#ifndef QEMU_H
#define QEMU_H
-#include <signal.h>
-#include <string.h>
-
+#include "hostdep.h"
#include "cpu.h"
+#include "exec/exec-all.h"
+#include "exec/cpu_ldst.h"
#undef DEBUG_REMAP
#ifdef DEBUG_REMAP
-#include <stdlib.h>
#endif /* DEBUG_REMAP */
#include "exec/user/abitypes.h"
#include "exec/user/thunk.h"
#include "syscall_defs.h"
-#include "syscall.h"
+#include "target_syscall.h"
#include "exec/gdbstub.h"
#include "qemu/queue.h"
#define THREAD __thread
+/* This is the size of the host kernel's sigset_t, needed where we make
+ * direct system calls that take a sigset_t pointer and a size.
+ */
+#define SIGSET_T_SIZE (_NSIG / 8)
+
/* This struct is used to hold certain information about the image.
* Basically, it replicates in user space what would be certain
* task_struct fields in the kernel
abi_ulong start_brk;
abi_ulong brk;
abi_ulong start_mmap;
- abi_ulong mmap;
- abi_ulong rss;
abi_ulong start_stack;
abi_ulong stack_limit;
abi_ulong entry;
abi_ulong auxv_len;
abi_ulong arg_start;
abi_ulong arg_end;
+ abi_ulong arg_strings;
+ abi_ulong env_strings;
+ abi_ulong file_string;
uint32_t elf_flags;
int personality;
#ifdef CONFIG_USE_FDPIC
#define MAX_SIGQUEUE_SIZE 1024
-struct sigqueue {
- struct sigqueue *next;
- target_siginfo_t info;
-};
-
struct emulated_sigtable {
int pending; /* true if signal is pending */
- struct sigqueue *first;
- struct sigqueue info; /* in order to always have memory for the
- first signal, we put it here */
+ target_siginfo_t info;
};
/* NOTE: we force a big alignment so that the stack stored after is
#endif
#if defined(TARGET_ARM) || defined(TARGET_M68K) || defined(TARGET_UNICORE32)
/* Extra fields for semihosted binaries. */
- uint32_t heap_base;
- uint32_t heap_limit;
+ abi_ulong heap_base;
+ abi_ulong heap_limit;
#endif
- uint32_t stack_base;
+ abi_ulong stack_base;
int used; /* non zero if used */
struct image_info *info;
struct linux_binprm *bprm;
+ struct emulated_sigtable sync_signal;
struct emulated_sigtable sigtab[TARGET_NSIG];
- struct sigqueue sigqueue_table[MAX_SIGQUEUE_SIZE]; /* siginfo queue */
- struct sigqueue *first_free; /* first free siginfo queue entry */
- int signal_pending; /* non zero if a signal may be pending */
+ /* This thread's signal mask, as requested by the guest program.
+ * The actual signal mask of this thread may differ:
+ * + we don't let SIGSEGV and SIGBUS be blocked while running guest code
+ * + sometimes we block all signals to avoid races
+ */
+ sigset_t signal_mask;
+ /* The signal mask imposed by a guest sigsuspend syscall, if we are
+ * currently in the middle of such a syscall
+ */
+ sigset_t sigsuspend_mask;
+ /* Nonzero if we're leaving a sigsuspend and sigsuspend_mask is valid. */
+ int in_sigsuspend;
+
+ /* Nonzero if process_pending_signals() needs to do something (either
+ * handle a pending signal or unblock signals).
+ * This flag is written from a signal handler so should be accessed via
+ * the atomic_read() and atomic_write() functions. (It is not accessed
+ * from multiple threads.)
+ */
+ int signal_pending;
+
} __attribute__((aligned(16))) TaskState;
extern char *exec_path;
extern unsigned long mmap_min_addr;
/* ??? See if we can avoid exposing so much of the loader internals. */
-/*
- * MAX_ARG_PAGES defines the number of pages allocated for arguments
- * and envelope for the new program. 32 should suffice, this gives
- * a maximum env+arg of 128kB w/4KB pages!
- */
-#define MAX_ARG_PAGES 33
/* Read a good amount of data initially, to hopefully get all the
program headers loaded. */
*/
struct linux_binprm {
char buf[BPRM_BUF_SIZE] __attribute__((aligned));
- void *page[MAX_ARG_PAGES];
abi_ulong p;
int fd;
int e_uid, e_gid;
void gemu_log(const char *fmt, ...) GCC_FMT_ATTR(1, 2);
extern THREAD CPUState *thread_cpu;
void cpu_loop(CPUArchState *env);
-char *target_strerror(int err);
+const char *target_strerror(int err);
int get_osversion(void);
void init_qemu_uname_release(void);
void fork_start(void);
#include "qemu/log.h"
+/* safe_syscall.S */
+
+/**
+ * safe_syscall:
+ * @int number: number of system call to make
+ * ...: arguments to the system call
+ *
+ * Call a system call if guest signal not pending.
+ * This has the same API as the libc syscall() function, except that it
+ * may return -1 with errno == TARGET_ERESTARTSYS if a signal was pending.
+ *
+ * Returns: the system call result, or -1 with an error code in errno
+ * (Errnos are host errnos; we rely on TARGET_ERESTARTSYS not clashing
+ * with any of the host errno values.)
+ */
+
+/* A guide to using safe_syscall() to handle interactions between guest
+ * syscalls and guest signals:
+ *
+ * Guest syscalls come in two flavours:
+ *
+ * (1) Non-interruptible syscalls
+ *
+ * These are guest syscalls that never get interrupted by signals and
+ * so never return EINTR. They can be implemented straightforwardly in
+ * QEMU: just make sure that if the implementation code has to make any
+ * blocking calls that those calls are retried if they return EINTR.
+ * It's also OK to implement these with safe_syscall, though it will be
+ * a little less efficient if a signal is delivered at the 'wrong' moment.
+ *
+ * Some non-interruptible syscalls need to be handled using block_signals()
+ * to block signals for the duration of the syscall. This mainly applies
+ * to code which needs to modify the data structures used by the
+ * host_signal_handler() function and the functions it calls, including
+ * all syscalls which change the thread's signal mask.
+ *
+ * (2) Interruptible syscalls
+ *
+ * These are guest syscalls that can be interrupted by signals and
+ * for which we need to either return EINTR or arrange for the guest
+ * syscall to be restarted. This category includes both syscalls which
+ * always restart (and in the kernel return -ERESTARTNOINTR), ones
+ * which only restart if there is no handler (kernel returns -ERESTARTNOHAND
+ * or -ERESTART_RESTARTBLOCK), and the most common kind which restart
+ * if the handler was registered with SA_RESTART (kernel returns
+ * -ERESTARTSYS). System calls which are only interruptible in some
+ * situations (like 'open') also need to be handled this way.
+ *
+ * Here it is important that the host syscall is made
+ * via this safe_syscall() function, and *not* via the host libc.
+ * If the host libc is used then the implementation will appear to work
+ * most of the time, but there will be a race condition where a
+ * signal could arrive just before we make the host syscall inside libc,
+ * and then then guest syscall will not correctly be interrupted.
+ * Instead the implementation of the guest syscall can use the safe_syscall
+ * function but otherwise just return the result or errno in the usual
+ * way; the main loop code will take care of restarting the syscall
+ * if appropriate.
+ *
+ * (If the implementation needs to make multiple host syscalls this is
+ * OK; any which might really block must be via safe_syscall(); for those
+ * which are only technically blocking (ie which we know in practice won't
+ * stay in the host kernel indefinitely) it's OK to use libc if necessary.
+ * You must be able to cope with backing out correctly if some safe_syscall
+ * you make in the implementation returns either -TARGET_ERESTARTSYS or
+ * EINTR though.)
+ *
+ * block_signals() cannot be used for interruptible syscalls.
+ *
+ *
+ * How and why the safe_syscall implementation works:
+ *
+ * The basic setup is that we make the host syscall via a known
+ * section of host native assembly. If a signal occurs, our signal
+ * handler checks the interrupted host PC against the addresse of that
+ * known section. If the PC is before or at the address of the syscall
+ * instruction then we change the PC to point at a "return
+ * -TARGET_ERESTARTSYS" code path instead, and then exit the signal handler
+ * (causing the safe_syscall() call to immediately return that value).
+ * Then in the main.c loop if we see this magic return value we adjust
+ * the guest PC to wind it back to before the system call, and invoke
+ * the guest signal handler as usual.
+ *
+ * This winding-back will happen in two cases:
+ * (1) signal came in just before we took the host syscall (a race);
+ * in this case we'll take the guest signal and have another go
+ * at the syscall afterwards, and this is indistinguishable for the
+ * guest from the timing having been different such that the guest
+ * signal really did win the race
+ * (2) signal came in while the host syscall was blocking, and the
+ * host kernel decided the syscall should be restarted;
+ * in this case we want to restart the guest syscall also, and so
+ * rewinding is the right thing. (Note that "restart" semantics mean
+ * "first call the signal handler, then reattempt the syscall".)
+ * The other situation to consider is when a signal came in while the
+ * host syscall was blocking, and the host kernel decided that the syscall
+ * should not be restarted; in this case QEMU's host signal handler will
+ * be invoked with the PC pointing just after the syscall instruction,
+ * with registers indicating an EINTR return; the special code in the
+ * handler will not kick in, and we will return EINTR to the guest as
+ * we should.
+ *
+ * Notice that we can leave the host kernel to make the decision for
+ * us about whether to do a restart of the syscall or not; we do not
+ * need to check SA_RESTART flags in QEMU or distinguish the various
+ * kinds of restartability.
+ */
+#ifdef HAVE_SAFE_SYSCALL
+/* The core part of this function is implemented in assembly */
+extern long safe_syscall_base(int *pending, long number, ...);
+
+#define safe_syscall(...) \
+ ({ \
+ long ret_; \
+ int *psp_ = &((TaskState *)thread_cpu->opaque)->signal_pending; \
+ ret_ = safe_syscall_base(psp_, __VA_ARGS__); \
+ if (is_error(ret_)) { \
+ errno = -ret_; \
+ ret_ = -1; \
+ } \
+ ret_; \
+ })
+
+#else
+
+/* Fallback for architectures which don't yet provide a safe-syscall assembly
+ * fragment; note that this is racy!
+ * This should go away when all host architectures have been updated.
+ */
+#define safe_syscall syscall
+
+#endif
+
/* syscall.c */
int host_to_target_waitstatus(int status);
abi_long arg1, abi_long arg2, abi_long arg3,
abi_long arg4, abi_long arg5, abi_long arg6);
void print_syscall_ret(int num, abi_long arg1);
+/**
+ * print_taken_signal:
+ * @target_signum: target signal being taken
+ * @tinfo: target_siginfo_t which will be passed to the guest for the signal
+ *
+ * Print strace output indicating that this signal is being taken by the guest,
+ * in a format similar to:
+ * --- SIGSEGV {si_signo=SIGSEGV, si_code=SI_KERNEL, si_addr=0} ---
+ */
+void print_taken_signal(int target_signum, const target_siginfo_t *tinfo);
extern int do_strace;
/* signal.c */
void process_pending_signals(CPUArchState *cpu_env);
void signal_init(void);
-int queue_signal(CPUArchState *env, int sig, target_siginfo_t *info);
+int queue_signal(CPUArchState *env, int sig, int si_type,
+ target_siginfo_t *info);
void host_to_target_siginfo(target_siginfo_t *tinfo, const siginfo_t *info);
void target_to_host_siginfo(siginfo_t *info, const target_siginfo_t *tinfo);
int target_to_host_signal(int sig);
long do_sigreturn(CPUArchState *env);
long do_rt_sigreturn(CPUArchState *env);
abi_long do_sigaltstack(abi_ulong uss_addr, abi_ulong uoss_addr, abi_ulong sp);
+int do_sigprocmask(int how, const sigset_t *set, sigset_t *oldset);
+/**
+ * block_signals: block all signals while handling this guest syscall
+ *
+ * Block all signals, and arrange that the signal mask is returned to
+ * its correct value for the guest before we resume execution of guest code.
+ * If this function returns non-zero, then the caller should immediately
+ * return -TARGET_ERESTARTSYS to the main loop, which will take the pending
+ * signal and restart execution of the syscall.
+ * If block_signals() returns zero, then the caller can continue with
+ * emulation of the system call knowing that no signals can be taken
+ * (and therefore that no race conditions will result).
+ * This should only be called once, because if it is called a second time
+ * it will always return non-zero. (Think of it like a mutex that can't
+ * be recursively locked.)
+ * Signals will be unblocked again by process_pending_signals().
+ *
+ * Return value: non-zero if there was a pending signal, zero if not.
+ */
+int block_signals(void); /* Returns non zero if signal pending */
#ifdef TARGET_I386
/* vm86.c */
int target_msync(abi_ulong start, abi_ulong len, int flags);
extern unsigned long last_brk;
extern abi_ulong mmap_next_start;
-void mmap_lock(void);
-void mmap_unlock(void);
abi_ulong mmap_find_vma(abi_ulong, abi_ulong);
-void cpu_list_lock(void);
-void cpu_list_unlock(void);
void mmap_fork_start(void);
void mmap_fork_end(int child);
__builtin_choose_expr(sizeof(*(hptr)) == 2, stw_##e##_p, \
__builtin_choose_expr(sizeof(*(hptr)) == 4, stl_##e##_p, \
__builtin_choose_expr(sizeof(*(hptr)) == 8, stq_##e##_p, abort)))) \
- ((hptr), (x)), 0)
+ ((hptr), (x)), (void)0)
#define __get_user_e(x, hptr, e) \
((x) = (typeof(*hptr))( \
__builtin_choose_expr(sizeof(*(hptr)) == 2, lduw_##e##_p, \
__builtin_choose_expr(sizeof(*(hptr)) == 4, ldl_##e##_p, \
__builtin_choose_expr(sizeof(*(hptr)) == 8, ldq_##e##_p, abort)))) \
- (hptr)), 0)
+ (hptr)), (void)0)
#ifdef TARGET_WORDS_BIGENDIAN
# define __put_user(x, hptr) __put_user_e(x, hptr, be)
({ \
abi_ulong __gaddr = (gaddr); \
target_type *__hptr; \
- abi_long __ret; \
+ abi_long __ret = 0; \
if ((__hptr = lock_user(VERIFY_WRITE, __gaddr, sizeof(target_type), 0))) { \
- __ret = __put_user((x), __hptr); \
+ __put_user((x), __hptr); \
unlock_user(__hptr, __gaddr, sizeof(target_type)); \
} else \
__ret = -TARGET_EFAULT; \
({ \
abi_ulong __gaddr = (gaddr); \
target_type *__hptr; \
- abi_long __ret; \
+ abi_long __ret = 0; \
if ((__hptr = lock_user(VERIFY_READ, __gaddr, sizeof(target_type), 1))) { \
- __ret = __get_user((x), __hptr); \
+ __get_user((x), __hptr); \
unlock_user(__hptr, __gaddr, 0); \
} else { \
/* avoid warning */ \
#ifdef DEBUG_REMAP
{
void *addr;
- addr = malloc(len);
+ addr = g_malloc(len);
if (copy)
memcpy(addr, g2h(guest_addr), len);
else
return;
if (len > 0)
memcpy(g2h(guest_addr), host_ptr, len);
- free(host_ptr);
+ g_free(host_ptr);
#endif
}
projects / mirror_qemu.git / blobdiff