| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
native library path, supported throughout DynamicLibraryBundle[Info]
Motivation: It is helpful to retrieve the actually used native library pathname,
since loading a library w/o absolute path but lookup through LD_LIBRARY_PATH
may render it hard for the user to determine which library is used.
+++
+++
Windows implementation simply can use GetModuleFileNameA() with the native library handle.
POSIX implementation may utilize a symbol-name to retrieve its address within the
loading native library used to retrieved the library information
via dladdr().
To support this feature throughout DynamicLibraryBundle and DynamicLibraryBundleInfo,
the custom DynamicLibraryBundleInfo specializations shall provide
optional symbol-names per each tool-library-name for the POSIX implementation,
see above.
public interface DynamicLibraryBundleInfo {
...
/**
* Returns optional list of optional symbol names per {@link #getToolLibNames()}
* in same order for an OS which requires the symbol's address to retrieve
* the path of the containing library.
*/
public List<String> getSymbolForToolLibPath();
...
}
|
|
|
|
|
|
|
|
| |
current monotonic milliseconds **since start of this application**
Also added TSPrinter, a simple millisecond timestamp prepending `print*()` wrapper for a {@link PrintStream}.
Test coverage via TestClock01.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Win32 clock_gettime() implementation.
Consider return code on failed native clock_gettime(..) call
- Return Instant.EPOCH for all Instant variations (essentially 0)
- Return 0 for all 'long' variations (ms, ns)
Add Win32 clock_gettime() implementation.
- Source: https://github.com/Alexpux/mingw-w64/blob/master/mingw-w64-libraries/winpthreads/src/clock.c
- Public Domain within mingw-w64, included here to simplify linkage.
- Tested on Win10 64bit w/ TestTextRendererNEWT00, all values are OK
|
|
|
|
| |
duration since module startup and not time.
|
|
|
|
|
|
|
|
|
|
|
| |
startup, retrievable via getMonotonicStartupTime(). (performance)
Settings two long fields in getMonotonicTime() and creating Instant and using Duration
for high-frequency counter is too expensive.
currentTimeNanos() subtracts the startup time from the current monotonic time and returns the
resulting duration in nanoseconds, which lasts for 292 years since module startup.
This satisfies performance counter requirements.
|
|
|
|
|
|
|
|
|
| |
(sec + nsec), currentTimeMillis() is also monotonic now, reused by Platform. Dropped Platform.currentTimeMicros()
Clock and its implementation was copied from jaulibs, a spin-off from Direct-BT.
The implementation uses `clock_gettime(CLOCK_MONOTONIC, &t)` and is considered safe and high-performant
as it avoids a kernel call via VDSO (GNU/Linux).
|
|
|
|
| |
to 'java.library.path', others are absolute
|
|
|
|
|
|
| |
is being resolved as absolute-canonical as required for System.load*()
Further, detailed DEBUG messages are added on -Djogamp.debug.NativeLibrary
|
|
|
|
| |
.. content
|
| |
|
|
|
|
| |
'long double' than regular x86_32_unix
|
|
|
|
|
|
|
|
|
| |
make/lib/gluegen-cpptask-android-<abi>.xml and scripts
All aligned now
- gluegen-cpptasks-android-aarch64.xml
- gluegen-cpptasks-android-armv6.xml (this has ld flag --no-undefined disabled, due to internal missing symbols)
- gluegen-cpptasks-android-x86.xml
|
|
|
|
|
|
|
|
|
|
| |
gluegen-rt -> gluegen_rt
- Fix finding native library for Elf parsing
- Fix one unit test
- Fix comments
Completes commit 330dad069dee5a0cc0480cf5cd9052000004223a
|
| |
|
|
|
|
| |
relax closeLibrary(..)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
libgluegen-rt[.so] -> libgluegen_rt[.so|.a]
Recognize Java9+ ..
- Recognize new Java9+ version string as of JEP 223
- Avoid Classpath's private findLibrary call
+++
Support JEP 178 static linkage (OpenJDK 1.8)
- Need to change native library basename: libgluegen-rt[.so] -> libgluegen_rt[.so|.a]
since the dash '-' is not supported in a ANSI-c function name.
- Added 'JNI_OnLoad_gluegen_rt' to recognize statical linked JNI code
- Added JNI_VERSION_1_8 to jni/jni.h
|
|
|
|
|
|
|
|
|
|
|
|
| |
Current build system for JogAmp iOS Build is:
- Build Machine: OSX Mojave 10.14
- Using own (still proprietary) OpenJDK 9 iOS Build
- OpenJDK 1.8 (This will be replaced by OpenJDK 11 soon)
- Xcode 10.2
|
|
|
|
| |
executables
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
and 'searchSystemPathFirst' attributes
NativeLibrary can be instantiate by defining
'searchSystemPath' and 'searchSystemPathFirst' arguments,
allowing to specify the system path role while looking up the library.
Since NativeLibrary is utilized via DynamicLibraryBundleInfo upstream,
the latter interface shall allow users to specify those attributes.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
PlatformPropsImpl.findSysLib(..).
This patch partially reverts of commit d12e4d4ea279998b27457691038e709879dcaca6.
NativeLibrary.open(..) requires search of system libraries,
since it loads the actual 'tool library' for which we generate the JNI binding.
The 'tool library' is preferably the system wide installed version,
e.g. libGL.so etc.
PlatformPropsImpl.findSysLib(..) also requires finding system libraries
as needed for PlatformPropsImpl.queryElfFile(..), i.e. using libjava.so etc.
Only the JNI 'glue library', glueing java calls to the 'tool library',
shall not use the system wide library search since we shall only use
JogAmp provided instances here.
This patch also reinstates binary compatibility w/ prev. GlueGen JARs
since NativeLibrary.enumerateLibraryPath(..) is public.
+++
Further more 'NativeLibrary.enumerateLibraryPath(..)'
now adds OSX system framework search _before_ the user path
in case 'searchSystemPath && searchSystemPathFirst'.
Original code always added this search to the end,
which does not match the intended behavior (-> bug).
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
src/java/com/jogamp/common/os/NativeLibrary.java
Due to commit for Bug 1145, bf4d8786cb732d86db333b43020ecf0af27f60bf
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
extension mechanism
NativeLibrary:
API change: Removed searchSystemPathFirst argument to the
open and enumerateLibraryPaths methods.
Removed the generic sun.boot.library.path system path and the
MacOS specific Frameworks paths from enumerateLibraryPaths.
JNILibLoaderBase, PlatformPropsImpl & TestElfReader01:
Updated to handle the NativeLibrary API change.
This change will prevent JogAmp modules to pickup and load unsupported
and old SUN JOGL 1 natives that may have been deployed with the JRE.
|
|/
|
|
|
|
| |
natives/os-arch/ + library names
Fixes Bug 1145 cc1 when using an unpacked fat-jar
|
|
|
|
|
|
|
| |
'toolGetProcAddressHandle' (Bug 1062)
Show 'toolGetProcAddressHandle' in DEBUG mode in DynamicLibraryBundle.toolDynamicLookupFunction(..),
allowing to validate source of symbols.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
commit 12feaa7d3b1544098f684d851e3caff1ec88cbc8
Regression of commit 12feaa7d3b1544098f684d851e3caff1ec88cbc8
'isToolLibComplete()' returned false if dynLinkGlobal is null,
even if no tool-lib has been used. In which case dynLinkGlobal
has not been initialized and hence is always null.
'isToolLibComplete()' now takes no tool-lib into consideration!
Currently only 'OVRDynamicLibraryBundleInfo' of JOGL's 'oculusvr' binding
used this configuration within JogAmp.
|
|
|
|
| |
SecurityException' decl., remove dead code, remove redundant check.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
matching StaticConfig at runtime; Fix PPC (Bug 1056) and MIPSLE (Bug 1014) issues.
Currently the StaticConfig is being queried
via the key[OSType, CPUType ..]
as pre-determined by Java properties or the ELF parser.
This adds complication to maintain different platforms
and the key query might not even be sufficient.
The MachineDescriptor's StaticConfig only purpose shall be
to speed-up native data size and offset/alignment retrieval.
This is done by using the StaticConfig index within
all StaticConfig[]s as a lookup-index for the precomputed
struct's size and offset tables.
+++
Solution:
Rename: MachineDescriptor -> MachineDataInfo
Rename: MachineDescriptorRuntime -> MachineDataInfoRuntime
After having defined os.and.arch (OSType, CPUType and ABIType)
w/ the optional help of the now self containing ELF Reader (Bug 1125),
the native gluegen-rt library gets loaded enabling JNI methods.
It is satisfactory to retrieve MachineDataInfo
at runtime w/ JNI and find the matching/compatible StaticConfig.
Only in case none is found, the program needs to abort.
Otherwise the found MachineDataInfo.StaticConfig and MachineDataInfo
are stored for further use (see above).
This removes above complication and key to StaticConfig mapping.
New platforms simply need to add a new unique entry into the
StaticConfig[] table.
++
Also fixes Bug 1056 (PPC), thanks to tmancill [@] debian [.] org,
and Bug 1014 (MIPSLE), thanks to Dejan Latinovic.
Parts of the patch for Bug 1014 from Dejan Latinovic are included.
also solved by this change set.
|
|
|
|
| |
detection
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ELF Reader 'jogamp.common.os.elf' currently uses
Platform's pre-determined OS_TYPE and CPUType.
It also uses the host platforms MachineDescription,
hence can not read ELF files from other machines.
This also forbids Platform to determine CPUType etc
w/o having a valid 'os.arch' property.
+++
ElfHeader should be split in
- ElfHeaderPart1 (CPUType independent)
- ElfHeaderPart2 (CPUType dependent)
Fix shall make the ELF Reader self containing
by only using ELF CPUType data, etc.
This requires customization of struct parsing,
where MachineDescription.Static index shall be
- defined in ElfHeaderPart1 using e_Ident's CPUType.
- used in ElfHeaderPart2 and all its struct types.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Android:
- Detect ABIType.EABI_GNU_ARMHF via 'armeabi-v7a-hard'
Platform.CPUType:
- contains is32Bit now
MachineDescription:
- Rename *x86_64_unix* -> *lp64_unix*, reflecting universal __LP64__ mode
- Remove is32Bit, which is determined by CPUType
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
and fix Android AArch64 BionicDynamicLinker (Bug 1122)
- Bulk Permissions
ProcAddressTable.reset(..) performs address lookup in one block.
Now claiming all permissions upfront once, and releasing them afterwards.
- Cleanup DynamicLinker impl.
Proper top-down impl. of DynamicLinkerImpl,
handling all security code and validations.
- Fix Android AArch64 BionicDynamicLinker (Bug 1122)
Dalvik uses diff RTLD_* defines for AArch64!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Add AArch64 detection via
- Elf Parser
- Android properties
- Java properties
- Android: Validate CPUType.Family _and_ ABIType
- MachineDescription
- Remove redundant Type ID and its field
- Reuse X86_64_UNIX for AArch64 (static config)
New ARCH 'aarch64' for
ant: <os arch>
armv8a
aarch64
New CPUType.ARM64 (ARM):
java: os.arch
aarch64
arm64
New CPUType.ARMv8_A (ARM):
java: os.arch
armv8-a
arm64-v8a
New ABIType:
EABI_AARCH64
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Uri:
- Add error correction (fix) of path at ctor
in case a a contained Uri is requested from an opaque one.
The latter might come from URI/URL and is not properly encoded!
See Uri.PARSE_HINT_FIX_PATH
- Simplify conversion methods:
- valueOf(URI/URL): Re-encode only if not opaque
- getNativeFilePath() -> toFile()
- Move IOUtil Uri related constants to class Uri
- Add DEBUG and DEBUG_SHOWFIX
- Complete adoption of URI -> Uri changes
- IOUtil, JarUtil and TempJarCache still holds some
left over deprecated methods, which will be removed
after officially starting 2.3.0 (JOGL dependencies)
- Otherwise not URI utilization left
- Tests
- Cleaned ip TestUri01 and TestUri03Resolving
- TestUri03Resolving also tests URL -> Uri,
and hence mentioned Uri.PARSE_HINT_FIX_PATH
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In situations, where the native-jar file is not located within
the same parent URI as it's java-jar file, our location mechanism fails.
This patch adds a classloader based native-jar file location mechanism as a fallback,
requiring the native jar file to be included in the users CLASSPATH.
Classloader based location algorithm in JNILibLoaderBase.addNativeJarLibsImpl(..):
- Extract the 'module-name' from the given classFromJavaJar's package name,
i.e. the last package-part: 'jogamp.common.Debug' -> 'common'
Hence it is important to pass a 'classFromJavaJar',
which last package segment reflects the module-name!
- <os.and.arch> -> <os_and_arch_dot>, e.g. linux-amd64 -> linux.amd64 (linux/amd64)
- Locate class 'jogamp.nativetag.<module-name>.<os_and_arch_dot>.TAG',
e.g. 'jogamp.nativetag.common.linux.amd64.TAG'
- Use located class's JarFile URI .. continue as usual
Injection of above mentioned TAG class via gluegen-cpptasks-base.xml macro 'native.tag.jar':
- Creates dummy TAG.java code
- Compiles TAG.java
- Creates the native-jar file
Example:
<native.tag.jar objdir="${build}/obj"
nativejarfile="${build}/gluegen-rt-natives-${os.and.arch}.jar"
manifestfile="${build}/Manifest-rt-natives.temp"
module="common"
includelibs="*gluegen-rt.${native.library.suffix}" />
Note that the manifest file uses a matching Extension-Name:
Extension-Name: jogamp.nativetag.common
|
|
|
|
|
|
|
| |
accessed from inner class
Otherwise the compiler has to fill in a wrapping access function for the inner
class to call, which is less efficient.
|
|\
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
src/java/com/jogamp/common/jvm/JNILibLoaderBase.java
src/java/com/jogamp/common/os/Platform.java
src/junit/com/jogamp/common/util/TestTempJarCache.java
(re-injected that api-doc of addNativeJarLibsImpl)
|
| |
| |
| |
| |
| | |
Bug: 1023
Bug: 1024
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
c47bc86ae2ee268a1f38c5580d11f93d7f8d6e74)
Code Clean-Up based on our Recommended Settings (jogamp-scripting c47bc86ae2ee268a1f38c5580d11f93d7f8d6e74)
- Change non static accesses to static members using declaring type
- Change indirect accesses to static members to direct accesses (accesses through subtypes)
- Add final modifier to private fields
- Add final modifier to method parameters
- Add final modifier to local variables
- Remove unnecessary casts
- Remove unnecessary '$NON-NLS$' tags
- Remove trailing white spaces on all lines
|
|/ |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'best' ABI
- Use 'os.arch' as a prelim CPUType for MachineDescription
- Always attempt to load a binary and parse it's elf header
- Linux: self-exe
- Android: gluegen-rt library
- Other: java lib
- Always use details (ABI) if ARM
- Android: Check CPU_TYPE and CPU_TYPE2
// FIXME / HACK:
// We use sCPUType for MachineDescriptionRuntime.getStatic()
// until we have determined the final CPU_TYPE, etc.
// MachineDescriptionRuntime gets notified via MachineDescriptionRuntime.notifyPropsInitialized() below.
//
// We could use Elf Ehdr's machine value to determine the bit-size
// used for it's offset table!
// However, 'os.arch' should be a good guess for this task.
Tested manually on
- Linux x86, x86_64, armhf (raspi)
- Android intel and arm
- Windows x86_64
- OSX x86_64
|
| |
|
|
|
|
| |
Signed-off-by: Harvey Harrison <[email protected]>
|
|
|
|
| |
Signed-off-by: Harvey Harrison <[email protected]>
|
|
|
|
|
|
|
|
|
| |
TempJarCache.bootstrapNativeLib(), allowing versatile use of 1st native jar file (big-java-jar w/ big-native-jar)
The remaining Platform dependency existed in IOUtil.copyStream2Stream(..), used by JarUtil.extract(..),
i.e. the MachineDescription's PAGE_SIZE.
Solved by using a const buffer size of 4096 bytes.
|
| |
|
|
|
|
| |
methods, based on 'gettimeofday(..)'
|