freebsd-skq/sys/dev/uart/uart_cpu_sparc64.c

308 lines
9.6 KiB
C
Raw Normal View History

/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2003, 2004 Marcel Moolenaar
* Copyright (c) 2004 - 2006 Marius Strobl <marius@FreeBSD.org>
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/systm.h>
#include <machine/bus.h>
#include <machine/bus_private.h>
#include <dev/ofw/openfirm.h>
#include <machine/ofw_machdep.h>
#include <dev/uart/uart.h>
#include <dev/uart/uart_cpu.h>
bus_space_tag_t uart_bus_space_io;
bus_space_tag_t uart_bus_space_mem;
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
static struct bus_space_tag bst_store[3];
/*
* Determine which channel of a SCC a device referenced by a full device
* path or as an alias is (in the latter case we try to look up the device
* path via the /aliases node).
* Only the device paths of devices which are used for TTYs really allow
* to do this as they look like these (taken from /aliases nodes):
* ttya: '/central/fhc/zs@0,902000:a'
* ttyc: '/pci@1f,0/pci@1,1/ebus@1/se@14,400000:a'
* Additionally, for device paths of SCCs which are connected to a RSC
* (Remote System Control) device we can hardcode the appropriate channel.
* Such device paths look like these:
* rsc: '/pci@1f,4000/ebus@1/se@14,200000:ssp'
* ttyc: '/pci@1f,4000/ebus@1/se@14,200000:ssp'
*/
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
static int
uart_cpu_channel(char *dev)
{
char alias[64];
phandle_t aliases;
int len;
const char *p;
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
strcpy(alias, dev);
if ((aliases = OF_finddevice("/aliases")) != -1)
(void)OF_getprop(aliases, dev, alias, sizeof(alias));
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
len = strlen(alias);
if ((p = strrchr(alias, ':')) == NULL)
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
return (0);
p++;
if (p - alias == len - 1 && (*p == 'a' || *p == 'b'))
return (*p - 'a' + 1);
if (strcmp(p, "ssp") == 0)
return (1);
return (0);
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
}
int
uart_cpu_eqres(struct uart_bas *b1, struct uart_bas *b2)
{
return ((b1->bsh == b2->bsh) ? 1 : 0);
}
/*
* Get the package handle of the UART that is selected as the console, if
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
* the console is an UART of course. Note that we enforce that both input
* and output are selected.
* Note that the currently active console (i.e. /chosen/stdout and
* /chosen/stdin) may not be the same as the device selected in the
* environment (ie /options/output-device and /options/input-device) because
* keyboard and screen were selected but the keyboard was unplugged or the
* user has changed the environment. In the latter case I would assume that
* the user expects that FreeBSD uses the new console setting.
* For weirder configurations, use ofw_console(4).
*/
static phandle_t
uart_cpu_getdev_console(phandle_t options, char *dev, size_t devsz)
{
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
char buf[sizeof("serial")];
ihandle_t inst;
phandle_t chosen, input, output;
if (OF_getprop(options, "input-device", dev, devsz) == -1)
return (-1);
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
input = OF_finddevice(dev);
if (OF_getprop(options, "output-device", dev, devsz) == -1)
return (-1);
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
output = OF_finddevice(dev);
if (input == -1 || output == -1 ||
OF_getproplen(input, "keyboard") >= 0) {
if ((chosen = OF_finddevice("/chosen")) == -1)
return (-1);
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
if (OF_getprop(chosen, "stdin", &inst, sizeof(inst)) == -1)
return (-1);
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
if ((input = OF_instance_to_package(inst)) == -1)
return (-1);
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
if (OF_getprop(chosen, "stdout", &inst, sizeof(inst)) == -1)
return (-1);
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
if ((output = OF_instance_to_package(inst)) == -1)
return (-1);
snprintf(dev, devsz, "ttya");
}
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
if (input != output)
return (-1);
if (OF_getprop(input, "device_type", buf, sizeof(buf)) == -1)
return (-1);
if (strcmp(buf, "serial") != 0)
return (-1);
/* For a Serengeti console device point to the bootbus controller. */
if (OF_getprop(input, "name", buf, sizeof(buf)) > 0 &&
!strcmp(buf, "sgcn")) {
if ((chosen = OF_finddevice("/chosen")) == -1)
return (-1);
if (OF_getprop(chosen, "iosram", &input, sizeof(input)) == -1)
return (-1);
}
return (input);
}
/*
* Get the package handle of the UART that's selected as the debug port.
* Since there's no place for this in the OF, we use the kernel environment
* variable "hw.uart.dbgport". Note however that the variable is not a
* list of attributes. It's single device name or alias, as known by
* the OF.
*/
static phandle_t
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
uart_cpu_getdev_dbgport(char *dev, size_t devsz)
{
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
char buf[sizeof("serial")];
phandle_t input;
if (!getenv_string("hw.uart.dbgport", dev, devsz))
return (-1);
if ((input = OF_finddevice(dev)) == -1)
return (-1);
if (OF_getprop(input, "device_type", buf, sizeof(buf)) == -1)
return (-1);
if (strcmp(buf, "serial") != 0)
return (-1);
return (input);
}
/*
* Get the package handle of the UART that is selected as the keyboard port,
* if it's actually used to connect the keyboard according to the OF. I.e.
* this will return the UART used to connect the keyboard regardless whether
* it's stdin or not, however not in case the user or the OF gave preference
* to e.g. a PS/2 keyboard by setting /aliases/keyboard accordingly.
*/
static phandle_t
uart_cpu_getdev_keyboard(char *dev, size_t devsz)
{
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
char buf[sizeof("serial")];
phandle_t input;
if ((input = OF_finddevice("keyboard")) == -1)
return (-1);
if (OF_getprop(input, "device_type", buf, sizeof(buf)) == -1)
return (-1);
if (strcmp(buf, "serial") != 0)
return (-1);
if (OF_getprop(input, "name", dev, devsz) == -1)
return (-1);
/*
* So far this also matched PS/2 keyboard nodes so make sure it's
* one of the SCCs/UARTs known to be used to connect keyboards.
*/
if (strcmp(dev, "su") && strcmp(dev, "su_pnp") && strcmp(dev, "zs"))
return (-1);
return (input);
}
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
int
uart_cpu_getdev(int devtype, struct uart_devinfo *di)
{
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
char buf[32], compat[32], dev[64];
struct uart_class *class;
phandle_t input, options;
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
bus_addr_t addr;
int baud, bits, error, range, space, stop;
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
char flag, par;
if ((options = OF_finddevice("/options")) == -1)
return (ENXIO);
switch (devtype) {
case UART_DEV_CONSOLE:
input = uart_cpu_getdev_console(options, dev, sizeof(dev));
break;
case UART_DEV_DBGPORT:
- Change the code that determines whether to use a serial console and which serial device to use in that case respectively to not rely on the OFW names of the input/output and stdin/stdout devices. Instead check whether input and output refers to the same device and is of type serial (uart(4) was already doing this) and for the fallback to a serial console in case a keyboard is the selected input device but unplugged do the same for stdin and stdout in case the input device is nonexistent (PS/2 and USB keyboards) or has a 'keyboard' property (RS232 keyboards). Additionally also check whether the OFW did a fallback to a serial console in the same way in case the output device is nonexistent. While at it save on some variables and for sys/boot/sparc64/loader/metadata.c move the code in question to a new function md_bootserial() so it can be kept in sync with uart_cpu_getdev_console() more easily. This fixes selecting a serial console and the appropriate device when using a device path for the 'input-device' and 'output-device' OFW environment variables instead of an alias for the serial device to use or when using a screen alias that additionally denotes a video mode (like e.g. 'screen:r1024x768x60') but no keyboard is plugged in (amongst others). It also makes the code select a serial console in case the OFW did the same due to a misconfiguration like both 'input-device' and 'output-device' set to 'keyboard' or to a nonexisting device (whether the OFW does a fallback to a serial console in case of a misconfiguration or one ends up with just no console at all highly depends on the OBP version however). - Reduce the size of buffers that only ever need to hold the string 'serial' accordingly. Double the size of buffers that may need to hold a device path as e.g. '/pci@8,700000/ebus@5/serial@1,400000:a' exceeds 32 chars. - Remove the package handle of the '/options' node from the argument list of uart_cpu_getdev_dbgport() as it's unused there and future use is also unlikely. MFC after: 1 week
2005-08-15 20:58:36 +00:00
input = uart_cpu_getdev_dbgport(dev, sizeof(dev));
break;
case UART_DEV_KEYBOARD:
input = uart_cpu_getdev_keyboard(dev, sizeof(dev));
break;
default:
input = -1;
break;
}
if (input == -1)
return (ENXIO);
error = OF_decode_addr(input, 0, &space, &addr);
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
if (error)
return (error);
/* Get the device class. */
if (OF_getprop(input, "name", buf, sizeof(buf)) == -1)
return (ENXIO);
if (OF_getprop(input, "compatible", compat, sizeof(compat)) == -1)
compat[0] = '\0';
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
di->bas.regshft = 0;
di->bas.rclk = 0;
class = NULL;
if (!strcmp(buf, "se") || !strcmp(buf, "FJSV,se") ||
!strcmp(compat, "sab82532")) {
class = &uart_sab82532_class;
/* SAB82532 are only known to be used for TTYs. */
if ((di->bas.chan = uart_cpu_channel(dev)) == 0)
return (ENXIO);
addr += uart_getrange(class) * (di->bas.chan - 1);
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
} else if (!strcmp(buf, "zs")) {
class = &uart_z8530_class;
if ((di->bas.chan = uart_cpu_channel(dev)) == 0) {
/*
* There's no way to determine from OF which
* channel has the keyboard. Should always be
* on channel 1 however.
*/
if (devtype == UART_DEV_KEYBOARD)
di->bas.chan = 1;
else
return (ENXIO);
}
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
di->bas.regshft = 1;
range = uart_getrange(class) << di->bas.regshft;
addr += range - range * (di->bas.chan - 1);
} else if (!strcmp(buf, "lom-console") || !strcmp(buf, "su") ||
!strcmp(buf, "su_pnp") || !strcmp(compat, "rsc-console") ||
!strcmp(compat, "su") || !strcmp(compat, "su16550") ||
!strcmp(compat, "su16552")) {
class = &uart_ns8250_class;
di->bas.chan = 0;
} else if (!strcmp(compat, "sgsbbc")) {
class = &uart_sbbc_class;
di->bas.chan = 0;
}
if (class == NULL)
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
return (ENXIO);
/* Fill in the device info. */
di->ops = uart_getops(class);
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
di->bas.bst = &bst_store[devtype];
di->bas.bsh = sparc64_fake_bustag(space, addr, di->bas.bst);
/* Get the line settings. */
if (devtype == UART_DEV_KEYBOARD)
di->baudrate = 1200;
else if (!strcmp(compat, "rsc-console"))
di->baudrate = 115200;
else
di->baudrate = 9600;
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
di->databits = 8;
di->stopbits = 1;
di->parity = UART_PARITY_NONE;
snprintf(buf, sizeof(buf), "%s-mode", dev);
if (OF_getprop(options, buf, buf, sizeof(buf)) == -1 &&
OF_getprop(input, "ssp-console-modes", buf, sizeof(buf)) == -1)
The uart(4) driver is an universal driver for various UART hardware. It improves on sio(4) in the following areas: o Fully newbusified to allow for memory mapped I/O. This is a must for ia64 and sparc64, o Machine dependent code to take full advantage of machine and firm- ware specific ways to define serial consoles and/or debug ports. o Hardware abstraction layer to allow the driver to be used with various UARTs, such as the well-known ns8250 family of UARTs, the Siemens sab82532 or the Zilog Z8530. This is especially important for pc98 and sparc64 where it's common to have different UARTs, o The notion of system devices to unkludge low-level consoles and remote gdb ports and provides the mechanics necessary to support the keyboard on sparc64 (which is UART based). o The notion of a kernel interface so that a UART can be tied to something other than the well-known TTY interface. This is needed on sparc64 to present the user with a device and ioctl handling suitable for a keyboard, but also allows us to cleanly hide an UART when used as a debug port. Following is a list of features and bugs/flaws specific to the ns8250 family of UARTs as compared to their support in sio(4): o The uart(4) driver determines the FIFO size and automaticly takes advantages of larger FIFOs and/or additional features. Note that since I don't have sufficient access to 16[679]5x UARTs, hardware flow control has not been enabled. This is almost trivial to do, provided one can test. The downside of this is that broken UARTs are more likely to not work correctly with uart(4). The need for tunables or knobs may be large enough to warrant their creation. o The uart(4) driver does not share the same bumpy history as sio(4) and will therefore not provide the necessary hooks, tweaks, quirks or work-arounds to deal with once common hardware. To that extend, uart(4) supports a subset of the UARTs that sio(4) supports. The question before us is whether the subset is sufficient for current hardware. o There is no support for multiport UARTs in uart(4). The decision behind this is that uart(4) deals with one EIA RS232-C interface. Packaging of multiple interfaces in a single chip or on a single expansion board is beyond the scope of uart(4) and is now mostly left for puc(4) to deal with. Lack of hardware made it impossible to actually implement such a dependency other than is present for the dual channel SAB82532 and Z8350 SCCs. The current list of missing features is: o No configuration capabilities. A set of tunables and sysctls is being worked out. There are likely not going to be any or much compile-time knobs. Such configuration does not fit well with current hardware. o No support for the PPS API. This is partly dependent on the ability to configure uart(4) and partly dependent on having sufficient information to implement it properly. As usual, the manpage is present but lacks the attention the software has gotten.
2003-09-06 23:13:47 +00:00
return (0);
if (sscanf(buf, "%d,%d,%c,%d,%c", &baud, &bits, &par, &stop, &flag)
!= 5)
return (0);
di->baudrate = baud;
di->databits = bits;
di->stopbits = stop;
di->parity = (par == 'n') ? UART_PARITY_NONE :
(par == 'o') ? UART_PARITY_ODD : UART_PARITY_EVEN;
return (0);
}