Home | History | Annotate | Download | only in usb
History log of /src/sys/dev/usb/xhci.c
RevisionDateAuthorComments
 1.190  24-Aug-2025  nat Handle special case of USBD_FORCE_SHORT_XFER.

This fixes devices such as urtwn(4) connected to an xhci port.

It works by sending a zero length transfer at the end of the data transfer
for the above mentioned flag.

Similar artwork in this space already exists for ohci/uhci/ehci.

OK jkallsch@.
 1.189  02-Aug-2025  mlelstv Fix computation of ISOC microframes.
Also, avoid crash on ISOC pipe restart (there is no upm_start method).
 1.188  30-Jan-2025  jmcneill branches: 1.188.2;
xhci: Do not clobber existing state in xhci_update_ep0_mps

When updating the MPS field, make sure to preserve existing configuration.
Without this, the XHCI IP on CIX CD8180 was returning a parameter error
in response to the command.
 1.187  30-Jan-2025  jmcneill xhci: Always do 32-bit accesses for platform (ACPI) XHCI devices.

The XHCI IP found in the CIX CD8180 throws an serror for any access that
is not 32-bit aligned. Let's add an XHCI_32BIT_ACCESS flag and set it
unconditionally for platform devices as there are only a handful of non-
32-bit accesses at init time.
 1.186  09-Jan-2025  jmcneill xhci: Don't use USBMALLOC_COHERENT

Use normal WB mappings with non-cache coherent tags, will improve
performance and seems to avoid stalling transfers on X1E.
 1.185  21-Dec-2024  skrll Relax the KASSERT in xhci_event_cmd to allow xhci_polling_p
 1.184  20-May-2024  riastradh branches: 1.184.2;
xhci(4): Narrow some more variable scopes in xhci_device_isoc_enter.

No functional change intended.
 1.183  20-May-2024  riastradh xhci(4): Narrow scope of variable.

Nix spurious initialization in wider scope.

No functional change intended.
 1.182  20-May-2024  riastradh xhci.c: Fix confusing line break.

No functionanl change intended.
 1.181  05-Apr-2024  riastradh usb *hci: Always set ux_status before usbd_xfer_schedule_timeout.

Add an assert to usbd_xfer_schedule_timeout to enforce this.

Since access to ux_status is serialized by the bus lock, and nothing
releases the bus lock in the interim, this doesn't make a functional
change. But it does reduce confusion by readers, who no longer have
to worry if some callers got the order wrong. It could also now
potentially be factored out in a subsequent commit.
 1.180  20-Jul-2023  riastradh xhci(4): Don't panic on suspend if previous suspend/resume failed.

Trying to resume again probably won't make the situation much worse,
but panicking can definitely make it worse.

XXX pullup-10
 1.179  05-Jun-2023  mlelstv - Fix clearing of EINT and other transient flags.

Patches from sc.dying in kern/56115:
- Set proper Max ESIT Payload value for interrupt/isoc endpoint context.
- Set proper Average TRB Length value.
- Not tested on superspeed/superspeedplus isochronous device.

- Add handling of some error paths for isochronous transfers.
 1.178  27-Apr-2023  skrll Remove an unnecessary cast in a KASSERTMSG
 1.177  09-Apr-2023  riastradh xhci(4): Avoid crash in suspend/resume/resume if first resume fails.

Rather than try to recover from this, just make new commands fail so
at least we don't deadlock.

XXX pullup-9
XXX pullup-10
 1.176  07-Apr-2023  riastradh xhci(4): Defer root intr xfers while polling.

Root intr xfers require taking adaptive locks, which is forbidden
while polling.

This is not great -- any USB transfer completion callbacks might try
to take adaptive locks, not just uhub_intr, and that will always
causes trouble. We get lucky with ukbd_intr because it's not
MP-safe, so it relies only on the kernel lock (a spin lock) anyway.
But this change brings xhci in line with ehci.

PR kern/57326

XXX pullup-8
XXX pullup-9
XXX pullup-10
 1.175  11-Oct-2022  msaitoh branches: 1.175.2;
KNF a bit. No functional change.
 1.174  11-Oct-2022  msaitoh "Add" number of ports because the same speed's ECR may appear multiple times.
 1.173  11-Oct-2022  msaitoh There is an xHCI device which has USB 2 port only. Support it.

- Example:

xhci4 at pci17 dev 0 function 0: AMD product 15b8 (rev. 0x00)
xhci4: 64-bit DMA
allocated pic msix10 type edge pin 0 level 6 to cpu0 slot 32 idt entry 107
xhci4: interrupting at msix10 vec 0
xhci4: xHCI version 1.20
xhci4: hcs1=1000840 hcs2=140000f1 hcs3=7000a
xhci4: hcc=0x110ffc5<XECP=0x110,MAXPSA=0xf,CFC,SEC,SPC,PAE,NSS,LTC,CSZ,AC64>
xhci4: xECP 440
xhci4: hcc2=0x3f<CIC,LEC,CTC,FSC,CMC,U3C>
xhci4: ECR: 0x00000401
xhci4: ECR: 0x02000402
xhci4: SP: 0x02000402 0x20425355 0x00180101 0x00000000
xhci4: hs ports 1 - 1
xhci4: ECR: 0x000f000a
xhci4: PAGESIZE 0x00000001
xhci4: sc_pgsz 0x00001000
xhci4: sc_maxslots 0x00000040
xhci4: sc_maxports 1
xhci4: sc_maxspbuf 2
xhci4: eventst: 0x000000013ee60fc0 0xffffb08826f5afc0 1000
xhci4: dcbaa: 0x000000013ee63000 0xffffb08826f5b000 1000
xhci4: current IMOD 0
(snip)
usb8 at xhci4: USB revision 3.1
usb9 at xhci4: USB revision 2.0
uhub8 at usb8: NetBSD (0x0000) xHCI root hub (0x0000), class 9/0, rev 3.00/1.00, addr 0
uhub8: 0 ports with 0 removable, self powered
uhub8: no ports, hub ignored
uhub8: WARNING: power management not supported
autoconfiguration error: usb8: root device is not a hub
usb8: WARNING: power management not supported
uhub9 at usb9: NetBSD (0x0000) xHCI root hub (0x0000), class 9/0, rev 2.00/1.00, addr 0
uhub9: 1 port with 1 removable, self powered

- To resolve this problem, keep number of ports of SS and HS and use
it to attach child device(s).
- Tested on ASUS TUF GAMING X670E-PLUS.
- OK'd by skrll@.
 1.172  25-Sep-2022  skrll Remove the bus_space_barrier I added with the commit

revision 1.103
date: 2019-01-11 15:43:51 +0000; author: skrll; state: Exp; lines: +12 -2; commitid: RYSytjLNOZrFCn7B;
Add a bs_barrier to make ThunderX xhci work. Not sure why this is
needed, but I'll work it out later.

Turns out bus_dma.c needed fixing which I did with

revision 1.115
date: 2019-06-14 10:09:12 +0100; author: skrll; state: Exp; lines: +39 -34; commitid: GKW2p7ijoslyu8rB;
Simplify the _ARM32_NEED_BUS_DMA_BOUNCE #ifdefs and rely on compiler
optimisation of the bouncing = false case.

Drain the write buf (aka DSB) in more cases

Catch all CPUs that support speculation. (thunderx isn't CPU_CORTEX)
 1.171  13-Sep-2022  riastradh xhci(4): After attach, access to sc_child/2 requires sc_intr_lock.

Serializes access with xhci_intr.

XXX Need to ensure the interrupt handler is quiesced at this point or
else it will trip over the assertion in xhci_intr about having at
least one child.
 1.170  13-Sep-2022  riastradh xhci(4): Resume commands even if USBSTS.SRE is set.

Commands might not work after this but let's at least not deadlock --
give them an opportunity to time out or fail.

XXX Maybe set sc_dying here to skip the timeout.
 1.169  01-Sep-2022  riastradh xhci(4): Make sure to destroy sc_rhlock on detach.
 1.168  23-Aug-2022  riastradh xhci(4): Fix error branch for failed suspend.

If suspend failed, at least we can stop it from blocking all
subsequent xhci commands or making a second suspend crash.
 1.167  24-May-2022  andvar fix various typos in comment, documentation and log messages.
 1.166  14-May-2022  riastradh xhci(4): Handle race between software abort and hardware stall.
 1.165  14-May-2022  riastradh xhci(4): Fix edge case in simultaneous xfer abort and failure.

On successful usbd_xfer_trycomplete, caller must set ux_status and
call usb_transfer_complete before releasing the pipe (bus) lock.
Failing to call usb_transfer_complete is a mistake. Presumably this
was intended to claim the xfer to complete it only on the last
packet.

I previously introduced the violation of this rule when the code
looked like

xfer->ux_status = err;
if (trb stuff)
usb_transfer_complete(xfer);

I mostly mechanically changed all the assignments of xfer->ux_status
to do usbd_xfer_trycomplete first and then usb_transfer_complete.

In the original, the extra assignment of xfer->ux_status in the event
we _don't_ immediately call usb_transfer_complete was likely
redundant and (except insofar as the abort protocol was broken)
harmless. But now it is a problem because of the contract between
usbd_xfer_trycomplete and usb_transfer_complete under the pipe (bus)
lock. In retrospect, the original probably should have been

if (trb stuff) {
xfer->ux_status = err;
usb_transfer_complete(xfer);
}

and my mechanical transformation should have worked, but also in
retrospect I should have put more thought into the change and done it
a little less mechanically.
 1.164  06-Apr-2022  mlelstv revert accidental last commit (except ukbd.c)
 1.163  06-Apr-2022  mlelstv remove debug printf
 1.162  13-Mar-2022  riastradh xhci(4): Serialize access to portsc registers.

Both xhci_roothub_ctrl and xhci_suspend/resume do r/m/w on them, so
use a mutex to serialize access to avoid stomping on each other.
 1.161  13-Mar-2022  riastradh xhci(4): Restore synchronous abort.

In revision 1.155, I made the logic to abort the hardware
asynchronous, under the misapprehension that it is necessary for
ubm_abortx not to release the bus lock.

Not only is this not necessary, but it is harmful to for the logic to
be asynchronous because the caller assumes the hardware won't use any
DMA buffers by the time ubm_abortx has returned so it is safe to
recycle them -- which is false if we don't synchronously wait for the
hardware to stop.
 1.160  09-Mar-2022  riastradh xhci(4): Avoid holding bus lock across usb_delay_ms.

We may still need a mechanism to serialize access to the portsc
registers between xhci_roothub_ctrl and xhci_suspend/resume, but the
bus lock is no longer that, and holding the bus lock across
usb_delay_ms may lead to deadlock by blocking the softints that wake
usb_delay_ms.
 1.159  09-Mar-2022  riastradh usb: Provisionally release bus lock around ubm_rhctrl.

This isn't quite correct, but it avoids a deadlock:

- *_roothub_ctrl holds bus lock, waits in usb_delay_ms for kpause
- softint waits for bus lock, holds up kpause wakeup

The deadlock is new since recent changes to hold the bus lock over
upm_start/upm_transfer. Making this change regresses to other
problems:

- *_suspend/resume and *_roothub_ctrl often touch the same portsc
registers

- roothub_ctrl_abort needs to wait for ubm_rhctrl to complete.

When the bus lock was held across both, a noop served here, but we
can't hold the bus lock across both, so that doesn't work.

However, these problems -- which we've had for a long time -- seem to
be less bad than the deadlock. So let's avoid the deadlock for now
and then work out another way to serialize suspend/resume/rhctrl and
aborts.

Candidate fix for PR kern/56739.
 1.158  03-Mar-2022  riastradh usb: Hold pipe lock across upm_transfer and upm_start.

This simplifies the code and fixes races with abort. Access to the
pipe's queue is now done exclusively while the pipe is locked.
 1.157  03-Mar-2022  riastradh usb: Factor usb_transfer_complete out of ubm_abortx method.
 1.156  03-Mar-2022  riastradh usb: Factor usb_insert_transfer out of upm_transfer and make private.

Almost every upm_transfer function starts with:

mutex_enter(&sc->sc_lock);
err = usb_insert_transfer(xfer);
mutex_exit(&sc->sc_lock);
if (err)
return err;

Some of them have debug messages sprinkled in here too, or assert
that err == USBD_NORMAL_COMPLETION (alternative is USBD_IN_PROGRESS,
only for pipes with up_running or up_serialise, presumably not
applicable for these types of pipes). Some of them also assert
xfer->ux_status == USBD_NOT_STARTED, which is guaranteed on entry and
preserved by usb_insert_transer.

Exceptions:

- arch/mips/adm5120/dev/ahci.c ahci_device_isoc_transfer just returns
USBD_NORMAL_COMPLETION, but I'm pretty sure this is and always has
been broken anyway, so won't make anything worse (if anything, might
make it better...)

- external/bsd/dwc2/dwc2.c dwc2_device_bulk_transfer and
dwc2_device_isoc_transfer _also_ issue dwc2_device_start(xfer)
under the lock. This is probably a better way to do it, but let's
do it uniformly across all HCIs at once.

- rump/dev/lib/libugenhc/ugenhc.c rumpusb_device_bulk_transfer
sometimes returns USBD_IN_PROGRESS _without_ queueing the transfer,
in the !rump_threads case. Not really sure how this is supposed to
work... If it actually breaks anything, we can figure it out.
 1.155  29-Jan-2022  riastradh xhci(4): Fix handling of endpoint reset/stop.

Use the same asynchronous task resetting a stalled/halted endpoint
and stopping a running endpoint -- either way we need to put the
endpoint back into a known state and, if there are transfers waiting
to run, start them up again.

- xhci_abortx must not drop the pipe (bus) lock -- usbdi(9) requires
this. So arrange to stop the endpoint and empty the queue
asynchronously.

- If the xhci softint claims an xfer for completion with
usbd_xfer_trycomplete, it must call usb_transfer_complete without
ever releasing the pipe (bus) lock -- it can't claim the xfer and
then defer the usb_transfer_complete to a task. So arrange to
reset the endpoint asynchronously, hold up new transfers until the
endpoint has been reset, and then do usb_transfer_complete
immediately.
 1.154  25-Jan-2022  msaitoh Accept USB 3.2 in xhci_id_protocols().
 1.153  16-Jan-2022  riastradh xhci(4): Fix citation in comment.

4.15.2.1 is device-initiated resume, but we are doing host-initiated.

No functional change.
 1.152  16-Jan-2022  riastradh xhci(4): Fix copypasto in resume error message.
 1.151  21-Dec-2021  skrll Change the usb_mem API to take a bus_dma_tag_t in usb_allocmem instead of
a struct usbd_bus *.

This allows an HCD to use more than one tag.
 1.150  23-Oct-2021  jakllsch xhci: Always read/write both halves of 64-bit registers

Access the 64-bit registers in their entirety at all times, independent
of the AC64 parameter of the hardware implementation.

Makes xhci(4) work on SolidRun MACCHIATObin (Marvell 8040 SoC)
 1.149  10-Oct-2021  jmcneill xhci: shrink port reset delay

Instead of waiting for USB_PORT_ROOT_RESET_DELAY (250ms) before checking
the port reset status, poll the PORTSC register and return early if the
reset has completed.
 1.148  17-Aug-2021  andvar fix multiplei repetitive typos in comments, messages and documentation. mainly because copy paste code big amount of files are affected.
 1.147  13-Aug-2021  andvar fix typos in words "pointer" and s/fram /frame/
 1.146  12-Jun-2021  riastradh usb(4): Fix racy endpoint reference counting.

Rules:

1. After usbd_setup_pipe*, must usbd_kill_pipe.
2. After usbd_open_pipe*, must usbd_close_pipe.

Still haven't merged the logic in usbd_kill_pipe and usbd_close_pipe,
but getting closer.
 1.145  12-Jun-2021  riastradh usb(4): Sprinkle kernel lock assertions.
 1.144  06-Jun-2021  jdolecek also reset xs->xs_xr[dci] to NULL when closing the pipe

avoids KASSERT() on next xhci_open()

should fix PR kern/56194 by nia
 1.143  29-May-2021  riastradh xhci(4): Wait USB_RESUME_WAIT ms, not 20 ms.

Better to use the named constant, and although the spec says 20 ms is
enough, apparently for some devices it's not.
 1.142  27-May-2021  skrll Fix error handling botch in 1.138. Thanks to jmcneill@ for spotting it.
 1.141  26-May-2021  riastradh xhci: Fix logic in waiting for command queue access.

_Either_ an existing command in progress, _or_ an existing suspend in
progress that is not done by us, should block us; the logic I wrote
previously erroneously blocked only if both conditions happened at
the same time.

Should fix issue reported by Andrius V in the PR kern/56050 followup
discussion.
 1.140  23-May-2021  riastradh xhci(4): Block commands and issue Stop Endpoint on suspend.
 1.139  23-May-2021  riastradh xhci(4): Draft suspend/resume.

Work almost entirely done and tested by maya@ based on xhci 1.2 spec;
tidied up and tweaked by me.

Not sure about issuing Stop Endpoint commands or ensuring the Command
Ring is in the Stopped or Idle state, but this seems to work as is,
so it's already an improvement over what we had before which was no
xhci suspend/resume at all.

In particular, it's not clear to us:

- if we don't have any pending USB activity whether we need to issue
the Stop Endpoints or quiesce the command ring; but

- if we do have any pending USB activity whether issuing Stop
Endpoint is enough or whether we also need to do anything to
synchronize with other software logic to quiesce it too.
 1.138  05-Jan-2021  skrll branches: 1.138.4; 1.138.6;
More converstion from usbd_status to int for function error reporting.
This time it's the turn of usb_allocmem.
 1.137  02-Jan-2021  jmcneill Use USBMALLOC_ZERO to ensure that all DMA memory is zero initialized.
 1.136  22-Dec-2020  riastradh usb: Omit bogus assertions about struct usbd_pipe::up_intrxfer.

These assertions were only valid for pipes at UE_IN_DIR, UE_INTERRUPT
endpoints created with usbd_open_pipe_intr, which uses up_intrxfer to
pass the struct usbd_xfer object to usbd_close_pipe to free later.

In contrast, for pipes at UE_OUT_DIR, UE_INTERRUPT endpoints,
up_intrxfer is never initialized, so the assertion cannot be right.
In principle we might even have more than one outstanding interrupt
transfer at a time, rendering the point of the assertion moot anyway.

Found by interrupting a uhidev write to a u2f device.

ok nick
 1.135  15-Oct-2020  jmcneill branches: 1.135.2;
Rename xhci_op_barrier to xhci_barrier and remove offset/length parameters.
Barriers will be applied to the entire XHCI register space.
 1.134  21-Aug-2020  jakllsch xhci(4): initial support for Isochronous pipes

Needs more cleanup, and more work, particularly transaction scheduling
may need to be examined with bus analyzer (to check if too frequent/
infrequent) or various isoc pipe consumers (to check if too infrequent)
and fixed if incorrect.

Begins to address PR kern/55574.
 1.133  21-Aug-2020  jakllsch xhci(4): copy HCCPARAMS regs to softc, use when needed

We'll need more than just the AC64 bit for the Isoc support coming up;
and we don't need to keep all the bits as bools in the softc.
 1.132  06-Jun-2020  skrll More __BITS. NFCI.
 1.131  04-Jun-2020  skrll Near complete conversion to __BITS
 1.130  01-Jun-2020  skrll Remove some unnecessary bit shifts by using appropriate read size
 1.129  21-May-2020  jakllsch xhci(4): only usbd_xfer_trycomplete() if certain to usb_transfer_complete()

Previously we'd cancel the timeout after the data stage of control xfers, and
then hope that the status stage would eventually interrupt.
 1.128  21-May-2020  jakllsch xhci(4): dynamically allocate xhci_xfer->xx_trb buffer

Allocation size remains the same for now.
 1.127  21-May-2020  jakllsch xhci(4): Check for overflows with KASSERT when putting TRBs on xfer TRB buffer.

Also more-correctly check for overruns when putting TRBs on a ring.
 1.126  21-May-2020  jakllsch xhci(4): Be more concise in putting the setup request in TRB immediate data.

Or: memcpy(),htole64() is replaced by functionally-identical le64dec().
 1.125  20-May-2020  jakllsch Use more-semantically-correct types in xhci allocx/freex functions.
 1.124  05-Apr-2020  skrll Switch USB to use non-coherent buffers for data transfers in the
same way as OpenBSD.

The use of coherent (uncacheable on ARM and other arches) mappings
for transfer buffers impacts performance, espcially where memcpys
are involved.

Audit the necessary usb_syncmem operations - a few were missing.
 1.123  02-Apr-2020  skrll Reduce the memory footprint by allocating a ring per endpoint/pipe on
pipe open.

From sc.dying on tech-kern
 1.122  14-Mar-2020  christos fix more broken kernhist formats (now I got them all).
 1.121  14-Mar-2020  christos revert the 0x% -> %# change for fixed width formats pointed out by uwe.
 1.120  13-Mar-2020  christos PR/55068: sc.dying: Fix printf formats:
- no %s/%p for kernel log
- 0x% -> %#
- always %j for kernel log
 1.119  15-Feb-2020  skrll Change bNbrPorts for loop to start from 1 to match others. NFCI.
 1.118  15-Feb-2020  riastradh Fix mistakes in previous sloppy change with root intr xfers.

- Make sure ux_status is set to USBD_IN_PROGRESS when started.
Otherwise, if it is still in flight when we abort the pipe,
usbd_ar_pipe will skip calling upm_abort.

- Initialize ux_status under the lock; in principle a completion
interrupt (or a delay) could race with the initialization.

- KASSERT that the xfer is in progress when we're about to complete
it.

Candidate fix for PR kern/54963 for other HCI drivers than uhci.

ok nick
ok phone

(This is the change that nick evidently MEANT to ok when he ok'd the
previous one!)
 1.117  12-Feb-2020  riastradh Fix steady state of root intr xfers.

Why?

- Avoid completing a root intr xfer multiple times in races.
- Avoid potential use-after-free in poll_hub callouts (uhci, ahci).

How?

- Use sc->sc_intr_xfer or equivalent to store only a pending xfer
that has not yet completed -- whether successfully, by timeout, or
by synchronous abort. When any of those happens, set it to null
under the lock, so the xfer is completed only once.

- For hci drivers that use a callout to poll the root hub (uhci, ahci):

. Pass the softc pointer, not the xfer, to the callout, so the
callout is not even tempted to use xfer after free -- if the
callout fires, but the xfer is synchronously aborted before the
callout can do anything, the xfer might be freed by the time the
callout starts to examine it.

. Teach the callout to do nothing if it is callout_pending after it
has fired. This way:

1. completion or synchronous abort can just callout_stop
2. start can just callout_schedule

If the callout had already fired before (1), and doesn't acquire
the bus lock until after (2), it may be tempted to abort the new
root intr xfer just after submission, which would be wrong -- so
instead we just have the callout do nothing if it notices it has
been rescheduled, since it will fire again after the appropriate
time has elapsed.
 1.116  12-Feb-2020  riastradh Factor out HCI-independent xfer completion logic.

New API for HCI drivers to synchronize hardware completion
interrupts, synchronous aborts, and asynchronous timeouts:

- When submitting an xfer to hardware, call
usbd_xfer_schedule_timeout(xfer).

- On HCI completion interrupt for xfer completion:

if (!usbd_xfer_trycomplete(xfer))
return; /* timed out or aborted, ignore it */

- In upm_abort methods, call usbd_xfer_abort(xfer).

For HCI drivers that use this API (not needed in drivers that don't,
or for xfers like root intr xfers that don't use it):

- New ubm_abortx method serves role of former *hci_abort_xfer, but
without any logic for wrangling timeouts/callouts/tasks -- caller
in usbd_xfer_abort has already handled them.

- New ubm_dying method, returns true if the device is in the process
of detaching, used by the timeout logic.

Converted and tested:
- ehci
- ohci

Converted and compile-tested:
- ahci (XXX did this ever work?)
- dwc2
- motg (XXX missing usbd_xfer_schedule_timeout in motg_*_start?)
- uhci
- xhci

Not changed:

- slhci (sys/dev/ic/sl811hs.c) -- doesn't use a separate per-xfer
callout for timeouts (XXX but maybe should?)

- ugenhc (sys/rump/dev/lib/libugenhc/ugenhc.c) -- doesn't manage its
own transfer timeouts

- vhci -- times transfers out only on detach; could be adapted easily
if we wanted to use the xfer->ux_callout
 1.115  29-Dec-2019  skrll branches: 1.115.2;
Acquire bus lock (if not polling) when changing xfer status, resetting
the timeout and ringing the doorbell in the start methods.
 1.114  08-Sep-2019  mrg avoid non-debug kernel build issues.
 1.113  07-Sep-2019  mrg fix ryzen usb issue: we set TD size to '1', where has xhci spec 4.11.2.4
says final TRB for a TD should have this set to '0'. since we currently
only generate sinel TRB TDs, set this to 0.

XXX: pullup-all

from sc.dying
 1.112  07-Sep-2019  mrg fix 32-bit debug build, and also vmstat -y. reported by sc.dying.
 1.111  28-Aug-2019  mrg introduce and use XHCIHIST_CALLARGS(). reduces lots of double logs,
and includes useful info in all cases instead of just "called".

add a couple of more logs that i've wanted while debugging ryzen3
vs USB issues.
 1.110  23-Aug-2019  mrg convert a debug message to an error message. will help more quickly
identify next time xhci version changes.
 1.109  21-Aug-2019  mrg match xhci version 3.10. allows properly finding all the USB
busses on new ryzen 3 based systems.

unfortunately, the USB busses are still non-functional.
 1.108  20-Aug-2019  skrll typo in error message
 1.107  08-May-2019  mrg branches: 1.107.2;
new comment explains this hack:

* XXXMRG: Stall task can run after slot is disabled when yanked.
* This hack notices that the xs has been memset() in
* xhci_disable_slot() and returns. Both xhci_reset_endpoint()
* and xhci_set_dequeue() rely upon a valid ring setup for correct
* operation, and the latter will fault, as would
* usb_transfer_complete() if it got that far.

this truly fixes yanking active ucom devices (uchcom, uplcom and
umcs all tested.) prior efforts largely worked by allowing the
stall task to run first, as a side effect to their actual work.

ok @skrll.

XXX: pullup.
 1.106  17-Feb-2019  rin Fix assertion failures triggered by usbdi.c,v 1.182, when devices
are detached.

This is because xfers of USBD_NOT_STARTED can be removed from queue
in an invisible way to host controller drivers.

Discussed on tech-kern.
 1.105  22-Jan-2019  skrll Revert the KASSERT
 1.104  22-Jan-2019  skrll Add a KASSERT
 1.103  11-Jan-2019  skrll Add a bs_barrier to make ThunderX xhci work. Not sure why this is needed,
but I'll work it out later.
 1.102  11-Jan-2019  skrll Style
 1.101  07-Jan-2019  jakllsch Seperate xHCI and xhci(4) TRB structs so as to avoid some of them
(the ones that the hardware doesn't touch) ending up less-aligned
than the compiler assumed.

Additionally, fix the most obvious problems that xhci(4) had on
big endian systems.

Briefly tested on a Jetson TK1 in LE and BE w/ evbarm GENERIC kernel.
 1.100  28-Oct-2018  mrg use xhci_polling_p(). this might miss when bus2 is active.

XXX: pullup-7, pullup-8.
 1.99  16-Sep-2018  mrg consolidate the handling of polling across HC drivers, and generic USB:
- don't take mutexes if polling
- normalise the code across all drivers
- add some not yet code to block discovery to/from polling
- minor CSE
- adjust comment for usbd_set_polling() to reality now i properly
understand what it is used for and why.

this, with a hack to make RB_ASKNAME to wait 5 seconds allows boot -a
work with USB keyboards. there are still multiple issues remaining:
- discovery and polling need to be mutually exclusive
- attachment of ukbd and wskbd is not handled by config_pending, and
the 5 second delay isn't going to always be enough.
 1.98  03-Sep-2018  riastradh Rename min/max -> uimin/uimax for better honesty.

These functions are defined on unsigned int. The generic name
min/max should not silently truncate to 32 bits on 64-bit systems.
This is purely a name change -- no functional change intended.

HOWEVER! Some subsystems have

#define min(a, b) ((a) < (b) ? (a) : (b))
#define max(a, b) ((a) > (b) ? (a) : (b))

even though our standard name for that is MIN/MAX. Although these
may invite multiple evaluation bugs, these do _not_ cause integer
truncation.

To avoid `fixing' these cases, I first changed the name in libkern,
and then compile-tested every file where min/max occurred in order to
confirm that it failed -- and thus confirm that nothing shadowed
min/max -- before changing it.

I have left a handful of bootloaders that are too annoying to
compile-test, and some dead code:

cobalt ews4800mips hp300 hppa ia64 luna68k vax
acorn32/if_ie.c (not included in any kernels)
macppc/if_gm.c (superseded by gem(4))

It should be easy to fix the fallout once identified -- this way of
doing things fails safe, and the goal here, after all, is to _avoid_
silent integer truncations, not introduce them.

Maybe one day we can reintroduce min/max as type-generic things that
never silently truncate. But we should avoid doing that for a while,
so that existing code has a chance to be detected by the compiler for
conversion to uimin/uimax without changing the semantics until we can
properly audit it all. (Who knows, maybe in some cases integer
truncation is actually intended!)
 1.97  30-Aug-2018  jakllsch silence "not known to be supported" message for all version 1.x xHCIs
 1.96  09-Aug-2018  mrg pull across abort fixes from nick-nhusb. add more abort fixes, using
ideas from Taylor and Nick, and myself. special thanks to both who
inspired much of the code here, if not wrote it directly.

among other problems, this assert should no longer trigger:

panic: kernel diagnostic assertion "xfer->ux_state == XFER_ONQU" failed: file "/current/src/sys/dev/usb/usbdi.c", line 914

using usbhist i was able to track down my instance of it being related
to userland close() beginning, dropping the sc_lock, and then the usb
softintr completes the transfer normally, and when it is done, the
abort path attempts to re-complete the transfer, and the above assert
is tripped.


changes from nhusb were commited with these logs:
--
Move the struct usb_task to struct usbd_xfer for everyone to use.
--
Set device transfer status to USBD_IN_PROGRESS if start methods succeeds
--
Actually set the transfer status on transfers in ohci_abort_xfer and
the controller is dying
--
Don't supply the lock to callout_halt when polling as it won't be held
--
Improve transfer abort
--
Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
--
#ifdef DIAGNOSTIC -> KASSERT and add another KASSERT
--
Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
--

additional changes include:
- initialise the usb abort task in the HCI allocx routine, so that it
can be safely usb_rem_task()'d.
- rework the handling of softintr vs cancellation vs timeout abort based
upon a scheme from Taylor:
when completing a transfer normally:
- if the status is not in progress, it must be cancelled or timed out,
and we should not process this xfer.
- set the status as normal.
- unconditionallly callout_stop() and usb_rem_task(). they're safe and
either aren't running, or will run and do nothing.
- finally call usb_transfer_complete().
when aborting a transfer:
- status should be cancelled or timed out.
- if cancelling, callout_halt and usb_rem_task_wait() to make sure the
timer is either done or cancelled.
- at this point, the ux_status must not be cancelled or timed out, and
if it is not in progress we're done.
- set the status.
- if the controller is dying, just return.
- perform HCI-specific tasks to abort this xfer.
- finally call usb_transfer_complete().
for the timeout and timeout task:
- if the HCI is not dying, and the ux_status is in progress, then
trigger the usb abort task.
- remove UXFER_ABORTWAIT and UXFER_ABORTING.

tested on:
- multiple PC systems with several types of devices: ugen/UPS, ucom,
umass with disk, ssd and cdrom backends, kbd, ms, using uhci, ehci
and xhci.
- erlite3: sd@umass on dwc2.
- sunblade2000: kbd/ms and umass disk on ohci.

untested:
- motg, slhci and ahci. motg has some portion of the new scheme
applied, but slhci and ahci require more study.

future work includes pushing a lot of the common abort handling into
usbdi.c and leaving upm_abort() for HC specific tasks, but this change
is pullup-able to netbsd-7 and netbsd-8 as it does not change any
external API, as well as removing over 100 lines of code while adding
over 30 new asserts.

XXX: pullup-7, pullup-8.
 1.95  18-Jul-2018  msaitoh Read xHCI 1.1's HCCPARAMS2 registar and print it with aprint_debug_dev().
e.g.: xhci0: hcc2=0x7d<ETC,CIC,LEC,CTC,FSC,U3C>
 1.94  16-Jul-2018  christos make typing on the console work during boot:
1. if polling avoid mutex
2. convert spin mutex calls to regular calls in xhci_poll so that we don't
spinout while waiting for the user to type.
Tested by phil@
 1.93  29-Jun-2018  msaitoh Detect USB 3.1.
 1.92  14-May-2018  jakllsch branches: 1.92.2;
With the hope it fixes PR #52911, be extra paranoid about assuring the
Device Context DMA memory is sync'd immediately before reading from it.
 1.91  02-May-2018  jmcneill "Root Hub" -> "root hub" for consistency with other HC drivers
 1.90  23-Apr-2018  jdolecek enable code to only trigger usb processing when EINT is set, to
avoid misinterpreting shared interrupt for another device

when clearing USBSTS, actually preserve the bits which spec requires to
preserve, and actually clear bit 1, which should be actually always
cleared to zero by spec

also #ifdef XHCI_DEBUG some unnecessary register reads

this should finally resolve PR kern/53066 also for Martin
 1.89  22-Apr-2018  jdolecek trigger the softint processing on that child bus which is not detached yet

fixes PR kern/53066 by Martin Husemann
 1.88  21-Apr-2018  jdolecek add KASSERT() that sc_child* is set to NULL after child detach; just for
readability, it's not immediatelly obvious this is done in xhci_childdet()

no functional changes
 1.87  09-Apr-2018  jakllsch Stop potential misuse of vendor names and USB vendor IDs in root hub
device and string descriptors.

Firstly: Few vendors have identical PCI-SIG vendor IDs and USB-IF vendor
IDs. As such, using the PCI vendor ID as a USB vendor ID may trample
on whomever is allocated that USB vendor ID.

Secondly: The vendor of the host controller hardware implementation has
little to nothing to do with our usbroothub implementation. Thus we
should not potentially associate any problems therewith to such third
party.

This change will result in root hubs being identified by USB Vendor ID
0x0000. Root hub vendor string will now be "NetBSD" (or, specifically:
ostype). Product ID (0x0000) and product strings remain unchanged.
 1.86  07-Feb-2018  prlw1 branches: 1.86.2;
xhci.c: avoid unused variable in non-DIAGNOSTIC build
 1.85  01-Feb-2018  msaitoh Fix KASSERT panic. When detaching USB host controoler, the following panic
occured:
kernel diagnostic assertion "sc->sc_intrxfer == xfer" failed: file "../../../../dev/usb/ohci.c", line 1720
Clear sc_intrxfer correctly. Adviced by Nick and tested by me.
 1.84  30-Jan-2018  msaitoh Avoid panic while detaching xhci. The xhci driver has both sc_child and
sc_child2 but xhci_childdet() only supported sc_child. OK'd by Nick.
 1.83  20-Dec-2017  skrll Fix the logic around TT Hub Slot ID, TT Port Number and Multi-TT in
xhci_setup_tthub. TT Hub/Port need to reference the upstream HS hub and
Multi-TT needs to reflect the capabilities of a HS hub itself or the same
upstream HS hub.
 1.82  19-Dec-2017  skrll Trailing whitespace
 1.81  10-Dec-2017  hannken Use USB_DEBUG, not XHCI_DEBUG to protect `dci', DPRINTFN() always uses it.
 1.80  09-Dec-2017  christos adjust for hexdump signature
 1.79  08-Dec-2017  christos coalesce the two copies of hexdump into libkern
 1.78  07-Dec-2017  christos PR/52791: Robert Sprowson: avoid duplicate tests (use accessor variables)
 1.77  17-Nov-2017  skrll s/PR_NOWAIT/PR_WAITOK/ in HCD allocx (allocate xfer) method
 1.76  09-Nov-2017  msaitoh Wait 1ms first. Existing Intel xHCI requies 1ms delay to prevent system hang
(Errata).

XXX pullup-[78]
 1.75  28-Oct-2017  pgoyette Update the kernhist(9) kernel history code to address issues identified
in PR kern/52639, as well as some general cleaning-up...

(As proposed on tech-kern@ with additional changes and enhancements.)

Details of changes:

* All history arguments are now stored as uintmax_t values[1], both in
the kernel and in the structures used for exporting the history data
to userland via sysctl(9). This avoids problems on some architectures
where passing a 64-bit (or larger) value to printf(3) can cause it to
process the value as multiple arguments. (This can be particularly
problematic when printf()'s format string is not a literal, since in
that case the compiler cannot know how large each argument should be.)

* Update the data structures used for exporting kernel history data to
include a version number as well as the length of history arguments.

* All [2] existing users of kernhist(9) have had their format strings
updated. Each format specifier now includes an explicit length
modifier 'j' to refer to numeric values of the size of uintmax_t.

* All [2] existing users of kernhist(9) have had their format strings
updated to replace uses of "%p" with "%#jx", and the pointer
arguments are now cast to (uintptr_t) before being subsequently cast
to (uintmax_t). This is needed to avoid compiler warnings about
casting "pointer to integer of a different size."

* All [2] existing users of kernhist(9) have had instances of "%s" or
"%c" format strings replaced with numeric formats; several instances
of mis-match between format string and argument list have been fixed.

* vmstat(1) has been modified to handle the new size of arguments in the
history data as exported by sysctl(9).

* vmstat(1) now provides a warning message if the history requested with
the -u option does not exist (previously, this condition was silently
ignored, with only a single blank line being printed).

* vmstat(1) now checks the version and argument length included in the
data exported via sysctl(9) and exits if they do not match the values
with which vmstat was built.

* The kernhist(9) man-page has been updated to note the additional
requirements imposed on the format strings, along with several other
minor changes and enhancements.

[1] It would have been possible to use an explicit length (for example,
uint64_t) for the history arguments. But that would require another
"rototill" of all the users in the future when we add support for an
architecture that supports a larger size. Also, the printf(3) format
specifiers for explicitly-sized values, such as "%"PRIu64, are much
more verbose (and less aesthetically appealing, IMHO) than simply
using "%ju".

[2] I've tried very hard to find "all [the] existing users of kernhist(9)"
but it is possible that I've missed some of them. I would be glad to
update any stragglers that anyone identifies.
 1.74  25-Sep-2017  jmcneill If the bus glue calls xhci_init with interrupts enabled, there is a window
between when xhci interrupts are enabled and the usb bus driver is
attached. If an irq occurs in this window, xhci will attempt to schedule
a softint with an invalid softint handle.

Add a quirk flag, XHCI_DEFERRED_START, that when set skips starting the
controller at the end of xhci_init. Bus glue that sets this is responsible
to call xhci_start after attaching the child usb devices.
 1.73  22-Aug-2017  skrll Check both buses for ub_usepolling and schedule the soft interrupt handler
or call it directly. Fixes usb keyboard in ddb when attached to xhci.
 1.72  01-Jun-2017  chs branches: 1.72.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.
 1.71  13-Apr-2017  skrll PR/52154: xHCI Slot Context Not Include in configure endpoint process

eXtensible Host Controller Interface Specification(rev 1.1), 4.6.6:

"A0 shall be set to 1 and refer to section 6.2.2.2 for the Slot Context fields
used by the Configure Endpoint Command."
 1.70  21-Jan-2017  skrll Fix a kmem_{zalloc,free} size
 1.69  19-Jan-2017  skrll Fix build
 1.68  19-Jan-2017  skrll Pull across xhci(4) improvemnts from nick-nhusb
 1.67  03-Sep-2016  skrll branches: 1.67.2;
Fix "spurious event" when command ring rolls over.
When the enqueue pointer of command ring points at the last TRB,
address of Link TRB was stored in sc_command_addr.
It should be address of 0th TRB of ring.

From t-hash
 1.66  03-Sep-2016  skrll Suppress stopping-endpoint in xhci_close_pipe if the endpoint is
already stopped.

From t-hash
 1.65  18-Aug-2016  skrll Put Link TRB always at the end of ring. Should fix ctrl xfer problem on
Intel xHC.

From t-hash.
 1.64  18-Aug-2016  skrll Byte swap the address in xhci_new_device appropriately. From t-hash
 1.63  17-Jul-2016  skrll Various improvement from t-hash (and a bit from me).

- Flesh out the abort routine
- Avoid using EVENT_DATA
- Miscellaneous debug
- Comment improvement
 1.62  08-Jul-2016  skrll branches: 1.62.2;
Don't ignore return value of xhci_update_ep0_mps and remove KASSERT at
the end of that function.

From t-hash
 1.61  08-Jul-2016  skrll More debug in new_device.

From t-hash.
 1.60  08-Jul-2016  skrll s/UBS/USB/

From t-hash
 1.59  07-Jul-2016  maya Turn a few KASSERTs into KASSERTMSG mentioning the tested values

Some of these values are optimized out, so knowing why the test failed
can be challenging.
 1.58  12-Jun-2016  skrll Remove KASSERTs that don't make sense.
 1.57  10-Jun-2016  skrll Add callout_stop while clearing stall. From t-hash.
 1.56  05-Jun-2016  skrll PR/51202: XHCI driver sends stale TRBs from a pipe previously opened

Use xhci_host_dequeue in xhci_set_dequeue
 1.55  05-Jun-2016  skrll PR/51202: XHCI driver sends stale TRBs from a pipe previously opened

A similar patch was sent to me by t-hash and the xhci_setup_ctx is taken
from that.
 1.54  05-Jun-2016  skrll Update some comments. Some from t-hash with updates from me.
 1.53  05-Jun-2016  skrll More debug
 1.52  05-Jun-2016  skrll Replace num of TRBs with xr->xr_ntrb. From t-hash.
 1.51  05-Jun-2016  skrll From t-hash
+ Gather up setting slot and device context into xhci_setup_ctx(),
and split out setting route string and TT hub params.
+ Apply htole{32,64} to contexts at once as possible.
+ Use xhci_setup_ctx in xhci_set_address and xhci_configure_endpoint.
 1.50  05-Jun-2016  skrll Wrap long lines and other whitespace. From t-hash.
 1.49  05-Jun-2016  skrll Reduce the scope of some variable. From t-hash
 1.48  05-Jun-2016  skrll From t-hash
+ Split out setting address device.
+ Split out freeing slot structure.
+ Use xhci_free_slot() in xhci_disable_slot().
 1.47  31-May-2016  mlelstv Flush memory correctly during ring put.
See PR 51199.

From Sprow, ok skrll@.
 1.46  10-May-2016  pooka sprinkle _KERNEL_OPT
 1.45  07-May-2016  skrll G/C more unused code
 1.44  06-May-2016  skrll Print LF in hexdump when datalen % 16 != 0. From t-hash.
 1.43  06-May-2016  skrll G/C some #if 0 code
 1.42  06-May-2016  skrll Remove magic numbers from xhci_hc_reset. From t-hash.
 1.41  30-Apr-2016  skrll Comment updates from t-hash
+ Add and update comments.
+ Update spec URLs.
 1.40  30-Apr-2016  skrll Updates from t-hash
+ Split out printing PSI and taking ownership.
+ Split out resetting HC.
+ Split out calculating TRB index.

XXX magic numbers
XXX waits
 1.39  30-Apr-2016  skrll Cancel command when command times out. From t-hash.

XXX interrupt context?
 1.38  30-Apr-2016  skrll Fix type of return value in xhci_do_command
 1.37  30-Apr-2016  skrll Updates from t-hash
+ Change xhci_init() returns errno.
+ Fix memory leaks in error paths.
+ Add more messages for critical errors.
 1.36  30-Apr-2016  skrll Wrap long line. From t-hash.
 1.35  30-Apr-2016  skrll Use XHCI_XFER2XXFER

From t-hash
 1.34  23-Apr-2016  skrll Merge nick-nhusb

- API / infrastructure changes to support memory management changes.
- Memory management improvements and bug fixes.
- HCDs should now be MP safe
- conversion to KERNHIST based debug
- FS/LS isoc support on ehci(4).
- conversion to kmem(9)
- Some USB 3 support - mostly from Takahiro HAYASHI (t-hash).
- interrupt transfers now get proper DMA operations
- general bug fixes
- kern/48308
- uhub status notification improvements
- umass(4) probe fix (applied to HEAD already)
- ohci(4) short transfer fix
 1.33  06-Jan-2016  skrll Get the iManufacturer, iProduct, and iSerialNumber strings before probing
for drivers and cache them for later use. This reduces bus transactions
and fixes attachment for at least two of my umass(4)s.
 1.32  10-Dec-2015  skrll Unwrap lines
 1.31  10-Dec-2015  skrll Fix indentation
 1.30  10-Dec-2015  skrll Trailing whitespace
 1.29  19-Aug-2015  skrll More IPL_SCHED -> IPL_USB
 1.28  18-Nov-2014  skrll branches: 1.28.2;
kern/49391: Fixes to XHCI driver command ring and status TRB

CRCR needs 64byte aligned address

Use usb_allocmem instead of usb_allocmem_flags(..., 0)

Correct status stage TRB in xhci_device_ctrl_start - direction logic was
wrong.

Correct a typo in a comment
 1.27  03-Oct-2014  skrll Convert to USB_HIST.

From Takahiro HAYASHI with tweaks from me.
 1.26  12-Aug-2014  skrll branches: 1.26.2;
Set the correct MPS for LS devices. From Takahiro HAYASHI.
 1.25  12-Aug-2014  skrll Serialise xhci_intr1 calls with sc_intr_lock. From Takahiro HAYASHI.
 1.24  11-Aug-2014  skrll PR/49091: xhci: wrong wMaxPacketSize value

While this is correct according to the specification only fixed sizes
are allowed, i.e. 512 for SS, etc. Maybe these should be used?
 1.23  05-Aug-2014  skrll branches: 1.23.2;
Remove #ifdef DIAGNOSTIC around sc from the root interrupt abort methods
now it's always used.
 1.22  05-Aug-2014  skrll Prevent the root hub sending updates once the root interrupt abort is
called.
 1.21  05-Aug-2014  skrll Simplify the freeing of the interrupt pipe transfer. Inspired by OpenBSD.
 1.20  26-Jul-2014  pgoyette Another __diagused
 1.19  15-Jul-2014  ozaki-r Unbreak the build for i386

__debugused => XHCI_DEBUG
 1.18  15-Jul-2014  ozaki-r Suppress another debug output

It's output when DDB.

Fix previous: DEBUG => XHCI_DEBUG
 1.17  14-Jul-2014  ozaki-r Suppress debug output of xhci_device_intr_done

It's annoying because it is output even on keyboard inputs.
 1.16  10-Mar-2014  skrll branches: 1.16.2; 1.16.4; 1.16.6;
Code style. No functional change.
 1.15  10-Mar-2014  skrll Don't abuse usbd_status in xhci_init
 1.14  10-Mar-2014  skrll Comment style.
 1.13  16-Jan-2014  dsl Add another XXX comment, I'm not in a postition to fix the code.
The 'cycle' bit of a LINK trb must not be changed until the cycle bit
of the following trb has been set.
 1.12  14-Dec-2013  jakllsch In xhci_init() use aprint_debug_dev() instead of device_printf().
 1.11  10-Dec-2013  dsl Add a few comments about some perversities of the xhci specification.
XXX: I can't see any code that ensures there is space in the rings.
XXX: Nothing stops buffer fragments crossing 64kB boundaries.
 1.10  17-Nov-2013  skrll Fix locking botch.

PR/48383
 1.9  13-Nov-2013  christos CID 1125827: Avoid buffer overrun (read past end of struct)
 1.8  10-Nov-2013  mrg add a default case in xhci_init_slot().
ok jakllsch.
 1.7  08-Nov-2013  christos fix unused variable warnings
 1.6  04-Nov-2013  skrll branches: 1.6.2;
PR/48366 Always zero the allocated memory in xhci_alloc, not just in
DIAGNOSTIC.
 1.5  28-Oct-2013  matt Add support for scratchpad buffers (required for some XHCI devices).
Fix an endian issue.
 1.4  18-Oct-2013  apb Wrap several unused static inline functions with #if 0.
I didn't delete them because I suspect that they might be needed later.
 1.3  18-Oct-2013  skrll Wrap some long lines.
 1.2  18-Oct-2013  apb Don't refer to uninitialised variable 'v'
in case C(UR_GET_DESCRIPTOR, UT_READ_CLASS_DEVICE)
in xhci_root_ctrl_start().
This code was apparently pasted from the ehci driver.

OK nick.
 1.1  14-Sep-2013  jakllsch Add work-in-progress xhci(4) driver code. Currently (mostly) supports
interrupt-driven control, interrupt and bulk transfers at the three USB
2.0 speeds on root hub ports.
 1.6.2.2  05-Nov-2013  matt Pull down xhci support from HEAD
 1.6.2.1  04-Nov-2013  matt file xhci.c was added on branch matt-nb5-mips64 on 2013-11-05 18:36:31 +0000
 1.16.6.2  22-May-2014  yamt sync with head.

for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.

this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
 1.16.6.1  10-Mar-2014  yamt file xhci.c was added on branch yamt-pagecache on 2014-05-22 11:40:37 +0000
 1.16.4.2  18-May-2014  rmind sync with head
 1.16.4.1  10-Mar-2014  rmind file xhci.c was added on branch rmind-smpnet on 2014-05-18 17:45:48 +0000
 1.16.2.1  10-Aug-2014  tls Rebase.
 1.23.2.9  17-Sep-2019  martin Pull up following revision(s) (requested by mrg in ticket #1707):

sys/dev/usb/xhci.c: revision 1.113
sys/dev/usb/xhci.c: revision 1.109

match xhci version 3.10. allows properly finding all the USB
busses on new ryzen 3 based systems.

unfortunately, the USB busses are still non-functional.

-

fix ryzen usb issue: we set TD size to '1', where has xhci spec 4.11.2.4
says final TRB for a TD should have this set to '0'. since we currently
only generate sinel TRB TDs, set this to 0.

XXX: pullup-all
from sc.dying
 1.23.2.8  25-Aug-2018  martin Pull up following revision(s) (requested by mrg in ticket #1632):

sys/dev/usb/usbdivar.h: revision 1.117
sys/external/bsd/dwc2/dwc2.c: revision 1.52
sys/dev/usb/xhcivar.h: revision 1.10
sys/dev/usb/motg.c: revision 1.22
sys/dev/usb/ehci.c: revision 1.260
sys/dev/usb/ehci.c: revision 1.261
sys/dev/usb/xhci.c: revision 1.96
sys/dev/usb/ohci.c: revision 1.282
sys/dev/usb/ohci.c: revision 1.283
sys/dev/usb/ehcivar.h: revision 1.45
sys/dev/usb/uhci.c: revision 1.281
sys/dev/usb/uhci.c: revision 1.282
sys/dev/usb/usbdi.c: revision 1.177
sys/dev/usb/ohcivar.h: revision 1.60
sys/dev/usb/uhcivar.h: revision 1.55
(all via patch)

pull across abort fixes from nick-nhusb. add more abort fixes, using
ideas from Taylor and Nick, and myself. special thanks to both who
inspired much of the code here, if not wrote it directly.

among other problems, this assert should no longer trigger:

panic: kernel diagnostic assertion "xfer->ux_state == XFER_ONQU" failed: file "/current/src/sys/dev/usb/usbdi.c", line 914

using usbhist i was able to track down my instance of it being related
to userland close() beginning, dropping the sc_lock, and then the usb
softintr completes the transfer normally, and when it is done, the
abort path attempts to re-complete the transfer, and the above assert
is tripped.

changes from nhusb were commited with these logs:
--
Move the struct usb_task to struct usbd_xfer for everyone to use.
--
Set device transfer status to USBD_IN_PROGRESS if start methods succeeds
--
Actually set the transfer status on transfers in ohci_abort_xfer and
the controller is dying
--
Don't supply the lock to callout_halt when polling as it won't be held
--
Improve transfer abort
--
Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
--
--
Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
--

additional changes include:
- initialise the usb abort task in the HCI allocx routine, so that it
can be safely usb_rem_task()'d.
- rework the handling of softintr vs cancellation vs timeout abort based
upon a scheme from Taylor:
when completing a transfer normally:
- if the status is not in progress, it must be cancelled or timed out,
and we should not process this xfer.
- set the status as normal.
- unconditionallly callout_stop() and usb_rem_task(). they're safe and
either aren't running, or will run and do nothing.
- finally call usb_transfer_complete().
when aborting a transfer:
- status should be cancelled or timed out.
- if cancelling, callout_halt and usb_rem_task_wait() to make sure the
timer is either done or cancelled.
- at this point, the ux_status must not be cancelled or timed out, and
if it is not in progress we're done.
- set the status.
- if the controller is dying, just return.
- perform HCI-specific tasks to abort this xfer.
- finally call usb_transfer_complete().
for the timeout and timeout task:
- if the HCI is not dying, and the ux_status is in progress, then
trigger the usb abort task.
- remove UXFER_ABORTWAIT and UXFER_ABORTING.

tested on:
- multiple PC systems with several types of devices: ugen/UPS, ucom,
umass with disk, ssd and cdrom backends, kbd, ms, using uhci, ehci
and xhci.
- erlite3: sd@umass on dwc2.
- sunblade2000: kbd/ms and umass disk on ohci.

untested:
- motg, slhci and ahci. motg has some portion of the new scheme
applied, but slhci and ahci require more study.

future work includes pushing a lot of the common abort handling into
usbdi.c and leaving upm_abort() for HC specific tasks, but this change
is pullup-able to netbsd-7 and netbsd-8 as it does not change any
external API, as well as removing over 100 lines of code while adding
over 30 new asserts.

XXX: pullup-7, pullup-8.

fix DIAGNOSTIC build by not copying ub_usepolling to stack before use

Sprinkle __diagused
 1.23.2.7  03-Jan-2018  snj Pull up following revision(s) (requested by skrll in ticket #1529):
sys/dev/usb/ehci.c: revision 1.257
sys/dev/usb/motg.c: revision 1.19
sys/dev/usb/ohci.c: revision 1.276
sys/dev/usb/uhci.c: revision 1.279
sys/dev/usb/xhci.c: revision 1.77
sys/external/bsd/dwc2/dwc2.c: revision 1.47
s/PR_NOWAIT/PR_WAITOK/ in HCD allocx (allocate xfer) method
 1.23.2.6  03-Jan-2018  snj Pull up following revision(s) (requested by msaitoh in ticket #1526):
sys/dev/usb/xhci.c: revision 1.76
Wait 1ms first. Existing Intel xHCI requies 1ms delay to prevent system hang
(Errata).
 1.23.2.5  05-Apr-2017  snj Pull up following revision(s) (requested by skrll in ticket #1395):
share/man/man4/axe.4: netbsd-7-nhusb
share/man/man4/axen.4: netbsd-7-nhusb
share/man/man4/cdce.4: netbsd-7-nhusb
share/man/man4/uaudio.4: netbsd-7-nhusb
share/man/man4/ucom.4: netbsd-7-nhusb
share/man/man4/uep.4: netbsd-7-nhusb
share/man/man4/urtw.4: netbsd-7-nhusb
share/man/man4/usb.4: netbsd-7-nhusb
share/man/man4/uyap.4: netbsd-7-nhusb
share/man/man4/xhci.4: netbsd-7-nhusb
share/man/man9/usbdi.9: netbsd-7-nhusb
sys/arch/amd64/conf/ALL: netbsd-7-nhusb
sys/arch/amd64/conf/GENERIC: netbsd-7-nhusb
sys/arch/amiga/dev/slhci_zbus.c: netbsd-7-nhusb
sys/arch/arm/allwinner/awin_otg.c: netbsd-7-nhusb
sys/arch/arm/allwinner/awin_usb.c: netbsd-7-nhusb
sys/arch/arm/amlogic/amlogic_dwctwo.c: netbsd-7-nhusb
sys/arch/arm/at91/at91ohci.c: netbsd-7-nhusb
sys/arch/arm/broadcom/bcm2835_dwctwo.c: netbsd-7-nhusb
sys/arch/arm/broadcom/bcm53xx_usb.c: netbsd-7-nhusb
sys/arch/arm/ep93xx/epohci.c: netbsd-7-nhusb
sys/arch/arm/gemini/obio_ehci.c: netbsd-7-nhusb
sys/arch/arm/imx/files.imx23: netbsd-7-nhusb
sys/arch/arm/imx/imxusb.c: netbsd-7-nhusb
sys/arch/arm/imx/imxusbreg.h: netbsd-7-nhusb
sys/arch/arm/omap/obio_ohci.c: netbsd-7-nhusb
sys/arch/arm/omap/omap3_ehci.c: netbsd-7-nhusb
sys/arch/arm/omap/omapl1x_ohci.c: netbsd-7-nhusb
sys/arch/arm/omap/tiotg.c: netbsd-7-nhusb
sys/arch/arm/s3c2xx0/ohci_s3c24x0.c: netbsd-7-nhusb
sys/arch/arm/samsung/exynos_usb.c: netbsd-7-nhusb
sys/arch/arm/xscale/pxa2x0_ohci.c: netbsd-7-nhusb
sys/arch/arm/zynq/zynq_usb.c: netbsd-7-nhusb
sys/arch/hpcarm/dev/nbp_slhci.c: netbsd-7-nhusb
sys/arch/hpcmips/dev/plumohci.c: netbsd-7-nhusb
sys/arch/i386/conf/ALL: netbsd-7-nhusb
sys/arch/i386/conf/GENERIC: netbsd-7-nhusb
sys/arch/i386/pci/gcscehci.c: netbsd-7-nhusb
sys/arch/luna68k/conf/GENERIC: netbsd-7-nhusb
sys/arch/mips/adm5120/dev/ahci.c: netbsd-7-nhusb
sys/arch/mips/adm5120/dev/ahcivar.h: netbsd-7-nhusb
sys/arch/mips/alchemy/dev/ohci_aubus.c: netbsd-7-nhusb
sys/arch/mips/atheros/dev/ehci_arbus.c: netbsd-7-nhusb
sys/arch/mips/atheros/dev/ohci_arbus.c: netbsd-7-nhusb
sys/arch/mips/conf/files.adm5120: netbsd-7-nhusb
sys/arch/mips/ralink/ralink_ehci.c: netbsd-7-nhusb
sys/arch/mips/ralink/ralink_ohci.c: netbsd-7-nhusb
sys/arch/mips/rmi/rmixl_ehci.c: netbsd-7-nhusb
sys/arch/mips/rmi/rmixl_ohci.c: netbsd-7-nhusb
sys/arch/playstation2/dev/ohci_sbus.c: netbsd-7-nhusb
sys/arch/powerpc/booke/dev/pq3ehci.c: netbsd-7-nhusb
sys/arch/powerpc/ibm4xx/dev/dwctwo_plb.c: netbsd-7-nhusb
sys/arch/x68k/dev/slhci_intio.c: netbsd-7-nhusb
sys/conf/files: netbsd-7-nhusb
sys/dev/cardbus/ehci_cardbus.c: netbsd-7-nhusb
sys/dev/cardbus/ohci_cardbus.c: netbsd-7-nhusb
sys/dev/cardbus/uhci_cardbus.c: netbsd-7-nhusb
sys/dev/ic/sl811hs.c: netbsd-7-nhusb
sys/dev/ic/sl811hsvar.h: netbsd-7-nhusb
sys/dev/isa/slhci_isa.c: netbsd-7-nhusb
sys/dev/marvell/ehci_mv.c: netbsd-7-nhusb
sys/dev/pci/ehci_pci.c: netbsd-7-nhusb
sys/dev/pci/ohci_pci.c: netbsd-7-nhusb
sys/dev/pci/uhci_pci.c: netbsd-7-nhusb
sys/dev/pci/xhci_pci.c: netbsd-7-nhusb
sys/dev/pcmcia/slhci_pcmcia.c: netbsd-7-nhusb
sys/dev/usb/Makefile.usbdevs: netbsd-7-nhusb
sys/dev/usb/TODO: netbsd-7-nhusb
sys/dev/usb/TODO.usbmp: netbsd-7-nhusb
sys/dev/usb/aubtfwl.c: netbsd-7-nhusb
sys/dev/usb/auvitek.c: netbsd-7-nhusb
sys/dev/usb/auvitek_audio.c: netbsd-7-nhusb
sys/dev/usb/auvitek_dtv.c: netbsd-7-nhusb
sys/dev/usb/auvitek_i2c.c: netbsd-7-nhusb
sys/dev/usb/auvitek_video.c: netbsd-7-nhusb
sys/dev/usb/auvitekvar.h: netbsd-7-nhusb
sys/dev/usb/ehci.c: netbsd-7-nhusb
sys/dev/usb/ehcireg.h: netbsd-7-nhusb
sys/dev/usb/ehcivar.h: netbsd-7-nhusb
sys/dev/usb/emdtv.c: netbsd-7-nhusb
sys/dev/usb/emdtv_dtv.c: netbsd-7-nhusb
sys/dev/usb/emdtv_ir.c: netbsd-7-nhusb
sys/dev/usb/emdtvvar.h: netbsd-7-nhusb
sys/dev/usb/ezload.c: netbsd-7-nhusb
sys/dev/usb/ezload.h: netbsd-7-nhusb
sys/dev/usb/files.usb: netbsd-7-nhusb
sys/dev/usb/hid.c: netbsd-7-nhusb
sys/dev/usb/hid.h: netbsd-7-nhusb
sys/dev/usb/if_athn_usb.c: netbsd-7-nhusb
sys/dev/usb/if_athn_usb.h: netbsd-7-nhusb
sys/dev/usb/if_atu.c: netbsd-7-nhusb
sys/dev/usb/if_atureg.h: netbsd-7-nhusb
sys/dev/usb/if_aue.c: netbsd-7-nhusb
sys/dev/usb/if_auereg.h: netbsd-7-nhusb
sys/dev/usb/if_axe.c: netbsd-7-nhusb
sys/dev/usb/if_axen.c: netbsd-7-nhusb
sys/dev/usb/if_axenreg.h: netbsd-7-nhusb
sys/dev/usb/if_axereg.h: netbsd-7-nhusb
sys/dev/usb/if_cdce.c: netbsd-7-nhusb
sys/dev/usb/if_cdcereg.h: netbsd-7-nhusb
sys/dev/usb/if_cue.c: netbsd-7-nhusb
sys/dev/usb/if_cuereg.h: netbsd-7-nhusb
sys/dev/usb/if_kue.c: netbsd-7-nhusb
sys/dev/usb/if_kuereg.h: netbsd-7-nhusb
sys/dev/usb/if_otus.c: netbsd-7-nhusb
sys/dev/usb/if_otusvar.h: netbsd-7-nhusb
sys/dev/usb/if_rum.c: netbsd-7-nhusb
sys/dev/usb/if_rumreg.h: netbsd-7-nhusb
sys/dev/usb/if_rumvar.h: netbsd-7-nhusb
sys/dev/usb/if_run.c: netbsd-7-nhusb
sys/dev/usb/if_runvar.h: netbsd-7-nhusb
sys/dev/usb/if_smsc.c: netbsd-7-nhusb
sys/dev/usb/if_smscreg.h: netbsd-7-nhusb
sys/dev/usb/if_smscvar.h: netbsd-7-nhusb
sys/dev/usb/if_udav.c: netbsd-7-nhusb
sys/dev/usb/if_udavreg.h: netbsd-7-nhusb
sys/dev/usb/if_upgt.c: netbsd-7-nhusb
sys/dev/usb/if_upgtvar.h: netbsd-7-nhusb
sys/dev/usb/if_upl.c: netbsd-7-nhusb
sys/dev/usb/if_ural.c: netbsd-7-nhusb
sys/dev/usb/if_uralreg.h: netbsd-7-nhusb
sys/dev/usb/if_uralvar.h: netbsd-7-nhusb
sys/dev/usb/if_url.c: netbsd-7-nhusb
sys/dev/usb/if_urlreg.h: netbsd-7-nhusb
sys/dev/usb/if_urndis.c: netbsd-7-nhusb
sys/dev/usb/if_urndisreg.h: netbsd-7-nhusb
sys/dev/usb/if_urtw.c: netbsd-7-nhusb
sys/dev/usb/if_urtwn.c: netbsd-7-nhusb
sys/dev/usb/if_urtwn_data.h: netbsd-7-nhusb
sys/dev/usb/if_urtwnreg.h: netbsd-7-nhusb
sys/dev/usb/if_urtwnvar.h: netbsd-7-nhusb
sys/dev/usb/if_urtwreg.h: netbsd-7-nhusb
sys/dev/usb/if_zyd.c: netbsd-7-nhusb
sys/dev/usb/if_zydreg.h: netbsd-7-nhusb
sys/dev/usb/irmce.c: netbsd-7-nhusb
sys/dev/usb/moscom.c: netbsd-7-nhusb
sys/dev/usb/motg.c: netbsd-7-nhusb
sys/dev/usb/motgvar.h: netbsd-7-nhusb
sys/dev/usb/ohci.c: netbsd-7-nhusb
sys/dev/usb/ohcireg.h: netbsd-7-nhusb
sys/dev/usb/ohcivar.h: netbsd-7-nhusb
sys/dev/usb/pseye.c: netbsd-7-nhusb
sys/dev/usb/slurm.c: netbsd-7-nhusb
sys/dev/usb/stuirda.c: netbsd-7-nhusb
sys/dev/usb/u3g.c: netbsd-7-nhusb
sys/dev/usb/uark.c: netbsd-7-nhusb
sys/dev/usb/uatp.c: netbsd-7-nhusb
sys/dev/usb/uaudio.c: netbsd-7-nhusb
sys/dev/usb/uberry.c: netbsd-7-nhusb
sys/dev/usb/ubsa.c: netbsd-7-nhusb
sys/dev/usb/ubsa_common.c: netbsd-7-nhusb
sys/dev/usb/ubsavar.h: netbsd-7-nhusb
sys/dev/usb/ubt.c: netbsd-7-nhusb
sys/dev/usb/uchcom.c: netbsd-7-nhusb
sys/dev/usb/ucom.c: netbsd-7-nhusb
sys/dev/usb/ucomvar.h: netbsd-7-nhusb
sys/dev/usb/ucycom.c: netbsd-7-nhusb
sys/dev/usb/udl.c: netbsd-7-nhusb
sys/dev/usb/udl.h: netbsd-7-nhusb
sys/dev/usb/udsbr.c: netbsd-7-nhusb
sys/dev/usb/udsir.c: netbsd-7-nhusb
sys/dev/usb/uep.c: netbsd-7-nhusb
sys/dev/usb/uftdi.c: netbsd-7-nhusb
sys/dev/usb/uftdireg.h: netbsd-7-nhusb
sys/dev/usb/ugen.c: netbsd-7-nhusb
sys/dev/usb/ugensa.c: netbsd-7-nhusb
sys/dev/usb/uhci.c: netbsd-7-nhusb
sys/dev/usb/uhcireg.h: netbsd-7-nhusb
sys/dev/usb/uhcivar.h: netbsd-7-nhusb
sys/dev/usb/uhid.c: netbsd-7-nhusb
sys/dev/usb/uhidev.c: netbsd-7-nhusb
sys/dev/usb/uhidev.h: netbsd-7-nhusb
sys/dev/usb/uhmodem.c: netbsd-7-nhusb
sys/dev/usb/uhso.c: netbsd-7-nhusb
sys/dev/usb/uhub.c: netbsd-7-nhusb
sys/dev/usb/uipad.c: netbsd-7-nhusb
sys/dev/usb/uipaq.c: netbsd-7-nhusb
sys/dev/usb/uirda.c: netbsd-7-nhusb
sys/dev/usb/uirdavar.h: netbsd-7-nhusb
sys/dev/usb/ukbd.c: netbsd-7-nhusb
sys/dev/usb/ukbdmap.c: netbsd-7-nhusb
sys/dev/usb/ukyopon.c: netbsd-7-nhusb
sys/dev/usb/ukyopon.h: netbsd-7-nhusb
sys/dev/usb/ulpt.c: netbsd-7-nhusb
sys/dev/usb/umass.c: netbsd-7-nhusb
sys/dev/usb/umass_isdata.c: netbsd-7-nhusb
sys/dev/usb/umass_isdata.h: netbsd-7-nhusb
sys/dev/usb/umass_quirks.c: netbsd-7-nhusb
sys/dev/usb/umass_quirks.h: netbsd-7-nhusb
sys/dev/usb/umass_scsipi.c: netbsd-7-nhusb
sys/dev/usb/umass_scsipi.h: netbsd-7-nhusb
sys/dev/usb/umassvar.h: netbsd-7-nhusb
sys/dev/usb/umcs.c: netbsd-7-nhusb
sys/dev/usb/umct.c: netbsd-7-nhusb
sys/dev/usb/umidi.c: netbsd-7-nhusb
sys/dev/usb/umidi_quirks.c: netbsd-7-nhusb
sys/dev/usb/umidi_quirks.h: netbsd-7-nhusb
sys/dev/usb/umodem.c: netbsd-7-nhusb
sys/dev/usb/umodem_common.c: netbsd-7-nhusb
sys/dev/usb/umodemvar.h: netbsd-7-nhusb
sys/dev/usb/ums.c: netbsd-7-nhusb
sys/dev/usb/uplcom.c: netbsd-7-nhusb
sys/dev/usb/urio.c: netbsd-7-nhusb
sys/dev/usb/urio.h: netbsd-7-nhusb
sys/dev/usb/usb.c: netbsd-7-nhusb
sys/dev/usb/usb.h: netbsd-7-nhusb
sys/dev/usb/usb_mem.c: netbsd-7-nhusb
sys/dev/usb/usb_mem.h: netbsd-7-nhusb
sys/dev/usb/usb_quirks.c: netbsd-7-nhusb
sys/dev/usb/usb_quirks.h: netbsd-7-nhusb
sys/dev/usb/usb_subr.c: netbsd-7-nhusb
sys/dev/usb/usbdevices.config: netbsd-7-nhusb
sys/dev/usb/usbdevs: netbsd-7-nhusb
sys/dev/usb/usbdevs.h: netbsd-7-nhusb
sys/dev/usb/usbdevs_data.h: netbsd-7-nhusb
sys/dev/usb/usbdi.c: netbsd-7-nhusb
sys/dev/usb/usbdi.h: netbsd-7-nhusb
sys/dev/usb/usbdi_util.c: netbsd-7-nhusb
sys/dev/usb/usbdi_util.h: netbsd-7-nhusb
sys/dev/usb/usbdivar.h: netbsd-7-nhusb
sys/dev/usb/usbhid.h: netbsd-7-nhusb
sys/dev/usb/usbhist.h: netbsd-7-nhusb
sys/dev/usb/usbroothub.c: netbsd-7-nhusb
sys/dev/usb/usbroothub.h: netbsd-7-nhusb
sys/dev/usb/usbroothub_subr.c: delete
sys/dev/usb/usbroothub_subr.h: delete
sys/dev/usb/uscanner.c: netbsd-7-nhusb
sys/dev/usb/uslsa.c: netbsd-7-nhusb
sys/dev/usb/usscanner.c: netbsd-7-nhusb
sys/dev/usb/ustir.c: netbsd-7-nhusb
sys/dev/usb/uthum.c: netbsd-7-nhusb
sys/dev/usb/utoppy.c: netbsd-7-nhusb
sys/dev/usb/uts.c: netbsd-7-nhusb
sys/dev/usb/uvideo.c: netbsd-7-nhusb
sys/dev/usb/uvisor.c: netbsd-7-nhusb
sys/dev/usb/uvscom.c: netbsd-7-nhusb
sys/dev/usb/uyap.c: netbsd-7-nhusb
sys/dev/usb/uyap_firmware.h: netbsd-7-nhusb
sys/dev/usb/uyurex.c: netbsd-7-nhusb
sys/dev/usb/x1input_rdesc.h: netbsd-7-nhusb
sys/dev/usb/xhci.c: netbsd-7-nhusb
sys/dev/usb/xhcireg.h: netbsd-7-nhusb
sys/dev/usb/xhcivar.h: netbsd-7-nhusb
sys/dev/usb/xinput_rdesc.h: netbsd-7-nhusb
sys/external/bsd/common/conf/files.linux: netbsd-7-nhusb
sys/external/bsd/common/include/linux/err.h: netbsd-7-nhusb
sys/external/bsd/common/include/linux/kernel.h: netbsd-7-nhusb
sys/external/bsd/common/include/linux/workqueue.h: netbsd-7-nhusb
sys/external/bsd/common/linux/linux_work.c: netbsd-7-nhusb
sys/external/bsd/drm2/dist/drm/radeon/atombios_encoders.c: netbsd-7-nhusb
sys/external/bsd/drm2/dist/drm/radeon/radeon_legacy_encoders.c: netbsd-7-nhusb
sys/external/bsd/drm2/drm/files.drmkms: netbsd-7-nhusb
sys/external/bsd/drm2/i915drm/files.i915drmkms: netbsd-7-nhusb
sys/external/bsd/drm2/include/linux/err.h: delete
sys/external/bsd/drm2/include/linux/workqueue.h: delete
sys/external/bsd/drm2/linux/files.drmkms_linux: netbsd-7-nhusb
sys/external/bsd/drm2/linux/linux_work.c: delete
sys/external/bsd/dwc2/dwc2.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dwc2.h: netbsd-7-nhusb
sys/external/bsd/dwc2/dwc2var.h: netbsd-7-nhusb
sys/external/bsd/dwc2/dwctwo2netbsd: netbsd-7-nhusb
sys/external/bsd/dwc2/conf/files.dwc2: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_core.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_core.h: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_coreintr.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_hcd.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_hcd.h: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_hcdddma.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_hcdintr.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_hcdqueue.c: netbsd-7-nhusb
sys/external/bsd/dwc2/dist/dwc2_hw.h: netbsd-7-nhusb
sys/modules/drmkms_linux/Makefile: netbsd-7-nhusb
sys/modules/i915drmkms/Makefile: netbsd-7-nhusb
sys/rump/dev/lib/libugenhc/ugenhc.c: netbsd-7-nhusb
sys/rump/dev/lib/libusb/Makefile: netbsd-7-nhusb
sys/rump/dev/lib/libusb/USB.ioconf: netbsd-7-nhusb
sys/rump/dev/lib/libusb/usb_at_ugenhc.c: delete
sys/rump/dev/lib/libusb/opt/opt_usb.h: delete
sys/rump/dev/lib/libusb/opt/opt_usbverbose.h: delete
sys/sys/mbuf.h: netbsd-7-nhusb
usr.sbin/usbdevs/usbdevs.8: netbsd-7-nhusb
usr.sbin/usbdevs/usbdevs.c: netbsd-7-nhusb
Merge netbsd-7-nhusb:
- API / infrastructure changes to support memory management changes.
- Memory management improvements and bug fixes.
- HCDs should now be MP safe
- conversion to KERNHIST based debug
- FS/LS isoc support on ehci(4).
- conversion to kmem(9)
- Some USB 3 support - mostly from Takahiro HAYASHI (t-hash).
- interrupt transfers now get proper DMA operations
- general bug fixes
- kern/48308
- uhub status notification improvements
- umass(4) probe fix (applied to HEAD already)
- ohci(4) short transfer fix
- Change the SOFTINT level from NET to SERIAL for the USB softint handler.
This gives the callback a chance of running when another softint handler
at SOFTINT_NET has blocked holding a lock, e.g. softnet_lock and most of
the network stack.
- kern/49065 - ifconfig tun0 ... sequence locks up system / lockup:
softnet_lock held across usb xfr
- kern/50491 - unkillable wait in usbd_transfer while using usmsc0
on raspberry pi 2
- kern/51395 - USB Ethernet makes xhci hang
- Various improvements to slhci(4)
- Various improvements to dwc2(4)
 1.23.2.4  06-Feb-2016  snj branches: 1.23.2.4.2;
Pull up following revision(s) (requested by skrll in ticket #1097):
sys/dev/usb/usb.c: revision 1.161
sys/dev/usb/usb_subr.c: revisions 1.207, 1.208
sys/dev/usb/usbdivar.h: revision 1.111
sys/dev/usb/xhci.c: revision 1.33
Get the iManufacturer, iProduct, and iSerialNumber strings before probing
for drivers and cache them for later use. This reduces bus transactions
and fixes attachment for at least two of my umass(4)s.
--
Need sys/kmem.h
 1.23.2.3  02-Jan-2015  martin Pull up following revision(s) (requested by skrll in ticket #372):
sys/dev/usb/xhcireg.h: revision 1.2
sys/dev/usb/xhci.c: revision 1.28
kern/49391: Fixes to XHCI driver command ring and status TRB
CRCR needs 64byte aligned address
Use usb_allocmem instead of usb_allocmem_flags(..., 0)
Correct status stage TRB in xhci_device_ctrl_start - direction logic was
wrong.
Correct a typo in a comment
 1.23.2.2  13-Aug-2014  riz Pull up following revision(s) (requested by skrll in ticket #11):
sys/dev/usb/xhci.c: revision 1.25
Serialise xhci_intr1 calls with sc_intr_lock. From Takahiro HAYASHI.
 1.23.2.1  11-Aug-2014  martin Pull up following revision(s) (requested by skrll in ticket #3):
sys/dev/usb/xhci.c: revision 1.24
PR/49091: xhci: wrong wMaxPacketSize value
While this is correct according to the specification only fixed sizes
are allowed, i.e. 512 for SS, etc. Maybe these should be used?
 1.23.2.4.2.3  26-Jan-2017  skrll Sync with HEAD/nhusb
 1.23.2.4.2.2  07-Sep-2016  skrll Sync with HEAD
 1.23.2.4.2.1  06-Sep-2016  skrll First pass at netbsd-7 updated with USB code from HEAD
 1.26.2.3  03-Dec-2017  jdolecek update from HEAD
 1.26.2.2  20-Aug-2014  tls Rebase to HEAD as of a few days ago.
 1.26.2.1  12-Aug-2014  tls file xhci.c was added on branch tls-maxphys on 2014-08-20 00:03:51 +0000
 1.28.2.86  28-Aug-2017  skrll Sync with HEAD
 1.28.2.85  05-Feb-2017  skrll Sync with HEAD
 1.28.2.84  03-Jan-2017  skrll Improve handling of roothub device and free up a bus address for LS/FS/HS
controllers.
 1.28.2.83  02-Jan-2017  skrll Parse the extended capabilies to and log each controller port to SS/HS
bus root hub ports.

Create/attach the two buses and adapt the xhci_roothub_ctrl to deal with
both buses and sets of roothub ports.

XXX the roothub ub_devices entry needs work to interact with usbdevs(1)
XXX correctly
 1.28.2.82  02-Jan-2017  skrll Use %s for __func__... oops
 1.28.2.81  02-Jan-2017  skrll Use __func__ in panic message
 1.28.2.80  02-Jan-2017  skrll Actually print the bus root hub device in debug output
 1.28.2.79  02-Jan-2017  skrll Remove unnecessary brackets
 1.28.2.78  29-Dec-2016  skrll Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
 1.28.2.77  05-Oct-2016  skrll Sync with HEAD
 1.28.2.76  09-Jul-2016  skrll Sync with HEAD
 1.28.2.75  12-Jun-2016  skrll Remove KASSERTs that don't make sense.
 1.28.2.74  10-Jun-2016  skrll Add callout_stop while clearing stall. From t-hash.
 1.28.2.73  10-Jun-2016  skrll Don't forget to destroy sc_cmdbusy_cv. From t-hash.
 1.28.2.72  10-Jun-2016  skrll Only call sc_vendor_init once. Spotted by t-hash.
 1.28.2.71  05-Jun-2016  skrll Sync with HEAD
 1.28.2.70  30-May-2016  skrll Serialise commands
 1.28.2.69  29-May-2016  skrll Sync with HEAD
 1.28.2.68  30-Apr-2016  skrll Move the struct usb_task to struct usbd_xfer for everyone to use.
 1.28.2.67  16-Apr-2016  skrll Remove an unnecessary ux_hcpriv assignement
 1.28.2.66  11-Apr-2016  skrll Don't use USB_IS_SS when there's a switch (speed) to use.
 1.28.2.65  10-Apr-2016  skrll Simplify xhci_do_command*
 1.28.2.64  10-Apr-2016  skrll route is a uint32_t in xhci_init_slot so pass it as such
 1.28.2.63  10-Apr-2016  skrll Free DMA buffer after disabling slot.
Fix bug DCBA was not cleared after disabling slot.

From Takahiro HAYASHI
 1.28.2.62  10-Apr-2016  skrll Fix set_port_feature U[12]_TIMEOUT to work when speed > 4.

From Takahiro HAYASHI
 1.28.2.61  10-Apr-2016  skrll Fix a DPRINTF where the route string and roothub port were swapped.

From Takahiro HAYASHI
 1.28.2.60  10-Apr-2016  skrll Remove variable v0 becasue ~XHCI_PS_CLEAR does not mask speed bits.

From Takahiro HAYASHI
 1.28.2.59  10-Apr-2016  skrll Fix assertion of upper limit of slot index.

From Takahiro HAYASHI
 1.28.2.58  10-Apr-2016  skrll Fix lock assertion in xhci_softintr() when entering DDB on ukbd.
Move this KASSERT to xhci_softintr() from xhci_handle_event().

From Takahiro HAYASHI
 1.28.2.57  10-Apr-2016  skrll Unwrap lines and other whitespace.

From Takahiro HAYASHI
 1.28.2.56  10-Apr-2016  skrll Fix merge botch
 1.28.2.55  08-Apr-2016  skrll Don't rely on XHCI_IMAN_INTR_PEND. Instead do as others do and simply
clear the bit and handle any pending events.

usb 3 now works on my toshiba laptop (and probably elsewhere)
 1.28.2.54  19-Mar-2016  skrll Sync with HEAD
 1.28.2.53  28-Feb-2016  skrll Centralise the up_repeat handling and use the standard pipe method to
start the next transfer. This allows the removal of a bunch of code
in the upm_done methods for interrupt transfers which had copies of
the upm_start method code.

At the same time we can perform the upm_done method before calling the
transfer callback allowing correct bus_dma(9) operations before
using the transfer DMA buffer.
 1.28.2.52  13-Feb-2016  skrll Improve transfer callout handling
 1.28.2.51  06-Feb-2016  skrll Remove unintended commit
 1.28.2.50  06-Feb-2016  skrll Use if (!expr) { ... } instead of a if (expr) goto label; ... ; label:
 1.28.2.49  10-Jan-2016  skrll Bring the following change from HEAD

Get the iManufacturer, iProduct, and iSerialNumber strings before probing
for drivers and cache them for later use. This reduces bus transactions
and fixes attachment for at least two of my umass(4)s.
 1.28.2.48  23-Dec-2015  skrll Revert previous... I got ahead of myself
 1.28.2.47  23-Dec-2015  skrll Transfer are only ever allocated from thread context now. Change IPL_USB
to IPL_NONE in pool_cache_init for the xfer struct pool.
 1.28.2.46  24-Oct-2015  skrll xhci_xspeed2speed is unused for now
 1.28.2.45  24-Oct-2015  skrll Do the following slightly differently to the diff from t-hash

+ Fix the bug that port feature UHF_PORT_{U1,U2}_TIMEOUT
were never set.
PS_SPEED bits were always cleared.
+ Fix the bug of logic inversion whether port speed is SS
when setting feature UHF_PORT_{U1,U2}_TIMEOUT.
 1.28.2.44  21-Oct-2015  skrll Comments
 1.28.2.43  21-Oct-2015  skrll Whitespace.
 1.28.2.42  20-Oct-2015  skrll Consistently providei/use *_{XFER,PIPE,BUS}2SC, etc macros
 1.28.2.41  11-Oct-2015  skrll Update ubm_allocx with the isoc frame count parameter and use it in
dwctwo(4)
 1.28.2.40  03-Oct-2015  skrll Report as USB 3.0
 1.28.2.39  29-Sep-2015  skrll sizeof KNF
 1.28.2.38  23-Sep-2015  skrll Grammar
 1.28.2.37  13-Sep-2015  skrll Various improvements from t-hash (updated by me)

+ Improve xhci_configure_endpoint().
+ Split off maxburst and interval calculation.
+ Start interval calculation with 10, not 11.
+ Put correct maxburst value.
+ Split off constructing endpoint context.
+ Improve xhci_event_transfer()
+ Remove case of unlikely completion codes I added.
+ Clear xr_cookies too when xhci_set_dequeue() clears xr_trb.
+ Return USBD_NO_ADDR if xhci_address_device fails with
XHCI_TRB_ERROR_NO_SLOTS.
+ Add aprint_debug xhci capability registers.
+ Add & update comments.

I checked HCCPARAMS1 against version 1.1 of the XHCI specification. Not
sure of previous version formats
 1.28.2.36  13-Sep-2015  skrll Split xhci_handle_event() into 3 functions. Whitespace.

From t-hash.
 1.28.2.35  13-Sep-2015  skrll Fix locking botch. From t-hash.
 1.28.2.34  31-Aug-2015  skrll Various improvements from t-hash (very very slightly tweaked by me)

+ Use usbd_xfer_isread().
+ Change mutex to be initialized at IPL_USB.
+ Add vendor init/portsc hook.
+ Modify xhci_trb_put() to take host byte order arguments and
convert them to little endian byte order.
+ Return PCI vendor ID of xhci instead of NetBSD(0x0) as a root hub
vendor ID like other HCs do.
+ Move sc_ih in struct xhci_softc to struct xhci_pci_softc.
+ Improve debug message.
 1.28.2.33  26-Jun-2015  skrll Make xhci_close_pipe void - nothing cares about any errors.

From t-hash.
 1.28.2.32  26-Jun-2015  skrll Add port range check in xhci_rhpsc().

KASSERT xfer->ux_pipe != NULL in xhci_handle_event().

Based on a diff from t-hash
 1.28.2.31  26-Jun-2015  skrll Revert previous (for now)
 1.28.2.30  25-Jun-2015  skrll Improve bInterval handling in xhci_configure_endpoint - should help
keyboards.

From t-hash
 1.28.2.29  23-Jun-2015  skrll Comment updates from t-hash.
 1.28.2.28  07-Jun-2015  skrll Update comment.

From t-hash.
 1.28.2.27  28-May-2015  skrll More changes from Takahiro HAYASHI

+ Add sc_statuspend that stores ports bitmap of pending interrupts
instead of sc_isxhciroothub hack til someone implements suspend pipe.
While sc_explorepending == 1, uhub_intr shall merge sc_status into
sc_statuspend.
+ Eliminate confusing UPS_SUPER_SPEED flag and introduce
UPS_OTHER_SPEED flag that indicates ud_speed of device is
super speed (or more).
uhub shall set this flag if ud_speed is super speed (or more).
+ Add the macro that checks ud_speed is super speed.
The codes shall use this macro to check ud_speed is super speed.
+ Add speed type conversion functions.
+ Include port link status in port_status if port is super speed.

Various other changes to support SS hubs and devices
 1.28.2.26  27-May-2015  skrll Add patch for intel chips quirk from ryoon@ in
https://mail-index.netbsd.org/netbsd-bugs/2014/08/31/msg038109.html
slightly modified to set quirks before calling xhci_init.

Set special IMOD value and route ports to xhci for intel chips.

From Takahiro HAYASHI.
 1.28.2.25  27-May-2015  skrll Rewrite if-else-if chains with switch-case.

From Takahiro HAYASHI.
 1.28.2.24  27-May-2015  skrll Add some sanity checks for xfer. [Ed: should really find the bugs]

From Takahiro HAYASHI.
 1.28.2.23  27-May-2015  skrll Fix slot leak when address_device fails.

From Takahiro HAYASHI.
 1.28.2.22  27-May-2015  skrll Don't abuse pipe->up_async_task for xhci_clear_endpoint_stall_async_task.
Add member struct usb_task xp_async_task to struct xhci_pipe.

From Takahiro HAYASHI.
 1.28.2.21  27-May-2015  skrll Add comments. Some cosmetic changes.

From Takahiro HAYASHI.
 1.28.2.20  07-Apr-2015  skrll Comments. From Takahiro HAYASHI.
 1.28.2.19  07-Apr-2015  skrll USB 3.0 snapshot from Takahiro HAYASHI
 1.28.2.18  06-Apr-2015  skrll Cosmetic/Whitespace changes from Takahiro HAYASHI
 1.28.2.17  06-Apr-2015  skrll *** empty log message ***
 1.28.2.16  22-Mar-2015  skrll Indentation and whitespace.
 1.28.2.15  22-Mar-2015  skrll Fix spellos
 1.28.2.14  19-Mar-2015  skrll Do the same as OpenBSD and get rid of the *_handle typedefs and use
plain structures insteads
 1.28.2.13  05-Dec-2014  skrll KNF. Remove ( ) from return statements.
 1.28.2.12  04-Dec-2014  skrll Rework roothub control transfers so that much of the code is shared
across HCDs.

I have retained the vendor/product reporting for each HCD for now,
but it maybe get removed later.

ahci(4) now reports a language table and uses the usb_makestrdesc
function instead of rolling its own version.
 1.28.2.11  03-Dec-2014  skrll Rename usbroothub_subr.[ch] to usbroothub.[ch]
 1.28.2.10  03-Dec-2014  skrll Provide a USETWD macro for use with USB words designated
initialisers.
 1.28.2.9  03-Dec-2014  skrll Remove #include <sys/malloc.h> where it's not (no longer) needed
 1.28.2.8  03-Dec-2014  skrll Replace malloc(9) with kmem(9)
 1.28.2.7  03-Dec-2014  skrll Use designated initializers for more descriptors.
 1.28.2.6  03-Dec-2014  skrll Use designated initialisers for usb_device_descriptor_t structs.
 1.28.2.5  03-Dec-2014  skrll The grand renaming of structure members.

No functional change.
 1.28.2.4  02-Dec-2014  skrll Step #1 of memory allocation re-organisation.

Centralised the buffer allocation routine which now supports DMA
and non-DMA capable host controllers. Remove the
ubm_{alloc,free}m methods from usbd_bus_methods.

The buffer allocation is only allowed in thread context and,
therefore, negates the usefulness of the reserve dma code which
is removed in this change.

USBD_NO_COPY is also no longer required as usbd_transfer and
usbd_transfer_complete now track buffer usage and handle any
copying.
 1.28.2.3  01-Dec-2014  skrll Add prefixes to method structures member names. No functional change.
 1.28.2.2  01-Dec-2014  skrll Remove usbd_bus no_intrs member it was (virtually) unused.
 1.28.2.1  30-Nov-2014  skrll Use C99 types. u_int{8,16,32,64}_t to uint{8,16,32,64}_t.

No functional change.
 1.62.2.3  26-Apr-2017  pgoyette Sync with HEAD
 1.62.2.2  20-Mar-2017  pgoyette Sync with HEAD
 1.62.2.1  26-Jul-2016  pgoyette Sync with HEAD
 1.67.2.1  21-Apr-2017  bouyer Sync with HEAD
 1.72.2.15  01-Aug-2023  martin Pull up following revision(s) (requested by riastradh in ticket #1869):

sys/dev/usb/xhci.c: revision 1.176
sys/dev/usb/xhcivar.h: revision 1.23

xhci(4): Defer root intr xfers while polling.

Root intr xfers require taking adaptive locks, which is forbidden
while polling.

This is not great -- any USB transfer completion callbacks might try
to take adaptive locks, not just uhub_intr, and that will always
causes trouble. We get lucky with ukbd_intr because it's not
MP-safe, so it relies only on the kernel lock (a spin lock) anyway.
But this change brings xhci in line with ehci.

PR kern/57326
 1.72.2.14  23-Jan-2023  martin Pull up the following revisions, requested by msaitoh in ticket #1787:

sys/dev/pci/xhci_pci.c 1.31 via patch
sys/dev/usb/xhci.c 1.173-1.175
sys/dev/usb/xhcivar.h 1.22

Support xHCI device which has USB 2 port only.
 1.72.2.13  16-Sep-2022  martin Pull up following revision(s) (requested by msaitoh in ticket #1765):

sys/dev/usb/xhci.c: revision 1.154

Accept USB 3.2 in xhci_id_protocols().
 1.72.2.12  16-Nov-2019  martin Pull up the following revisions, requested by msaitoh in ticket #1443:

sys/arch/arm/nvidia/tegra_xusb.c 1.13-1.14 via patch
sys/dev/pci/xhci_pci.c 1.13
sys/dev/usb/usb.c 1.169
sys/dev/usb/usbdivar.h 1.116
sys/dev/usb/xhci.c 1.93, 1.95, 1.97
sys/dev/usb/xhcireg.h 1.11-1.12

Detect USB 3.1
 1.72.2.11  17-Sep-2019  martin Pull up following revision(s) (requested by mrg in ticket #1377):

sys/dev/usb/xhci.c: revision 1.113
sys/dev/usb/xhci.c: revision 1.109

match xhci version 3.10. allows properly finding all the USB
busses on new ryzen 3 based systems.

unfortunately, the USB busses are still non-functional.

-

fix ryzen usb issue: we set TD size to '1', where has xhci spec 4.11.2.4
says final TRB for a TD should have this set to '0'. since we currently
only generate sinel TRB TDs, set this to 0.

XXX: pullup-all
from sc.dying
 1.72.2.10  04-Jan-2019  martin Pull up following revision(s) (requested by mrg in ticket #1155):

sys/dev/usb/xhci.c: revision 1.100

use xhci_polling_p(). this might miss when bus2 is active.
XXX: pullup-7, pullup-8.
 1.72.2.9  28-Sep-2018  martin Fixup for ticket #1037 - parts of the patch were accidently missing.
 1.72.2.8  27-Sep-2018  martin Pull up following revision(s) (requested by mrg in ticket #1037):

sys/dev/usb/uhub.c: revision 1.139
sys/external/bsd/dwc2/dwc2.c: revision 1.55
sys/ddb/db_output.c: revision 1.34
sys/ddb/db_command.c: revision 1.160
sys/dev/usb/ehci.c: revision 1.264
sys/dev/usb/xhci.c: revision 1.99
sys/dev/usb/ehci.c: revision 1.265
sys/kern/subr_userconf.c: revision 1.27
sys/dev/usb/ehcivar.h: revision 1.46
sys/dev/usb/ohci.c: revision 1.287
sys/dev/usb/uhci.c: revision 1.284
sys/dev/usb/usbdi.c: revision 1.178
sys/dev/usb/usb.c: revision 1.172
sys/dev/pci/xhci_pci.c: revision 1.14
sys/dev/usb/usb.c: revision 1.173
sys/dev/usb/usb.c: revision 1.174
share/man/man4/usb.4: revision 1.110
sys/ddb/db_command.c: revision 1.159
sys/dev/usb/usb_subr.c: revision 1.227
sys/dev/usb/uhcivar.h: revision 1.56
(all via patch)

consolidate the handling of polling across HC drivers, and generic USB:
- don't take mutexes if polling
- normalise the code across all drivers
- add some not yet code to block discovery to/from polling
- minor CSE
- adjust comment for usbd_set_polling() to reality now i properly
understand what it is used for and why.

this, with a hack to make RB_ASKNAME to wait 5 seconds allows boot -a
work with USB keyboards. there are still multiple issues remaining:
- discovery and polling need to be mutually exclusive
- attachment of ukbd and wskbd is not handled by config_pending, and
the 5 second delay isn't going to always be enough.

call cnpollc(1) and cnpollc(0) around cngetc().
(christos has a good idea to add a function that does all 3,
and we should switch all the callers in this sequence to use
it (and fix the MD ones missing it still). not all can, as
eg, line-grabbing functions can use cngetsn(), which only
calls cnpollc() twice.)


When this file is used when not building the kernel (eg: /usr/sbin/crash)
make cnpollc() go away.


reorder some struct members to remove holes.


add config_pending usage to uhub and general USB device attachment.
- call config_pending_incr() and config_pending_decr() around attaching
devices against "usbdevif" attribute.

uhub:
- convert sc_explorepending and sc_running to bool. add new sc_first_explore.
- call config_pending_incr() at the start of uhub_attach(). dropped in
uhub_explore(), if this is the first explore.


implement a gross hack to fix "boot -a" on systems with usb keyboards on
systems with ehci handover to uhci (and maybe ohci), and fix a similar
problem for "boot -s".

there is effort to ensure that all devices attached via USB are probed
before RB_ASKNAME or RB_SINGLE attempts to ask any questions on the console,
and largely this works, often by chance, today, for USB disks and root.
i've recently pushed this more into uhub and general USB device attachment
as well, and kept a config_pending reference across the first explore of
a bus. these fix many issues with directly attached hubs.

however, on systems where devices connected to ehci ports are handed over
to a companion uhci or ohci port, it may not be the first, or even second,
bus explore that finds the device finally before attachment, and at this
point all config_pending references are dropped.

there is no direct communication between drivers, the potentials are
looked up but their device_t is only used for generic things like the name,
so informing the correct companion to expect a device and deal with the
config_pending references is not possible without some fairly ugly layer
violations or multi-level callbacks (eg, we have "ehci0", and usually an
the relevant companion, eg, "uhci2", but it is the uhub that uhci2 has
attached that will deal with the device attachment.)

with the above fixes to generic USB code, the disown happens during the
first explore. the hack works by, at this point, checking if (a) root
is not mounted, (b) single user or ask name are set, and (c) if the hack
as not been triggered already. if all 3 conditions are true, then a
config_pending_incr() is called and a callback is triggered for (default)
5 seconds to call config_pending_decr(). ehci detach pauses waiting for
this callback if scheduled.

this allows enough time for the uhub and the ukbd/wskbd to attach before
the RK_ASKROOT prompts appear. testing shows it takes between 1.5 and
2 seconds for the keyboard to appear after the disown occurs.

Index: dev/usb/ehcivar.c
- new sc_compcallout, sc_compcallout, sc_complock, and a state for th
handover hack.

Index: dev/usb/ehci.c
ehci_init():
- use aprint_normal_dev() instead of manual device_xname().
- initialise sc_compcallout, sc_compcallout, sc_complock, and sc_comp_state.
ehci_detach():
- if there are companion controllers, tear own the above, including waiting
if there is a callback scheduled.
ehci_disown_callback():
- new callout to call config_pending_decr() in the the future.
schedule this ca
ehci_disown_sched_callback():
- if booting to single user or asking names, call config_pending_incr() and
schedule the callout above, default 5 second delay.
ehci_disown():
- if disowning a port call ehci_disown_sched_callback().
deal with partial attach failures in usb_attach vs usb_detach aka PR 53598.
- make sure xhci's sc->sc_ios is NULL if failure happens.
- rearrange usb_attach() / usb_doattach() to make it simpler to clean up.
- move usb_async_intr softint into usb_once_init(). previously, each USB
controller would start a new one, and leave the old one leaked.
- handle controller interrupts without a bus attached


remove usb(4)'s "flags 1" code. it has been dead for a while,
as it runs during the interrupts part of configuration now,
and all the devices try attach as early as possible, including
any root or boot required disk or keyboard device, which is
what this flag was for.
 1.72.2.7  25-Aug-2018  martin Pull up following revision(s) (requested by mrg in ticket #980):

sys/dev/usb/usbdivar.h: revision 1.117
sys/external/bsd/dwc2/dwc2.c: revision 1.52
sys/dev/usb/xhcivar.h: revision 1.10
sys/dev/usb/motg.c: revision 1.22
sys/dev/usb/ehci.c: revision 1.260
sys/dev/usb/ehci.c: revision 1.261
sys/dev/usb/xhci.c: revision 1.96
sys/dev/usb/ohci.c: revision 1.282
sys/dev/usb/ohci.c: revision 1.283
sys/dev/usb/ehcivar.h: revision 1.45
sys/dev/usb/uhci.c: revision 1.281
sys/dev/usb/uhci.c: revision 1.282
sys/dev/usb/usbdi.c: revision 1.177
sys/dev/usb/ohcivar.h: revision 1.60
sys/dev/usb/uhcivar.h: revision 1.55
(all via patch)

pull across abort fixes from nick-nhusb. add more abort fixes, using
ideas from Taylor and Nick, and myself. special thanks to both who
inspired much of the code here, if not wrote it directly.
among other problems, this assert should no longer trigger:

panic: kernel diagnostic assertion "xfer->ux_state == XFER_ONQU" failed: file "/current/src/sys/dev/usb/usbdi.c", line 914

using usbhist i was able to track down my instance of it being related
to userland close() beginning, dropping the sc_lock, and then the usb
softintr completes the transfer normally, and when it is done, the
abort path attempts to re-complete the transfer, and the above assert
is tripped.

changes from nhusb were commited with these logs:
--
Move the struct usb_task to struct usbd_xfer for everyone to use.
--
Set device transfer status to USBD_IN_PROGRESS if start methods succeeds
--
Actually set the transfer status on transfers in ohci_abort_xfer and
the controller is dying
--
Don't supply the lock to callout_halt when polling as it won't be held
--
Improve transfer abort
--
Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
--
--
Mark device transfers as USBD_IN_PROGRESS appropriately and improve
abort handling
--

additional changes include:
- initialise the usb abort task in the HCI allocx routine, so that it
can be safely usb_rem_task()'d.
- rework the handling of softintr vs cancellation vs timeout abort based
upon a scheme from Taylor:
when completing a transfer normally:
- if the status is not in progress, it must be cancelled or timed out,
and we should not process this xfer.
- set the status as normal.
- unconditionallly callout_stop() and usb_rem_task(). they're safe and
either aren't running, or will run and do nothing.
- finally call usb_transfer_complete().
when aborting a transfer:
- status should be cancelled or timed out.
- if cancelling, callout_halt and usb_rem_task_wait() to make sure the
timer is either done or cancelled.
- at this point, the ux_status must not be cancelled or timed out, and
if it is not in progress we're done.
- set the status.
- if the controller is dying, just return.
- perform HCI-specific tasks to abort this xfer.
- finally call usb_transfer_complete().
for the timeout and timeout task:
- if the HCI is not dying, and the ux_status is in progress, then
trigger the usb abort task.
- remove UXFER_ABORTWAIT and UXFER_ABORTING.

tested on:
- multiple PC systems with several types of devices: ugen/UPS, ucom,
umass with disk, ssd and cdrom backends, kbd, ms, using uhci, ehci
and xhci.
- erlite3: sd@umass on dwc2.
- sunblade2000: kbd/ms and umass disk on ohci.

untested:
- motg, slhci and ahci. motg has some portion of the new scheme
applied, but slhci and ahci require more study.

future work includes pushing a lot of the common abort handling into
usbdi.c and leaving upm_abort() for HC specific tasks, but this change
is pullup-able to netbsd-7 and netbsd-8 as it does not change any
external API, as well as removing over 100 lines of code while adding
over 30 new asserts.

XXX: pullup-7, pullup-8.

fix DIAGNOSTIC build by not copying ub_usepolling to stack before use

Sprinkle __diagused
 1.72.2.6  05-May-2018  martin Pull up following revision(s) (requested by jdolecek in ticket #787):

sys/dev/usb/xhci.c: revision 1.88-1.90
sys/dev/usb/xhcireg.h: revision 1.10

add KASSERT() that sc_child* is set to NULL after child detach; just for
readability, it's not immediatelly obvious this is done in xhci_childdet()
no functional changes

trigger the softint processing on that child bus which is not detached yet
fixes PR kern/53066 by Martin Husemann

enable code to only trigger usb processing when EINT is set, to
avoid misinterpreting shared interrupt for another device

when clearing USBSTS, actually preserve the bits which spec requires to
preserve, and actually clear bit 1, which should be actually always
cleared to zero by spec

also #ifdef XHCI_DEBUG some unnecessary register reads
this should finally resolve PR kern/53066 also for Martin
 1.72.2.5  21-Dec-2017  snj Pull up following revision(s) (requested by skrll in ticket #459):
sys/dev/usb/xhci.c: 1.78, 1.83
PR/52791: Robert Sprowson: avoid duplicate tests (use accessor variables)
--
Fix the logic around TT Hub Slot ID, TT Port Number and Multi-TT in
xhci_setup_tthub. TT Hub/Port need to reference the upstream HS hub and
Multi-TT needs to reflect the capabilities of a HS hub itself or the same
upstream HS hub.
 1.72.2.4  23-Nov-2017  martin Pull up following revision(s) (requested by skrll in ticket #385):
sys/dev/usb/xhci.c: revision 1.77
sys/external/bsd/dwc2/dwc2.c: revision 1.47
sys/dev/usb/motg.c: revision 1.19
sys/dev/usb/ehci.c: revision 1.257
sys/dev/usb/ohci.c: revision 1.276
sys/dev/usb/uhci.c: revision 1.279

s/PR_NOWAIT/PR_WAITOK/ in HCD allocx (allocate xfer) method
 1.72.2.3  17-Nov-2017  snj Pull up following revision(s) (requested by msaitoh in ticket #356):
sys/dev/usb/xhci.c: revision 1.76
Wait 1ms first. Existing Intel xHCI requies 1ms delay to prevent system hang
(Errata).
 1.72.2.2  02-Nov-2017  snj Pull up following revision(s) (requested by pgoyette in ticket #335):
share/man/man9/kernhist.9: 1.5-1.8
sys/arch/acorn26/acorn26/pmap.c: 1.39
sys/arch/arm/arm32/fault.c: 1.105 via patch
sys/arch/arm/arm32/pmap.c: 1.350, 1.359
sys/arch/arm/broadcom/bcm2835_bsc.c: 1.7
sys/arch/arm/omap/if_cpsw.c: 1.20
sys/arch/arm/omap/tiotg.c: 1.7
sys/arch/evbarm/conf/RPI2_INSTALL: 1.3
sys/dev/ic/sl811hs.c: 1.98
sys/dev/usb/ehci.c: 1.256
sys/dev/usb/if_axe.c: 1.83
sys/dev/usb/motg.c: 1.18
sys/dev/usb/ohci.c: 1.274
sys/dev/usb/ucom.c: 1.119
sys/dev/usb/uhci.c: 1.277
sys/dev/usb/uhub.c: 1.137
sys/dev/usb/umass.c: 1.160-1.162
sys/dev/usb/umass_quirks.c: 1.100
sys/dev/usb/umass_scsipi.c: 1.55
sys/dev/usb/usb.c: 1.168
sys/dev/usb/usb_mem.c: 1.70
sys/dev/usb/usb_subr.c: 1.221
sys/dev/usb/usbdi.c: 1.175
sys/dev/usb/usbdi_util.c: 1.67-1.70
sys/dev/usb/usbroothub.c: 1.3
sys/dev/usb/xhci.c: 1.75
sys/external/bsd/drm2/dist/drm/i915/i915_gem.c: 1.34
sys/kern/kern_history.c: 1.15
sys/kern/kern_xxx.c: 1.74
sys/kern/vfs_bio.c: 1.275-1.276
sys/miscfs/genfs/genfs_io.c: 1.71
sys/sys/kernhist.h: 1.21
sys/ufs/ffs/ffs_balloc.c: 1.63
sys/ufs/lfs/lfs_vfsops.c: 1.361
sys/ufs/lfs/ulfs_inode.c: 1.21
sys/ufs/lfs/ulfs_vnops.c: 1.52
sys/ufs/ufs/ufs_inode.c: 1.102
sys/ufs/ufs/ufs_vnops.c: 1.239
sys/uvm/pmap/pmap.c: 1.37-1.39
sys/uvm/pmap/pmap_tlb.c: 1.22
sys/uvm/uvm_amap.c: 1.108
sys/uvm/uvm_anon.c: 1.64
sys/uvm/uvm_aobj.c: 1.126
sys/uvm/uvm_bio.c: 1.91
sys/uvm/uvm_device.c: 1.66
sys/uvm/uvm_fault.c: 1.201
sys/uvm/uvm_km.c: 1.144
sys/uvm/uvm_loan.c: 1.85
sys/uvm/uvm_map.c: 1.353
sys/uvm/uvm_page.c: 1.194
sys/uvm/uvm_pager.c: 1.111
sys/uvm/uvm_pdaemon.c: 1.109
sys/uvm/uvm_swap.c: 1.175
sys/uvm/uvm_vnode.c: 1.103
usr.bin/vmstat/vmstat.c: 1.219
Reorder to test for null before null deref in debug code
--
Reorder to test for null before null deref in debug code
--
KNF
--
No need for '\n' in UVMHIST_LOG
--
normalise a BIOHIST log message
--
Update the kernhist(9) kernel history code to address issues identified
in PR kern/52639, as well as some general cleaning-up...
(As proposed on tech-kern@ with additional changes and enhancements.)
Details of changes:
* All history arguments are now stored as uintmax_t values[1], both in
the kernel and in the structures used for exporting the history data
to userland via sysctl(9). This avoids problems on some architectures
where passing a 64-bit (or larger) value to printf(3) can cause it to
process the value as multiple arguments. (This can be particularly
problematic when printf()'s format string is not a literal, since in
that case the compiler cannot know how large each argument should be.)
* Update the data structures used for exporting kernel history data to
include a version number as well as the length of history arguments.
* All [2] existing users of kernhist(9) have had their format strings
updated. Each format specifier now includes an explicit length
modifier 'j' to refer to numeric values of the size of uintmax_t.
* All [2] existing users of kernhist(9) have had their format strings
updated to replace uses of "%p" with "%#jx", and the pointer
arguments are now cast to (uintptr_t) before being subsequently cast
to (uintmax_t). This is needed to avoid compiler warnings about
casting "pointer to integer of a different size."
* All [2] existing users of kernhist(9) have had instances of "%s" or
"%c" format strings replaced with numeric formats; several instances
of mis-match between format string and argument list have been fixed.
* vmstat(1) has been modified to handle the new size of arguments in the
history data as exported by sysctl(9).
* vmstat(1) now provides a warning message if the history requested with
the -u option does not exist (previously, this condition was silently
ignored, with only a single blank line being printed).
* vmstat(1) now checks the version and argument length included in the
data exported via sysctl(9) and exits if they do not match the values
with which vmstat was built.
* The kernhist(9) man-page has been updated to note the additional
requirements imposed on the format strings, along with several other
minor changes and enhancements.
[1] It would have been possible to use an explicit length (for example,
uint64_t) for the history arguments. But that would require another
"rototill" of all the users in the future when we add support for an
architecture that supports a larger size. Also, the printf(3)
format
specifiers for explicitly-sized values, such as "%"PRIu64, are much
more verbose (and less aesthetically appealing, IMHO) than simply
using "%ju".
[2] I've tried very hard to find "all [the] existing users of
kernhist(9)"
but it is possible that I've missed some of them. I would be glad
to
update any stragglers that anyone identifies.
--
For some reason this single kernel seems to have outgrown its declared
size as a result of the kernhist(9) changes. Bump the size.
XXX The amount of increase may be excessive - anyone with more detailed
XXX knowledge please feel free to further adjust the value
appropriately.
--
Misssed one cast of pointer --> uintptr_t in previous kernhist(9) commit
--
And yet another one. :(
--
Use correct mark-up for NetBSD version.
--
More improvements in grammar and readability.
--
Remove a stray '"' (obvious typo) and add a couple of casts that are
probably needed.
--
And replace an instance of "%p" conversion with "%#jx"
--
Whitespace fix. Give Bl tag table a width. Fix Xr.
 1.72.2.1  25-Aug-2017  snj Pull up following revision(s) (requested by skrll in ticket #226):
sys/dev/usb/xhci.c: revision 1.73
Check both buses for ub_usepolling and schedule the soft interrupt handler
or call it directly. Fixes usb keyboard in ddb when attached to xhci.
 1.86.2.9  18-Jan-2019  pgoyette Synch with HEAD
 1.86.2.8  26-Nov-2018  pgoyette Sync with HEAD, resolve a couple of conflicts
 1.86.2.7  30-Sep-2018  pgoyette Ssync with HEAD
 1.86.2.6  06-Sep-2018  pgoyette Sync with HEAD

Resolve a couple of conflicts (result of the uimin/uimax changes)
 1.86.2.5  28-Jul-2018  pgoyette Sync with HEAD
 1.86.2.4  21-May-2018  pgoyette Sync with HEAD
 1.86.2.3  02-May-2018  pgoyette Synch with HEAD
 1.86.2.2  22-Apr-2018  pgoyette Sync with HEAD
 1.86.2.1  16-Apr-2018  pgoyette Sync with HEAD, resolve some conflicts
 1.92.2.2  13-Apr-2020  martin Mostly merge changes from HEAD upto 20200411
 1.92.2.1  10-Jun-2019  christos Sync with HEAD
 1.107.2.12  01-Aug-2023  martin Pull up following revision(s) (requested by riastradh in ticket #1687):

sys/dev/usb/xhci.c: revision 1.176
sys/dev/usb/xhcivar.h: revision 1.23

xhci(4): Defer root intr xfers while polling.

Root intr xfers require taking adaptive locks, which is forbidden
while polling.

This is not great -- any USB transfer completion callbacks might try
to take adaptive locks, not just uhub_intr, and that will always
causes trouble. We get lucky with ukbd_intr because it's not
MP-safe, so it relies only on the kernel lock (a spin lock) anyway.
But this change brings xhci in line with ehci.

PR kern/57326
 1.107.2.11  23-Jan-2023  martin Pull up the following revisions, requested by msaitoh in ticket #1570:

sys/dev/pci/xhci_pci.c 1.31 via patch
sys/dev/usb/xhci.c 1.173-1.175
sys/dev/usb/xhcivar.h 1.22

Support xHCI device which has USB 2 port only.
 1.107.2.10  16-Sep-2022  martin Pull up following revision(s) (requested by msaitoh in ticket #1525):

sys/dev/usb/xhci.c: revision 1.154

Accept USB 3.2 in xhci_id_protocols().
 1.107.2.9  22-Jun-2021  martin Fix pullup #1302 by adapting the changes to this branch
(there was a misunderstanding in the ticket handling)
 1.107.2.8  21-Jun-2021  martin Pull up following revision(s) (requested by riastradh in ticket #1301):

sys/dev/usb/xhci.c: revision 1.140
sys/dev/usb/xhci.c: revision 1.141
sys/dev/usb/xhci.c: revision 1.143
sys/dev/usb/xhcivar.h: revision 1.18
sys/dev/usb/xhcivar.h: revision 1.19
sys/dev/usb/xhcireg.h: revision 1.19
sys/dev/usb/xhci.c: revision 1.139

xhci(4): Draft suspend/resume.

Work almost entirely done and tested by maya@ based on xhci 1.2 spec;
tidied up and tweaked by me.

Not sure about issuing Stop Endpoint commands or ensuring the Command
Ring is in the Stopped or Idle state, but this seems to work as is,
so it's already an improvement over what we had before which was no
xhci suspend/resume at all.

In particular, it's not clear to us:
- if we don't have any pending USB activity whether we need to issue
the Stop Endpoints or quiesce the command ring; but
- if we do have any pending USB activity whether issuing Stop
Endpoint is enough or whether we also need to do anything to
synchronize with other software logic to quiesce it too.

xhci(4): Block commands and issue Stop Endpoint on suspend.

xhci: Fix logic in waiting for command queue access.
_Either_ an existing command in progress, _or_ an existing suspend in
progress that is not done by us, should block us; the logic I wrote
previously erroneously blocked only if both conditions happened at
the same time.

Should fix issue reported by Andrius V in the PR kern/56050 followup
discussion.

xhci(4): Wait USB_RESUME_WAIT ms, not 20 ms.
Better to use the named constant, and although the spec says 20 ms is
enough, apparently for some devices it's not.
 1.107.2.7  23-Dec-2020  martin Pull up following revision(s) (requested by riastradh in ticket #1159):

sys/dev/usb/ohci.c: revision 1.314
sys/external/bsd/dwc2/dwc2.c: revision 1.75
sys/dev/usb/ehci.c: revision 1.284
sys/dev/usb/uhci.c: revision 1.305
sys/dev/usb/xhci.c: revision 1.136

usb: Omit bogus assertions about struct usbd_pipe::up_intrxfer.

These assertions were only valid for pipes at UE_IN_DIR, UE_INTERRUPT
endpoints created with usbd_open_pipe_intr, which uses up_intrxfer to
pass the struct usbd_xfer object to usbd_close_pipe to free later.

In contrast, for pipes at UE_OUT_DIR, UE_INTERRUPT endpoints,
up_intrxfer is never initialized, so the assertion cannot be right.
In principle we might even have more than one outstanding interrupt
transfer at a time, rendering the point of the assertion moot anyway.

Found by interrupting a uhidev write to a u2f device.

ok nick
 1.107.2.6  25-May-2020  martin Pull up following revision(s) (requested by jakllsch in ticket #921):

sys/dev/usb/xhci.c: revision 1.129

xhci(4): only usbd_xfer_trycomplete() if certain to usb_transfer_complete()

Previously we'd cancel the timeout after the data stage of control xfers, and
then hope that the status stage would eventually interrupt.
 1.107.2.5  01-Mar-2020  martin Pull up following revision(s) (requested by riastradh in ticket #744):

sys/dev/usb/uhci.c: revision 1.292
sys/dev/usb/uhci.c: revision 1.293
sys/dev/usb/usbdi.h: revision 1.99
sys/dev/usb/motg.c: revision 1.26
sys/dev/usb/motg.c: revision 1.27
sys/dev/usb/motg.c: revision 1.28
sys/dev/usb/motg.c: revision 1.29
sys/external/bsd/dwc2/dwc2.c: revision 1.70
sys/external/bsd/dwc2/dwc2.c: revision 1.71
sys/dev/usb/usb.c: revision 1.181
sys/arch/mips/adm5120/dev/ahci.c: revision 1.20
sys/dev/usb/usb.c: revision 1.182
sys/dev/usb/xhci.c: revision 1.116
sys/dev/usb/xhci.c: revision 1.117
sys/dev/usb/xhci.c: revision 1.118
sys/dev/usb/uhci.c: revision 1.289
sys/dev/usb/usbdivar.h: revision 1.121
sys/dev/usb/usbdi.c: revision 1.190
sys/dev/usb/usbdivar.h: revision 1.122
sys/dev/usb/usbdi.c: revision 1.191
sys/dev/usb/usbdi.c: revision 1.192
sys/external/bsd/dwc2/dwc2var.h: revision 1.7
sys/dev/usb/motg.c: revision 1.30
sys/dev/usb/motg.c: revision 1.31
sys/dev/usb/motg.c: revision 1.32
sys/dev/usb/motg.c: revision 1.33
sys/external/bsd/dwc2/dwc2.c: revision 1.67
sys/external/bsd/dwc2/dwc2.c: revision 1.68
sys/dev/usb/ehci.c: revision 1.270
sys/external/bsd/dwc2/dwc2.c: revision 1.69
sys/dev/usb/usbdi.h: revision 1.100
sys/dev/usb/ehci.c: revision 1.271
sys/arch/mips/adm5120/dev/ahci.c: revision 1.18
sys/dev/usb/usbdi.h: revision 1.101
sys/dev/usb/ehci.c: revision 1.272
sys/dev/ic/sl811hs.c: revision 1.103
sys/arch/mips/adm5120/dev/ahci.c: revision 1.19
sys/dev/usb/ehci.c: revision 1.273
sys/dev/usb/ohci.c: revision 1.293
sys/dev/usb/uhci.c: revision 1.290
sys/dev/usb/ohci.c: revision 1.294
sys/dev/usb/uhci.c: revision 1.291
sys/dev/usb/ohci.c: revision 1.295

Teach usb_rem_task to return whether removed from queue or not.

New function usb_task_pending for diagnostic assertions.
Usable only for negative diagnostic assertions:

KASSERT(!usb_task_pending(dev, task))

If you can think of a better name for this than !usb_task_pending,
I'm all ears.

-

Nothing guarantees xfer's timeout has completed.

Wait for it when we free the xfer.

-


New xfer state variables ux_timeout_set and ux_timeout_reset.

These are needed because:
- The host controller interrupt cannot wait for the callout or task
to finish running.
- Nothing in the USBD API as is waits for the callout or task to
finish running.
- Callers expect to be able to resubmit USB xfers from xfer callbacks
without waiting for anything to finish running.

The variable ux_timeout_set can be used by a host controller to
decide on submission whether to schedule the callout or to ask an
already-scheduled callout or already-queued task to reschedule the
callout, by setting the variable ux_timeout_reset to true.

When the callout or task runs and sees that ux_timeout_reset is true,
rather than queue the task or abort the xfer, it can instead just
schedule the callout anew.

-

Fix steady state of timeouts in ehci.

This is complicated because:
1. There are three ways that an xfer can be completed:
(a) hardware interrupt completes xfer
(b) software decision aborts xfer with USBD_CANCELLED
(c) timeout aborts xfer with USBD_TIMEOUT
2. The timeout abort can't be done in callout because ehci_sync_hc,
called unconditionally by ehci_abort_xfer to wait until the device
has finished using any references to the xfer, may sleep. So we
have to schedule a callout that, when run, will schedule a usb_task.
3. The hardware completion interrupt can't sleep waiting for a callout
or task to finish -- can't use callout_halt or usb_rem_task_wait.
So the callout and usb_task must be able to run _after_ the hardware
completion interrupt, and recognize that they're late to the party.
(Note, though, that usbd_free_xfer does wait for the callout and
task to complete, so there's no danger they may use themselves after
free.)
4. The xfer may resubmitted -- and the timeout may be rescheduled --
immediately after the hardware completion interrupt, _while_ the
callout and/or usb_task may still be scheduled. Specifically, we
may have the following sequence of events:
(a) hardware completion interrupt
(b) callout or usb_task fires
(c) driver resubmits xfer
(d) callout or usb_task acquires lock and looks around dazed and
bewildered at the firehose of events like reading the news in 2019

The mechanism for sorting this out is that we have two bits of state:
- xfer->ux_timeout_set informs the driver, when submitting an xfer and
setting up its timeout, whether either the callout or usb_task is
already scheduled or not.
- xfer->ux_timeout_reset informs the callout or usb_task whether it
should reschedule the callout, because the xfer got resubmitted, or
not.

-

Factor out HCI-independent xfer completion logic.

New API for HCI drivers to synchronize hardware completion
interrupts, synchronous aborts, and asynchronous timeouts:
- When submitting an xfer to hardware, call
usbd_xfer_schedule_timeout(xfer).
- On HCI completion interrupt for xfer completion:
if (!usbd_xfer_trycomplete(xfer))
return; /* timed out or aborted, ignore it */
- In upm_abort methods, call usbd_xfer_abort(xfer).

For HCI drivers that use this API (not needed in drivers that don't,
or for xfers like root intr xfers that don't use it):
- New ubm_abortx method serves role of former *hci_abort_xfer, but
without any logic for wrangling timeouts/callouts/tasks -- caller
in usbd_xfer_abort has already handled them.
- New ubm_dying method, returns true if the device is in the process
of detaching, used by the timeout logic.

Converted and tested:
- ehci
- ohci

Converted and compile-tested:
- ahci (XXX did this ever work?)
- dwc2
- motg (XXX missing usbd_xfer_schedule_timeout in motg_*_start?)
- uhci
- xhci

Not changed:
- slhci (sys/dev/ic/sl811hs.c) -- doesn't use a separate per-xfer
callout for timeouts (XXX but maybe should?)
- ugenhc (sys/rump/dev/lib/libugenhc/ugenhc.c) -- doesn't manage its
own transfer timeouts

-

Fix steady state of root intr xfers.

Why?
- Avoid completing a root intr xfer multiple times in races.
- Avoid potential use-after-free in poll_hub callouts (uhci, ahci).

How?
- Use sc->sc_intr_xfer or equivalent to store only a pending xfer
that has not yet completed -- whether successfully, by timeout, or
by synchronous abort. When any of those happens, set it to null
under the lock, so the xfer is completed only once.
- For hci drivers that use a callout to poll the root hub (uhci, ahci):
. Pass the softc pointer, not the xfer, to the callout, so the
callout is not even tempted to use xfer after free -- if the
callout fires, but the xfer is synchronously aborted before the
callout can do anything, the xfer might be freed by the time the
callout starts to examine it.
. Teach the callout to do nothing if it is callout_pending after it
has fired. This way:
1. completion or synchronous abort can just callout_stop
2. start can just callout_schedule
If the callout had already fired before (1), and doesn't acquire
the bus lock until after (2), it may be tempted to abort the new
root intr xfer just after submission, which would be wrong -- so
instead we just have the callout do nothing if it notices it has
been rescheduled, since it will fire again after the appropriate
time has elapsed.

-

Initialize xfer->ux_status in uhci_root_intr_start.

Otherwise, it will be USBD_NOT_STARTED, so usbd_ar_pipe will skip
calling upm_abort.
Candidate fix for PR kern/54963, same problem as reported at:
href="https://mail-index.NetBSD.org/current-users/2020/02/13/msg037740.html

-

Set ux_isdone in uhci_poll_hub for DIAGNOSTIC.

-

Fix mistakes in previous sloppy change with root intr xfers.
- Make sure ux_status is set to USBD_IN_PROGRESS when started.
Otherwise, if it is still in flight when we abort the pipe,
usbd_ar_pipe will skip calling upm_abort.
- Initialize ux_status under the lock; in principle a completion
interrupt (or a delay) could race with the initialization.
- KASSERT that the xfer is in progress when we're about to complete
it.

Candidate fix for PR kern/54963 for other HCI drivers than uhci.
ok nick
ok phone
(This is the change that nick evidently MEANT to ok when he ok'd the
previous one!)

-

Fix build

-

Fix non-DIAGNOSTIC builds.

-

Fix wrong KASSERT in motg abort.
This has been wrong since last summer when we did the transition to
xfer->ux_status = USBD_CANCELLED earlier.
XXX pullup-9

-

Fix mistakes in timeout/abort/completion changes in motg(4).
- Call usbd_xfer_schedule_timeout so we actually do time out.
- Don't call usbd_xfer_trycomplete until all the data have been
transferred -- it commits to completion, not timeout.
- Use xfer->ux_status != USBD_IN_PROGRESS to test whether, after a
partial write, an xfer has been interrupted or timed out and need
not be continued.
- Remove wrong assertion.

-

Fix mistake in use of usbd_xfer_schedule_timeout in motg.

This code path is used both for xfers that are new, and xfers that
are being done piece by piece and are partway done. For the latter
case, skip usbd_xfer_schedule_timeout so we schedule it only once per
xfer.

-

Simplify some branches and kassert some redundant assignments.
 1.107.2.4  21-Jan-2020  martin Pull up following revision(s) (requested by taca in ticket #625):

sys/dev/usb/xhci.c: revision 1.115

Acquire bus lock (if not polling) when changing xfer status, resetting
the timeout and ringing the doorbell in the start methods.
 1.107.2.3  13-Sep-2019  martin Pull up following revision(s) (requested by mrg in ticket #196):

sys/dev/usb/xhci.c: revision 1.111
sys/dev/usb/xhci.c: revision 1.112
sys/dev/usb/xhci.c: revision 1.113
sys/dev/usb/xhci.c: revision 1.114

introduce and use XHCIHIST_CALLARGS(). reduces lots of double logs,
and includes useful info in all cases instead of just "called".
add a couple of more logs that i've wanted while debugging ryzen3
vs USB issues.

fix 32-bit debug build, and also vmstat -y. reported by sc.dying.

fix ryzen usb issue: we set TD size to '1', where has xhci spec 4.11.2.4
says final TRB for a TD should have this set to '0'. since we currently
only generate sinel TRB TDs, set this to 0.
from sc.dying

avoid non-debug kernel build issues.
 1.107.2.2  01-Sep-2019  martin Pull up following revision(s) (requested by mrg in ticket #135):

distrib/sets/lists/comp/mi 1.2279
distrib/sets/lists/modules/mi 1.123
share/man/man9/Makefile 1.438
share/man/man9/usbnet.9 1.1-1.9
sys/dev/ic/rndisreg.h 1.3
sys/dev/usb/TODO 1.47-1.52
sys/dev/usb/TODO.usbmp 1.15,1.16
sys/dev/usb/files.usb 1.157-1.167
sys/dev/usb/if_aue.c 1.155-1.161
sys/dev/usb/if_auereg.h 1.30-1.32
sys/dev/usb/if_axe.c 1.103-1.119
sys/dev/usb/if_axen.c 1.51-1.53,1.55-1.67
sys/dev/usb/if_axenreg.h 1.15
sys/dev/usb/if_cdce.c 1.54-1.67
sys/dev/usb/if_cue.c 1.85,1.86
sys/dev/usb/if_cuereg.h 1.23
sys/dev/usb/if_kue.c 1.97-1.100
sys/dev/usb/if_kuereg.h 1.23,1.24
sys/dev/usb/if_mue.c 1.51-1.55
sys/dev/usb/if_muereg.h 1.6
sys/dev/usb/if_muevar.h 1.9
sys/dev/usb/if_smsc.c 1.46-1.61
sys/dev/usb/if_smscreg.h 1.6
sys/dev/usb/if_smscvar.h delete
sys/dev/usb/if_udav.c 1.60-1.71
sys/dev/usb/if_udavreg.h 1.14,1.15
sys/dev/usb/if_upl.c 1.65,1.66
sys/dev/usb/if_ure.c 1.15-1.31
sys/dev/usb/if_urevar.h 1.4,1.5
sys/dev/usb/if_url.c 1.67-1.70
sys/dev/usb/if_urlreg.h 1.14
sys/dev/usb/if_urndis.c 1.22-1.33
sys/dev/usb/if_urtwn.c 1.72
sys/dev/usb/ohci.c 1.290
sys/dev/usb/uhub.c 1.143
sys/dev/usb/usb.c 1.180
sys/dev/usb/usb.h 1.118
sys/dev/usb/usb_mem.c 1.71
sys/dev/usb/usb_subr.c 1.238,1.239
sys/dev/usb/usbdevs 1.772
sys/dev/usb/usbdi.c 1.183,1.186
sys/dev/usb/usbdi.h 1.97
sys/dev/usb/usbdi_util.c 1.75
sys/dev/usb/usbhist.h 1.5,1.6
sys/dev/usb/usbnet.c 1.1-1.24
sys/dev/usb/usbnet.h 1.1-1.14
sys/dev/usb/usbroothub.c 1.9
sys/dev/usb/xhci.c 1.109,1.110
sys/modules/Makefile 1.223
sys/modules/usbnet/Makefile 1.1

usbnet(9): Add common framework for USB network devices.
This bring various safety fixes to all updated drivers,
and includes locking clean up, detach safety when being
used or not, separate rx/tx locks to improve performance,
porting to NET_MPSAFE, many edge/error case bugs in
drivers fixed, as well as resovling PRs 54303 and 54308.
These drivers are converted: axe(4), axen(4), aue(4),
cdce(4), cue(4), kue(4), mue(4), smsc(4), udav(4),
upl(4), ure(4), url(4), and urndis(4).
 1.107.2.1  21-Aug-2019  martin Pull up following revision(s) (requested by skrll in ticket #104):

sys/dev/usb/xhci.c: revision 1.108

typo in error message
 1.115.2.1  29-Feb-2020  ad Sync with head.
 1.135.2.2  03-Apr-2021  thorpej Sync with HEAD.
 1.135.2.1  03-Jan-2021  thorpej Sync w/ HEAD.
 1.138.6.1  31-May-2021  cjep sync with head
 1.138.4.1  17-Jun-2021  thorpej Sync w/ HEAD.
 1.175.2.4  23-Sep-2025  martin Pull up following revision(s) (requested by nat in ticket #1159):

sys/dev/usb/xhci.c: revision 1.190

Handle special case of USBD_FORCE_SHORT_XFER.

This fixes devices such as urtwn(4) connected to an xhci port.

It works by sending a zero length transfer at the end of the data transfer
for the above mentioned flag.

Similar artwork in this space already exists for ohci/uhci/ehci.

OK jkallsch@.
 1.175.2.3  20-Sep-2024  martin Pull up following revision(s) (requested by rin in ticket #885):

sys/dev/usb/xhcireg.h: revision 1.23
sys/dev/usb/xhci.c: revision 1.178
sys/dev/usb/xhci.c: revision 1.179

Remove an unnecessary cast in a KASSERTMSG
- Fix clearing of EINT and other transient flags.

Patches from sc.dying in kern/56115:
- Set proper Max ESIT Payload value for interrupt/isoc endpoint context.
- Set proper Average TRB Length value.
- Not tested on superspeed/superspeedplus isochronous device.
- Add handling of some error paths for isochronous transfers.
 1.175.2.2  01-Aug-2023  martin Pull up following revision(s) (requested by riastradh in ticket #281):

sys/dev/usb/xhci.c: revision 1.177
sys/dev/usb/xhci.c: revision 1.180
sys/dev/usb/xhcivar.h: revision 1.24

xhci(4): Avoid crash in suspend/resume/resume if first resume fails.
Rather than try to recover from this, just make new commands fail so
at least we don't deadlock.

xhci(4): Don't panic on suspend if previous suspend/resume failed.
Trying to resume again probably won't make the situation much worse,
but panicking can definitely make it worse.
 1.175.2.1  01-Aug-2023  martin Pull up following revision(s) (requested by riastradh in ticket #280):

sys/dev/usb/xhci.c: revision 1.176
sys/dev/usb/xhcivar.h: revision 1.23

xhci(4): Defer root intr xfers while polling.

Root intr xfers require taking adaptive locks, which is forbidden
while polling.

This is not great -- any USB transfer completion callbacks might try
to take adaptive locks, not just uhub_intr, and that will always
causes trouble. We get lucky with ukbd_intr because it's not
MP-safe, so it relies only on the kernel lock (a spin lock) anyway.
But this change brings xhci in line with ehci.

PR kern/57326
 1.184.2.1  02-Aug-2025  perseant Sync with HEAD
 1.188.2.1  12-Sep-2025  snj Pull up following revision(s) (requested by nat in ticket #30):

sys/dev/usb/xhci.c: revision 1.190

Handle special case of USBD_FORCE_SHORT_XFER.
This fixes devices such as urtwn(4) connected to an xhci port.
It works by sending a zero length transfer at the end of the data transfer
for the above mentioned flag.
Similar artwork in this space already exists for ohci/uhci/ehci.
OK jkallsch@.

RSS XML Feed