x86/pkeys: Make mprotect_key() mask off additional vm_flags
authorDave Hansen <dave.hansen@linux.intel.com>
Fri, 29 Jul 2016 16:30:13 +0000 (09:30 -0700)
committerThomas Gleixner <tglx@linutronix.de>
Fri, 9 Sep 2016 11:02:26 +0000 (13:02 +0200)
commita8502b67d739c1d7a4542c1da0a5d98a6a58c177
treeb5c3e39594eb7e23168bae4b3b58d84a0a1dfb34
parent7d06d9c9bd813fc956b9c7bffc1b9724009983eb
x86/pkeys: Make mprotect_key() mask off additional vm_flags

Today, mprotect() takes 4 bits of data: PROT_READ/WRITE/EXEC/NONE.
Three of those bits: READ/WRITE/EXEC get translated directly in to
vma->vm_flags by calc_vm_prot_bits().  If a bit is unset in
mprotect()'s 'prot' argument then it must be cleared in vma->vm_flags
during the mprotect() call.

We do this clearing today by first calculating the VMA flags we
want set, then clearing the ones we do not want to inherit from
the original VMA:

vm_flags = calc_vm_prot_bits(prot, key);
...
newflags = vm_flags;
newflags |= (vma->vm_flags & ~(VM_READ | VM_WRITE | VM_EXEC));

However, we *also* want to mask off the original VMA's vm_flags in
which we store the protection key.

To do that, this patch adds a new macro:

ARCH_VM_PKEY_FLAGS

which allows the architecture to specify additional bits that it would
like cleared.  We use that to ensure that the VM_PKEY_BIT* bits get
cleared.

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: Dave Hansen <dave@sr71.net>
Cc: arnd@arndb.de
Cc: linux-api@vger.kernel.org
Cc: linux-mm@kvack.org
Cc: luto@kernel.org
Cc: akpm@linux-foundation.org
Cc: torvalds@linux-foundation.org
Link: http://lkml.kernel.org/r/20160729163013.E48D6981@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
arch/x86/include/asm/pkeys.h
include/linux/pkeys.h
mm/mprotect.c