On 2022-04-05, Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote: > From the list of patches and the only one I am Cc'ed to I can't find > the answer to my question, It seems lore is having trouble picking up the series. Until then, I've pushed the series to github.com so you can see the code. > i.e. does it take care about console_unregister() cases at run time? > (We have some drivers to call ->exit() for console in that case, > perhaps it should do something before calling it.) Yes. In console_unregister(), the printing thread is shutdown before the actual console unregistration takes place. The thread is shutdown here: https://github.com/Linutronix/linux/blob/b27eb6c3d168d608b3a9d04578e030a3d79d889a/kernel/printk/printk.c#L3393 The console's exit() callback is below that: https://github.com/Linutronix/linux/blob/b27eb6c3d168d608b3a9d04578e030a3d79d889a/kernel/printk/printk.c#L3429 > Would be nice to see some pointers where I can find the answer and > maybe even describing in the cover-letter/commit > message/documentation. I suppose for the console thread patch [0] I could add a paragraph to the commit message mentioning kernel thread shutdown for console unregistration. [0] https://github.com/Linutronix/linux/commit/e26ae677da2e339dd268c1f871b81e61e782393f John
