This is a very small step towards a better interface for bind and
create, internally (rather than exposing thread-local storage!).

This also makes it clear that Binding, at least, does not have the RPC
proxy case.

Change-Id: I8571e655e12002794980448cbd5869b63b95c98b

Missed a few configs before. I hope to go back and tidy these helpers up
once we add a SafeInt library.

[22948->880/23885 ~58] CXX host-x64-linux-clang/obj/system/host/minfs/minfs.main.cpp.o
../system/host/minfs/main.cpp:117:12: warning: implicit conversion loses integer precision: 'ssize_t' (aka 'long') to 'zx_status_t' (aka 'int') [-Wshorten-64-to-32]
    return r;
    ~~~~~~ ^
../system/host/minfs/main.cpp:313:37: warning: implicit conversion loses integer precision: 'unsigned long' to 'uint32_t' (aka 'unsigned int') [-Wshorten-64-to-32]
    uint32_t dir_blocks = dir_count + (dir_bytes_ / minfs::kMinfsBlockSize);
             ~~~~~~~~~~   ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../system/host/minfs/main.cpp:318:37: warning: implicit conversion loses integer precision: 'long' to 'uint32_t' (aka 'unsigned int') [-Wshorten-64-to-32]
    info.block_count = data_blocks_ + dir_blocks;
                     ~ ~~~~~~~~~~~~~^~~~~~~~~~~~
../system/host/minfs/main.cpp:580:37: warning: implicit conversion loses integer precision: 'size_t' (aka 'unsigned long') to 'uint8_t' (aka 'unsigned char') [-Wimplicit-int-conversion]
    dir_bytes_ += minfs::DirentSize(strlen(last_slash));
                  ~~~~~             ^~~~~~~~~~~~~~~~~~
../system/host/minfs/main.cpp:585:67: warning: implicit conversion loses integer precision: 'long' to 'uint32_t' (aka 'unsigned int') [-Wshorten-64-to-32]
    uint32_t remaining = (file_size + minfs::kMinfsBlockSize - 1) / minfs::kMinfsBlockSize;
             ~~~~~~~~~   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
5 warnings generated.

BLD-361 #comment [gn][minfs] Fix a few minor warnings in new gn build
ZX-3415 #comment [gn][minfs] Fix a few minor warnings in new gn build

Test: CQ
Change-Id: I7e620abc62f9b9328a4066c465f33102ff331f62

[22813->1015/23885 ~58] CXX host-x64-linux-clang/obj/system/host/xdc-server/xdc-server.xdc-server.cpp.o
../system/host/xdc-server/xdc-server.cpp:82:30: warning: comparison of integers of different signs: 'ssize_t' (aka 'long') and 'size_t' (aka 'unsigned long') [-Wsign-compare]
        while (total_written < len_to_write) {
               ~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~
../system/host/xdc-server/xdc-server.cpp:95:45: warning: implicit conversion loses integer precision: 'long' to 'int' [-Wshorten-64-to-32]
            int offset = transfer->offset() + res;
                ~~~~~~   ~~~~~~~~~~~~~~~~~~~^~~~~
../system/host/xdc-server/xdc-server.cpp:119:21: warning: implicit conversion loses integer precision: 'ssize_t' (aka 'long') to 'int' [-Wshorten-64-to-32]
            int n = recv(fd(), buf, UsbHandler::Transfer::MAX_WRITE_DATA_SIZE, 0);
                ~   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../system/host/xdc-server/xdc-server.cpp:276:37: warning: implicit conversion loses integer precision: 'std::__2::vector<pollfd, std::__2::allocator<pollfd> >::size_type' (aka 'unsigned long') to 'int' [-Wshorten-64-to-32]
        int num_sockets = poll_fds_.size();
            ~~~~~~~~~~~   ~~~~~~~~~~^~~~~~
4 warnings generated.

BLD-361 #comment [gn][xdc-server] Fix a few warnings in new gn build for xdc-server
ZX-3415 #comment [gn][xdc-server] Fix a few warnings in new gn build for xdc-server

Test: CQ
Change-Id: I35f975285dffd647e2bc7834a6c636b14809caab

[22816->1012/23885 ~58] CXX host-x64-linux-clang/obj/system/host/banjo/formatter/banjo-format.main.cpp.o
../system/host/banjo/formatter/main.cpp:75:16: warning: comparison of integers of different signs: 'int' and 'std::__2::vector<std::__2::basic_string<char>, std::__2::allocator<std::__2::basic_string<char> > >::size_type' (aka 'unsigned long') [-Wsign-compare]
    while (pos < args.size() && args[pos] != "--" && args[pos].find("-") == 0) {
           ~~~ ^ ~~~~~~~~~~~
../system/host/banjo/formatter/main.cpp:87:13: warning: comparison of integers of different signs: 'int' and 'std::__2::vector<std::__2::basic_string<char>, std::__2::allocator<std::__2::basic_string<char> > >::size_type' (aka 'unsigned long') [-Wsign-compare]
    if (pos >= args.size()) {
        ~~~ ^  ~~~~~~~~~~~
../system/host/banjo/formatter/main.cpp:95:16: warning: comparison of integers of different signs: 'int' and 'std::__2::vector<std::__2::basic_string<char>, std::__2::allocator<std::__2::basic_string<char> > >::size_type' (aka 'unsigned long') [-Wsign-compare]
    for (; pos < args.size(); pos++) {
           ~~~ ^ ~~~~~~~~~~~
3 warnings generated.

BLD-361 #comment [gn][banjo] Fix some warnings in banjo gn build
ZX-3415 #comment [gn][banjo] Fix some warnings in banjo gn build

Test: CQ
Change-Id: If31eaabfd01fcb0502db9f1bdf20a5794ad72730

[22794->1034/23885 ~58] CXX host-x64-linux-clang/obj/system/host/fidl/formatter/fidl-format.main.cpp.o
../system/host/fidl/formatter/main.cpp:78:16: warning: comparison of integers of different signs: 'int' and 'std::__2::vector<std::__2::basic_string<char>, std::__2::allocator<std::__2::basic_string<char> > >::size_type' (aka 'unsigned long') [-Wsign-compare]
    while (pos < args.size() && args[pos] != "--" && args[pos].find("-") == 0) {
           ~~~ ^ ~~~~~~~~~~~
../system/host/fidl/formatter/main.cpp:92:13: warning: comparison of integers of different signs: 'int' and 'std::__2::vector<std::__2::basic_string<char>, std::__2::allocator<std::__2::basic_string<char> > >::size_type' (aka 'unsigned long') [-Wsign-compare]
    if (pos >= args.size()) {
        ~~~ ^  ~~~~~~~~~~~
2 warnings generated.

BLD-361 #comment [gn][fidl] Fix signedness warning in fidl formatter
ZX-3415 #comment [gn][fidl] Fix signedness warning in fidl formatter

Test: CQ
Change-Id: Ic29ea8a601910f991d3c74ca6f93978523b08168

ninja: Entering directory `build-gcc'
[4399->238/4694 ~58] CXX host-x64-linux-clang/obj/system/host/fvm/fvm.main.cpp.o
../system/host/fvm/main.cpp:50:28: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
    for (unsigned i = 0; i < argc; i += 2) {
                         ~ ^ ~~~~
../system/host/fvm/main.cpp:128:14: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
    while (i < argc) {
           ~ ^ ~~~~
../system/host/fvm/main.cpp:129:50: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
        if (!strcmp(argv[i], "--slice") && i + 1 < argc) {
                                           ~~~~~ ^ ~~~~
../system/host/fvm/main.cpp:140:58: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
        } else if (!strcmp(argv[i], "--offset") && i + 1 < argc) {
                                                   ~~~~~ ^ ~~~~
../system/host/fvm/main.cpp:145:58: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
        } else if (!strcmp(argv[i], "--length") && i + 1 < argc) {
                                                   ~~~~~ ^ ~~~~
5 warnings generated.

BLD-361 #comment [gn][fvm] Fix some warnings in host build
ZX-3415 #comment [gn][fvm] Fix some warnings in host build

Test: CQ
Change-Id: Ib9b92867d1abb025af77b4ce47638dbaf3519d17

[8979->855/9891 ~58] CXX host-x64-linux-clang/obj/system/host/fidl/lib/libfidl.lexer.cpp.o
../system/host/fidl/lib/lexer.cpp:284:9: warning: unannotated fall-through between switch labels [-Wimplicit-fallthrough]
        case '-':
        ^
../system/host/fidl/lib/lexer.cpp:284:9: note: insert '__FALLTHROUGH;' to silence this warning
        case '-':
        ^
        __FALLTHROUGH;
../system/host/fidl/lib/lexer.cpp:284:9: note: insert 'break;' to avoid fall-through
        case '-':
        ^
        break;
1 warning generated.

BLD-361 #comment [gn][fidlc] Fix warning in gn build in fidl lexer
ZX-3415 #comment [gn][fidlc] Fix warning in gn build in fidl lexer

Test: CQ
Change-Id: Ia19039629d5699bb5a1e6d750137940aef226a6a

build-zircon.sh and thus `fx build-zircon` now runs gn (if necessary) and
ninja (just once) rather than running `make` several times.  The extra
arguments previously taken as make arguments are now taken as either GN
build arguments or ninja arguments.

`fx set` now runs Zircon's `gn gen` before Fuchsia's `gn gen`, which
consumes files generated by the former.

Zircon-related GN build arguments are gone.  Zircon configuration
can be controlled via arguments to build-zircon.sh instead.

The new behavior is roughly equivalent to `zircon_boot_groups="core,misc"
zircon_system_groups=""`.  That is, the BOOTFS in the Fuchsia boot images
always contains all of the Zircon binaries, excluding the tests.  The
Zircon tests do not appear in either `/boot/test` or `/system/test`.

Bug: BLD-325 #comment Fuchsia GN integration adapted to Zircon GN legacy support
Test: manual
Change-Id: Ie0b8b76c31020d3cb236b781d496fb702779c0a8

This is almost certainly missing things, but replaces many docs
locations where Make commands or arguments are described.

Change-Id: I74e83c5606fdbbc6d4734c6002ee4db637a4899f

The obsolete scripts build-all-zircon, make-parallel, and
make-release are gone entirely.  The build-zircon script is
updated to drive the GN+Ninja build using approximately the
same command line UI.  The buildall script is updated to use
build-zircon.  The run-zircon script is updated to default to
the new locations used by build-zircon.

Bug: BLD-325
Test: ./scripts/buildall -c -A -r && ./scripts/run-zircon-x64
Change-Id: I4c8b71a804ef81b71bd535e6bdef28e683f22450

Bug: BLD-325
Change-Id: If88c836c805ab92657f091371c6a4033b37aa281

Cache the current head of pages in memory, so they are always available.

LE-692 #done

Change-Id: If76927381f0a5df99819437ee4e5d1b98c7c09a0

SO_TIMESTAMP was removed from transport endpoints in
https://github.com/google/gvisor/commit/a9cb3dc. We need to stop relying
on that support before upgrading our dependency on
github.com/google/netstack in order to achieve a soft transition.

Change-Id: I4883940df30a7d2b252323a3255f9cf33443be9f

Currently, this file keeps track of default persona id only. It is meant
to be extended later.

Tested: CQ

AUTH-106

Change-Id: I54aec96f1f6f8ebd3a3e778bd1a44774110e16bc

Test: read the docs

Changed HTML tables in .md files to native markdown.

Change-Id: I14944f0cb46832158679e021c6c74796678c74eb

Test: read the docs

This introduces the "bits" declaration.

Change-Id: I323cd742183911f8a10253c302c73af3360fd526

The format rules changed a couple of weeks back, this gets the diff out
of the way so it doesn't clutter future work.

Change-Id: Id51a945e706fb7622905b937b0ab2c3604ce57f4

Test: read the docs

Change-Id: I3817e82de52bf7032edeece2bb191f5297da3f98

Tested: CQ
Change-Id: I73810cdefc0da83f7cafce41e16c7c4b01617e10

TEST: fx run-test mediasession_tests
Change-Id: Iacdf3174c75ffe7ba618aa90b336b2a813ce1d51

IdAllocator can be used to alloc and free unique ids within range
[0-id_count). IdAllocator treats id like a resource. alloc and
free have O(logn) time complexities.

Around 128KiB should be sufficient for (1<<20) Ids. Approximately
1.016 bits per id are used to track id's state.

Tested: Added id-alloc-bitmap test cases.

Change-Id: I9e888ddc6dea4e71ce9c8f63caade8f716980945

In the time conversion macros (like ZX_SEC), don't cast intermediate
results. Perform the multiplication and only cast at the end. There
could be code that passes floats or doubles to these macros so delay
the cast to minimize loss of precision due to truncation.

Change I62c8386a0805a850c657a444041f5e69b99709c2 inadvertently altered
the behavior of these macros when they are called with float or double
values. Compiler flags used in zircon prohibit code from calling these
macros with floats or doubles, but that's not the case outside of
zircon. Because of that change, ZX_MSEC(0.5) would evaluate to 0
rather than 500000.

See also PT-98 and ZX-130.

Bug: ZX-3446 #done
Test: added a temp static_assert(ZX_MSEC(0.5), 500000) to garnet
Change-Id: I73ffe210d0702da4c41b4b77a0442bf1e3bc8f82

Change-Id: Ib91c6fd4e0f09fac16e96abc06a3440755c1a282

- testutil: Add DummyEventDispatcherBuilder, which is capable of
  storing the configuration of a network stack including forwarding
  table entries, devices and their assigned IP addresses, ARP table
  entries, etc; its build method produces a Context<DummyEventDispatcher>
  with all of the appropriate state configured
- testutil: Add DummyEventDispatcherConfig, which holds a simple
  configuration of an IPv4 network with a single local subnet,
  a single local interface, a single local IP address, and a single
  remote IP address
- ip: Add add_route function

Test: This functionality is used in tests in later commits

Change-Id: Ia2180037b61399472b1b033a7fd17bbec6103d37

- Adds special helper for device settings that allows setting up a
mocked device settings interface in emulated environments.

- Added new test for mocked service.
- Updated README with new helper.
- [bonus change] Updated sandbox to *ALWAYS* run "setup" before apps and tests (was
running setup after apps).

TEST: added new test, run with rx shell run_test_component
fuchsia-pkg: //fuchsia.com/netemul_sandbox_test#meta/mock_device_settings.cmx
Change-Id: Ia1a879e161a8fb33e8f103c8c242924ca72fba73

- DESIGN.md: Update the documentation relating to parsing and
  serializing packets to be up-to-date with our current architecture,
  which makes use of the utilities provided by the packet crate
- Move the section on control flow to the top of the document since
  it's more important

Change-Id: I11cd3670c5ba8357beacd875fb4f52d8fd6b2ed8

- Add dedicated "Performance Issues" section
- Add "Single-threaded execution" sub-section
- Add "`debug_err!` macro" sub-section
- Add "Pervasive use of `HashMap`s" sub-section
- Expand "Devices in `HashMap`s" sub-section

Change-Id: I6721b556cf21edf03f24fc21b3282b4e5fb64461

This CL removes a copy. Previously, all reads from the upstream source
were copied into a temporary buffer and then copied into
SparseByteBuffer with separate allocations and frees for regions.

Now, the upstream reader writes directly to an address provided by
SlidingBuffer.

This results in a measured framerate increase of 4.9% with --loop and
1.6% with --test-seek, which I assume are more significant than
suggested because the test used SW decode and SW colorspace conversion.

TEST: fx run-test mediaplayer_tests

Change-Id: I9b6665fedaf7e61ddc33b7be4905d25d3d786958

- transport::{tcp,udp}::receive_ip_packet: Return error in case of
  unrecognized connection in preparation for sending ICMP responses
- transport: Remove receive_ip_packet function
- ip: Call directly into transport::{tcp,udp}::receive_ip_packet
  rather than into transport::receive_ip_packet

Test: Tested in a later commit which adds support for ICMP responses
      in the case of this error

Change-Id: I8fdd552a3e1d27795cbb42f025a379588ed8bb07

Includes updating journal unittests and documentation.

Test: JournalTest.*
Change-Id: Ibaaba817facc9deb191f7d480821beb031a10a0c

sbase is now provided instead from a large number of standalone packages.

Change-Id: I335946449c1278b6ed73a795d6dc2eb860cf73b0

- Add drop_packet! macro, which takes care of the common operation
  of pulling the source IP, destination IP, protocol, and parse
  metadata out of a parsed packet, and then dropping the packet so
  that the underlying buffer it was parsed from can be used again
- Add drop_packet_and_undo_parse! macro, which calls drop_packet!,
  but also uses the parse metadata to undo the effects of parsing,
  leaving the underlying buffer containing the entire IP packet
  (including header)
- Replace instances of these patterns with calls to these macros

Change-Id: Ic7115fed579871e60ea7be15aa2afa19254e46dc

We'll soon want to be able to pass fuchsia.sysmem.SingleBufferSettings
into vulkan drivers, and they need to be able to read information about
the format. By adding magma functions we can allow that without causing
them to depend on the precise layout of the structure.

This takes in an encoded fidl message instead of a
fuchsia_sysmem_SingleBufferSettings struct because
fuchsia.sysmem.SingleBufferSettings will eventually change to being a
fidl table, and we want to use the wire format so it's as compatible as
possible.

Test: vim2:go/magma-tps#L0

Change-Id: Ie5b417761f61d22d49cd65a5e04c7bfe83b67053

Bug: ZX-3357 #comment atomics
Change-Id: I79ceadf572b0e2b3a983e7f25207376b08ae2427

All the clients have migrated to fdio_fd_clone and fdio_fd_transfer,
respectively.

Change-Id: I50d4814c3b1013edc8a885ba8b2af5ad764c3ced

- device::arp: Add 'insert' function
- device::ethernet: Add 'insert_arp_table_entry' function
- While we're at it...
  - device::ethernet: Make EthernetIpExt public
  - device: Make DeviceId::new_ethernet public

Test: This functionality is exercised by later commits

Change-Id: I54949cc740a280705ca597890b235707a78be00a

The boot-shim environment (MMU disabled) does not permit unaligned
memory accesses.  The optimized assembly routines rely on them.

Bug: BLD-325
Test: manual
Change-Id: Ic3f1f6706482df9b31d2178e799f5c1312cf578b

Bug: BLD-325
Test: gn gen && ninja
Change-Id: I263457bbde19928d0f5003622552785a48f270ed

- Add IpAddress type, which is an enum of Ipv4Addr and Ipv6Addr
- Add IpSubnet type, which is an enum of Subnet<A> where A is
  Ipv4Addr and Ipv6Addr
  - Add 'unsafe const fn new_unchecked' constructor; this will be
    removed once trait bounds other than Sized are supported in
    const fns

Change-Id: I8eb002ef73bf334a051b5bf75494ada0ce07e6d8

This application can launch on boot to communicate protected memory
information between the sysmem and the secmem TA.

Test: vkcube --protected_output on astro
MTWN-231 #comment

Change-Id: If0a89a3995671ed7d8039424a1c31983262908d2

Test: display-test on NUC
Change-Id: I78d6fb29da8a5da0c736ba8ff4568ef6cdf03cc0