[PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Mark Rutland <mark.rutland@arm.com> · 2016-11-01
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Dave Hansen <dave.hansen@linux.intel.com> · 2016-11-02
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Mark Rutland <mark.rutland@arm.com> · 2016-11-04
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Heiko Carstens <hidden> · 2016-11-08
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Russell King - ARM Linux <linux@armlinux.org.uk> · 2016-11-08
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Heiko Carstens <hidden> · 2016-11-08
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Dave Hansen <dave.hansen@linux.intel.com> · 2016-11-08
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Arnd Bergmann <arnd@arndb.de> · 2016-11-08
Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS · Heiko Carstens <hidden> · 2016-11-08

Re: [PATCH] mm: only enable sys_pkey* when ARCH_HAS_PKEYS

From: Heiko Carstens <hidden>
Date: 2016-11-08 12:05:52
Also in: linux-arch, linux-mm, lkml

On Tue, Nov 08, 2016 at 12:39:28PM +0100, Arnd Bergmann wrote:

On Tuesday, November 8, 2016 10:30:42 AM CET Heiko Carstens wrote:

quoted

Three architectures (parisc, powerpc, s390) decided to ignore the system
calls completely, but still have the pkey code linked into the kernel
image.

Wouldn't it actually make sense to hook this up to the storage keys
in the s390 page tables?

We have storage keys per _physical_ page. Not per page within the the table
entries. So this doesn't work unfortunately.