[PATCH v2] arm64: ptrace: don't warn on missing siginfo in a group-stop

[PATCH v2] arm64: ptrace: don't warn on missing siginfo in a group-stop

[Zhao Gongyi]

PTRACE_ATTACH to a group-stopped child can enter do_jobctl_trap(),
which calls ptrace_stop() without siginfo. If the tracer writes
NT_ARM_SYSTEM_CALL, system_call_set() can change syscallno from
NO_SYSCALL before calling update_syscall_orig_x0_after_ptrace(). The
WARN_ON_ONCE(!info) then fires, although this is a valid ptrace stop.

Remove the warning and guard the syscall-exit-stop check against a
missing siginfo. Keep recognizing both SIGTRAP and SIGTRAP | 0x80:
syscall-exit-stops use either value depending on whether the tracer
enabled PTRACE_O_TRACESYSGOOD. With no siginfo, allow the final
assignment to synchronize orig_x0 from x0.

Fixes: 88b839ce497c ("arm64: ptrace: Keep 'orig_x0' in-sync with x0 on syscall entry")
Cc: stable@vger.kernel.org
Signed-off-by: Zhao Gongyi <zhaogongyi@bytedance.com>
---
Changes since v1:
- Remove the warning and early return, as suggested by Oleg.
- Keep recognizing syscall-exit-stops with or without PTRACE_O_TRACESYSGOOD.

v1: https://lore.kernel.org/r/20261016024536.18938-1-zhaogongyi@bytedance.com

Tested on arm64 KVM at af32da41b032 with this v2 applied.
At a group-stop, PTRACE_GETSIGINFO returned -EINVAL and an
NT_ARM_SYSTEM_CALL write succeeded. Syscall-entry and seccomp
x0 rewrites worked. At syscall-exit-stops, NT_PRSTATUS writes succeeded
with and without PTRACE_O_TRACESYSGOOD (si_code 133 and 5,
respectively). No kernel warnings occurred.

 arch/arm64/kernel/ptrace.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index f743cbec1..22fb9eeed 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -573,10 +573,6 @@ static void update_syscall_orig_x0_after_ptrace(struct task_struct *target)
 	if (regs->syscallno == NO_SYSCALL)
 		return;
 
-	/* We should only be called when target is in a ptrace stop */
-	if (WARN_ON_ONCE(!info))
-		return;
-
 	/*
 	 * For compat tasks, orig_r0 is provided directly through GPR index
 	 * 17.
@@ -588,7 +584,7 @@ static void update_syscall_orig_x0_after_ptrace(struct task_struct *target)
 	 * Don't update orig_x0 for a syscall-exit-stop, as x0 now contains the
 	 * return value of the system call.
 	 */
-	if ((info->si_code & ~0x80) == SIGTRAP &&
+	if (info && (info->si_code & ~0x80) == SIGTRAP &&
 	    target->ptrace_message == PTRACE_EVENTMSG_SYSCALL_EXIT) {
 		return;
 	}

base-commit: 24a19b953e2a3d0ea5a8c4951a0ab8f32720c907
-- 
2.39.5 (Apple Git-154)

Re: [PATCH v2] arm64: ptrace: don't warn on missing siginfo in a group-stop

[Oleg Nesterov]

On 10/23, Zhao Gongyi wrote:
>
> --- a/arch/arm64/kernel/ptrace.c
> +++ b/arch/arm64/kernel/ptrace.c
> @@ -573,10 +573,6 @@ static void update_syscall_orig_x0_after_ptrace(struct task_struct *target)
>  	if (regs->syscallno == NO_SYSCALL)
>  		return;
>
> -	/* We should only be called when target is in a ptrace stop */
> -	if (WARN_ON_ONCE(!info))
> -		return;
> -
>  	/*
>  	 * For compat tasks, orig_r0 is provided directly through GPR index
>  	 * 17.
> @@ -588,7 +584,7 @@ static void update_syscall_orig_x0_after_ptrace(struct task_struct *target)
>  	 * Don't update orig_x0 for a syscall-exit-stop, as x0 now contains the
>  	 * return value of the system call.
>  	 */
> -	if ((info->si_code & ~0x80) == SIGTRAP &&
> +	if (info && (info->si_code & ~0x80) == SIGTRAP &&
>  	    target->ptrace_message == PTRACE_EVENTMSG_SYSCALL_EXIT) {
>  		return;

LGTM, but I don't understand this arm-specific code.

And... Do we need to check target->last_siginfo at all?

ptrace_signal() -> ptrace_stop() path sets ->ptrace_message = 0.

In all other cases: if info != NULL, then .si_code == target->exit_code.
ptrace_event(event) users always use event != 0 and != 0x80.

So it seems that we can just do

	if ((target->exit_code & ~0x80) == SIGTRAP &&
	    target->ptrace_message == PTRACE_EVENTMSG_SYSCALL_EXIT)

?

I'll recheck... If I am right, perhaps it makes sense to simplify
ptrace_get_syscall_info_op().

Anyway, up to you.

Oleg.