This seem to enhance the stability of the app_server by a magnitude (at least I can
now run apps like Installer and Media without an instant crash).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14549 a95241bf-73f2-0310-859d-f6bbb57e9c96
Turns out we are using a single FT_Face object for a given font size throughout the
app_server - but the FT_Face object is not designed for a multi-threaded access, AFAICT.
For example, it only has a single glyph slot, that we were using from different threads.
This fix does not cover the renderer, however, which also uses that shared object; IOW
even though it will crash less often it will still crash because of this.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14543 a95241bf-73f2-0310-859d-f6bbb57e9c96
* there is now a "Disable Hyper-Threading" safemode in the boot loader
* the SMP & HT menu items are now added in smp.cpp - and are only added
if the system supports one of them.
* more cleanup to smp_apic.h
* removed cpuid() from the boot loader's support.S - instead, it will now
use the one from the kernel.
* added a very weak HT detection: if the MP config only listed one CPU,
and this CPU supports HT, we enable the other logic processor manually -
as this currently doesn't work, it's disabled, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14536 a95241bf-73f2-0310-859d-f6bbb57e9c96
Return type of thread_func is now status_t instead of the previous int32.
Added some more cpuid_info fields - many fields aren't just reserved anymore.
"sysinfo" now supports the extended features of Intel CPUs, and also knows
the "HTT" bit of the standard features indicating the hyper-threading capability.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14533 a95241bf-73f2-0310-859d-f6bbb57e9c96
tested ok on Dano
note that _IO* macros were redefined for the R5 target, the ones in Haiku aren't compatible with R5
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14524 a95241bf-73f2-0310-859d-f6bbb57e9c96
The "data_ptr" parameter is now only freed if you specify SMP_MSG_FLAG_FREE_ARG
when sending the ICI message.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14519 a95241bf-73f2-0310-859d-f6bbb57e9c96
Extracted scheduler_init() from start_scheduler() (which is now called scheduler_start()).
Moved scheduler related function prototypes from thread.h to the new scheduler.h.
Cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14518 a95241bf-73f2-0310-859d-f6bbb57e9c96
part of the base table, not the extended table.
Renamed some structure fields, variables to be clearer and nicer to read.
Removed some unused stuff.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14502 a95241bf-73f2-0310-859d-f6bbb57e9c96
your system, I've also added a "Disable SMP" safemode option. The NO_SMP
define is still there, and will be removed once SMP works flawlessly.
Prints out infos about the interrupt entries in the MP config.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14499 a95241bf-73f2-0310-859d-f6bbb57e9c96
having the thread lock held and interrupts disabled.
Cleaned up the signal handling code, and fixed some minor bugs with blockable
vs. non-blockable signals.
thread_debug_info was using uint64 for signals sets instead of sigset_t.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14457 a95241bf-73f2-0310-859d-f6bbb57e9c96
that into account as well (they were reporting an error even though everything
went fine).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14450 a95241bf-73f2-0310-859d-f6bbb57e9c96
* got rid of bogus member variables in FontStyle (which were
already flags in the underlying FT_Face structure)
* disabled using the FreeType font cache -> I think from
my earlier tests, I can conclude that the cache was not
actually working. At least not giving any speed improvements.
The AGG engine contains a caching system, for now, it works ok.
I have no idea if this has anything to do with crashes in the
freetype code, but at least I have not seen any since this
change. But I have not tested much...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14448 a95241bf-73f2-0310-859d-f6bbb57e9c96
Added and implemented new functions cache_blocks_in_[sub_]transaction().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14431 a95241bf-73f2-0310-859d-f6bbb57e9c96
some heuristic: when you booted from a CD, CDs are preferred; else, volumes with
names like "Haiku" or "System" are preferred - if someone has better ideas, please
shout.
Note, this heuristic will only come into play if the boot loader was loaded from
an image (ie. floppy/CD/network), and you didn't choose any boot device.
Added evil methods to the Stack class that come in handy (you can now directly
access the array) for this.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14410 a95241bf-73f2-0310-859d-f6bbb57e9c96
and into its own file vfs_boot.cpp.
Added basic support for booting from CD - it doesn't give CDs a higher priority,
so you could end up booting from HD when you didn't explicetly select "CD-ROM"
in the boot loader. Eventually, it should only boot from HD in this case, if
booting from CD failed (because of a missing boot partition or whatever).
fs_mount(), _kern_mount(), and _user_mount() will now return the dev_t of the
mounted device, and not just B_OK. Maybe we should have fs_unmount() work on
a dev_t instead of a path as well...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14403 a95241bf-73f2-0310-859d-f6bbb57e9c96
"booted_from_image" in the kernel_args' boot_disk structure.
Also, added fields "cd" and "user_selected".
A CHOICE_MENU menu can now have a choice text - this is automatically updated
as entries in the menu get selected.
The boot volume menu now has the initial choice text "CD-ROM or hard drive"
in case the boot loader was loaded from an image. The "Rescan volumes" item
is no longer selected by default (only if there was no boot volume found) - but
it's still functionless anyway.
The TAR fs will now appear as "Boot from CD-ROM" in the boot volume menu.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14388 a95241bf-73f2-0310-859d-f6bbb57e9c96
like floppy or CD boot.
This allows it to reduce the number of scans needed to identify the boot
partition - when booted from a real floppy, this speeds up the boot
process by a magnitude.
Also, the loader now has a fall back in case there were no "boot" links
on the disk - the current boot floppy script doesn't create them.
With these changes, I was able to boot into a HD based Haiku installation
from a floppy disk. It's not yet enough to boot from CD (as the boot
device selection is a bit too simplistic right now), but it will eventually
come next. Testing is a lot slower here, though, as neither qemu nor
Bochs support multi-session CDs (at least I have no idea how to get them
to do this).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14380 a95241bf-73f2-0310-859d-f6bbb57e9c96
per file descriptor (a descriptor can be shared among several slots). There is now a
second table in the io_context structure that contains that information in a bitmap.
There are now two new (private) functions to control the close-on-exec flag, fd_close_on_exec(),
and fd_set_close_on_exec().
F_DUPFD, dup(), and dup2() are supposed to clear the close-on-exec flag on the duplicated
slot - this fixes bug #57 (no output after a redirect of a shell builtin).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14313 a95241bf-73f2-0310-859d-f6bbb57e9c96
socketpair() in the network stack is empty and still to be implemented for real
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14267 a95241bf-73f2-0310-859d-f6bbb57e9c96
moved static inline to extern inline to allow this change
please someone review this (ie are they platform dependent functions ?)
should fix bug #47
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14103 a95241bf-73f2-0310-859d-f6bbb57e9c96
improved GetBoundingBoxesAsString (just for fun, keep in mind we'll use AGGTextRenderer in the end :) )
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14091 a95241bf-73f2-0310-859d-f6bbb57e9c96
removed the locking it did, as that doesn't work anymore in the kernel debugger.
The function was not thought to be used outside the debugger, anyway.
Improved usage message from debugger command "ls".
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14050 a95241bf-73f2-0310-859d-f6bbb57e9c96
even if "kernel_startup" was "true". Page faults are now no longer allowed
during kernel startup (and could be easily avoided). The only situation where
we accept page faults with interrupts turned off now is during a kernel debugger
session.
Added a command debug_debugger_running() to test for that situation.
kernel_debugger() no longer sets kernel_startup while it's running; there should
be no situation when this could be helpful.
Interrupts are no longer enabled when a page fault happens in the kernel
debugger.
This potentially fixes all sorts of problems, and not only in the kernel debugger,
it could also have affected SMP (will test later).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14045 a95241bf-73f2-0310-859d-f6bbb57e9c96
- moved standard keyboard navigation into the BView::KeyDown() hook
- the window now only handles tab+option/command key
- B_COMMAND_KEY triggers group jumping, not B_CONTROL_KEY (that activates the switcher,
but directly in the app_server)
- fixed broken group navigation: (modifiers & B_COMMAND_KEY & B_SHIFT_KEY) is different
to (modifiers & (B_COMMAND_KEY | B_SHIFT_KEY)) and is just never true with these
constants.
That allows the tab key completion to be used again in Terminal.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14038 a95241bf-73f2-0310-859d-f6bbb57e9c96
from OS.h, sorry for triggering a complete rebuild.
To avoid doing a syscall for find_thread(0), the assembly
version is now in libroot. For BeOS R5 compatibility,
_kfind_thread_ is retained. This will fix some compile
problems, and provides a cleaner OS.h for future Haiku
versions.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14014 a95241bf-73f2-0310-859d-f6bbb57e9c96
Moved the scanning of individual font folders to FontServer
Implemented server-side code for update_font_families
Removed ClientFontList from the build
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13938 a95241bf-73f2-0310-859d-f6bbb57e9c96
working directory (instead of the full path).
Cleanup of some remaining "int" status variables (where it should have
been a "status_t").
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13924 a95241bf-73f2-0310-859d-f6bbb57e9c96
Also cleaned up Message3.h revert it if you think its a binary compatibility issue.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13920 a95241bf-73f2-0310-859d-f6bbb57e9c96
Sorry to pollute the repository with all those extra files, they will be (re-)moved when the decision about which implementation to use is made.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13916 a95241bf-73f2-0310-859d-f6bbb57e9c96
prevents the system to allocate caches for files that don't use or have
a file cache (ie. only those can be mmap()ed!).
Therefore, cache_prefetch() no longer crashes when trying to prefetch
files without a file cache.
read_into_cache() no longer does anything if the requested size is 0.
Fixed a bug in cache_prefetch_vnode(): if the cache couldn't be retrieved,
it put the vnode, but didn't own it (the caller does).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13904 a95241bf-73f2-0310-859d-f6bbb57e9c96
device becomes available.
Currently, it opens the "launch_speedup" module (if available), later it
should consult a settings file for what to do.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13894 a95241bf-73f2-0310-859d-f6bbb57e9c96
(dprintf() locks using acquire_spinlock() which can itself drop into the
kernel debugger, causing an endless loop (until the stack was full).
Removed debug_putchar().
The gdb interface is now calling arch_debug_serial_*() directly.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13882 a95241bf-73f2-0310-859d-f6bbb57e9c96
Removed vfs_vnode_release_ref(), as vfs_put_vnode() already does the same thing.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13867 a95241bf-73f2-0310-859d-f6bbb57e9c96
the other one with usual device/inode ID pair.
Both versions now accept an offset/size pair to specify the region of the
file to be prefetched - this may be turned into a file_vec_io array later on.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13866 a95241bf-73f2-0310-859d-f6bbb57e9c96
changed the way prefs is stored, so that we have default settings
general cleanup
time sliders are now linked together as on R5
password window is centered onscreen
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13849 a95241bf-73f2-0310-859d-f6bbb57e9c96
added ScreenSaver.cpp to ScreenSaver preferences to have Matrix working
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13828 a95241bf-73f2-0310-859d-f6bbb57e9c96
creation/deletion (and management) over to that class.
ServerApp now gets a desktop pointer, and no longer uses gDesktop.
Converted private MessageLooper::_MessagePort() to a public method MessagePort()
so that the looper can be addressed from elsewhere without using PostMessage().
Added a real basic message loop to MessageLooper::_MessageLoop().
BApplication now only asks the app_server to get its desktop object which should
now be used for everything that's not in the realm of the application.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13824 a95241bf-73f2-0310-859d-f6bbb57e9c96
The unflatten time is now reduced to about a third of the current BMessage implementation but it's still
about half as quick as R5 (we're talking about microseconds here).
A third version of BMessage that operates purely on a flat buffer is in the works. We'll see which one
will be faster for normal uses.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13821 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added the respective case statement in AppServer::DispatchMessage().
The code that actually activates the app is still missing.
* Removed the remnants of the old way of notifying the registrar about
what app got activated (the activated client window did that).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13820 a95241bf-73f2-0310-859d-f6bbb57e9c96
user application performs a division by zero or causes a general
protection fault. For some exceptions (e.g. machine check) I wasn't
quite sure whether they can be caused by user apps at all, so we panic()
in those cases. Wouldn't harm, if someone more knowledgable would check
this, though.
* Removed the unused fault handling stuff, respectively moved the little
that was used into x86/arch_int.c.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13795 a95241bf-73f2-0310-859d-f6bbb57e9c96
Passes all unit tests now and works on Haiku too. Speed is not yet optimized and Message2.cpp is still not cleaned up.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13785 a95241bf-73f2-0310-859d-f6bbb57e9c96
It's currently really broken so don't even try to test it, but you can still review it ;-).
Message2.cpp is not yet cleaned - more to come...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13784 a95241bf-73f2-0310-859d-f6bbb57e9c96
use it. It is not (yet) included in the build and won't break anything.
As we now only have thin wrappers around the *Data() functions, the templatized implementation does
not make much sense anymore and wouldn't work either.
I started this new implementation to be as clean as possible. Instead of using a std::map and
BDataBuffer it uses BList and BMallocIO. It passes the unit tests and it even seems to be a bit quicker
in some tests (but not as quick as the R5 one).
Flattening/Unflattening does not work yet so you can't use it under Haiku right now. It's completely work in progress (I started it just 4 hours ago).
Shout if you see something completely broken, reviews welcome.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13776 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Moved setting the default cursor from ServerScreen to Desktop
* Getting the default cursor is now done using the CursorManager
* Removed outdated setcursor from SysCursor.cpp (we have a new implementation by now)
* Renamed SysCursor.cpp to CursorSet.cpp as that's what it is
* Moved headers/private/app/SysCursor.h to headers/private/servers/app/CursorSet.h
* Removed some unneeded header includes along the way
There remains {set|get}_syscursor now in CursorSet.cpp. Serverside for these are not implemented
and they are obvious hacks. Do we need to keep them?
Also this commit _would_ break Appearance, but 1) all the related code is currently commented out
with the comment "cursor set management belongs in another app" and 2) it is already broken
because of ColorSet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13726 a95241bf-73f2-0310-859d-f6bbb57e9c96
Deskbar to initiate the shutdown process asynchronously. Couldn't test it,
because opening the Be menu doesn't work:
***PANIC: BW: Can't find view with ID: 19 !***
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13722 a95241bf-73f2-0310-859d-f6bbb57e9c96
the desktop's settings, you have to do something like this:
DesktopSettings settings(desktop);
settings.SetMouseMode(mode);
The advantage of this is that this object is fully locked, and cannot lead to
corrupted settings anymore. Also, the settings will stay the same until you
delete the object again.
Updated all accesses to use this new API.
Removed no longer used FFM messages.
Implemented AS_{GET|SET}_MENU_INFO for future use.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13719 a95241bf-73f2-0310-859d-f6bbb57e9c96
debug_server uses this feature to print stack traces with symbols.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13698 a95241bf-73f2-0310-859d-f6bbb57e9c96
is sent to a debugger when the debugged team has been successfully handed
over to another debugger.
* Fixed handling of B_DEBUG_MESSAGE_PREPARE_HANDOVER. We don't send a reply.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13683 a95241bf-73f2-0310-859d-f6bbb57e9c96
allocating a buffer and copying your data into it, when you have data to
send in several chunks (for example, this could be used by BMessage, as
suggested by Ingo Weinhold).
Code is untested, but should work.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13667 a95241bf-73f2-0310-859d-f6bbb57e9c96
This is needed in case the on-disk representation of a file changes (due to reorganization/defragmentation).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13660 a95241bf-73f2-0310-859d-f6bbb57e9c96
last commit was wrong as easily noticed by Jack Burton :)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13636 a95241bf-73f2-0310-859d-f6bbb57e9c96
Forgot to adapt the limit to switch to "GHz" in AboutHaiku.
Both bugs were reported by Herve W a.k.a. "V" - whatever
that means :-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13571 a95241bf-73f2-0310-859d-f6bbb57e9c96
someone with app server insight wants to do that. :-)
* Simplified main_thread_for() for Haiku.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13538 a95241bf-73f2-0310-859d-f6bbb57e9c96
Fixed bug in BShelf::ReplicantAt() - it is supposed to test if any of
the variables passed in is NULL before writing to them.
Got rid of the _PR3_COMPATIBLE_ stuff - we definitely don't need that :)
Refactored the code a bit - moved the identical parts of the three
DeleteReplicant() variants into a separate method _DeleteReplicant(),
moved _rep_data_ code out of the class definition, replaced names like
"fView2" and "aBool" with something more meaningful.
Prefixed private methods with '_'.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13471 a95241bf-73f2-0310-859d-f6bbb57e9c96
Basic support for the desktop window feel: WinBorder will now resize
a window with kDesktopWindowFeel to span over the whole screen.
Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13462 a95241bf-73f2-0310-859d-f6bbb57e9c96
Fixed handling in ServerWindow as stippi's latest commit broke it.
It's now properly done with a separate ServerWindow::SetTitle() method,
that will also take care to rename the window's thread.
Changed naming the window thread in the app_server to "w:<team>:<title>".
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13461 a95241bf-73f2-0310-859d-f6bbb57e9c96
(we can still remove it later if needed).
Added B_SERVER_TOKEN type.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13451 a95241bf-73f2-0310-859d-f6bbb57e9c96
any public header.
* Replaced a good deal of the MessageDeliverer's DeliverMessage() versions
by more general ones using the new interface MessagingTargetSet to represent
a set of targets. This simplifies the usage in cases where the caller doesn't
already have the targets in a supported representation.
* Implemented a first approximation of the shutdown process. There is no
GUI yet. Only superficially tested under R5.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13417 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Adjusted the shutdown command to use BRoster::ShutDown(). Removed
the alert.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13416 a95241bf-73f2-0310-859d-f6bbb57e9c96
