its death entry into the parent team's queue - we need to do this atomically.
As an effect, wait_for_thread() sometimes failed with B_BAD_THREAD_ID.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19808 a95241bf-73f2-0310-859d-f6bbb57e9c96
so that this can be done safely.
It was also needed, as it would call vm_cache_release_ref() on failure which requires you
to have no vm_cache_ref locks around (as it might deadlock in this case).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19806 a95241bf-73f2-0310-859d-f6bbb57e9c96
had the cache_ref locked, it also locked two refs in the wrong order (bottom-up);
there was even a TODO item for this...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19801 a95241bf-73f2-0310-859d-f6bbb57e9c96
* "sc"/"where"/"bt" now prints the area where the function of the stack frame
is located in case there is no other information (using the above function).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19800 a95241bf-73f2-0310-859d-f6bbb57e9c96
* even worse, in case of fork(), it was never initialized.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19799 a95241bf-73f2-0310-859d-f6bbb57e9c96
two cache_refs - it needs to count the consumers of the lower cache to find
its actual number of references; the upper cache could still be in use by
someone else.
* There were several locking bugs in the VM code; since cache_ref::cache can
change, we must not access it without having the cache_ref locked.
* As a result, map_backing_store() now requires you to have the lock of the
store's cache_ref held.
* And therefore, some functions in vm_cache.c must no longer lock the cache_ref
on their own, but require the caller to have it locked already.
* Added the -s option to the cache/cache_ref KDL commands: it will only print
the requested structure, and not its counterpart (useful if accessing one
structure results in a page fault, as was possible previously).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19796 a95241bf-73f2-0310-859d-f6bbb57e9c96
waiting for in case it was interrupted; but that could easily lead the thread_exit()
function to access invalid memory (and thus, crash the kernel): since we could not
remove our death entry from the thread, we have to make sure the thread (which might
still run even if not in the thread hash anymore) will access our death entry as
long as it is valid. IOW we must wait for the thread to delete its exit semaphore,
even if we were interrupted before.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19790 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Fixed dumping the area list of a cache I broke with the previous commit.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19789 a95241bf-73f2-0310-859d-f6bbb57e9c96
(you can still use both commands, but you'll see always the same output).
* The cache_ref's area list now also prints the owner of the area.
* Added "-p" option to "cache"/"cache_ref" that will show the pages of the cache; if you
omit it, it will now only present you a page counter.
* Nicer output for the commands above.
* Added "dl" to display memory in 64 bit values.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19787 a95241bf-73f2-0310-859d-f6bbb57e9c96
* send_signal_etc() (among others) used team::main_thread in an unsafe way; it is
possible for a team to be part of the team list and a group without having
a main thread very early in its creation process.
* Replaced most occurences of team->main_thread->id with team->id since the team
always inherits the ID of its main thread in Haiku for quite some time now.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19781 a95241bf-73f2-0310-859d-f6bbb57e9c96
the userland stack in an unsafe way - moved that stuff to arch_thread_enter_userspace(), too.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19778 a95241bf-73f2-0310-859d-f6bbb57e9c96
* This is now used for userland stack - they now always pre-commit two pages, enough
to initialize TLS and copy the user-thread-exit stub to that area.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19776 a95241bf-73f2-0310-859d-f6bbb57e9c96
turned off - accessing userland memory. Now, arch_thread_enter_userspace() does that
job, and as a result, may also fail.
* dump_thread() now directly prints the info of the current thread when used without
argument (rather than iterating the thread list to look for the current thread).
* If arch_thread_init_tls() fails upon thread creation, the function will now return
an error.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19775 a95241bf-73f2-0310-859d-f6bbb57e9c96
them to commit substantially less memory (we we're committing about 40 MB (!) too much
after a complete system boot). This means you'll run out of memory less likely now.
* fill_area_info() no longer filters out kernel protection flags - we may want to keep
filtering them when called from userland, though, dunno.
* Added new debugger command "avail" which shows how much memory has been committed, and
how much is regarded as free space.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19771 a95241bf-73f2-0310-859d-f6bbb57e9c96
to other kernel components.
* wait_for_thread_etc() will now search the team's death entries in case the
thread is already gone; also resume_thread() is now done later, and its return
code will no longer matter (as we already have our death entry, no matter if
the thread is gone now or not).
* The fibo_load_image test now works as expected (only tested with low numbers
yet, though - the mean testing comes later (first comes functionality) :-))
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19758 a95241bf-73f2-0310-859d-f6bbb57e9c96
with a lower priority than the current one.
* Further reduced the probability to skip a thread to roughly 4%, that makes around
13 skips per second, and about 40 msecs spend in lower priority threads per second
(with a 3 msec quantum). Might still be too much, but we'll see.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19719 a95241bf-73f2-0310-859d-f6bbb57e9c96
not a penalty to be the first thread in the run queue...
I've reduced it to about 12.5%, but that might still be too much.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19710 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Now using the new %f flag to print the level of support.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19584 a95241bf-73f2-0310-859d-f6bbb57e9c96
a very simplistic floating point output routine to snprintf() and friends, for
your convenience.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19583 a95241bf-73f2-0310-859d-f6bbb57e9c96
* vm_soft_fault() did not take into account that a cache's source can change while
traversing a cache chain.
* Now, we grab a reference to every cache we get before locking it, and
* no longer get the cache's source without having the cache locked.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19571 a95241bf-73f2-0310-859d-f6bbb57e9c96
for an extra check in vm_cache_release_ref() as the cache shouldn't have any areas
or consumers at this point.
* Fixed a locking problem in vm_cache_remove_consumer(): the cache was acquired after
its lock was gone, so someone else might have released in the mean time.
* if the cache's source is to be replaced, we now no longer release its lock after
having merged it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19570 a95241bf-73f2-0310-859d-f6bbb57e9c96
* the member will now also be shown when dumping a team.
* minor cleanup: moved fill_team_info() into the private functions part of the file.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19551 a95241bf-73f2-0310-859d-f6bbb57e9c96
-Turns out the area removal routine had a massive race condition inside
vm_put_area(). Basically the area was removed from the address space's
area list before the pages were unmapped, so the vm could (and would)
recycle the space before the pages were finally unmapped.
It was completely reproducable on my machine during initialization of a bunch
of storage drivers that were bringing the locked_pool module into and out of
existence, which caused a thread to be spawned and stopped in rapid sucession.
On a dual processor machine, it was possible for the new thread to be started
up while the old one was still shutting down, and the kernel stack of the new
one would get wiped out.
Note, there still is a page ref counting problem with this area removal code.
It doesn't decrement the ref count of the page as it unmaps it. Will have to
figure that out.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19549 a95241bf-73f2-0310-859d-f6bbb57e9c96
put a spinlock around the serial debug routines to keep multiple
cpus from interleaving their output.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19548 a95241bf-73f2-0310-859d-f6bbb57e9c96
commented the insertion of the attribute name in patterns in the case of a string attribute
notify_probe_by_file chooses a module based on a bus specific suffix
dm_register_child_device has a parameter to optionally check the support for the node
added scanning of bus devices after the boot filesystem is mounted
fixed dm_rescan, locking was misbehaving
fixed SYSTEM_DRIVER_REGISTRATION definition
added B_DRIVER_MAPPING attributes for PCI and ACPI devices:
%vendor%_%device% for PCI, hid_%hid% and type_%type% for ACPI
moved acpi_device_module_info definition to public ACPI.h
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19394 a95241bf-73f2-0310-859d-f6bbb57e9c96
added a user friendly mode for listdev: it works for PCI
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19262 a95241bf-73f2-0310-859d-f6bbb57e9c96
* added a generic syscall for device_manager
it enables to iterate the device manager tree from userland
* the listdev tool is now using it: it's still incomplete as it only dumps nodes and attributes
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19260 a95241bf-73f2-0310-859d-f6bbb57e9c96
This fixes bug #227 by closing that memory and semaphore leak.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19041 a95241bf-73f2-0310-859d-f6bbb57e9c96
address space of reserved areas - IOW address spaces were never freed upon
team exit.
* dump_cache() now prints a list of the cache's consumers.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19040 a95241bf-73f2-0310-859d-f6bbb57e9c96
* a vm_cache now maintains a list of its "consumer" caches.
* introduced to new functions that add/remove consumer to a cache (instead
of only maintaining the vm_cache::source field).
* fixed the incorrect reference counting when doing copy-on-write; we kept
one ref too many of the lower cache.
* minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19035 a95241bf-73f2-0310-859d-f6bbb57e9c96
This also fixes the issue of exporting a C++ API from the kernel.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18974 a95241bf-73f2-0310-859d-f6bbb57e9c96
of a team to avoid confusion. It now also accepts a NULL session pointer in
which case the actual group's session doesn't matter.
* Fixed the race condition in send_signal_etc() that could allow accessing an
invalid team pointer.
* Jerome's earlier change already fixed bug #841, and this also fixed bug #855.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18941 a95241bf-73f2-0310-859d-f6bbb57e9c96
move a 16 bit segment register into 32 bit memory location, and movzx (called movzwl within gcc) doesn't work with
segement registers.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18864 a95241bf-73f2-0310-859d-f6bbb57e9c96
the next step would be to rescan the partition tree with a job to recognize unrecognized partitions (asynchronously ?)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18737 a95241bf-73f2-0310-859d-f6bbb57e9c96
parameter during construction).
* Doing so will now result in a kernel panic whenever your file system tries to
write to a block.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18719 a95241bf-73f2-0310-859d-f6bbb57e9c96
will now grab a reference to the vnode as well if successful. This way, vfs_get_vnode_cache()
now actually works how it should: it will now always grab a reference to the cache and
its underlying vnode. This removes an extra reference to the vnode (and vm_cache) that
got ignored before and prevented volumes to be unmounted (or file caches to be removed).
Thanks to Korli for pointing this out.
* file_cache_create() is now aware of that extra vnode reference and releases it; unmounting
volumes is now working again as it should.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18716 a95241bf-73f2-0310-859d-f6bbb57e9c96
of waiting threads by ID (instead of just the queue head and tail pointers).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18688 a95241bf-73f2-0310-859d-f6bbb57e9c96
(that was part of the problem of bug #702).
* Fixed send_signal_etc() when you called it with a pid_t of zero: the signals should
go to all teams in the calling team's group, not only to the team (for -1, we do
the same for now).
* Made team_get_process_group_locked() public, and rewrote send_signal_etc() to use
it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18684 a95241bf-73f2-0310-859d-f6bbb57e9c96
some signals could go to some group when a negative value was passed in.
This actually fixes bug #702 where SoundPlay obviously does a kill_thread(-1)
at the end (which accidently killed all userspace applications, including the
input_server, app_server, ...).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18683 a95241bf-73f2-0310-859d-f6bbb57e9c96
hash. This also allows them to stay valid after the group leader died when there are
other teams left in it. This closes bug #1.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18622 a95241bf-73f2-0310-859d-f6bbb57e9c96
alarm, or 0 if there isn't any.
* setitimer() now also sets the previous timer values correctly, so that
alarm() and ualarm() now return the correct values as well - this fixes
the test fork_9-1 failing at alarm().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18545 a95241bf-73f2-0310-859d-f6bbb57e9c96
a waitpid() can no longer succeed; this fixes the occasional fork 3-1 test failures.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18544 a95241bf-73f2-0310-859d-f6bbb57e9c96
it waited for a specific child), as B_RELEASE_ALL opened up a race condition between
looking for an existing death entry, and waiting for the dead children semaphore.
Now we're counting all waiting threads for teams and groups separately.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18543 a95241bf-73f2-0310-859d-f6bbb57e9c96
wait_for_child() hanging less probable - there is a general problem with
this code, though, as we need to have a dedicated free counter for the
semaphore to remove all race conditions.
* Also, test fork_3-1 still sometimes fails because the thread is still
available to the public for a short time, even after its death entry
has been collected.
* Added a TODO in the code for these issues.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18529 a95241bf-73f2-0310-859d-f6bbb57e9c96
wanted to wait for a specific child; it always assumed you had specified WNOHANG.
This fixes the bug the fork_3-1 and fork_4-1 test applications reported. 3-1 still
sometimes fails, but that's a different problem (to be solved later).
* Also, it could return B_BAD_THREAD_ID instead of the expected ECHILD (for waitpid()).
* There was a race condition between testing for a thread, and checking its death
entry.
* wait_for_child() can now be interrupted in case it has to wait.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18528 a95241bf-73f2-0310-859d-f6bbb57e9c96
Fixed the build of most of targets using these rules. Though the build can be still broken, feel free to fix.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18521 a95241bf-73f2-0310-859d-f6bbb57e9c96
buffer to allow safe access of the user provided string - maybe we should
introduce a user_strdup() instead.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17960 a95241bf-73f2-0310-859d-f6bbb57e9c96
files, too.
* Added a temporary icon for the kernel until Stephan comes up with a better
one (hint hint!) :-))
* This even fixes bug #648.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17825 a95241bf-73f2-0310-859d-f6bbb57e9c96
arch dependent code (they will be removed as soon as someone else
asks for these interrupt lines).
* Added an interrupt driven keyboard handler to the kernel that uses
this technique. As a result, you can now press F12 to enter the kernel
debugger before the input_server has been started, and Control-Alt-Delete
should reboot the system (actually I did not test the latter yet).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17806 a95241bf-73f2-0310-859d-f6bbb57e9c96
vfs_get_vnode_cache() which did not acquire an extra reference to the
cache_ref when the cache had to be created.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17693 a95241bf-73f2-0310-859d-f6bbb57e9c96
mapping, but it connected the vm_cache objects, so if it failed later, and
thus called vm_cache_release_ref() the object could have been freed accidently.
Most uses of map_backing_store() explicetly acquired a cache_ref *after* the
call was successful, but _vm_map_file() did not do this.
_vm_map_file() might still not work correctly, though, need to have a closer
look at it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17692 a95241bf-73f2-0310-859d-f6bbb57e9c96
- add loading of kernel debugger/ modules
- add a kgets() exported func for use by nasty modules =)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17657 a95241bf-73f2-0310-859d-f6bbb57e9c96
let them eat death stack entries: after setting the next thread state to
THREAD_STATE_FREE_ON_RESCHED, interrupts were enabled again, which could
cause the thread to be rescheduled before having called put_death_stack().
This fixes bug #434.
* Note that the above change pretty much reverts revision 7865 that was supposed
to fix interrupt problem on thread exit (patch by Jack Burton almost 2 years
ago, that's how long this problem existed!).
* Made get_death_stack() and put_death_stack() symmetrical in that they don't
change interrupts. Also pulled out rescheduling from put_death_stack[_and_reschedule]()
and put it back into thread_exit2().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17652 a95241bf-73f2-0310-859d-f6bbb57e9c96
running team to be able to fill in the team_info::args field. Currently, only
the path is stored, there, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17646 a95241bf-73f2-0310-859d-f6bbb57e9c96
#603.
* Moved devfs_get_partition_info() into the devfs_ioctl() hook.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17622 a95241bf-73f2-0310-859d-f6bbb57e9c96
no longer try to access any MSRs when there are no MTRRs. Got it? :-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17574 a95241bf-73f2-0310-859d-f6bbb57e9c96
into more than MAX_FILE_IO_VECS pieces. This could have very weird consequences
like overwriting data outside the file (but on that same partition only).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17536 a95241bf-73f2-0310-859d-f6bbb57e9c96
reading the file map of certain files (hope I got that more or less
right)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17525 a95241bf-73f2-0310-859d-f6bbb57e9c96
* As suggested by Korli, the signal is now encoded in the "reason" field of
wait_for_child().
* waitpid() now sets the status passed in so that the signal can be read out
(but it still doesn't do it's full job).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17338 a95241bf-73f2-0310-859d-f6bbb57e9c96
of the one supplied to fs_mount() (but not the /dev/disk/virtual/... entry, as that
wouldn't be that clear to the user).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17335 a95241bf-73f2-0310-859d-f6bbb57e9c96
scan job in the calling thread).
* KDiskDeviceManager::InitialDeviceScan() now runs synchronously, so that
get_boot_partitions() doesn't need to do this ugly wait hack.
* KDiskDeviceManager::CreateFileDevice() can now run synchronously as well, which
fixes a deadlock in fs_mount() - note, mounting file devices still doesn't work,
though as Haiku's BFS doesn't allow this right now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17334 a95241bf-73f2-0310-859d-f6bbb57e9c96
* _ReadLock() will now fail with B_NO_MEMORY in case the ReadLockInfo couldn't
be created.
* Note, due to a design bug, we cannot guarantee that a previous read lock
can be reestablished after releasing a write lock in case of low memory.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17332 a95241bf-73f2-0310-859d-f6bbb57e9c96
when called with interrupts turned off, get_memory_map() will now call the new
vm_translation_map_ops::query_interrupt() call.
Under PPC, this is trivial (at least right now), but on x86 we need to make sure
we have access to the page table entry, ie. we need to create an area that points
to its own page table entry, so that we can map in the page table entry containing
the address we're looking for. It's not really nice, feel free to come up with
a cleaner solution :-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17280 a95241bf-73f2-0310-859d-f6bbb57e9c96
always call all handlers in this case, but we still try to return the correct
return code (ie. B_HANDLED_INTERRUPT and B_INVOKE_SCHEDULER).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17278 a95241bf-73f2-0310-859d-f6bbb57e9c96
the interrupt handlers, not before (like we do for edge triggered ones).
This should prevent hardware from issuing a second interrupt when the software driver
is still busy handling the first.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17276 a95241bf-73f2-0310-859d-f6bbb57e9c96
work correctly when on power supply (ie. when running at full speed). Turns out we
misdetected a spurious interrupt on line 15 - we now ignore them completely which
seems to fix that problem.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17275 a95241bf-73f2-0310-859d-f6bbb57e9c96
no longer needs to lock address space hash table - that also makes the lookup much
faster, too (and a direct pointer is used instead of a hash lookup).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17246 a95241bf-73f2-0310-859d-f6bbb57e9c96
dir_remove() did not normalize the path, and thus, could forward a "." as
name for the removed directory - which BFS didn't catch because it assumed
our VFS would work correctly...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17165 a95241bf-73f2-0310-859d-f6bbb57e9c96
vnode is not becoming unbusy (right now it even panics, but that can be removed
later on).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17161 a95241bf-73f2-0310-859d-f6bbb57e9c96
as stated in the BeBook - this fixes bug #458 as Tracker directly compared
with that error code.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17144 a95241bf-73f2-0310-859d-f6bbb57e9c96
* vfs_get_fs_node_from_path() now also work for absolute paths again (but
still for relative ones from the volume root) - it just tests if the
mount IDs fit, so it only returns successful if the path really is on
the desired mount.
* the Disk Device Manager publish functions now call devfs_publish_*()
correctly (by omitting the "/dev/" mount point).
* devfs_publish_partition() now accepts absolute device paths but relative
partition paths.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17138 a95241bf-73f2-0310-859d-f6bbb57e9c96
