Re: [PATCH 4.4 000/118] 4.4.22-stable review

From: Andrew Donnellan
Date: Wed Sep 28 2016 - 04:45:19 EST

On 24/09/16 01:30, Greg Kroah-Hartman wrote:
On Fri, Sep 23, 2016 at 05:11:48PM +0200, Greg Kroah-Hartman wrote:
On Fri, Sep 23, 2016 at 06:25:33AM -0700, Guenter Roeck wrote:
On 09/23/2016 01:08 AM, Greg Kroah-Hartman wrote:
On Thu, Sep 22, 2016 at 08:01:35PM -0700, Guenter Roeck wrote:
e5767a13d7cb ("cxl: use pcibios_free_controller_deferred() when removing vPHBs")
does not apply to v4.4 nor to v4.7 (and isn't marked stable ???).

I don't understand what you mean here. It did apply, and it was marked
for stable inclusion. It is commit
6f38a8b9a45833495dc878c335c5431cd98a16ed in Linus's tree. Does it break
the build?

drivers/misc/cxl/vphb.c: In function 'cxl_pci_vphb_add':
drivers/misc/cxl/vphb.c:263:9: error: 'pcibios_free_controller_deferred' undeclared

Yes, it does break the build, by calling a function which does not exist in 4.4 or 4.7.

Sorry, the Cc: stable was dropped in the 4.4/4.7 versions, so I didn't see it.
I meant to say "it doesn't build".

You will either have to drop e5767a13d7cb. or you'll also need commit 2dd9c11b9d4
("powerpc/pseries: use pci_host_bridge.release_fn() to kfree(phb)") from upstream.
I didn't check though if that applies (or works) with the older kernels.

I'm going to drop it, thanks for pointing this out, I should have caught
it earlier too...

Should be removed, hopefully all will build properly...

It would've been helpful if Mauricio and I were Cc'd on this, but in any case I was away for most of last week :)

Looks like I suggested that 2dd9c11b9d4 be Cc: stable on the list but that didn't quite make it in to the final commit. Probably didn't help that our patches were then applied to powerpc#fixes in the wrong order...

2dd9c11b9d4 applies cleanly on top of 4.4 and works for me in the cxl case. Mauricio - any objections from the pseries/dlpar side?


Andrew Donnellan OzLabs, ADL Canberra
andrew.donnellan@xxxxxxxxxxx IBM Australia Limited