This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Tidy i386 signal handler
- From: Andrew Haley <aph at redhat dot com>
- To: java at gcc dot gnu dot org
- Date: Thu, 11 Jan 2007 19:29:17 +0000
- Subject: Tidy i386 signal handler
The main purpose of this is to remove unnecessary differences between
i386 and x86_64 signal handlers accesses to undocumented fields on the
stack, and sundry strict-aliasing violations.
A side-effect is to use rt_sigaction() rather than sigaction(), and
this makes libgcj incompatible with pre-Linux 2.1.68 (30 Nov 1997)
kernels. Let me know if you care.
Andrew.
2007-01-11 Andrew Haley <aph@redhat.com>
* include/i386-signal.h: Rewrite to use rt_sigaction.
Index: include/i386-signal.h
===================================================================
--- include/i386-signal.h (revision 120638)
+++ include/i386-signal.h (working copy)
@@ -19,46 +19,51 @@
#define HANDLE_SEGV 1
#define HANDLE_FPE 1
-#define SIGNAL_HANDLER(_name) \
-static void _name (int _dummy __attribute__ ((__unused__)))
-
-#define MAKE_THROW_FRAME(_exception)
+#define SIGNAL_HANDLER(_name) \
+static void _Jv_##_name (int, siginfo_t *, \
+ void *_p __attribute__ ((__unused__)))
#define HANDLE_DIVIDE_OVERFLOW \
do \
{ \
- void **_p = (void **)&_dummy; \
- volatile struct sigcontext_struct *_regs = (struct sigcontext_struct *)++_p;\
- \
- register unsigned char *_eip = (unsigned char *)_regs->eip; \
+ struct ucontext *_uc = (struct ucontext *)_p; \
+ gregset_t &_gregs = _uc->uc_mcontext.gregs; \
+ unsigned char *_eip = (unsigned char *)_gregs[REG_EIP]; \
\
/* According to the JVM spec, "if the dividend is the negative \
- * integer of the smallest magnitude and the divisor is -1, then \
- * overflow occurs and the result is equal to the dividend. Despite \
- * the overflow, no exception occurs". \
+ * integer of largest possible magnitude for the type and the \
+ * divisor is -1, then overflow occurs and the result is equal to \
+ * the dividend. Despite the overflow, no exception occurs". \
\
* We handle this by inspecting the instruction which generated the \
- * signal and advancing eip to point to the following instruction. \
+ * signal and advancing ip to point to the following instruction. \
* As the instructions are variable length it is necessary to do a \
* little calculation to figure out where the following instruction \
* actually is. \
\
- */ \
+ */ \
\
+ /* Detect a signed division of Integer.MIN_VALUE. */ \
if (_eip[0] == 0xf7) \
{ \
+ bool _min_value_dividend = false; \
unsigned char _modrm = _eip[1]; \
\
- if (_regs->eax == 0x80000000 \
- && ((_modrm >> 3) & 7) == 7) /* Signed divide */ \
+ if (((_modrm >> 3) & 7) == 7) /* Signed divide */ \
+ { \
+ _min_value_dividend = \
+ _gregs[REG_EAX] == (greg_t)0x80000000UL; \
+ } \
+ \
+ if (_min_value_dividend) \
{ \
unsigned char _rm = _modrm & 7; \
- _regs->edx = 0; /* the remainder is zero */ \
+ _gregs[REG_EDX] = 0; /* the remainder is zero */ \
switch (_modrm >> 6) \
{ \
case 0: /* register indirect */ \
if (_rm == 5) /* 32-bit displacement */ \
- _eip += 4; \
+ _eip += 4; \
if (_rm == 4) /* A SIB byte follows the ModR/M byte */ \
_eip += 1; \
break; \
@@ -76,24 +81,30 @@
break; \
} \
_eip += 2; \
- _regs->eip = (unsigned long)_eip; \
+ _gregs[REG_EIP] = (greg_t)_eip; \
return; \
} \
} \
-} \
+ } \
while (0)
-/* We use old_kernel_sigaction here because we're calling the kernel
+/* We use kernel_sigaction here because we're calling the kernel
directly rather than via glibc. The sigaction structure that the
syscall uses is a different shape from the one in userland and not
visible to us in a header file so we define it here. */
-struct old_i386_kernel_sigaction {
- void (*k_sa_handler) (int);
- unsigned long k_sa_mask;
- unsigned long k_sa_flags;
- void (*sa_restorer) (void);
-};
+extern "C"
+{
+ struct kernel_sigaction
+ {
+ void (*k_sa_sigaction)(int,siginfo_t *,void *);
+ unsigned long k_sa_flags;
+ void (*k_sa_restorer) (void);
+ sigset_t k_sa_mask;
+ };
+}
+
+#define MAKE_THROW_FRAME(_exception)
#define RESTORE(name, syscall) RESTORE2 (name, syscall)
# define RESTORE2(name, syscall) \
@@ -101,38 +112,39 @@
( \
".text\n" \
".byte 0 # Yes, this really is necessary\n" \
- " .align 8\n" \
+ " .align 16\n" \
"__" #name ":\n" \
- " popl %eax\n" \
" movl $" #syscall ", %eax\n" \
" int $0x80" \
);
-RESTORE (restore, __NR_sigreturn)
-static void restore (void) asm ("__restore");
-
-#define INIT_SEGV \
-do \
- { \
- struct old_i386_kernel_sigaction kact; \
- kact.k_sa_handler = catch_segv; \
- kact.k_sa_mask = 0; \
- kact.k_sa_flags = 0x4000000; \
- kact.sa_restorer = restore; \
- syscall (SYS_sigaction, SIGSEGV, &kact, NULL); \
- } \
+/* The return code for realtime-signals. */
+RESTORE (restore_rt, __NR_rt_sigreturn)
+void restore_rt (void) asm ("__restore_rt")
+ __attribute__ ((visibility ("hidden")));
+
+#define INIT_SEGV \
+do \
+ { \
+ struct kernel_sigaction act; \
+ act.k_sa_sigaction = _Jv_catch_segv; \
+ sigemptyset (&act.k_sa_mask); \
+ act.k_sa_flags = SA_SIGINFO|0x4000000; \
+ act.k_sa_restorer = restore_rt; \
+ syscall (SYS_rt_sigaction, SIGSEGV, &act, NULL, _NSIG / 8); \
+ } \
while (0)
-#define INIT_FPE \
-do \
- { \
- struct old_i386_kernel_sigaction kact; \
- kact.k_sa_handler = catch_fpe; \
- kact.k_sa_mask = 0; \
- kact.k_sa_flags = 0x4000000; \
- kact.sa_restorer = restore; \
- syscall (SYS_sigaction, SIGFPE, &kact, NULL); \
- } \
+#define INIT_FPE \
+do \
+ { \
+ struct kernel_sigaction act; \
+ act.k_sa_sigaction = _Jv_catch_fpe; \
+ sigemptyset (&act.k_sa_mask); \
+ act.k_sa_flags = SA_SIGINFO|0x4000000; \
+ act.k_sa_restorer = restore_rt; \
+ syscall (SYS_rt_sigaction, SIGFPE, &act, NULL, _NSIG / 8); \
+ } \
while (0)
/* You might wonder why we use syscall(SYS_sigaction) in INIT_FPE