[PATCH v2] selftests/x86: Add a selftest for SYSRET to noncanonical addresses

From: Andy Lutomirski
Date: Mon Dec 26 2016 - 22:21:21 EST


SYSRET to a noncanonical address will blow up on Intel CPUs. Linux
needs to prevent this from happening in two major cases, and the
criteria will become more complicated when support for larger virtual
address spaces is added.

A fast-path SYSCALL will fallthrough to the following instruction
using SYSRET without any particular checking. To prevent fallthrough
to a noncanonical address, Linux prevents the highest canonical page
from being mapped. This test case checks a variety of possible maximum
addresses to make sure that either we can't map code there or that
SYSCALL fallthrough works.

A slow-path system call can return anywhere. Linux needs to make sure
that, if the return address is non-canonical, it won't use SYSRET.
This test cases causes sigreturn() to return to a variety of addresses
(with RCX == RIP) and makes sure that nothing explodes.

Some of this code comes from Kirill Shutemov.

Changes from v1:
- Get rid of some extra parentheses.
- Test more corner cases.
- Get rid of some duplicate '0x' printout.

Cc: "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>
Signed-off-by: Andy Lutomirski <luto@xxxxxxxxxx>
---
tools/testing/selftests/x86/Makefile | 2 +-
tools/testing/selftests/x86/sysret_rip.c | 195 +++++++++++++++++++++++++++++++
2 files changed, 196 insertions(+), 1 deletion(-)
create mode 100644 tools/testing/selftests/x86/sysret_rip.c

diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
index 8c1cb423cfe6..25d4067c11e4 100644
--- a/tools/testing/selftests/x86/Makefile
+++ b/tools/testing/selftests/x86/Makefile
@@ -10,7 +10,7 @@ TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt ptrace_sysc
TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault test_syscall_vdso unwind_vdso \
test_FCMOV test_FCOMI test_FISTTP \
vdso_restorer
-TARGETS_C_64BIT_ONLY := fsgsbase
+TARGETS_C_64BIT_ONLY := fsgsbase sysret_rip

TARGETS_C_32BIT_ALL := $(TARGETS_C_BOTHBITS) $(TARGETS_C_32BIT_ONLY)
TARGETS_C_64BIT_ALL := $(TARGETS_C_BOTHBITS) $(TARGETS_C_64BIT_ONLY)
diff --git a/tools/testing/selftests/x86/sysret_rip.c b/tools/testing/selftests/x86/sysret_rip.c
new file mode 100644
index 000000000000..d85ec5b3671c
--- /dev/null
+++ b/tools/testing/selftests/x86/sysret_rip.c
@@ -0,0 +1,195 @@
+/*
+ * sigreturn.c - tests that x86 avoids Intel SYSRET pitfalls
+ * Copyright (c) 2014-2016 Andrew Lutomirski
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ */
+
+#define _GNU_SOURCE
+
+#include <stdlib.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <inttypes.h>
+#include <sys/signal.h>
+#include <sys/ucontext.h>
+#include <sys/syscall.h>
+#include <err.h>
+#include <stddef.h>
+#include <stdbool.h>
+#include <setjmp.h>
+#include <sys/user.h>
+#include <sys/mman.h>
+#include <assert.h>
+
+
+asm (
+ ".pushsection \".text\", \"ax\"\n\t"
+ ".balign 4096\n\t"
+ "test_page: .globl test_page\n\t"
+ ".fill 4094,1,0xcc\n\t"
+ "test_syscall_insn:\n\t"
+ "syscall\n\t"
+ ".ifne . - test_page - 4096\n\t"
+ ".error \"test page is not one page long\"\n\t"
+ ".endif\n\t"
+ ".popsection"
+ );
+
+extern const char test_page[];
+static void const *current_test_page_addr = test_page;
+
+static void sethandler(int sig, void (*handler)(int, siginfo_t *, void *),
+ int flags)
+{
+ struct sigaction sa;
+ memset(&sa, 0, sizeof(sa));
+ sa.sa_sigaction = handler;
+ sa.sa_flags = SA_SIGINFO | flags;
+ sigemptyset(&sa.sa_mask);
+ if (sigaction(sig, &sa, 0))
+ err(1, "sigaction");
+}
+
+static void clearhandler(int sig)
+{
+ struct sigaction sa;
+ memset(&sa, 0, sizeof(sa));
+ sa.sa_handler = SIG_DFL;
+ sigemptyset(&sa.sa_mask);
+ if (sigaction(sig, &sa, 0))
+ err(1, "sigaction");
+}
+
+/* State used by our signal handlers. */
+static gregset_t initial_regs;
+
+static volatile unsigned long rip;
+
+static void sigsegv_for_sigreturn_test(int sig, siginfo_t *info, void *ctx_void)
+{
+ ucontext_t *ctx = (ucontext_t*)ctx_void;
+
+ if (rip != ctx->uc_mcontext.gregs[REG_RIP]) {
+ printf("[FAIL]\tRequested RIP=0x%lx but got RIP=0x%lx\n",
+ rip, (unsigned long)ctx->uc_mcontext.gregs[REG_RIP]);
+ fflush(stdout);
+ _exit(1);
+ }
+
+ memcpy(&ctx->uc_mcontext.gregs, &initial_regs, sizeof(gregset_t));
+
+ printf("[OK]\tGot SIGSEGV at RIP=0x%lx\n", rip);
+}
+
+static void sigusr1(int sig, siginfo_t *info, void *ctx_void)
+{
+ ucontext_t *ctx = (ucontext_t*)ctx_void;
+
+ memcpy(&initial_regs, &ctx->uc_mcontext.gregs, sizeof(gregset_t));
+
+ /* Set IP and CX to match so that SYSRET can happen. */
+ ctx->uc_mcontext.gregs[REG_RIP] = rip;
+ ctx->uc_mcontext.gregs[REG_RCX] = rip;
+
+ /* R11 and EFLAGS should already match. */
+ assert(ctx->uc_mcontext.gregs[REG_EFL] ==
+ ctx->uc_mcontext.gregs[REG_R11]);
+
+ sethandler(SIGSEGV, sigsegv_for_sigreturn_test, SA_RESETHAND);
+
+ return;
+}
+
+static void test_sigreturn_to(unsigned long ip)
+{
+ rip = ip;
+ printf("[RUN]\tsigreturn to 0x%lx\n", ip);
+ raise(SIGUSR1);
+}
+
+static jmp_buf jmpbuf;
+
+static void sigsegv_for_fallthrough(int sig, siginfo_t *info, void *ctx_void)
+{
+ ucontext_t *ctx = (ucontext_t*)ctx_void;
+
+ if (rip != ctx->uc_mcontext.gregs[REG_RIP]) {
+ printf("[FAIL]\tExpected SIGSEGV at 0x%lx but got RIP=0x%lx\n",
+ rip, (unsigned long)ctx->uc_mcontext.gregs[REG_RIP]);
+ fflush(stdout);
+ _exit(1);
+ }
+
+ siglongjmp(jmpbuf, 1);
+}
+
+static void test_syscall_fallthrough_to(unsigned long ip)
+{
+ void *new_address = (void *)(ip - 4096);
+ void *ret;
+
+ printf("[RUN]\tTrying a SYSCALL that falls through to 0x%lx\n", ip);
+
+ ret = mremap((void *)current_test_page_addr, 4096, 4096,
+ MREMAP_MAYMOVE | MREMAP_FIXED, new_address);
+ if (ret == MAP_FAILED) {
+ if (ip <= (1UL << 47) - PAGE_SIZE) {
+ err(1, "mremap to %p", new_address);
+ } else {
+ printf("[OK]\tmremap to %p failed\n", new_address);
+ return;
+ }
+ }
+
+ if (ret != new_address)
+ errx(1, "mremap malfunctioned: asked for %p but got %p\n",
+ new_address, ret);
+
+ current_test_page_addr = new_address;
+ rip = ip;
+
+ if (sigsetjmp(jmpbuf, 1) == 0) {
+ asm volatile ("call *%[syscall_insn]" :: "a" (SYS_getpid),
+ [syscall_insn] "rm" (ip - 2));
+ errx(1, "[FAIL]\tSyscall trampoline returned");
+ }
+
+ printf("[OK]\tWe survived\n");
+}
+
+int main()
+{
+ /*
+ * When the kernel returns from a slow-path syscall, it will
+ * detect whether SYSRET is appropriate. If it incorrectly
+ * thinks that SYSRET is appropriate when RIP is noncanonical,
+ * it'll crash on Intel CPUs.
+ */
+ sethandler(SIGUSR1, sigusr1, 0);
+ for (int i = 47; i < 64; i++)
+ test_sigreturn_to(1UL<<i);
+
+ clearhandler(SIGUSR1);
+
+ sethandler(SIGSEGV, sigsegv_for_fallthrough, 0);
+
+ /* One extra test to check that we didn't screw up the mremap logic. */
+ test_syscall_fallthrough_to((1UL << 47) - 2*PAGE_SIZE);
+
+ /* These are the interesting cases. */
+ for (int i = 47; i < 64; i++) {
+ test_syscall_fallthrough_to((1UL<<i) - PAGE_SIZE);
+ test_syscall_fallthrough_to(1UL<<i);
+ }
+
+ return 0;
+}
--
2.9.3