* packagefs_mount(): Initialize the fs_volume earlier, so it is more
usuable in Volume::Mount().
* The new mount parameter "shine-through" can be used to specify which
shine-through mode shall be used. Can be "system", "common", "home",
and "none". Depending on the setting it is decided which directories
of the underlying file system are bind-mounted on top of ours.
* Fix infinite loop in Volume::_RemovePackageContent().
* Use the RETURN_ERROR() macro in more places to help with debugging.
* Add support function vfs_get_mount_point(), so a file system can get
its own mount point (i.e. the node it covers). Re-added
fs_mount::covers_vnode for that purpose -- the root node isn't know to
the VFS before the mount() hook returns.
* Add function vfs_bind_mount_directory() which bind-mounts a directory
to another. The Vnode::covers/covered_by mechanism is used, so this
isn't true bind-mounting, but sufficient for what we need ATM and
cheaper as well. The vnodes connected thus aren't tracked yet, which
is needed for undoing the connection when unmounting.
* get_vnode_name(): Don't use dir_read() to read the directory. Since we
have already resolved vnode to the covered vnode, we don't want the
dirents to be "fixed" to refer to the covering nodes. Such a vnode
simply wouldn't be found.
* Introduce Vnode flags for covered and covering. Can be used as a quick
check when one doesn't already hold sVnodeLock.
* Rename resolve_mount_point_to_volume_root() to
resolve_vnode_to_covering_vnode().
* Adjust all code that deals with transitions between mount points and
volume root vnodes to generally support covered/covering vnodes.
After failing to start the input server by signature, the fallback
didn't append the input server name to the servers directory returned by
find_directory().
Pull AttributeDataReader and FDDataReader implementations out of
DataReader.cpp into own source files. Thus we can avoid dependencies
(e.g. to fs_attr code) we don't need/want.
On a non-Haiku build platform map openat(), fstat(), and
FileDescriptorCloser to _kern_open(), _kern_read_stat(), and
BuildFileDescriptorCloser respectively, so symlinks can be opened and
stat()ed.
PackageWriterImpl:
* Iterate through attributes using fs_read_attr_dir() instead of
readdir(). Makes it work correctly on the build platform.
* On the build platform look up the system licenses in their source
directory rather than based on find_directory().
Since USES_BE_API is set by default on all libroot_build sources, the
error.cpp
was broken, since it wouldn't be exempt from the error mapping. Define
the
BUILDING_HAIKU_ERROR_MAPPER for it directly in the Jamfile, now.
For the xattr/BSD (untyped) attribute backend implement fs_fopen_attr()
and fs_close_attr(). A new AttributeDescriptor is created. It is
currently used in write_pos() only.
* Resolve TODO: HOST_GCC_BASE_FLAGS should not be included in
HOST_LDFLAGS. Enable adding "-fno-strict-aliasing -fno-tree-vrp"
accordingly.
* Fix handling of HOST_PLATFORM_IS_64_BIT and HAIKU_HOST_USE_32BIT: The
former does now state whether the platform is effectively treated as
64 bit platform, i.e. it actually is 64 bit and the 32 bit mode is
not enforced. HAIKU_HOST_USE_32BIT is now only set when the platform
is actually 64 bit, but 32 bit mode is enforced.
* Map build variables HOST_CPU and HOST_ARCH to x86_64, if it they are
* x86 and
64 bit and define the __x86_64__ C macro instead of __INTEL__ in that
case.
* <OS.h>: Also handle __x86_64__.
* The sa_flags field of the structure passed to sigaction() wasn't
cleared, which resulted in the wrong type of handler function being
called and the second parameter not being the expected user data.
* Always check fCaughtDeadlySignal in the main loop, so the signal
doesn't have to hit the _kern_system_profiler_next_buffer() call
(though that's the most likely one anyway).
Now scheduling_recorder can be stopped with a single Ctrl-C again.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43219 a95241bf-73f2-0310-859d-f6bbb57e9c96
They were too small for all the fields added. This is why the system
profiler skipped "thread added" notifications.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43216 a95241bf-73f2-0310-859d-f6bbb57e9c96
* When encountering scheduling data referring to an unknown thread, add
a dummy thread for it. This prevents a crash when processing the
per-thread events later.
* Make warning output more uniform.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43214 a95241bf-73f2-0310-859d-f6bbb57e9c96
Allows to associate untyped data with the model, which the model free()s
upon destruction.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43213 a95241bf-73f2-0310-859d-f6bbb57e9c96
get_iframe_registers() in arch_user_debugger.cpp return the correct
eax/edx values also when the kernel was entered via an exception that
doesn't push an error code (e.g. page fault).
Fixes incorrect eax and edx values shown in Debugger and variable values
based on them.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43201 a95241bf-73f2-0310-859d-f6bbb57e9c96
influencing this script.
* If haikuRequiredLegacyGCCVersion has not been specified, assume that
the script was called manually and just use the version string from
the gcc sources.
* Make error output a bit more consistent.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43163 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Turn VMCache::consumers C list into a DoublyLinkedList.
* Use object caches for the different VMCache types and the VMCacheRefs.
The purpose is to reduce slab area fragmentation.
* Requires the introduction of a pure virtual VMCache::DeleteObject()
method, implemented in the derived classes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43133 a95241bf-73f2-0310-859d-f6bbb57e9c96
template function object_cache_delete() to be used to delete objects
constructed with it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43132 a95241bf-73f2-0310-859d-f6bbb57e9c96
* AllocationTrackingCallback::ProcessTrackingInfo(): Also pass the
allocation pointer.
* "allocations_per_caller" KDL command: Add option "-d". When given,
each allocation for the specified caller is printed, including the
respective stack trace.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43087 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Add TraceOutput::PrintArgs(), a va_list version of Print().
* Move code of TraceOutput::Print() to new private template function
print_stack_trace().
* Add public tracing_print_stack_trace().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43085 a95241bf-73f2-0310-859d-f6bbb57e9c96
Refactor the "allocations_per_caller" KDL command related functions.
They expect an instance of a class implementing the new
AllocationTrackingCallback interface, now. The only implementation ATM
is AllocationCollectorCallback, which does the work the now removed
slab_debug_add_allocation_for_caller() did before.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43082 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Fix build broken in r43078. The slab_debug_add_allocation_for_caller()
wasn't guarded correctly.
* slab_debug_add_allocation_for_caller(): Add bool resetAllocationInfos
parameter, which makes the function clear the allocation tracking
infos after processing the data.
* "allocations_per_caller" KDL command: Add option "-r" to reset the
allocation tracking infos. The next invocation of the command will
only show the allocations made after the reset.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43079 a95241bf-73f2-0310-859d-f6bbb57e9c96
* dump_allocations_per_caller(): Compute the total allocation count and
size from the caller infos instead of using return arguments in the
helper functions called.
* Move caller info update code from analyze_allocation_callers() to new
function slab_debug_add_allocation_for_caller(), so it can be reused.
* Add MemoryManager::AnalyzeAllocationCallers() to collect the
allocation information for the memory manager.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43078 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Add optional stack trace capturing for slab memory manager tracing.
* Add allocation tracking for the slab allocator (enabled via
SLAB_ALLOCATION_TRACKING). The allocation tracking requires tracing
with stack traces to be enabled for object caches and/or the memory
manager.
- Add class AllocationTrackingInfo that associates an allocation with
its respective tracing entry. The structure is added to the end of
an allocation done by the memory manager. For the object caches
there's a separate array for each slab.
- Add code range markers to the slab code, so that the first caller
into the slab code can be retrieved from the stack traces.
- Add KDL command "allocations_per_caller" that lists all allocations
summarized by caller.
* Move debug definitions from slab_private.h to slab_debug.h.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43072 a95241bf-73f2-0310-859d-f6bbb57e9c96
Add helper macros for placing markers in the source, so we can get the
address ranges of code we're interested in.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43071 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Move struct tracing_stack_trace to tracing.h header.
* Add tracing_find_caller_in_stack_trace(). Helper function to get the
first return address of a stack trace that is not in one of the given
address ranges.
* Add AbstractTracingEntryWithStackTrace::StackTrace() getter.
* Add tracing_is_entry_valid(). Checks, based on the additionally given
time, whether a tracing entry is (probably) still in the tracing
buffer.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@43070 a95241bf-73f2-0310-859d-f6bbb57e9c96
get clashes between target names e.g. between the "common/include/" for OpenSSL
and the one for CLucene. Should fix#7798.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42402 a95241bf-73f2-0310-859d-f6bbb57e9c96
introduced in r37701 was not taken into account for computing the short meta
chunk size. Fixes the reopened#6237.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42385 a95241bf-73f2-0310-859d-f6bbb57e9c96
specified the syscall error code B_WOULD_BLOCK would no longer be mapped to 0.
Fixes#7693.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42188 a95241bf-73f2-0310-859d-f6bbb57e9c96
most headers refer to can be found. Fixes the build under Haiku (ticket #7688).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42187 a95241bf-73f2-0310-859d-f6bbb57e9c96
arguments and environment. Fixes failures to start programs in case the space
available through rounding to full page sizes wasn't sufficient.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42184 a95241bf-73f2-0310-859d-f6bbb57e9c96
exit info with some generic status.
* team_create_thread_start(), common_thread_entry(): Initializes the team's
exit info (if that's the main thread) before calling thread_exit(). Fixes
#7686.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42183 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Only try to cancel the a transfer, if sUseUSBKeyboard is set.
* Set the "_usbPipe" parameter.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42181 a95241bf-73f2-0310-859d-f6bbb57e9c96
* libbe_build: Where possible we directly use the actual Haiku headers and
sources, now. In the headers/build headers we just include the respective
Haiku headers as needed. That still allows overrides where necessary. The
intention is to make it easier to keep the build stuff in sync.
* Fixed a few printf() format and signed/unsigned comparison warnings.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42179 a95241bf-73f2-0310-859d-f6bbb57e9c96
HAIKU_TARGET_PLATFORM_HAIKU, so the file can also be compiled for the host
platform.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42178 a95241bf-73f2-0310-859d-f6bbb57e9c96
[u]int32 and [u]int64 to avoid clashes with the int/unsigned int versions.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42165 a95241bf-73f2-0310-859d-f6bbb57e9c96
before assuming we have data. If the I/O port isn't valid (e.g. because there
is no serial port) 0xff seems to be a typical value to read. In that case fail.
Fixes KDL input on machines without serial port -- kgetc() would always think
it read something from the serial port, never trying any other input.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42153 a95241bf-73f2-0310-859d-f6bbb57e9c96
partition_module_info::uninitialize().
* Implemented the hook for BFS.
* Implemented KFileSystem::Uninitialize().
Fixes failure to initialize a BFS initialized device with an intel partition
map.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42142 a95241bf-73f2-0310-859d-f6bbb57e9c96
destroy the partitioning system's on-disk structure.
* Adjusted the existing partitioning system implementations accordingly.
Actually implemented the hook for the intel partitioning system.
* Added Uninitialize() method to KDiskSystem and KPartitioningSystem. The latter
implements the method calling the new module hook.
* _user_uninitialize_partition(): Also let the disk system uninitialize the
on-disk structure.
This fixes the failure to initialize a disk device with BFS, when it contains a
valid partition map with at least one partition.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42140 a95241bf-73f2-0310-859d-f6bbb57e9c96
disk system, if the partition already has a disk system set. This can happen
when the disk system's initialization function is lazy and just lets the DDM
rescan the partition. In bad cases the previous disk system has a higher
priority than the new one and, if its on-disk structures have not been
destroyed, it will win the rescan. Not setting the disk system in such cases at
least leaves the partition object in a consistent state.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42139 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Don't fall back to the disk device, if no partition was selected. No idea why
that was done. It led to operations (like initialize) being enabled that
would fail, if invoked.
* Don't check CanInitialize() on the selected partition with its current disk
system to determine whether to enable the initialization menu. No idea why
that was done either. It led to the menu being disabled, when there was no
reason for it.
* Also disable the initialization menu when no partition is selected.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42135 a95241bf-73f2-0310-859d-f6bbb57e9c96
its size.
* Added "Display current boot loader log" item to the "Debug Options" boot
loader menu. It displays what the boot loader has logged so far. Might be
interesting for early boot issues when serial debugging is not possible.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42134 a95241bf-73f2-0310-859d-f6bbb57e9c96
which don't wait for a character, but return -1 when no character is
available ATM. Implemented correctly for x86 only.
* Changed the semantics of the debugger_module_info::debugger_getchar() hook.
It is supposed to return immediately now.
* Adjusted usb_keyboard accordingly. Hacked UHCI's debug_process_transfer() to
achieve that. It does now start, check, or cancel a transfer. Split
UHCI::ProcessDebugTransfer() into StartDebugTransfer(), and
CheckDebugTransfer() accordingly, and also added a CancelDebugTransfer().
The latter seems to have issues. Michael, please have a look. I have no clue
what I'm doing. :-)
* Adjusted kgetc() to poll all possible inputs using the new
functions/semantics. This allows to use any input (USB, PS/2, serial) in KDL.
* Removed the no longer needed "serial_input" command.
* read_line(): Also support 0x7f as backspace code. That's what xterm sends.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42126 a95241bf-73f2-0310-859d-f6bbb57e9c96
trunk. The interesting part of r42122 was already fixed in the trunk, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42124 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Reorganized the kernel locking related to threads and teams.
* We now discriminate correctly between process and thread signals. Signal
handlers have been moved to teams. Fixes#5679.
* Implemented real-time signal support, including signal queuing, SA_SIGINFO
support, sigqueue(), sigwaitinfo(), sigtimedwait(), waitid(), and the addition
of the real-time signal range. Closes#1935 and #2695.
* Gave SIGBUS a separate signal number. Fixes#6704.
* Implemented <time.h> clock and timer support, and fixed/completed alarm() and
[set]itimer(). Closes#5682.
* Implemented support for thread cancellation. Closes#5686.
* Moved send_signal() from <signal.h> to <OS.h>. Fixes#7554.
* Lots over smaller more or less related changes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42116 a95241bf-73f2-0310-859d-f6bbb57e9c96
exceptions (page faults and the like). The handler dumps possibly interesting
information (registers, a (numerical) stack trace) to the serial output, and,
if possible, also to the screen. That should help debugging boot loader
crashes.
* For the IDT the boot loader sets up for the kernel the descriptors are set up
the same way, so until the kernel initializes the IDT itself (arch_init()) the
facility is still in place and can thus catch very early kernel boot crashes.
Unfortunately the on-screen output doesn't seem to work anymore at that point,
so the output only goes to the serial port...
* ... and to the debug syslog -- dprintf() does now append it there after
debug_cleanup() has been called. Seems to work fine in qemu, but when I tested
it on real hardware the debug syslog was gone.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42092 a95241bf-73f2-0310-859d-f6bbb57e9c96
Ingo Weinhold