On Wed, Dec 20, 2017 at 11:51:15AM -0800, Brian Norris wrote:
Hi Greg,
On Mon, Dec 18, 2017 at 04:47:58PM +0100, Greg Kroah-Hartman wrote:
4.4-stable review patch. If anyone has any objections, please let me know.
I'm sorry, but I already objected to this one during the discussion here:
https://patchwork.kernel.org/patch/10065483/ [PATCH 4.13 03/28] Bluetooth: btusb: fix QCA Rome suspend/resume
in which we pointed out a regression. The $subject patch does NOT actually resolve the previous regression, though it might help to mask it. The proper approach to resolve the above regression was to revert the patch, not to backport the $subject patch.
Regarding this patch, IIUC this is not a bugfix -- it's a feature addition (e.g., for helping with BLE mouse wakeup), and it has already been proven to break some user space (we have an internal bug tracking this, but suffice it to say that we've already tried and reverted this patch [1]). This patch massively increases the surface in which spurious bluetooth activity can wake the system, and in some cases we never can suspend the system at all.
Unfortunately, Matthias was on vacation when you sent the review request, so our team wasn't alerted properly. Can you please back this out of all -stable branches?
Or alternatively, if those I've added on CC disagree and are happy to deal with the fallout of this patch...well, then that's fine. We can revert this patch in our downstream kernels and reapply if/when we can account for it properly :)
As Linus's tree is also broken, being bug-compatible here is good, right? I can just apply the revert/fix patch when it lands in that tree, and all will be ok.
Or is Linus's tree not broken now? Sorry, this whole thread has been really confusing...
thanks,
greg k-h