linux-toradex.git/kernel/module.c, branch v2.6.25-rc2

Linux Kernel Markers: support multiple probes

2008-02-14T00:21:20+00:00

RCU style multiple probes support for the Linux Kernel Markers.  Common case
(one probe) is still fast and does not require dynamic allocation or a
supplementary pointer dereference on the fast path.

- Move preempt disable from the marker site to the callback.

Since we now have an internal callback, move the preempt disable/enable to the
callback instead of the marker site.

Since the callback change is done asynchronously (passing from a handler that
supports arguments to a handler that does not setup the arguments is no
arguments are passed), we can safely update it even if it is outside the
preempt disable section.

- Move probe arm to probe connection. Now, a connected probe is automatically
  armed.

Remove MARK_MAX_FORMAT_LEN, unused.

This patch modifies the Linux Kernel Markers API : it removes the probe
"arm/disarm" and changes the probe function prototype : it now expects a
va_list * instead of a "...".

If we want to have more than one probe connected to a marker at a given
time (LTTng, or blktrace, ssytemtap) then we need this patch. Without it,
connecting a second probe handler to a marker will fail.

It allow us, for instance, to do interesting combinations :

Do standard tracing with LTTng and, eventually, to compute statistics
with SystemTAP, or to have a special trigger on an event that would call
a systemtap script which would stop flight recorder tracing.

Signed-off-by: Mathieu Desnoyers 
Cc: Christoph Hellwig 
Cc: Mike Mason 
Cc: Dipankar Sarma 
Cc: David Smith 
Cc: "Paul E. McKenney" 
Cc: "Frank Ch. Eigler" 
Cc: Steven Rostedt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

fix "modules: make module_address_lookup() safe"

2008-02-08T17:22:24+00:00

Get the constness right, avoid nasty cast.

Cc: Ingo Molnar 
Cc: Kyle McMartin 
Cc: Rusty Russell 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

modules: include sections.h to avoid defining linker variables explicitly

2008-02-08T17:22:24+00:00

module.c should not define linker variables on its own. We have an include
file for that.

Signed-off-by: Christoph Lameter 
Cc: Rusty Russell 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

Modules: handle symbols that have a zero value

2008-02-08T17:22:24+00:00

The module subsystem cannot handle symbols that are zero.  If symbols are
present that have a zero value then the module resolver prints out a
message that these symbols are unresolved.

[akinobu.mita@gmail.com: fix __find_symbl() error checks]
Cc: Mathieu Desnoyers 
Cc: Kay Sievers 
Cc: Rusty Russell 
Cc: Andi Kleen 
Signed-off-by: Akinobu Mita 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

x86/non-x86: percpu, node ids, apic ids x86.git fixup

2008-01-30T12:33:32+00:00

Signed-off-by: Ingo Molnar 
Signed-off-by: Thomas Gleixner

Module: check to see if we have a built in module with the same name

2008-01-29T06:13:27+00:00

When trying to load a module with the same name as a built-in one, a
scary kobject backtrace comes up.  Prevent that from checking for this
condition and warning the user as to what exactly is going on.

Cc: Rusty Russell 
Cc: Linus Torvalds 
Cc: Andrew Morton 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Rusty Russell

module: add module taint on ndiswrapper

2008-01-29T06:13:26+00:00

The struct module taints member is supposed to store per-module taint
data. The kernel knows about certain specific external modules that will
taint the kernel, such as ndiswrapper. Use of ndiswrapper possibly
should set the per-module taint in addition to the global kernel
taint flag, unless we're arguing not because wrapper module itself
is not what actually causes the kernel to be tainted as such?

Signed-off-by: Jon Masters 
Signed-off-by: Rusty Russell

module: make module_address_lookup safe

2008-01-29T06:13:23+00:00

module_address_lookup releases preemption then returns a pointer into
the module space.  The only user (kallsyms) copies the result, so just
do that under the preempt disable.

Signed-off-by: Rusty Russell

module: better OOPS and lockdep coverage for loading modules

2008-01-29T06:13:22+00:00

If we put the module in the linked list *before* calling into to, we
get the module name and functions in the OOPS (is_module_address can
find the module).  It also helps lockdep in a similar way.

Acked-and-tested-by: Joern Engel 
Tested-by: Erez Zadok 
Signed-off-by: Rusty Russell

module: Fix gratuitous sprintf in module.c

2008-01-29T06:13:21+00:00

Andrew sent an older version of this patch: we shouldn't use sprintf
to copy a string.

Signed-off-by: Rusty Russell