linux-kernel - Re: [PATCH 02/14] perf trace: Apply new perf_mmap__read

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <44636157-9aaf-4e50-c12a-a2e60e01e415@linux.intel.com>
Date:   Mon, 5 Mar 2018 09:50:10 -0500
From:   "Liang, Kan" <kan.liang@...ux.intel.com>
To:     Arnaldo Carvalho de Melo <acme@...nel.org>,
        Jiri Olsa <jolsa@...hat.com>
Cc:     Ingo Molnar <mingo@...nel.org>, linux-kernel@...r.kernel.org,
        namhyung@...nel.org, wangnan0@...wei.com, ak@...ux.intel.com
Subject: Re: [PATCH 02/14] perf trace: Apply new perf_mmap__read_event()
 interface



On 3/5/2018 8:46 AM, Arnaldo Carvalho de Melo wrote:
> Em Mon, Mar 05, 2018 at 10:03:39AM -0300, Arnaldo Carvalho de Melo escreveu:
>> Em Sat, Mar 03, 2018 at 12:30:22AM +0100, Jiri Olsa escreveu:
>>> On Thu, Mar 01, 2018 at 06:08:59PM -0500, kan.liang@...ux.intel.com wrote:
>>>> From: Kan Liang <kan.liang@...ux.intel.com>
>>>> The perf trace still use the legacy interface.
>>>> +++ b/tools/perf/builtin-trace.c
>>>> @@ -2472,8 +2472,14 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
>>>>   
>>>>   	for (i = 0; i < evlist->nr_mmaps; i++) {
>>>>   		union perf_event *event;
>>>> +		struct perf_mmap *md;
>>>> +		u64 end, start;
>>>>   
>>>> -		while ((event = perf_evlist__mmap_read(evlist, i)) != NULL) {
>>>> +		md = &evlist->mmap[i];
>>>> +		if (perf_mmap__read_init(md, 0, &start, &end) < 0)
>>>> +			continue;
>>>> +
>>>> +		while ((event = perf_mmap__read_event(md, 0, &start, end)) != NULL) {
>>>>   			struct perf_sample sample;
>>>>   
>>>>   			++trace->nr_events;
>>>> @@ -2486,7 +2492,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
>>>>   
>>>>   			trace__handle_event(trace, event, &sample);
>>>>   next_event:
>>>> -			perf_evlist__mmap_consume(evlist, i);
>>>> +			perf_mmap__consume(md, 0);
>>>
>>> could you call this with 'false' instead of 0, it's 'bool overwrite'
>>> applies also to the rest of the patchset
>>
>> I'm doing this, the argument is 'bool', so the value should be false or
>> true, even '0' being way shorter...
> 
> While doing that I wonder why is that we can't do it without those
> explicit start/end variables in all the call sites and passing
> overwrite, start and end as parameters...
> 
> Can't we just add 'overwrite, 'start' and 'end' to struct perf_mmap,
> then have perf_mmap__read_init() with just 'md' and 'overwrite' as
> parameters, initialize md->{start,end} and set md->overwrite to the
> 'overwrite' parameter in 'perf_mmap__read_init()' and then use just
> md as parameters for both perf_mmap__read_event() and
> perf_mmap__consume()?
>

I don't see any reason we cannot do this.
I will do some test. If it works well, I will submit the follow-up patch 
to clean up the interface.

Thanks,
Kan

> What am I missing to have this much simpler without all this
> boilerplate?
> 
> Anyway, finishing the s/0/false/g, will leave this for later, but please
> comment on it.
> 
> - Arnaldo
>