Merge branch 'Fix accessing syscall arguments'

Ilya Leoshkevich says:

====================

libbpf now has macros to access syscall arguments in an
architecture-agnostic manner, but unfortunately they have a number of
issues on non-Intel arches, which this series aims to fix.

v1: https://lore.kernel.org/bpf/20220201234200.1836443-1-iii@linux.ibm.com/
v1 -> v2:
* Put orig_gpr2 in place of args[1] on s390 (Vasily).
* Fix arm64, powerpc and riscv (Heiko).

v2: https://lore.kernel.org/bpf/20220204041955.1958263-1-iii@linux.ibm.com/
v2 -> v3:
* Undo args[1] change (Andrii).
* Rename PT_REGS_SYSCALL to PT_REGS_SYSCALL_REGS (Andrii).
* Split the riscv patch (Andrii).

v3: https://lore.kernel.org/bpf/20220204145018.1983773-1-iii@linux.ibm.com/
v3 -> v4:
* Undo arm64's and s390's user_pt_regs changes.
* Use struct pt_regs when vmlinux.h is available (Andrii).
* Use offsetofend for accessing orig_gpr2 and orig_x0 (Andrii).
* Move libbpf's copy of offsetofend to a new header.
* Fix riscv's __PT_FP_REG.
* Use PT_REGS_SYSCALL_REGS in test_probe_user.c.
* Test bpf_syscall_macro with userspace headers.
* Use Naveen's suggestions and code in patches 5 and 6.
* Add warnings to arm64's and s390's ptrace.h (Andrii).

v4: https://lore.kernel.org/bpf/20220208051635.2160304-1-iii@linux.ibm.com/
v4 -> v5:
* Go back to v3.
* Do not touch arch headers.
* Use CO-RE struct flavors to access orig_x0 and orig_gpr2.
* Fail compilation if non-CO-RE macros are used to access the first
  syscall parameter on arm64 and s390.
* Fix accessing frame pointer on riscv.
====================
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>

tools/lib/bpf/bpf_tracing.h

@@ -70,13 +70,15 @@
 #define __PT_PARM2_REG si
 #define __PT_PARM3_REG dx
 #define __PT_PARM4_REG cx
-#define __PT_PARM4_REG_SYSCALL r10 /* syscall uses r10 */
 #define __PT_PARM5_REG r8
 #define __PT_RET_REG sp
 #define __PT_FP_REG bp
 #define __PT_RC_REG ax
 #define __PT_SP_REG sp
 #define __PT_IP_REG ip
+/* syscall uses r10 for PARM4 */
+#define PT_REGS_PARM4_SYSCALL(x) ((x)->r10)
+#define PT_REGS_PARM4_CORE_SYSCALL(x) BPF_CORE_READ(x, r10)
 
 #else
 
@@ -100,13 +102,15 @@
 #define __PT_PARM2_REG rsi
 #define __PT_PARM3_REG rdx
 #define __PT_PARM4_REG rcx
-#define __PT_PARM4_REG_SYSCALL r10 /* syscall uses r10 */
 #define __PT_PARM5_REG r8
 #define __PT_RET_REG rsp
 #define __PT_FP_REG rbp
 #define __PT_RC_REG rax
 #define __PT_SP_REG rsp
 #define __PT_IP_REG rip
+/* syscall uses r10 for PARM4 */
+#define PT_REGS_PARM4_SYSCALL(x) ((x)->r10)
+#define PT_REGS_PARM4_CORE_SYSCALL(x) BPF_CORE_READ(x, r10)
 
 #endif /* __i386__ */
 
@@ -114,6 +118,10 @@
 
 #elif defined(bpf_target_s390)
 
+struct pt_regs___s390 {
+	unsigned long orig_gpr2;
+};
+
 /* s390 provides user_pt_regs instead of struct pt_regs to userspace */
 #define __PT_REGS_CAST(x) ((const user_pt_regs *)(x))
 #define __PT_PARM1_REG gprs[2]
@@ -126,6 +134,8 @@
 #define __PT_RC_REG gprs[2]
 #define __PT_SP_REG gprs[15]
 #define __PT_IP_REG psw.addr
+#define PT_REGS_PARM1_SYSCALL(x) ({ _Pragma("GCC error \"use PT_REGS_PARM1_CORE_SYSCALL() instead\""); 0l; })
+#define PT_REGS_PARM1_CORE_SYSCALL(x) BPF_CORE_READ((const struct pt_regs___s390 *)(x), orig_gpr2)
 
 #elif defined(bpf_target_arm)
 
@@ -142,6 +152,10 @@
 
 #elif defined(bpf_target_arm64)
 
+struct pt_regs___arm64 {
+	unsigned long orig_x0;
+};
+
 /* arm64 provides struct user_pt_regs instead of struct pt_regs to userspace */
 #define __PT_REGS_CAST(x) ((const struct user_pt_regs *)(x))
 #define __PT_PARM1_REG regs[0]
@@ -154,6 +168,8 @@
 #define __PT_RC_REG regs[0]
 #define __PT_SP_REG sp
 #define __PT_IP_REG pc
+#define PT_REGS_PARM1_SYSCALL(x) ({ _Pragma("GCC error \"use PT_REGS_PARM1_CORE_SYSCALL() instead\""); 0l; })
+#define PT_REGS_PARM1_CORE_SYSCALL(x) BPF_CORE_READ((const struct pt_regs___arm64 *)(x), orig_x0)
 
 #elif defined(bpf_target_mips)
 
@@ -180,6 +196,8 @@
 #define __PT_RC_REG gpr[3]
 #define __PT_SP_REG sp
 #define __PT_IP_REG nip
+/* powerpc does not select ARCH_HAS_SYSCALL_WRAPPER. */
+#define PT_REGS_SYSCALL_REGS(ctx) ctx
 
 #elif defined(bpf_target_sparc)
 
@@ -208,10 +226,12 @@
 #define __PT_PARM4_REG a3
 #define __PT_PARM5_REG a4
 #define __PT_RET_REG ra
-#define __PT_FP_REG fp
+#define __PT_FP_REG s0
 #define __PT_RC_REG a5
 #define __PT_SP_REG sp
-#define __PT_IP_REG epc
+#define __PT_IP_REG pc
+/* riscv does not select ARCH_HAS_SYSCALL_WRAPPER. */
+#define PT_REGS_SYSCALL_REGS(ctx) ctx
 
 #endif
 
@@ -265,22 +285,22 @@ struct pt_regs;
 
 #endif
 
+#ifndef PT_REGS_PARM1_SYSCALL
 #define PT_REGS_PARM1_SYSCALL(x) PT_REGS_PARM1(x)
+#endif
 #define PT_REGS_PARM2_SYSCALL(x) PT_REGS_PARM2(x)
 #define PT_REGS_PARM3_SYSCALL(x) PT_REGS_PARM3(x)
-#ifdef __PT_PARM4_REG_SYSCALL
-#define PT_REGS_PARM4_SYSCALL(x) (__PT_REGS_CAST(x)->__PT_PARM4_REG_SYSCALL)
-#else /* __PT_PARM4_REG_SYSCALL */
+#ifndef PT_REGS_PARM4_SYSCALL
 #define PT_REGS_PARM4_SYSCALL(x) PT_REGS_PARM4(x)
 #endif
 #define PT_REGS_PARM5_SYSCALL(x) PT_REGS_PARM5(x)
 
+#ifndef PT_REGS_PARM1_CORE_SYSCALL
 #define PT_REGS_PARM1_CORE_SYSCALL(x) PT_REGS_PARM1_CORE(x)
+#endif
 #define PT_REGS_PARM2_CORE_SYSCALL(x) PT_REGS_PARM2_CORE(x)
 #define PT_REGS_PARM3_CORE_SYSCALL(x) PT_REGS_PARM3_CORE(x)
-#ifdef __PT_PARM4_REG_SYSCALL
-#define PT_REGS_PARM4_CORE_SYSCALL(x) BPF_CORE_READ(__PT_REGS_CAST(x), __PT_PARM4_REG_SYSCALL)
-#else /* __PT_PARM4_REG_SYSCALL */
+#ifndef PT_REGS_PARM4_CORE_SYSCALL
 #define PT_REGS_PARM4_CORE_SYSCALL(x) PT_REGS_PARM4_CORE(x)
 #endif
 #define PT_REGS_PARM5_CORE_SYSCALL(x) PT_REGS_PARM5_CORE(x)
@@ -326,6 +346,16 @@ struct pt_regs;
 
 #endif /* defined(bpf_target_defined) */
 
+/*
+ * When invoked from a syscall handler kprobe, returns a pointer to a
+ * struct pt_regs containing syscall arguments and suitable for passing to
+ * PT_REGS_PARMn_SYSCALL() and PT_REGS_PARMn_CORE_SYSCALL().
+ */
+#ifndef PT_REGS_SYSCALL_REGS
+/* By default, assume that the arch selects ARCH_HAS_SYSCALL_WRAPPER. */
+#define PT_REGS_SYSCALL_REGS(ctx) ((struct pt_regs *)PT_REGS_PARM1(ctx))
+#endif
+
 #ifndef ___bpf_concat
 #define ___bpf_concat(a, b) a ## b
 #endif

tools/testing/selftests/bpf/prog_tests/test_bpf_syscall_macro.c

@@ -33,7 +33,11 @@ void test_bpf_syscall_macro(void)
 
 	/* check whether args of syscall are copied correctly */
 	prctl(exp_arg1, exp_arg2, exp_arg3, exp_arg4, exp_arg5);
+#if defined(__aarch64__) || defined(__s390__)
+	ASSERT_NEQ(skel->bss->arg1, exp_arg1, "syscall_arg1");
+#else
 	ASSERT_EQ(skel->bss->arg1, exp_arg1, "syscall_arg1");
+#endif
 	ASSERT_EQ(skel->bss->arg2, exp_arg2, "syscall_arg2");
 	ASSERT_EQ(skel->bss->arg3, exp_arg3, "syscall_arg3");
 	/* it cannot copy arg4 when uses PT_REGS_PARM4 on x86_64 */

tools/testing/selftests/bpf/progs/bpf_syscall_macro.c

@@ -28,14 +28,19 @@ int BPF_KPROBE(handle_sys_prctl)
 {
 	struct pt_regs *real_regs;
 	pid_t pid = bpf_get_current_pid_tgid() >> 32;
+	unsigned long tmp = 0;
 
 	if (pid != filter_pid)
 		return 0;
 
-	real_regs = (struct pt_regs *)PT_REGS_PARM1(ctx);
+	real_regs = PT_REGS_SYSCALL_REGS(ctx);
 
 	/* test for PT_REGS_PARM */
-	bpf_probe_read_kernel(&arg1, sizeof(arg1), &PT_REGS_PARM1_SYSCALL(real_regs));
+
+#if !defined(bpf_target_arm64) && !defined(bpf_target_s390)
+	bpf_probe_read_kernel(&tmp, sizeof(tmp), &PT_REGS_PARM1_SYSCALL(real_regs));
+#endif
+	arg1 = tmp;
 	bpf_probe_read_kernel(&arg2, sizeof(arg2), &PT_REGS_PARM2_SYSCALL(real_regs));
 	bpf_probe_read_kernel(&arg3, sizeof(arg3), &PT_REGS_PARM3_SYSCALL(real_regs));
 	bpf_probe_read_kernel(&arg4_cx, sizeof(arg4_cx), &PT_REGS_PARM4(real_regs));