[Acked] [Xenial][SRU][PATCH 0/2] #LP1606147 Workaround for monitor bug slows down system

[Andy Whitcroft]

On Tue, Jul 26, 2016 at 04:33:15PM +0800, Phidias Chiang wrote:
> [Impact]
> 
> We've found a bug on some Apollo Lake system makes it runs very slow.
> From upstream it is caused by following reason:
> 
> "Monitored cached line may not wake up from mwait on certain Goldmont
> based CPUs. This patch will avoid calling current_set_polling_and_test()
> and thereby not set the TIF_ flag. The result is that we'll always send
> IPIs for wakeups."
> 
> [Fix]
> 
> Upstream already provided a workaround[1] to fix this problem. This
> patches haven't been merged in latest(v4.7) release but already in
> linux-next.
> 
> To fix this issue in Xenial it also needs another commit introduces
> macros use in the workaround:
> 
> commit 970442c599b22ccd644ebfe94d1d303bf6f87c05
> Author: Dave Hansen <dave at sr71.net>
> Date: Thu Jun 2 17:19:27 2016 -0700
> 
>     x86/cpu/intel: Introduce macros for Intel family numbers
> 
> This workaround checks CPU families and certain features, so no further
> regressions is expected
> 
> [1]: https://lkml.org/lkml/2016/7/6/469
> 
> Dave Hansen (1):
>   x86/cpu/intel: Introduce macros for Intel family numbers
> 
> Peter Zijlstra (1):
>   UBUNTU: SAUCE: x86/cpu: Add workaround for MONITOR instruction erratum
>     on Goldmont based CPUs
> 
>  arch/x86/include/asm/cpufeatures.h  |  1 +
>  arch/x86/include/asm/intel-family.h | 68 +++++++++++++++++++++++++++++++++++++
>  arch/x86/include/asm/mwait.h        |  2 +-
>  arch/x86/kernel/cpu/intel.c         |  5 +++
>  arch/x86/kernel/process.c           |  2 +-
>  5 files changed, 76 insertions(+), 2 deletions(-)
>  create mode 100644 arch/x86/include/asm/intel-family.h

Acked-by: Andy Whitcroft <apw at canonical.com>

-apw