Package Kit: Use an object_cache in kernel mode for decompression buffers.

The kernel heap only uses object caches for objects up to size 8192.
Larger allocations have to go through the raw allocator. That can
get pretty expensive.

Adding instrumentation around the malloc/free calls in this function
showed that on my machine, some 596ms during boot were spent on
*malloc/free alone*, all else aside. After this change, we are at
around 110ms, or a >5x improvement. Running an fgrep -R on /system/
after boot increased the cumulative time in memory functions to over
5 seconds, while after this change it is "only" 1170ms.

Honestly, it seems like the object depots should be able to be faster
than that, even if this function is called thousands of times. But that
is a problem for a different investigation.

It would be even faster for every consumer of this data in
packagefs just allocated one set of buffers up front, or at least
for a single "read session", but plumbing that all the way
through the myriad abstractions of the Package Kit will
not be easy, and is left for another time, as well.

headers/private/package/hpkg/PackageFileHeapAccessorBase.h

@@ -98,6 +98,9 @@ public:
 
 public:
 	static	const size_t		kChunkSize = 64 * 1024;
+#if defined(_KERNEL_MODE)
+	static	void*				sChunkCache;
+#endif
 
 protected:
 	virtual	status_t			ReadAndDecompressChunk(size_t chunkIndex,

src/add-ons/kernel/file_systems/packagefs/kernel_interface.cpp

@@ -14,9 +14,12 @@
 #include <fs_interface.h>
 #include <KernelExport.h>
 #include <io_requests.h>
+#include <slab/Slab.h>
 
 #include <AutoDeleter.h>
 
+#include <package/hpkg/PackageFileHeapAccessorBase.h>
+
 #include "AttributeCookie.h"
 #include "AttributeDirectoryCookie.h"
 #include "DebugSupport.h"
@@ -1055,6 +1058,8 @@ packagefs_rewind_query(fs_volume* fsVolume, void* cookie)
 static status_t
 packagefs_std_ops(int32 op, ...)
 {
+	using BPackageKit::BHPKG::BPrivate::PackageFileHeapAccessorBase;
+
 	switch (op) {
 		case B_MODULE_INIT:
 		{
@@ -1075,6 +1080,12 @@ packagefs_std_ops(int32 op, ...)
 				return error;
 			}
 
+			PackageFileHeapAccessorBase::sChunkCache =
+				create_object_cache_etc("packagefs heap buffers",
+					PackageFileHeapAccessorBase::kChunkSize, sizeof(void*),
+					0, /* magazine capacity, count */ 2, 1, 0, NULL,
+					NULL, NULL, NULL);
+
 			error = PackageFSRoot::GlobalInit();
 			if (error != B_OK) {
 				ERROR("Failed to init PackageFSRoot\n");
@@ -1091,6 +1102,8 @@ packagefs_std_ops(int32 op, ...)
 		{
 			PRINT("package_std_ops(): B_MODULE_UNINIT\n");
 			PackageFSRoot::GlobalUninit();
+			delete_object_cache((object_cache*)
+				PackageFileHeapAccessorBase::sChunkCache);
 			StringConstants::Cleanup();
 			StringPool::Cleanup();
 			exit_debugging();

src/kits/package/hpkg/PackageFileHeapAccessorBase.cpp

@@ -11,6 +11,9 @@
 
 #include <algorithm>
 #include <new>
+#ifdef _KERNEL_MODE
+#include <slab/Slab.h>
+#endif
 
 #include <ByteOrder.h>
 #include <DataIO.h>
@@ -27,6 +30,11 @@ namespace BHPKG {
 namespace BPrivate {
 
 
+#if defined(_KERNEL_MODE)
+void* PackageFileHeapAccessorBase::sChunkCache = NULL;
+#endif
+
+
 // #pragma mark - OffsetArray
 
 
@@ -209,10 +217,43 @@ PackageFileHeapAccessorBase::ReadDataToOutput(off_t offset, size_t size,
 	}
 
 	// allocate buffers for compressed and uncompressed data
-	uint16* compressedDataBuffer = (uint16*)malloc(kChunkSize);
-	uint16* uncompressedDataBuffer = (uint16*)malloc(kChunkSize);
-	MemoryDeleter compressedDataBufferDeleter(compressedDataBuffer);
-	MemoryDeleter uncompressedDataBufferDeleter(uncompressedDataBuffer);
+	uint16* compressedDataBuffer, *uncompressedDataBuffer;
+	MemoryDeleter compressedMemoryDeleter, uncompressedMemoryDeleter;
+
+#if defined(_KERNEL_MODE) && !defined(_BOOT_MODE)
+	struct ObjectCacheDeleter {
+		object_cache* cache;
+		void* object;
+
+		ObjectCacheDeleter(object_cache* c)
+			: cache(c)
+			, object(NULL)
+		{
+		}
+
+		~ObjectCacheDeleter()
+		{
+			if (cache != NULL && object != NULL)
+				object_cache_free(cache, object, 0);
+		}
+	};
+
+	ObjectCacheDeleter compressedCacheDeleter((object_cache*)sChunkCache),
+		uncompressedCacheDeleter((object_cache*)sChunkCache);
+	if (sChunkCache != NULL) {
+		compressedDataBuffer = (uint16*)object_cache_alloc((object_cache*)sChunkCache, 0);
+		uncompressedDataBuffer = (uint16*)object_cache_alloc((object_cache*)sChunkCache, 0);
+		compressedCacheDeleter.object = compressedDataBuffer;
+		uncompressedCacheDeleter.object = uncompressedDataBuffer;
+	} else
+#endif
+	{
+		compressedDataBuffer = (uint16*)malloc(kChunkSize);
+		uncompressedDataBuffer = (uint16*)malloc(kChunkSize);
+		compressedMemoryDeleter.SetTo(compressedDataBuffer);
+		uncompressedMemoryDeleter.SetTo(uncompressedDataBuffer);
+	}
+
 	if (compressedDataBuffer == NULL || uncompressedDataBuffer == NULL)
 		return B_NO_MEMORY;