At Tue, 18 Mar 2014 11:27:56 +0000, Mark Brown wrote:
On Tue, Mar 18, 2014 at 12:20:00PM +0100, Takashi Iwai wrote:
regmap deploys the spinlock for the protection when set up in fast_io mode. But some places are still left with possible sleep in the lock, typically allocating memory with GFP_KERNEL. This patch fixes these calls by moving the allocations out of the lock.
You should have split these two unrelated changes into two patches.
You're suggesting to split the patch to fix for regmap_bulk_write() and regmap_register_patch()? I don't mind to split, but basically both are the very same bug, and the solution is very same, too.
I already have a fix for the cache sync in my tree locally which I will test just a soon as I fix whatever broke the -next build for me.
I don't know why you're talking about the build fix...
Takashi