[RISCV][ASAN] implementation of internal syscalls wrappers for riscv64

[EccoTheDolphin]

Summary

implements glibc-like wrappers over Linux syscalls.

[3/11] patch series to port ASAN for riscv64

Depends On D87998

[jrtc27]

compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_riscv64.inc

183

"clobbering"?

[EccoTheDolphin]

yep... Thanks, will adjust

[EccoTheDolphin]

fixed typo

[eugenis]

compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_riscv64.inc

85

why not

return a0

[MaskRay]

compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_riscv64.inc

35

All long int can be simplified to long

[EccoTheDolphin]

updated implementation of syscall wrappers to follow style of existing codebase

[vitalybuka]

LGTM

[EccoTheDolphin]

rebasing/addressing comments

[eugenis]

LGTM

compiler-rt/lib/sanitizer_common/CMakeLists.txt

set(SANITIZER_IMPL_HEADERS
   sanitizer_symbolizer_internal.h
   sanitizer_symbolizer_libbacktrace.h
   sanitizer_symbolizer_mac.h
   sanitizer_symbolizer_rtems.h
   sanitizer_syscall_generic.inc
   sanitizer_syscall_linux_aarch64.inc
   sanitizer_syscall_linux_arm.inc
   sanitizer_syscall_linux_x86_64.inc
+  sanitizer_syscall_linux_riscv64.inc
   sanitizer_syscalls_netbsd.inc
   sanitizer_thread_registry.h
   sanitizer_tls_get_addr.h
   sanitizer_vector.h
   sanitizer_win.h
   sanitizer_win_defs.h
   sanitizer_win_dll_thunk.h
   sanitizer_win_weak_interception.h

compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_riscv64.inc

+//===-- sanitizer_syscall_linux_aarch64.inc --------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Implementations of internal_syscall and internal_iserror for Linux/riscv64.
+//
+//===----------------------------------------------------------------------===//
+
+// About local register variables:
+// https://gcc.gnu.org/onlinedocs/gcc/Local-Register-Variables.html#Local-Register-Variables
+//
+// Kernel ABI...
+// To my surprise I haven't found much information regarding it.
+// Kernel source and internet browsing shows that:
+//  syscall number is passed in a7
+//  (http://man7.org/linux/man-pages/man2/syscall.2.html) results are return in
+//  a0 and a1 (http://man7.org/linux/man-pages/man2/syscall.2.html) arguments
+//  are passed in: a0-a7 (see below)
+//
+//  Regarding the arguments. The only "documentation" I could find is
+//  this comment (!!!) by Bruce Hold on google forums (!!!):
+//    https://groups.google.com/a/groups.riscv.org/forum/#!topic/sw-dev/exbrzM3GZDQ
+//    Confirmed by inspecting glibc sources.
+//  Great way to document things.
+#define SYSCALL(name) __NR_##name
+
+#define INTERNAL_SYSCALL_CLOBBERS "memory"
+
+static uptr __internal_syscall(u64 nr) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0");
+  __asm__ volatile("ecall\n\t"
+                   : "=r"(a0)
+                   : "r"(a7)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall0(n) (__internal_syscall)(n)
+
+static uptr __internal_syscall(u64 nr, u64 arg1) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall1(n, a1) (__internal_syscall)(n, (u64)(a1))
+
+static uptr __internal_syscall(u64 nr, u64 arg1, long arg2) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  register u64 a1 asm("a1") = arg2;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7), "r"(a1)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall2(n, a1, a2) \
+  (__internal_syscall)(n, (u64)(a1), (long)(a2))
+
+static uptr __internal_syscall(u64 nr, u64 arg1, long arg2, long arg3) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  register u64 a1 asm("a1") = arg2;
+  register u64 a2 asm("a2") = arg3;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7), "r"(a1), "r"(a2)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall3(n, a1, a2, a3) \
+  (__internal_syscall)(n, (u64)(a1), (long)(a2), (long)(a3))
+
+static uptr __internal_syscall(u64 nr, u64 arg1, long arg2, long arg3,
+                               u64 arg4) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  register u64 a1 asm("a1") = arg2;
+  register u64 a2 asm("a2") = arg3;
+  register u64 a3 asm("a3") = arg4;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7), "r"(a1), "r"(a2), "r"(a3)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall4(n, a1, a2, a3, a4) \
+  (__internal_syscall)(n, (u64)(a1), (long)(a2), (long)(a3), (long)(a4))
+
+static uptr __internal_syscall(u64 nr, u64 arg1, long arg2, long arg3, u64 arg4,
+                               long arg5) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  register u64 a1 asm("a1") = arg2;
+  register u64 a2 asm("a2") = arg3;
+  register u64 a3 asm("a3") = arg4;
+  register u64 a4 asm("a4") = arg5;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7), "r"(a1), "r"(a2), "r"(a3), "r"(a4)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall5(n, a1, a2, a3, a4, a5)                       \
+  (__internal_syscall)(n, (u64)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (u64)(a5))
+
+static uptr __internal_syscall(u64 nr, u64 arg1, long arg2, long arg3, u64 arg4,
+                               long arg5, long arg6) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  register u64 a1 asm("a1") = arg2;
+  register u64 a2 asm("a2") = arg3;
+  register u64 a3 asm("a3") = arg4;
+  register u64 a4 asm("a4") = arg5;
+  register u64 a5 asm("a5") = arg6;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7), "r"(a1), "r"(a2), "r"(a3), "r"(a4), "r"(a5)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall6(n, a1, a2, a3, a4, a5, a6)                   \
+  (__internal_syscall)(n, (u64)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (u64)(a5), (long)(a6))
+
+static uptr __internal_syscall(u64 nr, u64 arg1, long arg2, long arg3, u64 arg4,
+                               long arg5, long arg6, long arg7) {
+  register u64 a7 asm("a7") = nr;
+  register u64 a0 asm("a0") = arg1;
+  register u64 a1 asm("a1") = arg2;
+  register u64 a2 asm("a2") = arg3;
+  register u64 a3 asm("a3") = arg4;
+  register u64 a4 asm("a4") = arg5;
+  register u64 a5 asm("a5") = arg6;
+  register u64 a6 asm("a6") = arg7;
+  __asm__ volatile("ecall\n\t"
+                   : "+r"(a0)
+                   : "r"(a7), "r"(a1), "r"(a2), "r"(a3), "r"(a4), "r"(a5),
+                     "r"(a6)
+                   : INTERNAL_SYSCALL_CLOBBERS);
+  return a0;
+}
+#define __internal_syscall7(n, a1, a2, a3, a4, a5, a6, a7)               \
+  (__internal_syscall)(n, (u64)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (u64)(a5), (long)(a6), (long)(a7))
+
+#define __SYSCALL_NARGS_X(a1, a2, a3, a4, a5, a6, a7, a8, n, ...) n
+#define __SYSCALL_NARGS(...) \
+  __SYSCALL_NARGS_X(__VA_ARGS__, 7, 6, 5, 4, 3, 2, 1, 0, )
+#define __SYSCALL_CONCAT_X(a, b) a##b
+#define __SYSCALL_CONCAT(a, b) __SYSCALL_CONCAT_X(a, b)
+#define __SYSCALL_DISP(b, ...) \
+  __SYSCALL_CONCAT(b, __SYSCALL_NARGS(__VA_ARGS__))(__VA_ARGS__)
+
+#define internal_syscall(...) __SYSCALL_DISP(__internal_syscall, __VA_ARGS__)
+
+// Helper function used to avoid clobbering of errno.
+bool internal_iserror(uptr retval, int *rverrno) {
+  if (retval >= (uptr)-4095) {
+    if (rverrno)
+      *rverrno = -retval;
+    return true;
+  }
+  return false;
+}