vm_cache_ref starts with a reference count of 1. When acquiring a vm_cache,
you no longer need to worry if that should go through the vm_store, or not;
as it now always does.
* map_backing_store() no longer needs to play with the vm_cache_ref
references.
* that simplified some code.
* vfs_get_vnode_cache() now grabs a reference to the cache, if successful.
* better balanced vnode ownership on vnode_store creation (vnode_store
released the vnode before if its creation failed).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15641 a95241bf-73f2-0310-859d-f6bbb57e9c96
this saves 4 bytes per page. To compensate the loss of bytes, the offset is now
stored in page size units, that's enough to address 2^44 or 16 TB (which is now
the maximal supported file size!).
* Renamed vm_page::ppn to physical_page_number.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15637 a95241bf-73f2-0310-859d-f6bbb57e9c96
arch/*/thread_struct.h to arch_thread_types.h, so that it can directly
be included without having to specify the architecure.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15616 a95241bf-73f2-0310-859d-f6bbb57e9c96
them apart (this even saves a pointer from vm_virtual_map to its address space)
* aspace -> address_space
* vm_create_address_space() did not check if creating the semaphore succeeded
* Removed team::kaspace - was not really needed (introduced a new vm_kernel_address_space()
function that doesn't grab a reference to the address space)
* Removed vm_address_space::name - it was just a copy of the team name, anyway,
and there is always only one address space per team
* Removed aspace_id - the address space is now using the team_id
* Some cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15609 a95241bf-73f2-0310-859d-f6bbb57e9c96
Committed libnetwork.so. I actually wanted to port BSD's libc/net resolver, but that one uses too many unsupported functions, so I basically merged libbind.so and libsocket.so into one library.
Moved the libbind.so sources to libnetwork/dns.
Cleaned up native part of libbind.so a little bit more. Though, the ugly hack remains. We nbetter pthread/mutex support. Added TODO that we also shousctl() into the kernel and libroot.so.
Improved DEVNOTES a little bit.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15605 a95241bf-73f2-0310-859d-f6bbb57e9c96
written the page, we now do it before, so that it cannot lose any changed data
anymore; it doesn't matter if the page is written to while writing it back, the
worst thing that can happen is that we write the same page twice. Also, we don't
rely on the PAGE_MODIFIED bit anymore, we now check all mappings of that page
to find all modified pages, no matter how far the (currently disabled) page
daemon had come.
Also, destroying an area will now result in writing back changed pages - this
is only really important for memory mapped files, though, and should probably
be avoided for other vm_store types.
Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15597 a95241bf-73f2-0310-859d-f6bbb57e9c96
Moved some headers around and cleaned them up.
Fixed threading code in libbind.so. Don't know why I hadn't noticed this before. Now it is reliable, but uses an even uglier hack.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15585 a95241bf-73f2-0310-859d-f6bbb57e9c96
mappings yet, that's a good way to test this functionality.
* Turns out the directory of the mappings must be known already - they
should be added automatically, but I've added them manually for now
(which is okay for the default system directory).
* Having more than one style with the same family in the mappings didn't
work as planned - it now does.
* On my current system, time spend in the font manager constructor went
down from 1.5 secs to around 12000 usecs (and I have only a moderate
number of fonts installed, or so I thought - looks like the mappings
were a good idea :-)) - and that directly cuts down the boot time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15580 a95241bf-73f2-0310-859d-f6bbb57e9c96
service depends on it, it doesn't make any sense to call it that early in the
game.
* The VFS now has a low memory handler for vnodes as well. If there is enough
memory left, it won't free any vnodes anymore.
* Potential crashing bug fix: some functions did not check if the FD passed
in belonged to the right type; they just assumed it had a valid vnode, but
it could have had a mount structure associated as well.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15566 a95241bf-73f2-0310-859d-f6bbb57e9c96
memory status.
Added new B_NO_LOW_MEMORY constant for the usual case.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15551 a95241bf-73f2-0310-859d-f6bbb57e9c96
kernel TLBs from being flushed on context switch.
* new arch_cpu_user_TLB_invalidate() that now does what arch_cpu_global_TLB_invalidate()
did before.
* arch_cpu_global_TLB_invalidate() will now flush all TLBs, even those from the
kernel.
* some cleanups.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15535 a95241bf-73f2-0310-859d-f6bbb57e9c96
* The boot loader now checks the CPU for the cpuid and rdtsc features, which we
currently both rely on.
* Removed old and no longer used stage2_priv.h header
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15534 a95241bf-73f2-0310-859d-f6bbb57e9c96
has not yet been tested, though - I'll do this after this commit):
* Removed the arch_memory_type stuff from vm_area; since there are only 8 memory
ranges on x86, it's simply overkill. The MTRR code now remembers the area ID
and finds the MTRR that way (it could also iterate over the existing MTRRs).
* Introduced some post_modules() init functions.
* If the other x86 CPUs out there don't differ a lot, MTRR functionality might
be put back into the kernel.
* x86_write_msr() was broken, it wrote the 64 bit number with the 32 bit words
switched - it took me some time (and lots of #GPs) to figure that one out.
* Removed the macro read_ebp() and introduced a function x86_read_ebp()
(it's not really a time critical call).
* Followed the Intel docs on how to change MTRRs (symmetrically on all CPUs
with caches turned off).
* Asking for memory types will automatically change the requested length to
a power of two - note that BeOS seems to behave in the same, although that's
not really very clean.
* fixed MTRRs are ignored for now - we should make sure at least, though,
that they are identical on all CPUs (or turn them off, even though I'd
prefer the BIOS stuff to be uncacheable, which we don't enforce yet, though).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15528 a95241bf-73f2-0310-859d-f6bbb57e9c96
but it compiles at least. :-)
* Pulled the architecture specific part out of <posix/setjmp.h> into
<posix/arch/<arch>/arch_setjmp.h>.
* Moved setjmp_save_sigs.c from the x86 specific implementation into a
"generic" sibling directory, since it is reusable (and actually used
by the PPC implementation).
* Added generic/longjmp_return.c containing a function __longjmp_return,
which is invoked at the end of siglongjmp(), resetting the signal mask
and validating the return value. It is used by the PPC implementation,
and should also be used by the x86 implementation, but I'll leave that
to someone who's motivated enough to also test it. :-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15479 a95241bf-73f2-0310-859d-f6bbb57e9c96
* finally nailed that much hated bug where the
wrong background underneath the cursor was restored.
I always looked for this in the wrong place -
of course, you cannot blit regions on top
of the cursor and expect the backup of the
background to be magically valid still... ;-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15461 a95241bf-73f2-0310-859d-f6bbb57e9c96
AttachedToWindow().
This fixes the double team entries in the Deskbar.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15398 a95241bf-73f2-0310-859d-f6bbb57e9c96
The BViews must be resized directly after every change in the window size - we
cannot wait until B_WINDOW_RESIZED, since subsequent calls have wrong view
sizes, then.
B_WINDOW_RESIZED is only really useful for app_server caused window resizing.
Added TODO to SetLook(): it may change the window size as well.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15385 a95241bf-73f2-0310-859d-f6bbb57e9c96
server.
This saves overhead on both sides, the server doesn't need to build the update message
for the client, and the client doesn't have to unflatten and parse another message.
* This code is actually needed for the new clipping code in the app_server, but shouldn't
do much harm for the old app_server, either.
* Also disabled getting the bounds from the server, as that is just never needed (and would
also break the code).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15353 a95241bf-73f2-0310-859d-f6bbb57e9c96
is known by the server, anyway.
* B_MODAL_SUBSET_WINDOW_FEEL now also works as expected.
* Renamed AS_REM_FROM_SUBSET to AS_REMOVE_FROM_SUBSET.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15285 a95241bf-73f2-0310-859d-f6bbb57e9c96
* renamed some Decorator functions like GetLook() to Look().
* renamed _look to fLook and _flags to fFlags.
* removed the feel from decorators - they don't need to know
or care about the feel.
* the DefaultDecorator didn't allow resizing of modal windows.
* DefaultDecorator::SetTitle() no longer clears the passed
in updateRegion, it just fills it (so that you can call
more than one of these methods after the other, and still
get a correct update region).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15266 a95241bf-73f2-0310-859d-f6bbb57e9c96
know what's going on, but this appears to fix it. And since we'll get a new glyph
layout backend sooner or later, I don't want to mess with that stuff again.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15176 a95241bf-73f2-0310-859d-f6bbb57e9c96
an app_server code. It's always up-to-date as it just reads the header
directly.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15148 a95241bf-73f2-0310-859d-f6bbb57e9c96
* RootLayer's mouse event processing is now at its minimum - the
EventDispatcher handles them now. As a result, a window will now
get only one message per event.
* RootLayer adds "_view_token" to mouse moved messages that specify
the view currently under the cursor.
* There is now a mouse event layer in RootLayer that gets preferred
when it's set - this is now used for the window moving instead of
the previous mechanism.
* changed the previous DistributeMessage() to an UnpackMessage()
method following Adi's suggestion.
* caveat: some things might be functionally broken in RootLayer now
because of removing the mouse notification stuff.
* "be:transit" handling is now done completely client side by
BWindow::_SanitizeMessage(() (similar to what the input_server does).
This should also make the mechanism pretty robust, since every
B_MOUSE_MOVED message can now trigger the view transit (in case a
message is lost). B_WINDOW_ACTIVATED messages should be generated
client side as well.
* renamed AS_LAYER_GET_MOUSE_COORDS to AS_GET_MOUSE as it's not a
layer specific command, and also gets the mouse buttons.
* B_MOUSE_* messages from the up server now contain only a "screen_where"
field; "where" (in window's coordinates) and "be:view_where" are
added in BMessage::_SanitizeMessage().
* messages that don't have a valid target in the looper are now
dropped instead of being sent to the looper - this should be done
in BLooper as well, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15087 a95241bf-73f2-0310-859d-f6bbb57e9c96
a suspend focus field to the message, there is now a "feed focus" field in
case the message should be forwarded).
* added a comment to the BPoint version of _FindView() (since it's broken)
* _DistributeMessage() is now called after _DetermineTarget() - so that it
can prevent sending the message twice to the focus view.
* removed BWindow::DoUpdate() as it's no longer used.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15066 a95241bf-73f2-0310-859d-f6bbb57e9c96
* attachView() is now called _CreateSelf() and creates the app_server
view counterpart for itself, and no longer for a child view.
* removed superfluous deleteView().
* moved drawing from BWindow::DoUpdate() to BView::_Draw().
* made the recursive hook call functions consistent.
* fixed BWindow::DispatchMessage() to send public messages to the intended
target (instead of always handling them itself directly).
* DispatchMessage() no longer eats unmapped key events that were targeted
at the window directly.
* B_KEY_DOWN and B_KEY_UP events are now send to the target view as well
(this couldn't work before as BMessages were broken with B_PREFERRED_TOKEN).
* the default button is now correctly targeted by BWindow::_DetermineTarget()
(previously, the enter key was hacked to get through via _HandleKeyDown()).
* removing a view now also makes sure it won't have focus any longer.
* also, the DetachedFromWindow() hooks are now called first, so that any
changes made there cannot mess up our window anymore.
* removed BView::SetPattern(); _UpdatePattern() now does its job.
* renamend private methods to match our preferred style.
* removed unused methods and variables.
* more consistent naming overall.
* removed _PR3_COMPATIBLE_ stuff - there is definitely no need for us to be
compatible to that one.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15060 a95241bf-73f2-0310-859d-f6bbb57e9c96
task_looper() again.
* removed BMessenger::fPreferred - whenever you had to specify "usePreferred" separately,
you don't have to do that anymore - use B_PREFERRED_TOKEN instead.
* fixed BTokenSpace::GetToken() semantics: it will no longer touch the "object" argument
in case of failure.
* Introduced a BWindow::_DistributeMessage() that will be part of the event dispatcher
counterpart to the app_server (the other will be _DetermineTarget()).
* Made it easier to use Michael's Message4 implementation: just add the following line
to your UserBuildConfig:
AppendToConfigVar DEFINES : HAIKU_TOP src : USING_MESSAGE4 : global ;
* Introduced ServerWindow::HandlerMessenger() and FocusMessenger() - the first will
target the client handler, while the other will target the preferred handler of the
client looper (usually the view having focus).
* Fixed dano message unflattening in the Message4 code.
* Changed BMessage::PrintToStream() to no longer use macros in the Message4 implementation.
* I hope that's all - it's a huge change, but it's all connected.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15046 a95241bf-73f2-0310-859d-f6bbb57e9c96
used once now. If tokens were reused, you could eventually target the wrong
(but valid again) messenger.
* removed unused stuff (like GetList()).
* some cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15028 a95241bf-73f2-0310-859d-f6bbb57e9c96
distribute any messages to the clients yet.
* removed the working thread from RootLayer - for now, its event handlers are
still called using input filters in the new event dispatcher, though (to
get things started).
* ServerApp is now using a BMessenger to identify its client, and no longer
stores the port/token separately.
* the input_server handshake is a bit simpler now, as it can now just reply
to the app_server message, removed unused code from ServerProtocol.h
* calmed down the MultiLocker (it always printed thread statistics on startup,
because it's compiled in debug mode).
* removed the cursor thread stuff from AppServer.cpp
* the new event dispatcher now uses a cursor thread when supported (only in
native mode, not in the test environment), although it improves cursor
movement under Qemu, the effect is not as good as expected - this might
need some more investigations (might just be a thread priority problem).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15012 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Rewrote event handling: instead of writing every single device message to
the event port, they are just queued in a list, and the event loop is only
notified if necessary (ie. if a notification has been sent already, new
events are just queued up until the input_server found the time to process
them).
* This also fixed a big memory leak: every message processed by EnqueueDeviceMessage()
(IOW every key or mouse event) was leaked!
* no longer abuses gInputMethodListLocker to lock the method event queue
(it now uses the standard event queue lock).
* removed the completely superfluous, weird and decelerating event caching mechanism
* tried to find a better distribution of work between _SanitizeEvents(),
_MethodizeEvents(), and _DispatchEvents().
* HandleSetMousePosition() now only does what it's supposed to do (this currently
causes the mouse to jump at the start, though).
* now uses the "Message4" for message sending if available.
* fixed "kb_mouse_settings.h" to include all headers it needs.
* some more cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14966 a95241bf-73f2-0310-859d-f6bbb57e9c96
cleanup.
* made BMessageQueue::IsLocked() const - the non-const version is still provided
for binary compatibility.
* Both BMessageQueue::FindMessage() versions are now thread safe, the queue's BLocker
is now mutable to allow for this.
* renamed BMessage::link to fQueueLink as the "Message4" implementation uses it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14955 a95241bf-73f2-0310-859d-f6bbb57e9c96
temporary handling code in the app_server.
* RootLayer no longer creates the input_server messaging port - this is now
the responsibility of the input_server.
* Moved AS_CREATE_[OFFSCREEN_]WINDOW from ServerApp::_MessageLooper() to
_DispatchMessage().
* The RootLayer thread is now started as soon as the input_server is there.
* removed or disabled any input_server stuff in the AppServer class.
* removed old message commmands to the app_server.
* Removed the R5_CURSOR_COMM and HAIKU_APPSERVER_COMM definitions: the
input_server is now automatically built correctly depending on the target.
* InputServer::EventLoop() plays now safe and checks for error conditions.
* InputServer::EnqueueDeviceMessage() seems to leak memory, added TODO about
this.
* InputServer event loop messaging uses ports for inner-app communication - why?
* The InputServer event loop thread is no longer killed on exit, it just quits
when its port is gone.
* Minor cleanup in input_server.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14949 a95241bf-73f2-0310-859d-f6bbb57e9c96
* removed unused stuff.
* shuffled methods around in the source, so that you don't have to jump so far
when working at a specific thing (ie. all draw functions are grouped together, ...)
* ResizeToPreferred() should now work as expected.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14934 a95241bf-73f2-0310-859d-f6bbb57e9c96
* fix deselecting items
* fix selection with the mouse and also holding down modifiers
* fix some more stuff I have already forgotten
* applied coding style in most places
* renamed private functions to be consistent
needs more work though, some stuff can definitely be
removed, some needs to be added
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14932 a95241bf-73f2-0310-859d-f6bbb57e9c96
* uses another virtual to implement WindowActivated() and draw itself in
disabled look just like in R5
* fixed _ButtonRectFor() off by one errors
* tiny problems remain when clicking on the ScrollBar (it is shifted)
* the _ValueFor() might not be perfect
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14929 a95241bf-73f2-0310-859d-f6bbb57e9c96
have different token spaces depending on the scope of its objects.
* removed TokenHandler - we're now using BTokenSpace instead.
* removed unused IPoint.cpp - if we ever need it again, it can still easily be
resurrected from the dead.
* some cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14925 a95241bf-73f2-0310-859d-f6bbb57e9c96
Axel Dörfler