ZX-2468 #comment patch

Tested: Turn on verbosity in trace_manager, then
fx shell trace record trace --help
and verify all devhosts are participating in the trace.
Then boot with driver.tracing.enable=0, rerun, and verify devhosts are absent.

Change-Id: I5cd56133f12331aa604e6a857ffb5fef4eb40c0e

After usb-hub does the port reset, XHCI sends TRB_CMD_RESET_DEVICE.
We also need to address the device again.

I extracted part of xhci_handle_enumerate_device into a new
function xhci_setup_slot, as it is the same thing as xhci_reset_device
has to do after the reset.

This is so we can load new firmware on the USB ML device.

ZX-3155 #comment

TEST= plugged in USB ML device, ran usb-fwloader -d -f firmware and
verified the device descriptor changed.

Also tried resetting in the usb audio driver and running audio -d 1 tone

Change-Id: Id3bd748c0d40838d98e233ad1f01589cb1042021

Adds interface for listening to the test lifecycle.
EventBroadcaster so we there is one entity in charge
of propagating all events.

Moved TestInfo and TestCase out of internal namespace.

TEST=zxtest-test

Change-Id: I4586571349cdfd5e523bde98e9eb6951a25c6dce

TEST= lsusb and verify string descriptors show correctly.

Change-Id: Ibe70381e858a8eba93d3985344e8be17b42cbef3

Test registering, filtering and shuffling.

TEST=boot/test/sys/zxtest-test

Change-Id: Iafe0dc6d542eee3a43fbf5ed0ccb5479978614c4

Test each of the report fields for the eve v2
touchpad.

ZX-3205

TEST: runtests -a -t hidparse-test

Change-Id: I33c49635fa6f7f5e85eb6535f0ca4804e84eafda

All clients have been transitioned over.

Tested: Compiles.
Change-Id: I5ad90421ee6a75ba17a5bebb62b99c5d5e539f1e

This class provides mechanism for instantiating registered tests.
Test only validate minimal correctness.

Tests are added into an array instead of being directly referenced in
main.

TEST=zxtest-tests

Change-Id: I3c7eca40f720c3b8eb52e68037f005ee9e6f2331

Add eve v2 touchpad data and a simple test to
ensure the parser works correctly

ZX-3205

TEST: runtests -a -t hidparse-test

Change-Id: Ic9ad19a3711a9b9ed3623b0193218eeb8d0ae345

SOC specific drivers should bind only against the VID and DID.
PID should not be used unless the driver only works on a specific board.

TEST: boot on eagle

Change-Id: Ic2a8140c09e7e4d8bdd7ed9b419e2612a50c7781

TEST: lsblk shows eMMC partitions on eagle
Change-Id: Idc492b50aa38978e09736eccc55877c4a076e04b

Add thread annotations for the following debug routines...
++ dump_thread
++ dump_all_threads
++ dump_thread_user_tid

Specifically, for each of these routines, make sure there are three
versions of the routine.  The base version (which demands that the
thread lock not be held, and will automatically obtain the lock), a
_locked version (which demands that the lock be held), and a
_during_panic version (which obtains no locks by going directly to the
_locked verion and skipping the compile time analysis)

Also, clean up the users in debug.cpp and x86/faults.cpp to use the
proper versions of each.

Test: Build and k thread (list|list_full) on QEMU
Change-Id: I88e3656e0ec64faf052a9efa32c1a20848456844

TEST= Plug in and remove USB device and checked lsusb.

Change-Id: I05d19b084924e5c2fee8235b66ceea640720b378

flash-cleo, flash-eagle, flash-imx8mevk, flash-imx8mmevk flash-imx8mmevk and flash-mt8167s_ref
now take -c option to add command line to the ZBI.

TEST: flash-cleo -c devmgr.suspend-timeout-debug=true
Change-Id: I9b6dfad3fff1c19eff44bfbecdb277ad91db6922

In order not to inappropriately line break in the "dm dump" output.

TEST: dm dump on cleo
Change-Id: If42757a84726a638b9a49433f93bd39afc6da2eb

This should prevent any regressions on the naming rules.

Test: CQ
Change-Id: I2fb21a131ad84d965c6cb198879ef3a4e0816d8c

Test: compiled
Change-Id: Id5260858bf54b4bd89a28e1120b8fafd0b327b8f

No longer allowed to be returned first. Assuming null terminated, so
output is (char*, size_t)

Tested: Manually checked protocol with string in output.
Change-Id: I1e23ff6e5a7e896a8d4109abadfa642eb579c65a

Tested: Compiles
Change-Id: I891e885d6b63862d39d625e2963d17561954a1cd

Both `dwarf_find_unwind_table` and `dwarf_as_find_unwind_table` have an
offsetting bug that causes them to set di.u.rti.table_data to point into
the EH frame header, rather than to the entries. This results in two
different bugs:

1) Because the table is pointing to the eh_frame_ptr word in the header,
a small value, basically any PC will always appear to be found. However,
no FDE actually exists for that PC and so we fail to unwind PCs that
aren't described by any FDEs instead of failing to find the PC and
falling back to frame-pointer-based unwinding.

2) Because the table is pointing into eh_frame_ptr, but the table_len is
correct, the final FDE cannot be resolved for PCs that are valid, but
are described by the final FDE. Instead, the penultimate FDE will be
chosen, leading to erroneous results.

The calls to `dwarf_read_encoded_pointer` have the side-effect of
advancing the `addr` argument, however this lives in a different address
space. The table data is indeed located at the load_base + section
offset + header length -- only, the dwarf_eh_frame_hdr struct doesn't
include the `eh_frame_ptr` and `fde_count` fields. So we end up with the
table_data pointing to `eh_frame_ptr` instead of the first table entry.

This patch solves the problem by using the last read address (which
should be at the end of the optional `eh_frame_ptr` and `fde_count`)
minus the address of the header as the offset.

Bug: TC-306 #done

Test: manual
Change-Id: Iaf4f0c45b02a2c7f4bacd9863887038401745c8e

... at least not currently.

Tested: CQ
Change-Id: I756bab3c1861bac3ab0799060d69bedb26959b69

Renaming all *Proxy to *Client. Once *Proxy is removed from drivers, we
will remove *Proxy generation code.

Tested: Builds
Change-Id: I923c90365d23d7d28d7edf82230dc6b3800e1750

This reverts commit 95288e7e.

Reason for revert: causing build failure in higher layers, see
e.g. https://logs.chromium.org/logs/fuchsia/buildbucket/cr-buildbucket.appspot.com/8926667935749990112/+/steps/build/0/steps/build_fuchsia/0/steps/ninja/0/stdout

Original change's description:
> [cobalt-fidl] Add new CobaltEvent methods
> 
> This will alow us to support multiple events at once as well as
> multiple event codes per event.
> 
> CB-197
> 
> Test: No functional changes, CQ
> Change-Id: Ifaf2f1ddf19a9f87299deff88692413208daafee

TBR=rudominer@google.com,azani@google.com,zmbush@google.com

Change-Id: I60cad9e595abe3ec4de7ee59fe2600c934a7cdb8
No-Presubmit: true
No-Tree-Checks: true
No-Try: true

Now that the parser is tested we can replace it
with the better HID library parser and it should
work the same

TEST: runtests -a -t hid-test

ZX-3148 #done

Change-Id: Icd3f6a8ed42b2167f28cd1d7319e21de4f25cf95

We accidentally cast to input_report_id_t instead of
input_report_size_t, which means that any values above
255 were being truncated.

TEST: runtests -a -t hid-test

ZX-3148

Change-Id: I0c79c5773e2f08d93221c9460c4c08975602fb8b

GCC doesn't do the desired optimization of the early exit test
in trace_acquire_context_for_category_cached(). This causes a stack
frame to be unnecessarily set up for the common case of tracing off.
Work around this by moving the tail of the function to a separate
no-inline function. GCC then does the desired optimization of
no stack frame for the early-exit test, and the tail call is just a branch.

Benchmark timings on a vim2 (little cores disabled):

TRACE_DURATION_BEGIN macro with 0 arguments (time per call):
 Before:
  Tracing off: 28ns
  Tracing on, category disabled: 28ns
  Tracing on, category enabled: 478ns
 After:
  Tracing off: 26ns
  Tracing on, category disabled: 26ns
  Tracing on, category enabled: 478ns

Note that while the speed improvement is measurable, another reason for
doing this is to remove the noise when trying to visualize how efficient
the tracing=off test is.

PT-84 #comment patch

Tested: CQ
Change-Id: I66753b3ad36a4fe50edb4926f85d571e3b0c9384

All drivers are using the new banjo USB protocol now.
Also removed unused usb_control(), which has been replaced by
usb_control_out() and usb_control_in().

TEST: paved garnet to Pixelbook, USB works.

Change-Id: I4247b53f9607111784c65e7b5db9046c6188f4e8

Tested: paved vim2
Change-Id: I8bd157aad195a61ee48f0b7c2a78e070664ea5d5

Tested: Paved vim2.
Change-Id: I4f32790d465cef96e1a93f87a87df37c1a668c87

This will alow us to support multiple events at once as well as
multiple event codes per event.

CB-197

Test: No functional changes, CQ
Change-Id: Ifaf2f1ddf19a9f87299deff88692413208daafee

Now the fake ddk can return protocols and a name from the parent device.
It also supports testing drivers that use zxlog() calls.

Test: manual

Change-Id: Iafecf51f42285ee47c5028b171c5162647df0b8f

Introduce TimerSlack to encapsulate timer slack.

Extend JobPolicy with TimerSlack

Plumb TimerSlack through from syscall layer to timer subsystem.

No behavior change. A future CL will:
- provide a mechanism for setting the timer slack in job policy
- integrate job policy slack with timers

Bug: ZX-931 #comment Add TimerSlack to JobPolicy
Test: booted, ran "k ut job_policy" and runtests on qemu and vim2
Change-Id: I1f88533563280c93a15d24c25bed936d04e903ad

Use the HID parsing lib to output the same report
sizes that this library also expects.
This is the first step to replacing the current
HID parsing code.

TEST: No test since the code is connected to anything
      yet

ZX-3148

Change-Id: I9e8472c4487cd2828ecae0156603f01a3934c5d8

Remove the intrusive contianer pointer traits defined for
ktl::unique_ptr.  Currently, ktl::unique_ptr is nothing by an alias
for std::unique_ptr, and std::unique_ptr now has official support in
the fbl:: instrusive containers, so the ktl:: definition is no longer
needed.

If ktl::unique_ptr ever becomes its own thing, different from
std::unique_ptr, a set of intrusive container pointer traits will need
to be defined, but that should become pretty obvious at compile time
if/when that day arrives.

Also, remove the un-needed #include for fbl::recycler.  All unique_ptr
users of fbl::recycle have been converted to a custom deleter pattern
instead of using fbl::recycle, so this is no longer needed.

Test: Build and runtests on QEMU
Change-Id: I4c8894175cee561d3fe325e102cf579d2d3c23f7

Test: Builds.
Change-Id: Ic2f83fdb9d18d4a697f71544b2fdd0475eb0b977

This function is a variation of trace_acquire_context_for_category
that adds call-site caching of enabled/disabled state.

The first call to a TRACE_*() macro invokes trace_acquire_context_for_category,
regardless of whether tracing is enabled or not, same as before.
But now the enabled/disabled state of that lookup is cached for each call
site, allowing subsequent calls to just do a load/test. This greatly
speeds up the disabled case, both when tracing is off and when tracing is
on but the category hasn't been selected by the user.

Benchmark timings on a NUC:
TRACE_DURATION_BEGIN macro with 0 arguments:
 Before:
  Tracing off: 8ns per call
  Tracing on, category disabled: 41ns per call
  Tracing on, category enabled: 107ns
 After:
  Tracing off: 5ns per call
  Tracing on, category disabled: 5ns per call
  Tracing on, category enabled: 112ns

PT-2877 #done

Tested: CQ

Change-Id: I9f454031e15c27c05da64dceb60dd472b7b76390

Land a few additional changes which fell out of code review feedback,
but somehow failed to get uploaded before landing.  Specifically...

++ Instead of discussing ways to approach a TODO in the TODO comment
   itself, file a bug instead and have the discussion there.
++ Remove the constexpr = default constructor from the custom deleter
   for MessagePackets.  While it is part of the std::default_delete,
   it is currently serving no valuable purpose in our version as far
   as I can tell, so just remove it instead of cargo culting the std::
   library.

Test: Build and runtests on QEMU
Change-Id: I9c4f0f859184f68a4ac0f6c52929a1c72cfa4806

This allieviates some pain for the folks working on the zircon GN
conversion.  The old copies will be removed once references are removed
from GN.

This is a relanding of fbbabe8e

Test: Compiled
Change-Id: Iad671b901b345923911e4f8364c2033f8d8b5fe6

Tested: Compiles
Change-Id: I16303280ddc30cd486caf5d765689a91443bbb30

Tested: paved vim2.
Change-Id: I1da11d74db85581bf50c565fd09d6f2845561f56