Messages in this thread | | | Date | Wed, 13 Feb 2013 10:47:00 -0800 | From | Greg KH <> | Subject | Re: [PATCH v2 8/9] pps: Use a single cdev |
| |
On Wed, Feb 13, 2013 at 01:35:29PM -0500, George Spelvin wrote: > > You forgot a Signed-off-by: line for this patch, so I can't apply it, or > > the 9/9 patch :( > > Oops, fixed. I don't see why the 9/9 patch depends on it, > though. They're not related or interdependent in any way. > > If you want to check the logic, I'd appreciate it. I'm not > really sure about the RCU stuff. My understanding is that: > - the idr code does the appropriate write locking when > modifying itself, so I don't need to do any. > - The pps_device returned from idr_find is itself refcounted, > so it can't go away, and the accesses don't have bo be > inside the RCU read "lock". It's only the IDR's internal > index nodes that might get reallocated by modificaitons of > a different part of the tree. > > > Care to resend just these two after fixing this up? > > I can, but if you think you need 9/9 resent (which *did* have a S-o-b), > I'm confused and wonder why...
I stopped at that point in the series, that's the only reason why, I didn't "check" to see if there was a dependancy, I just assumed there was...
So please resend, thanks.
greg k-h
| |