Use hash func to boost file creation and lookup

Introduce a hash-based mechanism to speed up file creation and lookup
operations. The hash function enables faster access to extent and logical
block extent index, improving overall filesystem performance.

hash_code = file_hash(file_name);
extent index = hash_code / SIMPLEFS_MAX_BLOCKS_PER_EXTENT
block index = hash_code % SIMPLEFS_MAX_BLOCKS_PER_EXTENT;

Author: RoyWFHuang

.config

@@ -0,0 +1,11 @@
+#
+# Automatically generated file; DO NOT EDIT.
+# Configuration
+#
+
+#
+# Simplefs Configuration
+#
+# CONFIG_SIMPLEFS_HASH_HASH64 is not set
+CONFIG_SIMPLEFS_HASH_FNV1A=y
+# CONFIG_SIMPLEFS_HASH_FULLNAME is not set

Kconfig

@@ -0,0 +1,17 @@
+menu "Simplefs Configuration"
+
+choice
+    prompt "Select hash method"
+    default SIMPLEFS_HASH_HASH64
+
+config SIMPLEFS_HASH_HASH64
+    bool "Use hash64 (CRC32 + hash_64)"
+
+config SIMPLEFS_HASH_FNV1A
+    bool "Use FNV-1a 64-bit hash"
+
+config SIMPLEFS_HASH_FULLNAME
+    bool "Use full_name_hash (Linux built-in)"
+endchoice
+
+endmenu

Makefile

@@ -1,11 +1,14 @@
 obj-m += simplefs.o
-simplefs-objs := fs.o super.o inode.o file.o dir.o extent.o
-
+simplefs-objs := fs.o super.o inode.o file.o dir.o extent.o symlink.o hash.o
 KDIR ?= /lib/modules/$(shell uname -r)/build
 
+CONFIG := .config
+CONFIG_H := config.h
+
 MKFS = mkfs.simplefs
+MENUCONFIG := kconfig-mconf
 
-all: $(MKFS)
+all: $(MKFS) $(CONFIG_H)
 	make -C $(KDIR) M=$(PWD) modules
 
 IMAGE ?= test.img
@@ -23,6 +26,26 @@ $(IMAGE): $(MKFS)
 	dd if=/dev/zero of=${IMAGE} bs=1M count=${IMAGESIZE}
 	./$< $(IMAGE)
 
+$(CONFIG_H): $(CONFIG)
+	@echo "/*" > $(CONFIG_H)
+	@echo " * Automatically generated file; DO NOT EDIT." >> $(CONFIG_H)
+	@echo " * Configuration" >> $(CONFIG_H)
+	@echo " */" >> $(CONFIG_H)
+	@echo "" >> $(CONFIG_H)
+	@sed \
+	    -e '/^$$/d' \
+	    -e '/^[[:space:]]*# *$$/d' \
+	    -e '/^[[:space:]]*# Automatically/d' \
+	    -e '/^[[:space:]]*# Configuration/d' \
+	    -e '/^[[:space:]]*# \(CONFIG_[A-Z0-9_]*\) is not set/ s//#undef \1/' \
+	    -e '/^[[:space:]]*CONFIG_[A-Z0-9_]*=y/ s/^[[:space:]]*\(CONFIG_[A-Z0-9_]*\)=y/#define \1 1/' \
+	    -e '/^[[:space:]]*CONFIG_[A-Z0-9_]*=[0-9]\+/ s/^[[:space:]]*\(CONFIG_[A-Z0-9_]*\)=\([0-9]*\)/#define \1 \2/' \
+	    -e '/^[[:space:]]*# / s/^[[:space:]]*# \(.*\)/\/\* \1 \*\//' \
+	    $(CONFIG) >> $(CONFIG_H)
+
+menuconfig:
+	$(MENUCONFIG) Kconfig
+
 journal: $(JOURNAL)
 
 $(JOURNAL):
@@ -35,6 +58,9 @@ check: all
 clean:
 	make -C $(KDIR) M=$(PWD) clean
 	rm -f *~ $(PWD)/*.ur-safe
-	rm -f $(MKFS) $(IMAGE) $(JOURNAL)
+	rm -f $(MKFS) $(IMAGE) $(JOURNAL) config.h
+
+distclean: clean
+	rm -f $(CONFIG)
 
-.PHONY: all clean journal
+.PHONY: all clean journal menuconfig distclean

bitmap.h

@@ -48,11 +48,11 @@ static inline uint32_t get_free_blocks(struct super_block *sb, uint32_t len)
     struct simplefs_sb_info *sbi = SIMPLEFS_SB(sb);
     uint32_t ret = get_first_free_bits(sbi->bfree_bitmap, sbi->nr_blocks, len);
     uint32_t i;
+    struct buffer_head *bh;
     if (!ret) /* No enough free blocks */
         return 0;
 
     sbi->nr_free_blocks -= len;
-    struct buffer_head *bh;
     for (i = 0; i < len; i++) {
         bh = sb_bread(sb, ret + i);
         if (!bh) {

dir.c

@@ -81,7 +81,7 @@ static int simplefs_iterate(struct file *dir, struct dir_context *ctx)
                 continue;
             }
 
-            for (fi = 0; fi < SIMPLEFS_FILES_PER_BLOCK;) {
+            for (fi = 0; fi < SIMPLEFS_FILES_PER_BLOCK && dblock->nr_files;) {
                 if (dblock->files[fi].inode != 0) {
                     if (offset) {
                         offset--;

hash.c

@@ -0,0 +1,40 @@
+#include <linux/crc32.h>
+#include <linux/hash.h>
+#include <linux/stringhash.h>
+#include <linux/types.h>
+#include "config.h"
+#include "simplefs.h"
+
+
+#if defined(CONFIG_SIMPLEFS_HASH_HASH64)
+static uint64_t _hash64(const char *name, size_t len)
+{
+    uint32_t val = crc32(0, name, len);
+    return hash_64(
+        val, ilog2(SIMPLEFS_MAX_EXTENTS * SIMPLEFS_MAX_BLOCKS_PER_EXTENT));
+}
+#elif defined(CONFIG_SIMPLEFS_HASH_FNV1A)
+static uint64_t _fnv1a_64(const char *str)
+{
+    uint64_t h = 0xcbf29ce484222325ULL;
+    while (*str) {
+        h ^= (unsigned char) (*str++);
+        h *= 0x100000001b3ULL;
+    }
+    return h;
+}
+#endif
+
+uint32_t simplefs_hash(struct dentry *dentry)
+{
+#if defined(CONFIG_SIMPLEFS_HASH_HASH64)
+    return _hash64(dentry->d_name.name, strlen(dentry->d_name.name));
+#elif defined(CONFIG_SIMPLEFS_HASH_FNV1A)
+    return _fnv1a_64(dentry->d_name.name);
+#elif defined(CONFIG_SIMPLEFS_HASH_FULLNAME)
+    return full_name_hash(dentry, dentry->d_name.name,
+                          strlen(dentry->d_name.name));
+#else
+#error "No hash method selected!"
+#endif
+}