GitLab

All scroll wheel events are now sent to PointerV 9 (WheelChange), along
with any extra mouse buttons. This will allow us to centralise Wimp
scrollwheel logic in the WindowScroll module.

Version 1.30. Tagged as 'USBDriver-1_30'

We want to add PointerReason_WheelChange to Hdr:Pointer, but doing so
will make USBDriver upset. Prematurely add a #ifndef guard to prevent
this.


Version 1.29. Not tagged

Detail:
  1) If the alt setting had been changed from 0 it couldn't be changed back to 0.
  2) The alt value that usbd_set_interface takes should be an index into
     the number of alternate settings (altidx) not idesc->bAlternateSetting.
  3) usbd_set_interface returns an error (USBD_IN_USE) if pipes are open
     on the interface when you try to change it. We should return this error
     if the alternate setting can't be changed as the alternate setting would
     be wrong.
  Replace uses of usbd_set_interface() with a new function which correctly maps alternate number to alternate index, and propagates errors back to the caller.
Admin:
  Submission from Colin Granville.

Version 1.29. Tagged as 'USBDriver-1_29'

Merged from NetBSD revision 1.745.

Version 1.28. Tagged as 'USBDriver-1_28'

Detail:
  Service_USB subreason 5 now announces the USBDriver has finalised, HCDs will be able to use this to do any cleanup they desire.
  usbmodule.c - Add service, remove and detach buses. Also, fix accidental dereference of address 1 in config_detach().
  USBDevFS.h & Doc/USB - Add service.
Admin:
  Requires TCPIPLibs-5_63 for the safe queue removal macros.
  Submission from Colin Granville.

Version 1.27. Tagged as 'USBDriver-1_27'

Detail:
  Add missing brace which otherwise meant only the line
    req.bmRequestType = UT_WRITE_ENDPOINT;
  was conditional, rather than the following 6 lines.
  The confusing set of #ifdef __riscos/#endif was masking a missed call to cleartoggle() which would leave the controller and device out of sync.
  Tidy this up, and apply similar to usbd_clear_endpoint_stall() for clarity.
Admin:
  usbdi.c - as detailed above, based on a submission from Colin Granville, with edits.
  usbmodule.c - delete unused DEVICEFSISBROKEN switch, the bug to which it refers was fixed in DeviceFS-0_63 in 2003.

Version 1.26. Tagged as 'USBDriver-1_26'

Detail:
  OHCI_DEBUG shouldn't be set here.
Admin:
  Submission for USB bounty.

Version 1.25. Tagged as 'USBDriver-1_25'

Should read endpoint descriptor or set protocol fail, free the softc.
Ref https://www.riscosopen.org/forum/forums/4/topics/9503

Version 1.24. Tagged as 'USBDriver-1_24'

Detail:
  Remove the (now) unused sources from when this component produced multiple targets. See Controllers/OHCIDriver & Controllers/EHCIDriver for the live copies.
  Remove unset USBHAL defines.
  Simplify the Makefile and !Mk* scripts.
  No longer export usbroothub_subr.h.
  Add a porting type "device_t" to usb_port.h.
Admin:
  Submission for USB bounty.

Version 1.23. Tagged as 'USBDriver-1_23'

The documentation of the Alias$@USBDevice variables barely made sense, and examples not that consistent - improve this.
Fix step 1 of the alias handling to match the docs (this was fixed previously in revision 1.34 but accidentally overwritten in 1.35 less than an hour later).
Add a check for malloc() failing in low memory situations.

Tested with a couple of USB devices and alias variables containing 'echo' commands.

Version 1.22. Tagged as 'USBDriver-1_22'

Detail:
  Export usbroothub_subr.h, for those HC's that might need it.
  Export C version of usbhal.h, in addition to the assembler Hdr version, for MUSBDriver.
  Eliminate the dependency on opt_usbverbose.h - an empty header file - by merging the change from usb_subr.c revision 1.186 from NetBSD which guards it with internal _KERNEL_OPT (unset).
Admin:
  Submission for USB bounty.

Version 1.21. Tagged as 'USBDriver-1_21'

Detail:
  Makefile and !Mk* updated, and corresponding unused files put in the attic.
  Add an interim OHCIHALLib and EHCIHALLib target so that the HAL library variant still builds.
  Split messages into CmdHelp and Messages per a normal C component.
  Remove unused wsconsio/wsmousevar porting headers, no longer used.
Admin:
  Submission for USB bounty.
  Note, the version number of the resulting module will jump to match the top level VersionNum file.

Version 1.20. Tagged as 'USBDriver-1_20'

Admin:
  This is a (server side) copy of the component previously called RiscOS/Sources/HWSupport/USB/NetBSD to be used for ongoing development, retaining the significant revision history.
  Submission for USB bounty.

Detail:
  During internationalisation, the last line lost its newline.
Admin:
  Submission from Colin Granville.

Version 1.19. Tagged as 'NetBSD-1_19'

Detail:
  USBDriver was performing unnecessary messages file reopening on Service_ResourceFSStarted (MessageTrans does this itself), remove this.
  Internationalised USBDriver, in particular its *Commands.
  Fixed standalone builds to output the messages file objects in the right place (Makefile mistake).
  Changed OHCIDriver and EHCIDriver to use allocated error bases rather than 0.
  Sync'd, where possible, the OHCIDriver and EHCIDriver sources which share a common heritage. Tentative shared interrupt support fixes for non-PCI attached controllers. Comment blocks & indentation improved.
  Export min() macro in usb_port.h.
  CMHG updated to not listen for unwanted services.
Admin:
  Submission for USB bounty.
  Tested on Pandaboard and Pi 2. OHCIDriver untested but low risk.

Version 1.18. Tagged as 'NetBSD-1_18'

Detail:
  DeRegisterBus->DeregisterBus. Since clients calling this SWI are compiled C or assembler, this doesn't affect existing modules which remain binary compatible.
Admin:
  Submission for USB bounty.

Version 1.17. Tagged as 'NetBSD-1_17'

Detail:
  build/c/usbmouse - Because the touchpad sends data that's consistent with a Boot protocol mouse, the data it sends is shorter than the expected size. This can result in bogus scroll wheel data being read from off the end of the buffer. Fix this by completely ignoring the report descriptor that the device returns and instead using a descriptor that's consistent with a Boot protocol mouse.
  build/Version - Increase USBDriver version number
Admin:
  Tested on TouchBook
  Fixes issue seen where garbage data at the end of the buffer would result in windows scrolling for every mouse interrupt


Version 1.16. Tagged as 'NetBSD-1_16'

The order of the new device enumeration means we have the initial device descriptor (hence bMaxPacketSize) but don't compute wMaxPacketSize until after addressing.
The XHCI HC needs this value by the addressing stage in order to setup the control endpoint context (other HC don't need it until later) so the specific case of a full speed device whose endpoint 0 has a non default max packet size would be set up using the default (of 64), then when the full device descriptor is requested the packet looks like a short transfer and the transaction ends having only collected (eg) 8 bytes. This would then appear as vendor &0000 product &0000 and not get any further.
Tested on Titanium.

Version 1.15. Tagged as 'NetBSD-1_15'

Detail:
  build/c/usbhal - Don't assume HAL_IRQClear will clear the timer interrupt, use HAL_TimerIRQClear as well.
Admin:
  Tested on BB-xM


Version 1.14. Tagged as 'NetBSD-1_14'

Found by pattern matching, this is the same change as DWCDriver-0_27.

Version 1.13. Tagged as 'NetBSD-1_13'

Version 1.12. Tagged as 'NetBSD-1_12'

Detail:
  build/c/port - Fix vtophys requesting that pages are made uncacheable. They're already uncacheable (all allocated via PCI_RAMAlloc), and once we're done with the pages we never do a 'make cacheable' call to balance it out
  build/Version - Bump version numbers of modules which use vtophys
Admin:
  Untested


Version 1.10. Tagged as 'NetBSD-1_10'

Admin:
  These changes are currently subject to rework via the normal submissions channel.

Admin:


Version 1.11. Tagged as 'NetBSD-1_11'

Detail:
Frontend
========

NetBSD.build.c.usbmodule
	frontend isochronous changes
	forces reads to always be multiple of maxpacket
NetBSD.dev.usb.c.usb_subr
	datatoggle fix
	panic fix
NetBSD.dev.usb.c.usbdi
	usbd_transfer changed so that bufrem is passed the xfer pointer instead of the buffer pointer
	fixed missing braces in usbd_clear_endpoint_stall

EHCI
====

The HS (USB2) isochronous as been rewritten and isochronous split transactions - FS (USB1) over a HS Connection - has been added

NetBSD.dev.usb.c.ehci
NetBSD.dev.usb.h.ehcireg
NetBSD.dev.usb.h.ehcivar

OHCI
====

Changes to the NetBSD backend to get isochronous transfers working on a OHCI device. While it did have isochronous support it wasn't working properly.

NetBSD.dev.usb.c.ohci

Admin:
  Submitted by Colin Granville
Tested in iMx6

Version 1.10. Tagged as 'NetBSD-1_10'

The safety struct fillers added in OHCIDriver 0.44 and EHCIDriver 0.30 respectively were being carefully replaced by veneers to nothing. Skip veneer generation is the method isn't implemented.

Version 1.09. Tagged as 'NetBSD-1_09'

Detail:
  Play safe, put NULLs in for the unused HC overrides.

Version 1.08. Not tagged

Detail:
  Support controllers such as the XHCI which perform some steps of enumeration themselves in hardware, thereby needing some steps of the software enumeration to be bypassed.
  usb.c:
    Internally permit addr of 0, to represent the root hub, since the controller is responsible for allocating 1+. Merged from NetBSD 1.144.
  usb_subr.c:
    Test pointers against NULL not 0. Merged from NetBSD 1.191.
    Add 4 new bus methods, for prior to enumeration, address assignment phase, post enumeration fixup, and device removal.
  usbdivar.h:
    Prototypes of 4 new bus methods.
Admin:
  HC API version number bumped.
  Submission from Elesar.

Version 1.08. Tagged as 'NetBSD-1_08'

On first registering a new bus in the the NetBSD build of the code usb_kthread_create(usb_create_event_thread, sc) is called which in turn creates the main event thread ("usb_event_thread").
The first thing usb_event_thread() does is set
  sc->sc_bus->needs_explore = 1;
  usb_discover(sc);
however in the RISC OS version of the code (usb.c around line 254) only the first action is performed, no discover happens.
This change schedules a callback via usb_needs_explore_callback() which in turn calls discover_callback() when the OS is next free, which calls usb_discover() for each bus in an equivalent way to NetBSD's event thread. There's then no longer any need for the special one-shot 11s rediscover callback, only the repeating 30s rediscover is needed.

For reference, the places that set needs_explore in this version are all in usb.c
  Line 254: set, but doesn't schedule a callback, as noted above
  Line 339: code not compiled for __riscos
  Line 548: set, does schedule a callback
  Line 727: cleared, in usb_discover()
  Line 743: set, does schedule a callback
  Line 756: code not called in RISC OS

NetBSD uses 500ms initial delay before discover, then 60s between discovers (or until requested by a hub, whichever is sooner).
RISC OS has a periodic callback of 30s (in usbmodule.c) and an initial discover of however long it is from the bus being registered until callbacks next fire usb_event_thread(), assumed ~500ms.

Version 1.07. Tagged as 'NetBSD-1_07'

Detail:
  In version 1.05, the USB library which is exported for the use of the HAL
  keyboard scan code unconditionally required the HAL_USBPortPower symbol.
  Not all HALs provide that entry point; for those that didn't, they would
  no longer link. Fixed by use of a bit of magic assembler.
Admin:
  Verified that OMAP3 ROM now builds again.

Version 1.06. Tagged as 'NetBSD-1_06'

Detail:
  When signalled as a feature of the HAL device, call HAL_USBPortPower to turn on the port.
Admin:
  Submission from Willi Theiss.

Version 1.05. Tagged as 'NetBSD-1_05'

Assign a proper error number and translate "Incompatible USBDriver version".
Don't bother internationalising the *commands which are for debug builds only.
Rephrase the help for USBQuirk so it makes sense, and trim the unmatched '<' from the syntax example.
Bracket the USBDiscover help with the same switch that governs the code, which has been disabled for over a decade.

Version 1.04. Tagged as 'NetBSD-1_04'

Detail:
  From NetBSD usb_quirks.c revision 1.81.
  usbmodule: NULL is for pointer types, not ints.
  usb_quirks: Group the RISC OS specific ones together, disable the non applicable UQ_APPLE_ISO ones, UQ_LOST_CS_DESC doesn't apply either but is only used for USB modem drivers not ported here. Untangle ifdefs around usbd_find_quirk().
Admin:
  Removed unused source file ums.c

Version 1.03. Tagged as 'NetBSD-1_03'

Minor makefile correction to enable o_hal component creation, and change to !MkExport to export EHCI and OHCI elements as well

Detail:
Admin:
  tested in iMx6 build


Version 1.02. Tagged as 'NetBSD-1_02'

In products using OHCIDriver/EHCIDriver when they came to do a header export would overwrite the USBDriver header with an empty one.

However, internally we get that info from the descriptor, so mask it off.
Submission from Colin Granville.

No version number bump, it's already been incremented today.

Version 1.01. Tagged as 'NetBSD-1_01'

From NetBSD dev/usb/usbdivar.h 1.109.
As these extra members change the struct layout, bump the API version number too. Having carefully analysed which headers are actually required by host controllers (in NetBSD-0_98) also be explicit about them in Hdr/USBDriver rather than the former vague comment in h/usbdivar.

Hdr/USBDriver:
  Relocate API define here.
usbmodule.c:
  Include USBDriver.h to get the API define.
usbdivar.h:
  Add struct members from recent NetBSD.

Version 1.00. Tagged as 'NetBSD-1_00'

Use -1 as the invalid mark, so that an accidental match on interface=0 alternate=0 is no longer possible.
Submission from Colin Granville.

Version 0.99. Tagged as 'NetBSD-0_99'

From NetBSD 1.698.

USBDriver was not exporting its SWI (and other) definitions to the export directory like other modules do, meaning the host controller drivers all have to reach into this component to get their hands on it.
Makefile:
  Export an assembler header, conversion to C, and SWI definitions from CMHG.
  Rationalise the dev/usb/h/* headers which are exported in addition to those that are pertinent to host controllers (was missing usbdivar, and was exporting unused usbdi_util).
  Export machine/h/bus too.
  As it's unusual for a CModule to output both a library and module and headers, the EXPDIR has been divided into EXPLIBDIR for those things in C:USB, and EXPDIR for the others.
ehcimodule.c/ohcimodule.c:
  Use the exported USBDriver.h

Version 0.98. Tagged as 'NetBSD-0_98'