Lines Matching defs:to
44 one machine. You'll have to increase the number of boards in sx.h
97 - Set sx_poll to 1 to poll every timer tick (10ms on Intel).
100 - set sx_slowpoll to 100 to do an extra poll once a second (on Intel). If
101 the driver misses an interrupt (report this if it DOES happen to you!)
102 everything will continue to work....
140 /* Set the mask to all-ones. This alas, only supports 32 interrupts.
168 features enabled, and disable them "runtime". That allows me to
169 instruct people with problems to enable debugging without requiring
170 them to recompile...
228 the card some time to breathe between accesses. (Otherwise the
229 processor on the card might not be able to access its OWN bus... */
357 printk(KERN_INFO "sx: Card doesn't respond to "
380 /* If you have one of those, someone will need to write
381 the equivalent of this, which will amount to about 3 lines. I don't
382 want to complicate this right now. -- REW
448 /* Note. The SX register is write-only. Therefore, we have to enable the
652 /* The baud rate is not set to 0, so we're enabeling DTR... -- REW */
729 Ha! figured it out. It is to allow you to get DTR active again
730 if you've dropped it with stty 0. Moved to set_baud, where it
731 belongs (next to the drop dtr if baud == 0) -- REW */
816 Other drivers use the macro "MIN" to calculate how much to copy.
821 I assign a value, and then only allow the value to decrease. This
846 /* Took me 5 minutes to deduce this formula.
885 It would allow others to start putting more data into the
949 It allows the card to start putting more data into the
1046 /* AAargh! The order in which to do these things is essential and
1068 actual interrupt routine!. (Well, used to when I wrote that
1157 /* Ehhm. I don't know how to fiddle with interrupts on the SX card. --REW */
1176 /* First transmit the characters that we're supposed to */
1301 printk(KERN_ERR "sx: Card didn't respond to LOPEN "
1334 /* Give the port 5 seconds to close down. */
1335 int to = 5 * HZ;
1343 while (to-- && (sx_read_channel_byte(port, hi_hstat) != HS_IDLE_CLOSED))
1357 5 * HZ - to - 1, port->gs.count);
1362 /*printk("%s SETTING port count to zero: %p count: %d\n",
1467 sx_dprintk(SX_DEBUG_FIRMWARE, "set board to %ld\n", arg);
1477 rc = -ENOENT; /* If we manage to miss one, return error. */
1667 * and CTS/RTS flow control itself. This means that all we have to
1668 * do when signalled by the upper tty layer to throttle/unthrottle is
1669 * to make a note of it here. When we come to read characters from the
1670 * rx buffers on the card (sx_receive_chars()) we look to see if the
1728 /* This resets the processor again, to make sure it didn't do any
1899 sx_dprintk(SX_DEBUG_PROBE, "Going to verify vpd prom at %p.\n",
1966 /* Specialix probes for this card at 32k increments from 640k to 16M.
1977 sx_dprintk(SX_DEBUG_PROBE, "Going to verify SI signature hw %lx at "
1999 but to prevent trouble, we'd better double check that we don't
2020 but to prevent trouble, we'd better double check that we don't
2043 /* Compared to the SX boards, it is a complete guess as to what
2044 this card is up to... */
2122 There is no reason not to allocate them dynamically.
2173 efficient. Neil tells me it isn't going to
2189 /* This has to be done earlier. */
2226 /* It is safe/allowed to del_timer a non-active timer */
2338 * This is to prevent the card locking up on Intel Xeon *
2489 printk(KERN_ERR "SX: Unable to register firmware loader "