void after turning off BAT for the segment containing itself.
The monster macro for the exception vector code was not really
elegant besides being too long for the 32 byte performance
monitor exception slot. Furthermore wasting three of the SPRG*
registers as cheap scratch memory wasn't that nice either.
We now have a three-step approach: The exception vectors
themselves contain only five instructions which branch to common
code at the beginning of the same physical page. That one sets
up BAT for itself, turns address translation back on and jumps
into the kernel. There we turn off BAT again, dump an iframe,
and enter the actual exception handler (/dispatcher). Upon return
the registers are restored from the iframe and we get back to the
place where the exception occurred.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15881 a95241bf-73f2-0310-859d-f6bbb57e9c96
and 24 bits the set upper bits were overwriting the upper bits of
the word at the target location (usually something like a branch
instruction).
This is the kind of bug that easily take six hours to find, especially
when working with interrupt code and being convinced that the problem
must be related.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15879 a95241bf-73f2-0310-859d-f6bbb57e9c96
* fixed weird pointer conversion in SetStyle()
* fixed a potential mix up in operator=() in case the
other ServerFont has fStyle == NULL
ServerWindow:
* the WindowLayer fTopLayer cannot be deleted by
client request, just for safety reasons
* the link is flushed if there is no drawing engine,
but this case is theoretical only
* deleting the ServerWindow object syncs with the
client, so that when BBitmaps are deleted, they
can be sure there are no pending messages (which
would be executed in a nother thread)
* there is no timeout anymore when sending messages
to the client, which made absolutely no sense
AGGTextRenderer:
* renamed fFontManager to fFontCache, because that's
what it really is
* fLastFamilyAndStyle defaulted to the system plain
font and therefor that font was never loaded when
the font never changed meanwhile
DrawingMode:
* I'm not quite sure but I think there was the
potential of a division by zero, at least I
had crashes with "divide error"
HWInterface:
* fix update when the cursor shape changed in
double buffered mode
ViewLayer:
* since the top layer is never really deleted
before its time has come, it is not necessary
to set it to NULL in the ViewLayer destructor
ViewLayer/WindowLayer:
* added a function to collect the view tokens
that are affected by an update session
EventDispatcher:
* use the importance of the message for the timeout
in _SendMessage()
* drop mouse moved events in the server if we're
lagging behind more than 5 ms (Axel, maybe review)
View:
* there were some problems with the locking
of the BWindow looper in RemoveSelf(), since
this is called from the window destructor,
also of BWindows from BBitmaps, which have
never been run (this might need review), at
least I seem to have solved the crashing
problems introduced by actually deleting the
view hirarchy in the BWindow destructor
* fixed _Draw() for being used non-recursively,
temporarily disabled DrawAfterChildren, which
didn't work yet anyways (because views cannot
draw over children in the server yet)
Window:
* small cleanup when deleting shortcuts
* sync with the server when having send
AS_DELETE_WINDOW (see ServerWindow above)
* fixed locking in Begin/EndViewTransaction()
* removed folding of _UPDATE_ messages, since
there is only one ever in the queue
* set the fInTransaction flag during an update,
I plan to use this in BView later to
flush the link when drawing outside of an
update
* BView::_Draw() is now called by view token,
this gives the next leap forward in speed,
the overhead because of drawing clean views
was considerable
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15878 a95241bf-73f2-0310-859d-f6bbb57e9c96
fixed some visual glitches (and gone around a bug in our menu implementation)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15874 a95241bf-73f2-0310-859d-f6bbb57e9c96
(as the shortcut list wasn't emptied), courtesy of Stippi.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15866 a95241bf-73f2-0310-859d-f6bbb57e9c96
The time base conversion factor is the 32 bit value
2^32 * 1000000 / time base frequency,
so the system time can be computed by
system time = time base * conversion factor / 2^32.
The expression in system_time() looks more complicated now, but is
actually much faster (factor 2.5 on my Mac mini). I'm positively
surprised, how good the assembly looks, that GCC 4 generates. There's
not that much potential for optimization by hand-coding the function.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15863 a95241bf-73f2-0310-859d-f6bbb57e9c96
and real_time_clock[_usecs]() finally seem to work (at least in
the kernel).
* Removed accidentially committed debug output.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15861 a95241bf-73f2-0310-859d-f6bbb57e9c96
is actually not that nice for PPC since we don't have large pages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15860 a95241bf-73f2-0310-859d-f6bbb57e9c96
parts, too. Fixed a potential overflow.
* The generic physical page mapper reserves the virtual address range
for the IO space now, so that noone can interfere until an area has
been created. The location of the IO space is no longer fixed; it
didn't look to me like it was necessary for x86, and we definitely
need to be flexible for PPC.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15855 a95241bf-73f2-0310-859d-f6bbb57e9c96
"IO space" out of the x86 specific source into arch/generic. We'll use
it for PPC as well.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15853 a95241bf-73f2-0310-859d-f6bbb57e9c96
* make sure layers are removed from the token space when the
WindowLayer is deleted
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15852 a95241bf-73f2-0310-859d-f6bbb57e9c96
* BViews set fOwner to NULL recursively when detached
* updated a comment
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15851 a95241bf-73f2-0310-859d-f6bbb57e9c96
now run under BeOS as well (as long as they don't use any functions that are not
available under R5).
The solution is a bit messy, but we have to live with it :-)
The runtime loader now patches the __gRuntimeLoader symbol in libroot.so to point
to its exported structure instead of passing it to the init functions as an
argument.
(Hax0red by axeld and bonefish on stippi's assimilated machine -- resistence is futile)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15848 a95241bf-73f2-0310-859d-f6bbb57e9c96
device in the Open Firmware implementation of boot loader and
pass its path to the kernel, where it's opened and used for
getting/setting the real time. The expensive atomic_*64() on PPC
32-bit make things a bit more complicated. Moreover, missing
64 bit multiplication and division instructions won't really
allow system_time() to be anywhere near as fast as on x86. :-/
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15837 a95241bf-73f2-0310-859d-f6bbb57e9c96
substructure now (that's the only member actually). The system time
offset is therefore accessed via architecture specific accessor
functions.
Note, that this commit breaks the PPC build. Since I want to rename at
least one file I've already changed, I can't avoid that.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15835 a95241bf-73f2-0310-859d-f6bbb57e9c96
the Deskbar anymore. It's still disabled for the libbe_test build, though
(as it will find incompatible add-ons via queries there).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15834 a95241bf-73f2-0310-859d-f6bbb57e9c96
it was good enough to fix SetMouseEventMask() to no longer drop into the debugger).
* Minor cleanup, renamed member variables to comply with our style guide.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15832 a95241bf-73f2-0310-859d-f6bbb57e9c96
