Messages in this thread | | | From | Andy Shevchenko <> | Date | Wed, 28 Jul 2021 14:31:43 +0300 | Subject | Re: LED subsystem lagging maintenance |
| |
On Wed, Jul 28, 2021 at 2:17 PM Pavel Machek <pavel@ucw.cz> wrote:
...
> > > To be honest, patches were not applied because they were not that > > > important to begin with, > > > > Reference counting disbalance is not critical, but what is then? > > Things with end-user impact. What is end-user impact here? How much > memory is leaked in usual config?
Not sure what "usual" means, but if the user has a device in question, then it's struct fwnode_handle (7 pointers + u8, unpacked) per each modprobe. Taking into account that there are usually not so many of the same LED devices in the system and the user rarely does rmmod/insmod cycle, I can say a few dozens of bytes.
-- With Best Regards, Andy Shevchenko
| |