Re: [ PATCH stable v4.19] EDAC: Drop per-memory controller buses

From: Manali Shukla (manashuk)
Date: Sun Apr 05 2020 - 14:01:48 EST


Hi Greg

With this patch , we are removing per-MC bus, this removes dependency on value of max number of controllers (EDAC_MAX_MCS) which is hardcoded to 2 * MAX_NUMNODES in all stable versions of kernel.

On two nodes system MAX_NUMNODES value is â1â , so value of max number of memory controller becomes â2â, this patch fixes this issue when there are only 2 nodes on the system and number of memory controllers are more than â2'

https://lore.kernel.org/linux-edac/20180926152752.GG5584@xxxxxxx/

-Manali

ïOn 17/03/20, 3:52 PM, "Greg KH" <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:

On Wed, Mar 11, 2020 at 10:22:01PM -0700, Manali K Shukla wrote:
> From: Borislav Petkov <bp@xxxxxxx>
>
> upstream 861e6ed667c83d64a42b0db41a22d6b4de4e913f commit
>
> ... and use the single edac_subsys object returned from
> subsys_system_register(). The idea is to have a single bus
> and multiple devices on it.
>
> Signed-off-by: Borislav Petkov <bp@xxxxxxx>
> Acked-by: Mauro Carvalho Chehab <mchehab+samsung@xxxxxxxxxx>
> CC: Aristeu Rozanski Filho <arozansk@xxxxxxxxxx>
> CC: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
> CC: Justin Ernst <justin.ernst@xxxxxxx>
> CC: linux-edac <linux-edac@xxxxxxxxxxxxxxx>
> CC: Mauro Carvalho Chehab <mchehab@xxxxxxxxxx>
> CC: Russ Anderson <rja@xxxxxxx>
> Cc: Tony Luck <tony.luck@xxxxxxxxx>
> Link: https://lkml.kernel.org/r/20180926152752.GG5584@xxxxxxx
> [Manali: backport to v4.19 -stable :
> - removing per-MC bus, this enables to get rid of memory controllers
> maximum number notion
> - value of max number of memory controllers is 2 * MAX_NUMNODES. On two nodes system MAX_NUMNODES value is â1â and
> so value of max number of memory controller becomes â2â, this patch fixes this issue when there are only 2 nodes on the system
> and number of memory controllers are more than â2â]
> (cherry picked from commit 861e6ed667c83d64a42b0db41a22d6b4de4e913f)
> Signed-off-by: Manali K Shukla <manashuk@xxxxxxxxx>

Why is this a patch for the stable trees? What problem does it solve?

thanks,

greg k-h-