linux-kernel - Re: [PATCH 4/9] perf session: Create kernel maps in the constructor

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 for Android: free password hash cracker in your pocket

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

Date:	Fri, 29 Jan 2010 02:01:58 -0500
From:	Masami Hiramatsu <mhiramat@...hat.com>
To:	Arnaldo Carvalho de Melo <acme@...radead.org>
CC:	Ingo Molnar <mingo@...e.hu>, linux-kernel@...r.kernel.org,
	Frédéric Weisbecker <fweisbec@...il.com>,
	Mike Galbraith <efault@....de>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Paul Mackerras <paulus@...ba.org>
Subject: Re: [PATCH 4/9] perf session: Create kernel maps in the constructor

Arnaldo Carvalho de Melo wrote:
> Em Thu, Jan 28, 2010 at 03:59:38PM -0500, Masami Hiramatsu escreveu:
>>> you'll get ipv6.ko located and loaded, mod->dso->long_name thus will
>>> have the file to use with libdwarf and sym->start will have the
>>> unrelocated address, and mod->start will have where ipv6 is loaded in
>>> kernel space
> 
>> Oh, nice :-) Is that available for the modules which aren't loaded?
> 
> No it isn't, what usecase do you see where you would like to look at non
> loaded modules? Traversing the /lib/modules passed can be made to just
> create DSOs, not maps if we need it. Maps when created using one of these
> DSOs would find them on the list.

One usecase I thought is probing module init code for checking/debugging
driver initialization routine. For this purpose, build-id checking also
should be solved. If the module is not loaded, we can't check build-id
in that time. I think we can compare it when loading the module,
in kernel, if we can pass target build-id to kprobe-tracer.


>> And yeah, I'd like to support modules and it will requires some
>> enhancement in kprobe-tracer too.
> 
> Ok
> 
> <SNIP>
>>> With the recent modifications done to untie the session and symbol
>>> layers, 'perf probe' now can use just the symbols layer.
> 
>> Could you remove session.kmap too?
>> other parts look good to me.
> 
> I'll haven't done that just to keep the patch as small as possible :-)
> But will do before submitting.

Thanks! :-)

-- 
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division

e-mail: mhiramat@...hat.com

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/