[PATCH] mtdcore: Infrastructure for the device release API

Martin Habets mhabets at solarflare.com
Tue Jul 18 08:51:32 PDT 2017


On 08/07/17 02:47, Brian Norris wrote:
> On Thu, Jun 22, 2017 at 02:54:07PM +0100, Martin Habets wrote:
>> Brian, any opinion about this?
> 
> I'm sure this is all awesome, but I haven't seen a patch that utilizes
> this. It's save me some mental energy of trying to figure out what all
> is supposed to be happening here if I had one.

I can cook up a patch for mtdpart.c, but I cannot test it since I do not have any
hardware that uses this. Or maybe this is enough reason that I should post my
patch for the sfc driver here.
In fact I'll probably post both those patches here.

> Or this can wait until I've spent more time trying to figure out what
> you're trying to do with this. And I haven't proven myself to spend a
> lot of time looking at your patches so far ;)

The easy step is to build a kernel with CONFIG_DEBUG_KOBJECT_RELEASE enabled,
the tricky bit is to analyse the output it spews into dmesg when a module that
uses mtd is unloaded.

Martin



More information about the linux-mtd mailing list