Skip to content

Commit

Permalink
bpf/selftests: Add bpf_get_task_stack retval bounds test_prog
Browse files Browse the repository at this point in the history
Add a libbpf test prog which feeds bpf_get_task_stack's return value
into seq_write after confirming it's positive. No attempt to bound the
value from above is made.

Load will fail if verifier does not refine retval range based on buf sz
input to bpf_get_task_stack.

Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20210416204704.2816874-4-davemarchevsky@fb.com
  • Loading branch information
Dave Marchevsky authored and Alexei Starovoitov committed Apr 20, 2021
1 parent bdc4e36 commit c77cec5
Show file tree
Hide file tree
Showing 2 changed files with 28 additions and 0 deletions.
1 change: 1 addition & 0 deletions tools/testing/selftests/bpf/prog_tests/bpf_iter.c
Original file line number Diff line number Diff line change
Expand Up @@ -147,6 +147,7 @@ static void test_task_stack(void)
return;

do_dummy_read(skel->progs.dump_task_stack);
do_dummy_read(skel->progs.get_task_user_stacks);

bpf_iter_task_stack__destroy(skel);
}
Expand Down
27 changes: 27 additions & 0 deletions tools/testing/selftests/bpf/progs/bpf_iter_task_stack.c
Original file line number Diff line number Diff line change
Expand Up @@ -35,3 +35,30 @@ int dump_task_stack(struct bpf_iter__task *ctx)

return 0;
}

SEC("iter/task")
int get_task_user_stacks(struct bpf_iter__task *ctx)
{
struct seq_file *seq = ctx->meta->seq;
struct task_struct *task = ctx->task;
uint64_t buf_sz = 0;
int64_t res;

if (task == (void *)0)
return 0;

res = bpf_get_task_stack(task, entries,
MAX_STACK_TRACE_DEPTH * SIZE_OF_ULONG, BPF_F_USER_STACK);
if (res <= 0)
return 0;

buf_sz += res;

/* If the verifier doesn't refine bpf_get_task_stack res, and instead
* assumes res is entirely unknown, this program will fail to load as
* the verifier will believe that max buf_sz value allows reading
* past the end of entries in bpf_seq_write call
*/
bpf_seq_write(seq, &entries, buf_sz);
return 0;
}

0 comments on commit c77cec5

Please sign in to comment.