Skip to contentSkip to conversation

arm64: Fix handling of tracer updates to x0 on syscall entry

PATCH Applied Will Deacon 1 participant 4 messages Jul 30 – Aug 2

Subsystems (2)
lore
Share & tools
lore

Will Deacon

author maintainer
#0 lore raw

v4 cover arm64: Fix handling of tracer updates to x0 on syscall entry

Hi all,

This is version 4 of the patch previously posted here:

v1: https://lore.kernel.org/r/[email protected]
v2: https://lore.kernel.org/r/[email protected]
v3: https://lore.kernel.org/r/[email protected]

Changes since v3 include:

- Add an additional patch to pass orig_x0 as the first syscall argument for native tasks (Jinjie)
- Move WARN_ON_ONCE() after checking regs->syscallno (Sashiko)
- Add Tested-by from Jinjie.

Cheers,

Will

Cc Kees Cook <[email protected]>
Cc Jinjie Ruan <[email protected]>
Cc Mark Rutland <[email protected]>

--->8

Will Deacon (2):
arm64: ptrace: Keep 'orig_x0' in-sync with x0 on syscall entry
arm64: syscall: Pass 'orig_x0' as first argument to native system call

 arch/arm64/include/asm/syscall_wrapper.h | 13 ++++--
 arch/arm64/kernel/ptrace.c               | 50 ++++++++++++++++++++++++
 2 files changed, 59 insertions(+), 4 deletions(-)
Signature · 1 line
2.55.0.508.g3f0d502094-goog
author maintainer
#1 lore raw

PATCH v4 1/2 arm64: ptrace: Keep 'orig_x0' in-sync with x0 on syscall entry

Commit e057b9477232 ("arm64: syscall: Ensure saved x0 is kept in-sync with tracer updates") attempted to resolve a long-standing issue with syscall entry tracing, where a tracer is able to manipulate the first syscall argument without being subjected to seccomp or audit checking.

Unfortunately, that fix was incomplete [1], as it failed to update 'orig_x0' between a tracer updating x0 during a seccomp ptrace exit (SECCOMP_RET_TRACE) and the seccomp filter being re-evaluated.

Rather than add hooks to the core seccomp code, instead move the synchronisation code into the ptrace GPR and syscall setting code so that 'orig_x0' is kept up to date with x0 whenever we're stopped on the syscall entry path.

Cc Kees Cook <[email protected]>
Cc Jinjie Ruan <[email protected]>
Cc Mark Rutland <[email protected]>
Fixes e057b9477232 ("arm64: syscall: Ensure saved x0 is kept in-sync with tracer updates")
Fixes a5cd110cb836 ("arm64/ptrace: run seccomp after ptrace")
---
 arch/arm64/kernel/ptrace.c | 50 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
arch/arm64/kernel/ptrace.c +50 -0
index 4d08598e2891..7a97865acb79 100644
@@ -560,6 +560,42 @@ static int gpr_get(struct task_struct *target,
 	return membuf_write(&to, uregs, sizeof(*uregs));
 }
 
+static void update_syscall_orig_x0_after_ptrace(struct task_struct *target)
+{
+	struct pt_regs *regs = task_pt_regs(target);
+	struct kernel_siginfo *info = target->last_siginfo;
+
+	/*
+	 * Skip the update for NO_SYSCALL (set either by the user or the
+	 * tracer), as regs[0] holds the return value (see the comment in
+	 * el0_svc_common()) and can be unwound using syscall_rollback().
+	 */
+	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.
+	 */
+	if (is_compat_thread(task_thread_info(target)))
+		return;
+
+	/*
+	 * 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 &&
+	    target->ptrace_message == PTRACE_EVENTMSG_SYSCALL_EXIT) {
+		return;
+	}
+
+	regs->orig_x0 = regs->regs[0];
+}
+
 static int gpr_set(struct task_struct *target, const struct user_regset *regset,
 		   unsigned int pos, unsigned int count,
 		   const void *kbuf, const void __user *ubuf)
@@ -575,6 +611,14 @@ static int gpr_set(struct task_struct *target, const struct user_regset *regset,
 		return -EINVAL;
 
 	task_pt_regs(target)->user_regs = newregs;
+
+	/*
+	 * Keep orig_x0 authoritative so that seccomp (via
+	 * syscall_get_arguments()), audit and the restart path all see the same
+	 * first argument the syscall is dispatched with, even if it has been
+	 * updated by a tracer.
+	 */
+	update_syscall_orig_x0_after_ptrace(target);
 	return 0;
 }
 
@@ -753,6 +797,12 @@ static int system_call_set(struct task_struct *target,
 		return ret;
 
 	task_pt_regs(target)->syscallno = syscallno;
+
+	/*
+	 * Re-sync orig_x0 in case the syscall number has been changed
+	 * from NO_SYSCALL.
+	 */
+	update_syscall_orig_x0_after_ptrace(target);
 	return ret;
 }
 
Signature · 1 line
2.55.0.508.g3f0d502094-goog
author maintainer
#2 lore raw

PATCH v4 2/2 arm64: syscall: Pass 'orig_x0' as first argument to native system call

In reply to Will Deacon

syscall_get_arguments() returns 'regs->orig_x0' for the first system call argument so as to avoid aliasing with the syscall return value in 'regs->regs[0]' on the return path, however the actual syscall invocation passes 'regs->regs[0]' as the first parameter.

Although the two registers should be kept in sync during syscall entry for native tasks, pass 'regs->orig_x0' as the first syscall parameter for consistency with the syscall argument APIs. Compat tasks continue to use 'regs->regs[0]' for compatibility with the behaviour of the 32-bit kernel.

---
 arch/arm64/include/asm/syscall_wrapper.h | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
arch/arm64/include/asm/syscall_wrapper.h +9 -4
index abb57bc54305..395152ef5372 100644
@@ -10,13 +10,13 @@
 
 #include <asm/ptrace.h>
 
-#define SC_ARM64_REGS_TO_ARGS(x, ...)				\
+#ifdef CONFIG_COMPAT
+
+#define COMPAT_SC_ARM64_REGS_TO_ARGS(x, ...)			\
 	__MAP(x,__SC_ARGS					\
 	      ,,regs->regs[0],,regs->regs[1],,regs->regs[2]	\
 	      ,,regs->regs[3],,regs->regs[4],,regs->regs[5])
 
-#ifdef CONFIG_COMPAT
-
 #define COMPAT_SYSCALL_DEFINEx(x, name, ...)						\
 	asmlinkage long __arm64_compat_sys##name(const struct pt_regs *regs);		\
 	ALLOW_ERROR_INJECTION(__arm64_compat_sys##name, ERRNO);				\
@@ -24,7 +24,7 @@
 	static inline long __do_compat_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__));	\
 	asmlinkage long __arm64_compat_sys##name(const struct pt_regs *regs)		\
 	{										\
-		return __se_compat_sys##name(SC_ARM64_REGS_TO_ARGS(x,__VA_ARGS__));	\
+		return __se_compat_sys##name(COMPAT_SC_ARM64_REGS_TO_ARGS(x,__VA_ARGS__));	\
 	}										\
 	static long __se_compat_sys##name(__MAP(x,__SC_LONG,__VA_ARGS__))		\
 	{										\
@@ -46,6 +46,11 @@
 
 #endif /* CONFIG_COMPAT */
 
+#define SC_ARM64_REGS_TO_ARGS(x, ...)				\
+	__MAP(x,__SC_ARGS					\
+	      ,,regs->orig_x0,,regs->regs[1],,regs->regs[2]	\
+	      ,,regs->regs[3],,regs->regs[4],,regs->regs[5])
+
 #define __SYSCALL_DEFINEx(x, name, ...)						\
 	asmlinkage long __arm64_sys##name(const struct pt_regs *regs);		\
 	ALLOW_ERROR_INJECTION(__arm64_sys##name, ERRNO);			\
Signature · 1 line
2.55.0.508.g3f0d502094-goog
author maintainer Applied
#3 lore raw
In reply to Will Deacon
Quoted 9 lines · Will Deacon

This is version 4 of the patch previously posted here:

v1: https://lore.kernel.org/r/[email protected]
v2: https://lore.kernel.org/r/[email protected]
v3: https://lore.kernel.org/r/[email protected]

Changes since v3 include:

[...]

I decided against applying these for v7.2 this late in the cycle, so I've queued them up for v7.3 for some more exposure.

Applied to arm64 (for-next/ptrace), thanks!

[1/2] arm64: ptrace: Keep 'orig_x0' in-sync with x0 on syscall entry

      https://git.kernel.org/arm64/c/88b839ce497c
[2/2] arm64: syscall: Pass 'orig_x0' as first argument to native system call
      https://git.kernel.org/arm64/c/5a87e8c7f370

Cheers,

Signature · 5 lines
Will

https://fixes.arm64.dev
https://next.arm64.dev
https://will.arm64.dev