* Add these to the usb_support_descriptor when adding from the quirky devices.
* Extend the SIXAXIS comment a bit.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42000 a95241bf-73f2-0310-859d-f6bbb57e9c96
descriptor. This allows us to support devices that aren't strictly HID or only
provide vendor specific interfaces by constructing report descriptors for their
data format.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41992 a95241bf-73f2-0310-859d-f6bbb57e9c96
absolute coordinates (qemu/kvm -usbdevice tablet) and various digitizers. If a
tablet-like usage is determined the MouseProtocolHandler publishes to
input/tablet instead of input/mouse and reports its info in a tablet_movement
structure instead of a mouse_movement structure, allowing for absolute
coordinates. Note that right now only the absolute coordinates are used, the
other tablet specific inputs (eraser, pressure, tilt and contact) aren't yet
filled in. I will add these as a second step.
I've only tested this with kvm where this makes the "seamless" mouse integration
work and therefore fixes#5989. Note that QEMU for BeOS/Haiku doesn't support
that mode (patch pending on my side) and will lock the pointer into the upper
left corner if -usbdevice tablet is used (whereas it previously just did
nothing).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41951 a95241bf-73f2-0310-859d-f6bbb57e9c96
pointing devices that publish entries into /dev/input/tablet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41950 a95241bf-73f2-0310-859d-f6bbb57e9c96
to build a new descriptor using a HIDWriter. This way a completely new
descriptor can be built on the fly or a fixed already patched blob can be
written to the writer if desired.
* Implemented writing a new report descriptor for Sony SIXAXIS controllers. The
descriptor also includes the pressure sensitive button states (exposed as 12
axes) and the accelerometers/gyroscope (exposed as another 4 axes). While the
motion sensing data is now available (and fun to look at), the usefulness is
fairly limited, as the values are not scaled/transformed in any way. It might
be possible to generically do such transformations using unit descriptions,
but those are not implemented yet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41925 a95241bf-73f2-0310-859d-f6bbb57e9c96
duplicate usages and simply adds more axes in that case. It also removes the
gaps that were previously put in place if there were higher numbered axes.
However, since the ordering of the axes now depends on the ordering inside the
HID collection, it is possible that some controllers won't have the X, Y and Z
axis mapped as the first three axes, which might confuse applications. I've not
encountered a report structure that would lead to such a situation yet, but then
again the amount of reports I was able to get hold of is fairly limited. If it
becomes a problem the mapping needs to be adjusted accordingly.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41923 a95241bf-73f2-0310-859d-f6bbb57e9c96
report descriptors. These descriptors aren't optimized for efficiency (the
writer doesn't attempt to avoid writing unneeded global items for example) but
they will only serve as input for the HIDParser that parses them correctly
regardless.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41921 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Only cancel the update thread if the last open instance is closed. This fixes
that other applications using the same joystick would stop getting updated
values as soon as one application was closed.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41905 a95241bf-73f2-0310-859d-f6bbb57e9c96
function that is triggered before the protocol handlers are added.
* Use the quirky device mechanism to support the Sony SIXAXIS controller (the
PS3 one). It requires a specific get_report to become operational. Note that
you still have to push the PS button to enable it after plugging in. Note also
that only the two analog sticks are reported as axis, the analog values of the
buttons as well as the motion sensors aren't described by the HID descriptor
and therefore aren't accessible. The digital button states (on/off only) work
however. A fixed descriptor could possibly be crafted to support the missing
features later on though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41902 a95241bf-73f2-0310-859d-f6bbb57e9c96
handlers. A full sized keyboard is assumed when LEDs are found and those are
preferred then. This makes USB keyboards work (at least those that worked
by voluntarily entering KDL before, i.e. those that are hooked up to UHCI root
ports) also when entering KDL by means of a panic. The first (few) keys are
swallowed though, but that's certainly better than it not working at all.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41898 a95241bf-73f2-0310-859d-f6bbb57e9c96
that mode when they support it. Right now there is only usb_hid though. With
this commit controllers with an arbitrary number of axes, hats and buttons are
now supported by BJoystick and by the BJoystick <-> driver protocol as well as
by usb_hid (usb_hid doesn't yet have support for hats though).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41893 a95241bf-73f2-0310-859d-f6bbb57e9c96
both, variable and extended joystick mode, using the same mechanism of mimicing
the extended_joystick structure with the variable one if required. The mode
used depends on the support flag coming in from BJoystick (currently off), so
with this commit the behaviour doesn't change. I'm going to flip that switch
next though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41891 a95241bf-73f2-0310-859d-f6bbb57e9c96
methods the members are populated, the flat data size is calcualted, the data
buffer is allocated and the convenience pointers are set up automatically. With
initialize() an arbitrary configuration can be created, with
initialize_to_extended_joystick() a configuration is created that exactly
matches the data structure of the extended_joystick struct.
Having them here makes it easy to use the structure from a driver as well
without the need to set up everything manually.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41888 a95241bf-73f2-0310-859d-f6bbb57e9c96
into joystick_driver.h as it is convenient to have also on the driver side.
* Added comments explaining the data structure and use case of the structure.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41886 a95241bf-73f2-0310-859d-f6bbb57e9c96
inside BJoystick.
* Add joystick_module_info flag to communicate support for variably sized reads.
* The variably sized data structure is set up to describe either the actual
amount of data, when variably sized reads are supported by the driver, or it
is set up so that it exactly matches the data layout of the extended_joystick
structure. This allows us to support both as input data, while only needing to
care about a single format inside BJoystick. Convenience pointers allow the
data to be retrieved without additional overhead or extra logic.
* Add some sanity checks and ensure some boundaries when dealing reading data
from the variably sized structure (as there might not be any buttons, hats,
axis at all now).
* Ensure that the extended_joystick structure doesn't change in size due to
padding by making it _PACKED (it wasn't padded though).
This is still supposed to work exactly as before. However, it opens up the
possibility to actually support arbitrary controllers with arbitrary axis, hat
and button counts. It therefore allows to actually deliver what the BJoystick
API was designed to handle all along.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41885 a95241bf-73f2-0310-859d-f6bbb57e9c96
be retrieved. I think the fixed layout will actually make it impossible to see
anything beyond the 9th button however... The coding style of this app is,
different, to say the least and I didn't really bother fixing it up right now,
so the changes just mimic what's there.
* Some automatic whitespace cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41883 a95241bf-73f2-0310-859d-f6bbb57e9c96
Get*Values() functions but retrieves an arbitrary amount of button states using
a boolean array. This overcomes the 32 button limitation on the API side. Right
now the function does simply set the first 32 states from the button bitmap, as
the BJoystick to driver interface hasn't been adapted yet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41882 a95241bf-73f2-0310-859d-f6bbb57e9c96
already anyway (as compared to the original BeOS ones) this won't introduce any
missing symbols.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41879 a95241bf-73f2-0310-859d-f6bbb57e9c96
declarations to group them more logically. No functional changes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41878 a95241bf-73f2-0310-859d-f6bbb57e9c96
documented in the BeBook. Sadly the success return is specified as "a positive
integer", so I didn't change it to B_OK.
* We actually want non-blocking mode, so don't reset the O_NONBLOCK flag. It was
ignored before anyway though, so this doesn't change anything.
* The legacy buttons 1 and 2 are in the pressed state when false, so initialize
them to true instead.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41876 a95241bf-73f2-0310-859d-f6bbb57e9c96
(which is uint32), but the legacy BJoystick axis values are just the same int16
as the ones in extended mode, so nothing has to be done besides assigning them.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41875 a95241bf-73f2-0310-859d-f6bbb57e9c96
to be updated with newly plugged in devices without the need to recreate a
BJoystick object.
* Do an initial (re-)scan on object creation instead of always scanning in
CountDevices() and by extension GetDeviceName().
* This makes CountDevices() and GetDeviceName() stable with regards to the
device list, so that an enumerating application doesn't suddenly get more/less
devices than it might expect.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41870 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Include sorting, header style, macro naming, variable naming, spacing, ...
* Some simplifications, early returns and the like.
* The device name list items were allocated with new but deleted with free. The
whole private BJoystick list being modified by a friend class is dubious
though and should be reworked. Ideally node monitoring would be implemented
so that the list can be re-populated on demand instead of scanning through
every time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41868 a95241bf-73f2-0310-859d-f6bbb57e9c96
with the BJoystick requirements:
* Make the Read() non-blocking. This is required as BJoystick is a polling
interface. A single current state is used that is updated by a separate thread
on report arrival. The thread is spawned as soon as the ProtocolHandler is
opened for the first time (and quit at the first wait return after the
ProtocolHandler is closed). With this we can simply return the current state
on read.
* Remove the ring buffer as it was not needed in the first place. This also
happens to solve the problem of sharing a JoystickProtocolHandler. Before,
concurrent reads would queue up the same result multiple times in the ring
buffer and then return stale data on the next update.
Solves most of #7629.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41865 a95241bf-73f2-0310-859d-f6bbb57e9c96
ever wait for reports on demand, there's no need for buffering at all.
This removes some unnecessary copying and ring buffer overhead, the ring buffer
will now in fact not be created at all.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41864 a95241bf-73f2-0310-859d-f6bbb57e9c96
the BJoystick API and joystick drivers.
* Also fix a few warnings in the code that were caused by using NULL instead of
0 for BWindow/BView flags.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41852 a95241bf-73f2-0310-859d-f6bbb57e9c96
This includes input scaling, so the resulting input ranges should always match
the ones of the BJoystick data (hence no calibration should be required).
* It supports joysticks, gamepads and multi-axis controllers. I've only tested
it with a Microsoft SideWinder Gamepad Pro so far, which now works as expected
with stickit.
* Fixes#7429.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41851 a95241bf-73f2-0310-859d-f6bbb57e9c96
info and values. Inspired by and in parts based on the patch by caz_haiku in
ticket #7429 (though rewritten completely due to the other changes). Thanks
for the pointers!
* Clean up the mixup of internal joystick info and the one from
joystick_driver.h so that BJoystick and the drivers talk about the same
structures.
* Extensive coding style cleanup, simplifications, NULL checks, early returns,
std::nothrow allocations, include sorting, argument naming, ... that kind of
stuff.
* Added some TODO notes for remaining stuff.
* Automatic (and manual) whitespace cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41849 a95241bf-73f2-0310-859d-f6bbb57e9c96
of returning B_ERROR to the ProtocolHandler base class. Not used yet, but will
be used for the BJoystick <-> JoystickProtocolHandler.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41835 a95241bf-73f2-0310-859d-f6bbb57e9c96
handle the different device classes. Handlers are now added based on the
application collections that the HID descriptor describes instead of by
enumerating the different report items inside the reports. This means that a
device is now logically treated as a mouse when it comes with an application
collection that designates it as a mouse, instead of when there is a report
that contains an X and a Y axis. This resolves the conflicts that gamepads
and joysticks were added as mice due to them containing such elements. This
therefore fixes#4499 and opens up the way to properly handle other device
types like joysticks (#7429), gamepads, tablets (#7354, #5989 and #7481) and
so on. I'll work on gamepads/joysticks next and see where we stand for tablets
later.
* Added a few enumeration functions to HIDCollection to support the above.
* Fix the root collection handling. A device doesn't describe a single root
collection and then adds everything as a child. Instead it just has multiple
collections on level 0. We account for that now by always creating an empty
logical collection as the root collection where all the collections of the
descriptor get added.
* Rename the {Mouse|Keyboard}Device.{cpp|h} to
{Mouse|Keyboard}ProtocolHandler.{cpp|h} as that more clearly describes their
purpose. These classes are protocol handlers, i.e. they handle the ioctl based
mouse and keyboard protocol between the driver and the input_server add-ons.
* Change a lot of stuff to use references instead of pointers where it makes
sense (not necessarily complete yet).
I've tested this successfully on a keyboard with extended keys, a combo device
with a keyboard with extended keys and a mouse, a mouse and a gamepad (that now
doesn't do anything anymore) and found no regressions. However, since there are
a lot of very varied ways how to describe such functions with HID, it's not too
unlikely that some more curiously described devices will now stop working. These
have to be handled case by case and their usages have to be added to the added
to the appropriate handlers (or new handlers have to be written). Please test
and create bug reports (preferrably including the report descriptor that is
written out to /tmp).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41794 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Fix a warning about pci_address::segment potentially being used uninitialized.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41760 a95241bf-73f2-0310-859d-f6bbb57e9c96
check if all of them are assigned with a routing entry. If not, as happens for
PCI add-on cards that ACPI isn't aware of for example, infer the routing by
calculating it from the PCI default routing and going up through the parents
until a matching routing entry is found. Fixes#7520.
* This will also panic (for now) in case there remain unroutable devices and
eventually will prevent the IO-APIC to be used in these cases.
* Enabled some debug output.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41758 a95241bf-73f2-0310-859d-f6bbb57e9c96
actually be a bridge. We always want to check if the device at hand is a PCI
bridge, because in that case we need to recurse further down. We recurse now
even if there is no routing table found for the current bridge, since that
doesn't mean that another bridge, further below, can't have a routing table
again. Possibly fixes#7520.
* Fix the matching function to just check for function 0 to be present instead
of classifying it. Previously if there happened to be a cardbus bridge on
function 0 of a multi function device, the rest of the functions of the device
would've been ignored because we returned early due to the unsupported bridge.
* Only hand down the current bus number when recursing as that makes it clearer
what's going on instead of handing in a partially filled out pci_address
structure and then filling in other parts inside the function.
* Commented out getting the segment number as we don't support it at all.
* Extended debug output.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41726 a95241bf-73f2-0310-859d-f6bbb57e9c96
options we need. This still requires gsed to be installed from ports.
Fixes#7563.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41719 a95241bf-73f2-0310-859d-f6bbb57e9c96
it with our new setting. Otherwise we might end up with an illegal value as the
interrupt threshold which might inhibit interrupt generation alltogether
depending on the controller implementation. This was the case for the ATI/AMD
SBx00 chipsets. Therefore fixes#5551.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41690 a95241bf-73f2-0310-859d-f6bbb57e9c96
interrupt handler every millisecond so it can check for interrupts to handle.
While this is certainly not ideal, it allows EHCI to run (surprisingly well
even) on chipsets that experience interrupt routing issues.
It can be enabled with the safemode setting "ehci_polling on" that can either
be entered using the new advanced debug option entry in the bootloader and/or
in the kernel settings file.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41688 a95241bf-73f2-0310-859d-f6bbb57e9c96
ones that aren't masked. This reversed the enabled interrupts and therefore
disabled the already installed ACPI SCI handler and wrongly enabled the other
interrupt pins instead on PIC to IO-APIC handover. Probably fixes#7525.
+r1alpha3
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41685 a95241bf-73f2-0310-859d-f6bbb57e9c96
generator). There are chipsets (namely ATI/AMD SBx00) where mode 2 doesn't seem
to work (the counter isn't reloaded after the countdown has completed). Mode 0
has the same resolution as mode 2 and, as interrupts are disabled at this point
in booting, is otherwise equivalent as well. I've tested this on 5 machines
available to me and it doesn't seem to have any negative effect.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41683 a95241bf-73f2-0310-859d-f6bbb57e9c96
With that we will usually start right away without having to wait for the
timer to be in the desired state, which removes roughly 116ms of spinning
around for a normal calibration run.
* Read back the value once after programming. The delay this introduces
accounts for the fact that the counter will only start counting down on the
next clock cycle.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41680 a95241bf-73f2-0310-859d-f6bbb57e9c96
bound to prevent spinning forever if we have a very unstable TSC that we
simply can't calibrate. Shouldn't really happen though.
* Set the gate low for channel 2 on exit if we used it as the counter.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41678 a95241bf-73f2-0310-859d-f6bbb57e9c96
copies of it with a call to that.
* Make the variable names more readable and spacing cleanup.
No functional change intended.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41670 a95241bf-73f2-0310-859d-f6bbb57e9c96
timer code has such definitions too, but they're not quite as verbose. Will
eventually make the latter use these too.
* Prepare for possibly using other PIT channels for the calibration. Right now
everything's the same still.
* Add disabled print of the resulting factors.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41669 a95241bf-73f2-0310-859d-f6bbb57e9c96
active config and update the info for the remaining shared link device
entries. Seen on KVM for the (PCI) ACPI interrupt, but happens where there is
no _SRS method for a device (as this one is optional). I find that a bit
strange however as in such a case no _PRS (possible resources) should be
present either, especially not one advertising a config different from the
current one.
* Print the routing table later, after enabling irq routing, so that possible
changes due to such a fallback are included.
* Fix a typo.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41652 a95241bf-73f2-0310-859d-f6bbb57e9c96
used for el-torito booting the CD and anyboot images. Due to the previous lack
of the ACPI module on these, IO-APICs wouldn't get enabled on CD or anyboot
images booted via CD drives.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41551 a95241bf-73f2-0310-859d-f6bbb57e9c96
pin configuration as the configuration only happens after preparing the
configuration where we already need the NMI mask.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41536 a95241bf-73f2-0310-859d-f6bbb57e9c96
the macros. This makes things more readable and also fixes the few remaining
80 char limit violations. No functional change intended.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41533 a95241bf-73f2-0310-859d-f6bbb57e9c96
specify the delivery mode.
* Use that function from ioapic_configure_io_interrupt() and use it when
configuring NMI sources to actually set the entries to NMI.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41531 a95241bf-73f2-0310-859d-f6bbb57e9c96
unavailable for routing so that we don't end up configuring anything on them.
I haven't seen these in actual use though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41530 a95241bf-73f2-0310-859d-f6bbb57e9c96
we use a 64 bit level triggered mask. In practice these don't exist as far as I
know. If we encounter them at a later stage we need to revisit the mask.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41529 a95241bf-73f2-0310-859d-f6bbb57e9c96
at the override entry to trigger the overriden vector so that we don't need
to configure any additional redirections.
* Also configures the polarity and trigger modes found in the override entry.
* When disabling the legacy PIC, retrieve the enabled interrupts and re-enable
then in the IO-APIC. This will for example make the ACPI SCI work that is
installed prior to switching interrupt models. Through the transparent support
for interrupt source overrides it'll also automatically relay from the old to
the new vector.
This should make ACPI interrupts work and should support relocating the ISA PIT
from irq 0 to a different global system interrupt (usually 2) so that it can
still work when IO-APICs are in use.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41528 a95241bf-73f2-0310-859d-f6bbb57e9c96
all reports so far have been positive. We fall back to legacy mode in the cases
where we can't figure out the correct routing.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41527 a95241bf-73f2-0310-859d-f6bbb57e9c96
the case, but maybe there's some BIOSes that disable them? More or less poking
around for #5551, doesn't do any harm in any case.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41514 a95241bf-73f2-0310-859d-f6bbb57e9c96
but maybe it helps on some chipsets to avoid interrupt storms during handover.
* Remove comments that aren't true anymore.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41513 a95241bf-73f2-0310-859d-f6bbb57e9c96
periodic list cache. This workaround is based on similar patches in NetBSD and
Linux and disables the (Advanced) Periodic List Cache on the affected devices
using ATI/AMD specifc registers (as documented in the corresponding AMD
register reference guide).
Also remove a ton of stray whitespace introduced by the isochronous patches.
Please take more care the next time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41512 a95241bf-73f2-0310-859d-f6bbb57e9c96
of bufferSize, corrupting entries when multiple items within the same menu were
checked.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41505 a95241bf-73f2-0310-859d-f6bbb57e9c96
all inter-CPU messaging. The previously described timekeeping is just an extra
function of local APICs. Before, if you'd select "disable local APIC" it would
still blindly write to invalid memory (targetting the non-mapped local APIC) and
then just hang waiting for the other CPUs (that were obviously not responding
to the init sequence that wasn't programmed).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41504 a95241bf-73f2-0310-859d-f6bbb57e9c96
configuring SMP via MP tables. We don't support that method anyhow, but it
should still report correctly.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41503 a95241bf-73f2-0310-859d-f6bbb57e9c96
kernel settings file. As pointed out by Rene, there's otherwise no way to
enable ACPI when the settings file is absent, as there's only a disable switch
in the boot menu.
* Remove MADT dumping as it isn't really implemented. This info can actually be
printed in the IO-APIC code now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41497 a95241bf-73f2-0310-859d-f6bbb57e9c96
at all and, since there can be multiple IO-APICs, we need to do the
enumeration again in the kernel anyway. Also only set ioapic_phys the first
time we encounter an IO-APIC object as it looks cleaner when we arrive at the
first IO-APIC default address.
* Therefore we don't have to worry about already mapped IO-APICs when
enumerating them in the kernel.
* Also remove the mapping function that is now not used anymore.
* We still use the ioapic_phys field of the kernel args to determine whether
there is an IO-APIC at all to avoid needlessly doing the enumeration again.
This fixes multi IO-APIC configurations, because before we would indeed map
the last IO-APIC listed in the MADT, but then in the kernel assumed we mapped
the first one. We'd end up with mapping the last listed IO-APIC twice and the
first IO-APIC never, always programming the last one when we actually targetted
the first one.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41476 a95241bf-73f2-0310-859d-f6bbb57e9c96
written first/last depending on the operation to avoid modifying entries that
are still unmasked or unmasking entries that aren't set up yet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41457 a95241bf-73f2-0310-859d-f6bbb57e9c96
have unexpected side effects once we shift them more than 30 bits.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41452 a95241bf-73f2-0310-859d-f6bbb57e9c96
informational only, but most of these entries actually need to be handled.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41450 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Actually program the IO-APIC ID when configuring the IO-APIC.
* Moved some debug output around.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41447 a95241bf-73f2-0310-859d-f6bbb57e9c96
mark the ISA interrupts as unusable and then use ioapic_is_interrupt_available
to determine if that vector is possibly taken by an IO-APIC. If IO-APICs are
not used, this will simply always return false, leaving all vectors free for
MSI use.
* The msi_init() now has to be done after a potential IO-APIC init, so it is now
done after ioapic_init() instead of inside apic_init().
* Add apic_disable_local_ints() to clear the local ints on the local APIC once
we are in APIC mode (i.e. the IO-APIC is set up and we don't need the external
routing anymore).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41445 a95241bf-73f2-0310-859d-f6bbb57e9c96
multiple IO-APICs every IO-APIC was initialized to the standard mapping before.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41434 a95241bf-73f2-0310-859d-f6bbb57e9c96
we got. If the routing is possible with the limited IO-APICs everythings good,
if not we will simply fail at the routing preparation stage where we can still
fall back gracefully. This makes things a bit more error resilient.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41432 a95241bf-73f2-0310-859d-f6bbb57e9c96
number. This accounts for possible gaps in the IO-APIC GSI mappings. Since most
of the time there will be only a single IO-APIC the extra overhead is relatively
small.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41431 a95241bf-73f2-0310-859d-f6bbb57e9c96
the same for the first IO-APIC in the bootloader (also using the MADT), but
here we can use the ACPI module. The enumerated IO-APICs are added to the list
and should be usable from there on. For lack of hardware I wasn't able to really
test that though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41430 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added some TODOs regarding register write order that might have bad side
effects.
* Some cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41428 a95241bf-73f2-0310-859d-f6bbb57e9c96
functional change intended.
* Use an appropriately sized sLevelTriggeredInterrupts for each controller type.
This also fixes an out of bound access for IO-APICs with more than 32 entries
and also returns the right mode in such cases.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41426 a95241bf-73f2-0310-859d-f6bbb57e9c96
already. This allows to detect invalid settings before starting to enable
IRQ routing and therefore allows to gracefully fall back to PIC mode on error.
* Actually read the number of IO-APIC entries before using that number in
routing preparation.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41416 a95241bf-73f2-0310-859d-f6bbb57e9c96
gracefully in such cases (resulting in the IO-APIC not being used).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41415 a95241bf-73f2-0310-859d-f6bbb57e9c96
Michael Lotz