- Moved out existing ISP Init code from MIPI
  and put it under ISP
- Will be adding other initialization stuff in
  subsequent CLs.

Test: Verified all 3 drivers getting loaded on Sherlock
Change-Id: Ibc9143e3916c6af5073bcc919b97a18b33b1850d

This is a reland of 0f31d965, now that
https://fuchsia-review.googlesource.com/c/garnet/+/248677 has landed
and all clients of libfs have correct header dependencies.

Original change's description:
> [vfs] Plumb a hash table around to track koids of tokens
>
> Rather than rely on set/get_cookie
>
> Test: fs unit tests
> Change-Id: I03d6b5de4eae90e8bbcef09958fb983b84e86e64

Test: fs unit tests
Change-Id: If1632111373282bd82be113e5760c255fc3ef0f0

If gfx_create_surface_from_display fails to allocate memory it'll
return nullptr. If this happens we're hosed, but it's better to fail
an assert than dereference a nullptr.

Test: booted Eve w/ gfxconsole.early, then "k gfx rgb_bars"
Change-Id: Ibdcfcf05800ab40a7cd37602877672a5688f4eef

TEST: No behavior change
Change-Id: I2ff1eb2e8daba447b3d6e5f1dda5f4f6ad6c9ff7

Link/unlink eports from their ports under the eport lock and have the
port call a new locked variant of PortMatches so the eport lock is
always acquired first.

Test: CQ
ZX-2517 #done

Change-Id: I96c24ac61052794113513a0b0aec266c5c78a704

Test: make tools -j12 HOST_USE_ASAN=true && ./build-x64/host_tests/fidl-compiler-test
Change-Id: Ic1d807524f6d220026addc0782197e72e6224d1f

In kernel PCI, an upstream node was an interface
for bridges and roots to allow for downstream allocation,
and hooking scanned devices into the topology. In the user
PCI world bus scanning and ownership will be handled by
the top level Bus class, but we still need to have a hierarchy
of allocators for device space downsteam. Unlike kernel PCI,
this version of upstream implements two kinds of allocators:

1. Root allocators which allocate address space over the Pciroot
   protocol. This type is used by Root nodes.
2. Region allocators which allocate address space from their own
   region allocators which were allocated from the node upstream
   of them.

Some method signatures have been included but are not implemented.
These may go away in the future as the Bus driver interface is
finalized. The bulk of the UpstreamNode interface related to downstream
devices will be added back to the class as pci::Device objects are
ported over to userspace to avoid CLs spanning thousands of lines.

ZX-3146

Test: Compiled and ran upci, no regressions detected.

Change-Id: I95dc56cab4ef6a1672c5eee7424a1e1995ba5f36

TEST: Pave garnet on cleo over USB
Change-Id: I506d721abb8dc95f76227e06d06be503c9b1bcb9

TEST: USB audio device binds on NUC.
Change-Id: I0252c2d421f595c8608894a8b813feaabec09bd2

FIDL-451 #progress

Test: this is a new test, run it
Change-Id: Ia76ef0073096b956fb7f24144b032dfc8a139ffb

This allows us to rearrange and simplify some of the code. It also makes
the devmgr startup sequence a little clearer.

ZX-3286

Test: Ran Fuchsia and /system/test/ddk tests.
Change-Id: I6ce509595ccfe1e9a8dfab0cfb78db79b455414b

Userspace should now always be using <atomic>. The kernel will migrate
to a ktl-exposed subset shortly.

Bug: ZX-3357 #comment
Test: no functional change
Change-Id: Idc986ef47a996d2080d15ab680dc0b2255bb5c8d

Move driver-specific code to a driver.h, and rename drivers.cpp to
driver.cpp for consistency. In addition, slightly simplify the
coordinator unit tests to use load_driver.

ZX-3286

Test: Ran Fuchsia and /system/test/ddk tests.
Change-Id: I873539a1b1c0ea6c0626834568ffe31146eee86d

Reduce the visibility of some of the Coordinator member functions so
that we can more aggressively reorganize and refactor it.

Also rename some of the member functions to match the FIDL interface, so
it's clearer that they are used to implement the interface.

ZX-3286

Test: Build Zircon.
Change-Id: I8d61aecf1794162caa920e19f10e3fee4c8c8096

Use a temporary buffer to accomodate any return value from the server
and truncate locally.

Test: fx run-image-test netstack_getsockname_test
Change-Id: I057f155fbbc0f9dffe21cb3931c5487c4eb7b6ba

Committing a pager vmo generates page requests to the backing pager
service. The kernel will attempt to batch requests for adjacent pages,
but the pager API makes no guarantees about exactly what the batching
behavior is.

For a single commit operation, the current implementation generates a
single batched page request for each run of non-resident pages which do
not already have outstanding page requests. For a commit operation
which is broken up into multiple requests, the requests are dispatched
serially, in the sense that the next request will not be dispatched
until the current request is fully supplied.

This implementation uses a missing page count to track how much work is
left for a pending page request. This can result in redundant page
requests if pages are removed from under a pending page request. Given
that it greatly simplifies request tracking and that redundant requests
are not technically a pager API violation, it seems like a worthwhile
trade-off. Furthermore, this should be rare in the current system, and
may become impossible if decommit becomes less forceful (and evict
handles it).

The commit-centric pager tests in core-tests do test the batching
behavior of the kernel, as the correctness of such behavior is
important for the efficiency of the system. However, that does not mean
that the behavior is part of the pager API's contract.

Test: /boot/test/core/pager-test
Test: Pave, boot, use session shell on nuc
Test: vmstress from sandbox/stevensd/userpager for 48+ hours on
nuc/vim
Test: repeatedly run and kill vmstress for 12+ hours on nuc
ZX-1480 #comment [kernel][vm] Implement commit on pager vmos

Change-Id: I65b594a9f3957f2d1e94086852fb3e322847afcd

This removes the control-plane portion of libzxs, moving the
implementation of socket control to a FIDL interface which communicates
over a socket's control plane.

NET-1865 #comment

Test: (in //garnet)
Change-Id: I62ba4c12995d05af8799d45c27b38d15620f7ad3

Added FIDL call to allow querying about minfs block allocations

test: utest/fs/test-minfs
Change-Id: Iaa24435b0ae636496dcd65dcc1c9fe3487d59033

Make the bind_devices unit test send back a BindDriver response.

ZX-3286

Test: Ran modified test.
Change-Id: I8433c447b9d49f3df89a5b026adfa1de9787de07

It cannot be properly implemented lock-free per core. This
can be provided in a lossy way via the usermode kcounter
app.

ZX-3337 #comment progress

Test: see bug above for manual test.
Change-Id: I28f035e3aad92791d91b8acf95624ad77602b820

The inner types all have default constructors and a way to check if they
are initialized so it was redundant to wrap them in std::optional.

Tested: runtests on vim2
Change-Id: I1d14b4b8d70b8cece6e4fac31ee1b49d1bc0c15b

Introduce Controller, a very simple interface between SCSI target
drivers (ex: Disks) and SCSI Controller drivers. It provides a
mechanism for targets to execute commands with a data-in/data-out
region and retrieve the status of the command.

This interface is purely synchronous and non-queued. It is only in place
to support development. In the future it should be replaced with a
Banjo interface.

The plan (see conflicting WIP CL) is that peripherals will be constructed
with a pointer to their Controller and will use this interface to execute
initial (probe, startup, enumerate) commands.

ZX-2314

Tested: Built

Change-Id: I6e25bbd80a79e6dbf652c6a28b980d850ef32d4f

Run `fx format-code` without changing "#pragma once" to header guards.
Also set the line limit to 100 columns in the Zircon clang-format
configuration.

Test: N/A
Change-Id: I7049fcec3ebe80a0ca910ed2be95bdbc0303a596

This is the correct way to move a value out of an optional.

Tested: runtests
Change-Id: I3c0b24eb85f5be0cc91a3fbebc145860ef1b854a

Test: builds / ftl-test

Change-Id: I6ceb6363b143153eee9ac137d2d601f9aabbbe51

Test: builds / ftl-test

Change-Id: Ibbfcf1778f932772d0e2f25f2cf08a35d2f53ff8

Use metadata to specify the size of a protected memory region, which
will be created using contigous memory. Portions of that memory region
can be handed out to clients as requested. Currently the memory isn't
actually marked as protected using the TEE, so it's no different from
regular memory.

Also convert the query of the current VID and PID to read from the
sysmem metadata, as pdev_get_device_info gets information about the
current device, which always PDEV_VID_GENERIC and PDEV_PID_GENERIC for
sysmem.

Test: vkcube on astro

Change-Id: I681f928ccde4e4d99787d2b0853e19ed9b9eaab0

This design isn't necessarily something that will be kept long term, but
it's useful for bring-up of the port while things are solidifying.

Test: Used addr() in debug output.

Change-Id: I69112ae7e678e23ff4841e053efba212127743d7

volume.cpp currently uses a variable length array of type char
allocated on the stack to pass to block_impl_queue. This is not guaranteed
to be aligned to the required ABI requirements, and is also unsafe.

This patch moves the dynamic memory allocation to the heap
and ensures the allocation is at least aligned to alignof(block_op_t)

Test: Ensure that all automated tests pass on ARM64 and x86_64.
Change-Id: I5ccab789b02d9411b84bcd2b05c44d722852429d

This is a reland of a3b4302f

This CL was reverted due to lack of support in measure.
The measure library support was added in:
https://fuchsia-review.googlesource.com/c/garnet/+/246312

TEST: The garnet-*-perf-* bot was failing before the support was added, but now it is green.

PT-72 #done

Original change's description:
> [tracing] Make TRACE_DURATION use trace duration complete events
>
> ... instead of a start and end events. Saving ~50% in buffer and JSON
> sizes.
>
> There is no noticeable change in performance for NTRACE, tracing
> disabled, category disabled and tracing enabled.
>
> The compiled code size is less than or nearly equal the old version for 4
> or less arguments but it is bigger for 8 arguments and more, issue
> tracked in PT-87.
>
> TEST: fx shell /system/test/sys/trace-test
> TEST: fx shell trace-benchmark
>
> PT-72 #done

Change-Id: I21510f2ffd1b7d79a622927bc999ce8cb9375396

This patch enables the USB peripheral driver to obtain its configuration through board-specific metadata.

Test: On Cleo, validate that this configuration works. Validate that the usb-virtual-bus does NOT set a default configuration.

Change-Id: I328931427a4fe30fc3e34f199aa02cfa1ba9cb08

This reverts commit 0f31d965.

Reason for revert: Breaking the zircon roller

Original change's description:
> [vfs] Plumb a hash table around to track koids of tokens
> 
> Rather than rely on set/get_cookie
> 
> Test: fs unit tests
> Change-Id: I03d6b5de4eae90e8bbcef09958fb983b84e86e64

TBR=kulakowski@google.com,smklein@google.com,abarth@google.com

Change-Id: Id223a343a627e8d91539a909b1ea4b7a2494396d
No-Presubmit: true
No-Tree-Checks: true
No-Try: true

Tested: Paved astro
Change-Id: I5bec4ec8cc5107be72a28c487485421b3a4ba446

TEST: USB works on NUC and in qemu with USB virtual bus

Change-Id: I98aeff7c3c0ae7bbdc6a6620261286da7271d5a1

Tested: runtests -t operation-test
Change-Id: I1b1897b3630998631a9b4c82b4e2de3737c849c2

This test goes through the full bind of a device with a devhost, and the
disconnection of the devhost.

ZX-3286

Test: Ran Fuchsia and /system/test/ddk tests.
Change-Id: Ibfc9bc5a2f041772130d9417cd9e032c75b7b587

Test: docs only

Change-Id: Ib67cf051fcb2f41167c2dbf715e5fb07388d2cad

* Return the status of AttemptBind in BindDevice and BindDriver
* Use ZX_CHANNEL_MAX_MSG_BYTES in HandleDeviceRead

ZX-3286

Test: Ran Fuchsia and /system/test/ddk tests.
Change-Id: I18aa9c488f988edd4d863dcaca8845f7148baefb

Rather than rely on set/get_cookie

Test: fs unit tests
Change-Id: I03d6b5de4eae90e8bbcef09958fb983b84e86e64

These bindings have parity with the remaining ioctl declarations.
For direct comparison, refer to: system/public/zircon/device/block.h

Test: Declarations only; bindings utilized in subsequent patches.

Change-Id: I6cfba4834b3d4688e1be92c04da20bd945df4afc