remove unneeded includes; ok mpi@

Remove unused device poll functions.

Also remove unneeded includes of <sys/poll.h> and <sys/select.h>.

Some addenda from jsg@.

OK miod@ mpi@

kernel: use gettime(9)/getuptime(9) in lieu of time_second(9)/time_uptime(9)

time_second(9) and time_uptime(9) are widely used in the kernel to
quickly get the system UTC or system uptime as a time_t. However,
time_t is 64-bit everywhere, so it is not generally safe to use them
on 32-bit platforms: you have a split-read problem if your hardware
cannot perform atomic 64-bit reads.

This patch replaces time_second(9) with gettime(9), a safer successor
interface, throughout the kernel. Similarly, time_uptime(9) is replaced
with getuptime(9).

There is a performance cost on 32-bit platforms in exchange for
eliminating the split-read problem: instead of two register reads you
now have a lockless read loop to pull the values from the timehands.
This is really not *too* bad in the grand scheme of things, but
compared to what we were doing before it is several times slower.

There is no performance cost on 64-bit (__LP64__) platforms.

With input from visa@, dlg@, and tedu@.

Several bugs squashed by visa@.

ok kettenis@

udcf(4): tvtohz(9)+timeout_add(9) -> timeout_add_msec(9); ok kettenis@

Delete unnecessary <sys/file.h> includes

ok millert@ krw@

Avoid calling usbd_set_config_index() in *_attach() and let the stack
do it instead.

If anything bad happen due to a malformed descriptor it makes no sense
to try to attach a driver, and bail before probing.

The Swiss Federal Government decided to shut down HBG at the end of 2011.
On 6 September 2012 at 12:02:00 UTC both antenna towers were demolished by
controlled explosives. So this is not coming back and we can tedu the
support for HBG form the DCF77 drivers.
Remided by mbalmer

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

tedu another bunch of DVACT_DEACTIVATE handlers that do nothing but set
the 'dying' flag.

Get rid of various 'typedef struct' definitions and use plain structure
definitions instead. We don't change usb.h for now to stay compatible
with userland.

Tested by mpi@ on macppc and myself on i386.

ok mpi@

proc.h is way too much header for usb to handle.

you probably don't need to include vnode.h or reboot.h. you most definitely
absolutely do not need to include vnode.h because it includes uvm_extern.h
and you want the idiotic TRUE FALSE defines from uvm.

Remove config_activate() and DVACT_ACTIVATE. PCMCIA's the only thing
that's ever used it, and it's long since been changed to use
DVACT_{QUIESCE,SUSPEND,RESUME} instead.

ok deraadt@, dlg@; miod@ also agreed with this idea when I brought it
up a few weeks ago

garbage collect "usb events". without /dev/usb there is no way to access
them from userland, and nothing in the kernel uses them.

ok krw@, miod@

- use usb_lookup() instead of handrolled macro

ok jakemsr@

* replace per-driver dying and/or other state variables with use of
usbd_deactivete() and usbd_is_dying()
* use usbd_deactivate() in activate()/DEACTIVATE
* convert a few more direct checks of the associated bus' dying flag
with usbd_is_dying()

check that a timeout(9) has been initialized before deleting it

ok phessler

create another kthread to run xfer abort tasks. xfer abort tasks cannot
be run from the generic task kthread, because xfers that need to be
aborted block newly queued tasks from running (i.e. the xfer to be aborted
blocks the abort of that task). as there are now three types of usb
tasks, add an argument to usb_init_task() and another member to struct
usb_task to specify the task type.

fixes boot hangs that are showing up because we now use usb tasks to
attach/detach usb devices.

remove dying flag in detach() function.
The dying flag will be set in activate()/DEACTIVATE.

ok deraadt@

Get rid of devact enum, substitute it with an int and coresponding defines.

This is needed for the addition of further suspend/resume actions.

Okay deraadt@, marco@.

in ca_attach() there is no need to set .value and .flags of sensors to 0,
since autoconf(9) allocates softc with M_ZERO; ok deraadt

Kill some whitespace, removed unneeded stuff at the end of a list.

Don't terminate the array of matching product/vendor IDs with { 0, 0 }
when usb_lookup is used. If a device indeed has 0 as vendor and product
ID, like the HP DL165 BMC Server Engines SE USB Device, a wrong driver
will attach and the machine will most pbly crash.

Problem reported Rivo Nurges <rix@estpak.ee>
cvs: ----------------------------------------------------------------------

Use UT_{READ|WRITE}_VENDOR_DEVICE instead of self defined values.

Add support for the new Expert mouseCLOCK USB II that uses an FTDI
FT232R chip to interface the receiver (instead of the NetCologne
chip used on older models). Only the DCF77 and HBG receivers are
supported.

A sidenote: Gude ADS not only provides me with receivers and
documentation, but as a result of my feedback on their older receivers
the new ones now have proper USB product IDs to distinguish between
DCF77, HBG, and MSF.

Make sure that variables used during detach are initialized during attach.
Discussed with dlg.

No need for process context to degrade the sensor status.
Discussed with kettenis.

treat usb vendor/product names as a locator, and have usbd_print handle it,
so that it shows up before the :. as a result, all the usb devices do not
need to have name printing code anymore. all this now works and prints
nicely because usbd_probe_and_attach() is serialized. ok kettenis

Remove the definition and usage of the USB_DECLARE_DRIVER_CLASS and
USB_DECLARE_DRIVER macros.

No binary change.

ok dlg.

Remove the definition and use of the device_ptr_t which was a struct device *.

No binary change.

ok mk.

Remove the definition and use of the USBDEVNAME macro.

(This might look easy, but it was a big diff. Thanks to dlg and especially jsg
for looking over it; we found at least four mistakes in the initial diff.)

ok jsg.

Remove the definition and use of the USBDEV macro. It only created confusion
and the address of it's argument: USBDEV(sc->sc_dev) yields &sc->sc_dev.

No binary changes.

ok jsg.

Mechanical removal of USBBASEDEVICE. No binary change.

Tested by thib and myself.

ok mbalmer jsg

Farewell USB_{ATTACH,MATCH,DETACH}* you will not be missed.
ok deraadt@ krw@ mbalmer@

Back out last. the world is not i386.

Remove unused header files.

split userland & kernel struct sensor/sensordev so that the addition
of new fields in the future is less disruptive. This is done similar
to how struct proc is handled for ps(1). ok jmc (man page changes)
tested fkr simon, and more suggestions from millert

return is not a function.

adapt to new two-level sensor api; Constantine A. Murenin

Make sure the bit collecting is restarted after signal loss.

restart collecting bits immediately after a parity error, i.e. use the frame
immediately following the corrupt one; there is no no need to skip it by
waiting for the next minute gap and the sensor value has valid data one
minute earlier than before.

Re-arm signal loss detection only after we received a bit, not in itself. This
prevents the signal loss function from being called repeatedly when we have
no signal.

Faster synchronization after an unexpected minute gap.
Only look at the skew value if it is != 0.
Remove some dead code.

When compiled with UDCF_DEBUG, only output the bits received when
udcfdebug > 1.

Add code to detect (and eliminate) statistical outliers. Tested by me and
naddy; many thanks to naddy for continous testing and feedback during the
last weeks.

ok naddy

Fix an integer type promotion that can lead to wrong offsets on 64 bit arches.
Problem and solution found by Christian "Naddy" Weisgerber <naddy@openbsd.org>,
thanks!

- Use better names for the variables containing the timeouts in Hz instead of
just t1-t10.
- Remove some unused variables.
- Restructure some functions to remove some levels of indentation.

udcf(4) degraded the sensor status from OK to WARNING on the first receiption
error, which is to fast as this does not give ntpd the opportunity to use the
sensor value.

If we decode a valid time information, it has to be available to applications
for some minimal time. I choose a value of 5 minutes. If we do not receive
any new valid time information during this period, the sensor will be degraded
to WARNING, and after another fifteen minutes it will be further degraded to
CRITICAL.

"makes sense to me" henning@, "sure" deraadt@

Apply some formatting.

Make the timedelta sensor appear right when a device is plugged-in, with a
status of SENSOR_S_UNKNOWN. Do not set the SENSOR_FINVALID bit, sysctl
hw.sensors does not list sensors with this flags set and we want to the sensor
when the device is present.

ok henning

In a drivers activate() entry point, if on DVACT_DEACTIVATE it does
something, then it MUST return zero on DVACT_ACTIVATE, not EOPNOTSUPP;
this very popular bug has been cut and pasted a lot of times...
ok deraadt@ mickey@

remove the no longer needed time calculations now that we have them in
sys/kern/clock_subr.c

move clock_subr.c to a better place, and now it is always in the kernel
so that things can use it; tested on all architectures; ok kettenis

udcf(4) no longer needs the file sys/dev/clock_subr.c to be compiled and linked
to the kernel, the two functions it used from this file are now included in
the driver itself (like in nmea(4)). udcf(4) can now be enabled to the zaurus,
too. in fact on all arches that have USB support.

Typos in comments.

cleanup.

- ntpd no longer needs the NTP identifier as part of a timedelta sensor
description.
- create the timedelta sensors as early as possible, but mark them invalid
as long as there is no real data.
- update docs accordingly

Change the timedelta sensor descritpion so that the first four bytes contain
the NTP identifier (filled with spaces), followed by a space and the station
name (which might be the same as the NTP identifier).

All timedelta sensor have to follow this scheme as OpenNTPD needs it.

Requested by henning.

There is no longer the need to create INVALID sensors on stock, the sensor
framework will create hotplug events for sensor addition/deletion. ntpd
will catch these events.

When the device is attached, immediately create the timedelta sensor with the
SENSOR_FINVALID flag set and without a description. We do not need a dummy
description as a sensor with the FINVALID flags set will not show up or be used
anyways. It's merely an indication to applications that there is a sensor that
will eventually have a correct value.

The SENSOR_FINVALID flag is removed and the description is set to the proper
clockname as soon as we have received the first valid time information.

This enables ntpd to check for hotplug events and rescan the list of timedelta
sensors after a hotplug event.

discussed with deraadt.

cope with recent changes in the sensor framework

- provide the timedelta in nanoseconds
- provide a timestamp in the sensor

support for HBG and DCF77:

- detect clocktype (DCF77 or HBG) and set it once is the sensor description

Simplify time calculations.

help and ok kettenis@

make udcf work as a SENSOR_TIMEDELTA type sensor. The sensor value is the
offset between the time of the device and the localtime in mSec.

the sensor status is set to OK once we received a complete and valid time
information.

help and ok deraadt

open/close/cdevsw code goes bye bye; ok mbalber

- remove the 'needs-flag' from the udcf entry in files.usb and the #ifdef NUDCF
idom in udcf.c
- declare t1 to t7 as static to prevent the risk of name clashes
- add two missing function prototypes

help and ok miod@

Driver to decode the DCF77 time code using a USB attached Gude Expert
mouseCLOCK device; without a mechanism to pass data out.

ok deraadt

Remove unused device poll functions.

Also remove unneeded includes of <sys/poll.h> and <sys/select.h>.

Some addenda from jsg@.

OK miod@ mpi@

kernel: use gettime(9)/getuptime(9) in lieu of time_second(9)/time_uptime(9)

time_second(9) and time_uptime(9) are widely used in the kernel to
quickly get the system UTC or system uptime as a time_t. However,
time_t is 64-bit everywhere, so it is not generally safe to use them
on 32-bit platforms: you have a split-read problem if your hardware
cannot perform atomic 64-bit reads.

This patch replaces time_second(9) with gettime(9), a safer successor
interface, throughout the kernel. Similarly, time_uptime(9) is replaced
with getuptime(9).

There is a performance cost on 32-bit platforms in exchange for
eliminating the split-read problem: instead of two register reads you
now have a lockless read loop to pull the values from the timehands.
This is really not *too* bad in the grand scheme of things, but
compared to what we were doing before it is several times slower.

There is no performance cost on 64-bit (__LP64__) platforms.

With input from visa@, dlg@, and tedu@.

Several bugs squashed by visa@.

ok kettenis@

udcf(4): tvtohz(9)+timeout_add(9) -> timeout_add_msec(9); ok kettenis@

Delete unnecessary <sys/file.h> includes

ok millert@ krw@

Avoid calling usbd_set_config_index() in *_attach() and let the stack
do it instead.

If anything bad happen due to a malformed descriptor it makes no sense
to try to attach a driver, and bail before probing.

The Swiss Federal Government decided to shut down HBG at the end of 2011.
On 6 September 2012 at 12:02:00 UTC both antenna towers were demolished by
controlled explosives. So this is not coming back and we can tedu the
support for HBG form the DCF77 drivers.
Remided by mbalmer

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

tedu another bunch of DVACT_DEACTIVATE handlers that do nothing but set
the 'dying' flag.

Get rid of various 'typedef struct' definitions and use plain structure
definitions instead. We don't change usb.h for now to stay compatible
with userland.

Tested by mpi@ on macppc and myself on i386.

ok mpi@

proc.h is way too much header for usb to handle.

you probably don't need to include vnode.h or reboot.h. you most definitely
absolutely do not need to include vnode.h because it includes uvm_extern.h
and you want the idiotic TRUE FALSE defines from uvm.

Remove config_activate() and DVACT_ACTIVATE. PCMCIA's the only thing
that's ever used it, and it's long since been changed to use
DVACT_{QUIESCE,SUSPEND,RESUME} instead.

ok deraadt@, dlg@; miod@ also agreed with this idea when I brought it
up a few weeks ago

garbage collect "usb events". without /dev/usb there is no way to access
them from userland, and nothing in the kernel uses them.

ok krw@, miod@

- use usb_lookup() instead of handrolled macro

ok jakemsr@

* replace per-driver dying and/or other state variables with use of
usbd_deactivete() and usbd_is_dying()
* use usbd_deactivate() in activate()/DEACTIVATE
* convert a few more direct checks of the associated bus' dying flag
with usbd_is_dying()

check that a timeout(9) has been initialized before deleting it

ok phessler

create another kthread to run xfer abort tasks. xfer abort tasks cannot
be run from the generic task kthread, because xfers that need to be
aborted block newly queued tasks from running (i.e. the xfer to be aborted
blocks the abort of that task). as there are now three types of usb
tasks, add an argument to usb_init_task() and another member to struct
usb_task to specify the task type.

fixes boot hangs that are showing up because we now use usb tasks to
attach/detach usb devices.

remove dying flag in detach() function.
The dying flag will be set in activate()/DEACTIVATE.

ok deraadt@

Get rid of devact enum, substitute it with an int and coresponding defines.

This is needed for the addition of further suspend/resume actions.

Okay deraadt@, marco@.

in ca_attach() there is no need to set .value and .flags of sensors to 0,
since autoconf(9) allocates softc with M_ZERO; ok deraadt

Kill some whitespace, removed unneeded stuff at the end of a list.

Don't terminate the array of matching product/vendor IDs with { 0, 0 }
when usb_lookup is used. If a device indeed has 0 as vendor and product
ID, like the HP DL165 BMC Server Engines SE USB Device, a wrong driver
will attach and the machine will most pbly crash.

Problem reported Rivo Nurges <rix@estpak.ee>
cvs: ----------------------------------------------------------------------

Use UT_{READ|WRITE}_VENDOR_DEVICE instead of self defined values.

Add support for the new Expert mouseCLOCK USB II that uses an FTDI
FT232R chip to interface the receiver (instead of the NetCologne
chip used on older models). Only the DCF77 and HBG receivers are
supported.

A sidenote: Gude ADS not only provides me with receivers and
documentation, but as a result of my feedback on their older receivers
the new ones now have proper USB product IDs to distinguish between
DCF77, HBG, and MSF.

Make sure that variables used during detach are initialized during attach.
Discussed with dlg.

No need for process context to degrade the sensor status.
Discussed with kettenis.

treat usb vendor/product names as a locator, and have usbd_print handle it,
so that it shows up before the :. as a result, all the usb devices do not
need to have name printing code anymore. all this now works and prints
nicely because usbd_probe_and_attach() is serialized. ok kettenis

Remove the definition and usage of the USB_DECLARE_DRIVER_CLASS and
USB_DECLARE_DRIVER macros.

No binary change.

ok dlg.

Remove the definition and use of the device_ptr_t which was a struct device *.

No binary change.

ok mk.

Remove the definition and use of the USBDEVNAME macro.

(This might look easy, but it was a big diff. Thanks to dlg and especially jsg
for looking over it; we found at least four mistakes in the initial diff.)

ok jsg.

Remove the definition and use of the USBDEV macro. It only created confusion
and the address of it's argument: USBDEV(sc->sc_dev) yields &sc->sc_dev.

No binary changes.

ok jsg.

Mechanical removal of USBBASEDEVICE. No binary change.

Tested by thib and myself.

ok mbalmer jsg

Farewell USB_{ATTACH,MATCH,DETACH}* you will not be missed.
ok deraadt@ krw@ mbalmer@

Back out last. the world is not i386.

Remove unused header files.

split userland & kernel struct sensor/sensordev so that the addition
of new fields in the future is less disruptive. This is done similar
to how struct proc is handled for ps(1). ok jmc (man page changes)
tested fkr simon, and more suggestions from millert

return is not a function.

adapt to new two-level sensor api; Constantine A. Murenin

Make sure the bit collecting is restarted after signal loss.

restart collecting bits immediately after a parity error, i.e. use the frame
immediately following the corrupt one; there is no no need to skip it by
waiting for the next minute gap and the sensor value has valid data one
minute earlier than before.

Re-arm signal loss detection only after we received a bit, not in itself. This
prevents the signal loss function from being called repeatedly when we have
no signal.

Faster synchronization after an unexpected minute gap.
Only look at the skew value if it is != 0.
Remove some dead code.

When compiled with UDCF_DEBUG, only output the bits received when
udcfdebug > 1.

Add code to detect (and eliminate) statistical outliers. Tested by me and
naddy; many thanks to naddy for continous testing and feedback during the
last weeks.

ok naddy

Fix an integer type promotion that can lead to wrong offsets on 64 bit arches.
Problem and solution found by Christian "Naddy" Weisgerber <naddy@openbsd.org>,
thanks!

- Use better names for the variables containing the timeouts in Hz instead of
just t1-t10.
- Remove some unused variables.
- Restructure some functions to remove some levels of indentation.

udcf(4) degraded the sensor status from OK to WARNING on the first receiption
error, which is to fast as this does not give ntpd the opportunity to use the
sensor value.

If we decode a valid time information, it has to be available to applications
for some minimal time. I choose a value of 5 minutes. If we do not receive
any new valid time information during this period, the sensor will be degraded
to WARNING, and after another fifteen minutes it will be further degraded to
CRITICAL.

"makes sense to me" henning@, "sure" deraadt@

Apply some formatting.

Make the timedelta sensor appear right when a device is plugged-in, with a
status of SENSOR_S_UNKNOWN. Do not set the SENSOR_FINVALID bit, sysctl
hw.sensors does not list sensors with this flags set and we want to the sensor
when the device is present.

ok henning

In a drivers activate() entry point, if on DVACT_DEACTIVATE it does
something, then it MUST return zero on DVACT_ACTIVATE, not EOPNOTSUPP;
this very popular bug has been cut and pasted a lot of times...
ok deraadt@ mickey@

remove the no longer needed time calculations now that we have them in
sys/kern/clock_subr.c

move clock_subr.c to a better place, and now it is always in the kernel
so that things can use it; tested on all architectures; ok kettenis

udcf(4) no longer needs the file sys/dev/clock_subr.c to be compiled and linked
to the kernel, the two functions it used from this file are now included in
the driver itself (like in nmea(4)). udcf(4) can now be enabled to the zaurus,
too. in fact on all arches that have USB support.

Typos in comments.

cleanup.

- ntpd no longer needs the NTP identifier as part of a timedelta sensor
description.
- create the timedelta sensors as early as possible, but mark them invalid
as long as there is no real data.
- update docs accordingly

Change the timedelta sensor descritpion so that the first four bytes contain
the NTP identifier (filled with spaces), followed by a space and the station
name (which might be the same as the NTP identifier).

All timedelta sensor have to follow this scheme as OpenNTPD needs it.

Requested by henning.

There is no longer the need to create INVALID sensors on stock, the sensor
framework will create hotplug events for sensor addition/deletion. ntpd
will catch these events.

When the device is attached, immediately create the timedelta sensor with the
SENSOR_FINVALID flag set and without a description. We do not need a dummy
description as a sensor with the FINVALID flags set will not show up or be used
anyways. It's merely an indication to applications that there is a sensor that
will eventually have a correct value.

The SENSOR_FINVALID flag is removed and the description is set to the proper
clockname as soon as we have received the first valid time information.

This enables ntpd to check for hotplug events and rescan the list of timedelta
sensors after a hotplug event.

discussed with deraadt.

cope with recent changes in the sensor framework

- provide the timedelta in nanoseconds
- provide a timestamp in the sensor

support for HBG and DCF77:

- detect clocktype (DCF77 or HBG) and set it once is the sensor description

Simplify time calculations.

help and ok kettenis@

make udcf work as a SENSOR_TIMEDELTA type sensor. The sensor value is the
offset between the time of the device and the localtime in mSec.

the sensor status is set to OK once we received a complete and valid time
information.

help and ok deraadt

open/close/cdevsw code goes bye bye; ok mbalber

- remove the 'needs-flag' from the udcf entry in files.usb and the #ifdef NUDCF
idom in udcf.c
- declare t1 to t7 as static to prevent the risk of name clashes
- add two missing function prototypes

help and ok miod@

Driver to decode the DCF77 time code using a USB attached Gude Expert
mouseCLOCK device; without a mechanism to pass data out.

ok deraadt

kernel: use gettime(9)/getuptime(9) in lieu of time_second(9)/time_uptime(9)

time_second(9) and time_uptime(9) are widely used in the kernel to
quickly get the system UTC or system uptime as a time_t. However,
time_t is 64-bit everywhere, so it is not generally safe to use them
on 32-bit platforms: you have a split-read problem if your hardware
cannot perform atomic 64-bit reads.

This patch replaces time_second(9) with gettime(9), a safer successor
interface, throughout the kernel. Similarly, time_uptime(9) is replaced
with getuptime(9).

There is a performance cost on 32-bit platforms in exchange for
eliminating the split-read problem: instead of two register reads you
now have a lockless read loop to pull the values from the timehands.
This is really not *too* bad in the grand scheme of things, but
compared to what we were doing before it is several times slower.

There is no performance cost on 64-bit (__LP64__) platforms.

With input from visa@, dlg@, and tedu@.

Several bugs squashed by visa@.

ok kettenis@

udcf(4): tvtohz(9)+timeout_add(9) -> timeout_add_msec(9); ok kettenis@

Delete unnecessary <sys/file.h> includes

ok millert@ krw@

Avoid calling usbd_set_config_index() in *_attach() and let the stack
do it instead.

If anything bad happen due to a malformed descriptor it makes no sense
to try to attach a driver, and bail before probing.

The Swiss Federal Government decided to shut down HBG at the end of 2011.
On 6 September 2012 at 12:02:00 UTC both antenna towers were demolished by
controlled explosives. So this is not coming back and we can tedu the
support for HBG form the DCF77 drivers.
Remided by mbalmer

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

tedu another bunch of DVACT_DEACTIVATE handlers that do nothing but set
the 'dying' flag.

Get rid of various 'typedef struct' definitions and use plain structure
definitions instead. We don't change usb.h for now to stay compatible
with userland.

Tested by mpi@ on macppc and myself on i386.

ok mpi@

proc.h is way too much header for usb to handle.

you probably don't need to include vnode.h or reboot.h. you most definitely
absolutely do not need to include vnode.h because it includes uvm_extern.h
and you want the idiotic TRUE FALSE defines from uvm.

Remove config_activate() and DVACT_ACTIVATE. PCMCIA's the only thing
that's ever used it, and it's long since been changed to use
DVACT_{QUIESCE,SUSPEND,RESUME} instead.

ok deraadt@, dlg@; miod@ also agreed with this idea when I brought it
up a few weeks ago

garbage collect "usb events". without /dev/usb there is no way to access
them from userland, and nothing in the kernel uses them.

ok krw@, miod@

- use usb_lookup() instead of handrolled macro

ok jakemsr@

* replace per-driver dying and/or other state variables with use of
usbd_deactivete() and usbd_is_dying()
* use usbd_deactivate() in activate()/DEACTIVATE
* convert a few more direct checks of the associated bus' dying flag
with usbd_is_dying()

check that a timeout(9) has been initialized before deleting it

ok phessler

create another kthread to run xfer abort tasks. xfer abort tasks cannot
be run from the generic task kthread, because xfers that need to be
aborted block newly queued tasks from running (i.e. the xfer to be aborted
blocks the abort of that task). as there are now three types of usb
tasks, add an argument to usb_init_task() and another member to struct
usb_task to specify the task type.

fixes boot hangs that are showing up because we now use usb tasks to
attach/detach usb devices.

remove dying flag in detach() function.
The dying flag will be set in activate()/DEACTIVATE.

ok deraadt@

Get rid of devact enum, substitute it with an int and coresponding defines.

This is needed for the addition of further suspend/resume actions.

Okay deraadt@, marco@.

in ca_attach() there is no need to set .value and .flags of sensors to 0,
since autoconf(9) allocates softc with M_ZERO; ok deraadt

Kill some whitespace, removed unneeded stuff at the end of a list.

Don't terminate the array of matching product/vendor IDs with { 0, 0 }
when usb_lookup is used. If a device indeed has 0 as vendor and product
ID, like the HP DL165 BMC Server Engines SE USB Device, a wrong driver
will attach and the machine will most pbly crash.

Problem reported Rivo Nurges <rix@estpak.ee>
cvs: ----------------------------------------------------------------------

Use UT_{READ|WRITE}_VENDOR_DEVICE instead of self defined values.

Add support for the new Expert mouseCLOCK USB II that uses an FTDI
FT232R chip to interface the receiver (instead of the NetCologne
chip used on older models). Only the DCF77 and HBG receivers are
supported.

A sidenote: Gude ADS not only provides me with receivers and
documentation, but as a result of my feedback on their older receivers
the new ones now have proper USB product IDs to distinguish between
DCF77, HBG, and MSF.

Make sure that variables used during detach are initialized during attach.
Discussed with dlg.

No need for process context to degrade the sensor status.
Discussed with kettenis.

treat usb vendor/product names as a locator, and have usbd_print handle it,
so that it shows up before the :. as a result, all the usb devices do not
need to have name printing code anymore. all this now works and prints
nicely because usbd_probe_and_attach() is serialized. ok kettenis

Remove the definition and usage of the USB_DECLARE_DRIVER_CLASS and
USB_DECLARE_DRIVER macros.

No binary change.

ok dlg.

Remove the definition and use of the device_ptr_t which was a struct device *.

No binary change.

ok mk.

Remove the definition and use of the USBDEVNAME macro.

(This might look easy, but it was a big diff. Thanks to dlg and especially jsg
for looking over it; we found at least four mistakes in the initial diff.)

ok jsg.

Remove the definition and use of the USBDEV macro. It only created confusion
and the address of it's argument: USBDEV(sc->sc_dev) yields &sc->sc_dev.

No binary changes.

ok jsg.

Mechanical removal of USBBASEDEVICE. No binary change.

Tested by thib and myself.

ok mbalmer jsg

Farewell USB_{ATTACH,MATCH,DETACH}* you will not be missed.
ok deraadt@ krw@ mbalmer@

Back out last. the world is not i386.

Remove unused header files.

split userland & kernel struct sensor/sensordev so that the addition
of new fields in the future is less disruptive. This is done similar
to how struct proc is handled for ps(1). ok jmc (man page changes)
tested fkr simon, and more suggestions from millert

return is not a function.

adapt to new two-level sensor api; Constantine A. Murenin

Make sure the bit collecting is restarted after signal loss.

restart collecting bits immediately after a parity error, i.e. use the frame
immediately following the corrupt one; there is no no need to skip it by
waiting for the next minute gap and the sensor value has valid data one
minute earlier than before.

Re-arm signal loss detection only after we received a bit, not in itself. This
prevents the signal loss function from being called repeatedly when we have
no signal.

Faster synchronization after an unexpected minute gap.
Only look at the skew value if it is != 0.
Remove some dead code.

When compiled with UDCF_DEBUG, only output the bits received when
udcfdebug > 1.

Add code to detect (and eliminate) statistical outliers. Tested by me and
naddy; many thanks to naddy for continous testing and feedback during the
last weeks.

ok naddy

Fix an integer type promotion that can lead to wrong offsets on 64 bit arches.
Problem and solution found by Christian "Naddy" Weisgerber <naddy@openbsd.org>,
thanks!

- Use better names for the variables containing the timeouts in Hz instead of
just t1-t10.
- Remove some unused variables.
- Restructure some functions to remove some levels of indentation.

udcf(4) degraded the sensor status from OK to WARNING on the first receiption
error, which is to fast as this does not give ntpd the opportunity to use the
sensor value.

If we decode a valid time information, it has to be available to applications
for some minimal time. I choose a value of 5 minutes. If we do not receive
any new valid time information during this period, the sensor will be degraded
to WARNING, and after another fifteen minutes it will be further degraded to
CRITICAL.

"makes sense to me" henning@, "sure" deraadt@

Apply some formatting.

Make the timedelta sensor appear right when a device is plugged-in, with a
status of SENSOR_S_UNKNOWN. Do not set the SENSOR_FINVALID bit, sysctl
hw.sensors does not list sensors with this flags set and we want to the sensor
when the device is present.

ok henning

In a drivers activate() entry point, if on DVACT_DEACTIVATE it does
something, then it MUST return zero on DVACT_ACTIVATE, not EOPNOTSUPP;
this very popular bug has been cut and pasted a lot of times...
ok deraadt@ mickey@

remove the no longer needed time calculations now that we have them in
sys/kern/clock_subr.c

move clock_subr.c to a better place, and now it is always in the kernel
so that things can use it; tested on all architectures; ok kettenis

udcf(4) no longer needs the file sys/dev/clock_subr.c to be compiled and linked
to the kernel, the two functions it used from this file are now included in
the driver itself (like in nmea(4)). udcf(4) can now be enabled to the zaurus,
too. in fact on all arches that have USB support.

Typos in comments.

cleanup.

- ntpd no longer needs the NTP identifier as part of a timedelta sensor
description.
- create the timedelta sensors as early as possible, but mark them invalid
as long as there is no real data.
- update docs accordingly

Change the timedelta sensor descritpion so that the first four bytes contain
the NTP identifier (filled with spaces), followed by a space and the station
name (which might be the same as the NTP identifier).

All timedelta sensor have to follow this scheme as OpenNTPD needs it.

Requested by henning.

There is no longer the need to create INVALID sensors on stock, the sensor
framework will create hotplug events for sensor addition/deletion. ntpd
will catch these events.

When the device is attached, immediately create the timedelta sensor with the
SENSOR_FINVALID flag set and without a description. We do not need a dummy
description as a sensor with the FINVALID flags set will not show up or be used
anyways. It's merely an indication to applications that there is a sensor that
will eventually have a correct value.

The SENSOR_FINVALID flag is removed and the description is set to the proper
clockname as soon as we have received the first valid time information.

This enables ntpd to check for hotplug events and rescan the list of timedelta
sensors after a hotplug event.

discussed with deraadt.

cope with recent changes in the sensor framework

- provide the timedelta in nanoseconds
- provide a timestamp in the sensor

support for HBG and DCF77:

- detect clocktype (DCF77 or HBG) and set it once is the sensor description

Simplify time calculations.

help and ok kettenis@

make udcf work as a SENSOR_TIMEDELTA type sensor. The sensor value is the
offset between the time of the device and the localtime in mSec.

the sensor status is set to OK once we received a complete and valid time
information.

help and ok deraadt

open/close/cdevsw code goes bye bye; ok mbalber

- remove the 'needs-flag' from the udcf entry in files.usb and the #ifdef NUDCF
idom in udcf.c
- declare t1 to t7 as static to prevent the risk of name clashes
- add two missing function prototypes

help and ok miod@

Driver to decode the DCF77 time code using a USB attached Gude Expert
mouseCLOCK device; without a mechanism to pass data out.

ok deraadt

udcf(4): tvtohz(9)+timeout_add(9) -> timeout_add_msec(9); ok kettenis@

Delete unnecessary <sys/file.h> includes

ok millert@ krw@

Avoid calling usbd_set_config_index() in *_attach() and let the stack
do it instead.

If anything bad happen due to a malformed descriptor it makes no sense
to try to attach a driver, and bail before probing.

The Swiss Federal Government decided to shut down HBG at the end of 2011.
On 6 September 2012 at 12:02:00 UTC both antenna towers were demolished by
controlled explosives. So this is not coming back and we can tedu the
support for HBG form the DCF77 drivers.
Remided by mbalmer

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

tedu another bunch of DVACT_DEACTIVATE handlers that do nothing but set
the 'dying' flag.

Get rid of various 'typedef struct' definitions and use plain structure
definitions instead. We don't change usb.h for now to stay compatible
with userland.

Tested by mpi@ on macppc and myself on i386.

ok mpi@

proc.h is way too much header for usb to handle.

you probably don't need to include vnode.h or reboot.h. you most definitely
absolutely do not need to include vnode.h because it includes uvm_extern.h
and you want the idiotic TRUE FALSE defines from uvm.

Remove config_activate() and DVACT_ACTIVATE. PCMCIA's the only thing
that's ever used it, and it's long since been changed to use
DVACT_{QUIESCE,SUSPEND,RESUME} instead.

ok deraadt@, dlg@; miod@ also agreed with this idea when I brought it
up a few weeks ago

garbage collect "usb events". without /dev/usb there is no way to access
them from userland, and nothing in the kernel uses them.

ok krw@, miod@

- use usb_lookup() instead of handrolled macro

ok jakemsr@

* replace per-driver dying and/or other state variables with use of
usbd_deactivete() and usbd_is_dying()
* use usbd_deactivate() in activate()/DEACTIVATE
* convert a few more direct checks of the associated bus' dying flag
with usbd_is_dying()

check that a timeout(9) has been initialized before deleting it

ok phessler

create another kthread to run xfer abort tasks. xfer abort tasks cannot
be run from the generic task kthread, because xfers that need to be
aborted block newly queued tasks from running (i.e. the xfer to be aborted
blocks the abort of that task). as there are now three types of usb
tasks, add an argument to usb_init_task() and another member to struct
usb_task to specify the task type.

fixes boot hangs that are showing up because we now use usb tasks to
attach/detach usb devices.

remove dying flag in detach() function.
The dying flag will be set in activate()/DEACTIVATE.

ok deraadt@

Get rid of devact enum, substitute it with an int and coresponding defines.

This is needed for the addition of further suspend/resume actions.

Okay deraadt@, marco@.

in ca_attach() there is no need to set .value and .flags of sensors to 0,
since autoconf(9) allocates softc with M_ZERO; ok deraadt

Kill some whitespace, removed unneeded stuff at the end of a list.

Don't terminate the array of matching product/vendor IDs with { 0, 0 }
when usb_lookup is used. If a device indeed has 0 as vendor and product
ID, like the HP DL165 BMC Server Engines SE USB Device, a wrong driver
will attach and the machine will most pbly crash.

Problem reported Rivo Nurges <rix@estpak.ee>
cvs: ----------------------------------------------------------------------

Use UT_{READ|WRITE}_VENDOR_DEVICE instead of self defined values.

Add support for the new Expert mouseCLOCK USB II that uses an FTDI
FT232R chip to interface the receiver (instead of the NetCologne
chip used on older models). Only the DCF77 and HBG receivers are
supported.

A sidenote: Gude ADS not only provides me with receivers and
documentation, but as a result of my feedback on their older receivers
the new ones now have proper USB product IDs to distinguish between
DCF77, HBG, and MSF.

Make sure that variables used during detach are initialized during attach.
Discussed with dlg.

No need for process context to degrade the sensor status.
Discussed with kettenis.

treat usb vendor/product names as a locator, and have usbd_print handle it,
so that it shows up before the :. as a result, all the usb devices do not
need to have name printing code anymore. all this now works and prints
nicely because usbd_probe_and_attach() is serialized. ok kettenis

Remove the definition and usage of the USB_DECLARE_DRIVER_CLASS and
USB_DECLARE_DRIVER macros.

No binary change.

ok dlg.

Remove the definition and use of the device_ptr_t which was a struct device *.

No binary change.

ok mk.

Remove the definition and use of the USBDEVNAME macro.

(This might look easy, but it was a big diff. Thanks to dlg and especially jsg
for looking over it; we found at least four mistakes in the initial diff.)

ok jsg.

Remove the definition and use of the USBDEV macro. It only created confusion
and the address of it's argument: USBDEV(sc->sc_dev) yields &sc->sc_dev.

No binary changes.

ok jsg.

Mechanical removal of USBBASEDEVICE. No binary change.

Tested by thib and myself.

ok mbalmer jsg

Farewell USB_{ATTACH,MATCH,DETACH}* you will not be missed.
ok deraadt@ krw@ mbalmer@

Back out last. the world is not i386.

Remove unused header files.

split userland & kernel struct sensor/sensordev so that the addition
of new fields in the future is less disruptive. This is done similar
to how struct proc is handled for ps(1). ok jmc (man page changes)
tested fkr simon, and more suggestions from millert

return is not a function.

adapt to new two-level sensor api; Constantine A. Murenin

Make sure the bit collecting is restarted after signal loss.

restart collecting bits immediately after a parity error, i.e. use the frame
immediately following the corrupt one; there is no no need to skip it by
waiting for the next minute gap and the sensor value has valid data one
minute earlier than before.

Re-arm signal loss detection only after we received a bit, not in itself. This
prevents the signal loss function from being called repeatedly when we have
no signal.

Faster synchronization after an unexpected minute gap.
Only look at the skew value if it is != 0.
Remove some dead code.

When compiled with UDCF_DEBUG, only output the bits received when
udcfdebug > 1.

Add code to detect (and eliminate) statistical outliers. Tested by me and
naddy; many thanks to naddy for continous testing and feedback during the
last weeks.

ok naddy

Fix an integer type promotion that can lead to wrong offsets on 64 bit arches.
Problem and solution found by Christian "Naddy" Weisgerber <naddy@openbsd.org>,
thanks!

- Use better names for the variables containing the timeouts in Hz instead of
just t1-t10.
- Remove some unused variables.
- Restructure some functions to remove some levels of indentation.

udcf(4) degraded the sensor status from OK to WARNING on the first receiption
error, which is to fast as this does not give ntpd the opportunity to use the
sensor value.

If we decode a valid time information, it has to be available to applications
for some minimal time. I choose a value of 5 minutes. If we do not receive
any new valid time information during this period, the sensor will be degraded
to WARNING, and after another fifteen minutes it will be further degraded to
CRITICAL.

"makes sense to me" henning@, "sure" deraadt@

Apply some formatting.

Make the timedelta sensor appear right when a device is plugged-in, with a
status of SENSOR_S_UNKNOWN. Do not set the SENSOR_FINVALID bit, sysctl
hw.sensors does not list sensors with this flags set and we want to the sensor
when the device is present.

ok henning

In a drivers activate() entry point, if on DVACT_DEACTIVATE it does
something, then it MUST return zero on DVACT_ACTIVATE, not EOPNOTSUPP;
this very popular bug has been cut and pasted a lot of times...
ok deraadt@ mickey@

remove the no longer needed time calculations now that we have them in
sys/kern/clock_subr.c

move clock_subr.c to a better place, and now it is always in the kernel
so that things can use it; tested on all architectures; ok kettenis

udcf(4) no longer needs the file sys/dev/clock_subr.c to be compiled and linked
to the kernel, the two functions it used from this file are now included in
the driver itself (like in nmea(4)). udcf(4) can now be enabled to the zaurus,
too. in fact on all arches that have USB support.

Typos in comments.

cleanup.

- ntpd no longer needs the NTP identifier as part of a timedelta sensor
description.
- create the timedelta sensors as early as possible, but mark them invalid
as long as there is no real data.
- update docs accordingly

Change the timedelta sensor descritpion so that the first four bytes contain
the NTP identifier (filled with spaces), followed by a space and the station
name (which might be the same as the NTP identifier).

All timedelta sensor have to follow this scheme as OpenNTPD needs it.

Requested by henning.

There is no longer the need to create INVALID sensors on stock, the sensor
framework will create hotplug events for sensor addition/deletion. ntpd
will catch these events.

When the device is attached, immediately create the timedelta sensor with the
SENSOR_FINVALID flag set and without a description. We do not need a dummy
description as a sensor with the FINVALID flags set will not show up or be used
anyways. It's merely an indication to applications that there is a sensor that
will eventually have a correct value.

The SENSOR_FINVALID flag is removed and the description is set to the proper
clockname as soon as we have received the first valid time information.

This enables ntpd to check for hotplug events and rescan the list of timedelta
sensors after a hotplug event.

discussed with deraadt.

cope with recent changes in the sensor framework

- provide the timedelta in nanoseconds
- provide a timestamp in the sensor

support for HBG and DCF77:

- detect clocktype (DCF77 or HBG) and set it once is the sensor description

Simplify time calculations.

help and ok kettenis@

make udcf work as a SENSOR_TIMEDELTA type sensor. The sensor value is the
offset between the time of the device and the localtime in mSec.

the sensor status is set to OK once we received a complete and valid time
information.

help and ok deraadt

open/close/cdevsw code goes bye bye; ok mbalber

- remove the 'needs-flag' from the udcf entry in files.usb and the #ifdef NUDCF
idom in udcf.c
- declare t1 to t7 as static to prevent the risk of name clashes
- add two missing function prototypes

help and ok miod@

Driver to decode the DCF77 time code using a USB attached Gude Expert
mouseCLOCK device; without a mechanism to pass data out.

ok deraadt

Delete unnecessary <sys/file.h> includes

ok millert@ krw@

Avoid calling usbd_set_config_index() in *_attach() and let the stack
do it instead.

If anything bad happen due to a malformed descriptor it makes no sense
to try to attach a driver, and bail before probing.

The Swiss Federal Government decided to shut down HBG at the end of 2011.
On 6 September 2012 at 12:02:00 UTC both antenna towers were demolished by
controlled explosives. So this is not coming back and we can tedu the
support for HBG form the DCF77 drivers.
Remided by mbalmer

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

tedu another bunch of DVACT_DEACTIVATE handlers that do nothing but set
the 'dying' flag.

Get rid of various 'typedef struct' definitions and use plain structure
definitions instead. We don't change usb.h for now to stay compatible
with userland.

Tested by mpi@ on macppc and myself on i386.

ok mpi@

proc.h is way too much header for usb to handle.

you probably don't need to include vnode.h or reboot.h. you most definitely
absolutely do not need to include vnode.h because it includes uvm_extern.h
and you want the idiotic TRUE FALSE defines from uvm.

Remove config_activate() and DVACT_ACTIVATE. PCMCIA's the only thing
that's ever used it, and it's long since been changed to use
DVACT_{QUIESCE,SUSPEND,RESUME} instead.

ok deraadt@, dlg@; miod@ also agreed with this idea when I brought it
up a few weeks ago

garbage collect "usb events". without /dev/usb there is no way to access
them from userland, and nothing in the kernel uses them.

ok krw@, miod@

- use usb_lookup() instead of handrolled macro

ok jakemsr@

* replace per-driver dying and/or other state variables with use of
usbd_deactivete() and usbd_is_dying()
* use usbd_deactivate() in activate()/DEACTIVATE
* convert a few more direct checks of the associated bus' dying flag
with usbd_is_dying()

check that a timeout(9) has been initialized before deleting it

ok phessler

create another kthread to run xfer abort tasks. xfer abort tasks cannot
be run from the generic task kthread, because xfers that need to be
aborted block newly queued tasks from running (i.e. the xfer to be aborted
blocks the abort of that task). as there are now three types of usb
tasks, add an argument to usb_init_task() and another member to struct
usb_task to specify the task type.

fixes boot hangs that are showing up because we now use usb tasks to
attach/detach usb devices.

remove dying flag in detach() function.
The dying flag will be set in activate()/DEACTIVATE.

ok deraadt@

Get rid of devact enum, substitute it with an int and coresponding defines.

This is needed for the addition of further suspend/resume actions.

Okay deraadt@, marco@.

in ca_attach() there is no need to set .value and .flags of sensors to 0,
since autoconf(9) allocates softc with M_ZERO; ok deraadt

Kill some whitespace, removed unneeded stuff at the end of a list.

Don't terminate the array of matching product/vendor IDs with { 0, 0 }
when usb_lookup is used. If a device indeed has 0 as vendor and product
ID, like the HP DL165 BMC Server Engines SE USB Device, a wrong driver
will attach and the machine will most pbly crash.

Problem reported Rivo Nurges <rix@estpak.ee>
cvs: ----------------------------------------------------------------------

Use UT_{READ|WRITE}_VENDOR_DEVICE instead of self defined values.

Add support for the new Expert mouseCLOCK USB II that uses an FTDI
FT232R chip to interface the receiver (instead of the NetCologne
chip used on older models). Only the DCF77 and HBG receivers are
supported.

A sidenote: Gude ADS not only provides me with receivers and
documentation, but as a result of my feedback on their older receivers
the new ones now have proper USB product IDs to distinguish between
DCF77, HBG, and MSF.

Make sure that variables used during detach are initialized during attach.
Discussed with dlg.

No need for process context to degrade the sensor status.
Discussed with kettenis.

treat usb vendor/product names as a locator, and have usbd_print handle it,
so that it shows up before the :. as a result, all the usb devices do not
need to have name printing code anymore. all this now works and prints
nicely because usbd_probe_and_attach() is serialized. ok kettenis

Remove the definition and usage of the USB_DECLARE_DRIVER_CLASS and
USB_DECLARE_DRIVER macros.

No binary change.

ok dlg.

Remove the definition and use of the device_ptr_t which was a struct device *.

No binary change.

ok mk.

Remove the definition and use of the USBDEVNAME macro.

(This might look easy, but it was a big diff. Thanks to dlg and especially jsg
for looking over it; we found at least four mistakes in the initial diff.)

ok jsg.

Remove the definition and use of the USBDEV macro. It only created confusion
and the address of it's argument: USBDEV(sc->sc_dev) yields &sc->sc_dev.

No binary changes.

ok jsg.

Mechanical removal of USBBASEDEVICE. No binary change.

Tested by thib and myself.

ok mbalmer jsg

Farewell USB_{ATTACH,MATCH,DETACH}* you will not be missed.
ok deraadt@ krw@ mbalmer@

Back out last. the world is not i386.

Remove unused header files.

split userland & kernel struct sensor/sensordev so that the addition
of new fields in the future is less disruptive. This is done similar
to how struct proc is handled for ps(1). ok jmc (man page changes)
tested fkr simon, and more suggestions from millert

return is not a function.

adapt to new two-level sensor api; Constantine A. Murenin

Make sure the bit collecting is restarted after signal loss.

restart collecting bits immediately after a parity error, i.e. use the frame
immediately following the corrupt one; there is no no need to skip it by
waiting for the next minute gap and the sensor value has valid data one
minute earlier than before.

Re-arm signal loss detection only after we received a bit, not in itself. This
prevents the signal loss function from being called repeatedly when we have
no signal.

Faster synchronization after an unexpected minute gap.
Only look at the skew value if it is != 0.
Remove some dead code.

When compiled with UDCF_DEBUG, only output the bits received when
udcfdebug > 1.

Add code to detect (and eliminate) statistical outliers. Tested by me and
naddy; many thanks to naddy for continous testing and feedback during the
last weeks.

ok naddy

Fix an integer type promotion that can lead to wrong offsets on 64 bit arches.
Problem and solution found by Christian "Naddy" Weisgerber <naddy@openbsd.org>,
thanks!

- Use better names for the variables containing the timeouts in Hz instead of
just t1-t10.
- Remove some unused variables.
- Restructure some functions to remove some levels of indentation.

udcf(4) degraded the sensor status from OK to WARNING on the first receiption
error, which is to fast as this does not give ntpd the opportunity to use the
sensor value.

If we decode a valid time information, it has to be available to applications
for some minimal time. I choose a value of 5 minutes. If we do not receive
any new valid time information during this period, the sensor will be degraded
to WARNING, and after another fifteen minutes it will be further degraded to
CRITICAL.

"makes sense to me" henning@, "sure" deraadt@

Apply some formatting.

Make the timedelta sensor appear right when a device is plugged-in, with a
status of SENSOR_S_UNKNOWN. Do not set the SENSOR_FINVALID bit, sysctl
hw.sensors does not list sensors with this flags set and we want to the sensor
when the device is present.

ok henning

In a drivers activate() entry point, if on DVACT_DEACTIVATE it does
something, then it MUST return zero on DVACT_ACTIVATE, not EOPNOTSUPP;
this very popular bug has been cut and pasted a lot of times...
ok deraadt@ mickey@

remove the no longer needed time calculations now that we have them in
sys/kern/clock_subr.c

move clock_subr.c to a better place, and now it is always in the kernel
so that things can use it; tested on all architectures; ok kettenis

udcf(4) no longer needs the file sys/dev/clock_subr.c to be compiled and linked
to the kernel, the two functions it used from this file are now included in
the driver itself (like in nmea(4)). udcf(4) can now be enabled to the zaurus,
too. in fact on all arches that have USB support.

Typos in comments.

cleanup.

- ntpd no longer needs the NTP identifier as part of a timedelta sensor
description.
- create the timedelta sensors as early as possible, but mark them invalid
as long as there is no real data.
- update docs accordingly

Change the timedelta sensor descritpion so that the first four bytes contain
the NTP identifier (filled with spaces), followed by a space and the station
name (which might be the same as the NTP identifier).

All timedelta sensor have to follow this scheme as OpenNTPD needs it.

Requested by henning.

There is no longer the need to create INVALID sensors on stock, the sensor
framework will create hotplug events for sensor addition/deletion. ntpd
will catch these events.

When the device is attached, immediately create the timedelta sensor with the
SENSOR_FINVALID flag set and without a description. We do not need a dummy
description as a sensor with the FINVALID flags set will not show up or be used
anyways. It's merely an indication to applications that there is a sensor that
will eventually have a correct value.

The SENSOR_FINVALID flag is removed and the description is set to the proper
clockname as soon as we have received the first valid time information.

This enables ntpd to check for hotplug events and rescan the list of timedelta
sensors after a hotplug event.

discussed with deraadt.

cope with recent changes in the sensor framework

- provide the timedelta in nanoseconds
- provide a timestamp in the sensor

support for HBG and DCF77:

- detect clocktype (DCF77 or HBG) and set it once is the sensor description

Simplify time calculations.

help and ok kettenis@

make udcf work as a SENSOR_TIMEDELTA type sensor. The sensor value is the
offset between the time of the device and the localtime in mSec.

the sensor status is set to OK once we received a complete and valid time
information.

help and ok deraadt

open/close/cdevsw code goes bye bye; ok mbalber

- remove the 'needs-flag' from the udcf entry in files.usb and the #ifdef NUDCF
idom in udcf.c
- declare t1 to t7 as static to prevent the risk of name clashes
- add two missing function prototypes

help and ok miod@

Driver to decode the DCF77 time code using a USB attached Gude Expert
mouseCLOCK device; without a mechanism to pass data out.

ok deraadt