USB: Add overall timeout to the PhysicalMemoryAllocator.

The within-an-attempt timeout did not successfully break up
deadlocks that occur on a system with a lot of USB transfers
going, as we may never hit 2 seconds in between wake-ups,
but the size requested may be un-fulfillable regardless.

So, now we have a 2-second overall timeout. This fixes
the system freeze in #15569, but now attached USB disk
drives enter a Stall state, making the system unusable
anyway.

USB: Rework the PhysicalMemoryAllocator to use MutexLocker.

No functional change, but will make some subsequent
changes easier and more fail-proof.

USB: Don't loop endlessly waiting for a physical buffer.

This is used quite a lot in critical transfer paths, so we don't
want to lock things up if no buffers are available for whatever reason.
Wait 2 seconds, and if we didn't get anything by then, return B_NO_MEMORY.

Possibly fixes or helps with certain USB-related lockups.

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: Remove BeOSCompatibility.h and associated cruft.

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

usb: fix format string

fDebugChunkSize is an uint32.

Signed-off-by: Paweł Dziepak <pdziepak@quarnos.org>

USB: Cleanup hardcoded bitmap size.

USB: Warn if debug allocation does not fit debug chunk size.

USB: Increase debug chunk size from 64 to 128 bytes.

The queue head structure in EHCI is bigger than the previous debug
chunk size. Multiple allocations therefore overwrote eachother.

Get USB functional on x86-64.

- Various changes to printf format specifiers to make them 32/64 safe.
- Adjustments to various descriptor structures since addr_t is not 32-bit
on x86-64, and consequently this breaks the physical structure that's
expected. Also adjusted padding to accomodate 32-byte boundaries when
64-bit pointers are in use.
- Adjustments to make use of phys_addr_t where needed.

Make the USB physical memory allocator more robust. Wait for space to become
available when there is none instead of spinning around. Might help with #8053
and #8058, though it's possible that leaks or interrupt issues are the cause
for these.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43125 a95241bf-73f2-0310-859d-f6bbb57e9c96

Replaced B_32_BIT_MEMORY by B_32_BIT_FULL_LOCK and B_32_BIT_CONTIGUOUS, so
the constraint can be expressed more precisely. ATM B_32_BIT_FULL_LOCK is
implemented as B_32_BIT_CONTIGUOUS when B_HAIKU_PHYSICAL_BITS > 32, though.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37226 a95241bf-73f2-0310-859d-f6bbb57e9c96

Force allocation of physical memory < 4 GB until the related TODOs are fixed.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37201 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

* 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

* Added simplified possibility to schedule UHCI transfers from within KDL.
* Added debugger commands to resolve usb_ids to pipes.
* Adjusted the physical memory allocator to be usable in a slimmed down mode
when running inside the kernel debugger.
* Implemented USB keyboard support for KDL through a kernel debugger add-on.
* Added kgetc() and made use of it where previously individual methods were used
to ensure that reading characters always goes through the kernel debugger
add-ons and the other methods.

This has some preconditions to meet though:
1) The keyboard must be in the boot protocol (currently the case but needs to
be revisited once we have a full usb_hid).
2) The keyboard must be attached to a UHCI root port (i.e. not use EHCI or OHCI,
also not through hubs unless those are USB 1.1).
3) the usb_hid driver has to be opened for this to work. This means that for the
time between initializing USB and when usb_hid is opened by the input_server
there is no keyboard support.

Also note that this has no way of detecting hot-plug, meaning that you can't
re-attach your USB keyboard from the hub to the root port once in KDL.

On the bright side of things, since this is a non-destructive mechanism it is
possible to enter and leave KDL without loosing the USB state.

Tested OK in QEMU, not tested on real hardware yet, will see in a few minutes.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29291 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed a ton of wrong usages of B_CONTIGUOUS + B_FULL_LOCK.
* The use of B_{READ|WRITE}_AREA throughout the drivers is surely alarming.
Defining these flags means that *every user* application can access these
buffers read/write, it becomes visible in userspace like any other memory
(just shared among all apps). I would like to ask each driver maintainer
to see if that is really wished here. If you only need one app to be able
to access it, cloning the area would be more appropriate.
* I came across the use of B_ANY_KERNEL_BLOCK_ADDRESS a number of times. This
is almost completely useless for most usages, as it tries to align the
virtual to a multiple of the size of the area. It just makes the allocation
more likely to fail. Please only use where appropriate, and please review
your code.
* Minor cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26858 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

* Added some more debug output to the hub code
* Report and clear yet unhandled port changes in hubs
* Fixed the way devices that vanish on port reset are handled
* Added a status field to the memory allocator and check it creating it

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19732 a95241bf-73f2-0310-859d-f6bbb57e9c96

Oops, this was the very conservative value. Testing with my laptop (that has a slow EHCI controller) reveals that we need some more room here.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18930 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added retries to the PhysicalMemoryAllocator. Memory is allocated and freed at a very high frequency, so low memory conditions shouldn't last very long.
* Added a separate thread for cleaning up EHCI transfers. This makes it possible to actually always wait for the async advance interrupt without hindering execution of transfers.
This pushes performance again and fixes the bug I introduced in the last change, that we could free yet cached descriptors when a previous async advance timed out.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18929 a95241bf-73f2-0310-859d-f6bbb57e9c96

Adding compatibility headers to enable R5 build of the USB stack again.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18856 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented most qTD and qH functions for EHCI async transfers.
* Implementing control and bulk (async) transfers. They don't work yet though.
* Corrected the legacy support handling in EHCI
* Fixed the bit mask for total bytes in qTDs and added the data toggle bit
* Changed the boolean low speed flags to usb_speed enum that allows for low, full and highspeed

Also wrote a memory allocator that takes over memory management from the stack. It is similar to a buddy allocator. This does on the one hand remove some ugly code from the stack and is on the other hand far more scalable than the previous free list approach. This scalability is needed for proper EHCI support where buffers can go up to a size of 5 pages. The allocator is not perfect and it's memory overhead is quite high compared to the managed memory (depends on the managed block sizes), but it's quik and works well so far.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18849 a95241bf-73f2-0310-859d-f6bbb57e9c96

usb: fix format string

fDebugChunkSize is an uint32.

Signed-off-by: Paweł Dziepak <pdziepak@quarnos.org>

USB: Cleanup hardcoded bitmap size.

USB: Warn if debug allocation does not fit debug chunk size.

USB: Increase debug chunk size from 64 to 128 bytes.

The queue head structure in EHCI is bigger than the previous debug
chunk size. Multiple allocations therefore overwrote eachother.

Get USB functional on x86-64.

- Various changes to printf format specifiers to make them 32/64 safe.
- Adjustments to various descriptor structures since addr_t is not 32-bit
on x86-64, and consequently this breaks the physical structure that's
expected. Also adjusted padding to accomodate 32-byte boundaries when
64-bit pointers are in use.
- Adjustments to make use of phys_addr_t where needed.

Make the USB physical memory allocator more robust. Wait for space to become
available when there is none instead of spinning around. Might help with #8053
and #8058, though it's possible that leaks or interrupt issues are the cause
for these.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43125 a95241bf-73f2-0310-859d-f6bbb57e9c96

Replaced B_32_BIT_MEMORY by B_32_BIT_FULL_LOCK and B_32_BIT_CONTIGUOUS, so
the constraint can be expressed more precisely. ATM B_32_BIT_FULL_LOCK is
implemented as B_32_BIT_CONTIGUOUS when B_HAIKU_PHYSICAL_BITS > 32, though.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37226 a95241bf-73f2-0310-859d-f6bbb57e9c96

Force allocation of physical memory < 4 GB until the related TODOs are fixed.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37201 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

* 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

* Added simplified possibility to schedule UHCI transfers from within KDL.
* Added debugger commands to resolve usb_ids to pipes.
* Adjusted the physical memory allocator to be usable in a slimmed down mode
when running inside the kernel debugger.
* Implemented USB keyboard support for KDL through a kernel debugger add-on.
* Added kgetc() and made use of it where previously individual methods were used
to ensure that reading characters always goes through the kernel debugger
add-ons and the other methods.

This has some preconditions to meet though:
1) The keyboard must be in the boot protocol (currently the case but needs to
be revisited once we have a full usb_hid).
2) The keyboard must be attached to a UHCI root port (i.e. not use EHCI or OHCI,
also not through hubs unless those are USB 1.1).
3) the usb_hid driver has to be opened for this to work. This means that for the
time between initializing USB and when usb_hid is opened by the input_server
there is no keyboard support.

Also note that this has no way of detecting hot-plug, meaning that you can't
re-attach your USB keyboard from the hub to the root port once in KDL.

On the bright side of things, since this is a non-destructive mechanism it is
possible to enter and leave KDL without loosing the USB state.

Tested OK in QEMU, not tested on real hardware yet, will see in a few minutes.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29291 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed a ton of wrong usages of B_CONTIGUOUS + B_FULL_LOCK.
* The use of B_{READ|WRITE}_AREA throughout the drivers is surely alarming.
Defining these flags means that *every user* application can access these
buffers read/write, it becomes visible in userspace like any other memory
(just shared among all apps). I would like to ask each driver maintainer
to see if that is really wished here. If you only need one app to be able
to access it, cloning the area would be more appropriate.
* I came across the use of B_ANY_KERNEL_BLOCK_ADDRESS a number of times. This
is almost completely useless for most usages, as it tries to align the
virtual to a multiple of the size of the area. It just makes the allocation
more likely to fail. Please only use where appropriate, and please review
your code.
* Minor cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26858 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

* Added some more debug output to the hub code
* Report and clear yet unhandled port changes in hubs
* Fixed the way devices that vanish on port reset are handled
* Added a status field to the memory allocator and check it creating it

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19732 a95241bf-73f2-0310-859d-f6bbb57e9c96

Oops, this was the very conservative value. Testing with my laptop (that has a slow EHCI controller) reveals that we need some more room here.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18930 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added retries to the PhysicalMemoryAllocator. Memory is allocated and freed at a very high frequency, so low memory conditions shouldn't last very long.
* Added a separate thread for cleaning up EHCI transfers. This makes it possible to actually always wait for the async advance interrupt without hindering execution of transfers.
This pushes performance again and fixes the bug I introduced in the last change, that we could free yet cached descriptors when a previous async advance timed out.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18929 a95241bf-73f2-0310-859d-f6bbb57e9c96

Adding compatibility headers to enable R5 build of the USB stack again.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18856 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Implemented most qTD and qH functions for EHCI async transfers.
* Implementing control and bulk (async) transfers. They don't work yet though.
* Corrected the legacy support handling in EHCI
* Fixed the bit mask for total bytes in qTDs and added the data toggle bit
* Changed the boolean low speed flags to usb_speed enum that allows for low, full and highspeed

Also wrote a memory allocator that takes over memory management from the stack. It is similar to a buddy allocator. This does on the one hand remove some ugly code from the stack and is on the other hand far more scalable than the previous free list approach. This scalability is needed for proper EHCI support where buffers can go up to a size of 5 pages. The allocator is not perfect and it's memory overhead is quite high compared to the managed memory (depends on the managed block sizes), but it's quik and works well so far.

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18849 a95241bf-73f2-0310-859d-f6bbb57e9c96