I don't think all this complexity is worth it for having a
marginally-more-realistic testing story. Using the pref just works and we should
do that, I think.
Differential Revision: https://phabricator.services.mozilla.com/D59980
This method always copies, and is redundant with the other getter methods on
IPDL unions. As there is only one caller, it can be removed to simplify the
code, and remove a source of complexity.
Differential Revision: https://phabricator.services.mozilla.com/D75349
Fix intermittent issues due to races.
We now run the MozPromise generated by the IPDL bindings to run their callbacks via a direct task dispatch.
This avoids a full trip to the back of the event queue for each additional asynchronous step when using MozPromise.
A consequence to this change is that each IPDL actor's thread must have an AbstractThread allocated if IPDL MozPromises are used.
It prevents unexpected racy behaviours when combining MozPromise with the other Resolve/Reject IPDL async declaration which was have lead to processing the events out of order.
Differential Revision: https://phabricator.services.mozilla.com/D71593
The hashes of certain strings is used indirectly in generating IDL C++ files. Before, we were using the `hash()` function to generate these hashes, which in Python 3 is non-deterministic over subsequent `python3` processes, causing bugs like bug 1635755. Instead, use a specific, deterministic hash to avoid spurious diff failures.
The `md5` hash function isn't completely cryptographically secure but the security of the hashes isn't important for us in this case since we're just using them as a per-string identifier. We could use a more robust hash function but there may be performance implications from doing so.
Differential Revision: https://phabricator.services.mozilla.com/D74446
`ply`, [by design](https://github.com/dabeaz/ply/issues/79), does not produce reproducible table files; hence bug 1633156. (Note that this was *always* true, but only became a problem once we switched to Python 3, which has more unpredictable dict iteration order than Python 2.7, at least prior to [3.7](https://docs.python.org/3/whatsnew/3.7.html#summary-release-highlights).)
In any other circumstance I would consider submitting a patch to `ply` to fix this, but as of the [in-progress version 4.0 of the library](https://github.com/dabeaz/ply/blob/master/CHANGES), it doesn't even emit this cached data any more, and indeed the [latest version of the code](1fac9fed64/ply) doesn't even call `open()` at all except to do logging or to read the text data to be parsed from `stdin`. So if we were going to pin our future on `ply` and upgrade to later versions of the library in the future, we would have to live in a world where `ply` doesn't generate cached table files for us anyway.
Emitting the cached table files so later build steps can consume them is an "optimization", but it's not clear exactly how much actual value that optimization provides overall. Quoth the `CHANGES` file from that repository:
```
PLY no longer writes cached table files. Honestly, the use of
the cached files made more sense when I was developing PLY on
my 200Mhz PC in 2001. It's not as much as an issue now. For small
to medium sized grammars, PLY should be almost instantaneous.
```
In practice, I have found this to be true; namely, `./mach build pre-export export` takes just about as long on my machine after this patch as it did before, and in a try push I performed, there's no noticeable performance regression from applying this patch. In local testing I also found that generating the LALR tables in calls to `yacc()` takes about 0.01s on my machine generally, and we generate these tables a couple dozen times total over the course of the `export` tier now. This isn't *nothing*, but in my opinion it's also not nearly long enough where it would be a concern given how long `export` already takes.
That `CHANGES` file also stresses that if caching this data is important, we have the option of doing so via `pickle`. If and when we decide that re-enabling this optimization is valuable for us, we should take control of this process and perform the generation in such a way that we can guarantee reproducibility.
Differential Revision: https://phabricator.services.mozilla.com/D73484
Adds IpdlQueue capability to PWebGL actors. The WebGLChild, used in content processes, implements SyncProducerActor and AsyncConsumerActor because it sends (sync and async) messages and receives responses to them that it reads as async messages. The WebGLParent, used in the compositor process, is a SyncConsumerActor and AsyncProducerActor for dual reasons.
Differential Revision: https://phabricator.services.mozilla.com/D68264
We need to separate WebGL actor construction and initialization since IpdlQueue initialization needs the actor to already exist.
Differential Revision: https://phabricator.services.mozilla.com/D68262
Adds IpdlQueue capability to PWebGL actors. The WebGLChild, used in content processes, implements SyncProducerActor and AsyncConsumerActor because it sends (sync and async) messages and receives responses to them that it reads as async messages. The WebGLParent, used in the compositor process, is a SyncConsumerActor and AsyncProducerActor for dual reasons.
Differential Revision: https://phabricator.services.mozilla.com/D68264
We need to separate WebGL actor construction and initialization since IpdlQueue initialization needs the actor to already exist.
Differential Revision: https://phabricator.services.mozilla.com/D68262
This commit:
- removes sendRpcMessage, which was unused;
- removes the CPOW argument to sendAsyncMessage, broadcastAsyncMessage, and
sendSyncMessage;
- removes the aIsSync argument used internally to distinguish sendRpcMessage
and sendSyncMessage;
- removes CPOW tests;
- updates the few remaining callsites that use more than 2 arguments in
sendAsyncMessage for the removal of the cpows argument.
Differential Revision: https://phabricator.services.mozilla.com/D71514
Before 1496578, URIFixup::keywordToURI used to do a synchronous IPC call to be
able to access search engines from the content process. Consumers of URIFixup
didn't care. Bug 1496578 moved the IPC messaging to the callers, in particular
nsDocShell, but assumed nsDocShellLoadState wasn't loading from content.
It looks like in some cases it does, so this adds another sync IPC call for
GetFixupURIInfo.
The total numer of sync IPCs should not change from before Bug 1496578, URIFIxup
was just doing it internally, while now it happens at the call point.
Note the long term plan would be for these docshell objects callers to just
handle URIs, while the UI code should do fixup.
Bug 1375244 tracks the removal of these sync IPC messages.
Differential Revision: https://phabricator.services.mozilla.com/D70607
--HG--
extra : moz-landing-system : lando
For Win32k lockdown, we need to remove the content processes' ability to
call GetICMProfileW(). Since it needs this to retrieve the output color
profile, a new synchronous call is added that allows it to request the
parent process to read this file on its behalf.
The contents of the file are now being cached as well, as this should help
ease some of the increased parent process I/O caused by the children not
being able to do this in their process anymore.
For performance reasons, during launch this information is passed directly
to the child through the SetXPCOMProcessAttributes call
Differential Revision: https://phabricator.services.mozilla.com/D66126
--HG--
extra : moz-landing-system : lando
This wasn't useful cross-platform.
ATK was the only consumer of this and it now uses ProxyAccessible::ChildAtPoint.
This also means the related aNeedsScreenCoords functionality in PDocAccessible::AccessibleAtPoint is no longer needed and has thus been removed.
Finally, this renames PDocAccessible::AccessibleAtPoint to PDocAccessible::ChildAtPoint for consistency with Accessible::ChildAtPoint now that the functionality is mirrored.
Differential Revision: https://phabricator.services.mozilla.com/D67987
--HG--
extra : moz-landing-system : lando
For Win32k lockdown, we need to remove the content processes' ability to
call GetICMProfileW(). Since it needs this to retrieve the output color
profile, a new synchronous call is added that allows it to request the
parent process to read this file on its behalf.
The contents of the file are now being cached as well, as this should help
ease some of the increased parent process I/O caused by the children not
being able to do this in their process anymore.
For performance reasons, during launch this information is passed directly
to the child through the SetXPCOMProcessAttributes call
Differential Revision: https://phabricator.services.mozilla.com/D66126
--HG--
extra : moz-landing-system : lando
Before this patch, PSM would register content type handlers to handle certain
certificate types. This was an easy way to install a client certificate after
generating a key with <keygen>, but keygen has been removed. This was also an
easy way to install root certificates, but that's actually a considerable
security risk. We kept this functionality for so long because it was the only
way to add a 3rd-party root certificate to Fennec's certificate store. Now that
Fennec is EOL, we can remove it. (Fenix will need a way to trust 3rd party root
certificates, but the path forward there is to implement the enterprise roots
feature for Android.)
Differential Revision: https://phabricator.services.mozilla.com/D66118
--HG--
extra : moz-landing-system : lando
For Win32k lockdown, we need to remove the content processes' ability to
call GetICMProfileW(). Since it needs this to retrieve the output color
profile, a new synchronous call is added that allows it to request the
parent process to read this file on its behalf.
The contents of the file are now being cached as well, as this should help
ease some of the increased parent process I/O caused by the children not
being able to do this in their process anymore.
Differential Revision: https://phabricator.services.mozilla.com/D66126
--HG--
extra : moz-landing-system : lando
We were using the wrong Accessible method to get the level. We also need to swap IPDL methods to use the right one.
Differential Revision: https://phabricator.services.mozilla.com/D65645
--HG--
extra : moz-landing-system : lando
WalkHistoryEntries function gets called by nsSHistory::CloneAndReplaceChild
and nsSHistory::SetChildHistoryEntry recursively, so those have to be moved
into the parent process. This eliminates many sync IPC calls.
To facilitate transition to a new session history design,
we are mirroring mOSHE and mLSHE SH entries from docshell to browsing context.
Whenever we update those entries in docshell, we will also update those in BC,
and vice versa.
Differential Revision: https://phabricator.services.mozilla.com/D56201
--HG--
extra : moz-landing-system : lando
We want to be able to send IPC messages from the translation in the parent. So
the simplest thing it move the top level actor parts of CanvasParent into
CanvasTranslator.
This patch also moves the canvas thread management parts out into a new
CanvasThreadHolder class and hopefully makes the lifecycle management of these
much more robust. This includes the use of a TaskQueue per CanvasTranslator to
manage serial processing on the canvas workers, instead of a boolean.
Differential Revision: https://phabricator.services.mozilla.com/D60887
--HG--
rename : gfx/layers/ipc/CanvasParent.cpp => gfx/layers/ipc/CanvasThread.cpp
rename : gfx/layers/ipc/CanvasParent.h => gfx/layers/ipc/CanvasThread.h
rename : gfx/layers/CanvasTranslator.cpp => gfx/layers/ipc/CanvasTranslator.cpp
rename : gfx/layers/CanvasTranslator.h => gfx/layers/ipc/CanvasTranslator.h
extra : moz-landing-system : lando
This removes the need for explicit #ifdef NS_BUILD_REFCNT_LOGGING without
introducing user-defined destructors when it is not defined.
Also, some uses of virtual for declaring destructors are replaced by the
appropriate override declaration through these changes.
Differential Revision: https://phabricator.services.mozilla.com/D62604
--HG--
extra : moz-landing-system : lando
This removes the need for explicit #ifdef NS_BUILD_REFCNT_LOGGING without
introducing user-defined destructors when it is not defined.
Also, some uses of virtual for declaring destructors are replaced by the
appropriate override declaration through these changes.
Differential Revision: https://phabricator.services.mozilla.com/D62604
--HG--
extra : moz-landing-system : lando
The default method implementations cause problems when trying to
override them with different types in a direct call class.
For the `Recv__delete__` case there's a simple solution: omit it if
there are any arguments, because it doesn't make much sense to specify
arguments and then completely ignore them, and the no-arg case isn't a
problem for overriding.
Differential Revision: https://phabricator.services.mozilla.com/D62977
--HG--
extra : moz-landing-system : lando
This change adds new "remote backbuffer" logic when compositing without
HW acceleration on Windows (IE compositing through Cairo using the Win32
GDI)
A new piece of shared memory is created between the GPU process and the UI
process, and the GPU process sends requests to the UI process to first "borrow"
a properly-sized buffer to draw into, and then sends a "present" request to
tell the UI process to actually blit the buffer to the Win32 window.
This is needed for the GPU sandbox to work, since Windows rightly doesn't
allow the untrusted GPU process to directly draw the contents of a window
owned by the trusted UI process.
Differential Revision: https://phabricator.services.mozilla.com/D61370
--HG--
extra : moz-landing-system : lando
This change adds new "remote backbuffer" logic when compositing without
HW acceleration on Windows (IE compositing through Cairo using the Win32
GDI)
A new piece of shared memory is created between the GPU process and the UI
process, and the GPU process sends requests to the UI process to first "borrow"
a properly-sized buffer to draw into, and then sends a "present" request to
tell the UI process to actually blit the buffer to the Win32 window.
This is needed for the GPU sandbox to work, since Windows rightly doesn't
allow the untrusted GPU process to directly draw the contents of a window
owned by the trusted UI process.
Differential Revision: https://phabricator.services.mozilla.com/D61370
--HG--
extra : moz-landing-system : lando
Replace all implicit copies by moves, or explicit clone operations.
Fixed a bug in /home/simon/work/ipdl/ipc/ipdl/ipdl/lower.py _cxxConstRefType,
which caused the const accessor to miss const on its return type for a moveonly
struct member.
Removed unused using directives for SerializedStructuredCloneBuffer in ipdl files.
Differential Revision: https://phabricator.services.mozilla.com/D59742
--HG--
extra : moz-landing-system : lando
Replace all implicit copies by moves, or explicit clone operations.
Fixed a bug in /home/simon/work/ipdl/ipc/ipdl/ipdl/lower.py _cxxConstRefType,
which caused the const accessor to miss const on its return type for a moveonly
struct member.
Removed unused using directives for SerializedStructuredCloneBuffer in ipdl files.
Differential Revision: https://phabricator.services.mozilla.com/D59742
--HG--
extra : moz-landing-system : lando
Splits WebGLContext into ClientWebGLContext and HostWebGLContext. The Client enables the JS-control of a WebGL context in a content procecss while the Host executes the WebGL graphics operations (via a WebGLContext that maintains much of the existing code) in the compositor process. At this point, the cross-process behavior is disabled -- this series of patches is an incremental step toward that final goal.
Differential Revision: https://phabricator.services.mozilla.com/D54018
--HG--
extra : moz-landing-system : lando
The inclusions were removed with the following very crude script and the
resulting breakage was fixed up by hand. The manual fixups did either
revert the changes done by the script, replace a generic header with a more
specific one or replace a header with a forward declaration.
find . -name "*.idl" | grep -v web-platform | grep -v third_party | while read path; do
interfaces=$(grep "^\(class\|interface\).*:.*" "$path" | cut -d' ' -f2)
if [ -n "$interfaces" ]; then
if [[ "$interfaces" == *$'\n'* ]]; then
regexp="\("
for i in $interfaces; do regexp="$regexp$i\|"; done
regexp="${regexp%%\\\|}\)"
else
regexp="$interfaces"
fi
interface=$(basename "$path")
rg -l "#include.*${interface%%.idl}.h" . | while read path2; do
hits=$(grep -v "#include.*${interface%%.idl}.h" "$path2" | grep -c "$regexp" )
if [ $hits -eq 0 ]; then
echo "Removing ${interface} from ${path2}"
grep -v "#include.*${interface%%.idl}.h" "$path2" > "$path2".tmp
mv -f "$path2".tmp "$path2"
fi
done
fi
done
Differential Revision: https://phabricator.services.mozilla.com/D55443
--HG--
extra : moz-landing-system : lando
Previously, we created TextureD3D11 objects in the content process to back surfaces created for the plugin process. Those objects were then composited by the async ImageBridge. In order to remove Win32 kernel operations from content (including DX/GDI operations), this patch bounces the requests from content to the compositor process. The compositor process maintains 2 textures to be used for all plugin composition -- one for the plugin process and one for display. The plugin process can freely write to its texture and request composition when it is done, which triggers a blit to the display texture. This mirrors pre-existing behavior.
Differential Revision: https://phabricator.services.mozilla.com/D46086
--HG--
extra : moz-landing-system : lando
These operations report whether certain async plugin drawing modes are supported on the host architecture. They use kernel graphics operations to decide this so they need to be removed from the content process for sandboxing. We just bounce the requests to the gpu process (or main process on systems without a GPU process).
Differential Revision: https://phabricator.services.mozilla.com/D46085
--HG--
extra : moz-landing-system : lando
`windowUtils.setCompositionRecording()` now returns a promise that is resolved
when the composition recorder is enabled (if given `true`) or when frames are
written to disk (if given `false`). To accomplish this, the
`WebRenderCompositionRecorder` now returns a `MozPromise` when writing frames
to disk begins that is resolved when that process finishes.
Differential Revision: https://phabricator.services.mozilla.com/D47300
--HG--
extra : moz-landing-system : lando
`windowUtils.setCompositionRecording()` now returns a promise that is resolved
when the composition recorder is enabled (if given `true`) or when frames are
written to disk (if given `false`). To accomplish this, the
`WebRenderCompositionRecorder` now returns a `MozPromise` when writing frames
to disk begins that is resolved when that process finishes.
Differential Revision: https://phabricator.services.mozilla.com/D47300
--HG--
extra : moz-landing-system : lando
Inside of nsDocShell::UpdateURLAndHistory, there are 4 sync IPC calls to
nsSHistory plus 1 static call, which contains at least one nsSHEntry::GetParent
sync IPC call. All of these calls can be moved inside of a new method
EvictContentViewersOrReplaceEntry on nsSHEntry, resulting in just 1 sync IPC
call.
Differential Revision: https://phabricator.services.mozilla.com/D32729
--HG--
extra : rebase_source : ad09a9061cd6fe8eb6796b2809ea191aceb3ac73
extra : source : 2cd5cd24763ff320719aedb2142a79822efd6de4
extra : histedit_source : fdc4f80cfd8807e46c2dc02e6ab82f2bd3acc391
Currently, nsDocShell::GetChildSHEntry calls 3 sync IPC calls on nsISHEntry,
and the method GetChildSHEntry only has one caller. By moving GetChildSHEntry
method to parent process for nsISHEntry, resulting in a new method
nsISHEntry::GetChildSHEntry, 3 sync IPC calls can be reduced to 1.
Differential Revision: https://phabricator.services.mozilla.com/D27633
--HG--
extra : rebase_source : 1e76a1b07d827b35bae7ed6acca25aa8732c9ed0
extra : source : 34d91f82faa36dd8d8ed721de365122d831f7b25
extra : histedit_source : 86592bc747d506c546ec660e57087e78b9719cf7
Inside of nsDocShell::OnNewURI there are 4 sync IPC calls
to nsSHistory that can be replaced with 1 sync IPC call
by adding a new method EnsureCorrectEntryAtCurrIndex to nsSHistory.
Differential Revision: https://phabricator.services.mozilla.com/D31539
--HG--
extra : rebase_source : 09d7738b2f2dc2334c8f6186e5918b9d0ea3e618
extra : source : 82a41bffcbbca24ad3e84b045d75e4cb01ae1445
extra : histedit_source : 96eb7d2dbed2ad8fe4cc2d37358ec2358f696442
In nsDocShell::LoadHistoryEntry method, when it gets called
from parent process, there are 13 sync IPC calls on nsISHEntry
that retrieve information from the session history entry and
create a doc shell load state object using the retrieved
information. By adding a new method 'CreateLoadInfo'on nsISHEntry,
inside of which the doc shell load state object will be
created (with appropriate data filled out) and returned,
we eliminate 12 sync IPC call, resulting in just 1 IPC call
to nsISHEntry::CreateLoadInfo.
Differential Revision: https://phabricator.services.mozilla.com/D26042
--HG--
extra : rebase_source : a4e1fa52932fd5caabb59bd133e9fbee7f4d0e4a
extra : amend_source : f4d9f01afac0337808ba347eb997ce83e6ae1488
extra : source : 6ad53b35c7b4be933a3db1e1d45fa3da8a57abad
extra : histedit_source : c08d0cebcc11a3a4f64d01566cb62d9a334a12ec
Currently, nsDocShell repeatedly calls nsISHEntry::GetChildAt and
nsISHEntry::RemoveChild in nsDocShell::AddToSessionHistory,
which results in twice as many IPC calls as the number of children
a session history entry has. Additionally, there is one extra
IPC sync call to nsISHEntry::AbandonBFCacheEntry after the loop.
With the proposed solution, there will only be 1 sync IPC call.
Differential Revision: https://phabricator.services.mozilla.com/D24978
--HG--
extra : rebase_source : ab98af417bcb28fdfcb1a115473b505fa8a70aa1
extra : source : 0270d6b952ffa1704690c777c55fcaed30125b38
extra : histedit_source : 9a031cb921ad8efa1bd96f7342e08337f3c0ca8a
Currently, nsDocShell repeatedly calls nsISHEntry::GetChildAt, which
results in as many IPC sync calls as the number of children
a session history entry has. Calling nsISHEntry::GetChildCount and
ChildSHistory::Index and incurs additional extra 2 sync IPC calls.
With the proposed solution, there will only be 1 sync IPC call.
Differential Revision: https://phabricator.services.mozilla.com/D24980
--HG--
extra : rebase_source : ffaaff5c9521fb4d3a53a85f4570bcb99f7317d1
extra : source : 3f4a092d8f6544a212ee8a80d9275ae68c6446d1
extra : histedit_source : b4166d997119092ba5657f9ca4216da62dc82d97
This takes the approach of directly tying the `AddRef()` and `Release()` calls
on the underlying actor object to the lifecycle of the ActorLifecycleProxy
object, by adding another virtual `ActorAlloc()` method as the counterpart to
`ActorDealloc()`.
The changes to the methods called from C++ are relatively minimal:
1. The `SendPFooConstructor(...)` methods have the same signature, but now will
AddRef() internally (through `ActorAlloc()`), so an `do_AddRef(foo).take()`
isn't needed. Same with the `{Bind,Open}PFooEndpoint` methods.
This was done to reduce the number of invasive internal changes in the
codegen which were required to pipe a passed-in `already_AddRefed` to the
underlying `ActorLifecycleProxy` which is created by generic code.
2. The `AllocPFoo{Parent,Child}` methods have been modified to return an
`already_AddRefed` by default, and the callsites will store the result in a
`RefPtr<T>` if the type is refcounted.
3. No `DeallocPFoo...` method is called anymore. Unfortunately due to
devirtualization, it won't be an error if an unused method is still present,
though it won't be called.
Differential Revision: https://phabricator.services.mozilla.com/D39502
--HG--
extra : moz-landing-system : lando
The syntax for the new construct looks like the following, and re-uses the
refcounted keyword from refcounted using statements:
async refcounted protocol PFoo { ... }
Differential Revision: https://phabricator.services.mozilla.com/D39501
--HG--
extra : moz-landing-system : lando
The previous behaviour of failing unconditionally was performed as, during
shutdown, the channel could become unable to send without worker threads having
a chance to react. This change keeps that behaviour, isolating async message
senders from impending IPC shutdown, while performing expected actor teardown if
the manager actor has already been destroyed, and should no longer send messages.
An alternate behaviour here could be to crash if !Manager()->CanSend(). That
behaviour may be preferable if a sufficient number of callsites don't check the
return value of the `SendPFooConstructor()` method.
Differential Revision: https://phabricator.services.mozilla.com/D39534
--HG--
extra : moz-landing-system : lando
Gijs Kruitbosch
Emilio Cobos Álvarez
Peter Van der Beken
Nika Layzell
Csoregi Natalia
Jean-Yves Avenard
Ricky Stewart
Simon Giesecke
Bogdan Tara
David Parks
Daniel Varga
Kershaw Chang
Makoto Kato
Chris Peterson
Marco Bonardo
Chris Martin
James Teh
Razvan Maries
Sylvestre Ledru
Mihai Alexandru Michis
Dana Keeler
Eitan Isaacson
Anny Gakhokidze
Jonathan Kew
Andrew McCreight
Ciure Andrei
Bob Owen
Tom Ritter
Dorel Luca
Jed Davis
Oana Pop Rus
Nathan Froyd
David Parks
Olli Pettay
Gabriele Svelto
Barret Rennie
Jim Porter
Aaron Klotz
Byron Campen [:bwc]
Andrea Marchesini