* vm_clone_area() does now set the B_SHARED_AREA flag on both the source
and the cloned area. This is necessary, since it would no longer be
guaranteed that areas are backed by leaf caches only (after
fork()ing), which doesn't work with our cache merging strategy.
Fixes#2605.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28296 a95241bf-73f2-0310-859d-f6bbb57e9c96
matches the program's image name. This is a special case for add-ons
that link against the application, with the application not having a
soname set. The concerned Pe add-ons (HeaderHeader and others) work now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28294 a95241bf-73f2-0310-859d-f6bbb57e9c96
This is not necessary, since userland teams' page directories also
contain the kernel mappings, and avoids unnecessary TLB flushes. To make
that possible the vm_translation_map_arch_info objects are reference
counted now.
This optimization reduces the kernel time of the Haiku build on my
machine with SMP disabled a few percent, but interestingly the total
time decreases only marginally. Haven't tested with SMP yet, but for
full impact CPU affinity would be needed.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28287 a95241bf-73f2-0310-859d-f6bbb57e9c96
that dprintf_no_syslog() can avoid sending something to the syslog.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28286 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added deferred_delete() that takes a DeferredDeletable and deletes it
asynchronously.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28285 a95241bf-73f2-0310-859d-f6bbb57e9c96
scheduler_set_thread_priority(). Setting the thread priority was the
only situation in which it was used.
* Renamed scheduler.cpp to scheduler_simple.cpp.
* The scheduler functions are no longer called directly. Instead there's
an operation vector now, which is initialized at kernel init time.
This allows for picking the most suitable scheduler for the machine
(e.g. a non-SMP scheduler on a non-SMP machine).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28262 a95241bf-73f2-0310-859d-f6bbb57e9c96
scheduler tracing and scheduler analysis code into separate source
files.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28253 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Turned the checks for all those macros to "#if"s instead of "#ifdef"s.
* Introduced macro KDEBUG_LEVEL which serves as a master setting.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28248 a95241bf-73f2-0310-859d-f6bbb57e9c96
index into the sLastCaller array is vint32, so after overflowing the
modulo operation would yield negative indices. This would cause the
256 bytes before the array to be overwritten. Might also be the cause of
#2866.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28245 a95241bf-73f2-0310-859d-f6bbb57e9c96
added vm_memcpy_from_physical() and vm_memcpy_physical_page(), and
added respective functions to the vm_translation_map operations. The
architecture specific implementation can now decide how to implement
them most efficiently. Added generic implementations that can be used,
though.
* Changed vm_{get,put}_physical_page(). The former no longer accepts
flags (the only flag PHYSICAL_PAGE_DONT_WAIT wasn't needed anymore).
Instead it returns an implementation-specific handle that has to be
passed to the latter. Added vm_{get,put}_physical_page_current_cpu()
and *_debug() variants, that work only for the current CPU,
respectively when in the kernel debugger. Also adjusted the
vm_translation_map operations accordingly.
* Made consequent use of the physical memory operations in the source
tree.
* Also adjusted the m68k and ppc implementations with respect to the
vm_translation_map operation changes, but they are probably broken,
nevertheless.
* For x86 the generic physical page mapper isn't used anymore. It is
suboptimal in any case. For systems with small memory it is too much
overhead, since one can just map the complete physical memory (that's
not done yet, though). For systems with large memory it counteracts
the VM strategy to reuse the least recently used pages. Since those
pages will most likely not be mapped by the page mapper anymore, it
will keep remapping chunks. This was also the reason why building
Haiku in Haiku was significantly faster with only 256 MB RAM (since
that much could be kept mapped all the time).
Now we're using a different strategy: We have small pools of virtual
page slots per CPU that are used for the physical page operations
(memset_physical(), memcpy_*_physical()) with CPU-pinned thread.
Furthermore we have four slots per translation map, which are used to
map page tables.
These changes speed up the Haiku image build in Haiku significantly. On
my Core2 Duo 2.2 GHz 2 GB machine about 40% to 20 min 40 s (KDEBUG
disabled, block cache debug disabled). Still more than factor 3 slower
than FreeBSD and Linux, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28244 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Introduced flag "invoke_scheduler" in the per CPU structure. It is
evaluated in hardware_interrupt() (x86 only ATM).
* Introduced SMP_MSG_RESCHEDULE_IF_IDLE message, which enters the
scheduler when the CPU currently runs an idle thread.
* Don't do dprintf() "CPU x halted!" when handling a SMP_MSG_CPU_HALT
ICI message. It uses nested spinlocks and could thus potentially
deadlock itself (acquire_spinlock() processes ICI messages, so it
could already hold one of the locks). This is a pretty likely scenario
on machines with more than two CPUs, but is also possible when the
panic()ing thread holds the threads spinlock. Probably fixes#2572.
* Reworked the way the kernel debugger is entered and added a "cpu"
command that allows switching the CPU once in KDL. It is thus possible
to get a stack trace of the thread not on the panic()ing CPU.
* When a thread is added to the run queue, we do now check, if another
CPU is idle and ask it to reschedule, if it is. Before this change, the
CPU was continuing to idle until the quantum of the idle thread
expired. Speeds up the libbe.so build about 8% on my machine (haven't
tested the full Haiku image build yet).
* When spinlock debugging is enabled (DEBUG_SPINLOCKS) we also record
the spinlock acquirer on non-smp machines. Added "spinlock" debugger
command to get the info.
* Added debugger commands "ici" and "ici_message", printing info on
pending ICI message respectively on a given one.
* Process not only a single ICI message in acquire_spinlock() and other
places, but all pending ones.
* Also process ICI messages when waiting for a free one -- avoids a
potential deadlock.
* Mask out non-existing CPUs in send_multicast_ici(). panic() instead of
just returning when there's no target CPU left.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28223 a95241bf-73f2-0310-859d-f6bbb57e9c96
* "iospace p" should only print the entries for actually existing
memory.
* Fixed output of "iospace v".
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28222 a95241bf-73f2-0310-859d-f6bbb57e9c96
{memset,memcpy_to}_physical() functions.
Mapping lots of physical pages at once as done before was an actual
problem on systems with enough RAM, as the physical page mapper can map
only 64 chunks at a time. So multiple threads could play dining
philosophers, each getting only one of two chopsticks, waiting for
another one to be freed.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28220 a95241bf-73f2-0310-859d-f6bbb57e9c96
the free queue.
* "thread" also prints the thread's I/O priority.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28218 a95241bf-73f2-0310-859d-f6bbb57e9c96
pages have to be written and how tight the memory situation is.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28198 a95241bf-73f2-0310-859d-f6bbb57e9c96
interrupted, another thread closing the other end of the pipe could
invoke thread_unblock() while the first thread already entered
mutex_lock(). This would make the first thread think it successfully
locked the mutex, without removing its (on-stack) wait entry from the
mutex queue, thus leading to crashes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28195 a95241bf-73f2-0310-859d-f6bbb57e9c96
PHYSICAL_PAGE_{NO,CAN}_WAIT into an actual flag
PHYSICAL_PAGE_DONT_WAIT.
* Pass the flags through to the chunk mapper callback.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27979 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Use the new VMCache::Read() flags parameter to directly read into the
physical page in the page fault handler instead of mapping it first.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27966 a95241bf-73f2-0310-859d-f6bbb57e9c96
* memset() is now available through the commpage.
* CPU modules can provide a model-optimized memset().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27952 a95241bf-73f2-0310-859d-f6bbb57e9c96
subdirectories. Also moved the x86 kernel arch_string.S there.
* Moved memcpy.c from src/system/libroot/posix/string into the
arch/generic subdirectory.
* Dealt with the consequences of moving things around. Affected are also
the boot loader and runtime loader builds.
Adjust the m68k and ppc parts, too, but only the x86 build is tested.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27947 a95241bf-73f2-0310-859d-f6bbb57e9c96
arch_vm_aspace_swap().
* The x86 implementation does now maintain a bit mask per
vm_translation_map_arch_info indicating on which CPUs the address
space is active. This allows flush_tmap() to avoid ICI for user
address spaces when the team isn't currently running on any other CPU.
In this context ICI is relatively expensive, particularly since we map
most pages via vm_map_page() and therefore invoke flush_tmap() pretty
much for every single page.
This optimization speeds up a "hello world" compilation about 20% on
my machine (KDEBUG turned off, freshly booted), but interestingly it
has virtually no effect on the "-j2" haiku build time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27912 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added smp_send_multicast_ici(), which sends the message to all CPUs
specified via a mask.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27910 a95241bf-73f2-0310-859d-f6bbb57e9c96
timed out thread has a higher priority than the currently running one.
Maybe we should even restrict this behavior to realtime threads.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27897 a95241bf-73f2-0310-859d-f6bbb57e9c96
need to cast explicitly before the multiplication/shift, since the
former is 32 bit and the latter 64 bit. The worst instance was in
swap_file_add(), where the page count was int32, so that swap file
sizes between 2 and 4 GB resulted in a negative available swap space
size. Fixes bug #2721.
* Fixed and added optional debug output.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27877 a95241bf-73f2-0310-859d-f6bbb57e9c96
HAIKU[_KERNEL]_WARNING_CCFLAGS.
* Added -Werror for various parts of the source tree for the gcc 2
build. Feel encouraged to add more.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27842 a95241bf-73f2-0310-859d-f6bbb57e9c96
Currently it only contains KDEBUG and the block cache debugging macros.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27816 a95241bf-73f2-0310-859d-f6bbb57e9c96
valgrind's callgrind format in the given directory. Those can be
analyzed with graphical tools like KCachegrind.
Recursive functions are probably not handled correctly yet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27804 a95241bf-73f2-0310-859d-f6bbb57e9c96
GetHitImages() of base class AbstractThreadProfileResult.
* gcc 4 build fixes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27797 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Moved the BasicThreadProfileResult class, its subclasses and related
code into separate files.
* Made ThreadImage an abstract base class. Pulled the meaty part into
new subclass BasicThreadImage.
* Templatized AbstractThreadProfileResult over the ThreadImage type.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27788 a95241bf-73f2-0310-859d-f6bbb57e9c96
AbstractThreadProfileResult. The latter only manages the images, now.
Moved the AddSamples() code into BasicThreadProfileResult subclasses that
handle it depending on whether the function hits are counted inclusively
or exclusively ("-f" option).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27782 a95241bf-73f2-0310-859d-f6bbb57e9c96
thread management. There's now a ThreadProfileResult (abstract base
class) object associated with a Thread object. Currently there's only
one (currently misnamed) derived class AbstractThreadProfileResult, but
some more refactoring will make the purpose clearer.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27780 a95241bf-73f2-0310-859d-f6bbb57e9c96
buffer format, the buffer can now also contain other events than just
stack traces. ATM these are only references to the image events
(created/deleted). Therefore we no longer have to flush the profiling
buffer after such an event, since the debugger can exactly match the
samples. Since we couldn't flush when the profiling timer hit while the
thread was in the kernel, that wasn't working that well anyway.
"profile -f" fails to translate stack trace addresses only very rarely,
now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27775 a95241bf-73f2-0310-859d-f6bbb57e9c96
can load the image files, we can still do symbol lookups just fine.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27774 a95241bf-73f2-0310-859d-f6bbb57e9c96
exec*()ed, the debugger never got notified that a runtime loader image
was created.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27773 a95241bf-73f2-0310-859d-f6bbb57e9c96
using it for something else screws stack traces while being in such a
function.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27772 a95241bf-73f2-0310-859d-f6bbb57e9c96
and thus, with a multiple jobs build, jobs.c could be compiled before
the header was generated.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27763 a95241bf-73f2-0310-859d-f6bbb57e9c96
case the cd_driver_info::io_scheduler field is not initialized and we
shouldn't try to access it. No idea why, but this only crashed with
the gcc4 build.
* Fixed warning with tracing enabled.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27762 a95241bf-73f2-0310-859d-f6bbb57e9c96
up module image paths in module_init_post_boot_device(). Not sure whether
it also could cause #2776.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27760 a95241bf-73f2-0310-859d-f6bbb57e9c96
* No longer send a SIGKILL when a page fault occurred, there was no
signal handler, and the debugger told us to continue as usual. Instead
we send a SIGSEGV. Instead in handle_signal() when not in the main
thread and there's no handler for the signal, we first send the main
thread a SIGKILL before letting the thread die.
So in cases where the main thread caused an unhandled page fault, the
team will die from the SIGSEGV, now. This fixes bug #2773.
* For the other hardware-caused signals we do now do the same as in case
of page faults, i.e. we first check whether the thread has a handler
for the signal in question. If so, we don't notify the debugger, but
send the signal right away.
* B_GENERAL_PROTECTION_FAULT is translated to SIGILL now. Seems better
than SIGKILL.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27757 a95241bf-73f2-0310-859d-f6bbb57e9c96
letting the boot loader provide full paths for the pre-loaded images.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27753 a95241bf-73f2-0310-859d-f6bbb57e9c96
been mounted, before anyone could try to load any modules from it.
Also pass it a flag whether the boot volume is where the boot loader
pre-loaded the modules from.
* module_init_post_boot_device() changes the pre-loaded module image
paths to normalized boot volume paths, now. Got rid of the code in
register_preloaded_module_image() which tried something like this.
* Changed module image ref counting. A referenced module has single
reference to its image, which is released when the module becomes
unreferenced.
* get_module() for a reference module will not try to re-get and re-set
the module's image anymore. That could lead to a similar module (from
different paths) being loaded at the same time. A module from a new
file can only be loaded when the old one has been put completely.
* Simplified B_KEEP_ALIVE module handling a bit. When the module is
initialized, we add another reference, which we'll never free. Thus
the module remains loaded without special handling. Removed
module_image::keep_loaded. A B_KEEP_ALIVE module remains referenced
and thus its image remains referenced, too.
* Removed module::file, a cached path to the module's image. An
optimization that wouldn't work with multiple root directories for
modules (/boot/beos/..., /boot/common/...) or when module files were
moved. get_module() does now always search the image file, when the
module is still unreferenced. This should be a bit slower than before,
but I didn't notice any difference in VMware at least. If it turns out
to be a problem we could introduce a more intelligent cache that stays
up to date by using node monitoring.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27752 a95241bf-73f2-0310-859d-f6bbb57e9c96
function (normalize_path()), now. There was some code duplication
before.
* Added "bool traverseLink" parameter to vfs_normalize_path(). When
true and the leaf component is a symlink, it will be resolved.
* KPath:
- Added similar leaf link traversal parameter to SetTo() and
SetPath().
- Added Normalize().
- Added DetachBuffer(), which returns the object's current buffer and
unsets itself.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27751 a95241bf-73f2-0310-859d-f6bbb57e9c96
code (like found in Firefox) that relies on get_next_image_info() to
return it first.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27732 a95241bf-73f2-0310-859d-f6bbb57e9c96
does overflow when the image ends at the end of the address space (like
the commpage).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27723 a95241bf-73f2-0310-859d-f6bbb57e9c96
Ingo Weinhold