FileCreatorHelper creates a FileBlobImpl on the main-thread and, because of
this, we end up executing I/O operations on that thread, slowing down other
components. With this patch, FileCreatorHelper logic is moved to PBackground.
That the 'type' getter is still called on the main-thread because FileBlobImpl
uses nsIMIMEService which is a non thread-safe component.
Differential Revision: https://phabricator.services.mozilla.com/D27641
--HG--
extra : moz-landing-system : lando
FileCreatorHelper creates a FileBlobImpl on the main-thread and, because of
this, we end up executing I/O operations on that thread, slowing down other
components. With this patch, FileCreatorHelper logic is moved to PBackground.
That the 'type' getter is still called on the main-thread because FileBlobImpl
uses nsIMIMEService which is a non thread-safe component.
Differential Revision: https://phabricator.services.mozilla.com/D27641
--HG--
extra : moz-landing-system : lando
This consolidates array, maybe and unique ptrs in IPDL into a single
"has base type" qualifier, for types that wrap another type. I'm not
sure this patch fixes everything, but I think it is at least more
correct.
It also adds checking for the stuff inside the UniquePtr<>, because
the intent seems to be to allow things like protocol types in there.
Differential Revision: https://phabricator.services.mozilla.com/D28571
--HG--
extra : moz-landing-system : lando
Currently, when deserialization fails, the error message contains the base
name of the C++ type, without template parameters; this means we can get
unhelpfully vague errors about `RefPtr` or `Maybe` or `nsTArray`. (The
identical error cases are then merged by the compiler, so the execution
path can't be recovered from the crash dump.)
This patch uses the IPDL type instead. It would be possible to invoke
the code generator to use the full C++ type, but more complicated for no
real benefit.
Differential Revision: https://phabricator.services.mozilla.com/D28401
--HG--
extra : moz-landing-system : lando
FileCreatorHelper creates a FileBlobImpl on the main-thread and, because of
this, we end up executing I/O operations on that thread, slowing down other
components. With this patch, FileCreatorHelper logic is moved to PBackground.
That the 'type' getter is still called on the main-thread because FileBlobImpl
uses nsIMIMEService which is a non thread-safe component.
Differential Revision: https://phabricator.services.mozilla.com/D27641
--HG--
extra : moz-landing-system : lando
PPluginInstance's __delete__ was ported from rpc to intr, and is the
only non-async IPDL deletion in the codebase. We should be able to
align this with the rest of the IPDL interfaces and remove the need
for specialized intr __delete__ handling.
Differential Revision: https://phabricator.services.mozilla.com/D25674
--HG--
extra : moz-landing-system : lando
Add a new nsExternalHelperAppService derived class named nsOSHelperAppServiceChild to be used for the MIME service, external helper app service, and external protocol service interfaces in child processes. nsOSHelperAppServiceChild overrides some methods used to get MIME and external protocol handler information from the OS and implements these methods by remoting the calls to the parent process.
This is necessary because, on Mac, querying the OS for helper application info from sandboxed content processes is unreliable and has buggy side effects.
For now, only use the new class on Mac.
Android and unix file changes r+ by gcp.
Windows files changes r+ by bobowen.
Sync messages review r+ by nfroyd.
MozReview-Commit-ID: 63BiS6VCxfn
Differential Revision: https://phabricator.services.mozilla.com/D15620
--HG--
extra : moz-landing-system : lando
Add a new nsExternalHelperAppService derived class named nsOSHelperAppServiceChild to be used for the MIME service, external helper app service, and external protocol service interfaces in child processes. nsOSHelperAppServiceChild overrides some methods used to get MIME and external protocol handler information from the OS and implements these methods by remoting the calls to the parent process.
This is necessary because, on Mac, querying the OS for helper application info from sandboxed content processes is unreliable and has buggy side effects.
For now, only use the new class on Mac.
Android and unix file changes r+ by gcp.
Windows files changes r+ by bobowen.
Sync messages review r+ by nfroyd.
MozReview-Commit-ID: 63BiS6VCxfn
Differential Revision: https://phabricator.services.mozilla.com/D15620
--HG--
extra : moz-landing-system : lando
Now that fields are packed nicely, we can take advantage of the
contiguous layout of POD fields and read/write all the POD fields of a
given size in a single read/write call. For many structs, this should
have little or no effect, but for large structs such as LoadInfoArgs,
this reduces the number of function calls by ~50%.
Differential Revision: https://phabricator.services.mozilla.com/D22001
--HG--
extra : moz-landing-system : lando
We're about to start depending on how the fields are packed in a future
patch, so we should add some compile-time checking that our assertions
are correct.
Differential Revision: https://phabricator.services.mozilla.com/D22000
--HG--
extra : moz-landing-system : lando
We're going to read and write sentinels slightly differently for
bulk-writing adjacent fields, so let's factor out some reusable code for
doing so.
Differential Revision: https://phabricator.services.mozilla.com/D21999
--HG--
extra : moz-landing-system : lando
This patch changes the layout of IPDL-defined structs to order the POD
members by decreasing size, which ensures everything is packed well.
This optimization is only applied to the internal representation; the
external interface (e.g. constructors) is entirely unchaged.
Differential Revision: https://phabricator.services.mozilla.com/D21998
--HG--
extra : moz-landing-system : lando
This patch computes an ordering for the fields of an IPDL structure decl
such that they are packed well in memory. We'll take advantage of this
ordering in future patches.
Differential Revision: https://phabricator.services.mozilla.com/D21997
--HG--
extra : moz-landing-system : lando
This actor won't be being used anymore, and acts only as a maintenance burden
for people working on this code (which we're doing pretty often these days!).
Differential Revision: https://phabricator.services.mozilla.com/D20549
--HG--
extra : moz-landing-system : lando
This actor won't be being used anymore, and acts only as a maintenance burden
for people working on this code (which we're doing pretty often these days!).
Differential Revision: https://phabricator.services.mozilla.com/D20549
--HG--
extra : moz-landing-system : lando
This follows the naming convention of the related functions _splitClassDeclDefn() and _splitFuncDeclDefn().
Depends on D20056
Differential Revision: https://phabricator.services.mozilla.com/D20057
--HG--
extra : moz-landing-system : lando
Bug 1428984 caused a regression where ipdlc will emit (empty) method definitions for pure methods. C++ allows a pure member function to have a definition (in case a derived class wants to call a default implementation in the base class), but we don't want this for ipdl's generated code.
Differential Revision: https://phabricator.services.mozilla.com/D20056
--HG--
extra : moz-landing-system : lando
nsSystemInfo is initialzied at first page load. Actually, content process uses
sync IPC to get Android OS information. But now, we can use Java code even if
on content process, so we should use JNI directly instead of sync IPC.
Also, nsSystemInfo still has unused extern android_sdk_version that is for
HoneyComp's DNS hack. So let's remote it.
Differential Revision: https://phabricator.services.mozilla.com/D20129
--HG--
extra : moz-landing-system : lando
Consequently, this removes:
- MOZ_LIBPRIO, which is now always enabled.
- non_msvc_compiler, which is now always true.
- The cl.py wrapper, since it's not used anymore.
- CL_INCLUDES_PREFIX, which was only used for the cl.py wrapper.
- NONASCII, which was only there to ensure CL_INCLUDES_PREFIX still
worked in non-ASCII cases.
This however keeps a large part of detecting and configuring for MSVC,
because we still do need it for at least headers, libraries, and midl.
Depends on D19614
Differential Revision: https://phabricator.services.mozilla.com/D19615
--HG--
extra : moz-landing-system : lando
The one exception is refcounted types, because std::move(RefPtr<T>) does not
coerce to T*, which is what the current IPC methods accept.
This does not rewrite all Recv/Answer methods to take advantage of move
semantics.
Differential Revision: https://phabricator.services.mozilla.com/D19669
--HG--
extra : moz-landing-system : lando
This removes all .flake8 files except for the one at the root of the repo.
Instead we use the new 'per-file-ignores' config introduced in 3.7. To ignore
specific errors in a subdirectory, add a line like this to the root .flake8:
[per-file-ignores]
path/to/subdir/*: E100, F200, ...
The reasons for this change are:
1. Unblock flake8 blacklist (bug 1367092).
2. Simplify configuration and code.
3. Encourage more consistent styling.
4. Improve performance.
5. Greater editor consistency.
Differential Revision: https://phabricator.services.mozilla.com/D18354
--HG--
extra : moz-landing-system : lando
This bumps flake8 to version 3.7.5.
This also ignores the new lint rules that were added in the new versions.
These rules are de-marked via comment so we know that they should be enabled at
some point (as opposed to the other rules that are (presumably) ignored
intentionally.
Differential Revision: https://phabricator.services.mozilla.com/D18353
--HG--
extra : moz-landing-system : lando
For cases where the class has direct calls (that is, we cast `this` to the
subclass before making the call) no longer declare Recv/Answer methods on the
base class at all. This should ensure that slots for them are not generated in
vtables, and also allow the derived class to choose the method signature (e.g.
whether it wants to take something by reference or by value).
Differential Revision: https://phabricator.services.mozilla.com/D18132
--HG--
extra : moz-landing-system : lando
For cases where the class has direct calls (that is, we cast `this` to the
subclass before making the call) no longer declare Alloc/Dealloc methods on the
base class at all. This should ensure that slots for them are not generated in
vtables, and also allow the derived class to choose the method signature (e.g.
whether it wants to take something by reference or by value).
Differential Revision: https://phabricator.services.mozilla.com/D18131
--HG--
extra : moz-landing-system : lando
When calling a Recv/Alloc/Dealloc method on most types, cast `this` to the
derived class.
There is a heuristic to figure out what the correct derived type is. There is a
blacklist of types which we can't do direct calls on for the moment, as well as
an override for types that do work with direct calls but which don't match the
heuristic.
Differential Revision: https://phabricator.services.mozilla.com/D16492
--HG--
extra : moz-landing-system : lando
In order to enable asynchronous launch, destruction of
GeckoChildProcessHost (and its subclasses) has to be delayed until after
launching (or anything else that might be made asynchronous in the
future) has completed, to prevent use-after-free. However, there are
other dependencies on process hosts always being destroyed on the I/O
thread, so refcounting would be difficult to use.
Instead, GeckoChildProcessHost now may not be destroyed directly, but
must go through a method that handles the scheduling.
There are also some minor cleanups to the affected headers (removed
duplicate access modifiers, and made PluginProcessParent final).
Depends on D18010
Differential Revision: https://phabricator.services.mozilla.com/D18011
--HG--
extra : moz-landing-system : lando
This file has text-substitution placeholders that aren't part of normal
C++ syntax; they were broken by auto-inserted whitespace. This patch
restores the original formatting and protects them from further change.
Differential Revision: https://phabricator.services.mozilla.com/D18010
--HG--
extra : moz-landing-system : lando
This commit adds categories to all markers. This way the profiler's
marker categories and frame label categories agree. There are a few
duplicate category properties on some of the marker payloads, but
this could be cleaned up in a follow-up if needed.
Differential Revision: https://phabricator.services.mozilla.com/D16864
--HG--
extra : moz-landing-system : lando
MozPromise most common use is to have an single or exclusive listener. By making the MozPromise generated by IPDL exclusive we can also use move semantics.
While at it, we also use move semantics for the ResponseRejectReason and via the callback's reject method so that the lambda used with the MozPromise::Then can be identical to the one used by the IPDL callback.
As it currently is, it provides no advantage over a copy as it's just an enum; however, this will facilitate future changes where it may not be.
Differential Revision: https://phabricator.services.mozilla.com/D13906
--HG--
extra : moz-landing-system : lando
When shutting down a content process, we call `Close` on the
`IToplevelProtocol`. This causes the MessageChannel to be `Close`-ed,
which in turn sends a `GOODBYE_MESSAGE`:
https://searchfox.org/mozilla-central/rev/876022232b15425bb9efde189caf747823b39567/ipc/glue/MessageChannel.cpp#2852
This message is intercepted on the I/O thread in the content process,
before any code is informed in content, and used to set the
`mChannelState` property to `ChannelClosing`:
https://searchfox.org/mozilla-central/rev/876022232b15425bb9efde189caf747823b39567/ipc/glue/MessageChannel.cpp#1176
Once this state has been set, which is performed as soon as the
message is received, whether or not other messages have been processed
yet, no messages can be sent back to the parent process. This is
usually what causes the 'Too late to send/recv' message spam in the
console, as we're still trying to send messages at this time.
Usually this is fine - the message send fails, but we gracefully
recover, and the process begins shutting down like normal.
Unfortunately, child actor constructors currently have code
automatically generated in them which causes a process crash if the
send fails. As it's impossible for the main thread to know that the
channel has been closed ahead of time (due to this happening
out-of-band), we can then cause random content process crashes
during shutdown due to actor construction.
Unfortunately, we can't just destroy the actor, as our caller may
(and often do) depend on the actor reference they gave us still being valid
after calling Send*Constructor. Fortunately, if a message send failed, it means
we're in the process of being shut down.
This patch handles this by ignoring ctor send errors, and treating them like
messages which successfully were queued to send, but got lost due to the other
side hanging up. The actor will be gracefully destroyed in DestroySubtree when
its manager is destroyed.
Differential Revision: https://phabricator.services.mozilla.com/D12695
- modify line wrap up to 80 chars; (tw=80)
- modify size of tab to 2 chars everywhere; (sts=2, sw=2)
--HG--
extra : rebase_source : 7eedce0311b340c9a5a1265dc42d3121cc0f32a0
extra : amend_source : 9cb4ffdd5005f5c4c14172390dd00b04b2066cd7
There's now an upper limit for snapshot prefilling. The value is configurable and is currently set to 4096 bytes.
Snapshots can operate in multiple modes depending on if all items have been loaded or all keys have been received. This should provide the best performance for each specific state.
This patch also adds support for creating explicit snapshots which can be used for testing.
This improves performance a lot in cases when multiple operations are invoked by a single JS function (number of sync IPC calls is reduced to a minimum). It also improves correctness since changes are not visible to other content processes until a JS function finishes.
The patch implements core infrastructure, all items are sent to content when a snapshot is initialized and everything is fully working. However, sending of all items at once is not optimal for bigger databases. Support for lazy loading of items is implemented in a following patch.
The implementation is based on a cache (datastore) living in the parent process and sync IPC calls initiated from content processes.
IPC communication is done using per principal/origin database actors which connect to the datastore.
The synchronous blocking of the main thread is done by creating a nested event target and spinning the event loop.
When shutting down a content process, we call `Close` on the
`IToplevelProtocol`. This causes the MessageChannel to be `Close`-ed,
which in turn sends a `GOODBYE_MESSAGE`:
https://searchfox.org/mozilla-central/rev/876022232b15425bb9efde189caf747823b39567/ipc/glue/MessageChannel.cpp#2852
This message is intercepted on the I/O thread in the content process,
before any code is informed in content, and used to set the
`mChannelState` property to `ChannelClosing`:
https://searchfox.org/mozilla-central/rev/876022232b15425bb9efde189caf747823b39567/ipc/glue/MessageChannel.cpp#1176
Once this state has been set, which is performed as soon as the
message is received, whether or not other messages have been processed
yet, no messages can be sent back to the parent process. This is
usually what causes the 'Too late to send/recv' message spam in the
console, as we're still trying to send messages at this time.
Usually this is fine - the message send fails, but we gracefully
recover, and the process begins shutting down like normal.
Unfortunately, child actor constructors currently have code
automatically generated in them which causes a process crash if the
send fails. As it's impossible for the main thread to know that the
channel has been closed ahead of time (due to this happening
out-of-band), we can then cause random content process crashes
during shutdown due to actor construction.
Unfortunately, we can't just destroy the actor, as our caller may
(and often do) depend on the actor reference they gave us still being valid
after calling Send*Constructor. Fortunately, if a message send failed, it means
we're in the process of being shut down.
This patch handles this by ignoring ctor send errors, and treating them like
messages which successfully were queued to send, but got lost due to the other
side hanging up. The actor will be gracefully destroyed in DestroySubtree when
its manager is destroyed.
Differential Revision: https://phabricator.services.mozilla.com/D12695
Previously there were many different inconsistent implementations of this
variable across many different actors. This adds a unified implementation of
this variable inside of IProtocol.
Differential Revision: https://phabricator.services.mozilla.com/D12956
Content process of Android uses sync IPC when initializing LookAndFeel. But
current e10s has LookAndFeel cache for start up of content process.
So we should use it, then remove sync IPC for start up performance
Differential Revision: https://phabricator.services.mozilla.com/D9750
--HG--
extra : moz-landing-system : lando
This patch adds the ability to use UniquePtr<T> in IPDL messages if T is serializable. The behavior works as expected -- a UniquePtr in a Send is cleared and is passed by move.
Some design points:
* The UniquePtr identification is done in the parser. This is because the parser immediately normalizes CxxTemplateInst -- an old version of the patch did hacky string-parsing to pull it apart. I, instead, created CxxUniquePtrInst.
* The implementation allows passing to Send... by move or by reference. This is valid UniquePtr API behavior but passing by reference is not really useful in this case (but not harmful). This could probably piggy-back on the "moveonly" IPDL work but that is newer than this work and will require some refactoring.
Differential Revision: https://phabricator.services.mozilla.com/D9143
--HG--
extra : moz-landing-system : lando
I'd like to use a struct over IPDL which is move only. This patch adds a modifier to
usings statements to indicate the C++ type should always be moved, similar to refcounted.
Differential Revision: https://phabricator.services.mozilla.com/D6781
--HG--
extra : rebase_source : 492309e2ec4f4cc6a6ffe639e3c6e4adfb6ecf08
This patch allows IPDL tests to compile -- in other words, it fixes the build when the --enable-ipdl-tests flag is present. Most of the fixes are simple API updates but we also needed to fix an issue with the constness of nsTArray<Shmem> parameters that are defined in IPDL structs.
This patch does not fix the tests themselves -- some IPDL tests will still fail when run.
Differential Revision: https://phabricator.services.mozilla.com/D6705
--HG--
extra : moz-landing-system : lando
The framework to simulate the setting change works as following;
- nsIDOMWindowUtils.setPrefersReducedMotion() calls an IPC function which ends
up calling nsChildView::SetPrefersReducedMotion() in the parent process
- nsChildView::SetPrefersReducedMotion() sets the given value into
nsLookAndFeel::mPrefersReducedMotionCached just like we set the value queried
via NSWorkspace.accessibilityDisplayShouldReduceMotion in the parent process
and send a notification which is the same notification MacOSX sends when the
system setting changed
- Normally the cached value is cleared before quering new values since the
cache value is stale, but in this case the value is up-to-date one, so
nsChildView::SetPrefersReducedMotion() tells that we don't need to clear the
cache, and nsIDOMWindowUtils.resetPrefersReducedMotion() resets that state
of 'we don't need to clear the cache'
There are two test cases with the framework in this commit, one is just setting
the value and checking the value queried by window.matchMedia. The other one is
receiving 'change' event and checking the value of the event target.
Note that to make this test works the patch for bug 1478212 is necessary since
the test runs in an iframe.
Depends on D5003
Differential Revision: https://phabricator.services.mozilla.com/D5004
--HG--
extra : moz-landing-system : lando
There's also a field named mState on IProtocol, and this reduces confusion.
Differential Revision: https://phabricator.services.mozilla.com/D4356
--HG--
extra : moz-landing-system : lando
The test case in this patch fails without the proper fix in the first patch
in this patch series.
In this patch two new nsIDOMWindowUtils APIs are introduced to change the
system font settins in tests. Currently the APIs work only on GTK+ platform.
Also to work the test case properly we need to open a new XUL window because we
don't propagate font changes into descendant documents yet (bug 1478212).
MozReview-Commit-ID: 4OLxEkEuF8d
--HG--
extra : rebase_source : 683e64f07c4d8820e5499d8c15b90975618559b8
On the compositor we store animation values in a hash table and the hash is
the compositor animation id which is a unique id for each property respectively.
So we can get the corresponding animation value for the given property.
In this patch there are lots of duplicated code, but they will be removed in the
next patch.
MozReview-Commit-ID: 7EboVcculcg
--HG--
extra : rebase_source : 304ea80849af8af72a07437736041aeabbe47eeb
On the compositor we store animation values in a hash table and the hash is
the compositor animation id which is a unique id for each property respectively.
So we can get the corresponding animation value for the given property.
In this patch there are lots of duplicated code, but they will be removed in the
next patch.
MozReview-Commit-ID: 7EboVcculcg
--HG--
extra : rebase_source : 304ea80849af8af72a07437736041aeabbe47eeb
Currently, in order to retrieve supported clipboard formats
DataTransfer::CacheExternalClipboardFormats repeatedly makes the same calls to
clipboard->HasDataMatchingFlavors.
In the case when aPlainTextOnly == true only 1 call is made -
clipboard->HasDataMatchingFlavors(kUnicodeMime, ...), and when
aPlainTextOnly == false we have 1 call made for every member of the list
{ kCustomTypesMime, kFileMime, kHTMLMime, kRTFMime, kURLMime, kURLDataMime,
kUnicodeMime, kPNGImageMime } - a total of 8 calls.
We can see that in nsClipboardProxy::HasDataMatchingFlavors, there is a call to
ContentChild::GetSingleton()->SendClipboardHasType.
So when aPlainTextOnly == true, we will have 1 sync message, and when
aPlainTextOnly == false, we will have 8 sync messages.
With the proposed solution, in DataTransfer::CacheExternalClipboardFormats
we will only have 1 sync message regardless of the case because
GetExternalClipboardFormats() will retrieve all supported clipboard
formats at once.
MozReview-Commit-ID: CAmBfqB459v
--HG--
extra : rebase_source : 27f1b420f2613e6a747ed63762f1583ab71ba3e0
This function returns the transform value modified by both OMTA and APZC.
Note that the transform conversion code is almost the same as the code dropped
in https://hg.mozilla.org/mozilla-central/rev/415811f3804f .
MozReview-Commit-ID: HmsMQp3O4n4
--HG--
extra : rebase_source : ac3994359d646dedaa5ff2f664b20787be8a75f6
This was done automatically replacing:
s/mozilla::Move/std::move/
s/ Move(/ std::move(/
s/(Move(/(std::move(/
Removing the 'using mozilla::Move;' lines.
And then with a few manual fixups, see the bug for the split series..
MozReview-Commit-ID: Jxze3adipUh
This defines three flushing functions that flush different parts of the
WR pipeline. Using all three guarantees that everything sent to WR will
have been flushed. This is what we need to do in SyncWithCompositor to
ensure that it meets the API contract.
In addition, this patch updates the existing FlushRendering function to
use the new functions (no functional changes intended here).
MozReview-Commit-ID: GzxwpF4JT04
--HG--
extra : rebase_source : 1a8cf434d1280902906da257ae63751da7ffd114
Instead of crashing the process inside Transition on a bad state transition,
propagate an error up the stack (and crash higher up).
MozReview-Commit-ID: JJmAeq6xSfe
--HG--
extra : rebase_source : caba8a53603f7b0ad841b657aa3d00e827e3c68c
This commit removes some mtime checking logic in ipdl.py that is made redundant
with make's logic to re-invoke ipdl.py when dependencies have changed.
MozReview-Commit-ID: FJuYIZv5uym
--HG--
extra : rebase_source : 5e250cd2dd57b8591db700a3d79ec36c88fa4dd4
This commit removes some mtime checking logic in ipdl.py that is made redundant
with make's logic to re-invoke ipdl.py when dependencies have changed.
MozReview-Commit-ID: FJuYIZv5uym
--HG--
extra : rebase_source : 9e11ead04bd69eb1f30160fdb15b8ac74b183f87
ByteBuf is a new IPDL built in type, so I wrote a very basic test for
it, based on shmem.ipdl. It was added in bug 1379680.
MozReview-Commit-ID: 4tbnljpUqCh
--HG--
extra : rebase_source : 6f423d7d9cf132aba4498be96548684e551b8e2c
Bug 1443954 added some new syntax to using, but I noticed that there
was not very much existing test coverage, so I wrote a new test that
covers all of the possible cases.
MozReview-Commit-ID: JRgHCtXHDLZ
--HG--
extra : rebase_source : caacc05ae5b1de0841f73e5fdba4891bfc0ee4bf
Failing IPDL parser tests require that the error message is specified.
MozReview-Commit-ID: IGNTVAb5r0Q
--HG--
extra : rebase_source : 54e75ae8820f4e9ae2083b1f5382ff4151c8d59e
The IPDL parser now expects a valid message-metadata to be passed in,
even though we don't need it for parser tests. Fix this by creating
and specifying a blank one. This fixes a regression from bug 1348591.
MozReview-Commit-ID: HJ34mqBdUyP
--HG--
extra : rebase_source : d444e5346f40399f1f1cf71b3d86b6dc15a4b731
The ipc/ipdl/test/cxx/ subdirectory is only built when MOZ_IPDL_TESTS
is set, because it needs special code to be built into the
binary. ipdl/ should always be built, because it does not need special
support. This gating is already done in the ipc/ipdl/test/ directory,
so it shouldn't be done in ipc/ipdl/, because this stops us from
building it in a regular build.
MozReview-Commit-ID: Jpp0CeXEenR
--HG--
extra : rebase_source : 3dbb6a5c9b7f6ec8cbe9840bff240a63342ecd49
We can move this information into ProtocolState and save having two
virtual functions for every protocol. Moving some bits out of the
codegen'd IPC code is a nice bonus, though we keep the strange setup
where toplevel protocols have two mChannel member variables.
ProtocolName() is only used for producing error messages and annotating
crash reports. But examining actual crash reports that would have used
the result of ProtocolName() indicates that we can always tell what the
erroring protocol is due to the stack backtrace. So having this virtual
function around just provides duplicate information, and it takes up too
much space in the vtable besides. Let's get rid of it.
lower.py generates repetitious:
SetManager(...);
Register(...); // Or RegisterID.
SetIPCChannel(...);
calls, which are moderately sized, given that the above call sequence
requires virtual calls in several places. Instead of codegenning this
sequence, let's consolidate the sequence into IProtocol and change the
code generator to call into the consolidated function instead.
This patch was reviewed in parts, however the intermediate states would not build:
Bug 1443954 - Part 3A: Strip pointers from the argument to WriteParam and WriteIPDLParam before selecting the ParamTraits impl, r=froydnj
Bug 1443954 - Part 3B: Move nsIAlertNotification serialization to the refcounted system, r=bz
Bug 1443954 - Part 3C: Move geolocation serialization to the refcounted system, r=bz
Bug 1443954 - Part 3D: Move nsIInputStream serialization to the refcounted system, r=baku
Bug 1443954 - Part 3E: Move BlobImpl serialization to the refcounted system, r=baku
Bug 1443954 - Part 3F: Correctly implement ParamTraits for actors after the ParamTraits changes, r=froydnj
This remotes the APZInputBridge interface over the PAPZInputBridge
protocol in the case of the GPU process, and makes the GPU process'
main thread act as the APZ controller thread in that process. If
there is no GPU process we continue as before and the APZInputBridge
interface implementation is the concrete APZCTreeManager instance
in the UI process.
The main changes in this patch are moving all the code associated with
these messages out of APZCTreeManager{Parent,Child} and into
APZInputBridge{Parent,Child}. APZCTreeManagerChild now returns an
APZInputBridgeChild instance via InputBridge(), instead of returning
itself. The SetControllerThread call in the GPU process is also updated.
MozReview-Commit-ID: M4AaIW1Q0h
--HG--
extra : rebase_source : e5a8f14e23be34229fe80a47f6789d19b19e0a9f
This code is unused now that ReadLockDescriptors are not sent in layer transactions.
--HG--
extra : rebase_source : 8cd25541b22c3151e2dbd2f8ea6d1119e2f26c94
extra : source : 99a2d26d1ba82ad34a6c27641500a424cda015c3
This is important as these includes are often used to define ParamTraits
implementations which are needed to serialize struct members and
parameters. The place where the code which uses these implementations
will be moving from the protocol where they are used to the file where
the structs are defined, which means we need to includes these files
there.
MozReview-Commit-ID: H1wqgyv5mel
The FunctionBroker actors allow the NPAPI process (child) to run methods on the main process (parent). Both the parent and the child run dedicated threads for this task -- this is a top-level protocol.
inline is never set so cgen never emits inline.
MozReview-Commit-ID: BDL6BV8906t
--HG--
extra : rebase_source : aa51cda34db36ba08622d9940402cea14617c74a
never_inline is never set so cgen never emits MOZ_NEVER_INLINE.
MozReview-Commit-ID: HS7qw7D4lBC
--HG--
extra : rebase_source : 2eae9e3a7a0c95d94986283fabe670d5decec3c3
pure or override implies virtual. static is mutually exclusive with virtual (and pure and override). Combining these types into a "method specifier" enum simplifies the code and prevents bogus states like virtual=0 and pure=1.
MozReview-Commit-ID: IFeuvbp1RIo
--HG--
extra : rebase_source : d423d168a8b3a3d60cabd0f5e58fd03c6f655baf
This removes an unnecessary level of indirection by replacing all
nsStringGlue.h instances with just nsString.h.
--HG--
extra : rebase_source : 340989240af4018f3ebfd92826ae11b0cb46d019
Also ensure we're more consistent about having two blank lines after these methods.
MozReview-Commit-ID: 2SXXjDuMW4u
--HG--
extra : rebase_source : 05279ceab3aa32d1e577ea02ccd113180bb55792
There's already a typedef for MessageChannel, so use that instead of
Channel. Also, use IProtocol and not ProtocolBase or ChannelListener,
for simplicity.
MozReview-Commit-ID: 2zCjTpPTW4L
--HG--
extra : rebase_source : eed8d6d51a512ac85f6b4962fc12591bf81c2981
This method is now trivial, so we can inline a few things. This
removes the only callers of sendSems(), so remove that, too.
MozReview-Commit-ID: Bb4EF9M56ut
--HG--
extra : rebase_source : dedf31aa225361e31122362aefc76940edbf25f3
Currently if you write an async IPDL method which has a return value, we expose
a SendXXX method which returns a MozPromise. This MozPromise can then be
->Then-ed to run code when it is resolved or rejected.
Unfortunately, using this API loses ordering guarantees which IPDL provides.
MozPromise::Then takes an event target, which the resolve runnable is dispatched
to. This means that the resolve callback's code doesn't have any ordering
guarantees relative to the processing of other IPC messages coming over the same
protocol.
This adds a new overload to SendXXX with two additional arguments, a lambda
callback which is called if the call succeeds, and a lambda callback which is
called if the call fails. These will be called in order with other IPC messages
sent over the same protocol.
MozReview-Commit-ID: FZHJJaSDoZy
And remove unreachable code after MOZ_CRASH().
MozReview-Commit-ID: 6ShBtPRKYlF
--HG--
extra : rebase_source : 0fe45a59411bda663828336e2686707b550144ae
extra : source : 8473fd7333d2abe1ea1cc176510c292a5b34df45
Currently the Gecko Profiler defines a moderate amount of stuff when
MOZ_GECKO_PROFILER is undefined. It also #includes various headers, including
JS ones. This is making it difficult to separate Gecko's media stack for
inclusion in Servo.
This patch greatly simplifies how things are exposed. The starting point is:
- GeckoProfiler.h can be #included unconditionally;
- everything else from the profiler must be guarded by MOZ_GECKO_PROFILER.
In practice this introduces way too many #ifdefs, so the patch loosens it by
adding no-op macros for a number of the most common operations.
The net result is that #ifdefs and macros are used a bit more, but almost
nothing is exposed in non-MOZ_GECKO_PROFILER builds (including
ProfilerMarkerPayload.h and GeckoProfiler.h), and understanding what is exposed
is much simpler than before.
Note also that in BHR, ThreadStackHelper is now entirely absent in
non-MOZ_GECKO_PROFILER builds.
Never store names in Message. One can get string names from
Message::name() or use IPC::StringFromIPCMessageType() when only
message id is available.
MozReview-Commit-ID: 15ksx6SE90c
--HG--
extra : rebase_source : 1a041dc365b7f42edd540d8c7a4dfd8912e48921
Each protocol in IPDL has a bunch of autogenerated functions that
instantiate IPC::Message with various parameters. Each of these
functions, then:
1) Pays the cost of calling malloc()
2) Setting up various parameters
3) Calling IPC::Message()
There's no reason that we should be duplicating 1) across all of these
autogenerated functions. In step 2), several of the parameters we're
setting up are common across all or nearly all calls: the message
segment size is almost always zero, and we're always indicating that
IPDL-generated messages should be recorded in telemetry.
Instead of duplicating that code several thousand times, we can add a
small helper function that takes the only interesting parameters for an
IPDL message. This helper function can then deal with calling malloc in
a single place and setting up the common parameters. For messages that
require a custom segment size, we'll have to use the old scheme, but
such messages are uncommon.
The previous changes are not required for this scheme to work, but they
do help significantly, as the helper function (Message::IPDLMessage) can
now take four parameters, which ensures that its arguments are passed
solely in registers on Win64 and ARM. The wins from this change are
also larger than they would be without the previous parts: ~100K on
x86-64 Linux (!) and ~80K on ARM Android.
The current IPC::Message constructor takes a large number of arguments,
three of which--the nesting level, the priority, and the
compression--are almost always constant by virtue of the vast majority
of Message construction being done by auto-generated IPDL code. But
then we take these constant values into the Message constructor, we
check them for various values, and then based on those values, we
perform a bunch of bitfield operations to store flags based on those
values. This is wasted work.
Furthermore, for replies to IPDL messages, we'll construct a Message
object, and then call mutating setters on the Message object that will
perform even more bitfield manipulations. Again, these operations are
performing tasks at runtime that are the same every single time, and use
information we already have at compile time.
The impact of these extra operations is not large, maybe 15-30K of extra
code, depending on platform. Nonetheless, we can easily make them go
away, and make everything cleaner to boot.
This patch adds a HeaderFlags class that encapsulates all the knowledge
about the various kinds of flags Message needs to know about. We can
construct HeaderFlags objects with strongly-typed enum arguments for the
various kinds of flags, and the compiler can take care of folding all of
those flags together into a constant when possible (and it is possible
for all the IPDL-generated code that instantiates Messages). The upshot
is that we do no unnecessary work in the Message constructor itself. We
can also remove various mutating operations on Message, as those
operations were only there to support post-constructor flag twiddling,
which is no longer necessary.
There's no need to be repeating 'IPC::Message::' prefixes or spreading
around more ExprVar calls than we need here. Let's try to improve the
signal-to-noise ratio of this code by introducing a helper function to
inject some of the boilerplate for us.
_generateMessageConstructor takes a lot of `md.FOO`-style parameters,
which could be derived inside the function by simply passing `md`.
Especially with the upcoming changes to calculate things like reply-ness
of messages, sync-ness, etc, we'd be wanting to pass even more
parameters like `md.FOO`. So let's just pass `md` in, and then we can
make all the necessary future changes in a single place.
The nsIU2FToken and its implementors are no longer needed; the soft token was
re-implemented into dom/webauthn/U2FSoftTokenManager.cpp during the WebAuthn
implementation. When the dom/u2f/ code changed to the implementation from
WebAuthn, the old synchronous version became dead code.
This patch removes the dead code.
MozReview-Commit-ID: 2yDD0tccgZr
--HG--
extra : rebase_source : 0f14d8de8f62599a41c13aa4d8fc9cdbc1fd79c7
The nsIU2FToken and its implementors are no longer needed; the soft token was
re-implemented into dom/webauthn/U2FSoftTokenManager.cpp during the WebAuthn
implementation. When the dom/u2f/ code changed to the implementation from
WebAuthn, the old synchronous version became dead code.
This patch removes the dead code.
MozReview-Commit-ID: 2yDD0tccgZr
--HG--
extra : transplant_source : %B3%96Te%E7%02%08%98%1A%B2%FA%1C%40%C4J%BC%B2%85j%81
Certain types (such as Shmem and Endpoint types) cannot be copied, and need to
be moved when passed around. When used with MozPromises, that means that the
promise needs to be non-shareable, and the resolve functions need to use the
correct ref qualifiers.
MozReview-Commit-ID: Kt4WZNsDErK
--HG--
extra : rebase_source : 5a96f9844df1646482aa223edf5081de9d5fc976
Telemetry and some performance profiles show that Msg_NotifyIMEFocus can take
a few seconds to complete, and jank the browser. With bug 1217700, it removes
the necessity of sync Msg_NotifyIMEFocus, so in this patch we make this async
for performance improvement.
MozReview-Commit-ID: 15eUwMJ2Q7H
--HG--
extra : rebase_source : b463e6e881ca5ebec00d0f76e29ca103059b3ddd
We only ever use these for passing them into Endpoint construction.
Let's remove them. Removing them also shows that the corresponding
field in Endpoint is essentially read-only, so we can completely avoid
passing in protocol IDs to Endpoint.
We removed all uses of GetBlocklistState in bug 1350640. This patch
removes the message and supporting functions from the PContent IPDL.
MozReview-Commit-ID: 4JtGAWZ0nPu
--HG--
extra : rebase_source : 4eb3c21e3768e9d8284d4eec129e099be5ef17d0
This protocol was changed in bug 1351148, but the whitelist was not
updated.
MozReview-Commit-ID: Btl5633et9T
--HG--
extra : rebase_source : 0ce70a649c83213a8a47e1b08c06d7d5743a9842
Since the default size is 64, we only care about message size which is > 64 bytes.
MozReview-Commit-ID: 2vUpcaUjlNP
--HG--
extra : rebase_source : 946ae82f8fc2febd05d8fc4323145643ec97b306
This commit ties it all together by dispatching keyboard actions to scroll targets
in response to keyboard inputs when we have current and valid focus state.
MozReview-Commit-ID: G7rZiS3FH5e
--HG--
extra : rebase_source : 10129d417fe8ef576cac5bda3157dd8f65b5843a
extra : histedit_source : be651a33f787f68bc764988ddc073d346e854491
Focus can change at any moment in a document. This causes non-determinism and
correctness problems for doing keyboard apz scrolling. To get around this, we
will maintain deterministic behavior for focus changes initiated by input events
and see if we can get away with more non-determinism for things like `setTimeout`
In order to do this, we disable async keyboard scrolling when an input event is
processed that could have a event listener. We then attach a sequence number to
that input event and dispatch it to content. In content, we record the highest
sequence number that we have processed from an event, and send that on each focus
update. Using this, we can determine in APZ if we have a current focus target or
if we are still waiting for an input event to be processed and focus to be
reconfirmed.
MozReview-Commit-ID: CWcu8YEFQz4
--HG--
extra : rebase_source : 8c54a619bd4f5ee892f0cc8768a10f3e1e4e0b59
extra : histedit_source : 601ca293a028787883841adc6b40e62c0cc829e5
This patch makes the following changes to the macros.
- Removes PROFILER_LABEL_FUNC. It's only suitable for use in functions outside
classes, due to PROFILER_FUNCTION_NAME not getting class names, and it was
mostly misused.
- Removes PROFILER_FUNCTION_NAME. It's no longer used, and __func__ is
universally available now anyway.
- Combines the first two string literal arguments of PROFILER_LABEL and
PROFILER_LABEL_DYNAMIC into a single argument. There was no good reason for
them to be separate, and it forced a '::' in the label, which isn't always
appropriate. Also, the meaning of the "name_space" argument was interpreted
in an interesting variety of ways.
- Adds an "AUTO_" prefix to PROFILER_LABEL and PROFILER_LABEL_DYNAMIC, to make
it clearer they construct RAII objects rather than just being function calls.
(I myself have screwed up the scoping because of this in the past.)
- Fills in the 'js::ProfileEntry::Category::' qualifier within the macro, so
the caller doesn't need to. This makes a *lot* more of the uses fit onto a
single line.
The patch also makes the following changes to the macro uses (beyond those
required by the changes described above).
- Fixes a bunch of labels that had gotten out of sync with the name of the
class and/or function that encloses them.
- Removes a useless PROFILER_LABEL use within a trivial scope in
EventStateManager::DispatchMouseOrPointerEvent(). It clearly wasn't serving
any useful purpose. It also serves as extra evidence that the AUTO_ prefix is
a good idea.
- Tweaks DecodePool::SyncRunIf{Preferred,Possible} so that the labelling is
done within them, instead of at their callsites, because that's a more
standard way of doing things.
--HG--
extra : rebase_source : 318d1bc6fc1425a94aacbf489dd46e4f83211de4
