On Sat, Mar 20, 2004 at 06:49:09PM +0100, Jean Delvare wrote: > Hi all, > > I have a question about the locking mechanism we have in all our i2c > client drivers, 2.4 and 2.6 as well. > > The data->update_lock semaphore is used in the *_update_dev function, > and only there. There are various other places all over the drivers > where we access the same internal data, (in get callback functions) and > even write to it (in store callback functions typically), and also > access the i2c bus (in store functions too). > > Why is it that we don't use the lock here? > > What are we trying to protect exactly? I remember trying to figure this out a while ago too, and didn't come up with an answer. I sent in some patches that I thought were correct, but others said no. It would be good to figure this out. thanks, greg k-h
