: [PATCH 07/11] iwlwifi: add skb address to tx cmd in trace events data

[Kalle Valo <kvalo@codeaurora.org>]

Luca Coelho <luca@coelho.fi> writes:

> On Wed, 2017-12-20 at 17:49 +0200, Kalle Valo wrote:
>> Luca Coelho <luca@coelho.fi> writes:
>> 
>> > From: Mordechay Goodstein <mordechay.goodstein@intel.com>
>> > 
>> > This helps matching tx cmd with other trace events, like
>> > net_dev_xmit
>> > and net_dev_queue etc.
>> > 
>> > Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com>
>> > Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
>> 
>> [...]
>> 
>> > @@ -120,9 +121,9 @@ TRACE_EVENT(iwlwifi_dev_tx,
>> >  				      __get_dynamic_array(buf1),
>> >  				      skb->len - hdr_len);
>> >  	),
>> > -	TP_printk("[%s] TX %.2x (%zu bytes)",
>> > +	TP_printk("[%s] TX %.2x (%zu bytes) skbaddr=%p",
>> >  		  __get_str(dev), ((u8
>> > *)__get_dynamic_array(buf0))[0],
>> > -		  __entry->framelen)
>> > +		  __entry->framelen, __entry->skbaddr)
>> >  );
>> 
>> Just so that you know there has been a lot of changes recently about
>> %p,
>> I think the addresses are nowadays hashed due to security reasons. No
>> idea if it afffects tracing events, but just wanted to point out
>> this.
>
> Yeah, I've seen some discussions and I even commented that in our
> internal review.  But I reckoned that if they are all hash, it should
> still be possible to match them.

Good point, matching is of course possible with hashes.

> It would be good to check if that's done also for traces...

Yeah, please let me know if you find out that.

-- 
Kalle Valo