pci: extend MSI interrupt vector number to 32 bits

Also increase MSI message data size to 32 bits according to PCIe spec.

Remove 0xff check for MSI interrupts because it is potentially valid
interrupt vector number. Reject 0xff only for legacy pin interrupts.

- MSI-X supports up to 2048 interrupts per device that do not fit to
`uint8`.

- Non-x86 systems may use separate interrupt vector ranges for
hard-wired interrupts and MSI interrupts so `uint8` is not enough to
represent all of them.

Change-Id: Iaf9ffb197ec23db0f97ffe3ea756d28d7bfc8705
Reviewed-on: https://review.haiku-os.org/c/haiku/+/7433
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>

USB: Make "buffer error" reporting consistent.

This error means that the controller failed to read or write data
to/from system memory fast enough. As a result, we should report it
as READ_ERROR/WRITE_ERROR, rather than something that sounds like
an error with the device itself.

Then, make BABBLE reporting consistent: this is a data overrun/underrun
of the device itself, not the buffers passed to the controller.
This then leaves B_DEV_FIFO_OVERRUN/UNDERRUN unused, and thus usable
for reporting ring overrun/underrun on isochronous transfers.

The API documentation already described DATA_OVERRUN/UNDERRUN basically
like they were babble errors, while FIFO_OVERRUN/UNDERRUN are
currently described as "internal errors" at present. No driver actually
checked for these.

OHCI: Remove unused variable following previous commit.

USB: Initialize packet descriptors inside the stack.

Allows deletion of generic code from bus drivers.

USB: Support physical-vector bulk requests.

Introduce a new utility method, "generic_memcpy", which takes
generic_addr_t plus indications of whether these specify virtual or
physical addresses (and potentially user addresess) and calls the
appropriate memcpy variant depending.

All bus drivers adjusted to support this at once. We don't actually
take advantage of the physical addresses in any way (yet), as USB
controllers have some pretty specific requirements that would have
to be carefully validated to use these directly.

All bus drivers tested and confirmed to still be working.

Change-Id: I66326667e148091147bb2b3d0843a26fb7e5bda6
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6479
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

USB: Use generic_io_vec in Transfer.

Prepares the way for actual physical-address request support,
which is not yet wired up.

Change-Id: I1b2d04a31e334a79b8361280fa0f3a5fbdb43d2b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6478
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

bus & drivers: drop PCI_x86

Change-Id: I494deaf24a4793a5e0fe9fa46ecdce32f65e616a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6226
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>

kernel: Drop non-standard GNU inline assignment syntax

* We needed this previously due to our gcc2 compiled kernel.
* Now that our kernel is always latest gcc, we can move to the
c++20 syntax for inline assignment.
* Improves compatibility with clang, less GNU-specific stuff

Change-Id: Ib7272a0a52554a31e9a0e788fd3f031db9049795
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5898
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>

usb: fix references on the module B_PCI_X86_MODULE_NAME

the dependency is optional, only the device modules depend on it.
should fix #18075

Change-Id: Ifddb9814ff5ff59ee7d80d06b14ba9fded04cd56
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5822
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

usb: register device nodes for hubs and devices

the device identifier is the USBID. no actual interface to be used by drivers.

Change-Id: I9fd1fdab0d02f1b82380c619b39bc3a12a9e2333
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5736
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
Reviewed-by: X512 <danger_mail@list.ru>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

usb: remove debug leftover

Change-Id: Idba6eaecb2c5b5773cd61794ac62fac1f7e3c1f9
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5715
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

ohci: switch to the new driver API

Change-Id: I065743b5ced8b75f9576b62421525824698a2245
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5710
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

USB bus drivers: Add missing iterator increments.

Should fix boot hang regressions introduced in hrev55455
in the fix for #17275.

USB bus drivers: Return error when combining other transfers with a fragmented one.

The drivers do not support this properly at present, they would
run the other transfers interspersed with fragments from the fragemented one,
which is obviously the wrong thing to do.

No USB device drivers seem to do this at present (it would cause data
corruption if they had.)

Fixes #17275.

USB: Rename Transfer::VectorLength to Transfer::FragmentLength.

It did not return the length of the vectors, but only the current
fragment's worth of vectors. It also modified the fFragmented flag,
which really should have been set in SetVectors in the first place.

As everything seemed to call IsFragmented after VectorLength,
this is not a behavioral change.

busses/usb: Print the PCI ID of the device with an invalid IRQ.

busses/usb: Handle getting/putting the PCI modules properly.

Previously it was possible that we put() them twice, leading to
assert-failure panics as their ref-counts would go below 0.
Now we get() them once for every bus object that we create, as
the destructors put() them.

Should fix the panic in #15004.

busses/usb: Don't reject devices with "invalid" IRQs immediately.

We may be able to allocate an MSI for them. This was done for XHCI
already in hrev52742.

Probably fixes #15004.

busses/usb: Move ::AddTo to be above the class constructors.

It is called before them (and is responsible for calling them),
so having it in the middle of the file does not make a lot of sense.

Already done for XHCI. Only one functional change -- removing the
set_dprintf_enabled call. Drivers probably shouldn't spuriously
re-enable that if it was specifically disabled.

USB: Properly assign IDs to USB Bus Managers (HCI drivers).

Previously they would just get -1, as the BusManager class
would request their ID before they had been added to the Stack.
Now we add them to the stack inside ::Start(), rather than letting
the individual drivers do that just after ::Start(), and then assign
the ID there directly.

drivers: remove some non-haiku support code.

Change-Id: Ic55bb4832adabeb807f763c87ad938e74fb3a97d
Reviewed-on: https://review.haiku-os.org/520
Reviewed-by: Axel Dörfler <axeld@pinc-software.de>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>

usb: map registers or allocate area for the kernel only.

ohci: Fix x86_64 usb tracing

USB OHCI: Fix finalizing of STALL-ed control request transfers

In case control transfer is finished with STALL answer from device only
the setup TD is retired into the Done Queue but other TDs (data and
status) are left in the ED queue. HC set the endpoint halted - so we
process such endpoint regardless of it's non-empty queue.

PS: cleanup some trailing spaces.

USB OHCI: Support for isochronous transfers

Finishing and refactoring the draft, initially implemented during April-May 2012

NOTE: startingFrameNumber returned to device contains the number of the
next free frame right after the last packed of submitted data. For more
details please Look into corresponding [haiku-development] discussion
started at 03 Jul 2013.

Partially fixes #1045.

OHCI USB: disable MSI on destruction

* also remove interrupt handler and put the PCI x86 module.

USB OHCI: Fix ownership handover from SMM to OS.

* Disabling all interrupts prior to ownership handover from SMM to OS
can prevent propper OHCI and PS/2 functionality as described in
#8987 and #8984. In that case SMM does not respond to the ownership
change request. On the other hand not disabling the interrupts can
lead to interrupt storms (discussed in #8085) since no interrupt
handler is installed at that moment. As suggested by mmlr this patch
attempts to address both issues by keeping the ownership change
request interrupt enabled.

* Removed an unnecessary reset upon non-responding SMM for now,
since we reset a few lines later anyway and added TODOs respectively.
That should safe a bit boot time.

Signed-off-by: Michael Lotz <mmlr@mlotz.ch>

Add MSI support to OHCI.

It looks like VirtualBox assumes MSIs when emulating a 64bit system so
this gets OHCI working there. Shouldn't harm if it's used anywhere else
either.

Fix compilation of OHCI on x86.

Fixes were needed to go with the USB stack changes. Haven't tested it
yet because I don't have a a machine with OHCI, but I doubt there
will be any problems.

* Disable interrupts in OHCI before doing BIOS handover. Shouldn't matter much
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

added traces for unimplemented isochronous transfers for ehci and ohci


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41194 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed physical_entry::{address,size} to phys_{addr,size}_t and changed
map_physical_memory()'s physicalAddress parameter type from void* to
phys_addr_t. This breaks source compatibility, but -- as long as
phys_{addr,size}_t remain 32 bit wide -- keeps binary compatibility with
BeOS.
* Adjusted all code using the affected interfaces (Oh what fun!). Added a few
TODOs in places where the wrong types (e.g. void* for physical addresses
are used). Looks like quite a few drivers aren't 64 bit safe and others
will break with PAE.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36960 a95241bf-73f2-0310-859d-f6bbb57e9c96

We need to guarantee that when cancel returns that the callback won't be called
anymore. It could happen that the transfer was already in the process of being
finished, so wasn't in the list of pending transfers anymore. Cancel would then
return even though the callback wasn't called yet. This could lead to a callback
being called after a driver was already unloaded (even after it cleaned up the
pipes it used).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33239 a95241bf-73f2-0310-859d-f6bbb57e9c96

When submitting transfers there is a race condition between adding the pending
transfer to the transfer list and scheduling the transfer descriptors on the
controller by switching the endpoint tail. Since we only check that the endpoint
head is equal to the endpoint tail to conclude that there are no active
transfers, we need to ensure that we don't accidently do that check before the
transfer descriptors are scheduled. Otherwise we could happen to processed a not
yet started transfer and finishing it before an actual transfer has taken place.
This would then lead to 0 byte transfers and toggle mismatches. To fix this we
now protect the transfer addition and tail switching as well as the check in the
finisher with a per-endpoint mutex. Note that we allocate the lock on the heap
and only store the pointer in the endpoint structure as this one is allocated
from the precious physical memory pool. Could fix #4067.
Also switched the Jamfile to UsePrivateKernelHeaders (for the MutexLocker) and
removed BeOS compatibility.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33129 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix subtle but important error in transfer length calculation. That calculation
does calculate the remaining bytes, not the transferred ones, so obviously that
number has to be subtracted from the expected buffer size! Albeit severe, this
wasn't hit that often, because this calculation will only be used when the
transferred size doesn't match the handed in buffer size exactly. So things like
HID or mass storage, that mostly know the exact transfer size weren't affected.
But it could very well explain #4101 where the size isn't previously know.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32680 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix typo.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32552 a95241bf-73f2-0310-859d-f6bbb57e9c96

We cannot use the condition codes to determine if a transfer descriptor is
processed. The controller may initialize it to any value when starting to
process the descriptor. If a controller did that it was possible that we thought
a transfer was already done even if it actually just started. Transfers would
then return that weren't processed, returning uninitialized buffers.
Instead of relying on the condition codes we now check that the head and tail
pointers are the same. This guarantees that all transfer descriptors of the
endpoint at hand are completed.
Reverted r32534 again as this one fixes the problem for real. The same things
that were mentioned there could happen here essentially, so in the worst case
the device or controller could stall because of freeing in-use structures.
Fixes #4067.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32551 a95241bf-73f2-0310-859d-f6bbb57e9c96

When adding a transfer, the current tail descriptor of the endpoint will be
reused and overwritten by the first descriptor of the new transfer and the first
descriptor will become the new tail. We anticipate this situation in
_AddPendingTransfer() and set the first_descriptor of the transfer data to the
tail already. Since the tail was pretty much cleared to zero, this introduced
a race condition. After adding the pending transfer it can already be found
in the finisher thread. If this happened before actually switching the tail
and first descriptor it would find a descriptor with a condition of 0, meaning
"No Error" and would process the transfer incorrectly. Depending on the count
of descriptors and the timing of the switch taking place this could have
resulted in aborted transfers with actual length 0 or with the correct actual
length but invalid data. In the very worst case it could have freed things still
in use by the controller, resulting in all sorts of device errors. Sadly it
doesn't fix #4067.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32534 a95241bf-73f2-0310-859d-f6bbb57e9c96

That's actually not how it works. OHCI does the dirty work of splitting up the
transfer descriptors into individual packets. The descriptors are more a
logical thing. As such we do not generate one descriptor per packet but one
for each block of two pages at max. Therefore we only set the initial toggle bit
to the one we stored on the first descriptor and let OHCI use the carry bit for
subsequet descriptors. That error would only be visible on transfers above two
pages in size though.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32532 a95241bf-73f2-0310-859d-f6bbb57e9c96

Even though the specs say the toggle carry field contains the value of the last
retired descriptor, it actually contains the value that is to be used next.
Confirmed that by cross-referencing FreeBSD.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32531 a95241bf-73f2-0310-859d-f6bbb57e9c96

A try at actual data toggle handling. Since we do not properly ensure that the
toggle carry stays valid, we better make sure we set it correctly. Additionally
even if the toggle carry wasn't clobbered, there are conditions the toggle is
reset without the controller being able to notice (as in clearing a halt state).
Since I don't have any OHCI hardware this is untested, but should fix bug #4067
if correct.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32510 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Add the kernel debugger extensions only for the haiku target.
* Acknowledge non-enabled interrupts in UHCI as done in EHCI as well.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30063 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Rework the USB tracing mechanism. Cleaned it up and make it more convenient to
use. It will now print out the usb_ids of the objects that generate the trace
messages. These IDs are unique compared to the device address used previously,
because device addresses are per bus while usb_ids are global. This makes
trace output from devices across multiple controllers distinguishable.
* Some cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29002 a95241bf-73f2-0310-859d-f6bbb57e9c96

When calculating the actual length for requests, we don't want to include the
8 byte request header. It was included in the case of outgoing request
transfers before. In fact we always just want the length of the data phase,
which for non-requests just happens to be the whole transfer.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26508 a95241bf-73f2-0310-859d-f6bbb57e9c96

Remove disabling and clearing interrupts again as it doesn't turn out to be a problem.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25935 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fix the check after waiting for owner change to happen and don't reset if the
handover indeed worked. Doesn't really matter though, as we reset anyway,
just not twice in the success case.
* Disable all interrupts and clear any pending active interrupts just after
taking over the controller. This might fix bug #2361.
* Also trace output when the ownership change is successful.
* Don't disable interrupts during reset.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25915 a95241bf-73f2-0310-859d-f6bbb57e9c96

When canceling transfers, do not call the callbacks with the lock held. This
prevents deadlocks in cases where a new transfer is scheduled from within the
cancel callback. This is an edge case, as generally you don't want to schedule
anything when explicitly canceling transfers, but there are a valid use cases
when you cancel because of a timeout and then have to reset the device for
example. Note that this hides bug #2353, where the cancel case is probably
not handled correctly.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25861 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added optional spinlock contention measurement feature. Enabled when
B_DEBUG_SPINLOCK_CONTENTION is defined to 1. It typedefs spinlock to a
structure (thus breaking BeOS binary compatibility), containing a
counter which is incremented whenever a thread has to wait for the
spinlock.
* Added macros for spinlock initialization and access and changed
code using spinlocks accordingly. This breaks compilation for BeOS --
the macros should be defined in the respective compatibility wrappers.
* Added generic syscall to get the spinlock counters for the thread and
the team spinlocks.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25752 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Replaced all instances of benaphores in the kernel code by mutexes.
* Removed kernel benaphores.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25690 a95241bf-73f2-0310-859d-f6bbb57e9c96

Unconditionally report initialization success for OHCI too.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25614 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Disable the root hub change interrupt when it gets triggered. As the change
status stays on until the port is reset, this causes the interrupt to be
constantly retriggered on some controllers.
* Added TODO that we need to reenable it once we want to take advantage of
status change notifications instead of polling. It should be reenabled when
clearing roothub status changes in ClearPortFeature().

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25578 a95241bf-73f2-0310-859d-f6bbb57e9c96

Enable interrupts only after installing the interrupt handler.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25571 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Correct removal of failed transfers by applying the address mask. As the head
pointer also contains the toggle carry and the halt bit (which was obviously
always set as the endpoint had an error) the address comparison failed and
the head was dead in the water with all further transfers timing out until
a subsequet cancel cleaned the mess up again.
* Add the OHCI host controller driver to the image as it should now be fully
functional except for isochronous transfers.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25565 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implement removal of failed transfers. The head pointer will be moved past
the failed transfer if the transfer in question is found to be the offending
one. Probably not quite enough and still untested.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25558 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Build the interrupt tree the same (imperfect) way as in EHCI to conserve
on resources but still satisfy the scheduling policy. This works well, but
effectifly limits the bandwidth available as the interrupts aren't spread
across different frames.
* Implement finding the interrupt endpoint head for a certain interval.
* Deactivate endpoints that are going away until removing them is implemented.
* Allow short packets for all data phases.
* Disable tracing by default.

There you go, interrupt transfers should work as well. Should enable mice,
keyboards, hubs and other devices that use interrupt pipes. Note that it's
still easily possible that a single failed request will hang an endpoint and
prevent a device from working.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25554 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Handle the data == first and last == first case as this can occure for
single td bulk transfers.
* Implement SubmitTransfer() for bulk transfers (interrupts will use the exact
same code path, but as their endpoints aren't yet setup, interrupt transfers
won't work).
* Handle the cancel case when finishing transfers. The descriptors of the
canceled transfers cannot be accessed by the controller so they can be freed.

Bulk transfers should work now, so devices only using control and bulk transfers
should too (anything using usb_disk for example). Note though that a transfer
error will cause the whole thing to fail miserably as the error case is not
yet handled correctly (failed descriptors aren't removed from the endpoint).
Therefore I suggest not testing with the memory stick with that important
presentation you haven't stored anywhere else...

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25549 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Completely rework how transfer finishing is implemented. It is now using
pretty much the same strategy as UHCI. This does not leverage the done queue
the hardware provides us, but as it saves a lot of other overhead and allows
the structures to be smaller I think it is overall worth it.
* Removed now unnecessary stuff from the transfer descriptor struct.
* Directly acknowledge the done head interrupt as we do not use it either.
* Activate control endpoints by removing the skip bit when a transfer is scheduled.
* Correct the way the first descriptor is determined (the current tail becomes
the first descriptor while the current first descriptor becomes the tail).
* Remove the head_logical_descriptor field from the endpoint structure, as this
would simply get out of sync with the first td with no real way of updating.
* Stub out RemoveTransferFromEndpoint() as this needs reworking.
* Correct CreateDescriptorChain() signature and implement the method.
* Add ReadDescriptorChain() and ReadActualLength() to process finished transfers.
* Prepare for actual data transfers by renaming/regrouping transfer types.
* Fully clear out the new tail when switching (for the sake of it)
* Add some helpers and definitions to the hardware header.
* Add debugging facilities.

With all this transfers can now actually be scheduled and they are processed
when done. This brings control transfers to a usable state so setting the
device address and reading out/setting the configuration works. That means you
can now view your device using usb_dev_info, but since any other transfer type
besides control transfers isn't implemented yet the device will most probably
not yet do anything useful.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25546 a95241bf-73f2-0310-859d-f6bbb57e9c96

Correctly clear an endpoint on cancel. Adjusted some comments/output.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25543 a95241bf-73f2-0310-859d-f6bbb57e9c96

Use a dedicated endpoint lock for endpoint manipulation. Fixes the guaranteed
deadlock with the BusManager on setup of the default pipes (for addressing).

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25542 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implement "tail switching" so that a new transfer descriptor chain can be
appended to an endpoint descriptor without locking/disabling the endpoint.
* Correct the direction of the data phase of a control transfer.
* Some minor cleanup.

Control requests without data phase (set address for example) might actually
work now, but this is still untested.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25541 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Complete the roothub get/set/clear port feature functions
* Reorder to match the usual order of the states/commands
* Fix some debug output and make it more informative

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25540 a95241bf-73f2-0310-859d-f6bbb57e9c96

Acknowledge OHCI interrupts.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25539 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Make the operational register memory a uint8 * instead of a uint32 * so the
driver has at least a chance of working (it previously always used wrong
offsets for register access).
* Remove the hash approach for now (I'm going to explore a few other ways of
doing that first).
* Reorder some stuff and check for errors in some more places.
* More cleanup (mostly whitespace again).

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25538 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Getting familiar with the existing code
* Cleanup (whitespace, nameing, code style)
* Move around methods so they match the header order
* Fix some obvious stuff
* Initialize all members
* Sync roothub code (this will be reworked to a common roothub)
* Only minor functional changes (to the worse for now)

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25527 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix building usb busses with tracing on.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24880 a95241bf-73f2-0310-859d-f6bbb57e9c96

Change just host_controller_info to usb_host_controller_info for the sake of
clearity as suggested by Philippe Houdoin once and brought to my attention by
Slavatore. As this definition is completely private to the USB stack it would
not have caused any conflicts though.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24828 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added necessary dummy descriptor to every endpoint created
* Implemented _RemoveTransferFromEndpoint
* Reworked _AppendChainDescriptorsToEndpoint as now every endpoint has (must have) a dummy descriptor



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23741 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented private method _AppendChainDescriptorsToEndpoint in order to avoid code duplication (lock needs to be added)
* Fixed _SubmitControlRequest in order to use _AppendChainDescriptorsToEndpoint


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23734 a95241bf-73f2-0310-859d-f6bbb57e9c96

* More work on the finisher thread
* Reworked SubmitTrasfer method
* Implemented CancelQueuedTransfers
* Minor clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23722 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Removed cast: not needed anymore since Interval() has been moved to the base class (Pipe)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23704 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented _AddPendingTransfer
* renamed trasfer_data memeber top to first_descriptor
* added data_descriptor member to transfer_data (useful for control transfer)



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23702 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed fFinishTransfer to fLastTransfer (I don't even know why I called it
so in the first place)
* Renamed local variable descriptor to current in the finisher thread
* Using next_done_descriptor instead of next_logical_descriptor, as the latest
is used to keep track of all descriptors that belongs to the transfer
* Added _UnlinkTransfer private method
* Implemented _FreeDescriptorChain
* Moved endpoint pointer from the ohci_general_td structur to the transfer_data_s
structure, ad the ohci_general_td has already a transfer_data_s pointer which can
be used to the endpoint pointer reducing ridundancy.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23701 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added definition for the hash related methods (not implemented)
* Continue working in the _FinishTransfer thread
* added next_done_descriptor in ohci_[general|isochronous]_td structure in order to handle
collisions.
* added next_logical_descriptor to ohci_isochronous_td structure



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23594 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Started working on the finisher thread
* Added some TRACE calls in the _InterruptHandler


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23266 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Partially implemented interrupt handler routine


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23175 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed init part of the controller


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22980 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Remove SubmitAsyncTransfer and SubmitPeriodicTransfer
* Added _SubmitControlRequest (almost implemented), _LinkDescriptors, _WriteDescriptorChain (basically copied from EHCI, and UHCI)
* Renamed ohci_general_descriptor to ohci_general_td, and ohci_isochronous_descriptor to ohci_isochronous_td
* Finished _CreateGeneralDescriptor and _FreeGeneralDescriptor
* Added buffer_size member to ohci_general_td instead and removed last_logical_byte_address as it looked unnecessary
* minor clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22968 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented _InsertEndpointForPipe
* usual clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22938 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Replaced snooze with spin (Thanks Axel)
* Fixing TRACE messages
* minor clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22935 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Moved disable_interrupt right before resetting the controller
* Reworked on NotifyPipeChange
* Added _RemoveEndpointForPipe (not implemented)



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22934 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Completed SubmitTransfer
* Added SubmitAsyncTransfer and SubmitPeriodicTransfer (not implemented)
* Removed unecessary spinlock
* Wrapped some lines to follow coding guidelines



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22932 a95241bf-73f2-0310-859d-f6bbb57e9c96

Completely redesign the USB explore process. Replaces the scary race conditions of the previous locking mechanism and simplifies handling of device changes by a more centralized approach.
Changes are now collected during explore and notifications as well as rescans are done at once. Through this a driver is also not rescanned multiple times anymore.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22929 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed some variables
* Reworking the interrupts endpoints tree parts
* Added spin_locker, semaphore, finisher thread and interrupt handler (not implemented)
* Made fInterruptEndpoints allocation dynamic instead of static
* Fixed Start method (it should be correct now)




git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22927 a95241bf-73f2-0310-859d-f6bbb57e9c96

* ups :)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22875 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Placed underscore to private methods like ReadReg and WriteReg
* Rewrote _AllocateEndpoint
* More work in the constructor


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22873 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Rewrote part of the constructor: most of the code is ported from FreeBSD
* Usual clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22625 a95241bf-73f2-0310-859d-f6bbb57e9c96

* License changed: I've successfully contacted Jan-Rixt Van Hoye, and he gave me his bless :)
* Usual clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22555 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Revert Copyright


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22527 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Clean up (removing unneeded comment)
* Renaming, fixing coding guidelines
* Re-writing AddTo method
* Removing ohci_software.h file (its content has been moved to ohci.h) but it will probably be deleted in the future.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22522 a95241bf-73f2-0310-859d-f6bbb57e9c96

More tracing in ohci. Hmm, existing isn't totally consistent (prefix, ...). Oh well.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21242 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fixed wrong usage of object types in OHCI. The types are bitmasks, not values.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19900 a95241bf-73f2-0310-859d-f6bbb57e9c96

Implemented endpoint creation for Pipes.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18986 a95241bf-73f2-0310-859d-f6bbb57e9c96

Now the ports are powered on and seem to work...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18928 a95241bf-73f2-0310-859d-f6bbb57e9c96

Implemented most of the roothub.

There remains only one issue: reading the Port registers seem to return a bogus value (0). I need to find out what's causing it.



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18918 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented usable thread synchronisation for UHCI and EHCI.
* Cleaned interrupt handlers for both busses, too.
* Fixed memory leak of the queue head stray descriptors in EHCI.
* Fixed warnings about assigning NULL to the module flags field.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18857 a95241bf-73f2-0310-859d-f6bbb57e9c96

Wrote the initialisation routine of the host controller
* Started cleaning up the style (WIP)
* I'll slowly phase out ohci_software.h, since it is BSD legacy
* Cleanup other BSD inspired naming. I like full names.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18723 a95241bf-73f2-0310-859d-f6bbb57e9c96

replaced with PCI.h definitions, added PCI_usb_ehci


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18704 a95241bf-73f2-0310-859d-f6bbb57e9c96

First commit to show that work on the ohci bus has restarted.
It doesn't do anything yet, but it allocates the io registers properly and
now the values that are returned when reading it make sense.

More to come...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18673 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Adding skeleton EHCI driver. Some parts filled from the specs and the UHCI driver.
* Corrected some errors I found while reading UHCI.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18626 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added USB API Version 3. It's put into USB3.h while the v2 API resides in USB2.h. USB.h just includes USB2.h for now.
* Rewrote both headers on the way.
* The usb module now exports both, the v2 and v3 module_info.
* Changed the internals of the USB Stack to give out usb_ids instead of opaque handles to internal classes.
* Cleaned up some more of the Stack by moving members into other classes and removing unused stuff.
* Updated the usb_raw driver from v2 to v3 API.

Since both usb_hid (which still uses the v2 API) and usb_raw (which now uses the v3 API) work, I'd call it a success ;-)

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18539 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fixed building of the UHCI and OHCI modules [reported by AndyBe].


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15914 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fixed the bugs I introduced earlier. original code was ok.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14244 a95241bf-73f2-0310-859d-f6bbb57e9c96

properly mask off bits


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14232 a95241bf-73f2-0310-859d-f6bbb57e9c96

it's not allowed to spin 250ms


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14231 a95241bf-73f2-0310-859d-f6bbb57e9c96

Quick merge of my development branch with mainline (for SVN switch).
The OHCI stuff was done by Jixt (thanks)./installusb
All is still very much work in progress

Patches applied:

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--base-0
tag of Niels.Reedijk@gmail.com--haiku-2005/usb-busses--mainline--0.1--base-0

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-1
Set the PORT_STATUS_LOW_SPEED flag correctly (using the register bitflag for low speed devices, instead of the proper lowspeed flag)

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-2
Clean up transfers when they're done

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-3


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-4


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-5


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-6


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-7
Remove obsolete ehci.c

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-8


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-9
Completed memory allocation + initialization

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-10
First attempt at performing control transfers - failed miserably

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-11
Trying to read a 16bit value into a 8bit integer, stupid

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-12
nicer way of getting the base io address

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-13
Take over control from BIOS

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-14
Revert change for disabling USB BIOS support


git-svn-id: file:///srv/svn/repos/haiku/trunk/current@11743 a95241bf-73f2-0310-859d-f6bbb57e9c96

USB OHCI: Fix finalizing of STALL-ed control request transfers

In case control transfer is finished with STALL answer from device only
the setup TD is retired into the Done Queue but other TDs (data and
status) are left in the ED queue. HC set the endpoint halted - so we
process such endpoint regardless of it's non-empty queue.

PS: cleanup some trailing spaces.

USB OHCI: Support for isochronous transfers

Finishing and refactoring the draft, initially implemented during April-May 2012

NOTE: startingFrameNumber returned to device contains the number of the
next free frame right after the last packed of submitted data. For more
details please Look into corresponding [haiku-development] discussion
started at 03 Jul 2013.

Partially fixes #1045.

OHCI USB: disable MSI on destruction

* also remove interrupt handler and put the PCI x86 module.

USB OHCI: Fix ownership handover from SMM to OS.

* Disabling all interrupts prior to ownership handover from SMM to OS
can prevent propper OHCI and PS/2 functionality as described in
#8987 and #8984. In that case SMM does not respond to the ownership
change request. On the other hand not disabling the interrupts can
lead to interrupt storms (discussed in #8085) since no interrupt
handler is installed at that moment. As suggested by mmlr this patch
attempts to address both issues by keeping the ownership change
request interrupt enabled.

* Removed an unnecessary reset upon non-responding SMM for now,
since we reset a few lines later anyway and added TODOs respectively.
That should safe a bit boot time.

Signed-off-by: Michael Lotz <mmlr@mlotz.ch>

Add MSI support to OHCI.

It looks like VirtualBox assumes MSIs when emulating a 64bit system so
this gets OHCI working there. Shouldn't harm if it's used anywhere else
either.

Fix compilation of OHCI on x86.

Fixes were needed to go with the USB stack changes. Haven't tested it
yet because I don't have a a machine with OHCI, but I doubt there
will be any problems.

* Disable interrupts in OHCI before doing BIOS handover. Shouldn't matter much
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

added traces for unimplemented isochronous transfers for ehci and ohci


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41194 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed physical_entry::{address,size} to phys_{addr,size}_t and changed
map_physical_memory()'s physicalAddress parameter type from void* to
phys_addr_t. This breaks source compatibility, but -- as long as
phys_{addr,size}_t remain 32 bit wide -- keeps binary compatibility with
BeOS.
* Adjusted all code using the affected interfaces (Oh what fun!). Added a few
TODOs in places where the wrong types (e.g. void* for physical addresses
are used). Looks like quite a few drivers aren't 64 bit safe and others
will break with PAE.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36960 a95241bf-73f2-0310-859d-f6bbb57e9c96

We need to guarantee that when cancel returns that the callback won't be called
anymore. It could happen that the transfer was already in the process of being
finished, so wasn't in the list of pending transfers anymore. Cancel would then
return even though the callback wasn't called yet. This could lead to a callback
being called after a driver was already unloaded (even after it cleaned up the
pipes it used).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33239 a95241bf-73f2-0310-859d-f6bbb57e9c96

When submitting transfers there is a race condition between adding the pending
transfer to the transfer list and scheduling the transfer descriptors on the
controller by switching the endpoint tail. Since we only check that the endpoint
head is equal to the endpoint tail to conclude that there are no active
transfers, we need to ensure that we don't accidently do that check before the
transfer descriptors are scheduled. Otherwise we could happen to processed a not
yet started transfer and finishing it before an actual transfer has taken place.
This would then lead to 0 byte transfers and toggle mismatches. To fix this we
now protect the transfer addition and tail switching as well as the check in the
finisher with a per-endpoint mutex. Note that we allocate the lock on the heap
and only store the pointer in the endpoint structure as this one is allocated
from the precious physical memory pool. Could fix #4067.
Also switched the Jamfile to UsePrivateKernelHeaders (for the MutexLocker) and
removed BeOS compatibility.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33129 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix subtle but important error in transfer length calculation. That calculation
does calculate the remaining bytes, not the transferred ones, so obviously that
number has to be subtracted from the expected buffer size! Albeit severe, this
wasn't hit that often, because this calculation will only be used when the
transferred size doesn't match the handed in buffer size exactly. So things like
HID or mass storage, that mostly know the exact transfer size weren't affected.
But it could very well explain #4101 where the size isn't previously know.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32680 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix typo.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32552 a95241bf-73f2-0310-859d-f6bbb57e9c96

We cannot use the condition codes to determine if a transfer descriptor is
processed. The controller may initialize it to any value when starting to
process the descriptor. If a controller did that it was possible that we thought
a transfer was already done even if it actually just started. Transfers would
then return that weren't processed, returning uninitialized buffers.
Instead of relying on the condition codes we now check that the head and tail
pointers are the same. This guarantees that all transfer descriptors of the
endpoint at hand are completed.
Reverted r32534 again as this one fixes the problem for real. The same things
that were mentioned there could happen here essentially, so in the worst case
the device or controller could stall because of freeing in-use structures.
Fixes #4067.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32551 a95241bf-73f2-0310-859d-f6bbb57e9c96

When adding a transfer, the current tail descriptor of the endpoint will be
reused and overwritten by the first descriptor of the new transfer and the first
descriptor will become the new tail. We anticipate this situation in
_AddPendingTransfer() and set the first_descriptor of the transfer data to the
tail already. Since the tail was pretty much cleared to zero, this introduced
a race condition. After adding the pending transfer it can already be found
in the finisher thread. If this happened before actually switching the tail
and first descriptor it would find a descriptor with a condition of 0, meaning
"No Error" and would process the transfer incorrectly. Depending on the count
of descriptors and the timing of the switch taking place this could have
resulted in aborted transfers with actual length 0 or with the correct actual
length but invalid data. In the very worst case it could have freed things still
in use by the controller, resulting in all sorts of device errors. Sadly it
doesn't fix #4067.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32534 a95241bf-73f2-0310-859d-f6bbb57e9c96

That's actually not how it works. OHCI does the dirty work of splitting up the
transfer descriptors into individual packets. The descriptors are more a
logical thing. As such we do not generate one descriptor per packet but one
for each block of two pages at max. Therefore we only set the initial toggle bit
to the one we stored on the first descriptor and let OHCI use the carry bit for
subsequet descriptors. That error would only be visible on transfers above two
pages in size though.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32532 a95241bf-73f2-0310-859d-f6bbb57e9c96

Even though the specs say the toggle carry field contains the value of the last
retired descriptor, it actually contains the value that is to be used next.
Confirmed that by cross-referencing FreeBSD.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32531 a95241bf-73f2-0310-859d-f6bbb57e9c96

A try at actual data toggle handling. Since we do not properly ensure that the
toggle carry stays valid, we better make sure we set it correctly. Additionally
even if the toggle carry wasn't clobbered, there are conditions the toggle is
reset without the controller being able to notice (as in clearing a halt state).
Since I don't have any OHCI hardware this is untested, but should fix bug #4067
if correct.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32510 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Add the kernel debugger extensions only for the haiku target.
* Acknowledge non-enabled interrupts in UHCI as done in EHCI as well.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30063 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Rework the USB tracing mechanism. Cleaned it up and make it more convenient to
use. It will now print out the usb_ids of the objects that generate the trace
messages. These IDs are unique compared to the device address used previously,
because device addresses are per bus while usb_ids are global. This makes
trace output from devices across multiple controllers distinguishable.
* Some cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29002 a95241bf-73f2-0310-859d-f6bbb57e9c96

When calculating the actual length for requests, we don't want to include the
8 byte request header. It was included in the case of outgoing request
transfers before. In fact we always just want the length of the data phase,
which for non-requests just happens to be the whole transfer.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26508 a95241bf-73f2-0310-859d-f6bbb57e9c96

Remove disabling and clearing interrupts again as it doesn't turn out to be a problem.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25935 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fix the check after waiting for owner change to happen and don't reset if the
handover indeed worked. Doesn't really matter though, as we reset anyway,
just not twice in the success case.
* Disable all interrupts and clear any pending active interrupts just after
taking over the controller. This might fix bug #2361.
* Also trace output when the ownership change is successful.
* Don't disable interrupts during reset.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25915 a95241bf-73f2-0310-859d-f6bbb57e9c96

When canceling transfers, do not call the callbacks with the lock held. This
prevents deadlocks in cases where a new transfer is scheduled from within the
cancel callback. This is an edge case, as generally you don't want to schedule
anything when explicitly canceling transfers, but there are a valid use cases
when you cancel because of a timeout and then have to reset the device for
example. Note that this hides bug #2353, where the cancel case is probably
not handled correctly.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25861 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added optional spinlock contention measurement feature. Enabled when
B_DEBUG_SPINLOCK_CONTENTION is defined to 1. It typedefs spinlock to a
structure (thus breaking BeOS binary compatibility), containing a
counter which is incremented whenever a thread has to wait for the
spinlock.
* Added macros for spinlock initialization and access and changed
code using spinlocks accordingly. This breaks compilation for BeOS --
the macros should be defined in the respective compatibility wrappers.
* Added generic syscall to get the spinlock counters for the thread and
the team spinlocks.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25752 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Replaced all instances of benaphores in the kernel code by mutexes.
* Removed kernel benaphores.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25690 a95241bf-73f2-0310-859d-f6bbb57e9c96

Unconditionally report initialization success for OHCI too.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25614 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Disable the root hub change interrupt when it gets triggered. As the change
status stays on until the port is reset, this causes the interrupt to be
constantly retriggered on some controllers.
* Added TODO that we need to reenable it once we want to take advantage of
status change notifications instead of polling. It should be reenabled when
clearing roothub status changes in ClearPortFeature().

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25578 a95241bf-73f2-0310-859d-f6bbb57e9c96

Enable interrupts only after installing the interrupt handler.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25571 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Correct removal of failed transfers by applying the address mask. As the head
pointer also contains the toggle carry and the halt bit (which was obviously
always set as the endpoint had an error) the address comparison failed and
the head was dead in the water with all further transfers timing out until
a subsequet cancel cleaned the mess up again.
* Add the OHCI host controller driver to the image as it should now be fully
functional except for isochronous transfers.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25565 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implement removal of failed transfers. The head pointer will be moved past
the failed transfer if the transfer in question is found to be the offending
one. Probably not quite enough and still untested.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25558 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Build the interrupt tree the same (imperfect) way as in EHCI to conserve
on resources but still satisfy the scheduling policy. This works well, but
effectifly limits the bandwidth available as the interrupts aren't spread
across different frames.
* Implement finding the interrupt endpoint head for a certain interval.
* Deactivate endpoints that are going away until removing them is implemented.
* Allow short packets for all data phases.
* Disable tracing by default.

There you go, interrupt transfers should work as well. Should enable mice,
keyboards, hubs and other devices that use interrupt pipes. Note that it's
still easily possible that a single failed request will hang an endpoint and
prevent a device from working.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25554 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Handle the data == first and last == first case as this can occure for
single td bulk transfers.
* Implement SubmitTransfer() for bulk transfers (interrupts will use the exact
same code path, but as their endpoints aren't yet setup, interrupt transfers
won't work).
* Handle the cancel case when finishing transfers. The descriptors of the
canceled transfers cannot be accessed by the controller so they can be freed.

Bulk transfers should work now, so devices only using control and bulk transfers
should too (anything using usb_disk for example). Note though that a transfer
error will cause the whole thing to fail miserably as the error case is not
yet handled correctly (failed descriptors aren't removed from the endpoint).
Therefore I suggest not testing with the memory stick with that important
presentation you haven't stored anywhere else...

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25549 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Completely rework how transfer finishing is implemented. It is now using
pretty much the same strategy as UHCI. This does not leverage the done queue
the hardware provides us, but as it saves a lot of other overhead and allows
the structures to be smaller I think it is overall worth it.
* Removed now unnecessary stuff from the transfer descriptor struct.
* Directly acknowledge the done head interrupt as we do not use it either.
* Activate control endpoints by removing the skip bit when a transfer is scheduled.
* Correct the way the first descriptor is determined (the current tail becomes
the first descriptor while the current first descriptor becomes the tail).
* Remove the head_logical_descriptor field from the endpoint structure, as this
would simply get out of sync with the first td with no real way of updating.
* Stub out RemoveTransferFromEndpoint() as this needs reworking.
* Correct CreateDescriptorChain() signature and implement the method.
* Add ReadDescriptorChain() and ReadActualLength() to process finished transfers.
* Prepare for actual data transfers by renaming/regrouping transfer types.
* Fully clear out the new tail when switching (for the sake of it)
* Add some helpers and definitions to the hardware header.
* Add debugging facilities.

With all this transfers can now actually be scheduled and they are processed
when done. This brings control transfers to a usable state so setting the
device address and reading out/setting the configuration works. That means you
can now view your device using usb_dev_info, but since any other transfer type
besides control transfers isn't implemented yet the device will most probably
not yet do anything useful.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25546 a95241bf-73f2-0310-859d-f6bbb57e9c96

Correctly clear an endpoint on cancel. Adjusted some comments/output.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25543 a95241bf-73f2-0310-859d-f6bbb57e9c96

Use a dedicated endpoint lock for endpoint manipulation. Fixes the guaranteed
deadlock with the BusManager on setup of the default pipes (for addressing).

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25542 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implement "tail switching" so that a new transfer descriptor chain can be
appended to an endpoint descriptor without locking/disabling the endpoint.
* Correct the direction of the data phase of a control transfer.
* Some minor cleanup.

Control requests without data phase (set address for example) might actually
work now, but this is still untested.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25541 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Complete the roothub get/set/clear port feature functions
* Reorder to match the usual order of the states/commands
* Fix some debug output and make it more informative

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25540 a95241bf-73f2-0310-859d-f6bbb57e9c96

Acknowledge OHCI interrupts.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25539 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Make the operational register memory a uint8 * instead of a uint32 * so the
driver has at least a chance of working (it previously always used wrong
offsets for register access).
* Remove the hash approach for now (I'm going to explore a few other ways of
doing that first).
* Reorder some stuff and check for errors in some more places.
* More cleanup (mostly whitespace again).

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25538 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Getting familiar with the existing code
* Cleanup (whitespace, nameing, code style)
* Move around methods so they match the header order
* Fix some obvious stuff
* Initialize all members
* Sync roothub code (this will be reworked to a common roothub)
* Only minor functional changes (to the worse for now)

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25527 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix building usb busses with tracing on.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24880 a95241bf-73f2-0310-859d-f6bbb57e9c96

Change just host_controller_info to usb_host_controller_info for the sake of
clearity as suggested by Philippe Houdoin once and brought to my attention by
Slavatore. As this definition is completely private to the USB stack it would
not have caused any conflicts though.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24828 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added necessary dummy descriptor to every endpoint created
* Implemented _RemoveTransferFromEndpoint
* Reworked _AppendChainDescriptorsToEndpoint as now every endpoint has (must have) a dummy descriptor



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23741 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented private method _AppendChainDescriptorsToEndpoint in order to avoid code duplication (lock needs to be added)
* Fixed _SubmitControlRequest in order to use _AppendChainDescriptorsToEndpoint


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23734 a95241bf-73f2-0310-859d-f6bbb57e9c96

* More work on the finisher thread
* Reworked SubmitTrasfer method
* Implemented CancelQueuedTransfers
* Minor clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23722 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Removed cast: not needed anymore since Interval() has been moved to the base class (Pipe)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23704 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented _AddPendingTransfer
* renamed trasfer_data memeber top to first_descriptor
* added data_descriptor member to transfer_data (useful for control transfer)



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23702 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed fFinishTransfer to fLastTransfer (I don't even know why I called it
so in the first place)
* Renamed local variable descriptor to current in the finisher thread
* Using next_done_descriptor instead of next_logical_descriptor, as the latest
is used to keep track of all descriptors that belongs to the transfer
* Added _UnlinkTransfer private method
* Implemented _FreeDescriptorChain
* Moved endpoint pointer from the ohci_general_td structur to the transfer_data_s
structure, ad the ohci_general_td has already a transfer_data_s pointer which can
be used to the endpoint pointer reducing ridundancy.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23701 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added definition for the hash related methods (not implemented)
* Continue working in the _FinishTransfer thread
* added next_done_descriptor in ohci_[general|isochronous]_td structure in order to handle
collisions.
* added next_logical_descriptor to ohci_isochronous_td structure



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23594 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Started working on the finisher thread
* Added some TRACE calls in the _InterruptHandler


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23266 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Partially implemented interrupt handler routine


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23175 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed init part of the controller


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22980 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Remove SubmitAsyncTransfer and SubmitPeriodicTransfer
* Added _SubmitControlRequest (almost implemented), _LinkDescriptors, _WriteDescriptorChain (basically copied from EHCI, and UHCI)
* Renamed ohci_general_descriptor to ohci_general_td, and ohci_isochronous_descriptor to ohci_isochronous_td
* Finished _CreateGeneralDescriptor and _FreeGeneralDescriptor
* Added buffer_size member to ohci_general_td instead and removed last_logical_byte_address as it looked unnecessary
* minor clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22968 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented _InsertEndpointForPipe
* usual clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22938 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Replaced snooze with spin (Thanks Axel)
* Fixing TRACE messages
* minor clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22935 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Moved disable_interrupt right before resetting the controller
* Reworked on NotifyPipeChange
* Added _RemoveEndpointForPipe (not implemented)



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22934 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Completed SubmitTransfer
* Added SubmitAsyncTransfer and SubmitPeriodicTransfer (not implemented)
* Removed unecessary spinlock
* Wrapped some lines to follow coding guidelines



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22932 a95241bf-73f2-0310-859d-f6bbb57e9c96

Completely redesign the USB explore process. Replaces the scary race conditions of the previous locking mechanism and simplifies handling of device changes by a more centralized approach.
Changes are now collected during explore and notifications as well as rescans are done at once. Through this a driver is also not rescanned multiple times anymore.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22929 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed some variables
* Reworking the interrupts endpoints tree parts
* Added spin_locker, semaphore, finisher thread and interrupt handler (not implemented)
* Made fInterruptEndpoints allocation dynamic instead of static
* Fixed Start method (it should be correct now)




git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22927 a95241bf-73f2-0310-859d-f6bbb57e9c96

* ups :)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22875 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Placed underscore to private methods like ReadReg and WriteReg
* Rewrote _AllocateEndpoint
* More work in the constructor


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22873 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Rewrote part of the constructor: most of the code is ported from FreeBSD
* Usual clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22625 a95241bf-73f2-0310-859d-f6bbb57e9c96

* License changed: I've successfully contacted Jan-Rixt Van Hoye, and he gave me his bless :)
* Usual clean up


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22555 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Revert Copyright


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22527 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Clean up (removing unneeded comment)
* Renaming, fixing coding guidelines
* Re-writing AddTo method
* Removing ohci_software.h file (its content has been moved to ohci.h) but it will probably be deleted in the future.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22522 a95241bf-73f2-0310-859d-f6bbb57e9c96

More tracing in ohci. Hmm, existing isn't totally consistent (prefix, ...). Oh well.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21242 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fixed wrong usage of object types in OHCI. The types are bitmasks, not values.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19900 a95241bf-73f2-0310-859d-f6bbb57e9c96

Implemented endpoint creation for Pipes.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18986 a95241bf-73f2-0310-859d-f6bbb57e9c96

Now the ports are powered on and seem to work...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18928 a95241bf-73f2-0310-859d-f6bbb57e9c96

Implemented most of the roothub.

There remains only one issue: reading the Port registers seem to return a bogus value (0). I need to find out what's causing it.



git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18918 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented usable thread synchronisation for UHCI and EHCI.
* Cleaned interrupt handlers for both busses, too.
* Fixed memory leak of the queue head stray descriptors in EHCI.
* Fixed warnings about assigning NULL to the module flags field.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18857 a95241bf-73f2-0310-859d-f6bbb57e9c96

Wrote the initialisation routine of the host controller
* Started cleaning up the style (WIP)
* I'll slowly phase out ohci_software.h, since it is BSD legacy
* Cleanup other BSD inspired naming. I like full names.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18723 a95241bf-73f2-0310-859d-f6bbb57e9c96

replaced with PCI.h definitions, added PCI_usb_ehci


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18704 a95241bf-73f2-0310-859d-f6bbb57e9c96

First commit to show that work on the ohci bus has restarted.
It doesn't do anything yet, but it allocates the io registers properly and
now the values that are returned when reading it make sense.

More to come...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18673 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Adding skeleton EHCI driver. Some parts filled from the specs and the UHCI driver.
* Corrected some errors I found while reading UHCI.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18626 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added USB API Version 3. It's put into USB3.h while the v2 API resides in USB2.h. USB.h just includes USB2.h for now.
* Rewrote both headers on the way.
* The usb module now exports both, the v2 and v3 module_info.
* Changed the internals of the USB Stack to give out usb_ids instead of opaque handles to internal classes.
* Cleaned up some more of the Stack by moving members into other classes and removing unused stuff.
* Updated the usb_raw driver from v2 to v3 API.

Since both usb_hid (which still uses the v2 API) and usb_raw (which now uses the v3 API) work, I'd call it a success ;-)

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18539 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fixed building of the UHCI and OHCI modules [reported by AndyBe].


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15914 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fixed the bugs I introduced earlier. original code was ok.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14244 a95241bf-73f2-0310-859d-f6bbb57e9c96

properly mask off bits


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14232 a95241bf-73f2-0310-859d-f6bbb57e9c96

it's not allowed to spin 250ms


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14231 a95241bf-73f2-0310-859d-f6bbb57e9c96

Quick merge of my development branch with mainline (for SVN switch).
The OHCI stuff was done by Jixt (thanks)./installusb
All is still very much work in progress

Patches applied:

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--base-0
tag of Niels.Reedijk@gmail.com--haiku-2005/usb-busses--mainline--0.1--base-0

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-1
Set the PORT_STATUS_LOW_SPEED flag correctly (using the register bitflag for low speed devices, instead of the proper lowspeed flag)

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-2
Clean up transfers when they're done

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-3


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-4


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-5


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-6


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-7
Remove obsolete ehci.c

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-8


* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-9
Completed memory allocation + initialization

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-10
First attempt at performing control transfers - failed miserably

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-11
Trying to read a 16bit value into a 8bit integer, stupid

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-12
nicer way of getting the base io address

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-13
Take over control from BIOS

* Niels.Reedijk@gmail.com--nielx-2005/usb-busses--development--0.1--patch-14
Revert change for disabling USB BIOS support


git-svn-id: file:///srv/svn/repos/haiku/trunk/current@11743 a95241bf-73f2-0310-859d-f6bbb57e9c96