Re: [PATCHv3 3/6] x86/arch_prctl/vdso: add ARCH_MAP_VDSO_*
From: Dmitry Safonov
Date: Wed Aug 31 2016 - 13:22:44 EST
On 08/31/2016 06:42 PM, Andy Lutomirski wrote:
On Wed, Aug 31, 2016 at 8:30 AM, Dmitry Safonov <dsafonov@xxxxxxxxxxxxx> wrote:
On 08/31/2016 06:00 PM, Andy Lutomirski wrote:
On Fri, Aug 26, 2016 at 10:13 AM, Dmitry Safonov <dsafonov@xxxxxxxxxxxxx>
wrote:
Add API to change vdso blob type with arch_prctl.
As this is usefull only by needs of CRIU, expose
this interface under CONFIG_CHECKPOINT_RESTORE.
Cc: Andy Lutomirski <luto@xxxxxxxxxx>
Cc: Oleg Nesterov <oleg@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: "H. Peter Anvin" <hpa@xxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: linux-mm@xxxxxxxxx
Cc: x86@xxxxxxxxxx
Cc: Cyrill Gorcunov <gorcunov@xxxxxxxxxx>
Cc: Pavel Emelyanov <xemul@xxxxxxxxxxxxx>
Signed-off-by: Dmitry Safonov <dsafonov@xxxxxxxxxxxxx>
---
arch/x86/entry/vdso/vma.c | 45
++++++++++++++++++++++++++++++---------
arch/x86/include/asm/vdso.h | 2 ++
arch/x86/include/uapi/asm/prctl.h | 6 ++++++
arch/x86/kernel/process_64.c | 25 ++++++++++++++++++++++
4 files changed, 68 insertions(+), 10 deletions(-)
diff --git a/arch/x86/entry/vdso/vma.c b/arch/x86/entry/vdso/vma.c
index 5bcb25a9e573..dad2b2d8ff03 100644
--- a/arch/x86/entry/vdso/vma.c
+++ b/arch/x86/entry/vdso/vma.c
@@ -176,6 +176,16 @@ static int vvar_fault(const struct
vm_special_mapping *sm,
return VM_FAULT_SIGBUS;
}
+static const struct vm_special_mapping vdso_mapping = {
+ .name = "[vdso]",
+ .fault = vdso_fault,
+ .mremap = vdso_mremap,
+};
+static const struct vm_special_mapping vvar_mapping = {
+ .name = "[vvar]",
+ .fault = vvar_fault,
+};
+
/*
* Add vdso and vvar mappings to current process.
* @image - blob to map
@@ -188,16 +198,6 @@ static int map_vdso(const struct vdso_image *image,
unsigned long addr)
unsigned long text_start;
int ret = 0;
- static const struct vm_special_mapping vdso_mapping = {
- .name = "[vdso]",
- .fault = vdso_fault,
- .mremap = vdso_mremap,
- };
- static const struct vm_special_mapping vvar_mapping = {
- .name = "[vvar]",
- .fault = vvar_fault,
- };
-
if (down_write_killable(&mm->mmap_sem))
return -EINTR;
@@ -256,6 +256,31 @@ static int map_vdso_randomized(const struct
vdso_image *image)
return map_vdso(image, addr);
}
+int map_vdso_once(const struct vdso_image *image, unsigned long addr)
+{
+ struct mm_struct *mm = current->mm;
+ struct vm_area_struct *vma;
+
+ down_write(&mm->mmap_sem);
+ /*
+ * Check if we have already mapped vdso blob - fail to prevent
+ * abusing from userspace install_speciall_mapping, which may
+ * not do accounting and rlimit right.
+ * We could search vma near context.vdso, but it's a slowpath,
+ * so let's explicitely check all VMAs to be completely sure.
+ */
+ for (vma = mm->mmap; vma; vma = vma->vm_next) {
+ if (vma->vm_private_data == &vdso_mapping ||
+ vma->vm_private_data == &vvar_mapping) {
Should probably also check that vm_ops == &special_mapping_vmops,
which means that maybe there should be a:
static inline bool vma_is_special_mapping(const struct vm_area_struct
*vma, const struct vm_special_mapping &sm);
Oh, I remember why I didn't do it also (except header changes):
uprobes uses &special_mapping_vmops in inserted XOL area.
That's fine. I think you should check *both* vm_ops and
vm_private_data to avoid (unlikely) confusion with some other VMA.
Oh, I thought you suggest to check either of them.
*Both* makes sence, since AFAIU, vm_private_data may be uninitialized on
other VMAs, will do.
--
Dmitry