Hi Andy, > > > I'm wondering if synchronize_irq() is enough. The free_irq() theoretically is > > > the best option, though I dunno which one suits in which cases better. > > > > In which scenario do you think synchronize_irq() is not enough? > > I think if the driver is using tasklets this is not enough. However, I might > miss the context in i2c case. Do you mean 'threaded irqs' here? If so, synchronize_irq() should be safe because synchronize_hardirq() is the function for not waiting for threaded handlers. If you mean something else, currently no I2C bus master driver is using a separate tasklet, so I'd think synchronize_irq() is a sufficient choice? But still open for criticism. I want the best solution! Kind regards, Wolfram
Attachment:
signature.asc
Description: PGP signature
