seccomp: Move copy_seccomp() to no failure path.

Our syzbot instance reported memory leaks in do_seccomp() [0], similar
to the report [1].  It shows that we miss freeing struct seccomp_filter
and some objects included in it.

We can reproduce the issue with the program below [2] which calls one
seccomp() and two clone() syscalls.

The first clone()d child exits earlier than its parent and sends a
signal to kill it during the second clone(), more precisely before the
fatal_signal_pending() test in copy_process().  When the parent receives
the signal, it has to destroy the embryonic process and return -EINTR to
user space.  In the failure path, we have to call seccomp_filter_release()
to decrement the filter's refcount.

Initially, we called it in free_task() called from the failure path, but
the commit 3a15fb6 ("seccomp: release filter after task is fully
dead") moved it to release_task() to notify user space as early as possible
that the filter is no longer used.

To keep the change and current seccomp refcount semantics, let's move
copy_seccomp() just after the signal check and add a WARN_ON_ONCE() in
free_task() for future debugging.

[0]:
unreferenced object 0xffff8880063add00 (size 256):
  comm "repro_seccomp", pid 230, jiffies 4294687090 (age 9.914s)
  hex dump (first 32 bytes):
    01 00 00 00 01 00 00 00 00 00 00 00 00 00 00 00  ................
    ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff  ................
  backtrace:
    do_seccomp (./include/linux/slab.h:600 ./include/linux/slab.h:733 kernel/seccomp.c:666 kernel/seccomp.c:708 kernel/seccomp.c:1871 kernel/seccomp.c:1991)
    do_syscall_64 (arch/x86/entry/common.c:50 arch/x86/entry/common.c:80)
    entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120)
unreferenced object 0xffffc90000035000 (size 4096):
  comm "repro_seccomp", pid 230, jiffies 4294687090 (age 9.915s)
  hex dump (first 32 bytes):
    01 00 00 00 00 00 00 00 00 00 00 00 05 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    __vmalloc_node_range (mm/vmalloc.c:3226)
    __vmalloc_node (mm/vmalloc.c:3261 (discriminator 4))
    bpf_prog_alloc_no_stats (kernel/bpf/core.c:91)
    bpf_prog_alloc (kernel/bpf/core.c:129)
    bpf_prog_create_from_user (net/core/filter.c:1414)
    do_seccomp (kernel/seccomp.c:671 kernel/seccomp.c:708 kernel/seccomp.c:1871 kernel/seccomp.c:1991)
    do_syscall_64 (arch/x86/entry/common.c:50 arch/x86/entry/common.c:80)
    entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120)
unreferenced object 0xffff888003fa1000 (size 1024):
  comm "repro_seccomp", pid 230, jiffies 4294687090 (age 9.915s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    bpf_prog_alloc_no_stats (./include/linux/slab.h:600 ./include/linux/slab.h:733 kernel/bpf/core.c:95)
    bpf_prog_alloc (kernel/bpf/core.c:129)
    bpf_prog_create_from_user (net/core/filter.c:1414)
    do_seccomp (kernel/seccomp.c:671 kernel/seccomp.c:708 kernel/seccomp.c:1871 kernel/seccomp.c:1991)
    do_syscall_64 (arch/x86/entry/common.c:50 arch/x86/entry/common.c:80)
    entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120)
unreferenced object 0xffff888006360240 (size 16):
  comm "repro_seccomp", pid 230, jiffies 4294687090 (age 9.915s)
  hex dump (first 16 bytes):
    01 00 37 00 76 65 72 6c e0 83 01 06 80 88 ff ff  ..7.verl........
  backtrace:
    bpf_prog_store_orig_filter (net/core/filter.c:1137)
    bpf_prog_create_from_user (net/core/filter.c:1428)
    do_seccomp (kernel/seccomp.c:671 kernel/seccomp.c:708 kernel/seccomp.c:1871 kernel/seccomp.c:1991)
    do_syscall_64 (arch/x86/entry/common.c:50 arch/x86/entry/common.c:80)
    entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120)
unreferenced object 0xffff8880060183e0 (size 8):
  comm "repro_seccomp", pid 230, jiffies 4294687090 (age 9.915s)
  hex dump (first 8 bytes):
    06 00 00 00 00 00 ff 7f                          ........
  backtrace:
    kmemdup (mm/util.c:129)
    bpf_prog_store_orig_filter (net/core/filter.c:1144)
    bpf_prog_create_from_user (net/core/filter.c:1428)
    do_seccomp (kernel/seccomp.c:671 kernel/seccomp.c:708 kernel/seccomp.c:1871 kernel/seccomp.c:1991)
    do_syscall_64 (arch/x86/entry/common.c:50 arch/x86/entry/common.c:80)
    entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120)

[1]: https://syzkaller.appspot.com/bug?id=2809bb0ac77ad9aa3f4afe42d6a610aba594a987

[2]:
#define _GNU_SOURCE
#include <sched.h>
#include <signal.h>
#include <unistd.h>
#include <sys/syscall.h>
#include <linux/filter.h>
#include <linux/seccomp.h>

void main(void)
{
	struct sock_filter filter[] = {
		BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_ALLOW),
	};
	struct sock_fprog fprog = {
		.len = sizeof(filter) / sizeof(filter[0]),
		.filter = filter,
	};
	long i, pid;

	syscall(__NR_seccomp, SECCOMP_SET_MODE_FILTER, 0, &fprog);

	for (i = 0; i < 2; i++) {
		pid = syscall(__NR_clone, CLONE_NEWNET | SIGKILL, NULL, NULL, 0);
		if (pid == 0)
			return;
	}
}

Fixes: 3a15fb6 ("seccomp: release filter after task is fully dead")
Reported-by: [email protected]
Reported-by: Ayushman Dutta <[email protected]>
Suggested-by: Kees Cook <[email protected]>
Signed-off-by: Kuniyuki Iwashima <[email protected]>
Reviewed-by: Christian Brauner (Microsoft) <[email protected]>
Signed-off-by: Kees Cook <[email protected]>
Link: https://lore.kernel.org/r/[email protected]

Author: q2ven

kernel/fork.c

@@ -535,6 +535,9 @@ void put_task_stack(struct task_struct *tsk)
 
 void free_task(struct task_struct *tsk)
 {
+#ifdef CONFIG_SECCOMP
+	WARN_ON_ONCE(tsk->seccomp.filter);
+#endif
 	release_user_cpus_ptr(tsk);
 	scs_release(tsk);
 
@@ -2406,12 +2409,6 @@ static __latent_entropy struct task_struct *copy_process(
 
 	spin_lock(&current->sighand->siglock);
 
-	/*
-	 * Copy seccomp details explicitly here, in case they were changed
-	 * before holding sighand lock.
-	 */
-	copy_seccomp(p);
-
 	rv_task_fork(p);
 
 	rseq_fork(p, clone_flags);
@@ -2428,6 +2425,14 @@ static __latent_entropy struct task_struct *copy_process(
 		goto bad_fork_cancel_cgroup;
 	}
 
+	/* No more failure paths after this point. */
+
+	/*
+	 * Copy seccomp details explicitly here, in case they were changed
+	 * before holding sighand lock.
+	 */
+	copy_seccomp(p);
+
 	init_task_pid_links(p);
 	if (likely(p->pid)) {
 		ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace);