[Acked] [SRU][Xenial][PATCH 0/3] Cleanups for CVE-2017-5715 (Spectre v2)
Andy Whitcroft
apw at canonical.com
Fri Nov 30 12:00:55 UTC 2018
On Wed, Nov 21, 2018 at 02:58:28PM +0100, Juerg Haefliger wrote:
> This patchset cleans up the Ubuntu-specific IBRS and IBPB runtime controls.
> The runtime controls from the embargoed patches are messy and spread all
> over the place. These patches consolidate the modifications into the
> proper places (commandline options in arch/x86/kernel/cpu/bugs.c instead of
> kernel/smp.c, speculation macros in arch/x86/include/asm/nospec-branch.h
> instead of open-coded) and merge them with the additional spectre-related
> changes that went in recently.
>
> In addtion, the 2nd patch adds an entry to
> /sys/devices/system/cpu/vulnerabilities/spectre_v2 when IBRS is enabled
> via procfs to return the full set of enabled mitigations.
>
> Compile-tested all architectures. Ran release tests to verify no
> regression is introduced. Fiddled with the ibrs_enabled and ibpb_enabled
> procfs controls to verify proper behaviour.
>
> Signed-off-by: Juerg Haefliger <juergh at canonical.com>
>
>
> Juerg Haefliger (3):
> UBUNTU: SAUCE: x86/speculation: Cleanup IBPB runtime control handling
> UBUNTU: SAUCE: x86/speculation: Cleanup IBRS runtime control handling
> UBUNTU: SAUCE: x86/speculation: Move RSB_CTXSW hunk
>
> arch/x86/include/asm/mwait.h | 6 +-
> arch/x86/include/asm/nospec-branch.h | 28 +++-
> arch/x86/include/asm/spec_ctrl.h | 11 +-
> arch/x86/kernel/cpu/amd.c | 5 +-
> arch/x86/kernel/cpu/bugs.c | 87 +++++++-----
> arch/x86/kernel/cpu/microcode/core.c | 23 ---
> arch/x86/kernel/process.c | 10 +-
> arch/x86/kernel/smpboot.c | 6 +-
> arch/x86/kvm/svm.c | 6 +-
> arch/x86/kvm/vmx.c | 3 +-
> arch/x86/lib/delay.c | 8 +-
> arch/x86/mm/tlb.c | 2 +-
> include/linux/smp.h | 83 -----------
> kernel/smp.c | 46 ------
> kernel/sysctl.c | 201 ++++++++++++++++-----------
> 15 files changed, 223 insertions(+), 302 deletions(-)
>
> --
I assume we have some testing we can do on top to confirm these work on
the final kernels. They look ok on face value and the desire is good.
Acked-by: Andy Whitcroft <apw at canonical.com>
-apw
More information about the kernel-team
mailing list