--- a/hurd/hurdselect.c 2013-01-08 09:42:54.000000000 +0100 +++ b/hurd/hurdselect.c 2013-02-10 14:51:32.000000000 +0100 @@ -33,6 +33,279 @@ from SELECT_ALL (which better not have the high bit set). */ #define SELECT_RETURNED ((SELECT_ALL << 1) & ~SELECT_ALL) +struct dfd +{ + struct hurd_userlink ulink; + struct hurd_fd *cell; + mach_port_t io_port; + int type; + mach_port_t reply_port; +}; + +/* Helper functions */ +int _io_select_request (int ispoll, int firstfd, int lastfd, + error_t *errvec, mach_port_t *portset, struct dfd *d) +{ + int i, got = 0; + + if (firstfd == -1) /* DELAY */ + /* But not if there were no ports to deal with at all. + We are just a pure timeout. */ + *portset = __mach_reply_port (); + else /* POLL || SELECT */ + { + *portset = MACH_PORT_NULL; + + for (i = firstfd; i <= lastfd; ++i) + { + int err = 0, err1 = 0; + int type = d[i].type; + if (type) + { + d[i].reply_port = __mach_reply_port (); + /* FIXME: Needed for poll to account for the round trip delay */ + if (ispoll) + err = __io_select (d[i].io_port, d[i].reply_port, 1, &type); + else + err = __io_select (d[i].io_port, d[i].reply_port, 0, &type); + switch (err) + { + case MACH_RCV_TIMED_OUT: + /* No immediate response. This is normal. */ + err = 0; + if (firstfd == lastfd) + /* When there's a single descriptor, we don't need a + portset, so just pretend we have one, but really + use the single reply port. */ + *portset = d[i].reply_port; + else if (got == 0) + /* We've got multiple reply ports, so we need a port set to + multiplex them. */ + { + /* We will wait again for a reply later. */ + if (*portset == MACH_PORT_NULL) + /* Create the portset to receive all the replies on. */ + err1 = __mach_port_allocate (__mach_task_self (), + MACH_PORT_RIGHT_PORT_SET, + portset); + if (! err1) + /* Put this reply port in the port set. */ + __mach_port_move_member (__mach_task_self (), + d[i].reply_port, *portset); + } + break; + + /* FIXME: Only for poll */ + if (ispoll) + { + case EPIPE: + case EIO: + errvec[i] = err; + //errno = ENOTCONN; + break; + } + + default: + /* No other error should happen. Callers of select + don't expect to see errors, so we simulate + readiness of the erring object and the next call + hopefully will get the error again. */ + type = SELECT_ALL; + /* FALLTHROUGH */ + + case 0: + /* We got an answer. */ + if ((type & SELECT_ALL) == 0) + /* Bogus answer; treat like an error, as a fake positive. */ + type = SELECT_ALL; + + /* This port is already ready already. */ + d[i].type &= type; + d[i].type |= SELECT_RETURNED; + ++got; + break; + } /* switch (err) */ + _hurd_port_free (&d[i].cell->port, &d[i].ulink, d[i].io_port); + + /* FIXME: Handle only errors EPIPE (pflocal), EIO (pfinet) for POLL*/ + if ((ispoll && (err != 0 && err != EPIPE && err != EIO)) || + (!ispoll && err != 0)) + //if ((ispoll && (err != 0 && err != ENOTCONN)) || (!ispoll && err != 0)) + //if ((err != 0) && ((ispoll && err != ENOTCONN) || !ispoll)) + { + errno = err; + errvec[i] = err; + got = -1; + return got; + } + } /* if (type) */ + } /* for */ + } /* else */ + return got; +} /* _io_select_request() */ + +int _wait_for_replies (int nfds, int firstfd, int lastfd, int got, + mach_msg_timeout_t to, + error_t err, mach_port_t portset, struct dfd *d, + const struct timespec *timeout, const sigset_t *sigmask, + sigset_t *oset) +{ + int i; + union typeword /* Use this to avoid unkosher casts. */ + { + mach_msg_type_t type; + uint32_t word; + }; + assert (sizeof (union typeword) == sizeof (mach_msg_type_t)); + assert (sizeof (uint32_t) == sizeof (mach_msg_type_t)); + + union + { + mach_msg_header_t head; +#ifdef MACH_MSG_TRAILER_MINIMUM_SIZE + struct + { + mach_msg_header_t head; + NDR_record_t ndr; + error_t err; + } error; + struct + { + mach_msg_header_t head; + NDR_record_t ndr; + error_t err; + int result; + mach_msg_trailer_t trailer; + } success; +#else + struct + { + mach_msg_header_t head; + union typeword err_type; + error_t err; + } error; + struct + { + mach_msg_header_t head; + union typeword err_type; + error_t err; + union typeword result_type; + int result; + } success; +#endif + } msg; + mach_msg_option_t options = (timeout == NULL ? 0 : MACH_RCV_TIMEOUT); + error_t msgerr; + +#define IO_SELECT_REPLY_MSGID (21012 + 100) /* XXX */ +#ifdef MACH_MSG_TYPE_BIT + const union typeword inttype = + { type: + { MACH_MSG_TYPE_INTEGER_T, sizeof (integer_t) * 8, 1, 1, 0, 0 } + }; +#endif + + /* Now wait for reply messages. */ + if (!err && got == 0) + { + /* Now wait for io_select_reply messages on PORT, + timing out as appropriate. */ + while ((msgerr = __mach_msg (&msg.head, + MACH_RCV_MSG | MACH_RCV_INTERRUPT | options, + 0, sizeof msg, portset, to, + MACH_PORT_NULL)) == MACH_MSG_SUCCESS) + { + /* We got a message. Decode it. */ + if (msg.head.msgh_id == IO_SELECT_REPLY_MSGID && + msg.head.msgh_size >= sizeof msg.error && + !(msg.head.msgh_bits & MACH_MSGH_BITS_COMPLEX) && +#ifdef MACH_MSG_TYPE_BIT + msg.error.err_type.word == inttype.word +#endif + ) + { + /* This is a properly formatted message so far. + See if it is a success or a failure. */ + if (msg.error.err == EINTR && + msg.head.msgh_size == sizeof msg.error) + { + /* EINTR response; poll for further responses + and then return quickly. */ + err = EINTR; + goto poll; + } + if (msg.error.err || + msg.head.msgh_size != sizeof msg.success || +#ifdef MACH_MSG_TYPE_BIT + msg.success.result_type.word != inttype.word || +#endif + (msg.success.result & SELECT_ALL) == 0) + { + /* Error or bogus reply. Simulate readiness. */ + __mach_msg_destroy (&msg.head); + msg.success.result = SELECT_ALL; + } + + /* Look up the respondent's reply port and record its + readiness. */ + { + int had = got; + if (firstfd != -1) + for (i = firstfd; i <= lastfd; ++i) + if (d[i].type + && d[i].reply_port == msg.head.msgh_local_port) + { + d[i].type &= msg.success.result; + d[i].type |= SELECT_RETURNED; + ++got; + } + assert (got > had); + } + } /* if (decode message) */ + + if (msg.head.msgh_remote_port != MACH_PORT_NULL) + __mach_port_deallocate (__mach_task_self (), + msg.head.msgh_remote_port); + + if (got) + poll: + { + /* Poll for another message. */ + to = 0; + options |= MACH_RCV_TIMEOUT; + } + } /* while */ + + if (msgerr == MACH_RCV_INTERRUPTED) + /* Interruption on our side (e.g. signal reception). */ + err = EINTR; + + if (got) + /* At least one descriptor is known to be ready now, so we will + return success. */ + err = 0; + } /* if (!err && got == 0) */ + + if (firstfd != -1) + for (i = firstfd; i <= lastfd; ++i) + if (d[i].type) + __mach_port_destroy (__mach_task_self (), d[i].reply_port); + if (firstfd == -1 || (firstfd != lastfd && portset != MACH_PORT_NULL)) + /* Destroy PORTSET, but only if it's not actually the reply port for a + single descriptor (in which case it's destroyed in the previous loop; + not doing it here is just a bit more efficient). */ + __mach_port_destroy (__mach_task_self (), portset); + + if (err) + { + if (sigmask) + __sigprocmask (SIG_SETMASK, oset, NULL); + errno = err; + got = -1; + } + return got; + } /* _wait_for_replies() */ + /* Check the first NFDS descriptors either in POLLFDS (if nonnnull) or in each of READFDS, WRITEFDS, EXCEPTFDS that is nonnull. If TIMEOUT is not NULL, time out after waiting the interval specified therein. Returns @@ -50,25 +323,24 @@ _hurd_select (int nfds, fd_set rfds, wfds, xfds; int firstfd, lastfd; mach_msg_timeout_t to = 0; - struct - { - struct hurd_userlink ulink; - struct hurd_fd *cell; - mach_port_t io_port; - int type; - mach_port_t reply_port; - } d[nfds]; + struct dfd d[nfds]; sigset_t oset; - union typeword /* Use this to avoid unkosher casts. */ - { - mach_msg_type_t type; - uint32_t word; - }; - assert (sizeof (union typeword) == sizeof (mach_msg_type_t)); - assert (sizeof (uint32_t) == sizeof (mach_msg_type_t)); + /* case POLL */ + int j, i_index[nfds]; + error_t errvec[nfds]; + + enum { + SELECT = 0, + POLL = 1 + } ispoll; - if (nfds < 0 || nfds > FD_SETSIZE) + if (pollfds) + ispoll = POLL; + else + ispoll = SELECT; + + if (nfds < 0 || nfds >= FD_SETSIZE) { errno = EINVAL; return -1; @@ -91,33 +363,45 @@ _hurd_select (int nfds, if (sigmask && __sigprocmask (SIG_SETMASK, sigmask, &oset)) return -1; - if (pollfds) + /* Send them all io_select request messages. */ + + err = 0; + got = 0; + + switch (ispoll) { + case POLL: /* Collect interesting descriptors from the user's `pollfd' array. We do a first pass that reads the user's array before taking any locks. The second pass then only touches our own stack, and gets the port references. */ for (i = 0; i < nfds; ++i) - if (pollfds[i].fd >= 0) - { - int type = 0; - if (pollfds[i].events & POLLIN) - type |= SELECT_READ; - if (pollfds[i].events & POLLOUT) - type |= SELECT_WRITE; - if (pollfds[i].events & POLLPRI) - type |= SELECT_URG; + { + errvec[i] = 0; + i_index[i] = -1; - d[i].io_port = pollfds[i].fd; - d[i].type = type; - } - else - d[i].type = 0; + if (pollfds[i].fd >= 0) + { + int type = 0; + if (pollfds[i].events & POLLIN) + type |= SELECT_READ; + if (pollfds[i].events & POLLOUT) + type |= SELECT_WRITE; + if (pollfds[i].events & POLLPRI) + type |= SELECT_URG; + + d[i].io_port = pollfds[i].fd; + d[i].type = type; + } + else + d[i].type = 0; + } HURD_CRITICAL_BEGIN; __mutex_lock (&_hurd_dtable_lock); + j = 0; for (i = 0; i < nfds; ++i) if (d[i].type != 0) { @@ -127,16 +411,17 @@ _hurd_select (int nfds, { d[i].cell = _hurd_dtable[fd]; d[i].io_port = _hurd_port_get (&d[i].cell->port, &d[i].ulink); - if (d[i].io_port != MACH_PORT_NULL) - continue; - } - /* If one descriptor is bogus, we fail completely. */ - while (i-- > 0) - if (d[i].type != 0) - _hurd_port_free (&d[i].cell->port, - &d[i].ulink, d[i].io_port); - break; + /* If one descriptor is bogus, mark and remove it. */ + if (d[i].io_port == MACH_PORT_NULL) + { + _hurd_port_free (&d[i].cell->port, + &d[i].ulink, d[i].io_port); + continue; /* Next i */ + } + i_index[j] = i; + j++; + } } __mutex_unlock (&_hurd_dtable_lock); @@ -150,11 +435,69 @@ _hurd_select (int nfds, return -1; } - lastfd = i - 1; - firstfd = i == 0 ? lastfd : 0; - } - else - { + lastfd = j - 1; + firstfd = (j == 0) ? lastfd : i_index[0]; + + got = _io_select_request (ispoll, firstfd, lastfd, errvec, &portset, d); + if (got == -1) + return -1; + + got = _wait_for_replies (nfds, firstfd, lastfd, got, to, + err, portset, d, + timeout, sigmask, &oset); + if (got == -1) + return -1; + + /* Fill in the `revents' members of the user's array. */ + for (i = 0; i < nfds; ++i) + { + int type = d[i].type; + int_fast16_t revents = 0; + int k = i_index[i]; + /* Clear all revents entries before filling in */ + pollfds[i].revents = 0; + + /* Return revents for good file descriptors */ + if (k == i) + { + if (type & SELECT_RETURNED) + { + if (type & SELECT_READ) + revents |= POLLIN; + if (type & SELECT_WRITE) + revents |= POLLOUT; + if (type & SELECT_URG) + revents |= POLLPRI; + pollfds[i].revents = revents; + } + /* Return revents for broken file descriptors */ + else + { + if (errvec[i] == EPIPE) + { + pollfds[i].revents = POLLHUP; + ++got; + } + if (errvec[i] == EIO) + { + pollfds[i].revents = POLLERR; + ++got; + } + } + } + /* Return revents for bad file descriptors */ + /* FIXME: Can be written as: if (k == -1 && pollfds[i].fd >= 0) */ + if (k == -1 && pollfds[i].fd >= 0) + { + //fprintf(stderr,"POLL: bad FDs i,k=%d,%d\n",i,k); + pollfds[i].revents = POLLNVAL; + ++got; + } + } + return got; + break; + + case SELECT: /* Collect interested descriptors from the user's fd_set arguments. Use local copies so we can't crash from user bogosity. */ @@ -174,13 +517,17 @@ _hurd_select (int nfds, HURD_CRITICAL_BEGIN; __mutex_lock (&_hurd_dtable_lock); + /* FIXME: limit nfds upwards */ if (nfds > _hurd_dtablesize) - nfds = _hurd_dtablesize; + { + nfds = _hurd_dtablesize; + } /* Collect the ports for interesting FDs. */ firstfd = lastfd = -1; for (i = 0; i < nfds; ++i) { + errvec[i] = 0; int type = 0; if (readfds != NULL && FD_ISSET (i, &rfds)) type |= SELECT_READ; @@ -218,250 +565,17 @@ _hurd_select (int nfds, errno = EBADF; return -1; } - } - - - err = 0; - got = 0; - /* Send them all io_select request messages. */ + got = _io_select_request (ispoll, firstfd, lastfd, errvec, &portset, d); + if (got == -1) + return -1; + + got = _wait_for_replies (nfds, firstfd, lastfd, got, to, + err, portset, d, + timeout, sigmask, &oset); + if (got == -1) + return -1; - if (firstfd == -1) - /* But not if there were no ports to deal with at all. - We are just a pure timeout. */ - portset = __mach_reply_port (); - else - { - portset = MACH_PORT_NULL; - - for (i = firstfd; i <= lastfd; ++i) - if (d[i].type) - { - int type = d[i].type; - d[i].reply_port = __mach_reply_port (); - err = __io_select (d[i].io_port, d[i].reply_port, 0, &type); - switch (err) - { - case MACH_RCV_TIMED_OUT: - /* No immediate response. This is normal. */ - err = 0; - if (firstfd == lastfd) - /* When there's a single descriptor, we don't need a - portset, so just pretend we have one, but really - use the single reply port. */ - portset = d[i].reply_port; - else if (got == 0) - /* We've got multiple reply ports, so we need a port set to - multiplex them. */ - { - /* We will wait again for a reply later. */ - if (portset == MACH_PORT_NULL) - /* Create the portset to receive all the replies on. */ - err = __mach_port_allocate (__mach_task_self (), - MACH_PORT_RIGHT_PORT_SET, - &portset); - if (! err) - /* Put this reply port in the port set. */ - __mach_port_move_member (__mach_task_self (), - d[i].reply_port, portset); - } - break; - - default: - /* No other error should happen. Callers of select - don't expect to see errors, so we simulate - readiness of the erring object and the next call - hopefully will get the error again. */ - type = SELECT_ALL; - /* FALLTHROUGH */ - - case 0: - /* We got an answer. */ - if ((type & SELECT_ALL) == 0) - /* Bogus answer; treat like an error, as a fake positive. */ - type = SELECT_ALL; - - /* This port is already ready already. */ - d[i].type &= type; - d[i].type |= SELECT_RETURNED; - ++got; - break; - } - _hurd_port_free (&d[i].cell->port, &d[i].ulink, d[i].io_port); - } - } - - /* Now wait for reply messages. */ - if (!err && got == 0) - { - /* Now wait for io_select_reply messages on PORT, - timing out as appropriate. */ - - union - { - mach_msg_header_t head; -#ifdef MACH_MSG_TRAILER_MINIMUM_SIZE - struct - { - mach_msg_header_t head; - NDR_record_t ndr; - error_t err; - } error; - struct - { - mach_msg_header_t head; - NDR_record_t ndr; - error_t err; - int result; - mach_msg_trailer_t trailer; - } success; -#else - struct - { - mach_msg_header_t head; - union typeword err_type; - error_t err; - } error; - struct - { - mach_msg_header_t head; - union typeword err_type; - error_t err; - union typeword result_type; - int result; - } success; -#endif - } msg; - mach_msg_option_t options = (timeout == NULL ? 0 : MACH_RCV_TIMEOUT); - error_t msgerr; - while ((msgerr = __mach_msg (&msg.head, - MACH_RCV_MSG | MACH_RCV_INTERRUPT | options, - 0, sizeof msg, portset, to, - MACH_PORT_NULL)) == MACH_MSG_SUCCESS) - { - /* We got a message. Decode it. */ -#define IO_SELECT_REPLY_MSGID (21012 + 100) /* XXX */ -#ifdef MACH_MSG_TYPE_BIT - const union typeword inttype = - { type: - { MACH_MSG_TYPE_INTEGER_T, sizeof (integer_t) * 8, 1, 1, 0, 0 } - }; -#endif - if (msg.head.msgh_id == IO_SELECT_REPLY_MSGID && - msg.head.msgh_size >= sizeof msg.error && - !(msg.head.msgh_bits & MACH_MSGH_BITS_COMPLEX) && -#ifdef MACH_MSG_TYPE_BIT - msg.error.err_type.word == inttype.word -#endif - ) - { - /* This is a properly formatted message so far. - See if it is a success or a failure. */ - if (msg.error.err == EINTR && - msg.head.msgh_size == sizeof msg.error) - { - /* EINTR response; poll for further responses - and then return quickly. */ - err = EINTR; - goto poll; - } - if (msg.error.err || - msg.head.msgh_size != sizeof msg.success || -#ifdef MACH_MSG_TYPE_BIT - msg.success.result_type.word != inttype.word || -#endif - (msg.success.result & SELECT_ALL) == 0) - { - /* Error or bogus reply. Simulate readiness. */ - __mach_msg_destroy (&msg.head); - msg.success.result = SELECT_ALL; - } - - /* Look up the respondent's reply port and record its - readiness. */ - { - int had = got; - if (firstfd != -1) - for (i = firstfd; i <= lastfd; ++i) - if (d[i].type - && d[i].reply_port == msg.head.msgh_local_port) - { - d[i].type &= msg.success.result; - d[i].type |= SELECT_RETURNED; - ++got; - } - assert (got > had); - } - } - - if (msg.head.msgh_remote_port != MACH_PORT_NULL) - __mach_port_deallocate (__mach_task_self (), - msg.head.msgh_remote_port); - - if (got) - poll: - { - /* Poll for another message. */ - to = 0; - options |= MACH_RCV_TIMEOUT; - } - } - - if (err == MACH_RCV_TIMED_OUT) - /* This is the normal value for ERR. We might have timed out and - read no messages. Otherwise, after receiving the first message, - we poll for more messages. We receive with a timeout of 0 to - effect a poll, so ERR is MACH_RCV_TIMED_OUT when the poll finds no - message waiting. */ - err = 0; - if (msgerr == MACH_RCV_INTERRUPTED) - /* Interruption on our side (e.g. signal reception). */ - err = EINTR; - - if (got) - /* At least one descriptor is known to be ready now, so we will - return success. */ - err = 0; - } - - if (firstfd != -1) - for (i = firstfd; i <= lastfd; ++i) - if (d[i].type) - __mach_port_destroy (__mach_task_self (), d[i].reply_port); - if (firstfd == -1 || (firstfd != lastfd && portset != MACH_PORT_NULL)) - /* Destroy PORTSET, but only if it's not actually the reply port for a - single descriptor (in which case it's destroyed in the previous loop; - not doing it here is just a bit more efficient). */ - __mach_port_destroy (__mach_task_self (), portset); - - if (err) - { - if (sigmask) - __sigprocmask (SIG_SETMASK, &oset, NULL); - return __hurd_fail (err); - } - - if (pollfds) - /* Fill in the `revents' members of the user's array. */ - for (i = 0; i < nfds; ++i) - { - int type = d[i].type; - int_fast16_t revents = 0; - - if (type & SELECT_RETURNED) - { - if (type & SELECT_READ) - revents |= POLLIN; - if (type & SELECT_WRITE) - revents |= POLLOUT; - if (type & SELECT_URG) - revents |= POLLPRI; - } - - pollfds[i].revents = revents; - } - else - { /* Below we recalculate GOT to include an increment for each operation allowed on each fd. */ got = 0; @@ -489,7 +603,9 @@ _hurd_select (int nfds, else if (exceptfds) FD_CLR (i, exceptfds); } - } + break; + + } /* switch (ispoll) */ if (sigmask && __sigprocmask (SIG_SETMASK, &oset, NULL)) return -1;