only if memory becomes tight, and then the range is used from end to
start.
This is useful to reserve heap address ranges.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12675 a95241bf-73f2-0310-859d-f6bbb57e9c96
along the protection flags).
Changed the handling of B_STACK_AREA types and anonymous vm_areas: now
every area can overcommit if B_OVERCOMMITTING_AREA was specified.
B_STACK_AREA areas are still automatically overcommitting, but
B_KERNEL_STACK_AREA areas no longer.
vm_store_anonymous_noswap.c now only tests for guard pages if there were
any specified which is only done for B_STACK_AREA areas.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12673 a95241bf-73f2-0310-859d-f6bbb57e9c96
(right now, it's still malloc/free, just encapsulated in an allocator class).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12662 a95241bf-73f2-0310-859d-f6bbb57e9c96
Also renamed some defines and structures, although the structure and some other names are still odd.
This module should probably be moved into the kernel anyway, as it provides basic and crucial services.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12646 a95241bf-73f2-0310-859d-f6bbb57e9c96
This reduces the stack usage quite a lot, and since file system calls could be
nested (for example, a mounted image on a volume), definitely a must, too.
Might be a good idea for user calls, too, though, although it's not urgent.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12636 a95241bf-73f2-0310-859d-f6bbb57e9c96
We might want to use it as a basis for the device manager documentation.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12634 a95241bf-73f2-0310-859d-f6bbb57e9c96
BEOS_ADD_ONS_KERNEL_MISC.
Added PCI IDE modules to the build and boot links.
The config_manager can now be handled like every other bus manager.
The config_manager appeared twice in the jam argument list.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12629 a95241bf-73f2-0310-859d-f6bbb57e9c96
Can also be run under R5 when TARGET_PLATFORM=r5 for comparison.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12567 a95241bf-73f2-0310-859d-f6bbb57e9c96
After all, there might be real graphics card which behave the same??
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12551 a95241bf-73f2-0310-859d-f6bbb57e9c96
partitions, because the right entry was found - but we always want to
visist all partitions, anyway.
Also, if a partition has not been mounted yet, no mount point will be
printed.
The device string is still missing.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12550 a95241bf-73f2-0310-859d-f6bbb57e9c96
in case BPartition::Name() was NULL or empty (ContentName() was ignored).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12549 a95241bf-73f2-0310-859d-f6bbb57e9c96
mounted, but no one told it about its volume ID - that caused "mountvolume"
not to be able to return a valid mount point.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12548 a95241bf-73f2-0310-859d-f6bbb57e9c96
file system has been mounted, it would have crashed (that for example
happened if you deleted the kernel settings file...).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12547 a95241bf-73f2-0310-859d-f6bbb57e9c96
- has_signals_pending() is not declared in Be headers, and it isn't declared in
our KernelExport.h anymore, too
- removed snooze_etc() from KernelExport.h, as it's already defined in OS.h
(and really exported by libroot.so).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12506 a95241bf-73f2-0310-859d-f6bbb57e9c96
The properties are now the same as in the Be version - there is no reason (IMO) for usage
comments like these; it's a public class, and its scripting usage should be documented in
the BeBook, if anywhere.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12503 a95241bf-73f2-0310-859d-f6bbb57e9c96
keyboard and mouse part), since for some reason, the app_server refuses to work
with the input_server for me now):
- keyboards are now actually detected, the keyboard device is only published if
there actually is a keyboard
- keyboard_close() no longer calls put_module() for the ISA bus manager; it's
sharing that resource with the mouse driver.
- rewrote the keyboard command handshake; now, the acknowledge byte is always
considered, and it's now also safe to call keyboard commands with the interrupt
handler installed (bytes read are now safely routed through the handler).
- keyboard_open() now also sets the PS/2 command byte to enable keyboard
interrupts (this may actually have been the source of all evil - ie. keyboard
not working after having touched the mouse)
- many commands now try again several times if they failed
- ps2_get_command_byte() is now safe to be called when the keyboard driver
is up and running, too (it's now locally buffered)
- more and better error checks
- possibly more I forgot about...
- general cleanup.
Note, there is still some functionality missing from the keyboard driver:
- ctrl/alt/del handling
- key repeat (delay, time, non-repeating)
Repeat stuff might better be moved to the input_server add-on, anyway, for
more flexibility and better performance.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12498 a95241bf-73f2-0310-859d-f6bbb57e9c96
Lots of cleanup:
- moved B_NO_ENABLE_COUNTER flag definition out of int.h to KernelExport.h, as it's
described in the BeBook (although it's probably not really used that often :))
- int.c no longer has any platform dependent code (+ 0x20 on interrupt numbers is gone);
it's now entirely handled in the arch/x86/ section.
- the io_vectors[] is now statically initialized, instead of allocated from the heap
- removed {install|remove}_interrupt_handler(); they weren't that useful, arch_smp_init()
is now calling install_io_interrupt_handler() correctly instead
- introduced a new arch_int.h header file that currently contains NUM_IO_VECTORS only
(though on x86, it also has ARCH_INTERRUPT_BASE == 0x20).
- changed the return type from {install|remove}_io_interrupt_handler() from "long" to
"status_t"
- rearranged and cleaned the PIC initialization code, made the PIC code more prominent
- changed comments that talk about a non existing 8239 (the PIC chip is actually 8259)
- moved arch/x86/interrupts.h to the source directory, as it's not used outside
- added BeOS compatible interrupts_enabled() function, that should replace our
equivalent (and private) are_interrupts_enabled()
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12477 a95241bf-73f2-0310-859d-f6bbb57e9c96
so we shouldn't overwrite it with zero...
I must have tested an older version of the driver by accident; now it
actually works again under R5.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12475 a95241bf-73f2-0310-859d-f6bbb57e9c96
Apparently, the net_server does not use read()/write() but the kernel
syscalls directly. Even worse, it obviously ignores error values like
B_WOULD_BLOCK - the return length has to be set even in case of error.
The driver did not work at all under R5 net_server since then...
Some other minor changes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12474 a95241bf-73f2-0310-859d-f6bbb57e9c96
"AT Keyboard" and "USB Keyboard", instead of "at|usb Keyboard".
Also, the mouse input device now reports a "PS/2 Mouse" instead of
a "ps2 Mouse".
Generally, both now capitalize the device name when reporting the name.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12465 a95241bf-73f2-0310-859d-f6bbb57e9c96
Decreased the pipe buffer to 32768 bytes (it's 4096 on BeOS).
pipefs_select/deselect() did no locking.
Switched to using the BenaphoreLocker where possible.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12381 a95241bf-73f2-0310-859d-f6bbb57e9c96
Fixed location of the ldscript.
Is this driver needed anymore at all?
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12366 a95241bf-73f2-0310-859d-f6bbb57e9c96
called kernel_${arch} like on disk. A "jam kernel" will still work, though.
Also updated all those Jamfiles I forgot to update before.
Since makehdimage now runs through (and even builds a working image), I guess
most of the consequences of the "kernel/core" -> "system/kernel" rename task
are solved.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12364 a95241bf-73f2-0310-859d-f6bbb57e9c96
If anything is still broken (and was not before :)), please shout.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12361 a95241bf-73f2-0310-859d-f6bbb57e9c96
of the kernel. Also fixed unsafe userland communication, and other minor issues.
The packet_buffer encapsulates the ring_buffer and adds some handy functionality
(like locking).
Removed the global open masks - no reason for them to be defined globally.
Lots of cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12358 a95241bf-73f2-0310-859d-f6bbb57e9c96
fix the keyboard/mouse problems of most (because they probably didn't know
about this flag). Still does not work for me, though.
Note, if you want consoled the use the input_server, or if you want to use
the input_server as an R5 replacement, you have to define this flag again.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12356 a95241bf-73f2-0310-859d-f6bbb57e9c96
"consoled" now uses this information to tell the TTY its size.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12355 a95241bf-73f2-0310-859d-f6bbb57e9c96
of months now. This solves (probably) all those strange problems you might have
experienced using it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12354 a95241bf-73f2-0310-859d-f6bbb57e9c96
the advantage that module_image::ref_count does not get messed up, so we could still
determine when to unload such a module (with an extra function).
get_module() now calls put_module_image() if B_MODULE_INIT failed - failing modules
were never unloaded before...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12343 a95241bf-73f2-0310-859d-f6bbb57e9c96
arbitrarily into the current team address space if there are pending bytes
in the chain.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12329 a95241bf-73f2-0310-859d-f6bbb57e9c96
- it's restore_interrupts() not enable_interrupts()
- B_READ_AREA|B_WRITE_AREA gives every *userland* application full access
- fixed coding style
- the PCI module was never put away (but get_module() was called almost everywhere)
- added comments where the code looked buggy on first sight
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12326 a95241bf-73f2-0310-859d-f6bbb57e9c96
- removed libnetapi.so from $NETWORK_LIBS - it's not used by anyone anyway,
and it's definitely not necessary to link against it by default.
Note, this might cause problems in some of the mail add-ons; I haven't
tested this.
- route/ping/... now also link against $SELECT_UNAME_ETC_LIB
makehdimage should now work again under all BeOS platforms.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12325 a95241bf-73f2-0310-859d-f6bbb57e9c96
- pending_chain was set to -1, then MAIL:pending_chain was read into the variable "chain",
but later pending_chain was used again for comparison - and if that was not enough,
"chain" was later used again as well... (Nathan please check this)
- pending_chain was int32, Chain()->ID() uint32
- renamed pending_chain to pendingChain
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12323 a95241bf-73f2-0310-859d-f6bbb57e9c96
Removed GPLd network add-ons from the target image for now as they are only
compilable when the tree was configured to support GPL add-ons.
(I am not even sure if GPLd drivers are a legal problem for us, anyway)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12322 a95241bf-73f2-0310-859d-f6bbb57e9c96
this way, it won't mess with the MMU anymore after mmu_init_for_kernel()
has been called.
No other real changes; still disabled, still doesn't work (on my machine
at least).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12284 a95241bf-73f2-0310-859d-f6bbb57e9c96
by other sources as well. Maybe we can remove that parameter from
mmu_map_physical_memory() altogether, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12283 a95241bf-73f2-0310-859d-f6bbb57e9c96
it doesn't really do anything, it just passes the initial frame buffer
on to the app_server).
While it seems to work on real hardware (if you set the video mode to
640x480x32, app_server restriction), under Bochs, the app_server crashes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12273 a95241bf-73f2-0310-859d-f6bbb57e9c96
Also changed readlink() to be POSIX compliant with those changes.
"ls -l" does now resolve links properly again (the new coreutils version outlined the problems).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12263 a95241bf-73f2-0310-859d-f6bbb57e9c96
to ours - pretty hefty bugs, I wonder how it could work before at all.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12261 a95241bf-73f2-0310-859d-f6bbb57e9c96
already before mapping it. That's necessary because the vm_soft_fault()
function cannot guaranty to call the fault handler only once.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12260 a95241bf-73f2-0310-859d-f6bbb57e9c96
RUN_WITHOUT_APP_SERVER is no longer needed.
This reduces the hassle when switching between the real app_server and
consoled.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12256 a95241bf-73f2-0310-859d-f6bbb57e9c96
know it yet - this fixes an endless loop in the VM fault handler.
set_area_protection() no longer disallows settings B_EXECUTE_AREA and B_WRITE_AREA
at the same time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12252 a95241bf-73f2-0310-859d-f6bbb57e9c96
might be a better choice, this error suggests a kernel error - it shouldn't be
possible to achieve this with the public API).
Added fix_protection() function to make the area protection valid (ie. add
missing kernel protections), and also to make it consistent throughout the
API.
set_area_protection() no longer disallows setting B_EXECUTE_AREA and B_WRITE_AREA
at the same time: a) you can create areas with these flags, and b) it was thought
as a security feature, but it would be very weak (you could simply clone the
area read/write aynway).
The internal APIs now ask the arch VM if it can support the protection the
caller asked for.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12251 a95241bf-73f2-0310-859d-f6bbb57e9c96
create_area() and friends should fail if it's not supported.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12248 a95241bf-73f2-0310-859d-f6bbb57e9c96
with the IDs for the calling team, ie. setpgid(getpid(), getpid())
because it's essentially the same thing.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12247 a95241bf-73f2-0310-859d-f6bbb57e9c96
This fixes the "setpgrp() failed in child" message in gdb (that happened because
it lied and called setpgid() instead).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12246 a95241bf-73f2-0310-859d-f6bbb57e9c96
In case of an unknown error, it will now also take the error base definitions
into account (ie. it could return "Unknown Application Kit Error (x)").
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12237 a95241bf-73f2-0310-859d-f6bbb57e9c96
you choose it in the menu.
The boot loader now always switches into a graphics mode unless it's in
debug mode (currently triggered by pressing the escape key on startup).
That means, "Standard VGA" now actually means mode 0x12 (640x480x4)
instead of text mode.
Since the current limited boot logo code would do stupid things with VGA
planes, there will now only be some colored blocks visible during boot.
Implemented support for monochrome modes in the frame buffer console.
Additionally, the frame buffer console now has support for the VGA mode; it
will treat all 4 bit modes as monochrome for now (the VGA 0x12 mode is a
planar mode so this works nicely).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12235 a95241bf-73f2-0310-859d-f6bbb57e9c96
buffer resolution changes during execution (ie. if the boot menu is opened late
and you change the video mode).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@12233 a95241bf-73f2-0310-859d-f6bbb57e9c96
Axel Dörfler