1 files changed, 247 insertions, 294 deletions

diff --git a/arch/um/os-Linux/start_up.c b/arch/um/os-Linux/start_up.c
index 337518c5042a..054ac03bbf5e 100644
--- a/arch/um/os-Linux/start_up.c
+++ b/arch/um/os-Linux/start_up.c
@@ -1,6 +1,7 @@
+// SPDX-License-Identifier: GPL-2.0
 /*
+ * Copyright (C) 2021 Benjamin Berg <benjamin@sipsolutions.net>
  * Copyright (C) 2000 - 2007 Jeff Dike (jdike@{addtoit,linux.intel}.com)
- * Licensed under the GPL
  */
 
 #include <stdio.h>
@@ -17,14 +18,24 @@
 #include <sys/wait.h>
 #include <sys/time.h>
 #include <sys/resource.h>
+#include <asm/ldt.h>
 #include <asm/unistd.h>
 #include <init.h>
 #include <os.h>
+#include <smp.h>
+#include <kern_util.h>
 #include <mem_user.h>
 #include <ptrace_user.h>
+#include <stdbool.h>
+#include <stub-data.h>
+#include <sys/prctl.h>
+#include <linux/seccomp.h>
+#include <linux/filter.h>
+#include <sysdep/mcontext.h>
+#include <sysdep/stub.h>
 #include <registers.h>
 #include <skas.h>
-#include <skas_ptrace.h>
+#include "internal.h"
 
 static void ptrace_child(void)
 {
@@ -95,6 +106,8 @@ static int start_ptraced_child(void)
 {
 	int pid, n, status;
 
+	fflush(stdout);
+
 	pid = fork();
 	if (pid == 0)
 		ptrace_child();
@@ -111,140 +124,32 @@ static int start_ptraced_child(void)
 	return pid;
 }
 
-/* When testing for SYSEMU support, if it is one of the broken versions, we
- * must just avoid using sysemu, not panic, but only if SYSEMU features are
- * broken.
- * So only for SYSEMU features we test mustpanic, while normal host features
- * must work anyway!
- */
-static int stop_ptraced_child(int pid, int exitcode, int mustexit)
+static void stop_ptraced_child(int pid, int exitcode)
 {
-	int status, n, ret = 0;
+	int status, n;
+
+	if (ptrace(PTRACE_CONT, pid, 0, 0) < 0)
+		fatal_perror("stop_ptraced_child : ptrace failed");
 
-	if (ptrace(PTRACE_CONT, pid, 0, 0) < 0) {
-		perror("stop_ptraced_child : ptrace failed");
-		return -1;
-	}
 	CATCH_EINTR(n = waitpid(pid, &status, 0));
 	if (!WIFEXITED(status) || (WEXITSTATUS(status) != exitcode)) {
 		int exit_with = WEXITSTATUS(status);
-		if (exit_with == 2)
-			non_fatal("check_ptrace : child exited with status 2. "
-				  "\nDisabling SYSEMU support.\n");
-		non_fatal("check_ptrace : child exited with exitcode %d, while "
-			  "expecting %d; status 0x%x\n", exit_with,
-			  exitcode, status);
-		if (mustexit)
-			exit(1);
-		ret = -1;
+		fatal("stop_ptraced_child : child exited with exitcode %d, "
+		      "while expecting %d; status 0x%x\n", exit_with,
+		      exitcode, status);
 	}
-
-	return ret;
-}
-
-/* Changed only during early boot */
-int ptrace_faultinfo;
-static int disable_ptrace_faultinfo;
-
-int ptrace_ldt;
-static int disable_ptrace_ldt;
-
-int proc_mm;
-static int disable_proc_mm;
-
-int have_switch_mm;
-static int disable_switch_mm;
-
-int skas_needs_stub;
-
-static int __init skas0_cmd_param(char *str, int* add)
-{
-	disable_ptrace_faultinfo = 1;
-	disable_ptrace_ldt = 1;
-	disable_proc_mm = 1;
-	disable_switch_mm = 1;
-
-	return 0;
 }
 
-/* The two __uml_setup would conflict, without this stupid alias. */
-
-static int __init mode_skas0_cmd_param(char *str, int* add)
-	__attribute__((alias("skas0_cmd_param")));
-
-__uml_setup("skas0", skas0_cmd_param,
-"skas0\n"
-"    Disables SKAS3 and SKAS4 usage, so that SKAS0 is used\n\n");
-
-__uml_setup("mode=skas0", mode_skas0_cmd_param,
-"mode=skas0\n"
-"    Disables SKAS3 and SKAS4 usage, so that SKAS0 is used.\n\n");
-
-/* Changed only during early boot */
-static int force_sysemu_disabled = 0;
-
-static int __init nosysemu_cmd_param(char *str, int* add)
-{
-	force_sysemu_disabled = 1;
-	return 0;
-}
-
-__uml_setup("nosysemu", nosysemu_cmd_param,
-"nosysemu\n"
-"    Turns off syscall emulation patch for ptrace (SYSEMU) on.\n"
-"    SYSEMU is a performance-patch introduced by Laurent Vivier. It changes\n"
-"    behaviour of ptrace() and helps reducing host context switch rate.\n"
-"    To make it working, you need a kernel patch for your host, too.\n"
-"    See http://perso.wanadoo.fr/laurent.vivier/UML/ for further \n"
-"    information.\n\n");
-
 static void __init check_sysemu(void)
 {
-	unsigned long regs[MAX_REG_NR];
 	int pid, n, status, count=0;
 
-	non_fatal("Checking syscall emulation patch for ptrace...");
-	sysemu_supported = 0;
+	os_info("Checking syscall emulation for ptrace...");
 	pid = start_ptraced_child();
 
-	if (ptrace(PTRACE_SYSEMU, pid, 0, 0) < 0)
-		goto fail;
-
-	CATCH_EINTR(n = waitpid(pid, &status, WUNTRACED));
-	if (n < 0)
-		fatal_perror("check_sysemu : wait failed");
-	if (!WIFSTOPPED(status) || (WSTOPSIG(status) != SIGTRAP))
-		fatal("check_sysemu : expected SIGTRAP, got status = %d\n",
-		      status);
-
-	if (ptrace(PTRACE_GETREGS, pid, 0, regs) < 0)
-		fatal_perror("check_sysemu : PTRACE_GETREGS failed");
-	if (PT_SYSCALL_NR(regs) != __NR_getpid) {
-		non_fatal("check_sysemu got system call number %d, "
-			  "expected %d...", PT_SYSCALL_NR(regs), __NR_getpid);
-		goto fail;
-	}
-
-	n = ptrace(PTRACE_POKEUSER, pid, PT_SYSCALL_RET_OFFSET, os_getpid());
-	if (n < 0) {
-		non_fatal("check_sysemu : failed to modify system call "
-			  "return");
-		goto fail;
-	}
-
-	if (stop_ptraced_child(pid, 0, 0) < 0)
-		goto fail_stopped;
-
-	sysemu_supported = 1;
-	non_fatal("OK\n");
-	set_using_sysemu(!force_sysemu_disabled);
-
-	non_fatal("Checking advanced syscall emulation patch for ptrace...");
-	pid = start_ptraced_child();
-
-	if ((ptrace(PTRACE_OLDSETOPTIONS, pid, 0,
+	if ((ptrace(PTRACE_SETOPTIONS, pid, 0,
 		   (void *) PTRACE_O_TRACESYSGOOD) < 0))
-		fatal_perror("check_sysemu: PTRACE_OLDSETOPTIONS failed");
+		fatal_perror("check_sysemu: PTRACE_SETOPTIONS failed");
 
 	while (1) {
 		count++;
@@ -277,32 +182,27 @@ static void __init check_sysemu(void)
 			goto fail;
 		}
 	}
-	if (stop_ptraced_child(pid, 0, 0) < 0)
-		goto fail_stopped;
+	stop_ptraced_child(pid, 0);
 
-	sysemu_supported = 2;
-	non_fatal("OK\n");
+	os_info("OK\n");
 
-	if (!force_sysemu_disabled)
-		set_using_sysemu(sysemu_supported);
 	return;
 
 fail:
-	stop_ptraced_child(pid, 1, 0);
-fail_stopped:
-	non_fatal("missing\n");
+	stop_ptraced_child(pid, 1);
+	fatal("missing\n");
 }
 
 static void __init check_ptrace(void)
 {
 	int pid, syscall, n, status;
 
-	non_fatal("Checking that ptrace can change system call numbers...");
+	os_info("Checking that ptrace can change system call numbers...");
 	pid = start_ptraced_child();
 
-	if ((ptrace(PTRACE_OLDSETOPTIONS, pid, 0,
+	if ((ptrace(PTRACE_SETOPTIONS, pid, 0,
 		   (void *) PTRACE_O_TRACESYSGOOD) < 0))
-		fatal_perror("check_ptrace: PTRACE_OLDSETOPTIONS failed");
+		fatal_perror("check_ptrace: PTRACE_SETOPTIONS failed");
 
 	while (1) {
 		if (ptrace(PTRACE_SYSCALL, pid, 0, 0) < 0)
@@ -328,215 +228,268 @@ static void __init check_ptrace(void)
 			break;
 		}
 	}
-	stop_ptraced_child(pid, 0, 1);
-	non_fatal("OK\n");
+	stop_ptraced_child(pid, 0);
+	os_info("OK\n");
 	check_sysemu();
 }
 
-extern void check_tmpexec(void);
+extern unsigned long host_fp_size;
+extern unsigned long exec_regs[MAX_REG_NR];
+extern unsigned long *exec_fp_regs;
 
-static void __init check_coredump_limit(void)
+__initdata static struct stub_data *seccomp_test_stub_data;
+
+static void __init sigsys_handler(int sig, siginfo_t *info, void *p)
 {
-	struct rlimit lim;
-	int err = getrlimit(RLIMIT_CORE, &lim);
+	ucontext_t *uc = p;
 
-	if (err) {
-		perror("Getting core dump limit");
-		return;
-	}
+	/* Stow away the location of the mcontext in the stack */
+	seccomp_test_stub_data->mctx_offset = (unsigned long)&uc->uc_mcontext -
+					      (unsigned long)&seccomp_test_stub_data->sigstack[0];
 
-	printf("Core dump limits :\n\tsoft - ");
-	if (lim.rlim_cur == RLIM_INFINITY)
-		printf("NONE\n");
-	else printf("%lu\n", lim.rlim_cur);
+	/* Prevent libc from clearing memory (mctx_offset in particular) */
+	syscall(__NR_exit, 0);
+}
 
-	printf("\thard - ");
-	if (lim.rlim_max == RLIM_INFINITY)
-		printf("NONE\n");
-	else printf("%lu\n", lim.rlim_max);
+static int __init seccomp_helper(void *data)
+{
+	static struct sock_filter filter[] = {
+		BPF_STMT(BPF_LD | BPF_W | BPF_ABS,
+			 offsetof(struct seccomp_data, nr)),
+		BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, __NR_clock_nanosleep, 1, 0),
+		BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_ALLOW),
+		BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_TRAP),
+	};
+	static struct sock_fprog prog = {
+		.len = ARRAY_SIZE(filter),
+		.filter = filter,
+	};
+	struct sigaction sa;
+
+	/* close_range is needed for the stub */
+	if (stub_syscall3(__NR_close_range, 1, ~0U, 0))
+		exit(1);
+
+	set_sigstack(seccomp_test_stub_data->sigstack,
+			sizeof(seccomp_test_stub_data->sigstack));
+
+	sa.sa_flags = SA_ONSTACK | SA_NODEFER | SA_SIGINFO;
+	sa.sa_sigaction = (void *) sigsys_handler;
+	sa.sa_restorer = NULL;
+	if (sigaction(SIGSYS, &sa, NULL) < 0)
+		exit(2);
+
+	prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0);
+	if (syscall(__NR_seccomp, SECCOMP_SET_MODE_FILTER,
+			SECCOMP_FILTER_FLAG_TSYNC, &prog) != 0)
+		exit(3);
+
+	sleep(0);
+
+	/* Never reached. */
+	_exit(4);
 }
 
-void __init os_early_checks(void)
+static bool __init init_seccomp(void)
 {
 	int pid;
+	int status;
+	int n;
+	unsigned long sp;
 
-	/* Print out the core dump limits early */
-	check_coredump_limit();
+	/*
+	 * We check that we can install a seccomp filter and then exit(0)
+	 * from a trapped syscall.
+	 *
+	 * Note that we cannot verify that no seccomp filter already exists
+	 * for a syscall that results in the process/thread to be killed.
+	 */
 
-	check_ptrace();
+	os_info("Checking that seccomp filters can be installed...");
 
-	/* Need to check this early because mmapping happens before the
-	 * kernel is running.
-	 */
-	check_tmpexec();
+	seccomp_test_stub_data = mmap(0, sizeof(*seccomp_test_stub_data),
+				      PROT_READ | PROT_WRITE,
+				      MAP_SHARED | MAP_ANON, 0, 0);
 
-	pid = start_ptraced_child();
-	if (init_registers(pid))
-		fatal("Failed to initialize default registers");
-	stop_ptraced_child(pid, 1, 1);
-}
+	/* Use the syscall data area as stack, we just need something */
+	sp = (unsigned long)&seccomp_test_stub_data->syscall_data +
+	     sizeof(seccomp_test_stub_data->syscall_data) -
+	     sizeof(void *);
+	pid = clone(seccomp_helper, (void *)sp, CLONE_VFORK | CLONE_VM, NULL);
 
-static int __init noprocmm_cmd_param(char *str, int* add)
-{
-	disable_proc_mm = 1;
-	return 0;
-}
+	if (pid < 0)
+		fatal_perror("check_seccomp : clone failed");
 
-__uml_setup("noprocmm", noprocmm_cmd_param,
-"noprocmm\n"
-"    Turns off usage of /proc/mm, even if host supports it.\n"
-"    To support /proc/mm, the host needs to be patched using\n"
-"    the current skas3 patch.\n\n");
+	CATCH_EINTR(n = waitpid(pid, &status, __WCLONE));
+	if (n < 0)
+		fatal_perror("check_seccomp : waitpid failed");
 
-static int __init noptracefaultinfo_cmd_param(char *str, int* add)
-{
-	disable_ptrace_faultinfo = 1;
-	return 0;
-}
+	if (WIFEXITED(status) && WEXITSTATUS(status) == 0) {
+		struct uml_pt_regs *regs;
+		unsigned long fp_size;
+		int r;
 
-__uml_setup("noptracefaultinfo", noptracefaultinfo_cmd_param,
-"noptracefaultinfo\n"
-"    Turns off usage of PTRACE_FAULTINFO, even if host supports\n"
-"    it. To support PTRACE_FAULTINFO, the host needs to be patched\n"
-"    using the current skas3 patch.\n\n");
+		/* Fill in the host_fp_size from the mcontext. */
+		regs = calloc(1, sizeof(struct uml_pt_regs));
+		get_stub_state(regs, seccomp_test_stub_data, &fp_size);
+		host_fp_size = fp_size;
+		free(regs);
 
-static int __init noptraceldt_cmd_param(char *str, int* add)
-{
-	disable_ptrace_ldt = 1;
-	return 0;
-}
+		/* Repeat with the correct size */
+		regs = calloc(1, sizeof(struct uml_pt_regs) + host_fp_size);
+		r = get_stub_state(regs, seccomp_test_stub_data, NULL);
 
-__uml_setup("noptraceldt", noptraceldt_cmd_param,
-"noptraceldt\n"
-"    Turns off usage of PTRACE_LDT, even if host supports it.\n"
-"    To support PTRACE_LDT, the host needs to be patched using\n"
-"    the current skas3 patch.\n\n");
+		/* Store as the default startup registers */
+		exec_fp_regs = malloc(host_fp_size);
+		memcpy(exec_regs, regs->gp, sizeof(exec_regs));
+		memcpy(exec_fp_regs, regs->fp, host_fp_size);
 
-static inline void check_skas3_ptrace_faultinfo(void)
-{
-	struct ptrace_faultinfo fi;
-	int pid, n;
+		munmap(seccomp_test_stub_data, sizeof(*seccomp_test_stub_data));
 
-	non_fatal("  - PTRACE_FAULTINFO...");
-	pid = start_ptraced_child();
+		free(regs);
+
+		if (r) {
+			os_info("failed to fetch registers: %d\n", r);
+			return false;
+		}
 
-	n = ptrace(PTRACE_FAULTINFO, pid, 0, &fi);
-	if (n < 0) {
-		if (errno == EIO)
-			non_fatal("not found\n");
-		else
-			perror("not found");
-	} else if (disable_ptrace_faultinfo)
-		non_fatal("found but disabled on command line\n");
-	else {
-		ptrace_faultinfo = 1;
-		non_fatal("found\n");
+		os_info("OK\n");
+		return true;
 	}
 
-	stop_ptraced_child(pid, 1, 1);
+	if (WIFEXITED(status) && WEXITSTATUS(status) == 2)
+		os_info("missing\n");
+	else
+		os_info("error\n");
+
+	munmap(seccomp_test_stub_data, sizeof(*seccomp_test_stub_data));
+	return false;
 }
 
-static inline void check_skas3_ptrace_ldt(void)
+
+static void __init check_coredump_limit(void)
 {
-#ifdef PTRACE_LDT
-	int pid, n;
-	unsigned char ldtbuf[40];
-	struct ptrace_ldt ldt_op = (struct ptrace_ldt) {
-		.func = 2, /* read default ldt */
-		.ptr = ldtbuf,
-		.bytecount = sizeof(ldtbuf)};
-
-	non_fatal("  - PTRACE_LDT...");
-	pid = start_ptraced_child();
+	struct rlimit lim;
+	int err = getrlimit(RLIMIT_CORE, &lim);
 
-	n = ptrace(PTRACE_LDT, pid, 0, (unsigned long) &ldt_op);
-	if (n < 0) {
-		if (errno == EIO)
-			non_fatal("not found\n");
-		else
-			perror("not found");
-	} else if (disable_ptrace_ldt)
-		non_fatal("found, but use is disabled\n");
-	else {
-		ptrace_ldt = 1;
-		non_fatal("found\n");
+	if (err) {
+		perror("Getting core dump limit");
+		return;
 	}
 
-	stop_ptraced_child(pid, 1, 1);
-#endif
+	os_info("Core dump limits :\n\tsoft - ");
+	if (lim.rlim_cur == RLIM_INFINITY)
+		os_info("NONE\n");
+	else
+		os_info("%llu\n", (unsigned long long)lim.rlim_cur);
+
+	os_info("\thard - ");
+	if (lim.rlim_max == RLIM_INFINITY)
+		os_info("NONE\n");
+	else
+		os_info("%llu\n", (unsigned long long)lim.rlim_max);
 }
 
-static inline void check_skas3_proc_mm(void)
+void  __init get_host_cpu_features(
+		void (*flags_helper_func)(char *line),
+		void (*cache_helper_func)(char *line))
 {
-	non_fatal("  - /proc/mm...");
-	if (access("/proc/mm", W_OK) < 0)
-		perror("not found");
-	else if (disable_proc_mm)
-		non_fatal("found but disabled on command line\n");
-	else {
-		proc_mm = 1;
-		non_fatal("found\n");
+	FILE *cpuinfo;
+	char *line = NULL;
+	size_t len = 0;
+	int done_parsing = 0;
+
+	cpuinfo = fopen("/proc/cpuinfo", "r");
+	if (cpuinfo == NULL) {
+		os_info("Failed to get host CPU features\n");
+	} else {
+		while ((getline(&line, &len, cpuinfo)) != -1) {
+			if (strstr(line, "flags")) {
+				flags_helper_func(line);
+				done_parsing++;
+			}
+			if (strstr(line, "cache_alignment")) {
+				cache_helper_func(line);
+				done_parsing++;
+			}
+			free(line);
+			line = NULL;
+			if (done_parsing > 1)
+				break;
+		}
+		fclose(cpuinfo);
 	}
 }
 
-void can_do_skas(void)
-{
-	non_fatal("Checking for the skas3 patch in the host:\n");
+static int seccomp_config __initdata;
 
-	check_skas3_proc_mm();
-	check_skas3_ptrace_faultinfo();
-	check_skas3_ptrace_ldt();
+static int __init uml_seccomp_config(char *line, int *add)
+{
+	*add = 0;
+
+	if (strcmp(line, "off") == 0)
+		seccomp_config = 0;
+	else if (strcmp(line, "auto") == 0)
+		seccomp_config = 1;
+	else if (strcmp(line, "on") == 0)
+		seccomp_config = 2;
+	else
+		fatal("Invalid seccomp option '%s', expected on/auto/off\n",
+		      line);
 
-	if (!proc_mm || !ptrace_faultinfo || !ptrace_ldt)
-		skas_needs_stub = 1;
+	return 0;
 }
 
-int __init parse_iomem(char *str, int *add)
+__uml_setup("seccomp=", uml_seccomp_config,
+"seccomp=<on/auto/off>\n"
+"    Configure whether or not SECCOMP is used. With SECCOMP, userspace\n"
+"    processes work collaboratively with the kernel instead of being\n"
+"    traced using ptrace. All syscalls from the application are caught and\n"
+"    redirected using a signal. This signal handler in turn is permitted to\n"
+"    do the selected set of syscalls to communicate with the UML kernel and\n"
+"    do the required memory management.\n"
+"\n"
+"    This method is overall faster than the ptrace based userspace, primarily\n"
+"    because it reduces the number of context switches for (minor) page faults.\n"
+"\n"
+"    However, the SECCOMP filter is not (yet) restrictive enough to prevent\n"
+"    userspace from reading and writing all physical memory. Userspace\n"
+"    processes could also trick the stub into disabling SIGALRM which\n"
+"    prevents it from being interrupted for scheduling purposes.\n"
+"\n"
+"    This is insecure and should only be used with a trusted userspace\n\n"
+);
+
+void __init os_early_checks(void)
 {
-	struct iomem_region *new;
-	struct stat64 buf;
-	char *file, *driver;
-	int fd, size;
-
-	driver = str;
-	file = strchr(str,',');
-	if (file == NULL) {
-		fprintf(stderr, "parse_iomem : failed to parse iomem\n");
-		goto out;
-	}
-	*file = '\0';
-	file++;
-	fd = open(file, O_RDWR, 0);
-	if (fd < 0) {
-		perror("parse_iomem - Couldn't open io file");
-		goto out;
-	}
+	int pid;
 
-	if (fstat64(fd, &buf) < 0) {
-		perror("parse_iomem - cannot stat_fd file");
-		goto out_close;
-	}
+	/* Print out the core dump limits early */
+	check_coredump_limit();
+
+	/* Need to check this early because mmapping happens before the
+	 * kernel is running.
+	 */
+	check_tmpexec();
+
+	if (seccomp_config) {
+		if (init_seccomp()) {
+			using_seccomp = 1;
+			return;
+		}
 
-	new = malloc(sizeof(*new));
-	if (new == NULL) {
-		perror("Couldn't allocate iomem_region struct");
-		goto out_close;
+		if (seccomp_config == 2)
+			fatal("SECCOMP userspace requested but not functional!\n");
 	}
 
-	size = (buf.st_size + UM_KERN_PAGE_SIZE) & ~(UM_KERN_PAGE_SIZE - 1);
+	if (uml_ncpus > 1)
+		fatal("SMP is not supported with PTRACE userspace.\n");
 
-	*new = ((struct iomem_region) { .next		= iomem_regions,
-					.driver		= driver,
-					.fd		= fd,
-					.size		= size,
-					.phys		= 0,
-					.virt		= 0 });
-	iomem_regions = new;
-	iomem_size += new->size + UM_KERN_PAGE_SIZE;
+	using_seccomp = 0;
+	check_ptrace();
 
-	return 0;
- out_close:
-	close(fd);
- out:
-	return 1;
+	pid = start_ptraced_child();
+	if (init_pid_registers(pid))
+		fatal("Failed to initialize default registers");
+	stop_ptraced_child(pid, 1);
 }