As we saw in bug 1598119, 32-bit nop-space patches aren't compatible with clang 11's CFG because they return to the second instruction of the hooked function.
The `FunctionHook`s for plugins were pulling raw function pointers out of the interceptor stubs, so they didn't get the benefit of the stub's `operator()` that already has the CFG annotation.
As a bandaid, this patch marks all users of `BasicFunctionHook::OriginalFunction()` with the CFG annotation as well. A more thorough fix might be to somehow pass through to the stub's `operator()`, but we need something before merge day and I'm not confident in being able to do that regression-free in time.
Differential Revision: https://phabricator.services.mozilla.com/D90629
CLOSED TREE
We don't need these macros anymore, for two reasons:
1. We have static analysis to provide the same sort of checks via `MOZ_RAII`
and friends.
2. clang now warns for the "temporary that should have been a declaration" case.
The extra requirements on class construction also show up during debug tests
as performance problems.
This change was automated by using the following sed script:
```
# Remove declarations in classes.
/MOZ_DECL_USE_GUARD_OBJECT_NOTIFIER/d
/MOZ_GUARD_OBJECT_NOTIFIER_INIT/d
# Remove individual macros, carefully.
{
# We don't have to worry about substrings here because the closing
# parenthesis "anchors" the match.
s/MOZ_GUARD_OBJECT_NOTIFIER_PARAM)/)/g;
s/MOZ_GUARD_OBJECT_NOTIFIER_PARAM_TO_PARENT)/)/g;
s/MOZ_GUARD_OBJECT_NOTIFIER_PARAM_IN_IMPL)/)/g;
s/MOZ_GUARD_OBJECT_NOTIFIER_ONLY_PARAM_IN_IMPL)/)/g;
# Remove the longer identifier first.
s/MOZ_GUARD_OBJECT_NOTIFIER_ONLY_PARAM_TO_PARENT//g;
s/MOZ_GUARD_OBJECT_NOTIFIER_ONLY_PARAM//g;
}
# Remove the actual include.
\@# *include "mozilla/GuardObjects.h"@d
```
and running:
```
find . -name \*.cpp -o -name \*.h | grep -v 'GuardObjects.h' |xargs sed -i -f script 2>/dev/null
mach clang-format
```
Differential Revision: https://phabricator.services.mozilla.com/D85168
We don't need these macros anymore, for two reasons:
1. We have static analysis to provide the same sort of checks via `MOZ_RAII`
and friends.
2. clang now warns for the "temporary that should have been a declaration" case.
The extra requirements on class construction also show up during debug tests
as performance problems.
This change was automated by using the following sed script:
```
# Remove declarations in classes.
/MOZ_DECL_USE_GUARD_OBJECT_NOTIFIER/d
/MOZ_GUARD_OBJECT_NOTIFIER_INIT/d
# Remove individual macros, carefully.
{
# We don't have to worry about substrings here because the closing
# parenthesis "anchors" the match.
s/MOZ_GUARD_OBJECT_NOTIFIER_PARAM)/)/g;
s/MOZ_GUARD_OBJECT_NOTIFIER_PARAM_TO_PARENT)/)/g;
s/MOZ_GUARD_OBJECT_NOTIFIER_PARAM_IN_IMPL)/)/g;
s/MOZ_GUARD_OBJECT_NOTIFIER_ONLY_PARAM_IN_IMPL)/)/g;
# Remove the longer identifier first.
s/MOZ_GUARD_OBJECT_NOTIFIER_ONLY_PARAM_TO_PARENT//g;
s/MOZ_GUARD_OBJECT_NOTIFIER_ONLY_PARAM//g;
}
# Remove the actual include.
\@# *include "mozilla/GuardObjects.h"@d
```
and running:
```
find . -name \*.cpp -o -name \*.h | grep -v 'GuardObjects.h' |xargs sed -i -f script 2>/dev/null
mach clang-format
```
Differential Revision: https://phabricator.services.mozilla.com/D85168
With these changes, on my Linux analysis with ClangBuildAnalyzer, the
top two expensive headers, DOMTypes.h and TabMessageUtils.h are no longer
among the 30 most expensive headers.
Differential Revision: https://phabricator.services.mozilla.com/D82935
In all those cases, the current nsISerialEventTarget is either the main thread or the MessageChannel's nsISerialEventTarget (since bug 1634846)
Differential Revision: https://phabricator.services.mozilla.com/D81966
There's no use case for stateful comparators, so they can be just plain
function pointers.
This is used in some hot places like CSS selector matching.
Differential Revision: https://phabricator.services.mozilla.com/D77084
During the last cycle collection, state watchers may attempt to dispatch tasks to the current AbstractThread, so we can't clear the TLS entry until that step has completed.
We want the object to be deleted last; ordering in call to ClearOnShutdown() can't guarantee it.
Differential Revision: https://phabricator.services.mozilla.com/D75498
Crash annotations in content processes are currently sent over IPC via
shared memory buffers. To pave the way for the Rust rewrite of the exception
handler we are removing this code and gathering all the crash annotations
within the content processes themselves. This patch causes annotations to be
stored in the global table of each content process. They are then streamed
out to the parent process by the exception handler together with the
exception-time annotations.
This has a number of benefits:
* we have one less channel to exchange data between content processes and
the parent process
* we save memory because we don't need to allocate the shared memory buffers
* annotations are faster because we don't stream them all out every time one
changes
* we won't truncate annotations anymore if we run out of space in the shared
segment.
* we don't need delayed annotations anymore, so we can get rid of the
associated machinery
As I refactored the code I tried to adjust all the obsolete comments,
consolidate shared code and remove the redundant steps that were sometimes
present. In many places we had two entire crash annotation tables we merged to
change just a couple; that comes from the fact that historically we loaded
them from disk. Now it doesn't matter anymore and we can just go ahead and
change the ones we care about.
Differential Revision: https://phabricator.services.mozilla.com/D62586
--HG--
extra : moz-landing-system : lando
This code always intended to force wmode to be one of these modes (bug 1340276) but was not robust to unexpected wmode values (e.g. "default"). Unexpected values are now promoted to "direct".
Differential Revision: https://phabricator.services.mozilla.com/D66477
--HG--
extra : moz-landing-system : lando
Crash annotations in content processes are currently sent over IPC via
shared memory buffers. To pave the way for the Rust rewrite of the exception
handler we are removing this code and gathering all the crash annotations
within the content processes themselves. This patch causes annotations to be
stored in the global table of each content process. They are then streamed
out to the parent process by the exception handler together with the
exception-time annotations.
This has a number of benefits:
* we have one less channel to exchange data between content processes and
the parent process
* we save memory because we don't need to allocate the shared memory buffers
* annotations are faster because we don't stream them all out every time one
changes
* we won't truncate annotations anymore if we run out of space in the shared
segment.
* we don't need delayed annotations anymore, so we can get rid of the
associated machinery
As I refactored the code I tried to adjust all the obsolete comments,
consolidate shared code and remove the redundant steps that were sometimes
present. In many places we had two entire crash annotation tables we merged to
change just a couple; that comes from the fact that historically we loaded
them from disk. Now it doesn't matter anymore and we can just go ahead and
change the ones we care about.
Differential Revision: https://phabricator.services.mozilla.com/D62586
--HG--
extra : moz-landing-system : lando
Specifically, this renames
* nsTArray_CopyChooser to nsTArray_RelocationStrategy
* the Copy template argument of nsTArray_base to RelocationStrategy
* nsTArray_CopyWithConstructors to nsTArray_RelocateUsingMoveConstructor
* nsTArray_CopyWithMemutils to nsTArray_RelocateUsingMemutils
* DECLARE_USE_COPY_CONSTRUCTORS to MOZ_DECLARE_RELOCATE_USING_MOVE_CONSTRUCTOR
Differential Revision: https://phabricator.services.mozilla.com/D66243
--HG--
extra : moz-landing-system : lando
Crash annotations in content processes are currently sent over IPC via
shared memory buffers. To pave the way for the Rust rewrite of the exception
handler we are removing this code and gathering all the crash annotations
within the content processes themselves. This patch causes annotations to be
stored in the global table of each content process. They are then streamed
out to the parent process by the exception handler together with the
exception-time annotations.
This has a number of benefits:
* we have one less channel to exchange data between content processes and
the parent process
* we save memory because we don't need to allocate the shared memory buffers
* annotations are faster because we don't stream them all out every time one
changes
* we won't truncate annotations anymore if we run out of space in the shared
segment.
* we don't need delayed annotations anymore, so we can get rid of the
associated machinery
As I refactored the code I tried to adjust all the obsolete comments,
consolidate shared code and remove the redundant steps that were sometimes
present. In many places we had two entire crash annotation tables we merged to
change just a couple; that comes from the fact that historically we loaded
them from disk. Now it doesn't matter anymore and we can just go ahead and
change the ones we care about.
Differential Revision: https://phabricator.services.mozilla.com/D62586
--HG--
extra : moz-landing-system : lando
gfx::Color is currently misused in many places. The DrawTargets expect
the color space to be in device space, e.g. what we are actually going
to draw using. Everything sitting above generally deals with sRGB, as
specified in CSS. Sometimes we missed the conversion from sRGB to device
space when issuing draw calls, and similarly sometimes we converted the
color to device space twice.
This patch splits the type in two. sRGBColor and DeviceColor now
represent sRGB and device color spaces respectively. DrawTarget only
accepts DeviceColor, and one can get a DeviceColor from an sRGBColor via
the ToDeviceColor helper API. The reftests now pass with color
management enabled for everything (e.g. CSS) instead of just tagged
raster images.
There will be a follow up patch to enable color management everywhere by
default on all supported platforms.
Differential Revision: https://phabricator.services.mozilla.com/D64771
--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
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 also adds an explicit warning when this happens so that the child processes don't crash without leaving a trace.
Differential Revision: https://phabricator.services.mozilla.com/D60650
--HG--
extra : moz-landing-system : lando
Michael Froman
David Major
Nathan Froyd
Mihai Alexandru Michis
sotaro
Simon Giesecke
Jean-Yves Avenard
Butkovits Atila
Bas Schouten
Bogdan Tara
Emilio Cobos Álvarez
Martin Stransky
Narcis Beleuzu
Dorel Luca
Gabriele Svelto
David Parks
Sylvestre Ledru
Brindusan Cristian
Andrew Osmond
Eric Rahm
Jed Davis
Nika Layzell
pbz