Orange Pi5 kernel

^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   3)  *  fs/ext4/mballoc.h
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   4)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   5)  *  Written by: Alex Tomas <alex@clusterfs.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   6)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   7)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   8) #ifndef _EXT4_MBALLOC_H
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   9) #define _EXT4_MBALLOC_H
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  10) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  11) #include <linux/time.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  12) #include <linux/fs.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  13) #include <linux/namei.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  14) #include <linux/quotaops.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  15) #include <linux/buffer_head.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  16) #include <linux/module.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  17) #include <linux/swap.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  18) #include <linux/proc_fs.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  19) #include <linux/pagemap.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  20) #include <linux/seq_file.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  21) #include <linux/blkdev.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  22) #include <linux/mutex.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  23) #include "ext4_jbd2.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  24) #include "ext4.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  25) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  26) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  27)  * mb_debug() dynamic printk msgs could be used to debug mballoc code.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  28)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  29) #ifdef CONFIG_EXT4_DEBUG
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  30) #define mb_debug(sb, fmt, ...)						\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  31) 	pr_debug("[%s/%d] EXT4-fs (%s): (%s, %d): %s: " fmt,		\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  32) 		current->comm, task_pid_nr(current), sb->s_id,		\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  33) 	       __FILE__, __LINE__, __func__, ##__VA_ARGS__)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  34) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  35) #define mb_debug(sb, fmt, ...)	no_printk(fmt, ##__VA_ARGS__)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  36) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  37) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  38) #define EXT4_MB_HISTORY_ALLOC		1	/* allocation */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  39) #define EXT4_MB_HISTORY_PREALLOC	2	/* preallocated blocks used */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  40) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  41) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  42)  * How long mballoc can look for a best extent (in found extents)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  43)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  44) #define MB_DEFAULT_MAX_TO_SCAN		200
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  45) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  46) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  47)  * How long mballoc must look for a best extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  48)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  49) #define MB_DEFAULT_MIN_TO_SCAN		10
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  50) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  51) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  52)  * with 'ext4_mb_stats' allocator will collect stats that will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  53)  * shown at umount. The collecting costs though!
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  54)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  55) #define MB_DEFAULT_STATS		0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  56) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  57) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  58)  * files smaller than MB_DEFAULT_STREAM_THRESHOLD are served
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  59)  * by the stream allocator, which purpose is to pack requests
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  60)  * as close each to other as possible to produce smooth I/O traffic
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  61)  * We use locality group prealloc space for stream request.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  62)  * We can tune the same via /proc/fs/ext4/<parition>/stream_req
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  63)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  64) #define MB_DEFAULT_STREAM_THRESHOLD	16	/* 64K */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  65) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  66) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  67)  * for which requests use 2^N search using buddies
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  68)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  69) #define MB_DEFAULT_ORDER2_REQS		2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  70) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  71) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  72)  * default group prealloc size 512 blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  73)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  74) #define MB_DEFAULT_GROUP_PREALLOC	512
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  75) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  76) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  77)  * maximum length of inode prealloc list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  78)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  79) #define MB_DEFAULT_MAX_INODE_PREALLOC	512
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  80) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  81) struct ext4_free_data {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  82) 	/* this links the free block information from sb_info */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  83) 	struct list_head		efd_list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  84) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  85) 	/* this links the free block information from group_info */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  86) 	struct rb_node			efd_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  87) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  88) 	/* group which free block extent belongs */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  89) 	ext4_group_t			efd_group;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  90) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  91) 	/* free block extent */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  92) 	ext4_grpblk_t			efd_start_cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  93) 	ext4_grpblk_t			efd_count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  94) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  95) 	/* transaction which freed this extent */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  96) 	tid_t				efd_tid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  97) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  98) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  99) struct ext4_prealloc_space {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) 	struct list_head	pa_inode_list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) 	struct list_head	pa_group_list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) 	union {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) 		struct list_head pa_tmp_list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) 		struct rcu_head	pa_rcu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) 	} u;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) 	spinlock_t		pa_lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) 	atomic_t		pa_count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) 	unsigned		pa_deleted;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) 	ext4_fsblk_t		pa_pstart;	/* phys. block */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) 	ext4_lblk_t		pa_lstart;	/* log. block */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) 	ext4_grpblk_t		pa_len;		/* len of preallocated chunk */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) 	ext4_grpblk_t		pa_free;	/* how many blocks are free */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) 	unsigned short		pa_type;	/* pa type. inode or group */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) 	spinlock_t		*pa_obj_lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) 	struct inode		*pa_inode;	/* hack, for history only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) enum {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) 	MB_INODE_PA = 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) 	MB_GROUP_PA = 1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) struct ext4_free_extent {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) 	ext4_lblk_t fe_logical;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) 	ext4_grpblk_t fe_start;	/* In cluster units */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) 	ext4_group_t fe_group;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) 	ext4_grpblk_t fe_len;	/* In cluster units */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131)  * Locality group:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132)  *   we try to group all related changes together
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133)  *   so that writeback can flush/allocate them together as well
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134)  *   Size of lg_prealloc_list hash is determined by MB_DEFAULT_GROUP_PREALLOC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135)  *   (512). We store prealloc space into the hash based on the pa_free blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136)  *   order value.ie, fls(pa_free)-1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) #define PREALLOC_TB_SIZE 10
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) struct ext4_locality_group {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) 	/* for allocator */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) 	/* to serialize allocates */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) 	struct mutex		lg_mutex;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) 	/* list of preallocations */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) 	struct list_head	lg_prealloc_list[PREALLOC_TB_SIZE];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) 	spinlock_t		lg_prealloc_lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) struct ext4_allocation_context {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) 	struct inode *ac_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) 	struct super_block *ac_sb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) 	/* original request */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) 	struct ext4_free_extent ac_o_ex;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) 	/* goal request (normalized ac_o_ex) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) 	struct ext4_free_extent ac_g_ex;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) 	/* the best found extent */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) 	struct ext4_free_extent ac_b_ex;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) 	/* copy of the best found extent taken before preallocation efforts */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) 	struct ext4_free_extent ac_f_ex;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) 	__u16 ac_groups_scanned;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) 	__u16 ac_found;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) 	__u16 ac_tail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) 	__u16 ac_buddy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) 	__u16 ac_flags;		/* allocation hints */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) 	__u8 ac_status;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) 	__u8 ac_criteria;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) 	__u8 ac_2order;		/* if request is to allocate 2^N blocks and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) 				 * N > 0, the field stores N, otherwise 0 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) 	__u8 ac_op;		/* operation, for history only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) 	struct page *ac_bitmap_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) 	struct page *ac_buddy_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) 	struct ext4_prealloc_space *ac_pa;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) 	struct ext4_locality_group *ac_lg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) #define AC_STATUS_CONTINUE	1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) #define AC_STATUS_FOUND		2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) #define AC_STATUS_BREAK		3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) struct ext4_buddy {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) 	struct page *bd_buddy_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) 	void *bd_buddy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) 	struct page *bd_bitmap_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) 	void *bd_bitmap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) 	struct ext4_group_info *bd_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) 	struct super_block *bd_sb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) 	__u16 bd_blkbits;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) 	ext4_group_t bd_group;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) static inline ext4_fsblk_t ext4_grp_offs_to_block(struct super_block *sb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) 					struct ext4_free_extent *fex)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) 	return ext4_group_first_block_no(sb, fex->fe_group) +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) 		(fex->fe_start << EXT4_SB(sb)->s_cluster_bits);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) typedef int (*ext4_mballoc_query_range_fn)(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) 	struct super_block		*sb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) 	ext4_group_t			agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) 	ext4_grpblk_t			start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) 	ext4_grpblk_t			len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) 	void				*priv);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) ext4_mballoc_query_range(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) 	struct super_block		*sb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) 	ext4_group_t			agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) 	ext4_grpblk_t			start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) 	ext4_grpblk_t			end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) 	ext4_mballoc_query_range_fn	formatter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) 	void				*priv);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) #endif