[S390] vdso: dont map at mmap_base

The vdso object is currently always mapped with mm->mmap_base used as
requested address. In case of flexible mmap layout this means it gets
mapped above mmap_base and therefore potentially stealing a bit of
address space that is reserved for the stack.
In case of flexible mmap layout the object should be mapped below
mmap base. For legacy mmap layout above.
To fix this just don't request any specific address and let the mmap
code figure out an address that fits.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

1 files changed, 1 insertions, 3 deletions

diff --git a/arch/s390/kernel/vdso.c b/arch/s390/kernel/vdso.c
index e3150dd2fe74..f438d74dedbd 100644
--- a/arch/s390/kernel/vdso.c
+++ b/arch/s390/kernel/vdso.c
@@ -203,7 +203,6 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
         if (!uses_interp)
                 return 0;
 
-        vdso_base = mm->mmap_base;
 #ifdef CONFIG_64BIT
         vdso_pagelist = vdso64_pagelist;
         vdso_pages = vdso64_pages;
@@ -233,8 +232,7 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
          * fail and end up putting it elsewhere.
          */
         down_write(&mm->mmap_sem);
-        vdso_base = get_unmapped_area(NULL, vdso_base,
-                                      vdso_pages << PAGE_SHIFT, 0, 0);
+        vdso_base = get_unmapped_area(NULL, 0, vdso_pages << PAGE_SHIFT, 0, 0);
         if (IS_ERR_VALUE(vdso_base)) {
                 rc = vdso_base;
                 goto out_up;