recursive lock. I haven't investigated it closer, but the previous
implementation was even broken -- "strace /bin/true" showed two
release_sem() calls, but no acquire_sem().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34342 a95241bf-73f2-0310-859d-f6bbb57e9c96
as a mutex, but allocates its semaphore lazily. This comes at the cost of an
additional atomic_add() when the semaphore has actually to be acquired, but
saves the semaphore creation completely in single-threaded programs and in
any program when there's no lock contention.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34339 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Fixed check in hoardSbrk(): resize_area() was invoked, even if the area was
already large enough.
* Increased the initial heap size to 64 pages. Apparently the hoard
implementation is rather generous and the first malloc() (caused by
__init_heap()) already required enlarging the area.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34336 a95241bf-73f2-0310-859d-f6bbb57e9c96
preprocessed <syscalls.h> header in sources compiled for the build platform.
Instead we're generating macros for the return type and parameter type sizes
via the CreateAsmStructOffsetsHeader rule (no template magic anymore) and
thus get a clear separation of host and target code, resulting in better
portability (although surprisingly enough the build already worked on 64 bit
hosts).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34323 a95241bf-73f2-0310-859d-f6bbb57e9c96
architectures, there's now a arch/generic/generic_gensyscalls.h which is
included by the otherwise empty arch_gensyscalls.h files.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34320 a95241bf-73f2-0310-859d-f6bbb57e9c96
a file suitable for the CreateAsmStructOffsetsHeader rule. The file is then
used to generate the actual matypes.h header. This fixes issues when
cross-compiling from a 64 bit platform.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34312 a95241bf-73f2-0310-859d-f6bbb57e9c96
use for the asm_offsets.cpp file, so it can be reused elsewhere.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34311 a95241bf-73f2-0310-859d-f6bbb57e9c96
macros. This fixes the broken keymaps when building on a 64 bit host.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34283 a95241bf-73f2-0310-859d-f6bbb57e9c96
release branch. This completes the build system part of #5036.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34275 a95241bf-73f2-0310-859d-f6bbb57e9c96
building on Haiku, since those will be defined already -- by way of the host
platform's <BeBuild.h>, which currently is included by the POSIX headers..
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34272 a95241bf-73f2-0310-859d-f6bbb57e9c96
needed -- the code has its own complete platform abstraction layer -- and
only causes problems.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34270 a95241bf-73f2-0310-859d-f6bbb57e9c96
conditionally define some of the new stuff there. That fixes the build under
a r34172 Haiku, though it also causes some problems that needed working around.
Haven't tested under newer Haiku revisions yet. Will do tomorrow.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34267 a95241bf-73f2-0310-859d-f6bbb57e9c96
DoublyLinkedListLink. That actually seems to get the bfs_shell working
properly.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34257 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Removed AutoDeleter class and use the shared one instead.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34244 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added FSSH_[S]SIZE_MAX to headers/private/fs_shell/fssh_types.h.
* Fixed various 64 bit compiler warnings. Nothing too serious, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34241 a95241bf-73f2-0310-859d-f6bbb57e9c96
That's just leading to problems.
* Fixed various 64 bit warnings when building libbe_build.so. One of the more
serious issues, that might bite us, is that 64 bit Linux defines dev_t to
unsigned long, while Haiku code assumes that it is signed and 32 bit. We'll
see...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34227 a95241bf-73f2-0310-859d-f6bbb57e9c96
couldn't bring myself to replace the types everywhere they are used, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34215 a95241bf-73f2-0310-859d-f6bbb57e9c96
and types.h. The idea is to provide a basic architecture/compiler
abstraction by defining types and macros that allow the posix/ and os/
headers to be mostly architecture/compiler agnostic.
* Adjusted the posix/ and os/ headers accordingly.
* <SupportDefs.h>: Introduced B_PRI* and B_SCN* macros similar to the PRI*
and SCN* macros defined in <inttypes.h>, just for the BeOS/Haiku [u]int*
types and some POSIX types (e.g. off_t, dev_t, ino_t) that don't have POSIX
macros. Also the B_PRI* and B_SCN* macros are available unconditionally,
unlike the <inttypes.h> macros, which require __STDC_FORMAT_MACROS to be
defined in C++ mode.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34214 a95241bf-73f2-0310-859d-f6bbb57e9c96
* BuildSetup: Set jam variable HOST_PLATFORM_IS_64_BIT and add host define
HAIKU_HOST_PLATFORM_64_BIT if the host platform is 64 bit. Removed the
check from BeOSBuildCompatibility.h.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34211 a95241bf-73f2-0310-859d-f6bbb57e9c96
go, but for the time being at least don't cause the BFS sources to be compiled
twice regardless of which target is built.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34204 a95241bf-73f2-0310-859d-f6bbb57e9c96
build_cross_tools[_gcc4] script which in turn passes it to make. Cores and
hyperthreads are plentiful these days; no need to let most of them idle when
building the cross tools.
* Sorted the configure options alphabetically again.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34199 a95241bf-73f2-0310-859d-f6bbb57e9c96
all MTRRs at once.
* Added a respective x86_set_mtrrs() kernel function.
* x86 CPU module:
- Implemented the new hook.
- Prefixed most debug output with the CPU index. Otherwise it gets quite
confusing with multiple CPUs.
- generic_init_mtrrs(): No longer clear all MTRRs, if they are already
enabled. This lets us benefit from the BIOS's setup until we install our
own -- otherwise with caching disabled things are *really* slow.
* arch_vm.cpp: Completely rewrote the MTRR handling as the old one was not
only slow (O(2^n)), but also broken (resulting in incorrect setups (e.g.
with cachable ranges larger than requested)), and not working by design for
certain cases (subtractive setups intersecting ranges added later).
Now we maintain an array with the successfully set ranges. When a new range
is added, we recompute the complete MTRR setup as we need to. The new
algorithm analyzing the ranges has linear complexity and also handles range
base addresses with an alignment not matching the range size (e.g. a range
at address 0x1000 with size 0x2000) and joining of adjacent/overlapping
ranges of the same type.
This fixes the slow graphics on my 4 GB machine (though unfortunately the
8 MTRRs aren't enough to fully cover the complete frame buffer (about 35
pixel lines remain uncachable), but that can't be helped without rounding up
the frame buffer size, for which we don't have enough information). It might
also fix#1823.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34197 a95241bf-73f2-0310-859d-f6bbb57e9c96
This is something must must not do in an idle thread or we get the scheduler
into trouble.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34196 a95241bf-73f2-0310-859d-f6bbb57e9c96
* LeafNode/Directory: We make sure that the added PackageNode with the newest
modified time is always at the head of the list. This is the package node
whose stat, attribute data and, in case of non-directory nodes, also content
data will be returned. This results in a well-defined algorithm what happens
when multiple directories contain the same file -- the newest file wins.
* Resolved most TODOs regarding error handling when adding a package domain,
package, or package node.
* Volume::_AddPackageContentRootNode(): Fixed the back tracking loop. It would
just continue with siblings of the given root node, causing them to be added
more than once, leading to data structure corruption.
* We now listen to the packages directory. When packages are added or removed
their contents is automatically added/removed to the file system. We don't
listen to the package files themselves yet. I.e. modifying an "installed"
package in place respectively copying/downloading it there will probably not
work correctly.
Due to a limitation in the userlandfs the mechanism works correctly only when
using the kernel module. When running the userland version, one better leaves
the packages directory alone while packagefs is mounted.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34172 a95241bf-73f2-0310-859d-f6bbb57e9c96
to be cloned.
* Added "flags" parameter to the SetTo(const void*,...) version.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34155 a95241bf-73f2-0310-859d-f6bbb57e9c96
data to a DataOutput.
* Implemented packagefs' io() hook and changed the read() hook implementation
to use the file cache. It's now possible to mmap() files and thus execute
programs.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34138 a95241bf-73f2-0310-859d-f6bbb57e9c96
to physical memory whose address would accidentally satisfy the
IS_USER_ADDRESS() check.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34136 a95241bf-73f2-0310-859d-f6bbb57e9c96
parameter and request->IsWrite(). The parameter means whether we want to
write to the request's I/O buffer (therefore renamed it to writeToRequest),
while request->IsWrite() indicates whether the request is a write request.
One can only write to a read request's buffer and vice versa.
IOBuffer::LockMemory() also wants to know whether the request is a write
request, not whether we want to write to the memory.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34135 a95241bf-73f2-0310-859d-f6bbb57e9c96
PackageNodeAttribute to the PackageNodes.
* Implemented the attribute FS interface hooks.
* Made Package and PackageNode BReferenceable.
* Fixed the reference management for Node.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34119 a95241bf-73f2-0310-859d-f6bbb57e9c96
option to cache a freed buffer -- and implementations
BlockBufferCache{NoLock,Kernel}.
* ZlibPackageDataReader does now dynamically get its read and uncompress
buffers from a provided BufferCache when needed.
* Allocating the buffers once and keeping them over the whole life time was a
bit too memory heavy, since we create a reader for every file for which a
vnode is created. A FS module global factory provides a buffer cache.
* Added a mutex to PackageFile::DataAccessor which guards the access to the
data reader which isn't thread safe.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34115 a95241bf-73f2-0310-859d-f6bbb57e9c96
file once, tracking an open count.
* Added VFSInit()/VFSUninit() methods to the Node and PackageNode class
hierarchies, called by the {get,put}_vnode() PackageFile implements them to
set up/tear down access to the file data. Also added a Read() reading the
data.
* Implemented the open(), free_cookie(), and read() FS hooks for real. Reading
files works now. Executing doesn't yet -- it requires working with
IORequests in a way not supported by the userlandfs.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34102 a95241bf-73f2-0310-859d-f6bbb57e9c96
start of the chunk, which is wrong for non chunk aligned offsets.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34101 a95241bf-73f2-0310-859d-f6bbb57e9c96
* We now report more correct stat data. The information are retrieved from the
Package* objects.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34098 a95241bf-73f2-0310-859d-f6bbb57e9c96
PackageSymlink with a "fat" interface.
* Replaced File and Symlink by fat class LeafNode.
* Added Package* object management to Directory and LeafNode.
* Implemented packagefs_read_symlink(), so symlinks work now.
* Added some missing locking the kernel interface.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34096 a95241bf-73f2-0310-859d-f6bbb57e9c96
* The FS parses package files from "/boot/common/packages" and builds an
node tree representation of the contained entries. Merging directories should
work. Lots of other stuff is missing yet, though.
* Implemented the hooks for directory entry lookup and iteration. So it's
possible to see the contents of the FS at least.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34088 a95241bf-73f2-0310-859d-f6bbb57e9c96
them in the PackageReader instead of fprintf().
* Got rid of unconditional printf().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34084 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added support for compressing/decompressing the TOC section of the package
to PackageWriter/PackageReader.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34063 a95241bf-73f2-0310-859d-f6bbb57e9c96
table), resulting in invalid archives when storing files > 64 KB.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34058 a95241bf-73f2-0310-859d-f6bbb57e9c96
Ingo Weinhold