bpf, lsm: Fix the file_mprotect LSM test.
Subsystems (3)
PATCH 1/1 bpf, lsm: Fix the file_mprotect LSM test.
From: KP Singh <[email protected]>
The test was previously using an mprotect on the heap memory allocated using malloc and was expecting the allocation to be always using sbrk(2). This is, however, not always true and in certain conditions malloc may end up using anonymous mmaps for heap alloctions. This means that the following condition that is used in the "lsm/file_mprotect" program is not sufficent to detect all mprotect calls done on heap memory:
is_heap = (vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk);
The test is updated to use an mprotect on memory allocated on the stack. While this would result in the splitting of the vma, this happens only after the security_file_mprotect hook. So, the condition used in the BPF program holds true.
--- .../selftests/bpf/prog_tests/test_lsm.c | 18 +++++++++--------- tools/testing/selftests/bpf/progs/lsm.c | 8 ++++---- 2 files changed, 13 insertions(+), 13 deletions(-)
tools/testing/selftests/bpf/prog_tests/test_lsm.c +9 -9
@@ -15,7 +15,10 @@char *CMD_ARGS [] = {"true" ,NULL }; -int heap_mprotect (void ) +#define GET_PAGE_ADDR (ADDR ,PAGE_SIZE ) \ + (char *)(((unsigned long) (ADDR + PAGE_SIZE)) & ~(PAGE_SIZE-1)) + +int stack_mprotect (void ) {void *buf;long sz;
@@ -25,12 +28,9 @@ int heap_mprotect(void)if (sz <0 )return sz; - buf =memalign (sz,2 * sz); -if (buf ==NULL ) -return -ENOMEM ; - - ret =mprotect (buf, sz,PROT_READ |PROT_WRITE |PROT_EXEC ); -free (buf); + buf =alloca (sz *3 ); + ret =mprotect (GET_PAGE_ADDR (buf, sz), sz, +PROT_READ |PROT_WRITE |PROT_EXEC );return ret; }
@@ -73,8 +73,8 @@ void test_test_lsm(void) skel->bss->monitored_pid =getpid (); - err =heap_mprotect (); -if (CHECK (errno !=EPERM ,"heap_mprotect" ,"want errno=EPERM, got %d\n" , + err =stack_mprotect (); +if (CHECK (errno !=EPERM ,"stack_mprotect" ,"want err=EPERM, got %d\n" , errno)) goto close_prog;
tools/testing/selftests/bpf/progs/lsm.c +4 -4
@@ -23,12 +23,12 @@ int BPF_PROG(test_int_hook, struct vm_area_struct *vma,return ret;__u32 pid =bpf_get_current_pid_tgid () >>32 ; -int is_heap =0 ; +int is_stack =0 ; - is_heap = (vma->vm_start >= vma->vm_mm->start_brk && - vma->vm_end <= vma->vm_mm->brk); + is_stack = (vma->vm_start <= vma->vm_mm->start_stack && + vma->vm_end >= vma->vm_mm->start_stack); -if (is_heap && monitored_pid == pid) { +if (is_stack && monitored_pid == pid) { mprotect_count++; ret = -EPERM ; }
Signature · 1 line
2.20.1
Alexei Starovoitov
AppliedQuoted 14 lines · KP Singh
From: KP Singh <[email protected]>
The test was previously using an mprotect on the heap memory allocated using malloc and was expecting the allocation to be always using sbrk(2). This is, however, not always true and in certain conditions malloc may end up using anonymous mmaps for heap alloctions. This means that the following condition that is used in the "lsm/file_mprotect" program is not sufficent to detect all mprotect calls done on heap memory:
is_heap = (vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk);
Quoted 8 lines
The test is updated to use an mprotect on memory allocated on the stack. While this would result in the splitting of the vma, this happens only after the
security_file_mprotecthook. So, the condition used in the BPF program holds true.Signed-off-by KP Singh <[email protected]>Reported-by Alexei Starovoitov <[email protected]>Fixes 03e54f100d57 ("bpf: lsm: Add selftests for BPF_PROG_TYPE_LSM")
Applied. Thanks