Skip to content
Snippets Groups Projects
  1. Jan 28, 2019
  2. Jan 27, 2019
    • Sean Klein's avatar
      [ramdisk][fidl] Convert ramdisk IOCTLs to FIDL · d84a57a6
      Sean Klein authored
      Additionally, convert all clients within Zircon.
      
      Test: /boot/test/sys/ramdisk-test
      ZX-2674 #comment In Progress
      ZX-2866 #done
      
      Change-Id: Iff7751579a2a2aa496cc71da57ec002f39f15792
      d84a57a6
    • Sean Klein's avatar
      [ramdisk] Avoid usage of 'block' interface for accessing Ramdisk ioctls · 9eba95a2
      Sean Klein authored
      Additionally, redirects all usage of "ramdisk_ioctl" interfaces through
      the ramdisk library (currently located within ulib/fs-management,
      pending relocation). This will make it easier to convert the ioctl
      interfaces to FIDL without modifying client code.
      
      Test: /boot/test/sys/ramdisk-test
      ZX-2674 #comment In Progress
      
      Change-Id: Ia24192df10b2cb3bb5ee2a95f0b6ef647d515d69
      9eba95a2
    • Pascal Perez's avatar
      [fidlc] Deferring Type Construction Post Raw to Flat Conversion · 68817767
      Pascal Perez authored
      Introducing a generic representation named TypeConstructor to capture
      all areas where we reference types. This makes it possible to have a
      single representation, share code around it, and most importantly, defer
      type construction to a later stage in the compiler, once all constants
      have been resolved.
      
      For instance, with the example:
      
          const uint32 MAX_SIZE = 1024;
          struct MyStruct {
            string:MAX_SIZE desc;
            string:1024 other_desc;
          };
      
      We must first resolve the MAX_SIZE constant to understand that the type
      for both desc field, and other_desc field are the same.
      
      Type constructors create types using type templates, which are looked up
      by name. So for instance, the name "uint32" looks up a type template,
      which is an instance of PrimitiveTypeTemplate, and which knows how to
      instantiate a PrimitiveType representing a uint32.
      
      All this machinery moves us closer to 1. canonicalizing types (because
      their creation is now centralized, and soon easy to memoize), and 2.
      introducing user-defined templates such as
      
          struct pair<L, R> { ....
      
      Which would get materialized dynamically.
      
      This also makes it natural to extend type aliasing to use any types, not
      just primitives.
      
      Test: make tools -j12 HOST_USE_ASAN=true && ./build-x64/host_tests/fidl-compiler-test
      Change-Id: Id97c6751b8944b5ee04ff813d2cd0dcc47711f17
      68817767
    • Mike Voydanoff's avatar
      [dev][light-sensor] Move system/dev/light to system/dev/light-sensor · bca77f6b
      Mike Voydanoff authored
      So we can use system/dev/light for light drivers
      
      TEST: still builds
      Change-Id: Ia7a6684c704390d26cedee8209f076125e7a6fec
      bca77f6b
  3. Jan 26, 2019
    • Jocelyn Dang's avatar
      [dev][usb-tester] Add option to specify num packets. · 8ee09ae4
      Jocelyn Dang authored
      This will make it easier to add the cb_on_error_only tests,
      where we'll want to pass in which requests we want callbacks / errors on.
      
      TEST= plugged in FX3 and ran usb-fwloader && runtests -t usb-test
      
      Change-Id: I6078180c520ea8f61574f04316ff5801e809ece0
      8ee09ae4
    • George Kulakowski's avatar
      [fbl][type_support] Remove unused is_void, is_null_pointer, and is_function · 31ad9c8d
      George Kulakowski authored
      Bug: ZX-3357 #comment
      Test: full system build
      Change-Id: If6e600783164f7ca5a5f7924b5ea7598913b57dc
      31ad9c8d
    • George Kulakowski's avatar
      [intel-hda] Directly use std::atomic<State> · 9e17829f
      George Kulakowski authored
      Now that we are using std::atomic, which supports enums, this can be
      more direct.
      
      Test: no functional change
      Change-Id: Ifc08480efdddfd71cc7f336782eab4cff8a3f998
      9e17829f
    • Wez .'s avatar
      Revert "[syscalls] Remove vestigal vmar_..._old() calls." · d33e6153
      Wez . authored
      This reverts commit 70f3d16f.
      
      Reason for revert: Chromium and some of its third-party dependencies were apparently still using the old names, so we'll need to update the
      dependencies, and Chromium, and wait for everything to roll, before we can re-land this. :(
      
      Original change's description:
      > [syscalls] Remove vestigal vmar_..._old() calls.
      > 
      > Test: CQ
      > Change-Id: I8b1f1dd01c5cae31de2da4b9ba0e9f41c0fca137
      
      TBR=kulakowski@google.com,davemoore@google.com
      
      Change-Id: Idee154db1da8111ccdd0b738ff45d5745f20ee1e
      No-Presubmit: true
      No-Tree-Checks: true
      No-Try: true
      d33e6153
    • Sean Klein's avatar
      [third_party][zstd] Add rules.mk file, enabling compilation · f88d16ea
      Sean Klein authored
      Test: Manually included as third_party/ulib/zstd, compiles.
      
      Change-Id: I96ede69e0431a3a2ee5c7f16ceaab8a29f9f4a89
      f88d16ea
    • George Kulakowski's avatar
      [display] Migrate from fbl to std::atomic · fd5d40ff
      George Kulakowski authored
      fbl::atomic is being removed in lieu of std::atomic.
      
      The underlying implementations are identical builtins, so there should
      be no behavior change.
      
      Bug: ZX-3357 #comment
      Test: no functional change
      Change-Id: I731a336aacb23be59d113f3855351ce351ba6474
      fd5d40ff
    • George Kulakowski's avatar
      [userpager] Migrate from fbl to std::atomic · 9cfefcb7
      George Kulakowski authored
      fbl::atomic is being removed in lieu of std::atomic.
      
      The underlying implementations are identical builtins, so there should
      be no behavior change.
      
      Bug: ZX-3357 #comment
      Test: no functional change
      Change-Id: Ibbbbc8350ccd4a34d8d152295f1393a18465b163
      9cfefcb7
    • George Kulakowski's avatar
      [ramdisk] Migrate from fbl to std::atomic · a909b49e
      George Kulakowski authored
      fbl::atomic is being removed in lieu of std::atomic.
      
      The underlying implementations are identical builtins, so there should
      be no behavior change.
      
      Bug: ZX-3357 #comment
      Test: no functional change
      Change-Id: I74a9b26a14174228f7269eb9dfde5e63edc2dbe5
      a909b49e
    • Suraj Malhotra's avatar
      [pdev] Remove pdev_map_mmio_buffer2 · 48cc8f3a
      Suraj Malhotra authored
      Tested: Compiles.
      Change-Id: I5d119c226392edeb52a5bbc9c2060f7326f8bd1d
      48cc8f3a
    • Dustin Green's avatar
      [sysmem] sysmem proxy · 99b7e159
      Dustin Green authored
      Tested: CQ
      Change-Id: Ifc5039eb2274b98dbc701e3291d23877d287e349
      99b7e159
    • Christopher Anderson's avatar
      [upci] Remove binding.c, move ddk binding to Bus.cpp · 0edc040b
      Christopher Anderson authored
      Test: Built/run upci and verified the binding.
      
      Change-Id: I4d30dda6f2a8aa943d77fce58a155d0e5c94f027
      0edc040b
    • Dustin Green's avatar
      [sysmem] Sysmem as a Zircon driver. · cd8b1496
      Dustin Green authored
      Because sysmem has a Zircon driver client (display driver) and a Zircon
      process client (virtcon), sysmem needs to be a Zircon driver.  This CL
      creates the sysmem driver and gets it loaded for all boards.
      
      Future CLs will broker Zircon sysmem service requests to the driver, and
      will broker garnet+ sysmem service requests to the driver (via the
      Zircon service probably).
      
      Currently the only way to contact the sysmem driver is to talk to
      /dev/class/sysmem/000 directly.  See previous paragraph for how that
      will change soon.
      
      Some of the interfaces are being updated.  For now, the old sysmem
      remains in place and still implements the old interfaces, but we'll move
      those clients over soon.
      
      For now, sysmem runs in the platform bus driver devhost (or ACPI
      devhost) to permit sysmem to be contacted (soon) from devhost(s) that
      are children of the platform bus driver devhost (the only way to do
      driver-to-driver coms so far).  In particular, the display driver will
      need to contact the sysmem driver.
      
      Tested: sysmem_tests
      Change-Id: I886ad89e6f166a136aa75f3f1128f71851928f0c
      cd8b1496
    • Christopher Anderson's avatar
      [acpi][upci] Fix acpi build breakage caused by client_remote · c5f32d54
      Christopher Anderson authored
      Test: Verified upci builds again.
      
      Change-Id: If99b40a83631e54eafb79fc3ec23f3c10ddca295
      c5f32d54
    • Marie Janssen's avatar
      [bt][uart] Fix size of acl data buffer · 99f75ef1
      Marie Janssen authored
      The ACL data buffer wasn't large enough to actually contain some data
      packets now that we can send things other than events along it (A2DP).
      
      Increase the buffer size to the max frame size.
      
      Bug: BT-701
      Test: run bt-a2dp-sink, connect from phone and play music 🎶
      Change-Id: I83d1a0bad45e0a22fea1e8b0b324ce4ee9e35958
      99f75ef1
  4. Jan 25, 2019
    • Ed Coyne's avatar
      [kernel][acpi] Increase max IRQs from 64 to 224. · 98968017
      Ed Coyne authored
      On the threadripper 2970wx the kernel sees over 64 IRQs and panics.
      Increase to 224 which is the max supported by LAPICS.
      
      Test: ran on 2970wx, ran 'k ut all' on qemu x86.
      Change-Id: I393b999ef7495fef868c1b4a49e11fe9cb6600d0
      98968017
    • Stephen Demos's avatar
      [fs-test-utils] add fsck function · f304ce1e
      Stephen Demos authored
      this adds an Fsck function to the test utils to allow tests to check the
      integrity of the disk represented by the fixture.
      
      Test: fs-test-utils-test
      Change-Id: I2adcb0dd0044dc1ed56fbc0977d227c9f1002ca9
      f304ce1e
    • Nick Maniscalco's avatar
      [zircon][counters] Add help option to kcounter command · 97155893
      Nick Maniscalco authored
      Add help option to kcounter command and update comments to reference
      "kcounter" instead of "k counters".
      
      Test: Ran "kcounter --help" and "kcounter -h"
      Change-Id: I287c65d81f43f43e0f8fa2675aaf867b48960763
      97155893
    • Pascal Perez's avatar
      [fidlc] Splitting CodedTypesGenerator into its own file · 519033d4
      Pascal Perez authored
      Test: make tools -j12 HOST_USE_ASAN=true && ./build-x64/host_tests/fidl-compiler-test
      Change-Id: I470f516b8a38ce242987562173aa3d4f140ed471
      519033d4
    • Braden Kell's avatar
      [mt8167][sdmmc] Disable HS200 mode · 4dad44e2
      Braden Kell authored
      Paving doesn't work using HS200. Disable this mode
      so that HSDDR (52 MHz) is used instead. HS200 will
      be re-enabled once the stability issues are fixed.
      
      Test: Paving works without hack
            (Id8e0c9c3f52a7b3d2cf484d0fbf2bbda2a701492).
      Change-Id: Ic51e7f655c3f1bcab77a1fa6fa4e7d66b9e090ef
      4dad44e2
    • Braden Kell's avatar
      [mt8167][sdmmc] Set clock mode when initializing · ad54958a
      Braden Kell authored
      Set the clock mode to single data rate in Init().
      After paving it could still be set to DDR which
      causes block reads to return garbage.
      
      Test: 'lsblk' works after paving.
      Test: runtests -t mtk-sdmmc-test
      Change-Id: I61d94b38e24e4c0788fb8ce63c5ce54f7ab77a2b
      ad54958a
    • Tamir Duberstein's avatar
      [docs] fix `object_signal_peer` link · f6edcad5
      Tamir Duberstein authored
      Test: None
      Change-Id: I8a40ab21e9c3a0c4276b913d3d29dadba1a4e8e8
      f6edcad5
    • Pascal Perez's avatar
      [fidlc] Refactor tables generator, and add unit tests · b995c0c4
      Pascal Perez authored
      We split the TablesGenerator into a CodedTypesGenerator which does smart
      things (calculates coded types), and the TablesGenerator which emits
      code (that part is expected to be straightforward string manipulation).
      
      This split makes it possible to test the CodedTypesGenerator more
      easily, and clarifies the surface which needs to be tested.
      
      To reduce the diff size, making the TablesGenerator extend the CodedTypesGenerator
      (vs delegating to it). We may want to refactor that separately, but
      think this is fine as is.
      
      Test: make tools -j12 HOST_USE_ASAN=true && ./build-x64/host_tests/fidl-compiler-test
      Change-Id: I407892448ba706012456a99b2ab03c9e991fb584
      b995c0c4
    • Suraj Malhotra's avatar
      [pdev] Remove map_mmio and update map_mmio_buffer. · 0a021e34
      Suraj Malhotra authored
      map_mmio is no longer necessary as we expect client to perform mapping
      via a different helper library (mmio_buffer_t or ddk::MmioBuffer). In
      addition, the io-buffer version of map_mmio_buffer was removed in favor
      of the mmio_buffer_t variant. All clients were migrated as necessary.
      
      Tested: Compiles.
      Change-Id: I6565c4eef9d575683d2a02e1d60350394e67a76f
      0a021e34
    • Dave Moore's avatar
      [syscalls] Remove vestigal vmar_..._old() calls. · 70f3d16f
      Dave Moore authored
      Test: CQ
      Change-Id: I8b1f1dd01c5cae31de2da4b9ba0e9f41c0fca137
      70f3d16f
    • Pascal Perez's avatar
      [fidlc] Explicit sharing for multi-library tests · 732739cc
      Pascal Perez authored
      Before, the error reporter, typespace, and 'all libraries' was
      different for two dependent libraries. Since we're soon adding state to
      these objects which much be shared, we need to explicitly manage this
      sharing.
      
      Test: make tools -j12 HOST_USE_ASAN=true && ./build-x64/host_tests/fidl-compiler-test
      Change-Id: I74320e1e0d39ffaf2af74b2aaa71c06e25a4ef12
      732739cc
    • David Gilhooley's avatar
      [zircon][hid-parser] Fix unit API arguments · 695729f8
      David Gilhooley authored
      Fix the unit function arguments such that the output
      variable is the last argument and is a pointer instead
      of a reference.
      
      TEST: fx shell runtests -a -t hidparse-test
      
      Change-Id: I2d089bbfd980c53ea68f6cc06938b41f48956cc2
      695729f8
    • Nick Maniscalco's avatar
      [syscalls] Report blocked reason in object_wait_many when count == 0 · 548634aa
      Nick Maniscalco authored
      When zx_object_wait_many is called with no objects, be sure to set the
      blocked reason.
      
      Test: added a new test to /boot/test/sys/thread-state-test
      Change-Id: If532e18cc9cd54425a74deb2c925094357d0f28e
      548634aa
    • Nick Maniscalco's avatar
      [syscalls] Use policy slack in timer_set, wait_many, and nanosleep · 508016b0
      Nick Maniscalco authored
      Use job policy timer slack in zx_timer_set, zx_object_wait_many, and
      zx_nanosleep.
      
      timer_set - Prior to this change, timer_set applied whatever slack was
      passed to the syscall. Now, timer_set applies the maximum of the slack
      argument and the job policy minimum.
      
      wait_many and nanosleep - Prior to this change, zx_object_wait_many
      and zx_nanosleep automatically incorporated late slack between 1
      microsecond and 1 second depending on how far into the future their
      deadline was set. Now they use the default job policy slack, which is
      currently set to 0.
      
      In a future CL, the default job policy slack will change.
      
      Test: Did the following:
      
      1. Zircon - Built and booted a zircon-only system with this change.
      Ran runtests, observed no timer coalescing (via kcounters).  Hacked
      the default to be 10ms of late slack for all jobs, re-ran runtests and
      observed some coalescing.
      
      2. Full System - Built and booted the *full* system on Astro with this
      change.  Saw some timer coalescing (via kcounters).  Hacked the
      default to be 10ms of late slack for all jobs, re-ran runtests and
      observed lots more coalescing.  Verified that general functionality
      still worked.
      
      Bug: ZX-931 #comment use job policy slack
      Change-Id: I55b3f70af7dc29d6776adf4a6ec814396038cdc8
      508016b0
Loading