On 04/06/2017 12:25 PM, Borislav Petkov wrote:
Hi Brijesh,
On Thu, Apr 06, 2017 at 09:05:03AM -0500, Brijesh Singh wrote:
I looked into arch/x86/mm/init_{32,64}.c and as you pointed the file contains
routines to do basic page splitting. I think it sufficient for our usage.
Good :)
I should be able to drop the memblock patch from the series and update the
Patch 15 [1] to use the kernel_physical_mapping_init().
The kernel_physical_mapping_init() creates the page table mapping using
default KERNEL_PAGE attributes, I tried to extend the function by passing
'bool enc' flags to hint whether to clr or set _PAGE_ENC when splitting the
pages. The code did not looked clean hence I dropped that idea.
Or, you could have a
__kernel_physical_mapping_init_prot(..., prot)
helper which gets a protection argument and hands it down. The lower
levels already hand down prot which is good.
I did thought about prot idea but ran into another corner case which may require
us changing the signature of phys_pud_init and phys_pmd_init. The paddr_start
and paddr_end args into kernel_physical_mapping_init() should be aligned on PMD
level down (see comment [1]). So, if we encounter a case where our address range
is part of large page but we need to clear only one entry (i.e asked to clear just
one page into 2M region). In that case, now we need to pass additional arguments
into kernel_physical_mapping, phys_pud_init and phys_pmd_init to hint the splitting
code that it should use our prot for specific entries and all other entries will use
the old_prot.
[1] http://lxr.free-electrons.com/source/arch/x86/mm/init_64.c#L546
The interface kernel_physical_mapping_init() will then itself call:
__kernel_physical_mapping_init_prot(..., PAGE_KERNEL);
for the normal cases.
That in a pre-patch of course.
How does that sound?
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
