// to be used only from a BView or BWindow bool nsToolkit::CallMethodAsync(MethodInfo *info) { #ifdef DEBUG printf("CMA\n"); #endif ThreadInterfaceData id; GetInterface(); id.data = info; id.waitingThread = 0; // Check message count to not exceed the port's capacity. // There seems to be a BeOS bug that allows more // messages on a port than its capacity. port_info portinfo; if (get_port_info(eventport, &portinfo) != B_OK) { return false; } if (port_count(eventport) < portinfo.capacity - 20) { if(write_port_etc(eventport, WM_CALLMETHOD, &id, sizeof(id), B_TIMEOUT, 0) == B_OK) { return true; } } return false; }
// Synchronous version, not in use at the moment bool nsToolkit::CallMethod(MethodInfo *info) { #ifdef DEBUG printf("TK-CM\n"); #endif ThreadInterfaceData id; GetInterface(); id.data = info; id.waitingThread = find_thread(NULL); // Check message count to not exceed the port's capacity. // There seems to be a BeOS bug that allows more // messages on a port than its capacity. port_info portinfo; if (get_port_info(eventport, &portinfo) != B_OK) { return false; } if (port_count(eventport) < portinfo.capacity - 20) { // If we cannot write inside 5 seconds, something is really wrong. if(write_port_etc(eventport, WM_CALLMETHOD, &id, sizeof(id), B_TIMEOUT, 5000000) == B_OK) { // semantics for CallMethod are that it should be synchronous suspend_thread(id.waitingThread); return true; } } return false; }
void port_test() { char testdata[5]; thread_id t; int res; int32 dummy; int32 dummy2; strcpy(testdata, "abcd"); dprintf("porttest: port_create()\n"); test_p1 = port_create(1, "test port #1"); test_p2 = port_create(10, "test port #2"); test_p3 = port_create(1024, "test port #3"); test_p4 = port_create(1024, "test port #4"); dprintf("porttest: port_find()\n"); dprintf("'test port #1' has id %d (should be %d)\n", port_find("test port #1"), test_p1); dprintf("porttest: port_write() on 1, 2 and 3\n"); port_write(test_p1, 1, &testdata, sizeof(testdata)); port_write(test_p2, 666, &testdata, sizeof(testdata)); port_write(test_p3, 999, &testdata, sizeof(testdata)); dprintf("porttest: port_count(test_p1) = %d\n", port_count(test_p1)); dprintf("porttest: port_write() on 1 with timeout of 1 sec (blocks 1 sec)\n"); port_write_etc(test_p1, 1, &testdata, sizeof(testdata), PORT_FLAG_TIMEOUT, 1000000); dprintf("porttest: port_write() on 2 with timeout of 1 sec (wont block)\n"); res = port_write_etc(test_p2, 777, &testdata, sizeof(testdata), PORT_FLAG_TIMEOUT, 1000000); dprintf("porttest: res=%d, %s\n", res, res == 0 ? "ok" : "BAD"); dprintf("porttest: port_read() on empty port 4 with timeout of 1 sec (blocks 1 sec)\n"); res = port_read_etc(test_p4, &dummy, &dummy2, sizeof(dummy2), PORT_FLAG_TIMEOUT, 1000000); dprintf("porttest: res=%d, %s\n", res, res == ERR_PORT_TIMED_OUT ? "ok" : "BAD"); dprintf("porttest: spawning thread for port 1\n"); t = thread_create_kernel_thread("port_test", port_test_thread_func, NULL); // resume thread thread_resume_thread(t); dprintf("porttest: write\n"); port_write(test_p1, 1, &testdata, sizeof(testdata)); // now we can write more (no blocking) dprintf("porttest: write #2\n"); port_write(test_p1, 2, &testdata, sizeof(testdata)); dprintf("porttest: write #3\n"); port_write(test_p1, 3, &testdata, sizeof(testdata)); dprintf("porttest: waiting on spawned thread\n"); thread_wait_on_thread(t, NULL); dprintf("porttest: close p1\n"); port_close(test_p2); dprintf("porttest: attempt write p1 after close\n"); res = port_write(test_p2, 4, &testdata, sizeof(testdata)); dprintf("porttest: port_write ret %d\n", res); dprintf("porttest: testing delete p2\n"); port_delete(test_p2); dprintf("porttest: end test main thread\n"); }
int32 user_port_count(port_id port) { return port_count(port); }
ssize_t _user_port_count(port_id port) { return port_count(port); }
static void do_syscall_sw( uuprocess_t *u, struct trap_state *st) { #if HAVE_UNIX int callno = st->eax; /* unsigned int mina = 0, maxa = 0; { extern struct real_descriptor ldt[]; struct real_descriptor dsd = ldt[ st->ds/8 ]; mina = dsd.base_low | (dsd.base_med << 16) | (dsd.base_high << 24); maxa = (dsd.limit_low | dsd.limit_high << 16); if( dsd.granularity & SZ_G) maxa *= PAGE_SIZE; maxa += mina; } */ //phantom_thread_t *t = GET_CURRENT_THREAD(); //uuprocess_t *u = t->u; /* tid_t tid = get_current_tid(); pid_t pid; assert( !t_get_pid( tid, &pid )); uuprocess_t *u = proc_by_pid(pid); */ assert( u != 0 ); unsigned int mina = (int)u->mem_start, maxa = (int)u->mem_end; // trap state is good for interrupt. // call gate does not push eflags int user_esp = st->eflags; #ifndef ARCH_ia32 # warning machdep ia32 arg pass convention #endif // list of user syscsall arguments int *uarg = adjustin( user_esp, st ); uarg++; // syscall func return addr //SHOW_FLOW( 10, "Syscall pid %2d tid %2d, our esp %p, user esp %p, t kstack %p", u->pid, t->tid, &st, user_esp, t->kstack_top ); SHOW_FLOW( 8, "Syscall %d args %x, %x, %x", callno, uarg[0], uarg[1],uarg[2] ); int ret = 0; int err = 0; switch(callno) { case SYS__exit: // TODO housekeeping? SHOW_FLOW( 2, "exit %d", uarg[0] ); hal_exit_kernel_thread(); err = ENOSYS; break; case SYS_ssyslog: syslog( uarg[0], "%s", (const char *)adjustin( uarg[1], st ) ); SHOW_FLOW( 2, "syslog %d %s", uarg[0], (const char *)adjustin( uarg[1], st ) ); //SHOW_ERROR0( 0, "syslog not impl" ); break; case SYS_sleepmsec: //SHOW_FLOW( 2, "sleepmsec %d", uarg[0] ); hal_sleep_msec(uarg[0]); break; case SYS_getpagesize: ret = PAGE_SIZE; break; case SYS_personality: if( ((unsigned) uarg[0]) == 0xffffffff) { ret = 0; // Say we're Linux... well...to some extent. break; } err = EINVAL; break; case SYS_uname: err = copyout(uarg[0], mina, maxa, &phantom_uname, sizeof(phantom_uname) ); if( err ) ret = -1; break; case SYS_getuid: ret = u->uid; break; case SYS_getuid32: ret = u->uid;; break; case SYS_getegid: ret = u->egid; break; case SYS_getegid32: ret = u->egid; break; case SYS_geteuid: ret = u->euid; break; case SYS_geteuid32: ret = u->euid; break; case SYS_getgid: ret = u->gid; break; case SYS_getgid32: ret = u->gid; break; case SYS_gettid: ret = get_current_tid(); break; case SYS_getpid: ret = u->pid; break; case SYS_getpgrp: ret = u->pgrp_pid; break; case SYS_getppid: ret = u->ppid; break; case SYS_getpgid: goto unimpl; case SYS_time: { time_t t = time(0); AARG(time_t *, tp, 0, sizeof(time_t)); *tp = t; ret = t; break; } case SYS_nanosleep: // todo check for POSIX compliance, make it interruptible by signals //int nanosleep(const struct timespec *, struct timespec *); { goto unimpl; //AARG(struct timespec *, req_time, 0, sizeof(struct timespec)); //AARG(struct timespec *, rest_time, 1, sizeof(struct timespec)); //ret = usys_nanosleep( &err, u, req_time, rest_time ); //break; } case SYS_sync: case SYS_sysinfo: case SYS_sysfs: case SYS_klogctl: case SYS_shutdown: case SYS_reboot: case SYS_getitimer: case SYS_setitimer: case SYS_gettimeofday: case SYS_setuid: case SYS_setuid32: CHECK_CAP(CAP_SETUID); u->euid = uarg[0]; break; case SYS_setgid: case SYS_setgid32: CHECK_CAP(CAP_SETGID); u->egid = uarg[0]; break; case SYS_setgroups: case SYS_setgroups32: goto unimpl; case SYS_setpgid: case SYS_setregid: case SYS_setregid32: case SYS_setresgid: case SYS_setresgid32: case SYS_setresuid: case SYS_setresuid32: case SYS_setreuid: case SYS_setreuid32: goto unimpl; case SYS_open: ret = usys_open(&err, u, adjustin( uarg[0], st ), uarg[1], uarg[2] ); break; case SYS_close: ret = usys_close(&err, u, uarg[0] ); break; case SYS_read: { int count = uarg[2]; void *addr = adjustin( uarg[1], st ); CHECKA(addr,count); ret = usys_read(&err, u, uarg[0], addr, count ); break; } case SYS_write: { int count = uarg[2]; void *addr = adjustin( uarg[1], st ); CHECKA(addr,count); ret = usys_write(&err, u, uarg[0], addr, count ); break; } case SYS_lseek: ret = usys_lseek(&err, u, uarg[0], uarg[1], uarg[2] ); break; case SYS_creat: ret = usys_creat(&err, u, adjustin( uarg[0], st ), uarg[1] ); break; case SYS_chdir: { AARG(const char *, path, 0, 0); ret = usys_chdir(&err, u, path ); break; } case SYS_pipe: { AARG(int *, fds, 0, sizeof(int) * 2); ret = usys_pipe( &err, u, fds ); break; } case SYS_rmdir: case SYS_unlink: { AARG(const char *, name, 0, 1); ret = usys_rm( &err, u, name ); break; } case SYS_dup: ret = usys_dup( &err, u, uarg[0] ); break; case SYS_dup2: ret = usys_dup2( &err, u, uarg[0], uarg[1] ); break; case SYS_symlink: { AARG(const char *, n1, 0, 1); AARG(const char *, n2, 1, 1); ret = usys_symlink( &err, u, n1, n2 ); break; } case SYS_getcwd: { AARG(char *, buf, 0, uarg[1]); ret = usys_getcwd( &err, u, buf, uarg[1] ); break; } case SYS_mkdir: { AARG(const char *, path, 0, 2); ret = usys_mkdir( &err, u, path ); break; } case SYS_link: case SYS__llseek: case SYS_chroot: case SYS_lstat64: case SYS_mknod: goto unimpl; case SYS_mount: { const char *source = adjustin( uarg[0], st ); CHECKA(source,0); const char *target = adjustin( uarg[1], st ); CHECKA(target,0); const char *fstype = adjustin( uarg[2], st ); CHECKA(target,0); const void *data = adjustin( uarg[4], st ); CHECKA(data,0); ret = usys_mount(&err, u, source, target, fstype, uarg[3], data ); break; } case SYS_umount: { const char *target = adjustin( uarg[0], st ); CHECKA(target,0); ret = usys_umount(&err, u, target, 0 ); break; } /* case SYS_umount2: { const char *target = adjustin( uarg[0], st ); CHECKA(target); ret = usys_umount(&err, u, target, uarg[1] ); break; } */ case SYS_truncate: { AARG(const char *, path, 0, 0); ret = usys_truncate(&err, u, path, uarg[1] ); break; } case SYS_ftruncate: { ret = usys_ftruncate(&err, u, uarg[0], uarg[1] ); break; } case SYS_truncate64: case SYS_ftruncate64: goto unimpl; case SYS_fchdir: ret = usys_fchdir( &err, u, uarg[0] ); break; case SYS_readdir: { AARG(struct dirent *, ent, 1, sizeof(struct dirent)); ret = usys_readdir( &err, u, uarg[0], ent ); break; } case SYS_fchmod: ret = usys_fchmod( &err, u, uarg[0], uarg[1] ); break; case SYS_fchown: case SYS_fchown32: ret = usys_fchown( &err, u, uarg[0], uarg[1], uarg[2] ); break; case SYS_fcntl: { //AARG(void *, str, 2, sizeof(int)); // FIXME size of arg? ret = usys_fcntl( &err, u, uarg[0], uarg[1], uarg[0] ); } break; case SYS_fcntl64: case SYS_fdatasync: case SYS_flock: case SYS_fstat64: case SYS_fstatfs: case SYS_fsync: case SYS_utime: case SYS_chmod: case SYS_chown: case SYS_chown32: case SYS_access: case SYS_lchown: case SYS_lchown32: case SYS_pread: case SYS_pwrite: goto unimpl; case SYS_readv: { int fd = uarg[0]; int iovcnt = uarg[2]; AARG(struct iovec *, list, 1, sizeof(struct iovec) * iovcnt); unsigned int onerc; int lp; for( lp = 0; lp < iovcnt; lp++ ) { void *addr = adjustin( list[lp].iov_base, st ); CHECKA(addr,list[lp].iov_len); onerc = usys_read(&err, u, fd, addr, list[lp].iov_len ); /* if( onerc < 0 ) { ret = -1; goto err_ret; } */ ret += onerc; if( onerc < list[lp].iov_len ) break; } break; } case SYS_writev: { int fd = uarg[0]; int iovcnt = uarg[2]; AARG(struct iovec *, list, 1, sizeof(struct iovec) * iovcnt); unsigned int onerc; int lp; for( lp = 0; lp < iovcnt; lp++ ) { void *addr = adjustin( list[lp].iov_base, st ); CHECKA(addr,list[lp].iov_len); onerc = usys_write(&err, u, fd, addr, list[lp].iov_len ); /* if( onerc < 0 ) { ret = -1; goto err_ret; } */ ret += onerc; if( onerc < list[lp].iov_len ) break; } break; } case SYS_readlink: goto unimpl; case SYS_gethostname: { int len = uarg[1]; char *target = adjustin( uarg[0], st ); CHECKA(target,len); ret = usys_gethostname(&err, u, target, len ); break; } case SYS_sethostname: { int len = uarg[1]; const char *target = adjustin( uarg[0], st ); CHECKA(target,len); ret = usys_sethostname(&err, u, target, len ); break; } case SYS_socket: ret = usys_socket( &err, u, uarg[0], uarg[1], uarg[2] ); break; case SYS_setsockopt: { socklen_t optlen = uarg[4]; AARG(const void *, optval, 3, optlen); ret = usys_setsockopt( &err, u, uarg[0], uarg[1], uarg[2], optval, optlen); break; } case SYS_getsockopt: { AARG(socklen_t *, optlen, 4, sizeof(socklen_t)); AARG(void *, optval, 3, *optlen); ret = usys_getsockopt( &err, u, uarg[0], uarg[1], uarg[2], optval, optlen); break; } case SYS_getpeername: { AARG(socklen_t *, namelen, 2, sizeof(socklen_t)); AARG(struct sockaddr *, name, 1, *namelen); ret = usys_getpeername( &err, u, uarg[0], name, namelen); break; } case SYS_getsockname: { AARG(socklen_t *, namelen, 2, sizeof(socklen_t)); AARG(struct sockaddr *, name, 1, *namelen); ret = usys_getsockname( &err, u, uarg[0], name, namelen); break; } case SYS_bind: { AARG(const struct sockaddr *, addr, 1, sizeof(struct sockaddr)); ret = usys_bind( &err, u, uarg[0], addr, uarg[2] ); break; } case SYS_listen: ret = usys_listen( &err, u, uarg[0], uarg[1] ); break; case SYS_accept: { AARG( socklen_t *, len, 2, sizeof(socklen_t)); AARG( struct sockaddr *, acc_addr, 1, *len ); ret = usys_accept( &err, u, uarg[0], acc_addr, len ); break; } case SYS_recv: { int len = uarg[2]; AARG( void *, buf, 0, len ); ret = usys_recv( &err, u, uarg[0], buf, len, uarg[3] ); break; } case SYS_recvmsg: { AARG( struct msghdr *, msg, 0, sizeof(struct msghdr) ); ret = usys_recvmsg( &err, u, uarg[0], msg, uarg[2] ); break; } case SYS_send: { int len = uarg[2]; AARG( const void *, buf, 0, len ); ret = usys_send( &err, u, uarg[0], buf, len, uarg[3] ); break; } case SYS_sendmsg: { AARG( const struct msghdr *, msg, 0, sizeof(struct msghdr) ); ret = usys_sendmsg( &err, u, uarg[0], msg, uarg[2] ); break; } case SYS_sendto: { socklen_t tolen = uarg[5]; AARG( const struct sockaddr *, to, 4, tolen ); int len = uarg[2]; AARG( const void *, buf, 0, len ); ret = usys_sendto( &err, u, uarg[0], buf, len, uarg[3], to, tolen ); break; } case SYS_recvfrom: { AARG( socklen_t *, fromlen, 5, sizeof(socklen_t) ); AARG( struct sockaddr *, from, 4, *fromlen ); int len = uarg[2]; AARG( void *, buf, 0, len ); ret = usys_recvfrom( &err, u, uarg[0], buf, len, uarg[3], from, fromlen ); break; } case SYS_connect: // int connect(int socket, const struct sockaddr *address, socklen_t address_len); { int len = uarg[2]; AARG( struct sockaddr *, acc_addr, 1, len ); ret = usys_connect( &err, u, uarg[0], acc_addr, len ); break; } case SYS_socketpair: { AARG( int *, sv, 3, sizeof(int) * 2 ); ret = usys_socketpair( &err, u, uarg[0], uarg[1], uarg[2], sv ); break; } case SYS_sendfile: case SYS_socketcall: goto unimpl; case SYS_nice: { // int nice = uarg[0]; // set thr prio // break; goto unimpl; } case SYS_brk: goto unimpl; /*{ ret = usys_sbrk( &err, u, uarg[0] ); break; }*/ case SYS_fork: case SYS_vfork: goto unimpl; case SYS_ioctl: { void *data = adjustin( uarg[2], st ); CHECKA(data,0); ret = usys_ioctl( &err, u, uarg[0], uarg[1], data, uarg[3] ); break; } int statlink; case SYS_lstat: statlink = 1; goto dostat; case SYS_stat: statlink = 0; goto dostat; dostat: { const char *path = adjustin( uarg[0], st ); CHECKA(path,0); struct stat *data = adjustin( uarg[1], st ); CHECKA(data,sizeof(struct stat)); ret = usys_stat( &err, u, path, data, statlink ); break; } case SYS_fstat: { struct stat *data = adjustin( uarg[1], st ); CHECKA(data,sizeof(struct stat)); ret = usys_fstat( &err, u, uarg[0], data, 0 ); break; } case SYS_umask: { ret = u->umask; u->umask = uarg[0]; break; } case SYS_kill: { ret = usys_kill( &err, u, uarg[0], uarg[1] ); break; } case SYS_waitpid: { AARG(int *, addr, 1, sizeof(int)); ret = usys_waitpid( &err, u, uarg[0], addr, uarg[2] ); break; } case SYS_wait: { AARG(int *, addr, 0, sizeof(int)); ret = usys_waitpid( &err, u, -1, addr, 0 ); break; } case SYS_clone: goto unimpl; case SYS_madvise: case SYS_mincore: case SYS_mlock: case SYS_mlockall: case SYS_mmap: case SYS_mprotect: case SYS_mremap: case SYS_msync: case SYS_munlock: case SYS_munlockall: case SYS_munmap: goto unimpl; // NewOS/BeOS/Haiku case SYS_create_port: { // TODO check string length and final addr to be valid AARG(const char *, name, 1, 0); ret = port_create( uarg[0], name ); break; } case SYS_close_port: ret = port_close( uarg[0] ); break; case SYS_delete_port: ret = port_delete( uarg[0] ); break; case SYS_find_port: { // TODO check string length and final addr to be valid AARG(const char *, name, 0, 0); ret = port_find(name); break; } case SYS_get_port_info: { AARG(struct port_info *, info, 1, sizeof(struct port_info)); ret = port_get_info( uarg[0], info); } case SYS_get_port_bufsize: ret = port_buffer_size( uarg[0] ); break; case SYS_get_port_bufsize_etc: ret = port_buffer_size_etc( uarg[0], uarg[1], uarg[2] ); break; case SYS_get_port_count: ret = port_count( uarg[0] ); break; case SYS_read_port: { AARG(int32_t *, msg_code, 1, sizeof(int32_t)); AARG(void *, msg_buffer, 2, uarg[3]); ret = port_read( uarg[0], msg_code, msg_buffer, uarg[3]); break; } case SYS_write_port: { //AARG(int32_t *, msg_code, 1, sizeof(int32_t)); //AARG(int32_t, msg_code, 1, sizeof(int32_t)); AARG(void *, msg_buffer, 2, uarg[3]); ret = port_write( uarg[0], uarg[1], msg_buffer, uarg[3]); break; } case SYS_read_port_etc: case SYS_write_port_etc: case SYS_set_port_owner: case SYS_get_next_port_info: goto unimpl; case SYS_phantom_run: { // extern int phantom_run(const char *fname, const char **argv, const char **envp, int flags); AARG(const char *, fname, 0, 1); AARG(const char **, uav, 1, 4); AARG(const char **, uep, 2, 4); if( 0 == uarg[1] ) uav = 0; if( 0 == uarg[2] ) uep = 0; SHOW_FLOW( 2, "run %s flags 0x%b", fname, uarg[3], "\020\1<WAIT>\2<NEWWIN>\3<NEWPGRP>" ); char *a[1024]; char *e[1024]; SHOW_FLOW( 2, "run %s load args", fname ); if( user_args_load( mina, maxa, a, 1024, uav ) || user_args_load( mina, maxa, e, 1024, uep ) ) { ret = -1; err = EFAULT; SHOW_ERROR( 0, "fault reading args for %s", fname ); goto err_ret; } ret = usys_run( &err, u, fname, (const char**)a, (const char**)e, uarg[3] ); break; } case SYS_phantom_method: // AARG(const char *, m_name, 0, 1); // int nfd = aarg[1]; // AARG(int *, fds, 0, sizeof(int)*nfd); // ret = usys_pmethod( &err, u, m_name, int nfd, int fds[] ); case SYS_phantom_toobject: case SYS_phantom_fromobject: case SYS_phantom_intmethod: case SYS_phantom_strmethod: goto unimpl; // extern int phantom_runclass(const char *cname, int nmethod, int flags); case SYS_phantom_runclass: { AARG(const char *, cname, 0, 1); unsigned flags = uarg[2]; if(flags) SHOW_ERROR( 0, "SYS_phantom_runclass: unknown flags %x" , flags ); usys_phantom_runclass( &err, u, cname, uarg[1] ); ret = err; } break; case SYS_setproperty: { AARG(const char *, pName, 1, 1); // TODO check zero term string! AARG(const char *, pValue, 2, 1); usys_setproperty( &err, u, uarg[0], pName, pValue ); ret = err; } break; case SYS_getproperty: { AARG(const char *, pName, 1, 1); // TODO check zero term string! AARG(char *, pValue, 2, uarg[3]); usys_getproperty( &err, u, uarg[0], pName, pValue, uarg[3] ); ret = err; } break; case SYS_listproperties: { AARG(char *, buf, 2, uarg[3]); usys_listproperties( &err, u, uarg[0], uarg[1], buf, uarg[3] ); ret = err; } break; case SYS_name2ip: { AARG(in_addr_t *, out, 0, sizeof(in_addr_t)); AARG(const char *, name, 1, 1); ret = name2ip( out, name, uarg[2] ); if( ret != 0 ) err = ret; } break; case SYS_sigpending: { AARG(sigset_t *, set, 0, sizeof(sigset_t *)); ret = usys_sigpending( &err, u, set); break; } case SYS_signal: { #if 0 AARG(sighandler_t, hand, 1, sizeof(sighandler_t)); hand = usys_signal( &err, u, uarg[0], hand); // FIXME 64 bit error ret = ((int)hand) - mina; // Convert pointer back #else // We do not use pointer (ret and uarg 1), so we don't have to convert it ret = (int)usys_signal( &err, u, uarg[0], (void *)uarg[1]); #endif break; } case SYS_sigprocmask: //case raise: case SYS_sigaction: case SYS_siginterrupt: case SYS_sigsuspend: goto unimpl; unimpl: SHOW_ERROR( 0, "Unimplemented syscall %d called", callno ); err = ENOSYS; break; default: SHOW_ERROR( 0, "Unknown syscall %d called", callno ); err = ENOSYS; break; } #else // HAVE_UNIX int err = ENOSYS; int ret = -1; goto err_ret; // to clear warning #endif // HAVE_UNIX err_ret: #ifdef ARCH_ia32 #define _RET_OK st->eax = ret; st->edx = err; #endif #ifdef ARCH_mips #define _RET_OK st->r2 = ret; // v0 (normal ret register, low) st->r3 = err; // v1 (normal ret register, hi) #endif #ifdef ARCH_arm #define _RET_OK st->r0 = ret; // normal ret register st->r1 = err; // arg1 reg, we can safely use for return of errno #endif #ifndef _RET_OK #error arch ret #endif }
bool LinkReceiver::HasMessages() const { return fDataSize - (fRecvStart + fReplySize) > 0 || port_count(fReceivePort) > 0; }