Blame - docs/jni-tips.html - platform/dalvik

blob: 6d7fa5535394d4738469f66a1b8d9c02745dc781 [file] [log] [blame]

The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	1	<html>
				2	<head>
				3	<title>Android JNI Tips</title>
				4	<link rel=stylesheet href="android.css">
				5	</head>
				6
				7	<body>
				8	<h1><a name="JNI_Tips"></a>Android JNI Tips</h1>
				9	<p>
				10	</p><p>
				11	</p><ul>
				12	<li> <a href="#What_s_JNI_">What's JNI?</a>
				13	</li>
				14	<li> <a href="#JavaVM_and_JNIEnv">JavaVM and JNIEnv</a>
				15
				16	</li>
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	17	<li> <a href="#jclass_jmethodID_and_jfieldID">jclass, jmethodID, and jfieldID</a>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	18	</li>
				19	<li> <a href="#local_vs_global_references">Local vs. Global References</a>
				20	</li>
				21	<li> <a href="#UTF_8_and_UTF_16_strings">UTF-8 and UTF-16 Strings</a>
				22	</li>
				23	<li> <a href="#Arrays">Primitive Arrays</a>
				24	</li>
				25	<li> <a href="#RegionCalls">Region Calls</a>
				26	</li>
				27	<li> <a href="#Exceptions">Exceptions</a>
				28	</li>
				29
				30	<li> <a href="#Extended_checking">Extended Checking</a>
				31	</li>
				32	<li> <a href="#Native_Libraries">Native Libraries</a>
				33	</li>
				34	<li> <a href="#64bit">64-bit Considerations</a>
				35	</li>
				36
				37	<li> <a href="#Unsupported">Unsupported Features</a>
				38	</ul>
				39	<p>
				40	<noautolink>
				41	</noautolink></p><p>
				42	</p><h2><a name="What_s_JNI_"> </a> What's JNI? </h2>
				43	<p>
				44
				45	JNI is the Java Native Interface. It defines a way for code written in the
				46	Java programming language to interact with native
				47	code, e.g. functions written in C/C++. It's VM-neutral, has support for loading code from
				48	dynamic shared libraries, and while cumbersome at times is reasonably efficient.
				49	</p><p>
				50	You really should read through the
				51	<a href="http://java.sun.com/javase/6/docs/technotes/guides/jni/spec/jniTOC.html">JNI spec for J2SE 1.6</a>
				52	to get a sense for how JNI works and what features are available. Some
				53	aspects of the interface aren't immediately obvious on
				54	first reading, so you may find the next few sections handy.
				55	The more detailed <i>JNI Programmer's Guide and Specification</i> can be found
				56	<a href="http://java.sun.com/docs/books/jni/html/jniTOC.html">here</a>.
				57	</p><p>
				58	</p><p>
				59	</p><h2><a name="JavaVM_and_JNIEnv"> </a> JavaVM and JNIEnv </h2>
				60	<p>
				61	JNI defines two key data structures, "JavaVM" and "JNIEnv". Both of these are essentially
				62	pointers to pointers to function tables. (In the C++ version, it's a class whose sole member
				63	is a pointer to a function table.) The JavaVM provides the "invocation interface" functions,
				64	which allow you to create and destroy the VM. In theory you can have multiple VMs per process,
				65	but Android's VMs only allow one.
				66	</p><p>
				67	The JNIEnv provides most of the JNI functions. Your native functions all receive a JNIEnv as
				68	the first argument.
				69	</p><p>
				70
				71	On some VMs, the JNIEnv is used for thread-local storage. For this reason, <strong>you cannot share a JNIEnv between threads</strong>.
				72	If a piece of code has no other way to get its JNIEnv, you should share
				73	the JavaVM, and use JavaVM->GetEnv to discover the thread's JNIEnv.
				74	</p><p>
				75	The C and C++ declarations of JNIEnv and JavaVM are different. "jni.h" provides different typedefs
				76	depending on whether it's included into ".c" or ".cpp". For this reason it's a bad idea to
				77	include JNIEnv arguments in header files included by both languages. (Put another way: if your
				78	header file requires "#ifdef __cplusplus", you may have to do some extra work if anything in
				79	that header refers to JNIEnv.)
				80	</p><p>
				81	</p><p>
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	82	</p><h2><a name="jclass_jmethodID_and_jfieldID"> jclass, jmethodID, and jfieldID </a></h2>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	83	<p>
				84	If you want to access an object's field from native code, you would do the following:
				85	</p><p>
				86	</p><ul>
				87	<li> Get the class object reference for the class with <code>FindClass</code>
				88	</li>
				89	<li> Get the field ID for the field with <code>GetFieldID</code>
				90	</li>
				91	<li> Get the contents of the field with something appropriate, e.g.
				92	<code>GetIntField</code>
				93	</li>
				94	</ul>
				95	<p>
				96	Similarly, to call a method, you'd first get a class object reference and then a method ID. The IDs are often just
				97	pointers to internal VM data structures. Looking them up may require several string
				98	comparisons, but once you have them the actual call to get the field or invoke the method
				99	is very quick.
				100	</p><p>
				101	If performance is important, it's useful to look the values up once and cache the results
				102	in your native code. Because we are limiting ourselves to one VM per process, it's reasonable
				103	to store this data in a static local structure.
				104	</p><p>
				105	The class references, field IDs, and method IDs are guaranteed valid until the class is unloaded. Classes
				106	are only unloaded if all classes associated with a ClassLoader can be garbage collected,
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	107	which is rare but will not be impossible in our system. Note however that
				108	the <code>jclass</code>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	109	is a class reference and <strong>must be protected</strong> with a call
				110	to <code>NewGlobalRef</code> (see the next section).
				111	</p><p>
				112	If you would like to cache the IDs when a class is loaded, and automatically re-cache them
				113	if the class is ever unloaded and reloaded, the correct way to initialize
				114	the IDs is to add a piece of code that looks like this to the appropriate class:
				115	</p><p>
				116
				117	</p><pre> /*
				118	* We use a class initializer to allow the native code to cache some
				119	* field offsets.
				120	*/
				121
				122	/*
				123	* A native function that looks up and caches interesting
				124	* class/field/method IDs for this class. Returns false on failure.
				125	*/
				126	native private static boolean nativeClassInit();
				127
				128	/*
				129	* Invoke the native initializer when the class is loaded.
				130	*/
				131	static {
				132	if (!nativeClassInit())
				133	throw new RuntimeException("native init failed");
				134	}
				135	</pre>
				136	<p>
				137	Create a nativeClassInit method in your C/C++ code that performs the ID lookups. The code
				138	will be executed once, when the class is initialized. If the class is ever unloaded and
				139	then reloaded, it will be executed again. (See the implementation of java.io.FileDescriptor
				140	for an example in our source tree.)
				141	</p><p>
				142	</p><p>
				143	</p><p>
				144	</p><h2><a name="local_vs_global_references"> Local vs. Global References </a></h2>
				145	<p>
				146	Every object that JNI returns is a "local reference". This means that it's valid for the
				147	duration of the current native method in the current thread.
				148	<strong>Even if the object itself continues to live on after the native method returns, the reference is not valid.</strong>
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	149	This applies to all sub-classes of <code>jobject</code>, including
				150	<code>jclass</code>, <code>jstring</code>, and <code>jarray</code>.
				151	(Dalvik VM will warn you about most reference mis-uses when extended JNI
				152	checks are enabled.)
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	153	</p><p>
				154
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	155	If you want to hold on to a reference for a longer period, you must use
				156	a "global" reference. The <code>NewGlobalRef</code> function takes the
				157	local reference as an argument and returns a global one.
				158	The global reference is guaranteed to be valid until you call
				159	<code>DeleteGlobalRef</code>.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	160
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	161	</p><p>
				162	This pattern is commonly used when caching copies of class objects obtained
				163	from <code>FindClass</code>, e.g.:
				164	<p><pre>jclass* localClass = env->FindClass("MyClass");
				165	jclass* globalClass = (jclass*) env->NewGlobalRef(localClass);
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	166	</pre>
				167
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	168	</p><p>
				169	All JNI methods accept both local and global references as arguments.
Andy McFadden	5aca603	2009-05-06 16:48:33 -0700	[diff] [blame]	170	It's possible for references to the same object to have different values;
				171	for example, the return values from consecutive calls to
				172	<code>NewGlobalRef</code> on the same object may be different.
				173	<strong>To see if two references refer to the same object,
				174	you must use the <code>IsSameObject</code> function.</strong> Never compare
				175	references with "==" in native code.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	176	</p><p>
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	177	One consequence of this is that you
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	178	<strong>must not assume object references are constant or unique</strong>
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	179	in native code. The 32-bit value representing an object may be different
				180	from one invocation of a method to the next, and it's possible that two
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	181	different objects could have the same 32-bit value on consecutive calls. Do
				182	not use <code>jobject</code> values as keys.
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	183	</p><p>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	184	Programmers are required to "not excessively allocate" local references. In practical terms this means
				185	that if you're creating large numbers of local references, perhaps while running through an array of
				186	Objects, you should free them manually with
				187	<code>DeleteLocalRef</code> instead of letting JNI do it for you. The
				188	VM is only required to reserve slots for
				189	16 local references, so if you need more than that you should either delete as you go or use
				190	<code>EnsureLocalCapacity</code> to reserve more.
				191	</p><p>
				192	Note: method and field IDs are just 32-bit identifiers, not object
				193	references, and should not be passed to <code>NewGlobalRef</code>. The raw data
				194	pointers returned by functions like <code>GetStringUTFChars</code>
				195	and <code>GetByteArrayElements</code> are also not objects.
				196	</p><p>
				197	One unusual case deserves separate mention. If you attach a native
				198	thread to the VM with AttachCurrentThread, the code you are running will
				199	never "return" to the VM until the thread detaches from the VM. Any local
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	200	references you create will have to be deleted manually unless you're going
				201	to detach the thread soon.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	202	</p><p>
				203	</p><p>
				204	</p><p>
				205	</p><h2><a name="UTF_8_and_UTF_16_strings"> </a> UTF-8 and UTF-16 Strings </h2>
				206	<p>
				207	The Java programming language uses UTF-16. For convenience, JNI provides methods that work with "modified UTF-8" encoding
				208	as well. (Some VMs use the modified UTF-8 internally to store strings; ours do not.) The
				209	modified encoding only supports the 8- and 16-bit forms, and stores ASCII NUL values in a 16-bit encoding.
				210	The nice thing about it is that you can count on having C-style zero-terminated strings,
				211	suitable for use with standard libc string functions. The down side is that you cannot pass
				212	arbitrary UTF-8 data into the VM and expect it to work correctly.
				213	</p><p>
				214	It's usually best to operate with UTF-16 strings. With our current VMs, the
				215	<code>GetStringChars</code> method
				216	does not require a copy, whereas <code>GetStringUTFChars</code> requires a malloc and a UTF conversion. Note that
				217	<strong>UTF-16 strings are not zero-terminated</strong>, and \u0000 is allowed,
				218	so you need to hang on to the string length as well as
				219	the string pointer.
				220
				221	</p><p>
Andy McFadden	5aca603	2009-05-06 16:48:33 -0700	[diff] [blame]	222	<strong>Don't forget to Release the strings you Get</strong>. The
				223	string functions return <code>jchar</code> or <code>jbyte</code>, which
				224	are C-style pointers to primitive data rather than local references. They
				225	are guaranteed valid until Release is called, which means they are not
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	226	released when the native method returns.
				227	</p><p>
				228	</p><p>
				229
				230
				231	</p><h2><a name="Arrays"> </a> Primitive Arrays </h2>
				232	<p>
				233	JNI provides functions for accessing the contents of array objects.
				234	While arrays of objects must be accessed one entry at a time, arrays of
				235	primitives can be read and written directly as if they were declared in C.
				236	</p><p>
				237	To make the interface as efficient as possible without constraining
				238	the VM implementation,
				239	the <code>Get<PrimitiveType>ArrayElements</code> family of calls
				240	allows the VM to either return a pointer to the actual elements, or
				241	allocate some memory and make a copy. Either way, the raw pointer returned
				242	is guaranteed to be valid until the corresponding <code>Release</code> call
				243	is issued (which implies that, if the data wasn't copied, the array object
				244	will be pinned down and can't be relocated as part of compacting the heap).
				245	<strong>You must Release every array you Get.</strong> Also, if the Get
				246	call fails, you must ensure that your code doesn't try to Release a NULL
				247	pointer later.
				248	</p><p>
				249	You can determine whether or not the data was copied by passing in a
				250	non-NULL pointer for the <code>isCopy</code> argument. This is rarely
				251	useful.
				252	</p><p>
				253	The <code>Release</code> call takes a <code>mode</code> argument that can
				254	have one of three values. The actions performed by the VM depend upon
				255	whether it returned a pointer to the actual data or a copy of it:
				256	<ul>
				257	<li><code>0</code>
				258	<ul>
				259	<li>Actual: the array object is un-pinned.
				260	<li>Copy: data is copied back. The buffer with the copy is freed.
				261	</ul>
				262	<li><code>JNI_COMMIT</code>
				263	<ul>
				264	<li>Actual: does nothing.
				265	<li>Copy: data is copied back. The buffer with the copy
				266	<strong>is not freed</strong>.
				267	</ul>
				268	<li><code>JNI_ABORT</code>
				269	<ul>
				270	<li>Actual: the array object is un-pinned. Earlier
				271	writes are <strong>not</strong> aborted.
				272	<li>Copy: the buffer with the copy is freed; any changes to it are lost.
				273	</ul>
				274	</ul>
				275	</p><p>
				276	One reason for checking the <code>isCopy</code> flag is to know if
				277	you need to call <code>Release</code> with <code>JNI_COMMIT</code>
				278	after making changes to an array -- if you're alternating between making
				279	changes and executing code that uses the contents of the array, you may be
				280	able to
				281	skip the no-op commit. Another possible reason for checking the flag is for
				282	efficient handling of <code>JNI_ABORT</code>. For example, you might want
				283	to get an array, modify it in place, pass pieces to other functions, and
				284	then discard the changes. If you know that JNI is making a new copy for
				285	you, there's no need to create another "editable" copy. If JNI is passing
				286	you the original, then you do need to make your own copy.
				287	</p><p>
				288	Some have asserted that you can skip the <code>Release</code> call if
				289	<code>*isCopy</code> is false. This is not the case. If no copy buffer was
				290	allocated, then the original memory must be pinned down and can't be moved by
				291	the garbage collector.
				292	</p><p>
				293	Also note that the <code>JNI_COMMIT</code> flag does NOT release the array,
				294	and you will need to call <code>Release</code> again with a different flag
				295	eventually.
				296	</p><p>
				297	</p><p>
				298
				299
				300	</p><h2><a name="RegionCalls"> Region Calls </a></h2>
				301
				302	<p>
				303	There is an alternative to calls like <code>Get<Type>ArrayElements</code>
				304	and <code>GetStringChars</code> that may be very helpful when all you want
				305	to do is copy data in or out. Consider the following:
				306	<pre>
				307	jbyte* data = env->GetByteArrayElements(array, NULL);
				308	if (data != NULL) {
				309	memcpy(buffer, data, len);
				310	env->ReleaseByteArrayElements(array, data, JNI_ABORT);
				311	}
				312	</pre>
				313	<p>
				314	This grabs the array, copies the first <code>len</code> byte
				315	elements out of it, and then releases the array. Depending upon the VM
				316	policies the <code>Get</code> call will either pin or copy the array contents.
				317	We copy the data (for perhaps a second time), then call Release; in this case
				318	we use <code>JNI_ABORT</code> so there's no chance of a third copy.
				319	</p><p>
				320	We can accomplish the same thing with this:
				321	<pre>
				322	env->GetByteArrayRegion(array, 0, len, buffer);
				323	</pre>
				324	</p><p>
				325	This accomplishes the same thing, with several advantages:
				326	<ul>
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	327	<li>Requires one JNI call instead of 2, reducing overhead.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	328	<li>Doesn't require pinning or extra data copies.
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	329	<li>Reduces the risk of programmer error -- no risk of forgetting
				330	to call <code>Release</code> after something fails.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	331	</ul>
				332	</p><p>
				333	Similarly, you can use the <code>Set<Type>ArrayRegion</code> call
				334	to copy data into an array, and <code>GetStringRegion</code> or
				335	<code>GetStringUTFRegion</code> to copy characters out of a
				336	<code>String</code>.
				337
				338
				339	</p><h2><a name="Exceptions"> Exceptions </a></h2>
				340	<p>
				341	<strong>You may not call most JNI functions while an exception is pending.</strong>
				342	Your code is expected to notice the exception (via the function's return value,
				343	<code>ExceptionCheck()</code>, or <code>ExceptionOccurred()</code>) and return,
				344	or clear the exception and handle it.
				345	</p><p>
				346	The only JNI functions that you are allowed to call while an exception is
				347	pending are:
				348	<font size="-1"><ul>
				349	<li>DeleteGlobalRef
				350	<li>DeleteLocalRef
				351	<li>DeleteWeakGlobalRef
				352	<li>ExceptionCheck
				353	<li>ExceptionClear
				354	<li>ExceptionDescribe
				355	<li>ExceptionOccurred
				356	<li>MonitorExit
				357	<li>PopLocalFrame
				358	<li>PushLocalFrame
Elliott Hughes	b7906ed	2009-09-09 15:16:59 -0700	[diff] [blame]	359	<li>Release<PrimitiveType>ArrayElements
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	360	<li>ReleasePrimitiveArrayCritical
				361	<li>ReleaseStringChars
				362	<li>ReleaseStringCritical
				363	<li>ReleaseStringUTFChars
				364	</ul></font>
				365	</p><p>
				366	Note that exceptions thrown by interpreted code do not "leap over" native code,
				367	and C++ exceptions thrown by native code are not handled by Dalvik.
				368	The JNI <code>Throw</code> and <code>ThrowNew</code> instructions just
				369	set an exception pointer in the current thread. Upon returning to the VM from
				370	native code, the exception will be noted and handled appropriately.
				371	</p><p>
				372	Native code can "catch" an exception by calling <code>ExceptionCheck</code> or
				373	<code>ExceptionOccurred</code>, and clear it with
				374	<code>ExceptionClear</code>. As usual,
				375	discarding exceptions without handling them can lead to problems.
				376	</p><p>
				377	There are no built-in functions for manipulating the Throwable object
				378	itself, so if you want to (say) get the exception string you will need to
				379	find the Throwable class, look up the method ID for
				380	<code>getMessage "()Ljava/lang/String;"</code>, invoke it, and if the result
				381	is non-NULL use <code>GetStringUTFChars</code> to get something you can
				382	hand to printf or a LOG macro.
				383
				384	</p><p>
				385	</p><p>
				386	</p><h2><a name="Extended_checking"> Extended Checking </a></h2>
				387	<p>
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	388	JNI does very little error checking. Calling <code>SetIntField</code>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	389	on an Object field will succeed, even if the field is marked
				390	<code>private</code> and <code>final</code>. The
				391	goal is to minimize the overhead on the assumption that, if you've written it in native code,
				392	you probably did it for performance reasons.
				393	</p><p>
				394	Some VMs support extended checking with the "<code>-Xcheck:jni</code>" flag. If the flag is set, the VM
				395	puts a different table of functions into the JavaVM and JNIEnv pointers. These functions do
				396	an extended series of checks before calling the standard implementation.
				397
				398	</p><p>
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	399	Some things that may be checked:
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	400	</p><p>
				401	</p>
				402	<ul>
				403	<li> Check for null pointers where not allowed.
				404	<li>
				405	<li> Verify argument type correctness (jclass is a class object,
				406	jfieldID points to field data, jstring is a java.lang.String).
				407	</li>
				408	<li> Field type correctness, e.g. don't store a HashMap in a String field.
				409	</li>
				410	<li> Check to see if an exception is pending on calls where pending exceptions are not legal.
				411	</li>
				412	<li> Check for calls to inappropriate functions between Critical get/release calls.
				413	</li>
				414	<li> Check that JNIEnv structs aren't being shared between threads.
				415
				416	</li>
				417	<li> Make sure local references aren't used outside their allowed lifespan.
				418	</li>
				419	<li> UTF-8 strings contain valid "modified UTF-8" data.
				420	</li>
				421	</ul>
				422	<p>Accessibility of methods and fields (i.e. public vs. private) is not
				423	checked.
				424	<p>
				425	The Dalvik VM supports the <code>-Xcheck:jni</code> flag. For a
				426	description of how to enable it for Android apps, see
				427	<a href="embedded-vm-control.html">Controlling the Embedded VM</a>.
				428	It's currently enabled by default in the Android emulator and on
				429	"engineering" device builds.
				430
				431	</p><p>
				432	JNI checks can be modified with the <code>-Xjniopts</code> command-line
				433	flag. Currently supported values include:
				434	</p>
				435	<blockquote><dl>
				436	<dt>forcecopy
				437	<dd>When set, any function that can return a copy of the original data
				438	(array of primitive values, UTF-16 chars) will always do so. The buffers
				439	are over-allocated and surrounded with a guard pattern to help identify
				440	code writing outside the buffer, and the contents are erased before the
				441	storage is freed to trip up code that uses the data after calling Release.
Andy McFadden	c7659ec	2009-09-18 16:14:41 -0700	[diff] [blame]	442	This will have a noticeable performance impact on some applications.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	443	<dt>warnonly
				444	<dd>By default, JNI "warnings" cause the VM to abort. With this flag
				445	it continues on.
				446	</dl></blockquote>
				447
				448
				449	</p><p>
				450	</p><h2><a name="Native_Libraries"> Native Libraries </a></h2>
				451	<p>
				452	You can load native code from shared libraries with the standard
				453	<code>System.loadLibrary()</code> call. The
				454	preferred way to get at your native code is:
				455	</p><p>
				456	</p><ul>
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	457	<li> Call <code>System.loadLibrary()</code> from a static class
				458	initializer. (See the earlier example, where one is used to call
				459	<code>nativeClassInit()</code>.) The argument is the "undecorated"
				460	library name, e.g. to load "libfubar.so" you would pass in "fubar".
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	461
				462	</li>
				463	<li> Provide a native function: <code><strong>jint JNI_OnLoad(JavaVM* vm, void* reserved)</strong></code>
				464	</li>
				465	<li>In <code>JNI_OnLoad</code>, register all of your native methods. You
				466	should declare
				467	the methods "static" so the names don't take up space in the symbol table
				468	on the device.
				469	</li>
				470	</ul>
				471	<p>
				472	The <code>JNI_OnLoad</code> function should look something like this if
				473	written in C:
				474	</p><blockquote><pre>jint JNI_OnLoad(JavaVM* vm, void* reserved)
				475	{
				476	JNIEnv* env;
				477	if ((vm)->GetEnv(vm, (void*) &env, JNI_VERSION_1_4) != JNI_OK)
				478	return -1;
				479
				480	/* get class with (env)->FindClass /
				481	/* register methods with (env)->RegisterNatives /
				482
				483	return JNI_VERSION_1_4;
				484	}
				485	</pre></blockquote>
				486	</p><p>
				487	You can also call <code>System.load()</code> with the full path name of the
Andy McFadden	40f3271	2009-09-22 13:55:42 -0700	[diff] [blame]	488	shared library. For Android apps, you may find it useful to get the full
				489	path to the application's private data storage area from the context object.
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	490	</p><p>
Andy McFadden	278c12b	2009-09-21 16:28:39 -0700	[diff] [blame]	491	This is the recommended approach, but not the only approach. The VM does
				492	not require explicit registration, nor that you provide a
				493	<code>JNI_OnLoad</code> function.
				494	You can instead use "discovery" of native methods that are named in a
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	495	specific way (see <a href="http://java.sun.com/javase/6/docs/technotes/guides/jni/spec/design.html#wp615">
Andy McFadden	278c12b	2009-09-21 16:28:39 -0700	[diff] [blame]	496	the JNI spec</a> for details), though this is less desirable.
				497	It requires more space in the shared object symbol table,
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	498	loading is slower because it requires string searches through all of the
				499	loaded shared libraries, and if a method signature is wrong you won't know
				500	about it until the first time the method is actually used.
				501	</p><p>
Andy McFadden	278c12b	2009-09-21 16:28:39 -0700	[diff] [blame]	502	One other note about <code>JNI_OnLoad</code>: any <code>FindClass</code>
				503	calls you make from there will happen in the context of the class loader
				504	that was used to load the shared library. Normally <code>FindClass</code>
				505	uses the loader associated with the method at the top of the interpreted
				506	stack, or if there isn't one (because the thread was just attached to
				507	the VM) it uses the "system" class loader.
				508	</p><p>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	509
				510
				511	</p><h2><a name="64bit"> 64-bit Considerations </a></h2>
				512
				513	<p>
				514	Android is currently expected to run on 32-bit platforms. In theory it
				515	could be built for a 64-bit system, but that is not a goal at this time.
				516	For the most part this isn't something that you will need to worry about
				517	when interacting with native code,
				518	but it becomes significant if you plan to store pointers to native
				519	structures in integer fields in an object. To support architectures
				520	that use 64-bit pointers, <strong>you need to stash your native pointers in a
				521	<code>long</code> field rather than an <code>int</code></strong>.
				522
				523
				524	</p><h2><a name="Unsupported"> Unsupported Features </a></h2>
				525	<p>All JNI 1.6 features are supported, with the following exceptions:
				526	<ul>
				527	<li><code>DefineClass</code> is not implemented. Dalvik does not use
				528	Java bytecodes or class files, so passing in binary class data
				529	doesn't work. Translation facilities may be added in a future
				530	version of the VM.</li>
Andy McFadden	b18992f	2009-09-25 10:42:15 -0700	[diff] [blame^]	531	<li>"Weak global" references are implemented, but may only be passed
				532	to <code>NewLocalRef</code>, <code>NewGlobalRef</code>, and
				533	<code>DeleteWeakGlobalRef</code>. (The spec strongly encourages
				534	programmers to create hard references to weak globals before doing
				535	anything with them, so this should not be at all limiting.)</li>
The Android Open Source Project	f6c3871	2009-03-03 19:28:47 -0800	[diff] [blame]	536	<li><code>GetObjectRefType</code> (new in 1.6) is implemented but not fully
				537	functional -- it can't always tell the difference between "local" and
				538	"global" references.</li>
				539	</ul>
				540
				541	</p>
				542
				543	<address>Copyright © 2008 The Android Open Source Project</address>
				544
				545	</body>
				546	</html>