2007-07-18 03:37:04 +02:00
|
|
|
/*
|
|
|
|
* Xen mmu operations
|
|
|
|
*
|
|
|
|
* This file contains the various mmu fetch and update operations.
|
|
|
|
* The most important job they must perform is the mapping between the
|
|
|
|
* domain's pfn and the overall machine mfns.
|
|
|
|
*
|
|
|
|
* Xen allows guests to directly update the pagetable, in a controlled
|
|
|
|
* fashion. In other words, the guest modifies the same pagetable
|
|
|
|
* that the CPU actually uses, which eliminates the overhead of having
|
|
|
|
* a separate shadow pagetable.
|
|
|
|
*
|
|
|
|
* In order to allow this, it falls on the guest domain to map its
|
|
|
|
* notion of a "physical" pfn - which is just a domain-local linear
|
|
|
|
* address - into a real "machine address" which the CPU's MMU can
|
|
|
|
* use.
|
|
|
|
*
|
|
|
|
* A pgd_t/pmd_t/pte_t will typically contain an mfn, and so can be
|
|
|
|
* inserted directly into the pagetable. When creating a new
|
|
|
|
* pte/pmd/pgd, it converts the passed pfn into an mfn. Conversely,
|
|
|
|
* when reading the content back with __(pgd|pmd|pte)_val, it converts
|
|
|
|
* the mfn back into a pfn.
|
|
|
|
*
|
|
|
|
* The other constraint is that all pages which make up a pagetable
|
|
|
|
* must be mapped read-only in the guest. This prevents uncontrolled
|
|
|
|
* guest updates to the pagetable. Xen strictly enforces this, and
|
|
|
|
* will disallow any pagetable update which will end up mapping a
|
|
|
|
* pagetable page RW, and will disallow using any writable page as a
|
|
|
|
* pagetable.
|
|
|
|
*
|
|
|
|
* Naively, when loading %cr3 with the base of a new pagetable, Xen
|
|
|
|
* would need to validate the whole pagetable before going on.
|
|
|
|
* Naturally, this is quite slow. The solution is to "pin" a
|
|
|
|
* pagetable, which enforces all the constraints on the pagetable even
|
|
|
|
* when it is not actively in use. This menas that Xen can be assured
|
|
|
|
* that it is still valid when you do load it into %cr3, and doesn't
|
|
|
|
* need to revalidate it.
|
|
|
|
*
|
|
|
|
* Jeremy Fitzhardinge <jeremy@xensource.com>, XenSource Inc, 2007
|
|
|
|
*/
|
2007-07-18 03:37:06 +02:00
|
|
|
#include <linux/sched.h>
|
2007-07-18 03:37:05 +02:00
|
|
|
#include <linux/highmem.h>
|
2008-08-21 02:02:19 +02:00
|
|
|
#include <linux/debugfs.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
#include <linux/bug.h>
|
2010-03-26 23:37:50 +01:00
|
|
|
#include <linux/vmalloc.h>
|
2009-05-12 22:31:40 +02:00
|
|
|
#include <linux/module.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 09:04:11 +01:00
|
|
|
#include <linux/gfp.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
|
|
|
|
#include <asm/pgtable.h>
|
|
|
|
#include <asm/tlbflush.h>
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
#include <asm/fixmap.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
#include <asm/mmu_context.h>
|
2009-01-28 23:35:01 +01:00
|
|
|
#include <asm/setup.h>
|
2007-07-18 03:37:05 +02:00
|
|
|
#include <asm/paravirt.h>
|
2010-02-19 19:31:06 +01:00
|
|
|
#include <asm/e820.h>
|
2008-07-09 00:06:27 +02:00
|
|
|
#include <asm/linkage.h>
|
2009-02-09 21:05:46 +01:00
|
|
|
#include <asm/page.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
|
|
|
|
#include <asm/xen/hypercall.h>
|
2007-07-18 03:37:05 +02:00
|
|
|
#include <asm/xen/hypervisor.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2010-02-04 23:46:34 +01:00
|
|
|
#include <xen/xen.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
#include <xen/page.h>
|
|
|
|
#include <xen/interface/xen.h>
|
2010-06-17 15:22:52 +02:00
|
|
|
#include <xen/interface/hvm/hvm_op.h>
|
2009-01-28 23:35:01 +01:00
|
|
|
#include <xen/interface/version.h>
|
2010-02-04 23:46:34 +01:00
|
|
|
#include <xen/interface/memory.h>
|
2009-01-28 23:35:01 +01:00
|
|
|
#include <xen/hvc-console.h>
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
#include "multicalls.h"
|
2007-07-18 03:37:04 +02:00
|
|
|
#include "mmu.h"
|
2008-08-21 02:02:19 +02:00
|
|
|
#include "debugfs.h"
|
|
|
|
|
|
|
|
#define MMU_UPDATE_HISTO 30
|
|
|
|
|
2009-02-09 21:05:46 +01:00
|
|
|
/*
|
|
|
|
* Protects atomic reservation decrease/increase against concurrent increases.
|
|
|
|
* Also protects non-atomic updates of current_pages and driver_pages, and
|
|
|
|
* balloon lists.
|
|
|
|
*/
|
|
|
|
DEFINE_SPINLOCK(xen_reservation_lock);
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
#ifdef CONFIG_XEN_DEBUG_FS
|
|
|
|
|
|
|
|
static struct {
|
|
|
|
u32 pgd_update;
|
|
|
|
u32 pgd_update_pinned;
|
|
|
|
u32 pgd_update_batched;
|
|
|
|
|
|
|
|
u32 pud_update;
|
|
|
|
u32 pud_update_pinned;
|
|
|
|
u32 pud_update_batched;
|
|
|
|
|
|
|
|
u32 pmd_update;
|
|
|
|
u32 pmd_update_pinned;
|
|
|
|
u32 pmd_update_batched;
|
|
|
|
|
|
|
|
u32 pte_update;
|
|
|
|
u32 pte_update_pinned;
|
|
|
|
u32 pte_update_batched;
|
|
|
|
|
|
|
|
u32 mmu_update;
|
|
|
|
u32 mmu_update_extended;
|
|
|
|
u32 mmu_update_histo[MMU_UPDATE_HISTO];
|
|
|
|
|
|
|
|
u32 prot_commit;
|
|
|
|
u32 prot_commit_batched;
|
|
|
|
|
|
|
|
u32 set_pte_at;
|
|
|
|
u32 set_pte_at_batched;
|
|
|
|
u32 set_pte_at_pinned;
|
|
|
|
u32 set_pte_at_current;
|
|
|
|
u32 set_pte_at_kernel;
|
|
|
|
} mmu_stats;
|
|
|
|
|
|
|
|
static u8 zero_stats;
|
|
|
|
|
|
|
|
static inline void check_zero(void)
|
|
|
|
{
|
|
|
|
if (unlikely(zero_stats)) {
|
|
|
|
memset(&mmu_stats, 0, sizeof(mmu_stats));
|
|
|
|
zero_stats = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#define ADD_STATS(elem, val) \
|
|
|
|
do { check_zero(); mmu_stats.elem += (val); } while(0)
|
|
|
|
|
|
|
|
#else /* !CONFIG_XEN_DEBUG_FS */
|
|
|
|
|
|
|
|
#define ADD_STATS(elem, val) do { (void)(val); } while(0)
|
|
|
|
|
|
|
|
#endif /* CONFIG_XEN_DEBUG_FS */
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2009-01-28 23:35:01 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Identity map, in addition to plain kernel map. This needs to be
|
|
|
|
* large enough to allocate page table pages to allocate the rest.
|
|
|
|
* Each page can map 2MB.
|
|
|
|
*/
|
|
|
|
static pte_t level1_ident_pgt[PTRS_PER_PTE * 4] __page_aligned_bss;
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
/* l3 pud for userspace vsyscall mapping */
|
|
|
|
static pud_t level3_user_vsyscall[PTRS_PER_PUD] __page_aligned_bss;
|
|
|
|
#endif /* CONFIG_X86_64 */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Note about cr3 (pagetable base) values:
|
|
|
|
*
|
|
|
|
* xen_cr3 contains the current logical cr3 value; it contains the
|
|
|
|
* last set cr3. This may not be the current effective cr3, because
|
|
|
|
* its update may be being lazily deferred. However, a vcpu looking
|
|
|
|
* at its own cr3 can use this value knowing that it everything will
|
|
|
|
* be self-consistent.
|
|
|
|
*
|
|
|
|
* xen_current_cr3 contains the actual vcpu cr3; it is set once the
|
|
|
|
* hypercall to set the vcpu cr3 is complete (so it may be a little
|
|
|
|
* out of date, but it will never be set early). If one vcpu is
|
|
|
|
* looking at another vcpu's cr3 value, it should use this variable.
|
|
|
|
*/
|
|
|
|
DEFINE_PER_CPU(unsigned long, xen_cr3); /* cr3 stored as physaddr */
|
|
|
|
DEFINE_PER_CPU(unsigned long, xen_current_cr3); /* actual vcpu cr3 */
|
|
|
|
|
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
/*
|
|
|
|
* Just beyond the highest usermode address. STACK_TOP_MAX has a
|
|
|
|
* redzone above it, so round it up to a PGD boundary.
|
|
|
|
*/
|
|
|
|
#define USER_LIMIT ((STACK_TOP_MAX + PGDIR_SIZE - 1) & PGDIR_MASK)
|
|
|
|
|
|
|
|
|
2008-05-27 00:31:18 +02:00
|
|
|
#define P2M_ENTRIES_PER_PAGE (PAGE_SIZE / sizeof(unsigned long))
|
2008-05-27 00:31:20 +02:00
|
|
|
#define TOP_ENTRIES (MAX_DOMAIN_PAGES / P2M_ENTRIES_PER_PAGE)
|
2008-05-27 00:31:18 +02:00
|
|
|
|
2008-05-27 00:31:20 +02:00
|
|
|
/* Placeholder for holes in the address space */
|
2008-07-09 00:06:27 +02:00
|
|
|
static unsigned long p2m_missing[P2M_ENTRIES_PER_PAGE] __page_aligned_data =
|
2008-05-27 00:31:20 +02:00
|
|
|
{ [ 0 ... P2M_ENTRIES_PER_PAGE-1 ] = ~0UL };
|
|
|
|
|
|
|
|
/* Array of pointers to pages containing p2m entries */
|
2008-07-09 00:06:27 +02:00
|
|
|
static unsigned long *p2m_top[TOP_ENTRIES] __page_aligned_data =
|
2008-05-27 00:31:20 +02:00
|
|
|
{ [ 0 ... TOP_ENTRIES - 1] = &p2m_missing[0] };
|
2008-05-27 00:31:18 +02:00
|
|
|
|
2008-05-27 00:31:22 +02:00
|
|
|
/* Arrays of p2m arrays expressed in mfns used for save/restore */
|
2008-07-09 00:06:27 +02:00
|
|
|
static unsigned long p2m_top_mfn[TOP_ENTRIES] __page_aligned_bss;
|
2008-05-27 00:31:22 +02:00
|
|
|
|
2008-07-09 00:06:27 +02:00
|
|
|
static unsigned long p2m_top_mfn_list[TOP_ENTRIES / P2M_ENTRIES_PER_PAGE]
|
|
|
|
__page_aligned_bss;
|
2008-05-27 00:31:22 +02:00
|
|
|
|
2008-05-27 00:31:18 +02:00
|
|
|
static inline unsigned p2m_top_index(unsigned long pfn)
|
|
|
|
{
|
2008-05-27 00:31:19 +02:00
|
|
|
BUG_ON(pfn >= MAX_DOMAIN_PAGES);
|
2008-05-27 00:31:18 +02:00
|
|
|
return pfn / P2M_ENTRIES_PER_PAGE;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline unsigned p2m_index(unsigned long pfn)
|
|
|
|
{
|
|
|
|
return pfn % P2M_ENTRIES_PER_PAGE;
|
|
|
|
}
|
|
|
|
|
2008-05-27 00:31:22 +02:00
|
|
|
/* Build the parallel p2m_top_mfn structures */
|
2009-11-21 12:32:49 +01:00
|
|
|
void xen_build_mfn_list_list(void)
|
2008-05-27 00:31:22 +02:00
|
|
|
{
|
|
|
|
unsigned pfn, idx;
|
|
|
|
|
2008-12-16 20:56:06 +01:00
|
|
|
for (pfn = 0; pfn < MAX_DOMAIN_PAGES; pfn += P2M_ENTRIES_PER_PAGE) {
|
2008-05-27 00:31:22 +02:00
|
|
|
unsigned topidx = p2m_top_index(pfn);
|
|
|
|
|
|
|
|
p2m_top_mfn[topidx] = virt_to_mfn(p2m_top[topidx]);
|
|
|
|
}
|
|
|
|
|
2008-12-16 20:56:06 +01:00
|
|
|
for (idx = 0; idx < ARRAY_SIZE(p2m_top_mfn_list); idx++) {
|
2008-05-27 00:31:22 +02:00
|
|
|
unsigned topidx = idx * P2M_ENTRIES_PER_PAGE;
|
|
|
|
p2m_top_mfn_list[idx] = virt_to_mfn(&p2m_top_mfn[topidx]);
|
|
|
|
}
|
2009-02-28 00:34:59 +01:00
|
|
|
}
|
2008-05-27 00:31:22 +02:00
|
|
|
|
2009-02-28 00:34:59 +01:00
|
|
|
void xen_setup_mfn_list_list(void)
|
|
|
|
{
|
2008-05-27 00:31:22 +02:00
|
|
|
BUG_ON(HYPERVISOR_shared_info == &xen_dummy_shared_info);
|
|
|
|
|
|
|
|
HYPERVISOR_shared_info->arch.pfn_to_mfn_frame_list_list =
|
|
|
|
virt_to_mfn(p2m_top_mfn_list);
|
|
|
|
HYPERVISOR_shared_info->arch.max_pfn = xen_start_info->nr_pages;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set up p2m_top to point to the domain-builder provided p2m pages */
|
2008-05-27 00:31:18 +02:00
|
|
|
void __init xen_build_dynamic_phys_to_machine(void)
|
|
|
|
{
|
|
|
|
unsigned long *mfn_list = (unsigned long *)xen_start_info->mfn_list;
|
2008-05-27 00:31:19 +02:00
|
|
|
unsigned long max_pfn = min(MAX_DOMAIN_PAGES, xen_start_info->nr_pages);
|
2008-05-27 00:31:22 +02:00
|
|
|
unsigned pfn;
|
2008-05-27 00:31:18 +02:00
|
|
|
|
2008-12-16 20:56:06 +01:00
|
|
|
for (pfn = 0; pfn < max_pfn; pfn += P2M_ENTRIES_PER_PAGE) {
|
2008-05-27 00:31:18 +02:00
|
|
|
unsigned topidx = p2m_top_index(pfn);
|
|
|
|
|
|
|
|
p2m_top[topidx] = &mfn_list[pfn];
|
|
|
|
}
|
2009-02-28 00:34:59 +01:00
|
|
|
|
|
|
|
xen_build_mfn_list_list();
|
2008-05-27 00:31:18 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
unsigned long get_phys_to_machine(unsigned long pfn)
|
|
|
|
{
|
|
|
|
unsigned topidx, idx;
|
|
|
|
|
2008-05-27 00:31:19 +02:00
|
|
|
if (unlikely(pfn >= MAX_DOMAIN_PAGES))
|
|
|
|
return INVALID_P2M_ENTRY;
|
|
|
|
|
2008-05-27 00:31:18 +02:00
|
|
|
topidx = p2m_top_index(pfn);
|
|
|
|
idx = p2m_index(pfn);
|
|
|
|
return p2m_top[topidx][idx];
|
|
|
|
}
|
2008-06-02 13:20:11 +02:00
|
|
|
EXPORT_SYMBOL_GPL(get_phys_to_machine);
|
2008-05-27 00:31:18 +02:00
|
|
|
|
2009-02-27 00:48:33 +01:00
|
|
|
/* install a new p2m_top page */
|
|
|
|
bool install_p2mtop_page(unsigned long pfn, unsigned long *p)
|
2008-05-27 00:31:18 +02:00
|
|
|
{
|
2009-02-27 00:48:33 +01:00
|
|
|
unsigned topidx = p2m_top_index(pfn);
|
|
|
|
unsigned long **pfnp, *mfnp;
|
2008-05-27 00:31:18 +02:00
|
|
|
unsigned i;
|
|
|
|
|
2009-02-27 00:48:33 +01:00
|
|
|
pfnp = &p2m_top[topidx];
|
|
|
|
mfnp = &p2m_top_mfn[topidx];
|
2008-05-27 00:31:18 +02:00
|
|
|
|
2008-12-16 20:56:06 +01:00
|
|
|
for (i = 0; i < P2M_ENTRIES_PER_PAGE; i++)
|
2008-05-27 00:31:18 +02:00
|
|
|
p[i] = INVALID_P2M_ENTRY;
|
|
|
|
|
2009-02-27 00:48:33 +01:00
|
|
|
if (cmpxchg(pfnp, p2m_missing, p) == p2m_missing) {
|
2008-05-27 00:31:22 +02:00
|
|
|
*mfnp = virt_to_mfn(p);
|
2009-02-27 00:48:33 +01:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
2008-05-27 00:31:18 +02:00
|
|
|
}
|
|
|
|
|
2009-02-27 00:48:33 +01:00
|
|
|
static void alloc_p2m(unsigned long pfn)
|
2008-05-27 00:31:18 +02:00
|
|
|
{
|
2009-02-27 00:48:33 +01:00
|
|
|
unsigned long *p;
|
2008-05-27 00:31:18 +02:00
|
|
|
|
2009-02-27 00:48:33 +01:00
|
|
|
p = (void *)__get_free_page(GFP_KERNEL | __GFP_NOFAIL);
|
|
|
|
BUG_ON(p == NULL);
|
|
|
|
|
|
|
|
if (!install_p2mtop_page(pfn, p))
|
|
|
|
free_page((unsigned long)p);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Try to install p2m mapping; fail if intermediate bits missing */
|
|
|
|
bool __set_phys_to_machine(unsigned long pfn, unsigned long mfn)
|
|
|
|
{
|
|
|
|
unsigned topidx, idx;
|
2008-05-27 00:31:19 +02:00
|
|
|
|
|
|
|
if (unlikely(pfn >= MAX_DOMAIN_PAGES)) {
|
|
|
|
BUG_ON(mfn != INVALID_P2M_ENTRY);
|
2009-02-27 00:48:33 +01:00
|
|
|
return true;
|
2008-05-27 00:31:18 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
topidx = p2m_top_index(pfn);
|
2008-05-27 00:31:20 +02:00
|
|
|
if (p2m_top[topidx] == p2m_missing) {
|
2008-05-27 00:31:18 +02:00
|
|
|
if (mfn == INVALID_P2M_ENTRY)
|
2009-02-27 00:48:33 +01:00
|
|
|
return true;
|
|
|
|
return false;
|
2008-05-27 00:31:18 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
idx = p2m_index(pfn);
|
|
|
|
p2m_top[topidx][idx] = mfn;
|
2009-02-27 00:48:33 +01:00
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
void set_phys_to_machine(unsigned long pfn, unsigned long mfn)
|
|
|
|
{
|
|
|
|
if (unlikely(xen_feature(XENFEAT_auto_translated_physmap))) {
|
|
|
|
BUG_ON(pfn != mfn && mfn != INVALID_P2M_ENTRY);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (unlikely(!__set_phys_to_machine(pfn, mfn))) {
|
|
|
|
alloc_p2m(pfn);
|
|
|
|
|
|
|
|
if (!__set_phys_to_machine(pfn, mfn))
|
|
|
|
BUG();
|
|
|
|
}
|
2008-05-27 00:31:18 +02:00
|
|
|
}
|
|
|
|
|
2009-02-27 18:19:26 +01:00
|
|
|
unsigned long arbitrary_virt_to_mfn(void *vaddr)
|
|
|
|
{
|
|
|
|
xmaddr_t maddr = arbitrary_virt_to_machine(vaddr);
|
|
|
|
|
|
|
|
return PFN_DOWN(maddr.maddr);
|
|
|
|
}
|
|
|
|
|
2008-07-09 00:06:55 +02:00
|
|
|
xmaddr_t arbitrary_virt_to_machine(void *vaddr)
|
2007-07-18 03:37:04 +02:00
|
|
|
{
|
2008-07-09 00:06:55 +02:00
|
|
|
unsigned long address = (unsigned long)vaddr;
|
2008-02-09 23:24:08 +01:00
|
|
|
unsigned int level;
|
xen: fix Xen domU boot with batched mprotect
Impact: fix guest kernel boot crash on certain configs
Recent i686 2.6.27 kernels with a certain amount of memory (between
736 and 855MB) have a problem booting under a hypervisor that supports
batched mprotect (this includes the RHEL-5 Xen hypervisor as well as
any 3.3 or later Xen hypervisor).
The problem ends up being that xen_ptep_modify_prot_commit() is using
virt_to_machine to calculate which pfn to update. However, this only
works for pages that are in the p2m list, and the pages coming from
change_pte_range() in mm/mprotect.c are kmap_atomic pages. Because of
this, we can run into the situation where the lookup in the p2m table
returns an INVALID_MFN, which we then try to pass to the hypervisor,
which then (correctly) denies the request to a totally bogus pfn.
The right thing to do is to use arbitrary_virt_to_machine, so that we
can be sure we are modifying the right pfn. This unfortunately
introduces a performance penalty because of a full page-table-walk,
but we can avoid that penalty for pages in the p2m list by checking if
virt_addr_valid is true, and if so, just doing the lookup in the p2m
table.
The attached patch implements this, and allows my 2.6.27 i686 based
guest with 768MB of memory to boot on a RHEL-5 hypervisor again.
Thanks to Jeremy for the suggestions about how to fix this particular
issue.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-10-24 02:40:25 +02:00
|
|
|
pte_t *pte;
|
|
|
|
unsigned offset;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
xen: fix Xen domU boot with batched mprotect
Impact: fix guest kernel boot crash on certain configs
Recent i686 2.6.27 kernels with a certain amount of memory (between
736 and 855MB) have a problem booting under a hypervisor that supports
batched mprotect (this includes the RHEL-5 Xen hypervisor as well as
any 3.3 or later Xen hypervisor).
The problem ends up being that xen_ptep_modify_prot_commit() is using
virt_to_machine to calculate which pfn to update. However, this only
works for pages that are in the p2m list, and the pages coming from
change_pte_range() in mm/mprotect.c are kmap_atomic pages. Because of
this, we can run into the situation where the lookup in the p2m table
returns an INVALID_MFN, which we then try to pass to the hypervisor,
which then (correctly) denies the request to a totally bogus pfn.
The right thing to do is to use arbitrary_virt_to_machine, so that we
can be sure we are modifying the right pfn. This unfortunately
introduces a performance penalty because of a full page-table-walk,
but we can avoid that penalty for pages in the p2m list by checking if
virt_addr_valid is true, and if so, just doing the lookup in the p2m
table.
The attached patch implements this, and allows my 2.6.27 i686 based
guest with 768MB of memory to boot on a RHEL-5 hypervisor again.
Thanks to Jeremy for the suggestions about how to fix this particular
issue.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-10-24 02:40:25 +02:00
|
|
|
/*
|
|
|
|
* if the PFN is in the linear mapped vaddr range, we can just use
|
|
|
|
* the (quick) virt_to_machine() p2m lookup
|
|
|
|
*/
|
|
|
|
if (virt_addr_valid(vaddr))
|
|
|
|
return virt_to_machine(vaddr);
|
|
|
|
|
|
|
|
/* otherwise we have to do a (slower) full page-table walk */
|
2007-07-18 03:37:04 +02:00
|
|
|
|
xen: fix Xen domU boot with batched mprotect
Impact: fix guest kernel boot crash on certain configs
Recent i686 2.6.27 kernels with a certain amount of memory (between
736 and 855MB) have a problem booting under a hypervisor that supports
batched mprotect (this includes the RHEL-5 Xen hypervisor as well as
any 3.3 or later Xen hypervisor).
The problem ends up being that xen_ptep_modify_prot_commit() is using
virt_to_machine to calculate which pfn to update. However, this only
works for pages that are in the p2m list, and the pages coming from
change_pte_range() in mm/mprotect.c are kmap_atomic pages. Because of
this, we can run into the situation where the lookup in the p2m table
returns an INVALID_MFN, which we then try to pass to the hypervisor,
which then (correctly) denies the request to a totally bogus pfn.
The right thing to do is to use arbitrary_virt_to_machine, so that we
can be sure we are modifying the right pfn. This unfortunately
introduces a performance penalty because of a full page-table-walk,
but we can avoid that penalty for pages in the p2m list by checking if
virt_addr_valid is true, and if so, just doing the lookup in the p2m
table.
The attached patch implements this, and allows my 2.6.27 i686 based
guest with 768MB of memory to boot on a RHEL-5 hypervisor again.
Thanks to Jeremy for the suggestions about how to fix this particular
issue.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-10-24 02:40:25 +02:00
|
|
|
pte = lookup_address(address, &level);
|
|
|
|
BUG_ON(pte == NULL);
|
|
|
|
offset = address & ~PAGE_MASK;
|
2008-07-09 00:06:54 +02:00
|
|
|
return XMADDR(((phys_addr_t)pte_mfn(*pte) << PAGE_SHIFT) + offset);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void make_lowmem_page_readonly(void *vaddr)
|
|
|
|
{
|
|
|
|
pte_t *pte, ptev;
|
|
|
|
unsigned long address = (unsigned long)vaddr;
|
2008-02-09 23:24:08 +01:00
|
|
|
unsigned int level;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-01-30 13:33:43 +01:00
|
|
|
pte = lookup_address(address, &level);
|
2007-07-18 03:37:04 +02:00
|
|
|
BUG_ON(pte == NULL);
|
|
|
|
|
|
|
|
ptev = pte_wrprotect(*pte);
|
|
|
|
|
|
|
|
if (HYPERVISOR_update_va_mapping(address, ptev, 0))
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
void make_lowmem_page_readwrite(void *vaddr)
|
|
|
|
{
|
|
|
|
pte_t *pte, ptev;
|
|
|
|
unsigned long address = (unsigned long)vaddr;
|
2008-02-09 23:24:08 +01:00
|
|
|
unsigned int level;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-01-30 13:33:43 +01:00
|
|
|
pte = lookup_address(address, &level);
|
2007-07-18 03:37:04 +02:00
|
|
|
BUG_ON(pte == NULL);
|
|
|
|
|
|
|
|
ptev = pte_mkwrite(*pte);
|
|
|
|
|
|
|
|
if (HYPERVISOR_update_va_mapping(address, ptev, 0))
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2008-08-19 22:34:22 +02:00
|
|
|
static bool xen_page_pinned(void *ptr)
|
2008-05-31 02:24:27 +02:00
|
|
|
{
|
|
|
|
struct page *page = virt_to_page(ptr);
|
|
|
|
|
|
|
|
return PagePinned(page);
|
|
|
|
}
|
|
|
|
|
2010-02-04 23:46:34 +01:00
|
|
|
static bool xen_iomap_pte(pte_t pte)
|
|
|
|
{
|
2010-02-19 19:31:06 +01:00
|
|
|
return pte_flags(pte) & _PAGE_IOMAP;
|
2010-02-04 23:46:34 +01:00
|
|
|
}
|
|
|
|
|
2009-02-09 21:05:49 +01:00
|
|
|
void xen_set_domain_pte(pte_t *ptep, pte_t pteval, unsigned domid)
|
2010-02-04 23:46:34 +01:00
|
|
|
{
|
|
|
|
struct multicall_space mcs;
|
|
|
|
struct mmu_update *u;
|
|
|
|
|
|
|
|
mcs = xen_mc_entry(sizeof(*u));
|
|
|
|
u = mcs.args;
|
|
|
|
|
|
|
|
/* ptep might be kmapped when using 32-bit HIGHPTE */
|
|
|
|
u->ptr = arbitrary_virt_to_machine(ptep).maddr;
|
|
|
|
u->val = pte_val_ma(pteval);
|
|
|
|
|
2009-02-09 21:05:49 +01:00
|
|
|
MULTI_mmu_update(mcs.mc, mcs.args, 1, NULL, domid);
|
2010-02-04 23:46:34 +01:00
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
}
|
2009-02-09 21:05:49 +01:00
|
|
|
EXPORT_SYMBOL_GPL(xen_set_domain_pte);
|
|
|
|
|
|
|
|
static void xen_set_iomap_pte(pte_t *ptep, pte_t pteval)
|
|
|
|
{
|
|
|
|
xen_set_domain_pte(ptep, pteval, DOMID_IO);
|
|
|
|
}
|
2010-02-04 23:46:34 +01:00
|
|
|
|
2008-08-19 22:34:22 +02:00
|
|
|
static void xen_extend_mmu_update(const struct mmu_update *update)
|
2007-07-18 03:37:04 +02:00
|
|
|
{
|
2007-07-18 03:37:06 +02:00
|
|
|
struct multicall_space mcs;
|
|
|
|
struct mmu_update *u;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-06-16 13:30:03 +02:00
|
|
|
mcs = xen_mc_extend_args(__HYPERVISOR_mmu_update, sizeof(*u));
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
if (mcs.mc != NULL) {
|
|
|
|
ADD_STATS(mmu_update_extended, 1);
|
|
|
|
ADD_STATS(mmu_update_histo[mcs.mc->args[1]], -1);
|
|
|
|
|
2008-06-16 13:30:03 +02:00
|
|
|
mcs.mc->args[1]++;
|
2008-08-21 02:02:19 +02:00
|
|
|
|
|
|
|
if (mcs.mc->args[1] < MMU_UPDATE_HISTO)
|
|
|
|
ADD_STATS(mmu_update_histo[mcs.mc->args[1]], 1);
|
|
|
|
else
|
|
|
|
ADD_STATS(mmu_update_histo[0], 1);
|
|
|
|
} else {
|
|
|
|
ADD_STATS(mmu_update, 1);
|
2008-06-16 13:30:03 +02:00
|
|
|
mcs = __xen_mc_entry(sizeof(*u));
|
|
|
|
MULTI_mmu_update(mcs.mc, mcs.args, 1, NULL, DOMID_SELF);
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(mmu_update_histo[1], 1);
|
2008-06-16 13:30:03 +02:00
|
|
|
}
|
2007-07-18 03:37:06 +02:00
|
|
|
|
|
|
|
u = mcs.args;
|
2008-06-16 13:30:03 +02:00
|
|
|
*u = *update;
|
|
|
|
}
|
|
|
|
|
|
|
|
void xen_set_pmd_hyper(pmd_t *ptr, pmd_t val)
|
|
|
|
{
|
|
|
|
struct mmu_update u;
|
|
|
|
|
|
|
|
preempt_disable();
|
|
|
|
|
|
|
|
xen_mc_batch();
|
|
|
|
|
2008-07-09 00:06:55 +02:00
|
|
|
/* ptr may be ioremapped for 64-bit pagetable setup */
|
|
|
|
u.ptr = arbitrary_virt_to_machine(ptr).maddr;
|
2008-06-16 13:30:03 +02:00
|
|
|
u.val = pmd_val_ma(val);
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_extend_mmu_update(&u);
|
2007-07-18 03:37:06 +02:00
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pmd_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU);
|
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
|
|
|
|
preempt_enable();
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
void xen_set_pmd(pmd_t *ptr, pmd_t val)
|
|
|
|
{
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pmd_update, 1);
|
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
/* If page is not pinned, we can just update the entry
|
|
|
|
directly */
|
2008-08-19 22:34:22 +02:00
|
|
|
if (!xen_page_pinned(ptr)) {
|
2008-05-31 02:24:27 +02:00
|
|
|
*ptr = val;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pmd_update_pinned, 1);
|
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
xen_set_pmd_hyper(ptr, val);
|
|
|
|
}
|
|
|
|
|
2007-07-18 03:37:04 +02:00
|
|
|
/*
|
|
|
|
* Associate a virtual page frame with a given physical page frame
|
|
|
|
* and protection flags for that frame.
|
|
|
|
*/
|
|
|
|
void set_pte_mfn(unsigned long vaddr, unsigned long mfn, pgprot_t flags)
|
|
|
|
{
|
2008-07-09 00:06:58 +02:00
|
|
|
set_pte_vaddr(vaddr, mfn_pte(mfn, flags));
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void xen_set_pte_at(struct mm_struct *mm, unsigned long addr,
|
|
|
|
pte_t *ptep, pte_t pteval)
|
|
|
|
{
|
2010-02-04 23:46:34 +01:00
|
|
|
if (xen_iomap_pte(pteval)) {
|
|
|
|
xen_set_iomap_pte(ptep, pteval);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(set_pte_at, 1);
|
|
|
|
// ADD_STATS(set_pte_at_pinned, xen_page_pinned(ptep));
|
|
|
|
ADD_STATS(set_pte_at_current, mm == current->mm);
|
|
|
|
ADD_STATS(set_pte_at_kernel, mm == &init_mm);
|
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
if (mm == current->mm || mm == &init_mm) {
|
paravirt: clean up lazy mode handling
Currently, the set_lazy_mode pv_op is overloaded with 5 functions:
1. enter lazy cpu mode
2. leave lazy cpu mode
3. enter lazy mmu mode
4. leave lazy mmu mode
5. flush pending batched operations
This complicates each paravirt backend, since it needs to deal with
all the possible state transitions, handling flushing, etc. In
particular, flushing is quite distinct from the other 4 functions, and
seems to just cause complication.
This patch removes the set_lazy_mode operation, and adds "enter" and
"leave" lazy mode operations on mmu_ops and cpu_ops. All the logic
associated with enter and leaving lazy states is now in common code
(basically BUG_ONs to make sure that no mode is current when entering
a lazy mode, and make sure that the mode is current when leaving).
Also, flush is handled in a common way, by simply leaving and
re-entering the lazy mode.
The result is that the Xen, lguest and VMI lazy mode implementations
are much simpler.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Zach Amsden <zach@vmware.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Avi Kivity <avi@qumranet.com>
Cc: Anthony Liguory <aliguori@us.ibm.com>
Cc: "Glauber de Oliveira Costa" <glommer@gmail.com>
Cc: Jun Nakajima <jun.nakajima@intel.com>
2007-10-16 20:51:29 +02:00
|
|
|
if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU) {
|
2007-07-18 03:37:06 +02:00
|
|
|
struct multicall_space mcs;
|
|
|
|
mcs = xen_mc_entry(0);
|
|
|
|
|
|
|
|
MULTI_update_va_mapping(mcs.mc, addr, pteval, 0);
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(set_pte_at_batched, 1);
|
2007-07-18 03:37:06 +02:00
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
2008-04-02 19:54:10 +02:00
|
|
|
goto out;
|
2007-07-18 03:37:06 +02:00
|
|
|
} else
|
|
|
|
if (HYPERVISOR_update_va_mapping(addr, pteval, 0) == 0)
|
2008-04-02 19:54:10 +02:00
|
|
|
goto out;
|
2007-07-18 03:37:06 +02:00
|
|
|
}
|
|
|
|
xen_set_pte(ptep, pteval);
|
2008-04-02 19:54:10 +02:00
|
|
|
|
2009-02-18 08:53:19 +01:00
|
|
|
out: return;
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2008-12-16 20:56:06 +01:00
|
|
|
pte_t xen_ptep_modify_prot_start(struct mm_struct *mm,
|
|
|
|
unsigned long addr, pte_t *ptep)
|
2008-03-18 00:37:09 +01:00
|
|
|
{
|
2008-06-16 13:30:02 +02:00
|
|
|
/* Just return the pte as-is. We preserve the bits on commit */
|
|
|
|
return *ptep;
|
|
|
|
}
|
|
|
|
|
|
|
|
void xen_ptep_modify_prot_commit(struct mm_struct *mm, unsigned long addr,
|
|
|
|
pte_t *ptep, pte_t pte)
|
|
|
|
{
|
2008-06-16 13:30:03 +02:00
|
|
|
struct mmu_update u;
|
2008-06-16 13:30:02 +02:00
|
|
|
|
2008-06-16 13:30:03 +02:00
|
|
|
xen_mc_batch();
|
2008-03-18 00:37:09 +01:00
|
|
|
|
xen: fix Xen domU boot with batched mprotect
Impact: fix guest kernel boot crash on certain configs
Recent i686 2.6.27 kernels with a certain amount of memory (between
736 and 855MB) have a problem booting under a hypervisor that supports
batched mprotect (this includes the RHEL-5 Xen hypervisor as well as
any 3.3 or later Xen hypervisor).
The problem ends up being that xen_ptep_modify_prot_commit() is using
virt_to_machine to calculate which pfn to update. However, this only
works for pages that are in the p2m list, and the pages coming from
change_pte_range() in mm/mprotect.c are kmap_atomic pages. Because of
this, we can run into the situation where the lookup in the p2m table
returns an INVALID_MFN, which we then try to pass to the hypervisor,
which then (correctly) denies the request to a totally bogus pfn.
The right thing to do is to use arbitrary_virt_to_machine, so that we
can be sure we are modifying the right pfn. This unfortunately
introduces a performance penalty because of a full page-table-walk,
but we can avoid that penalty for pages in the p2m list by checking if
virt_addr_valid is true, and if so, just doing the lookup in the p2m
table.
The attached patch implements this, and allows my 2.6.27 i686 based
guest with 768MB of memory to boot on a RHEL-5 hypervisor again.
Thanks to Jeremy for the suggestions about how to fix this particular
issue.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-10-24 02:40:25 +02:00
|
|
|
u.ptr = arbitrary_virt_to_machine(ptep).maddr | MMU_PT_UPDATE_PRESERVE_AD;
|
2008-06-16 13:30:03 +02:00
|
|
|
u.val = pte_val_ma(pte);
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_extend_mmu_update(&u);
|
2008-03-18 00:37:09 +01:00
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(prot_commit, 1);
|
|
|
|
ADD_STATS(prot_commit_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU);
|
|
|
|
|
2008-06-16 13:30:02 +02:00
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
/* Assume pteval_t is equivalent to all the other *val_t types. */
|
|
|
|
static pteval_t pte_mfn_to_pfn(pteval_t val)
|
2008-03-18 00:37:09 +01:00
|
|
|
{
|
2008-06-17 00:01:56 +02:00
|
|
|
if (val & _PAGE_PRESENT) {
|
2008-07-22 07:59:42 +02:00
|
|
|
unsigned long mfn = (val & PTE_PFN_MASK) >> PAGE_SHIFT;
|
2008-07-22 07:59:56 +02:00
|
|
|
pteval_t flags = val & PTE_FLAGS_MASK;
|
2008-07-04 07:10:18 +02:00
|
|
|
val = ((pteval_t)mfn_to_pfn(mfn) << PAGE_SHIFT) | flags;
|
2008-06-17 00:01:56 +02:00
|
|
|
}
|
2008-03-18 00:37:09 +01:00
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
return val;
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
static pteval_t pte_pfn_to_mfn(pteval_t val)
|
2008-03-18 00:37:09 +01:00
|
|
|
{
|
2008-06-17 00:01:56 +02:00
|
|
|
if (val & _PAGE_PRESENT) {
|
2008-07-22 07:59:42 +02:00
|
|
|
unsigned long pfn = (val & PTE_PFN_MASK) >> PAGE_SHIFT;
|
2008-07-22 07:59:56 +02:00
|
|
|
pteval_t flags = val & PTE_FLAGS_MASK;
|
2008-07-04 07:10:18 +02:00
|
|
|
val = ((pteval_t)pfn_to_mfn(pfn) << PAGE_SHIFT) | flags;
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
return val;
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
|
|
|
|
2010-02-04 23:46:34 +01:00
|
|
|
static pteval_t iomap_pte(pteval_t val)
|
|
|
|
{
|
|
|
|
if (val & _PAGE_PRESENT) {
|
|
|
|
unsigned long pfn = (val & PTE_PFN_MASK) >> PAGE_SHIFT;
|
|
|
|
pteval_t flags = val & PTE_FLAGS_MASK;
|
|
|
|
|
|
|
|
/* We assume the pte frame number is a MFN, so
|
|
|
|
just use it as-is. */
|
|
|
|
val = ((pteval_t)pfn << PAGE_SHIFT) | flags;
|
|
|
|
}
|
|
|
|
|
|
|
|
return val;
|
|
|
|
}
|
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
pteval_t xen_pte_val(pte_t pte)
|
2008-03-18 00:37:09 +01:00
|
|
|
{
|
2010-02-04 23:46:34 +01:00
|
|
|
if (xen_initial_domain() && (pte.pte & _PAGE_IOMAP))
|
|
|
|
return pte.pte;
|
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
return pte_mfn_to_pfn(pte.pte);
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_pte_val);
|
2008-03-18 00:37:09 +01:00
|
|
|
|
|
|
|
pgdval_t xen_pgd_val(pgd_t pgd)
|
|
|
|
{
|
2008-06-17 00:01:56 +02:00
|
|
|
return pte_mfn_to_pfn(pgd.pgd);
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_pgd_val);
|
2008-03-18 00:37:09 +01:00
|
|
|
|
|
|
|
pte_t xen_make_pte(pteval_t pte)
|
|
|
|
{
|
2010-02-19 19:31:06 +01:00
|
|
|
phys_addr_t addr = (pte & PTE_PFN_MASK);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Unprivileged domains are allowed to do IOMAPpings for
|
|
|
|
* PCI passthrough, but not map ISA space. The ISA
|
|
|
|
* mappings are just dummy local mappings to keep other
|
|
|
|
* parts of the kernel happy.
|
|
|
|
*/
|
|
|
|
if (unlikely(pte & _PAGE_IOMAP) &&
|
|
|
|
(xen_initial_domain() || addr >= ISA_END_ADDRESS)) {
|
2010-02-04 23:46:34 +01:00
|
|
|
pte = iomap_pte(pte);
|
2010-02-19 19:31:06 +01:00
|
|
|
} else {
|
|
|
|
pte &= ~_PAGE_IOMAP;
|
2010-02-04 23:46:34 +01:00
|
|
|
pte = pte_pfn_to_mfn(pte);
|
2010-02-19 19:31:06 +01:00
|
|
|
}
|
2010-02-04 23:46:34 +01:00
|
|
|
|
2008-06-17 00:01:56 +02:00
|
|
|
return native_make_pte(pte);
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_make_pte);
|
2008-03-18 00:37:09 +01:00
|
|
|
|
|
|
|
pgd_t xen_make_pgd(pgdval_t pgd)
|
|
|
|
{
|
2008-06-17 00:01:56 +02:00
|
|
|
pgd = pte_pfn_to_mfn(pgd);
|
|
|
|
return native_make_pgd(pgd);
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_make_pgd);
|
2008-03-18 00:37:09 +01:00
|
|
|
|
|
|
|
pmdval_t xen_pmd_val(pmd_t pmd)
|
|
|
|
{
|
2008-06-17 00:01:56 +02:00
|
|
|
return pte_mfn_to_pfn(pmd.pmd);
|
2008-03-18 00:37:09 +01:00
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_pmd_val);
|
2008-05-09 13:05:57 +02:00
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
void xen_set_pud_hyper(pud_t *ptr, pud_t val)
|
2007-07-18 03:37:05 +02:00
|
|
|
{
|
2008-06-16 13:30:03 +02:00
|
|
|
struct mmu_update u;
|
2007-07-18 03:37:05 +02:00
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
preempt_disable();
|
|
|
|
|
2008-06-16 13:30:03 +02:00
|
|
|
xen_mc_batch();
|
|
|
|
|
2008-07-09 00:06:55 +02:00
|
|
|
/* ptr may be ioremapped for 64-bit pagetable setup */
|
|
|
|
u.ptr = arbitrary_virt_to_machine(ptr).maddr;
|
2008-06-16 13:30:03 +02:00
|
|
|
u.val = pud_val_ma(val);
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_extend_mmu_update(&u);
|
2007-07-18 03:37:06 +02:00
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pud_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU);
|
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
|
|
|
|
preempt_enable();
|
2007-07-18 03:37:05 +02:00
|
|
|
}
|
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
void xen_set_pud(pud_t *ptr, pud_t val)
|
|
|
|
{
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pud_update, 1);
|
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
/* If page is not pinned, we can just update the entry
|
|
|
|
directly */
|
2008-08-19 22:34:22 +02:00
|
|
|
if (!xen_page_pinned(ptr)) {
|
2008-05-31 02:24:27 +02:00
|
|
|
*ptr = val;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pud_update_pinned, 1);
|
|
|
|
|
2008-05-31 02:24:27 +02:00
|
|
|
xen_set_pud_hyper(ptr, val);
|
|
|
|
}
|
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
void xen_set_pte(pte_t *ptep, pte_t pte)
|
|
|
|
{
|
2010-02-04 23:46:34 +01:00
|
|
|
if (xen_iomap_pte(pte)) {
|
|
|
|
xen_set_iomap_pte(ptep, pte);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pte_update, 1);
|
|
|
|
// ADD_STATS(pte_update_pinned, xen_page_pinned(ptep));
|
|
|
|
ADD_STATS(pte_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU);
|
|
|
|
|
2008-07-09 00:06:38 +02:00
|
|
|
#ifdef CONFIG_X86_PAE
|
2007-07-18 03:37:05 +02:00
|
|
|
ptep->pte_high = pte.pte_high;
|
|
|
|
smp_wmb();
|
|
|
|
ptep->pte_low = pte.pte_low;
|
2008-07-09 00:06:38 +02:00
|
|
|
#else
|
|
|
|
*ptep = pte;
|
|
|
|
#endif
|
2007-07-18 03:37:05 +02:00
|
|
|
}
|
|
|
|
|
2008-07-09 00:06:38 +02:00
|
|
|
#ifdef CONFIG_X86_PAE
|
2007-07-18 03:37:04 +02:00
|
|
|
void xen_set_pte_atomic(pte_t *ptep, pte_t pte)
|
|
|
|
{
|
2010-02-04 23:46:34 +01:00
|
|
|
if (xen_iomap_pte(pte)) {
|
|
|
|
xen_set_iomap_pte(ptep, pte);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2008-07-09 00:06:38 +02:00
|
|
|
set_64bit((u64 *)ptep, native_pte_val(pte));
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void xen_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
|
|
|
|
{
|
|
|
|
ptep->pte_low = 0;
|
|
|
|
smp_wmb(); /* make sure low gets written first */
|
|
|
|
ptep->pte_high = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void xen_pmd_clear(pmd_t *pmdp)
|
|
|
|
{
|
2008-05-31 02:24:27 +02:00
|
|
|
set_pmd(pmdp, __pmd(0));
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
2008-07-09 00:06:38 +02:00
|
|
|
#endif /* CONFIG_X86_PAE */
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-03-18 00:37:07 +01:00
|
|
|
pmd_t xen_make_pmd(pmdval_t pmd)
|
2007-07-18 03:37:04 +02:00
|
|
|
{
|
2008-06-17 00:01:56 +02:00
|
|
|
pmd = pte_pfn_to_mfn(pmd);
|
2008-03-18 00:37:09 +01:00
|
|
|
return native_make_pmd(pmd);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_make_pmd);
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-07-09 00:06:38 +02:00
|
|
|
#if PAGETABLE_LEVELS == 4
|
|
|
|
pudval_t xen_pud_val(pud_t pud)
|
|
|
|
{
|
|
|
|
return pte_mfn_to_pfn(pud.pud);
|
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_pud_val);
|
2008-07-09 00:06:38 +02:00
|
|
|
|
|
|
|
pud_t xen_make_pud(pudval_t pud)
|
|
|
|
{
|
|
|
|
pud = pte_pfn_to_mfn(pud);
|
|
|
|
|
|
|
|
return native_make_pud(pud);
|
|
|
|
}
|
2009-01-28 23:35:07 +01:00
|
|
|
PV_CALLEE_SAVE_REGS_THUNK(xen_make_pud);
|
2008-07-09 00:06:38 +02:00
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
pgd_t *xen_get_user_pgd(pgd_t *pgd)
|
2008-07-09 00:06:38 +02:00
|
|
|
{
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
pgd_t *pgd_page = (pgd_t *)(((unsigned long)pgd) & PAGE_MASK);
|
|
|
|
unsigned offset = pgd - pgd_page;
|
|
|
|
pgd_t *user_ptr = NULL;
|
2008-07-09 00:06:38 +02:00
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
if (offset < pgd_index(USER_LIMIT)) {
|
|
|
|
struct page *page = virt_to_page(pgd_page);
|
|
|
|
user_ptr = (pgd_t *)page->private;
|
|
|
|
if (user_ptr)
|
|
|
|
user_ptr += offset;
|
|
|
|
}
|
2008-07-09 00:06:38 +02:00
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
return user_ptr;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __xen_set_pgd_hyper(pgd_t *ptr, pgd_t val)
|
|
|
|
{
|
|
|
|
struct mmu_update u;
|
2008-07-09 00:06:38 +02:00
|
|
|
|
|
|
|
u.ptr = virt_to_machine(ptr).maddr;
|
|
|
|
u.val = pgd_val_ma(val);
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_extend_mmu_update(&u);
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Raw hypercall-based set_pgd, intended for in early boot before
|
|
|
|
* there's a page structure. This implies:
|
|
|
|
* 1. The only existing pagetable is the kernel's
|
|
|
|
* 2. It is always pinned
|
|
|
|
* 3. It has no user pagetable attached to it
|
|
|
|
*/
|
|
|
|
void __init xen_set_pgd_hyper(pgd_t *ptr, pgd_t val)
|
|
|
|
{
|
|
|
|
preempt_disable();
|
|
|
|
|
|
|
|
xen_mc_batch();
|
|
|
|
|
|
|
|
__xen_set_pgd_hyper(ptr, val);
|
2008-07-09 00:06:38 +02:00
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
|
|
|
void xen_set_pgd(pgd_t *ptr, pgd_t val)
|
|
|
|
{
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
pgd_t *user_ptr = xen_get_user_pgd(ptr);
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pgd_update, 1);
|
|
|
|
|
2008-07-09 00:06:38 +02:00
|
|
|
/* If page is not pinned, we can just update the entry
|
|
|
|
directly */
|
2008-08-19 22:34:22 +02:00
|
|
|
if (!xen_page_pinned(ptr)) {
|
2008-07-09 00:06:38 +02:00
|
|
|
*ptr = val;
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
if (user_ptr) {
|
2008-08-19 22:34:22 +02:00
|
|
|
WARN_ON(xen_page_pinned(user_ptr));
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
*user_ptr = val;
|
|
|
|
}
|
2008-07-09 00:06:38 +02:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
ADD_STATS(pgd_update_pinned, 1);
|
|
|
|
ADD_STATS(pgd_update_batched, paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU);
|
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
/* If it's pinned, then we can at least batch the kernel and
|
|
|
|
user updates together. */
|
|
|
|
xen_mc_batch();
|
|
|
|
|
|
|
|
__xen_set_pgd_hyper(ptr, val);
|
|
|
|
if (user_ptr)
|
|
|
|
__xen_set_pgd_hyper(user_ptr, val);
|
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
2008-07-09 00:06:38 +02:00
|
|
|
}
|
|
|
|
#endif /* PAGETABLE_LEVELS == 4 */
|
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
/*
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
* (Yet another) pagetable walker. This one is intended for pinning a
|
|
|
|
* pagetable. This means that it walks a pagetable and calls the
|
|
|
|
* callback function on each page it finds making up the page table,
|
|
|
|
* at every level. It walks the entire pagetable, but it only bothers
|
|
|
|
* pinning pte pages which are below limit. In the normal case this
|
|
|
|
* will be STACK_TOP_MAX, but at boot we need to pin up to
|
|
|
|
* FIXADDR_TOP.
|
|
|
|
*
|
|
|
|
* For 32-bit the important bit is that we don't pin beyond there,
|
|
|
|
* because then we start getting into Xen's ptes.
|
|
|
|
*
|
|
|
|
* For 64-bit, we must skip the Xen hole in the middle of the address
|
|
|
|
* space, just after the big x86-64 virtual hole.
|
|
|
|
*/
|
2008-11-21 11:21:33 +01:00
|
|
|
static int __xen_pgd_walk(struct mm_struct *mm, pgd_t *pgd,
|
|
|
|
int (*func)(struct mm_struct *mm, struct page *,
|
|
|
|
enum pt_level),
|
|
|
|
unsigned long limit)
|
2007-07-18 03:37:04 +02:00
|
|
|
{
|
2007-07-18 03:37:05 +02:00
|
|
|
int flush = 0;
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
unsigned hole_low, hole_high;
|
|
|
|
unsigned pgdidx_limit, pudidx_limit, pmdidx_limit;
|
|
|
|
unsigned pgdidx, pudidx, pmdidx;
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
/* The limit is the last byte to be touched */
|
|
|
|
limit--;
|
|
|
|
BUG_ON(limit >= FIXADDR_TOP);
|
2007-07-18 03:37:04 +02:00
|
|
|
|
|
|
|
if (xen_feature(XENFEAT_auto_translated_physmap))
|
2007-07-18 03:37:05 +02:00
|
|
|
return 0;
|
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
/*
|
|
|
|
* 64-bit has a great big hole in the middle of the address
|
|
|
|
* space, which contains the Xen mappings. On 32-bit these
|
|
|
|
* will end up making a zero-sized hole and so is a no-op.
|
|
|
|
*/
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
hole_low = pgd_index(USER_LIMIT);
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
hole_high = pgd_index(PAGE_OFFSET);
|
|
|
|
|
|
|
|
pgdidx_limit = pgd_index(limit);
|
|
|
|
#if PTRS_PER_PUD > 1
|
|
|
|
pudidx_limit = pud_index(limit);
|
|
|
|
#else
|
|
|
|
pudidx_limit = 0;
|
|
|
|
#endif
|
|
|
|
#if PTRS_PER_PMD > 1
|
|
|
|
pmdidx_limit = pmd_index(limit);
|
|
|
|
#else
|
|
|
|
pmdidx_limit = 0;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
for (pgdidx = 0; pgdidx <= pgdidx_limit; pgdidx++) {
|
2007-07-18 03:37:05 +02:00
|
|
|
pud_t *pud;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
if (pgdidx >= hole_low && pgdidx < hole_high)
|
|
|
|
continue;
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
if (!pgd_val(pgd[pgdidx]))
|
2007-07-18 03:37:04 +02:00
|
|
|
continue;
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
pud = pud_offset(&pgd[pgdidx], 0);
|
2007-07-18 03:37:04 +02:00
|
|
|
|
|
|
|
if (PTRS_PER_PUD > 1) /* not folded */
|
2008-10-08 22:01:39 +02:00
|
|
|
flush |= (*func)(mm, virt_to_page(pud), PT_PUD);
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
for (pudidx = 0; pudidx < PTRS_PER_PUD; pudidx++) {
|
2007-07-18 03:37:05 +02:00
|
|
|
pmd_t *pmd;
|
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
if (pgdidx == pgdidx_limit &&
|
|
|
|
pudidx > pudidx_limit)
|
|
|
|
goto out;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
if (pud_none(pud[pudidx]))
|
2007-07-18 03:37:04 +02:00
|
|
|
continue;
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
pmd = pmd_offset(&pud[pudidx], 0);
|
2007-07-18 03:37:04 +02:00
|
|
|
|
|
|
|
if (PTRS_PER_PMD > 1) /* not folded */
|
2008-10-08 22:01:39 +02:00
|
|
|
flush |= (*func)(mm, virt_to_page(pmd), PT_PMD);
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
for (pmdidx = 0; pmdidx < PTRS_PER_PMD; pmdidx++) {
|
|
|
|
struct page *pte;
|
|
|
|
|
|
|
|
if (pgdidx == pgdidx_limit &&
|
|
|
|
pudidx == pudidx_limit &&
|
|
|
|
pmdidx > pmdidx_limit)
|
|
|
|
goto out;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
if (pmd_none(pmd[pmdidx]))
|
2007-07-18 03:37:04 +02:00
|
|
|
continue;
|
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
pte = pmd_page(pmd[pmdidx]);
|
2008-10-08 22:01:39 +02:00
|
|
|
flush |= (*func)(mm, pte, PT_PTE);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2008-08-19 22:32:51 +02:00
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
out:
|
2008-08-19 22:32:51 +02:00
|
|
|
/* Do the top level last, so that the callbacks can use it as
|
|
|
|
a cue to do final things like tlb flushes. */
|
2008-10-08 22:01:39 +02:00
|
|
|
flush |= (*func)(mm, virt_to_page(pgd), PT_PGD);
|
2007-07-18 03:37:05 +02:00
|
|
|
|
|
|
|
return flush;
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2008-11-21 11:21:33 +01:00
|
|
|
static int xen_pgd_walk(struct mm_struct *mm,
|
|
|
|
int (*func)(struct mm_struct *mm, struct page *,
|
|
|
|
enum pt_level),
|
|
|
|
unsigned long limit)
|
|
|
|
{
|
|
|
|
return __xen_pgd_walk(mm, mm->pgd, func, limit);
|
|
|
|
}
|
|
|
|
|
2008-08-19 22:34:22 +02:00
|
|
|
/* If we're using split pte locks, then take the page's lock and
|
|
|
|
return a pointer to it. Otherwise return NULL. */
|
2008-10-08 22:01:39 +02:00
|
|
|
static spinlock_t *xen_pte_lock(struct page *page, struct mm_struct *mm)
|
2007-10-16 20:51:30 +02:00
|
|
|
{
|
|
|
|
spinlock_t *ptl = NULL;
|
|
|
|
|
2008-09-10 00:43:22 +02:00
|
|
|
#if USE_SPLIT_PTLOCKS
|
2007-10-16 20:51:30 +02:00
|
|
|
ptl = __pte_lockptr(page);
|
2008-10-08 22:01:39 +02:00
|
|
|
spin_lock_nest_lock(ptl, &mm->page_table_lock);
|
2007-10-16 20:51:30 +02:00
|
|
|
#endif
|
|
|
|
|
|
|
|
return ptl;
|
|
|
|
}
|
|
|
|
|
2008-08-19 22:34:22 +02:00
|
|
|
static void xen_pte_unlock(void *v)
|
2007-10-16 20:51:30 +02:00
|
|
|
{
|
|
|
|
spinlock_t *ptl = v;
|
|
|
|
spin_unlock(ptl);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_do_pin(unsigned level, unsigned long pfn)
|
|
|
|
{
|
|
|
|
struct mmuext_op *op;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
|
|
|
|
mcs = __xen_mc_entry(sizeof(*op));
|
|
|
|
op = mcs.args;
|
|
|
|
op->cmd = level;
|
|
|
|
op->arg1.mfn = pfn_to_mfn(pfn);
|
|
|
|
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
|
|
|
|
}
|
|
|
|
|
2008-10-08 22:01:39 +02:00
|
|
|
static int xen_pin_page(struct mm_struct *mm, struct page *page,
|
|
|
|
enum pt_level level)
|
2007-07-18 03:37:05 +02:00
|
|
|
{
|
2008-04-28 11:12:51 +02:00
|
|
|
unsigned pgfl = TestSetPagePinned(page);
|
2007-07-18 03:37:05 +02:00
|
|
|
int flush;
|
|
|
|
|
|
|
|
if (pgfl)
|
|
|
|
flush = 0; /* already pinned */
|
|
|
|
else if (PageHighMem(page))
|
|
|
|
/* kmaps need flushing if we found an unpinned
|
|
|
|
highpage */
|
|
|
|
flush = 1;
|
|
|
|
else {
|
|
|
|
void *pt = lowmem_page_address(page);
|
|
|
|
unsigned long pfn = page_to_pfn(page);
|
|
|
|
struct multicall_space mcs = __xen_mc_entry(0);
|
2007-10-16 20:51:30 +02:00
|
|
|
spinlock_t *ptl;
|
2007-07-18 03:37:05 +02:00
|
|
|
|
|
|
|
flush = 0;
|
|
|
|
|
2008-08-19 22:32:51 +02:00
|
|
|
/*
|
|
|
|
* We need to hold the pagetable lock between the time
|
|
|
|
* we make the pagetable RO and when we actually pin
|
|
|
|
* it. If we don't, then other users may come in and
|
|
|
|
* attempt to update the pagetable by writing it,
|
|
|
|
* which will fail because the memory is RO but not
|
|
|
|
* pinned, so Xen won't do the trap'n'emulate.
|
|
|
|
*
|
|
|
|
* If we're using split pte locks, we can't hold the
|
|
|
|
* entire pagetable's worth of locks during the
|
|
|
|
* traverse, because we may wrap the preempt count (8
|
|
|
|
* bits). The solution is to mark RO and pin each PTE
|
|
|
|
* page while holding the lock. This means the number
|
|
|
|
* of locks we end up holding is never more than a
|
|
|
|
* batch size (~32 entries, at present).
|
|
|
|
*
|
|
|
|
* If we're not using split pte locks, we needn't pin
|
|
|
|
* the PTE pages independently, because we're
|
|
|
|
* protected by the overall pagetable lock.
|
|
|
|
*/
|
2007-10-16 20:51:30 +02:00
|
|
|
ptl = NULL;
|
|
|
|
if (level == PT_PTE)
|
2008-10-08 22:01:39 +02:00
|
|
|
ptl = xen_pte_lock(page, mm);
|
2007-10-16 20:51:30 +02:00
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
MULTI_update_va_mapping(mcs.mc, (unsigned long)pt,
|
|
|
|
pfn_pte(pfn, PAGE_KERNEL_RO),
|
2007-10-16 20:51:30 +02:00
|
|
|
level == PT_PGD ? UVMF_TLB_FLUSH : 0);
|
|
|
|
|
2008-08-19 22:32:51 +02:00
|
|
|
if (ptl) {
|
2007-10-16 20:51:30 +02:00
|
|
|
xen_do_pin(MMUEXT_PIN_L1_TABLE, pfn);
|
|
|
|
|
|
|
|
/* Queue a deferred unlock for when this batch
|
|
|
|
is completed. */
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_mc_callback(xen_pte_unlock, ptl);
|
2007-10-16 20:51:30 +02:00
|
|
|
}
|
2007-07-18 03:37:05 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
return flush;
|
|
|
|
}
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
/* This is called just after a mm has been created, but it has not
|
|
|
|
been used yet. We need to make sure that its pagetable is all
|
|
|
|
read-only, and can be pinned. */
|
2008-10-08 22:01:39 +02:00
|
|
|
static void __xen_pgd_pin(struct mm_struct *mm, pgd_t *pgd)
|
2007-07-18 03:37:04 +02:00
|
|
|
{
|
2007-07-18 03:37:05 +02:00
|
|
|
xen_mc_batch();
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-11-21 11:21:33 +01:00
|
|
|
if (__xen_pgd_walk(mm, pgd, xen_pin_page, USER_LIMIT)) {
|
2008-10-28 09:23:06 +01:00
|
|
|
/* re-enable interrupts for flushing */
|
2007-07-18 03:37:06 +02:00
|
|
|
xen_mc_issue(0);
|
2008-10-28 09:23:06 +01:00
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
kmap_flush_unused();
|
2008-10-28 09:23:06 +01:00
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
xen_mc_batch();
|
|
|
|
}
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
{
|
|
|
|
pgd_t *user_pgd = xen_get_user_pgd(pgd);
|
|
|
|
|
|
|
|
xen_do_pin(MMUEXT_PIN_L4_TABLE, PFN_DOWN(__pa(pgd)));
|
|
|
|
|
|
|
|
if (user_pgd) {
|
2008-10-08 22:01:39 +02:00
|
|
|
xen_pin_page(mm, virt_to_page(user_pgd), PT_PGD);
|
2008-12-16 20:56:06 +01:00
|
|
|
xen_do_pin(MMUEXT_PIN_L4_TABLE,
|
|
|
|
PFN_DOWN(__pa(user_pgd)));
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
#else /* CONFIG_X86_32 */
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
/* Need to make sure unshared kernel PMD is pinnable */
|
2008-11-06 22:48:24 +01:00
|
|
|
xen_pin_page(mm, pgd_page(pgd[pgd_index(TASK_SIZE)]),
|
2008-10-08 22:01:39 +02:00
|
|
|
PT_PMD);
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
#endif
|
2008-05-09 13:05:57 +02:00
|
|
|
xen_do_pin(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(pgd)));
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
#endif /* CONFIG_X86_64 */
|
2007-07-18 03:37:05 +02:00
|
|
|
xen_mc_issue(0);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2008-10-08 22:01:39 +02:00
|
|
|
static void xen_pgd_pin(struct mm_struct *mm)
|
|
|
|
{
|
|
|
|
__xen_pgd_pin(mm, mm->pgd);
|
|
|
|
}
|
|
|
|
|
2008-05-27 00:31:27 +02:00
|
|
|
/*
|
|
|
|
* On save, we need to pin all pagetables to make sure they get their
|
|
|
|
* mfns turned into pfns. Search the list for any unpinned pgds and pin
|
|
|
|
* them (unpinned pgds are not currently in use, probably because the
|
|
|
|
* process is under construction or destruction).
|
2008-10-08 22:01:39 +02:00
|
|
|
*
|
|
|
|
* Expected to be called in stop_machine() ("equivalent to taking
|
|
|
|
* every spinlock in the system"), so the locking doesn't really
|
|
|
|
* matter all that much.
|
2008-05-27 00:31:27 +02:00
|
|
|
*/
|
|
|
|
void xen_mm_pin_all(void)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
struct page *page;
|
2007-10-16 20:51:30 +02:00
|
|
|
|
2008-05-27 00:31:27 +02:00
|
|
|
spin_lock_irqsave(&pgd_lock, flags);
|
2007-07-18 03:37:05 +02:00
|
|
|
|
2008-05-27 00:31:27 +02:00
|
|
|
list_for_each_entry(page, &pgd_list, lru) {
|
|
|
|
if (!PagePinned(page)) {
|
2008-10-08 22:01:39 +02:00
|
|
|
__xen_pgd_pin(&init_mm, (pgd_t *)page_address(page));
|
2008-05-27 00:31:27 +02:00
|
|
|
SetPageSavePinned(page);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&pgd_lock, flags);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2008-07-09 00:06:24 +02:00
|
|
|
/*
|
|
|
|
* The init_mm pagetable is really pinned as soon as its created, but
|
|
|
|
* that's before we have page structures to store the bits. So do all
|
|
|
|
* the book-keeping now.
|
|
|
|
*/
|
2008-10-08 22:01:39 +02:00
|
|
|
static __init int xen_mark_pinned(struct mm_struct *mm, struct page *page,
|
|
|
|
enum pt_level level)
|
2007-07-18 03:37:04 +02:00
|
|
|
{
|
2007-07-18 03:37:05 +02:00
|
|
|
SetPagePinned(page);
|
|
|
|
return 0;
|
|
|
|
}
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2009-03-17 21:30:55 +01:00
|
|
|
static void __init xen_mark_init_mm_pinned(void)
|
2007-07-18 03:37:05 +02:00
|
|
|
{
|
2008-10-08 22:01:39 +02:00
|
|
|
xen_pgd_walk(&init_mm, xen_mark_pinned, FIXADDR_TOP);
|
2007-07-18 03:37:05 +02:00
|
|
|
}
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-10-08 22:01:39 +02:00
|
|
|
static int xen_unpin_page(struct mm_struct *mm, struct page *page,
|
|
|
|
enum pt_level level)
|
2007-07-18 03:37:05 +02:00
|
|
|
{
|
2008-04-28 11:12:51 +02:00
|
|
|
unsigned pgfl = TestClearPagePinned(page);
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
if (pgfl && !PageHighMem(page)) {
|
|
|
|
void *pt = lowmem_page_address(page);
|
|
|
|
unsigned long pfn = page_to_pfn(page);
|
2007-10-16 20:51:30 +02:00
|
|
|
spinlock_t *ptl = NULL;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
|
2008-08-19 22:32:51 +02:00
|
|
|
/*
|
|
|
|
* Do the converse to pin_page. If we're using split
|
|
|
|
* pte locks, we must be holding the lock for while
|
|
|
|
* the pte page is unpinned but still RO to prevent
|
|
|
|
* concurrent updates from seeing it in this
|
|
|
|
* partially-pinned state.
|
|
|
|
*/
|
2007-10-16 20:51:30 +02:00
|
|
|
if (level == PT_PTE) {
|
2008-10-08 22:01:39 +02:00
|
|
|
ptl = xen_pte_lock(page, mm);
|
2007-10-16 20:51:30 +02:00
|
|
|
|
2008-08-19 22:32:51 +02:00
|
|
|
if (ptl)
|
|
|
|
xen_do_pin(MMUEXT_UNPIN_TABLE, pfn);
|
2007-10-16 20:51:30 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
mcs = __xen_mc_entry(0);
|
2007-07-18 03:37:05 +02:00
|
|
|
|
|
|
|
MULTI_update_va_mapping(mcs.mc, (unsigned long)pt,
|
|
|
|
pfn_pte(pfn, PAGE_KERNEL),
|
2007-10-16 20:51:30 +02:00
|
|
|
level == PT_PGD ? UVMF_TLB_FLUSH : 0);
|
|
|
|
|
|
|
|
if (ptl) {
|
|
|
|
/* unlock when batch completed */
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_mc_callback(xen_pte_unlock, ptl);
|
2007-10-16 20:51:30 +02:00
|
|
|
}
|
2007-07-18 03:37:05 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0; /* never need to flush on unpin */
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2007-07-18 03:37:05 +02:00
|
|
|
/* Release a pagetables pages back as normal RW */
|
2008-10-08 22:01:39 +02:00
|
|
|
static void __xen_pgd_unpin(struct mm_struct *mm, pgd_t *pgd)
|
2007-07-18 03:37:05 +02:00
|
|
|
{
|
|
|
|
xen_mc_batch();
|
|
|
|
|
2007-10-16 20:51:30 +02:00
|
|
|
xen_do_pin(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(pgd)));
|
2007-07-18 03:37:05 +02:00
|
|
|
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
{
|
|
|
|
pgd_t *user_pgd = xen_get_user_pgd(pgd);
|
|
|
|
|
|
|
|
if (user_pgd) {
|
2008-12-16 20:56:06 +01:00
|
|
|
xen_do_pin(MMUEXT_UNPIN_TABLE,
|
|
|
|
PFN_DOWN(__pa(user_pgd)));
|
2008-10-08 22:01:39 +02:00
|
|
|
xen_unpin_page(mm, virt_to_page(user_pgd), PT_PGD);
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
/* Need to make sure unshared kernel PMD is unpinned */
|
2008-11-06 22:48:24 +01:00
|
|
|
xen_unpin_page(mm, pgd_page(pgd[pgd_index(TASK_SIZE)]),
|
2008-10-08 22:01:39 +02:00
|
|
|
PT_PMD);
|
xen: rework pgd_walk to deal with 32/64 bit
Rewrite pgd_walk to deal with 64-bit address spaces. There are two
notible features of 64-bit workspaces:
1. The physical address is only 48 bits wide, with the upper 16 bits
being sign extension; kernel addresses are negative, and userspace is
positive.
2. The Xen hypervisor mapping is at the negative-most address, just above
the sign-extension hole.
1. means that we can't easily use addresses when traversing the space,
since we must deal with sign extension. This rewrite expresses
everything in terms of pgd/pud/pmd indices, which means we don't need
to worry about the exact configuration of the virtual memory space.
This approach works equally well in 32-bit.
To deal with 2, assume the hole is between the uppermost userspace
address and PAGE_OFFSET. For 64-bit this skips the Xen mapping hole.
For 32-bit, the hole is zero-sized.
In all cases, the uppermost kernel address is FIXADDR_TOP.
A side-effect of this patch is that the upper boundary is actually
handled properly, exposing a long-standing bug in 32-bit, which failed
to pin kernel pmd page. The kernel pmd is not shared, and so must be
explicitly pinned, even though the kernel ptes are shared and don't
need pinning.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:06 +02:00
|
|
|
#endif
|
xen64: allocate and manage user pagetables
Because the x86_64 architecture does not enforce segment limits, Xen
cannot protect itself with them as it does in 32-bit mode. Therefore,
to protect itself, it runs the guest kernel in ring 3. Since it also
runs the guest userspace in ring3, the guest kernel must maintain a
second pagetable for its userspace, which does not map kernel space.
Naturally, the guest kernel pagetables map both kernel and userspace.
The userspace pagetable is attached to the corresponding kernel
pagetable via the pgd's page->private field. It is allocated and
freed at the same time as the kernel pgd via the
paravirt_pgd_alloc/free hooks.
Fortunately, the user pagetable is almost entirely shared with the
kernel pagetable; the only difference is the pgd page itself. set_pgd
will populate all entries in the kernel pagetable, and also set the
corresponding user pgd entry if the address is less than
STACK_TOP_MAX.
The user pagetable must be pinned and unpinned with the kernel one,
but because the pagetables are aliased, pgd_walk() only needs to be
called on the kernel pagetable. The user pgd page is then
pinned/unpinned along with the kernel pgd page.
xen_write_cr3 must write both the kernel and user cr3s.
The init_mm.pgd pagetable never has a user pagetable allocated for it,
because it can never be used while running usermode.
One awkward area is that early in boot the page structures are not
available. No user pagetable can exist at that point, but it
complicates the logic to avoid looking at the page structure.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-09 00:07:13 +02:00
|
|
|
|
2008-11-21 11:21:33 +01:00
|
|
|
__xen_pgd_walk(mm, pgd, xen_unpin_page, USER_LIMIT);
|
2007-07-18 03:37:05 +02:00
|
|
|
|
|
|
|
xen_mc_issue(0);
|
|
|
|
}
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-10-08 22:01:39 +02:00
|
|
|
static void xen_pgd_unpin(struct mm_struct *mm)
|
|
|
|
{
|
|
|
|
__xen_pgd_unpin(mm, mm->pgd);
|
|
|
|
}
|
|
|
|
|
2008-05-27 00:31:27 +02:00
|
|
|
/*
|
|
|
|
* On resume, undo any pinning done at save, so that the rest of the
|
|
|
|
* kernel doesn't see any unexpected pinned pagetables.
|
|
|
|
*/
|
|
|
|
void xen_mm_unpin_all(void)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
struct page *page;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&pgd_lock, flags);
|
|
|
|
|
|
|
|
list_for_each_entry(page, &pgd_list, lru) {
|
|
|
|
if (PageSavePinned(page)) {
|
|
|
|
BUG_ON(!PagePinned(page));
|
2008-10-08 22:01:39 +02:00
|
|
|
__xen_pgd_unpin(&init_mm, (pgd_t *)page_address(page));
|
2008-05-27 00:31:27 +02:00
|
|
|
ClearPageSavePinned(page);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&pgd_lock, flags);
|
|
|
|
}
|
|
|
|
|
2007-07-18 03:37:04 +02:00
|
|
|
void xen_activate_mm(struct mm_struct *prev, struct mm_struct *next)
|
|
|
|
{
|
2007-07-18 03:37:05 +02:00
|
|
|
spin_lock(&next->page_table_lock);
|
2008-10-08 22:01:39 +02:00
|
|
|
xen_pgd_pin(next);
|
2007-07-18 03:37:05 +02:00
|
|
|
spin_unlock(&next->page_table_lock);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void xen_dup_mmap(struct mm_struct *oldmm, struct mm_struct *mm)
|
|
|
|
{
|
2007-07-18 03:37:05 +02:00
|
|
|
spin_lock(&mm->page_table_lock);
|
2008-10-08 22:01:39 +02:00
|
|
|
xen_pgd_pin(mm);
|
2007-07-18 03:37:05 +02:00
|
|
|
spin_unlock(&mm->page_table_lock);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
/* Another cpu may still have their %cr3 pointing at the pagetable, so
|
|
|
|
we need to repoint it somewhere else before we can unpin it. */
|
|
|
|
static void drop_other_mm_ref(void *info)
|
|
|
|
{
|
|
|
|
struct mm_struct *mm = info;
|
2008-07-09 00:06:40 +02:00
|
|
|
struct mm_struct *active_mm;
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2009-01-18 16:38:57 +01:00
|
|
|
active_mm = percpu_read(cpu_tlbstate.active_mm);
|
2008-07-09 00:06:40 +02:00
|
|
|
|
|
|
|
if (active_mm == mm)
|
2007-07-18 03:37:06 +02:00
|
|
|
leave_mm(smp_processor_id());
|
2007-10-16 20:51:30 +02:00
|
|
|
|
|
|
|
/* If this cpu still has a stale cr3 reference, then make sure
|
|
|
|
it has been flushed. */
|
2009-02-18 08:24:03 +01:00
|
|
|
if (percpu_read(xen_current_cr3) == __pa(mm->pgd))
|
2007-10-16 20:51:30 +02:00
|
|
|
load_cr3(swapper_pg_dir);
|
2007-07-18 03:37:06 +02:00
|
|
|
}
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2008-08-19 22:34:22 +02:00
|
|
|
static void xen_drop_mm_ref(struct mm_struct *mm)
|
2007-07-18 03:37:06 +02:00
|
|
|
{
|
2008-12-17 02:34:05 +01:00
|
|
|
cpumask_var_t mask;
|
2007-10-16 20:51:30 +02:00
|
|
|
unsigned cpu;
|
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
if (current->active_mm == mm) {
|
|
|
|
if (current->mm == mm)
|
|
|
|
load_cr3(swapper_pg_dir);
|
|
|
|
else
|
|
|
|
leave_mm(smp_processor_id());
|
2007-10-16 20:51:30 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Get the "official" set of cpus referring to our pagetable. */
|
2008-12-17 02:34:05 +01:00
|
|
|
if (!alloc_cpumask_var(&mask, GFP_ATOMIC)) {
|
|
|
|
for_each_online_cpu(cpu) {
|
2009-09-24 17:34:51 +02:00
|
|
|
if (!cpumask_test_cpu(cpu, mm_cpumask(mm))
|
2008-12-17 02:34:05 +01:00
|
|
|
&& per_cpu(xen_current_cr3, cpu) != __pa(mm->pgd))
|
|
|
|
continue;
|
|
|
|
smp_call_function_single(cpu, drop_other_mm_ref, mm, 1);
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
2009-09-24 17:34:51 +02:00
|
|
|
cpumask_copy(mask, mm_cpumask(mm));
|
2007-10-16 20:51:30 +02:00
|
|
|
|
|
|
|
/* It's possible that a vcpu may have a stale reference to our
|
|
|
|
cr3, because its in lazy mode, and it hasn't yet flushed
|
|
|
|
its set of pending hypercalls yet. In this case, we can
|
|
|
|
look at its actual current cr3 value, and force it to flush
|
|
|
|
if needed. */
|
|
|
|
for_each_online_cpu(cpu) {
|
|
|
|
if (per_cpu(xen_current_cr3, cpu) == __pa(mm->pgd))
|
2008-12-17 02:34:05 +01:00
|
|
|
cpumask_set_cpu(cpu, mask);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
|
|
|
|
2008-12-17 02:34:05 +01:00
|
|
|
if (!cpumask_empty(mask))
|
|
|
|
smp_call_function_many(mask, drop_other_mm_ref, mm, 1);
|
|
|
|
free_cpumask_var(mask);
|
2007-07-18 03:37:06 +02:00
|
|
|
}
|
|
|
|
#else
|
2008-08-19 22:34:22 +02:00
|
|
|
static void xen_drop_mm_ref(struct mm_struct *mm)
|
2007-07-18 03:37:06 +02:00
|
|
|
{
|
|
|
|
if (current->active_mm == mm)
|
|
|
|
load_cr3(swapper_pg_dir);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/*
|
|
|
|
* While a process runs, Xen pins its pagetables, which means that the
|
|
|
|
* hypervisor forces it to be read-only, and it controls all updates
|
|
|
|
* to it. This means that all pagetable updates have to go via the
|
|
|
|
* hypervisor, which is moderately expensive.
|
|
|
|
*
|
|
|
|
* Since we're pulling the pagetable down, we switch to use init_mm,
|
|
|
|
* unpin old process pagetable and mark it all read-write, which
|
|
|
|
* allows further operations on it to be simple memory accesses.
|
|
|
|
*
|
|
|
|
* The only subtle point is that another CPU may be still using the
|
|
|
|
* pagetable because of lazy tlb flushing. This means we need need to
|
|
|
|
* switch all CPUs off this pagetable before we can unpin it.
|
|
|
|
*/
|
|
|
|
void xen_exit_mmap(struct mm_struct *mm)
|
|
|
|
{
|
|
|
|
get_cpu(); /* make sure we don't move around */
|
2008-08-19 22:34:22 +02:00
|
|
|
xen_drop_mm_ref(mm);
|
2007-07-18 03:37:06 +02:00
|
|
|
put_cpu();
|
2007-07-18 03:37:04 +02:00
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
spin_lock(&mm->page_table_lock);
|
2007-09-25 20:50:00 +02:00
|
|
|
|
|
|
|
/* pgd may not be pinned in the error exit path of execve */
|
2008-08-19 22:34:22 +02:00
|
|
|
if (xen_page_pinned(mm->pgd))
|
2008-10-08 22:01:39 +02:00
|
|
|
xen_pgd_unpin(mm);
|
2007-10-16 20:51:30 +02:00
|
|
|
|
2007-07-18 03:37:06 +02:00
|
|
|
spin_unlock(&mm->page_table_lock);
|
2007-07-18 03:37:04 +02:00
|
|
|
}
|
2008-08-21 02:02:19 +02:00
|
|
|
|
2009-01-28 23:35:01 +01:00
|
|
|
static __init void xen_pagetable_setup_start(pgd_t *base)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
2009-08-20 13:13:52 +02:00
|
|
|
static void xen_post_allocator_init(void);
|
|
|
|
|
2009-01-28 23:35:01 +01:00
|
|
|
static __init void xen_pagetable_setup_done(pgd_t *base)
|
|
|
|
{
|
|
|
|
xen_setup_shared_info();
|
2009-08-20 13:13:52 +02:00
|
|
|
xen_post_allocator_init();
|
2009-01-28 23:35:01 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_write_cr2(unsigned long cr2)
|
|
|
|
{
|
|
|
|
percpu_read(xen_vcpu)->arch.cr2 = cr2;
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long xen_read_cr2(void)
|
|
|
|
{
|
|
|
|
return percpu_read(xen_vcpu)->arch.cr2;
|
|
|
|
}
|
|
|
|
|
|
|
|
unsigned long xen_read_cr2_direct(void)
|
|
|
|
{
|
|
|
|
return percpu_read(xen_vcpu_info.arch.cr2);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_flush_tlb(void)
|
|
|
|
{
|
|
|
|
struct mmuext_op *op;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
|
|
|
|
preempt_disable();
|
|
|
|
|
|
|
|
mcs = xen_mc_entry(sizeof(*op));
|
|
|
|
|
|
|
|
op = mcs.args;
|
|
|
|
op->cmd = MMUEXT_TLB_FLUSH_LOCAL;
|
|
|
|
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
|
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_flush_tlb_single(unsigned long addr)
|
|
|
|
{
|
|
|
|
struct mmuext_op *op;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
|
|
|
|
preempt_disable();
|
|
|
|
|
|
|
|
mcs = xen_mc_entry(sizeof(*op));
|
|
|
|
op = mcs.args;
|
|
|
|
op->cmd = MMUEXT_INVLPG_LOCAL;
|
|
|
|
op->arg1.linear_addr = addr & PAGE_MASK;
|
|
|
|
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
|
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_flush_tlb_others(const struct cpumask *cpus,
|
|
|
|
struct mm_struct *mm, unsigned long va)
|
|
|
|
{
|
|
|
|
struct {
|
|
|
|
struct mmuext_op op;
|
|
|
|
DECLARE_BITMAP(mask, NR_CPUS);
|
|
|
|
} *args;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
|
2009-03-05 02:36:57 +01:00
|
|
|
if (cpumask_empty(cpus))
|
|
|
|
return; /* nothing to do */
|
2009-01-28 23:35:01 +01:00
|
|
|
|
|
|
|
mcs = xen_mc_entry(sizeof(*args));
|
|
|
|
args = mcs.args;
|
|
|
|
args->op.arg2.vcpumask = to_cpumask(args->mask);
|
|
|
|
|
|
|
|
/* Remove us, and any offline CPUS. */
|
|
|
|
cpumask_and(to_cpumask(args->mask), cpus, cpu_online_mask);
|
|
|
|
cpumask_clear_cpu(smp_processor_id(), to_cpumask(args->mask));
|
|
|
|
|
|
|
|
if (va == TLB_FLUSH_ALL) {
|
|
|
|
args->op.cmd = MMUEXT_TLB_FLUSH_MULTI;
|
|
|
|
} else {
|
|
|
|
args->op.cmd = MMUEXT_INVLPG_MULTI;
|
|
|
|
args->op.arg1.linear_addr = va;
|
|
|
|
}
|
|
|
|
|
|
|
|
MULTI_mmuext_op(mcs.mc, &args->op, 1, NULL, DOMID_SELF);
|
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_MMU);
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long xen_read_cr3(void)
|
|
|
|
{
|
|
|
|
return percpu_read(xen_cr3);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void set_current_cr3(void *v)
|
|
|
|
{
|
|
|
|
percpu_write(xen_current_cr3, (unsigned long)v);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __xen_write_cr3(bool kernel, unsigned long cr3)
|
|
|
|
{
|
|
|
|
struct mmuext_op *op;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
unsigned long mfn;
|
|
|
|
|
|
|
|
if (cr3)
|
|
|
|
mfn = pfn_to_mfn(PFN_DOWN(cr3));
|
|
|
|
else
|
|
|
|
mfn = 0;
|
|
|
|
|
|
|
|
WARN_ON(mfn == 0 && kernel);
|
|
|
|
|
|
|
|
mcs = __xen_mc_entry(sizeof(*op));
|
|
|
|
|
|
|
|
op = mcs.args;
|
|
|
|
op->cmd = kernel ? MMUEXT_NEW_BASEPTR : MMUEXT_NEW_USER_BASEPTR;
|
|
|
|
op->arg1.mfn = mfn;
|
|
|
|
|
|
|
|
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
|
|
|
|
|
|
|
|
if (kernel) {
|
|
|
|
percpu_write(xen_cr3, cr3);
|
|
|
|
|
|
|
|
/* Update xen_current_cr3 once the batch has actually
|
|
|
|
been submitted. */
|
|
|
|
xen_mc_callback(set_current_cr3, (void *)cr3);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_write_cr3(unsigned long cr3)
|
|
|
|
{
|
|
|
|
BUG_ON(preemptible());
|
|
|
|
|
|
|
|
xen_mc_batch(); /* disables interrupts */
|
|
|
|
|
|
|
|
/* Update while interrupts are disabled, so its atomic with
|
|
|
|
respect to ipis */
|
|
|
|
percpu_write(xen_cr3, cr3);
|
|
|
|
|
|
|
|
__xen_write_cr3(true, cr3);
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
{
|
|
|
|
pgd_t *user_pgd = xen_get_user_pgd(__va(cr3));
|
|
|
|
if (user_pgd)
|
|
|
|
__xen_write_cr3(false, __pa(user_pgd));
|
|
|
|
else
|
|
|
|
__xen_write_cr3(false, 0);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_CPU); /* interrupts restored */
|
|
|
|
}
|
|
|
|
|
|
|
|
static int xen_pgd_alloc(struct mm_struct *mm)
|
|
|
|
{
|
|
|
|
pgd_t *pgd = mm->pgd;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
BUG_ON(PagePinned(virt_to_page(pgd)));
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
{
|
|
|
|
struct page *page = virt_to_page(pgd);
|
|
|
|
pgd_t *user_pgd;
|
|
|
|
|
|
|
|
BUG_ON(page->private != 0);
|
|
|
|
|
|
|
|
ret = -ENOMEM;
|
|
|
|
|
|
|
|
user_pgd = (pgd_t *)__get_free_page(GFP_KERNEL | __GFP_ZERO);
|
|
|
|
page->private = (unsigned long)user_pgd;
|
|
|
|
|
|
|
|
if (user_pgd != NULL) {
|
|
|
|
user_pgd[pgd_index(VSYSCALL_START)] =
|
|
|
|
__pgd(__pa(level3_user_vsyscall) | _PAGE_TABLE);
|
|
|
|
ret = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
BUG_ON(PagePinned(virt_to_page(xen_get_user_pgd(pgd))));
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_pgd_free(struct mm_struct *mm, pgd_t *pgd)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
pgd_t *user_pgd = xen_get_user_pgd(pgd);
|
|
|
|
|
|
|
|
if (user_pgd)
|
|
|
|
free_page((unsigned long)user_pgd);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2009-02-02 22:58:06 +01:00
|
|
|
#ifdef CONFIG_X86_32
|
|
|
|
static __init pte_t mask_rw_pte(pte_t *ptep, pte_t pte)
|
|
|
|
{
|
|
|
|
/* If there's an existing pte, then don't allow _PAGE_RW to be set */
|
|
|
|
if (pte_val_ma(*ptep) & _PAGE_PRESENT)
|
|
|
|
pte = __pte_ma(((pte_val_ma(*ptep) & _PAGE_RW) | ~_PAGE_RW) &
|
|
|
|
pte_val_ma(pte));
|
|
|
|
|
|
|
|
return pte;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Init-time set_pte while constructing initial pagetables, which
|
|
|
|
doesn't allow RO pagetable pages to be remapped RW */
|
|
|
|
static __init void xen_set_pte_init(pte_t *ptep, pte_t pte)
|
|
|
|
{
|
|
|
|
pte = mask_rw_pte(ptep, pte);
|
|
|
|
|
|
|
|
xen_set_pte(ptep, pte);
|
|
|
|
}
|
|
|
|
#endif
|
2009-01-28 23:35:01 +01:00
|
|
|
|
2009-03-17 21:30:55 +01:00
|
|
|
static void pin_pagetable_pfn(unsigned cmd, unsigned long pfn)
|
|
|
|
{
|
|
|
|
struct mmuext_op op;
|
|
|
|
op.cmd = cmd;
|
|
|
|
op.arg1.mfn = pfn_to_mfn(pfn);
|
|
|
|
if (HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF))
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
2009-01-28 23:35:01 +01:00
|
|
|
/* Early in boot, while setting up the initial pagetable, assume
|
|
|
|
everything is pinned. */
|
|
|
|
static __init void xen_alloc_pte_init(struct mm_struct *mm, unsigned long pfn)
|
|
|
|
{
|
2009-03-17 21:30:55 +01:00
|
|
|
#ifdef CONFIG_FLATMEM
|
|
|
|
BUG_ON(mem_map); /* should only be used early */
|
|
|
|
#endif
|
|
|
|
make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
|
|
|
|
pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Used for pmd and pud */
|
|
|
|
static __init void xen_alloc_pmd_init(struct mm_struct *mm, unsigned long pfn)
|
|
|
|
{
|
2009-01-28 23:35:01 +01:00
|
|
|
#ifdef CONFIG_FLATMEM
|
|
|
|
BUG_ON(mem_map); /* should only be used early */
|
|
|
|
#endif
|
|
|
|
make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Early release_pte assumes that all pts are pinned, since there's
|
|
|
|
only init_mm and anything attached to that is pinned. */
|
2009-03-17 21:30:55 +01:00
|
|
|
static __init void xen_release_pte_init(unsigned long pfn)
|
2009-01-28 23:35:01 +01:00
|
|
|
{
|
2009-03-17 21:30:55 +01:00
|
|
|
pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn);
|
2009-01-28 23:35:01 +01:00
|
|
|
make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
|
|
|
|
}
|
|
|
|
|
2009-03-17 21:30:55 +01:00
|
|
|
static __init void xen_release_pmd_init(unsigned long pfn)
|
2009-01-28 23:35:01 +01:00
|
|
|
{
|
2009-03-17 21:30:55 +01:00
|
|
|
make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
|
2009-01-28 23:35:01 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/* This needs to make sure the new pte page is pinned iff its being
|
|
|
|
attached to a pinned pagetable. */
|
|
|
|
static void xen_alloc_ptpage(struct mm_struct *mm, unsigned long pfn, unsigned level)
|
|
|
|
{
|
|
|
|
struct page *page = pfn_to_page(pfn);
|
|
|
|
|
|
|
|
if (PagePinned(virt_to_page(mm->pgd))) {
|
|
|
|
SetPagePinned(page);
|
|
|
|
|
|
|
|
if (!PageHighMem(page)) {
|
|
|
|
make_lowmem_page_readonly(__va(PFN_PHYS((unsigned long)pfn)));
|
|
|
|
if (level == PT_PTE && USE_SPLIT_PTLOCKS)
|
|
|
|
pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
|
|
|
|
} else {
|
|
|
|
/* make sure there are no stray mappings of
|
|
|
|
this page */
|
|
|
|
kmap_flush_unused();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_alloc_pte(struct mm_struct *mm, unsigned long pfn)
|
|
|
|
{
|
|
|
|
xen_alloc_ptpage(mm, pfn, PT_PTE);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_alloc_pmd(struct mm_struct *mm, unsigned long pfn)
|
|
|
|
{
|
|
|
|
xen_alloc_ptpage(mm, pfn, PT_PMD);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* This should never happen until we're OK to use struct page */
|
|
|
|
static void xen_release_ptpage(unsigned long pfn, unsigned level)
|
|
|
|
{
|
|
|
|
struct page *page = pfn_to_page(pfn);
|
|
|
|
|
|
|
|
if (PagePinned(page)) {
|
|
|
|
if (!PageHighMem(page)) {
|
|
|
|
if (level == PT_PTE && USE_SPLIT_PTLOCKS)
|
|
|
|
pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn);
|
|
|
|
make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
|
|
|
|
}
|
|
|
|
ClearPagePinned(page);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_release_pte(unsigned long pfn)
|
|
|
|
{
|
|
|
|
xen_release_ptpage(pfn, PT_PTE);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_release_pmd(unsigned long pfn)
|
|
|
|
{
|
|
|
|
xen_release_ptpage(pfn, PT_PMD);
|
|
|
|
}
|
|
|
|
|
|
|
|
#if PAGETABLE_LEVELS == 4
|
|
|
|
static void xen_alloc_pud(struct mm_struct *mm, unsigned long pfn)
|
|
|
|
{
|
|
|
|
xen_alloc_ptpage(mm, pfn, PT_PUD);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xen_release_pud(unsigned long pfn)
|
|
|
|
{
|
|
|
|
xen_release_ptpage(pfn, PT_PUD);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
void __init xen_reserve_top(void)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_32
|
|
|
|
unsigned long top = HYPERVISOR_VIRT_START;
|
|
|
|
struct xen_platform_parameters pp;
|
|
|
|
|
|
|
|
if (HYPERVISOR_xen_version(XENVER_platform_parameters, &pp) == 0)
|
|
|
|
top = pp.virt_start;
|
|
|
|
|
|
|
|
reserve_top_address(-top);
|
|
|
|
#endif /* CONFIG_X86_32 */
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Like __va(), but returns address in the kernel mapping (which is
|
|
|
|
* all we have until the physical memory mapping has been set up.
|
|
|
|
*/
|
|
|
|
static void *__ka(phys_addr_t paddr)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
return (void *)(paddr + __START_KERNEL_map);
|
|
|
|
#else
|
|
|
|
return __va(paddr);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Convert a machine address to physical address */
|
|
|
|
static unsigned long m2p(phys_addr_t maddr)
|
|
|
|
{
|
|
|
|
phys_addr_t paddr;
|
|
|
|
|
|
|
|
maddr &= PTE_PFN_MASK;
|
|
|
|
paddr = mfn_to_pfn(maddr >> PAGE_SHIFT) << PAGE_SHIFT;
|
|
|
|
|
|
|
|
return paddr;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Convert a machine address to kernel virtual */
|
|
|
|
static void *m2v(phys_addr_t maddr)
|
|
|
|
{
|
|
|
|
return __ka(m2p(maddr));
|
|
|
|
}
|
|
|
|
|
|
|
|
static void set_page_prot(void *addr, pgprot_t prot)
|
|
|
|
{
|
|
|
|
unsigned long pfn = __pa(addr) >> PAGE_SHIFT;
|
|
|
|
pte_t pte = pfn_pte(pfn, prot);
|
|
|
|
|
|
|
|
if (HYPERVISOR_update_va_mapping((unsigned long)addr, pte, 0))
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn)
|
|
|
|
{
|
|
|
|
unsigned pmdidx, pteidx;
|
|
|
|
unsigned ident_pte;
|
|
|
|
unsigned long pfn;
|
|
|
|
|
|
|
|
ident_pte = 0;
|
|
|
|
pfn = 0;
|
|
|
|
for (pmdidx = 0; pmdidx < PTRS_PER_PMD && pfn < max_pfn; pmdidx++) {
|
|
|
|
pte_t *pte_page;
|
|
|
|
|
|
|
|
/* Reuse or allocate a page of ptes */
|
|
|
|
if (pmd_present(pmd[pmdidx]))
|
|
|
|
pte_page = m2v(pmd[pmdidx].pmd);
|
|
|
|
else {
|
|
|
|
/* Check for free pte pages */
|
|
|
|
if (ident_pte == ARRAY_SIZE(level1_ident_pgt))
|
|
|
|
break;
|
|
|
|
|
|
|
|
pte_page = &level1_ident_pgt[ident_pte];
|
|
|
|
ident_pte += PTRS_PER_PTE;
|
|
|
|
|
|
|
|
pmd[pmdidx] = __pmd(__pa(pte_page) | _PAGE_TABLE);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Install mappings */
|
|
|
|
for (pteidx = 0; pteidx < PTRS_PER_PTE; pteidx++, pfn++) {
|
|
|
|
pte_t pte;
|
|
|
|
|
|
|
|
if (pfn > max_pfn_mapped)
|
|
|
|
max_pfn_mapped = pfn;
|
|
|
|
|
|
|
|
if (!pte_none(pte_page[pteidx]))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
pte = pfn_pte(pfn, PAGE_KERNEL_EXEC);
|
|
|
|
pte_page[pteidx] = pte;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (pteidx = 0; pteidx < ident_pte; pteidx += PTRS_PER_PTE)
|
|
|
|
set_page_prot(&level1_ident_pgt[pteidx], PAGE_KERNEL_RO);
|
|
|
|
|
|
|
|
set_page_prot(pmd, PAGE_KERNEL_RO);
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
static void convert_pfn_mfn(void *v)
|
|
|
|
{
|
|
|
|
pte_t *pte = v;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
/* All levels are converted the same way, so just treat them
|
|
|
|
as ptes. */
|
|
|
|
for (i = 0; i < PTRS_PER_PTE; i++)
|
|
|
|
pte[i] = xen_make_pte(pte[i].pte);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Set up the inital kernel pagetable.
|
|
|
|
*
|
|
|
|
* We can construct this by grafting the Xen provided pagetable into
|
|
|
|
* head_64.S's preconstructed pagetables. We copy the Xen L2's into
|
|
|
|
* level2_ident_pgt, level2_kernel_pgt and level2_fixmap_pgt. This
|
|
|
|
* means that only the kernel has a physical mapping to start with -
|
|
|
|
* but that's enough to get __va working. We need to fill in the rest
|
|
|
|
* of the physical mapping once some sort of allocator has been set
|
|
|
|
* up.
|
|
|
|
*/
|
|
|
|
__init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd,
|
|
|
|
unsigned long max_pfn)
|
|
|
|
{
|
|
|
|
pud_t *l3;
|
|
|
|
pmd_t *l2;
|
|
|
|
|
|
|
|
/* Zap identity mapping */
|
|
|
|
init_level4_pgt[0] = __pgd(0);
|
|
|
|
|
|
|
|
/* Pre-constructed entries are in pfn, so convert to mfn */
|
|
|
|
convert_pfn_mfn(init_level4_pgt);
|
|
|
|
convert_pfn_mfn(level3_ident_pgt);
|
|
|
|
convert_pfn_mfn(level3_kernel_pgt);
|
|
|
|
|
|
|
|
l3 = m2v(pgd[pgd_index(__START_KERNEL_map)].pgd);
|
|
|
|
l2 = m2v(l3[pud_index(__START_KERNEL_map)].pud);
|
|
|
|
|
|
|
|
memcpy(level2_ident_pgt, l2, sizeof(pmd_t) * PTRS_PER_PMD);
|
|
|
|
memcpy(level2_kernel_pgt, l2, sizeof(pmd_t) * PTRS_PER_PMD);
|
|
|
|
|
|
|
|
l3 = m2v(pgd[pgd_index(__START_KERNEL_map + PMD_SIZE)].pgd);
|
|
|
|
l2 = m2v(l3[pud_index(__START_KERNEL_map + PMD_SIZE)].pud);
|
|
|
|
memcpy(level2_fixmap_pgt, l2, sizeof(pmd_t) * PTRS_PER_PMD);
|
|
|
|
|
|
|
|
/* Set up identity map */
|
|
|
|
xen_map_identity_early(level2_ident_pgt, max_pfn);
|
|
|
|
|
|
|
|
/* Make pagetable pieces RO */
|
|
|
|
set_page_prot(init_level4_pgt, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(level3_ident_pgt, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(level3_kernel_pgt, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(level3_user_vsyscall, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(level2_kernel_pgt, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(level2_fixmap_pgt, PAGE_KERNEL_RO);
|
|
|
|
|
|
|
|
/* Pin down new L4 */
|
|
|
|
pin_pagetable_pfn(MMUEXT_PIN_L4_TABLE,
|
|
|
|
PFN_DOWN(__pa_symbol(init_level4_pgt)));
|
|
|
|
|
|
|
|
/* Unpin Xen-provided one */
|
|
|
|
pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(pgd)));
|
|
|
|
|
|
|
|
/* Switch over */
|
|
|
|
pgd = init_level4_pgt;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* At this stage there can be no user pgd, and no page
|
|
|
|
* structure to attach it to, so make sure we just set kernel
|
|
|
|
* pgd.
|
|
|
|
*/
|
|
|
|
xen_mc_batch();
|
|
|
|
__xen_write_cr3(true, __pa(pgd));
|
|
|
|
xen_mc_issue(PARAVIRT_LAZY_CPU);
|
|
|
|
|
|
|
|
reserve_early(__pa(xen_start_info->pt_base),
|
|
|
|
__pa(xen_start_info->pt_base +
|
|
|
|
xen_start_info->nr_pt_frames * PAGE_SIZE),
|
|
|
|
"XEN PAGETABLES");
|
|
|
|
|
|
|
|
return pgd;
|
|
|
|
}
|
|
|
|
#else /* !CONFIG_X86_64 */
|
|
|
|
static pmd_t level2_kernel_pgt[PTRS_PER_PMD] __page_aligned_bss;
|
|
|
|
|
|
|
|
__init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd,
|
|
|
|
unsigned long max_pfn)
|
|
|
|
{
|
|
|
|
pmd_t *kernel_pmd;
|
|
|
|
|
x86: add brk allocation for very, very early allocations
Impact: new interface
Add a brk()-like allocator which effectively extends the bss in order
to allow very early code to do dynamic allocations. This is better than
using statically allocated arrays for data in subsystems which may never
get used.
The space for brk allocations is in the bss ELF segment, so that the
space is mapped properly by the code which maps the kernel, and so
that bootloaders keep the space free rather than putting a ramdisk or
something into it.
The bss itself, delimited by __bss_stop, ends before the brk area
(__brk_base to __brk_limit). The kernel text, data and bss is reserved
up to __bss_stop.
Any brk-allocated data is reserved separately just before the kernel
pagetable is built, as that code allocates from unreserved spaces
in the e820 map, potentially allocating from any unused brk memory.
Ultimately any unused memory in the brk area is used in the general
kernel memory pool.
Initially the brk space is set to 1MB, which is probably much larger
than any user needs (the largest current user is i386 head_32.S's code
to build the pagetables to map the kernel, which can get fairly large
with a big kernel image and no PSE support). So long as the system
has sufficient memory for the bootloader to reserve the kernel+1MB brk,
there are no bad effects resulting from an over-large brk.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
2009-02-27 02:35:44 +01:00
|
|
|
max_pfn_mapped = PFN_DOWN(__pa(xen_start_info->pt_base) +
|
|
|
|
xen_start_info->nr_pt_frames * PAGE_SIZE +
|
|
|
|
512*1024);
|
2009-01-28 23:35:01 +01:00
|
|
|
|
|
|
|
kernel_pmd = m2v(pgd[KERNEL_PGD_BOUNDARY].pgd);
|
|
|
|
memcpy(level2_kernel_pgt, kernel_pmd, sizeof(pmd_t) * PTRS_PER_PMD);
|
|
|
|
|
|
|
|
xen_map_identity_early(level2_kernel_pgt, max_pfn);
|
|
|
|
|
|
|
|
memcpy(swapper_pg_dir, pgd, sizeof(pgd_t) * PTRS_PER_PGD);
|
|
|
|
set_pgd(&swapper_pg_dir[KERNEL_PGD_BOUNDARY],
|
|
|
|
__pgd(__pa(level2_kernel_pgt) | _PAGE_PRESENT));
|
|
|
|
|
|
|
|
set_page_prot(level2_kernel_pgt, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(swapper_pg_dir, PAGE_KERNEL_RO);
|
|
|
|
set_page_prot(empty_zero_page, PAGE_KERNEL_RO);
|
|
|
|
|
|
|
|
pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(pgd)));
|
|
|
|
|
|
|
|
xen_write_cr3(__pa(swapper_pg_dir));
|
|
|
|
|
|
|
|
pin_pagetable_pfn(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(swapper_pg_dir)));
|
|
|
|
|
2009-05-07 20:56:44 +02:00
|
|
|
reserve_early(__pa(xen_start_info->pt_base),
|
|
|
|
__pa(xen_start_info->pt_base +
|
|
|
|
xen_start_info->nr_pt_frames * PAGE_SIZE),
|
|
|
|
"XEN PAGETABLES");
|
|
|
|
|
2009-01-28 23:35:01 +01:00
|
|
|
return swapper_pg_dir;
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_X86_64 */
|
|
|
|
|
2009-04-09 19:55:33 +02:00
|
|
|
static void xen_set_fixmap(unsigned idx, phys_addr_t phys, pgprot_t prot)
|
2009-01-28 23:35:01 +01:00
|
|
|
{
|
|
|
|
pte_t pte;
|
|
|
|
|
|
|
|
phys >>= PAGE_SHIFT;
|
|
|
|
|
|
|
|
switch (idx) {
|
|
|
|
case FIX_BTMAP_END ... FIX_BTMAP_BEGIN:
|
|
|
|
#ifdef CONFIG_X86_F00F_BUG
|
|
|
|
case FIX_F00F_IDT:
|
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_X86_32
|
|
|
|
case FIX_WP_TEST:
|
|
|
|
case FIX_VDSO:
|
|
|
|
# ifdef CONFIG_HIGHMEM
|
|
|
|
case FIX_KMAP_BEGIN ... FIX_KMAP_END:
|
|
|
|
# endif
|
|
|
|
#else
|
|
|
|
case VSYSCALL_LAST_PAGE ... VSYSCALL_FIRST_PAGE:
|
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_X86_LOCAL_APIC
|
|
|
|
case FIX_APIC_BASE: /* maps dummy local APIC */
|
|
|
|
#endif
|
2009-03-08 08:48:41 +01:00
|
|
|
case FIX_TEXT_POKE0:
|
|
|
|
case FIX_TEXT_POKE1:
|
|
|
|
/* All local page mappings */
|
2009-01-28 23:35:01 +01:00
|
|
|
pte = pfn_pte(phys, prot);
|
|
|
|
break;
|
|
|
|
|
2010-02-04 23:46:34 +01:00
|
|
|
case FIX_PARAVIRT_BOOTMAP:
|
|
|
|
/* This is an MFN, but it isn't an IO mapping from the
|
|
|
|
IO domain */
|
2009-01-28 23:35:01 +01:00
|
|
|
pte = mfn_pte(phys, prot);
|
|
|
|
break;
|
2010-02-04 23:46:34 +01:00
|
|
|
|
|
|
|
default:
|
|
|
|
/* By default, set_fixmap is used for hardware mappings */
|
|
|
|
pte = mfn_pte(phys, __pgprot(pgprot_val(prot) | _PAGE_IOMAP));
|
|
|
|
break;
|
2009-01-28 23:35:01 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
__native_set_fixmap(idx, pte);
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
/* Replicate changes to map the vsyscall page into the user
|
|
|
|
pagetable vsyscall mapping. */
|
|
|
|
if (idx >= VSYSCALL_LAST_PAGE && idx <= VSYSCALL_FIRST_PAGE) {
|
|
|
|
unsigned long vaddr = __fix_to_virt(idx);
|
|
|
|
set_pte_vaddr_pud(level3_user_vsyscall, vaddr, pte);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2009-08-20 13:13:52 +02:00
|
|
|
static __init void xen_post_allocator_init(void)
|
2009-01-28 23:35:01 +01:00
|
|
|
{
|
|
|
|
pv_mmu_ops.set_pte = xen_set_pte;
|
|
|
|
pv_mmu_ops.set_pmd = xen_set_pmd;
|
|
|
|
pv_mmu_ops.set_pud = xen_set_pud;
|
|
|
|
#if PAGETABLE_LEVELS == 4
|
|
|
|
pv_mmu_ops.set_pgd = xen_set_pgd;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/* This will work as long as patching hasn't happened yet
|
|
|
|
(which it hasn't) */
|
|
|
|
pv_mmu_ops.alloc_pte = xen_alloc_pte;
|
|
|
|
pv_mmu_ops.alloc_pmd = xen_alloc_pmd;
|
|
|
|
pv_mmu_ops.release_pte = xen_release_pte;
|
|
|
|
pv_mmu_ops.release_pmd = xen_release_pmd;
|
|
|
|
#if PAGETABLE_LEVELS == 4
|
|
|
|
pv_mmu_ops.alloc_pud = xen_alloc_pud;
|
|
|
|
pv_mmu_ops.release_pud = xen_release_pud;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
SetPagePinned(virt_to_page(level3_user_vsyscall));
|
|
|
|
#endif
|
|
|
|
xen_mark_init_mm_pinned();
|
|
|
|
}
|
|
|
|
|
2009-02-18 08:46:21 +01:00
|
|
|
static void xen_leave_lazy_mmu(void)
|
|
|
|
{
|
2009-02-21 08:01:26 +01:00
|
|
|
preempt_disable();
|
2009-02-18 08:46:21 +01:00
|
|
|
xen_mc_flush();
|
|
|
|
paravirt_leave_lazy_mmu();
|
2009-02-21 08:01:26 +01:00
|
|
|
preempt_enable();
|
2009-02-18 08:46:21 +01:00
|
|
|
}
|
2009-01-28 23:35:01 +01:00
|
|
|
|
2009-08-20 14:30:02 +02:00
|
|
|
static const struct pv_mmu_ops xen_mmu_ops __initdata = {
|
2009-01-28 23:35:01 +01:00
|
|
|
.read_cr2 = xen_read_cr2,
|
|
|
|
.write_cr2 = xen_write_cr2,
|
|
|
|
|
|
|
|
.read_cr3 = xen_read_cr3,
|
|
|
|
.write_cr3 = xen_write_cr3,
|
|
|
|
|
|
|
|
.flush_tlb_user = xen_flush_tlb,
|
|
|
|
.flush_tlb_kernel = xen_flush_tlb,
|
|
|
|
.flush_tlb_single = xen_flush_tlb_single,
|
|
|
|
.flush_tlb_others = xen_flush_tlb_others,
|
|
|
|
|
|
|
|
.pte_update = paravirt_nop,
|
|
|
|
.pte_update_defer = paravirt_nop,
|
|
|
|
|
|
|
|
.pgd_alloc = xen_pgd_alloc,
|
|
|
|
.pgd_free = xen_pgd_free,
|
|
|
|
|
|
|
|
.alloc_pte = xen_alloc_pte_init,
|
|
|
|
.release_pte = xen_release_pte_init,
|
2009-03-17 21:30:55 +01:00
|
|
|
.alloc_pmd = xen_alloc_pmd_init,
|
2009-01-28 23:35:01 +01:00
|
|
|
.alloc_pmd_clone = paravirt_nop,
|
2009-03-17 21:30:55 +01:00
|
|
|
.release_pmd = xen_release_pmd_init,
|
2009-01-28 23:35:01 +01:00
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
.set_pte = xen_set_pte,
|
|
|
|
#else
|
|
|
|
.set_pte = xen_set_pte_init,
|
|
|
|
#endif
|
|
|
|
.set_pte_at = xen_set_pte_at,
|
|
|
|
.set_pmd = xen_set_pmd_hyper,
|
|
|
|
|
|
|
|
.ptep_modify_prot_start = __ptep_modify_prot_start,
|
|
|
|
.ptep_modify_prot_commit = __ptep_modify_prot_commit,
|
|
|
|
|
2009-01-28 23:35:07 +01:00
|
|
|
.pte_val = PV_CALLEE_SAVE(xen_pte_val),
|
|
|
|
.pgd_val = PV_CALLEE_SAVE(xen_pgd_val),
|
2009-01-28 23:35:01 +01:00
|
|
|
|
2009-01-28 23:35:07 +01:00
|
|
|
.make_pte = PV_CALLEE_SAVE(xen_make_pte),
|
|
|
|
.make_pgd = PV_CALLEE_SAVE(xen_make_pgd),
|
2009-01-28 23:35:01 +01:00
|
|
|
|
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
.set_pte_atomic = xen_set_pte_atomic,
|
|
|
|
.pte_clear = xen_pte_clear,
|
|
|
|
.pmd_clear = xen_pmd_clear,
|
|
|
|
#endif /* CONFIG_X86_PAE */
|
|
|
|
.set_pud = xen_set_pud_hyper,
|
|
|
|
|
2009-01-28 23:35:07 +01:00
|
|
|
.make_pmd = PV_CALLEE_SAVE(xen_make_pmd),
|
|
|
|
.pmd_val = PV_CALLEE_SAVE(xen_pmd_val),
|
2009-01-28 23:35:01 +01:00
|
|
|
|
|
|
|
#if PAGETABLE_LEVELS == 4
|
2009-01-28 23:35:07 +01:00
|
|
|
.pud_val = PV_CALLEE_SAVE(xen_pud_val),
|
|
|
|
.make_pud = PV_CALLEE_SAVE(xen_make_pud),
|
2009-01-28 23:35:01 +01:00
|
|
|
.set_pgd = xen_set_pgd_hyper,
|
|
|
|
|
2009-03-17 21:30:55 +01:00
|
|
|
.alloc_pud = xen_alloc_pmd_init,
|
|
|
|
.release_pud = xen_release_pmd_init,
|
2009-01-28 23:35:01 +01:00
|
|
|
#endif /* PAGETABLE_LEVELS == 4 */
|
|
|
|
|
|
|
|
.activate_mm = xen_activate_mm,
|
|
|
|
.dup_mmap = xen_dup_mmap,
|
|
|
|
.exit_mmap = xen_exit_mmap,
|
|
|
|
|
|
|
|
.lazy_mode = {
|
|
|
|
.enter = paravirt_enter_lazy_mmu,
|
2009-02-18 08:46:21 +01:00
|
|
|
.leave = xen_leave_lazy_mmu,
|
2009-01-28 23:35:01 +01:00
|
|
|
},
|
|
|
|
|
|
|
|
.set_fixmap = xen_set_fixmap,
|
|
|
|
};
|
|
|
|
|
2009-08-20 14:30:02 +02:00
|
|
|
void __init xen_init_mmu_ops(void)
|
|
|
|
{
|
|
|
|
x86_init.paging.pagetable_setup_start = xen_pagetable_setup_start;
|
|
|
|
x86_init.paging.pagetable_setup_done = xen_pagetable_setup_done;
|
|
|
|
pv_mmu_ops = xen_mmu_ops;
|
2010-03-26 23:37:50 +01:00
|
|
|
|
|
|
|
vmap_lazy_unmap = false;
|
2009-08-20 14:30:02 +02:00
|
|
|
}
|
2009-01-28 23:35:01 +01:00
|
|
|
|
2009-02-09 21:05:46 +01:00
|
|
|
/* Protected by xen_reservation_lock. */
|
|
|
|
#define MAX_CONTIG_ORDER 9 /* 2MB */
|
|
|
|
static unsigned long discontig_frames[1<<MAX_CONTIG_ORDER];
|
|
|
|
|
|
|
|
#define VOID_PTE (mfn_pte(0, __pgprot(0)))
|
|
|
|
static void xen_zap_pfn_range(unsigned long vaddr, unsigned int order,
|
|
|
|
unsigned long *in_frames,
|
|
|
|
unsigned long *out_frames)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
struct multicall_space mcs;
|
|
|
|
|
|
|
|
xen_mc_batch();
|
|
|
|
for (i = 0; i < (1UL<<order); i++, vaddr += PAGE_SIZE) {
|
|
|
|
mcs = __xen_mc_entry(0);
|
|
|
|
|
|
|
|
if (in_frames)
|
|
|
|
in_frames[i] = virt_to_mfn(vaddr);
|
|
|
|
|
|
|
|
MULTI_update_va_mapping(mcs.mc, vaddr, VOID_PTE, 0);
|
|
|
|
set_phys_to_machine(virt_to_pfn(vaddr), INVALID_P2M_ENTRY);
|
|
|
|
|
|
|
|
if (out_frames)
|
|
|
|
out_frames[i] = virt_to_pfn(vaddr);
|
|
|
|
}
|
|
|
|
xen_mc_issue(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Update the pfn-to-mfn mappings for a virtual address range, either to
|
|
|
|
* point to an array of mfns, or contiguously from a single starting
|
|
|
|
* mfn.
|
|
|
|
*/
|
|
|
|
static void xen_remap_exchanged_ptes(unsigned long vaddr, int order,
|
|
|
|
unsigned long *mfns,
|
|
|
|
unsigned long first_mfn)
|
|
|
|
{
|
|
|
|
unsigned i, limit;
|
|
|
|
unsigned long mfn;
|
|
|
|
|
|
|
|
xen_mc_batch();
|
|
|
|
|
|
|
|
limit = 1u << order;
|
|
|
|
for (i = 0; i < limit; i++, vaddr += PAGE_SIZE) {
|
|
|
|
struct multicall_space mcs;
|
|
|
|
unsigned flags;
|
|
|
|
|
|
|
|
mcs = __xen_mc_entry(0);
|
|
|
|
if (mfns)
|
|
|
|
mfn = mfns[i];
|
|
|
|
else
|
|
|
|
mfn = first_mfn + i;
|
|
|
|
|
|
|
|
if (i < (limit - 1))
|
|
|
|
flags = 0;
|
|
|
|
else {
|
|
|
|
if (order == 0)
|
|
|
|
flags = UVMF_INVLPG | UVMF_ALL;
|
|
|
|
else
|
|
|
|
flags = UVMF_TLB_FLUSH | UVMF_ALL;
|
|
|
|
}
|
|
|
|
|
|
|
|
MULTI_update_va_mapping(mcs.mc, vaddr,
|
|
|
|
mfn_pte(mfn, PAGE_KERNEL), flags);
|
|
|
|
|
|
|
|
set_phys_to_machine(virt_to_pfn(vaddr), mfn);
|
|
|
|
}
|
|
|
|
|
|
|
|
xen_mc_issue(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Perform the hypercall to exchange a region of our pfns to point to
|
|
|
|
* memory with the required contiguous alignment. Takes the pfns as
|
|
|
|
* input, and populates mfns as output.
|
|
|
|
*
|
|
|
|
* Returns a success code indicating whether the hypervisor was able to
|
|
|
|
* satisfy the request or not.
|
|
|
|
*/
|
|
|
|
static int xen_exchange_memory(unsigned long extents_in, unsigned int order_in,
|
|
|
|
unsigned long *pfns_in,
|
|
|
|
unsigned long extents_out,
|
|
|
|
unsigned int order_out,
|
|
|
|
unsigned long *mfns_out,
|
|
|
|
unsigned int address_bits)
|
|
|
|
{
|
|
|
|
long rc;
|
|
|
|
int success;
|
|
|
|
|
|
|
|
struct xen_memory_exchange exchange = {
|
|
|
|
.in = {
|
|
|
|
.nr_extents = extents_in,
|
|
|
|
.extent_order = order_in,
|
|
|
|
.extent_start = pfns_in,
|
|
|
|
.domid = DOMID_SELF
|
|
|
|
},
|
|
|
|
.out = {
|
|
|
|
.nr_extents = extents_out,
|
|
|
|
.extent_order = order_out,
|
|
|
|
.extent_start = mfns_out,
|
|
|
|
.address_bits = address_bits,
|
|
|
|
.domid = DOMID_SELF
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
BUG_ON(extents_in << order_in != extents_out << order_out);
|
|
|
|
|
|
|
|
rc = HYPERVISOR_memory_op(XENMEM_exchange, &exchange);
|
|
|
|
success = (exchange.nr_exchanged == extents_in);
|
|
|
|
|
|
|
|
BUG_ON(!success && ((exchange.nr_exchanged != 0) || (rc == 0)));
|
|
|
|
BUG_ON(success && (rc != 0));
|
|
|
|
|
|
|
|
return success;
|
|
|
|
}
|
|
|
|
|
|
|
|
int xen_create_contiguous_region(unsigned long vstart, unsigned int order,
|
|
|
|
unsigned int address_bits)
|
|
|
|
{
|
|
|
|
unsigned long *in_frames = discontig_frames, out_frame;
|
|
|
|
unsigned long flags;
|
|
|
|
int success;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Currently an auto-translated guest will not perform I/O, nor will
|
|
|
|
* it require PAE page directories below 4GB. Therefore any calls to
|
|
|
|
* this function are redundant and can be ignored.
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (xen_feature(XENFEAT_auto_translated_physmap))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (unlikely(order > MAX_CONTIG_ORDER))
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
memset((void *) vstart, 0, PAGE_SIZE << order);
|
|
|
|
|
|
|
|
spin_lock_irqsave(&xen_reservation_lock, flags);
|
|
|
|
|
|
|
|
/* 1. Zap current PTEs, remembering MFNs. */
|
|
|
|
xen_zap_pfn_range(vstart, order, in_frames, NULL);
|
|
|
|
|
|
|
|
/* 2. Get a new contiguous memory extent. */
|
|
|
|
out_frame = virt_to_pfn(vstart);
|
|
|
|
success = xen_exchange_memory(1UL << order, 0, in_frames,
|
|
|
|
1, order, &out_frame,
|
|
|
|
address_bits);
|
|
|
|
|
|
|
|
/* 3. Map the new extent in place of old pages. */
|
|
|
|
if (success)
|
|
|
|
xen_remap_exchanged_ptes(vstart, order, NULL, out_frame);
|
|
|
|
else
|
|
|
|
xen_remap_exchanged_ptes(vstart, order, in_frames, 0);
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xen_reservation_lock, flags);
|
|
|
|
|
|
|
|
return success ? 0 : -ENOMEM;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(xen_create_contiguous_region);
|
|
|
|
|
|
|
|
void xen_destroy_contiguous_region(unsigned long vstart, unsigned int order)
|
|
|
|
{
|
|
|
|
unsigned long *out_frames = discontig_frames, in_frame;
|
|
|
|
unsigned long flags;
|
|
|
|
int success;
|
|
|
|
|
|
|
|
if (xen_feature(XENFEAT_auto_translated_physmap))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (unlikely(order > MAX_CONTIG_ORDER))
|
|
|
|
return;
|
|
|
|
|
|
|
|
memset((void *) vstart, 0, PAGE_SIZE << order);
|
|
|
|
|
|
|
|
spin_lock_irqsave(&xen_reservation_lock, flags);
|
|
|
|
|
|
|
|
/* 1. Find start MFN of contiguous extent. */
|
|
|
|
in_frame = virt_to_mfn(vstart);
|
|
|
|
|
|
|
|
/* 2. Zap current PTEs. */
|
|
|
|
xen_zap_pfn_range(vstart, order, NULL, out_frames);
|
|
|
|
|
|
|
|
/* 3. Do the exchange for non-contiguous MFNs. */
|
|
|
|
success = xen_exchange_memory(1, order, &in_frame, 1UL << order,
|
|
|
|
0, out_frames, 0);
|
|
|
|
|
|
|
|
/* 4. Map new pages in place of old pages. */
|
|
|
|
if (success)
|
|
|
|
xen_remap_exchanged_ptes(vstart, order, out_frames, 0);
|
|
|
|
else
|
|
|
|
xen_remap_exchanged_ptes(vstart, order, NULL, in_frame);
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&xen_reservation_lock, flags);
|
2009-08-20 14:30:02 +02:00
|
|
|
}
|
2009-02-09 21:05:46 +01:00
|
|
|
EXPORT_SYMBOL_GPL(xen_destroy_contiguous_region);
|
2009-01-28 23:35:01 +01:00
|
|
|
|
2010-07-29 15:37:48 +02:00
|
|
|
#ifdef CONFIG_XEN_PVHVM
|
2010-06-17 15:22:52 +02:00
|
|
|
static void xen_hvm_exit_mmap(struct mm_struct *mm)
|
|
|
|
{
|
|
|
|
struct xen_hvm_pagetable_dying a;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
a.domid = DOMID_SELF;
|
|
|
|
a.gpa = __pa(mm->pgd);
|
|
|
|
rc = HYPERVISOR_hvm_op(HVMOP_pagetable_dying, &a);
|
|
|
|
WARN_ON_ONCE(rc < 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int is_pagetable_dying_supported(void)
|
|
|
|
{
|
|
|
|
struct xen_hvm_pagetable_dying a;
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
a.domid = DOMID_SELF;
|
|
|
|
a.gpa = 0x00;
|
|
|
|
rc = HYPERVISOR_hvm_op(HVMOP_pagetable_dying, &a);
|
|
|
|
if (rc < 0) {
|
|
|
|
printk(KERN_DEBUG "HVMOP_pagetable_dying not supported\n");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
void __init xen_hvm_init_mmu_ops(void)
|
|
|
|
{
|
|
|
|
if (is_pagetable_dying_supported())
|
|
|
|
pv_mmu_ops.exit_mmap = xen_hvm_exit_mmap;
|
|
|
|
}
|
2010-07-29 15:37:48 +02:00
|
|
|
#endif
|
2010-06-17 15:22:52 +02:00
|
|
|
|
2009-05-21 11:09:46 +02:00
|
|
|
#define REMAP_BATCH_SIZE 16
|
|
|
|
|
|
|
|
struct remap_data {
|
|
|
|
unsigned long mfn;
|
|
|
|
pgprot_t prot;
|
|
|
|
struct mmu_update *mmu_update;
|
|
|
|
};
|
|
|
|
|
|
|
|
static int remap_area_mfn_pte_fn(pte_t *ptep, pgtable_t token,
|
|
|
|
unsigned long addr, void *data)
|
|
|
|
{
|
|
|
|
struct remap_data *rmd = data;
|
|
|
|
pte_t pte = pte_mkspecial(pfn_pte(rmd->mfn++, rmd->prot));
|
|
|
|
|
|
|
|
rmd->mmu_update->ptr = arbitrary_virt_to_machine(ptep).maddr;
|
|
|
|
rmd->mmu_update->val = pte_val_ma(pte);
|
|
|
|
rmd->mmu_update++;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int xen_remap_domain_mfn_range(struct vm_area_struct *vma,
|
|
|
|
unsigned long addr,
|
|
|
|
unsigned long mfn, int nr,
|
|
|
|
pgprot_t prot, unsigned domid)
|
|
|
|
{
|
|
|
|
struct remap_data rmd;
|
|
|
|
struct mmu_update mmu_update[REMAP_BATCH_SIZE];
|
|
|
|
int batch;
|
|
|
|
unsigned long range;
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
prot = __pgprot(pgprot_val(prot) | _PAGE_IOMAP);
|
|
|
|
|
|
|
|
vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;
|
|
|
|
|
|
|
|
rmd.mfn = mfn;
|
|
|
|
rmd.prot = prot;
|
|
|
|
|
|
|
|
while (nr) {
|
|
|
|
batch = min(REMAP_BATCH_SIZE, nr);
|
|
|
|
range = (unsigned long)batch << PAGE_SHIFT;
|
|
|
|
|
|
|
|
rmd.mmu_update = mmu_update;
|
|
|
|
err = apply_to_page_range(vma->vm_mm, addr, range,
|
|
|
|
remap_area_mfn_pte_fn, &rmd);
|
|
|
|
if (err)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
err = -EFAULT;
|
|
|
|
if (HYPERVISOR_mmu_update(mmu_update, batch, NULL, domid) < 0)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
nr -= batch;
|
|
|
|
addr += range;
|
|
|
|
}
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
out:
|
|
|
|
|
|
|
|
flush_tlb_all();
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(xen_remap_domain_mfn_range);
|
|
|
|
|
2008-08-21 02:02:19 +02:00
|
|
|
#ifdef CONFIG_XEN_DEBUG_FS
|
|
|
|
|
|
|
|
static struct dentry *d_mmu_debug;
|
|
|
|
|
|
|
|
static int __init xen_mmu_debugfs(void)
|
|
|
|
{
|
|
|
|
struct dentry *d_xen = xen_init_debugfs();
|
|
|
|
|
|
|
|
if (d_xen == NULL)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
d_mmu_debug = debugfs_create_dir("mmu", d_xen);
|
|
|
|
|
|
|
|
debugfs_create_u8("zero_stats", 0644, d_mmu_debug, &zero_stats);
|
|
|
|
|
|
|
|
debugfs_create_u32("pgd_update", 0444, d_mmu_debug, &mmu_stats.pgd_update);
|
|
|
|
debugfs_create_u32("pgd_update_pinned", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pgd_update_pinned);
|
|
|
|
debugfs_create_u32("pgd_update_batched", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pgd_update_pinned);
|
|
|
|
|
|
|
|
debugfs_create_u32("pud_update", 0444, d_mmu_debug, &mmu_stats.pud_update);
|
|
|
|
debugfs_create_u32("pud_update_pinned", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pud_update_pinned);
|
|
|
|
debugfs_create_u32("pud_update_batched", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pud_update_pinned);
|
|
|
|
|
|
|
|
debugfs_create_u32("pmd_update", 0444, d_mmu_debug, &mmu_stats.pmd_update);
|
|
|
|
debugfs_create_u32("pmd_update_pinned", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pmd_update_pinned);
|
|
|
|
debugfs_create_u32("pmd_update_batched", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pmd_update_pinned);
|
|
|
|
|
|
|
|
debugfs_create_u32("pte_update", 0444, d_mmu_debug, &mmu_stats.pte_update);
|
|
|
|
// debugfs_create_u32("pte_update_pinned", 0444, d_mmu_debug,
|
|
|
|
// &mmu_stats.pte_update_pinned);
|
|
|
|
debugfs_create_u32("pte_update_batched", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.pte_update_pinned);
|
|
|
|
|
|
|
|
debugfs_create_u32("mmu_update", 0444, d_mmu_debug, &mmu_stats.mmu_update);
|
|
|
|
debugfs_create_u32("mmu_update_extended", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.mmu_update_extended);
|
|
|
|
xen_debugfs_create_u32_array("mmu_update_histo", 0444, d_mmu_debug,
|
|
|
|
mmu_stats.mmu_update_histo, 20);
|
|
|
|
|
|
|
|
debugfs_create_u32("set_pte_at", 0444, d_mmu_debug, &mmu_stats.set_pte_at);
|
|
|
|
debugfs_create_u32("set_pte_at_batched", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.set_pte_at_batched);
|
|
|
|
debugfs_create_u32("set_pte_at_current", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.set_pte_at_current);
|
|
|
|
debugfs_create_u32("set_pte_at_kernel", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.set_pte_at_kernel);
|
|
|
|
|
|
|
|
debugfs_create_u32("prot_commit", 0444, d_mmu_debug, &mmu_stats.prot_commit);
|
|
|
|
debugfs_create_u32("prot_commit_batched", 0444, d_mmu_debug,
|
|
|
|
&mmu_stats.prot_commit_batched);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
fs_initcall(xen_mmu_debugfs);
|
|
|
|
|
|
|
|
#endif /* CONFIG_XEN_DEBUG_FS */
|