Don't hold gMutex when calling HandleSharePortsMessage() from PortServerThread to avoid deadlock.
Differential Revision: https://phabricator.services.mozilla.com/D31694
--HG--
extra : moz-landing-system : lando
This patch adds two things:
1. An optional fixed_address argument to SharedMemoryBasic::Map, which
is the address to map the shared memory at.
2. A FindFreeAddressSpace function that callers can use to find a
contiguous block of free address space, which can then be used to
determine an address to pass in to Map that is likely to be free.
Patches in bug 1474793 will use these to place the User Agent style
sheets in a shared memory buffer in the parent process at an address
that is also likely to be free in content processes.
Differential Revision: https://phabricator.services.mozilla.com/D15057
--HG--
extra : moz-landing-system : lando
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 : 67f6fee8b89933561a48e6f7f531b6969893a574
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
We were first allocating and mapping a virtual memory area using mach_vm_allocate
(similar to mmap with MAP_ANON) and then obtaining a shareable capability for the
underlying VM object using mach_make_memory_entry_64. However the memory mapping
is fragmented into multiple objects if it's over 128mb. Larger memory allocations
than 128mb weren't possible. To fix this, we are calling mach_make_memory_entry_64
with MAP_MEM_NAMED_CREATE. That will create a new memory object and return a port
for it.
MozReview-Commit-ID: 5LLiaqJx175
--HG--
extra : rebase_source : 7ac964a1093eaf8ee30f319f5d21132c5d884362
Change mozilla::Smprintf and friends to return a UniquePtr, rather than
relying on manual memory management. (Though after this patch there are
still a handful of spots needing SmprintfFree.)
MozReview-Commit-ID: COa4nzIX5qa
--HG--
extra : rebase_source : ab4a11b4d2e758099bd0794d5c25d799a7e42680
Haik Aftandilian
Tarek Ziadé
Sylvestre Ledru
Cameron McCormack
Doug Thayer
Cosmin Sabou
Margareta Eliza Balazs
Nazım Can Altınova
Stephen A Pohl
Tom Tromey
Andrew Osmond
Andi-Bogdan Postelnicu
George Wright
Lee Salzman
George Wright
Gian-Carlo Pascutto
Ted Mielczarek