In the MinGW browser build job, we're going to use -fms-extensions,
which will tell clang to start processing these comments. Clang
cannot process them correctly (it's an upstream bug) but it doesn't
need to, because we include the libs we need in moz.build files.
So we exclude them for MinGW builds. mingw-clang gets them wrong and
mingw-gcc (which doesn't even work anymore on -central) ignored them.
In the future, with a llvm fix, we could clean up the moz.build
files and re-enable these comments.
Differential Revision: https://phabricator.services.mozilla.com/D3527
--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
Closures are nice but -- as pointed out in bug 1481978 comment #2 --
it's a footgun to take a std::function argument in a context where heap
allocation isn't safe.
Fortunately, non-capturing closures convert to C function pointers,
so a C-style interface with a void* context can still be relatively
ergonomic.
Correctness improvements:
* UTF errors are handled safely per spec instead of dangerously truncating
strings.
* There are fewer converter implementations.
Performance improvements:
* The old code did exact buffer length math, which meant doing UTF math twice
on each input string (once for length calculation and another time for
conversion). Exact length math is more complicated when handling errors
properly, which the old code didn't do. The new code does UTF math on the
string content only once (when converting) but risks allocating more than
once. There are heuristics in place to lower the probability of
reallocation in cases where the double math avoidance isn't enough of a
saving to absorb an allocation and memcpy.
* Previously, in UTF-16 <-> UTF-8 conversions, an ASCII prefix was optimized
but a single non-ASCII code point pessimized the rest of the string. The
new code tries to get back on the fast ASCII path.
* UTF-16 to Latin1 conversion guarantees less about handling of out-of-range
input to eliminate an operation from the inner loop on x86/x86_64.
* When assigning to a pre-existing string, the new code tries to reuse the
old buffer instead of first releasing the old buffer and then allocating a
new one.
* When reallocating from the new code, the memcpy covers only the data that
is part of the logical length of the old string instead of memcpying the
whole capacity. (For old callers old excess memcpy behavior is preserved
due to bogus callers. See bug 1472113.)
* UTF-8 strings in XPConnect that are in the Latin1 range are passed to
SpiderMonkey as Latin1.
New features:
* Conversion between UTF-8 and Latin1 is added in order to enable faster
future interop between Rust code (or otherwise UTF-8-using code) and text
node and SpiderMonkey code that uses Latin1.
MozReview-Commit-ID: JaJuExfILM9
This introduces the machinery needed to generate crash annotations from a YAML
file. The relevant C++ functions are updated to take a typed enum. JavaScript
calls are unaffected but they will throw if the string argument does not
correspond to one of the known entries in the C++ enum. The existing whitelists
and blacklists of annotations are also generated from the YAML file and all
duplicate code related to them has been consolidated. Once written out to the
.extra file the annotations are converted in string form and are no different
than the existing ones.
All existing annotations have been included in the list (and some obsolete ones
have been removed) and all call sites have been updated including tests where
appropriate.
--HG--
extra : source : 4f6c43f2830701ec5552e08e3f1b06fe6d045860
This introduces the machinery needed to generate crash annotations from a YAML
file. The relevant functions are updated to take a typed enum (in C++) and an
integer constant (in JavaScript). A JavaScript wrapper around the crash
reporter service is provided to hold the constants. The existing whitelists
and blacklists of annotations are also generated from the YAML file and the
existing duplicate code has been consolidated. Once written out to the .extra
file the annotations are converted in string form and are no different than
the existing ones.
All existing annotations have been included (and some obsolete ones removed)
and all call sites have been updated including tests.
--HG--
extra : rebase_source : b4f0d4bf83c64851028c271d3fab3ebcb6fbcd3e
This introduces the machinery needed to generate crash annotations from a YAML
file. The relevant functions are updated to take a typed enum (in C++) and an
integer constant (in JavaScript). A JavaScript wrapper around the crash
reporter service is provided to hold the constants. The existing whitelists
and blacklists of annotations are also generated from the YAML file and the
existing duplicate code has been consolidated. Once written out to the .extra
file the annotations are converted in string form and are no different than
the existing ones.
All existing annotations have been included (and some obsolete ones removed)
and all call sites have been updated including tests.
--HG--
extra : rebase_source : f0e8d229581ac5c0daa0e0454cb258746108e28d
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
This replaces using file_util to open and unlink temporary files
(/dev/shm on Linux, $TMPDIR or /tmp otherwise) with the POSIX shm_open
API, or ashmem on Android (which doesn't implement shm_open).
glibc maps shm_open/shm_unlink to open and unlink in /dev/shm (as does
musl libc), so the Linux situation is mostly unchanged except we aren't
duplicating code from system libraries. Other OSes may (and some do)
use more efficient implementations than temporary files.
FreeBSD's SHM_ANON extension is used if available. Sadly, it's not
standard; it would make this patch much simpler if it were.
This patch changes the shm file names; they now start with "org.mozilla"
instead of "org.chromium" because the original Chromium code is mostly
gone at this point. When running as a Snap package, the required
filename prefix is added; other container/sandbox environments using
AppArmor to restrict the allowed filenames may need to be adjusted.
The shm names now include the creating process's pid, to allow
using sandboxing to prevent interfering with shm belonging to other
applications or other processes within the same browser instance.
MozReview-Commit-ID: 7PirIlcblh4
This adds an about:memory branch, "ipc-channels", which counts
cross-process IPC channels (ProcessLink, IPC::Channel) broken down by
the top-level IPDL actor name; these use OS resources which may be
limited (file descriptors on Linux and Mac). Intra-process use of IPC
(ThreadLink) is not counted.
The maximum channel count for each actor type is reported in another
branch, "ipc-channels-peak". This might be useful if there are
conditions that cause transient fd exhaustion, for example.
This patch also works around a problem where MessageChannel was trying
to register reporters too early in child processes, and failing.
MozReview-Commit-ID: CGEwny2ipcu
--HG--
extra : rebase_source : ad526f03bbef891f4474225a8e36a7ed9e3acdab
nsIURIWithPrincipal is currently used to retrieve the nsIPrincipal from a
BlobURL object. BlobURLProtocolHandler has a hashtable containing, for each
blobURL, a BlobImpl and its nsIPrincipal. This patch introduces
BlobURLProtocolHandler::GetBlobURLPrincipal() that retrieves the nsIPrincipal
from this hashtable.
This patch fixes also a bug in how the revocation of blobURLs is broadcasted to
other processes. This should be done immediately because each process creates
its own timer to revoke them after 5 seconds.
An important change is related to NS_SecurityCompareURIs() where, if 1 (or
both) of the 2 URIs to compare, is a revoked BlobURL, we will QI its URL to
nsIStandardURL and fail out at that point.
The current code is somewhat non-obvious to a first-time reader, and
OS_TEST is a bizarre thing anyway, since it's actually the name of the
CPU we're running on. We'd do well to minimize the use of OS_TEST.
Note that the complete nuking of the xptcall/md/unix/moz.build lines are
because we don't support OS X/x86 anymore.
There's a lot going on here, but it all fits under the idea of
being able to communicate about texture locking statuses
without spinning on IsReadLocked. This is a bit of a trade -
we could just always allocate/grab a texture from the pool,
which would put a smaller cap on the amount of time we can
possibly spend when a texture is locked. However, this eats
up more CPU and memory than waiting on the textures to unlock,
and could take longer, especially if there were a large number
of textures which we just need to wait for for a short amount
of time. In any case, we very rarely hit the case where we
actually need to wait on the sync IPC to the compositor - most
of the time the textures are already unlocked.
There is also an async IPC call in here, which we make before
flushing async paints. This just causes the compositor to
check whether the GPU is done with its textures or not and
unlock them if it is. This helps us avoid the case where we
take a long time painting asynchronously, turn IPC back on at
the end of that, and then have to wait for the compositor
to to get into TiledLayerBufferComposite::UseTiles before
getting a response. Specifically this eliminates several talos
regressions which use ASAP mode.
Lastly, there seem to be no other cases of static Monitors
being used. This seems like it falls under similar use cases
as StaticMutexes, so I added it in. I can move it into its own
file if we think it might be generally useful in the future.
MozReview-Commit-ID: IYQLwUqMxg2
--HG--
extra : rebase_source : 4f05832f51dae6db98773dcad03cb008a80eca6c
There's a lot going on here, but it all fits under the idea of
being able to communicate about texture locking statuses
without spinning on IsReadLocked. This is a bit of a trade -
we could just always allocate/grab a texture from the pool,
which would put a smaller cap on the amount of time we can
possibly spend when a texture is locked. However, this eats
up more CPU and memory than waiting on the textures to unlock,
and could take longer, especially if there were a large number
of textures which we just need to wait for for a short amount
of time. In any case, we very rarely hit the case where we
actually need to wait on the sync IPC to the compositor - most
of the time the textures are already unlocked.
There is also an async IPC call in here, which we make before
flushing async paints. This just causes the compositor to
check whether the GPU is done with its textures or not and
unlock them if it is. This helps us avoid the case where we
take a long time painting asynchronously, turn IPC back on at
the end of that, and then have to wait for the compositor
to to get into TiledLayerBufferComposite::UseTiles before
getting a response. Specifically this eliminates several talos
regressions which use ASAP mode.
Lastly, there seem to be no other cases of static Monitors
being used. This seems like it falls under similar use cases
as StaticMutexes, so I added it in. I can move it into its own
file if we think it might be generally useful in the future.
MozReview-Commit-ID: IYQLwUqMxg2
--HG--
extra : rebase_source : 3624ad04aa01dac1cd38efb47764dc3a8fbd5fbd
There's a lot going on here, but it all fits under the idea of
being able to communicate about texture locking statuses
without spinning on IsReadLocked. This is a bit of a trade -
we could just always allocate/grab a texture from the pool,
which would put a smaller cap on the amount of time we can
possibly spend when a texture is locked. However, this eats
up more CPU and memory than waiting on the textures to unlock,
and could take longer, especially if there were a large number
of textures which we just need to wait for for a short amount
of time. In any case, we very rarely hit the case where we
actually need to wait on the sync IPC to the compositor - most
of the time the textures are already unlocked.
There is also an async IPC call in here, which we make before
flushing async paints. This just causes the compositor to
check whether the GPU is done with its textures or not and
unlock them if it is. This helps us avoid the case where we
take a long time painting asynchronously, turn IPC back on at
the end of that, and then have to wait for the compositor
to to get into TiledLayerBufferComposite::UseTiles before
getting a response. Specifically this eliminates several talos
regressions which use ASAP mode.
Lastly, there seem to be no other cases of static Monitors
being used. This seems like it falls under similar use cases
as StaticMutexes, so I added it in. I can move it into its own
file if we think it might be generally useful in the future.
MozReview-Commit-ID: IYQLwUqMxg2
--HG--
extra : rebase_source : 67f6fee8b89933561a48e6f7f531b6969893a574
At the moment this isn't actually async because we immediately require
the pid and block on launch anyway. It also crashes the entire browser
on otherwise recoverable launch errors, because code that wants the pid
isn't set up to handle that operation failing.
MozReview-Commit-ID: 5favGu34QCv
--HG--
extra : rebase_source : 3c81c53e1eb8ead353ef3477ed3ceea0f5edcbbe
These are no longer providing useful information. There are still a
noticeable number of failures on Windows, but we've narrowed them down to
within SandboxBroker::LaunchApp.
MozReview-Commit-ID: 9srWLNZq1Wo
--HG--
extra : rebase_source : db44114a7623e75f9efd629046d2118748352ed1
This patch adds a new IPDL protocol PBackgroundLocalStorageCache. It is used by LocalStorageCache object to broadcast changes in local storage cache to other content processes. Each origin has its own PBackgroundLocalStorageCache, so now we can notify content processes that actually have a local storage cache for given origin. This greatly improves performance and reduces memory footprint especialy when local storage changes carry big strings and/or happen very quickly (before this patch all child processes were blindly notified).