mm: Optional full ASLR for mmap(), mremap(), vdso, stack and heap
Subsystems (12)
RFC PATCH v6 1/1 mm: Optional full ASLR for mmap(), mremap(), vdso, stack and heap
Writing a new value of 3 to /proc/sys/kernel/randomize_va_space enables full randomization of memory mappings. With 2, the base of the VMA used for such mappings is random, but the mappings are created in predictable places within the VMA and in sequential order. With 3, new VMAs are created to fully randomize the mappings.
Mappings created with mmap(NULL, ...) are randomized and mremap(..., MREMAP_MAYMOVE) will move the mappings even if not necessary. The locations of heap (memory allocated with brk()), stack and vdso are also randomized.
On 32 bit systems this may cause problems due to increased VM fragmentation if the address space gets crowded.
On all systems, it will reduce performance and increase memory and cache usage due to less efficient use of page tables and inability to merge adjacent VMAs with compatible attributes. On x86_64 with 5 level page tables, in the worst case, additional page table entries of up to 4 pages are created for each mapping, so with small mappings there's considerable penalty.
By lowering the lowest address for mapping the main executable from 2/3 of the address space to sysctl.vm.mmap_min_addr, it's possible to use the full 35 bits available on x86_64 for ASLR.
The method is to randomize the new address without considering VMAs. If the address fails checks because of overlap with the stack area (or in case of mremap(), overlap with the old mapping), the operation is retried a few times before falling back to old method.
In this example with sysctl.kernel.randomize_va_space = 2, main executable, heap allocated with brk(), locale-archive, libc, dynamic loader, some anonymous memory reserved with mmap(), stack and vdso are located in three groups and inside each group the mappings are close to each other:
$ cat /proc/self/maps (only first line for each object shown for brevity) 55d61f2ac000-55d61f2ae000 r--p 00000000 fe:0c 1868624 /usr/bin/cat 7f9124f40000-7f91254a2000 r--p 00000000 fe:0c 2474005 /usr/lib/locale/locale-archive 7f91254a2000-7f91255a2000 rw-p 00000000 00:00 0 7f91255a2000-7f91255c7000 r--p 00000000 fe:0c 2402332 /usr/lib/x86_64-linux-gnu/libc-2.31.so 7f9125763000-7f9125769000 rw-p 00000000 00:00 0 7f9125795000-7f9125796000 r--p 00000000 fe:0c 2400754 /usr/lib/x86_64-linux-gnu/ld-2.31.so 7f91257c1000-7f91257c2000 rw-p 00000000 00:00 0 7ffdf983d000-7ffdf985e000 rw-p 00000000 00:00 0 [stack] 7ffdf9897000-7ffdf989b000 r--p 00000000 00:00 0 [vvar] 7ffdf989b000-7ffdf989d000 r-xp 00000000 00:00 0 [vdso]
With sysctl.kernel.randomize_va_space = 3, they are located at unrelated addresses and the order is random:
$ echo 3 > /proc/sys/kernel/randomize_va_space $ cat /proc/self/maps (only first line for each object shown for brevity) bc5ed961000-bc5eda61000 rw-p 00000000 00:00 0 2968e14a4000-2968e14c5000 rw-p 00000000 00:00 0 [stack] 30f80fb63000-30f80fb65000 r--p 00000000 fe:0c 1868624 /usr/bin/cat 381de5bfa000-381de5bfe000 r--p 00000000 00:00 0 [vvar] 381de5bfe000-381de5c00000 r-xp 00000000 00:00 0 [vdso] 42cd1060d000-42cd10632000 r--p 00000000 fe:0c 2402332 /usr/lib/x86_64-linux-gnu/libc-2.31.so 42cd107ce000-42cd107d2000 rw-p 00000000 00:00 0 547f9c21b000-547f9c21c000 r--p 00000000 fe:0c 2400754 /usr/lib/x86_64-linux-gnu/ld-2.31.so 547f9c247000-547f9c248000 rw-p 00000000 00:00 0 743548368000-7435488ca000 r--p 00000000 fe:0c 2474005 /usr/lib/locale/locale-archive 7dd3a185f000-7dd3a1861000 rw-p 00000000 00:00 0
---
v2: also randomize mremap(..., MREMAP_MAYMOVE)
v3: avoid stack area and retry in case of bad random address (Jann Horn), improve description in kernel.rst (Matthew Wilcox) v4:
- use /proc/$pid/maps in the example (Mike Rapaport)
- CCs (Andrew Morton)
- only check randomize_va_space == 3
v5: randomize also vdso and stack
v6:
- randomize also heap
- use 35 bits for ASLR on x86_64
- RFC due to temporarily disabling mremap() randomization
--- Documentation/admin-guide/hw-vuln/spectre.rst | 6 ++--- Documentation/admin-guide/sysctl/kernel.rst | 22 +++++++++++++++++ arch/x86/Kconfig | 2 +- arch/x86/entry/vdso/vma.c | 7 +++++- arch/x86/kernel/process.c | 5 +++- arch/x86/mm/mmap.c | 3 +++ fs/binfmt_elf.c | 7 +++++- init/Kconfig | 2 +- mm/mmap.c | 24 ++++++++++++++++++- mm/mremap.c | 10 ++++++++ mm/util.c | 14 ++++++++++- 11 files changed, 92 insertions(+), 10 deletions(-)
Documentation/admin-guide/hw-vuln/spectre.rst +3 -3
@@ -254,7 +254,7 @@ Spectre variant 2 left by the previous process will also be cleared. User programs should use address space randomization to make attacks - more difficult (Set /proc/sys/kernel/randomize_va_space = 1 or 2). + more difficult (Set /proc/sys/kernel/randomize_va_space = 1, 2 or 3). 3. A virtualized guest attacking the host ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
@@ -499,8 +499,8 @@ Spectre variant 2 more overhead and run slower. User programs should use address space randomization - (/proc/sys/kernel/randomize_va_space = 1 or 2) to make attacks more - difficult. + (/proc/sys/kernel/randomize_va_space = 1, 2 or 3) to make attacks + more difficult. 3. VM mitigation ^^^^^^^^^^^^^^^^
Documentation/admin-guide/sysctl/kernel.rst +22 -0
@@ -1060,6 +1060,28 @@ that support this feature. Systems with ancient and/or broken binaries should be configured with ``CONFIG_COMPAT_BRK`` enabled, which excludes the heap from process address space randomization. + +3 Additionally enable full randomization of memory mappings. With 2, + the base of the VMA used for such mappings may be random, but the + mappings are created in predictable places within the VMA and in + sequential order. With 3, new VMAs are created to fully randomize + the mappings. + + Mappings created with mmap(NULL, ...) are randomized and + mremap(..., MREMAP_MAYMOVE) will move the mappings even if not + necessary. The locations of heap (memory allocated with brk()), + stack and vdso are also randomized. + + On 32 bit systems this may cause problems due to increased VM + fragmentation if the address space gets crowded. + + On all systems, it will reduce performance and increase memory and + cache usage due to less efficient use of page tables and inability + to merge adjacent VMAs with compatible attributes. On x86_64 with + 5 level page tables, in the worst case, additional page table + entries of up to 4 pages are created for each mapping, so with + small mappings there's considerable penalty. + == ===========================================================================
arch/x86/Kconfig +1 -1
@@ -266,7 +266,7 @@ config ARCH_MMAP_RND_BITS_MIN default 8 config ARCH_MMAP_RND_BITS_MAX - default 32 if 64BIT + default 35 if 64BIT default 16 config ARCH_MMAP_RND_COMPAT_BITS_MIN
arch/x86/entry/vdso/vma.c +6 -1
@@ -361,7 +361,12 @@ static unsigned long vdso_addr(unsigned long start, unsigned len)static int map_vdso_randomized (const struct vdso_image *image) { -unsigned long addr =vdso_addr (current->mm->start_stack, image->size-image->sym_vvar_start); +unsigned long addr; + +if (randomize_va_space ==3 ) + addr =0 ;/* let get_unmapped_area() pick the address */ +else + addr =vdso_addr (current->mm->start_stack, image->size-image->sym_vvar_start);return map_vdso (image, addr); }
arch/x86/kernel/process.c +4 -1
@@ -905,7 +905,10 @@ unsigned long arch_align_stack(unsigned long sp)unsigned long arch_randomize_brk (struct mm_struct *mm) { -return randomize_page (mm->brk,0x02000000 ); +if (randomize_va_space ==3 ) +return arch_mmap_rnd (); +else +return randomize_page (mm->brk,0x02000000 ); }/*
arch/x86/mm/mmap.c +3 -0
@@ -21,6 +21,7 @@#include <linux/elf-randomize.h> #include <asm/elf.h> #include <asm/io.h> +#include <linux/security.h> #include "physaddr.h"
@@ -122,6 +123,8 @@ static void arch_pick_mmap_base(unsigned long *base, unsigned long *legacy_base, *legacy_base =mmap_legacy_base (random_factor, task_size);if (mmap_is_legacy ()) *base = *legacy_base; +else if (randomize_va_space ==3 ) + *base = mmap_min_addr;else *base =mmap_base (random_factor, task_size, rlim_stack); }
fs/binfmt_elf.c +6 -1
@@ -1108,7 +1108,12 @@ static int load_elf_binary(struct linux_binprm *bprm)* without MAP_FIXED). */ if (interpreter) { - load_bias =ELF_ET_DYN_BASE ; +if (randomize_va_space ==3 ) +/* more space for randomization */ + load_bias = mmap_min_addr; +else + load_bias =ELF_ET_DYN_BASE ; +if (current->flags &PF_RANDOMIZE ) load_bias +=arch_mmap_rnd (); alignment =maximum_alignment (elf_phdata, elf_ex->e_phnum);
init/Kconfig +1 -1
@@ -1873,7 +1873,7 @@ config COMPAT_BRK also breaks ancient binaries (including anything libc5 based). This option changes the bootup default to heap randomization disabled, and can be overridden at runtime by setting - /proc/sys/kernel/randomize_va_space to 2. + /proc/sys/kernel/randomize_va_space to 2 or 3. On non-ancient distros (post-2000 ones) N is usually a safe choice.
mm/mmap.c +23 -1
@@ -47,6 +47,7 @@#include <linux/pkeys.h> #include <linux/oom.h> #include <linux/sched/mm.h> +#include <linux/elf-randomize.h> #include <linux/uaccess.h> #include <asm/cacheflush.h>
@@ -73,6 +74,8 @@ const int mmap_rnd_compat_bits_max = CONFIG_ARCH_MMAP_RND_COMPAT_BITS_MAX;int mmap_rnd_compat_bits__read_mostly =CONFIG_ARCH_MMAP_RND_COMPAT_BITS ;#endif +#define MAX_RANDOM_MMAP_RETRIES 5 +static bool ignore_rlimit_data;core_param (ignore_rlimit_data, ignore_rlimit_data, bool,0644 );
@@ -206,7 +209,7 @@ SYSCALL_DEFINE1(brk, unsigned long, brk)#ifdef CONFIG_COMPAT_BRK /* * CONFIG_COMPAT_BRK can still be overridden by setting -* randomize_va_space to 2, which will still cause mm->start_brk +* randomize_va_space to >= 2, which will still cause mm->start_brk * to be arbitrarily shifted */ if (current->brk_randomized)
@@ -2281,10 +2284,29 @@ get_unmapped_area(struct file *file, unsigned long addr, unsigned long len, get_area = shmem_get_unmapped_area; } +/* Pick a random address even outside current VMAs? */ +if (!addr && randomize_va_space ==3 ) { +int i; +unsigned long new_addr; + +/* Try a few times to find a free area */ +for (i =0 ; i <MAX_RANDOM_MMAP_RETRIES ; i++) { + new_addr =arch_mmap_rnd (); + + new_addr =get_area (file, new_addr, len, pgoff, flags |MAP_FIXED ); +if (!IS_ERR_VALUE (new_addr)) { + addr = new_addr; + goto found; + } + } +/* failed, retry with original addr */ + } + addr =get_area (file, addr, len, pgoff, flags);if (IS_ERR_VALUE (addr))return addr; + found:if (addr >TASK_SIZE - len)return -ENOMEM ;if (offset_in_page (addr))
mm/mremap.c +10 -0
@@ -738,6 +738,10 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len, }else if (retval ==1 ) downgraded = true; ret = addr; + +//if (randomize_va_space == 3) +// goto maybe_also_move; + goto out; }
@@ -770,6 +774,10 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len, new_addr = addr; } ret = addr; + +//if (randomize_va_space == 3) +// goto maybe_also_move; + goto out; } }
@@ -779,6 +787,8 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len,* we need to create a new one and move it.. */ ret = -ENOMEM ; + +//maybe_also_move: if (flags &MREMAP_MAYMOVE ) {unsigned long map_flags =0 ;if (vma->vm_flags &VM_MAYSHARE )
mm/util.c +13 -1
@@ -319,6 +319,12 @@ unsigned long randomize_stack_top(unsigned long stack_top) {unsigned long random_variable =0 ; +/* +* Randomize stack address. +*/ +if (randomize_va_space ==3 ) +return arch_mmap_rnd (); +if (current->flags &PF_RANDOMIZE ) { random_variable =get_random_long (); random_variable &=STACK_RND_MASK ;
@@ -334,6 +340,9 @@ unsigned long randomize_stack_top(unsigned long stack_top)#ifdef CONFIG_ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT unsigned long arch_randomize_brk (struct mm_struct *mm) { +if (randomize_va_space ==3 ) +return arch_mmap_rnd (); +/* Is the current task 32bit ? */ if (!IS_ENABLED (CONFIG_64BIT ) ||is_compat_task ())return randomize_page (mm->brk,SZ_32M );
@@ -412,7 +421,10 @@ void arch_pick_mmap_layout(struct mm_struct *mm, struct rlimit *rlim_stack)#elif defined(CONFIG_MMU ) && !defined(HAVE_ARCH_PICK_MMAP_LAYOUT )void arch_pick_mmap_layout (struct mm_struct *mm,struct rlimit *rlim_stack) { - mm->mmap_base =TASK_UNMAPPED_BASE ; +if (randomize_va_space ==3 ) + mm->mmap_base = mmap_min_addr; +else + mm->mmap_base =TASK_UNMAPPED_BASE ; mm->get_unmapped_area = arch_get_unmapped_area; }#endif
base-commit: 7f376f1917d7461e05b648983e8d2aea9d0712b2
Signature · 1 line
2.29.2