+++ Steven Rostedt [10/10/19 08:58 -0400]:
On Wed, 9 Oct 2019 22:36:38 -0400
Steven Rostedt wrote:
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -2029,6 +2029,37 @@ static void module_enable_nx(const struct module *mod)
frob_writable_data(&mod->init_layout, set_memory_nx);
}
Also,
On Thu, Oct 10, 2019 at 01:20:13PM -0400, Steven Rostedt wrote:
> Hmm, I'm lost at what the below is trying to do with respect to the
> above.
The below is an alternative approach for the module load issue. It
accepts we patch 'late' and then uses text_poke_bp().
It works. We can then look at mov
On Thu, 10 Oct 2019 19:01:11 +0200
Peter Zijlstra wrote:
> On Thu, Oct 10, 2019 at 10:55:15AM -0400, Steven Rostedt wrote:
>
> > OK, so basically this moves the enabling of function tracing from
> > within the ftrace_module_enable() code without releasing the
> > ftrace_lock mutex.
> >
> > But
On Thu, Oct 10, 2019 at 10:55:15AM -0400, Steven Rostedt wrote:
> OK, so basically this moves the enabling of function tracing from
> within the ftrace_module_enable() code without releasing the
> ftrace_lock mutex.
>
> But we have an issue with the state of the module here, as it is still
> set
On Thu, 10 Oct 2019 10:55:15 -0400
Steven Rostedt wrote:
I'll try to add more detail (mapping to functions)
>
> OK, so basically this moves the enabling of function tracing from
> within the ftrace_module_enable() code without releasing the
> ftrace_lock mutex.
>
> But we have an issue with th
On Thu, 10 Oct 2019 10:55:15 -0400
Steven Rostedt wrote:
> But we have an issue with the state of the module here, as it is still
> set as MODULE_STATE_UNFORMED. Let's look at what happens if we have:
>
>
> CPU0CPU1
>
On Thu, 10 Oct 2019 14:29:09 +0200
Peter Zijlstra wrote:
> Yes, your code is anal about checking the NOPs, so you first have to
Yep, because being paranoid about modifying code is always good ;-)
> write NOPs before you can write CALLs, if it is enabled. But afaict you
> really can do all that
On Thu, Oct 10, 2019 at 08:50:11AM -0400, Steven Rostedt wrote:
> On Thu, 10 Oct 2019 11:36:50 +0200
> Peter Zijlstra wrote:
> > > load_module()
> > > ...
> > > complete_formation()
> > > mutex_lock(&module_mutex);
> > > ...
> > > module_enable_ro();
> > > module_enable_nx();
On Wed, 9 Oct 2019 22:36:38 -0400
Steven Rostedt wrote:
> --- a/kernel/module.c
> +++ b/kernel/module.c
> @@ -2029,6 +2029,37 @@ static void module_enable_nx(const struct module *mod)
> frob_writable_data(&mod->init_layout, set_memory_nx);
> }
>
Also, if you are worried about these bein
On Thu, 10 Oct 2019 11:36:50 +0200
Peter Zijlstra wrote:
> > > > The reason is that ftrace_module_enable() is called after the module
> > > > has set its text to read-only. There's subtle reasons that this needs
> > > > to be called afterward, and we need to continue to do so.
> > >
> > > Plea
On Thu, Oct 10, 2019 at 11:36:50AM +0200, Peter Zijlstra wrote:
> On Thu, Oct 10, 2019 at 11:33:29AM +0200, Peter Zijlstra wrote:
> > I don't see any reason what so ever..
> >
> > load_module()
> > ...
> > complete_formation()
> > mutex_lock(&module_mutex);
> > ...
> > module_enab
On Thu, Oct 10, 2019 at 11:33:29AM +0200, Peter Zijlstra wrote:
> On Thu, Oct 10, 2019 at 09:31:21AM +0200, Peter Zijlstra wrote:
> > On Wed, Oct 09, 2019 at 10:36:38PM -0400, Steven Rostedt wrote:
> > > From: Steven Rostedt (VMware)
> > >
> > > In the process of using text_poke_bp() for ftrace o
On Thu, Oct 10, 2019 at 09:31:21AM +0200, Peter Zijlstra wrote:
> On Wed, Oct 09, 2019 at 10:36:38PM -0400, Steven Rostedt wrote:
> > From: Steven Rostedt (VMware)
> >
> > In the process of using text_poke_bp() for ftrace on x86, when
> > performing the following action:
> >
> > # rmmod snd_hda
On Thu, Oct 10, 2019 at 09:31:21AM +0200, Peter Zijlstra wrote:
> On Wed, Oct 09, 2019 at 10:36:38PM -0400, Steven Rostedt wrote:
> > From: Steven Rostedt (VMware)
> >
> > In the process of using text_poke_bp() for ftrace on x86, when
> > performing the following action:
> >
> > # rmmod snd_hda
On Wed, Oct 09, 2019 at 10:36:38PM -0400, Steven Rostedt wrote:
> From: Steven Rostedt (VMware)
>
> In the process of using text_poke_bp() for ftrace on x86, when
> performing the following action:
>
> # rmmod snd_hda_codec_hdmi
> # echo function > /sys/kernel/tracing/current_tracer
> # modpr
From: Steven Rostedt (VMware)
In the process of using text_poke_bp() for ftrace on x86, when
performing the following action:
# rmmod snd_hda_codec_hdmi
# echo function > /sys/kernel/tracing/current_tracer
# modprobe snd_hda_codec_hdmi
It triggered this:
BUG: unable to handle page fault fo
16 matches
Mail list logo