| Message ID | 20241121015751.2300234-2-chao@kernel.org (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [f2fs-dev,1/2] f2fs: fix to shrink read extent node in batches | expand |
Chao Yu via Linux-f2fs-devel <linux-f2fs-devel@lists.sourceforge.net> 于2024年11月21日周四 10:01写道： > > Quoted: > "at this time, there are still 1086911 extent nodes in this zombie > extent tree that need to be cleaned up. > > crash_arm64_sprd_v8.0.3++> extent_tree.node_cnt ffffff80896cc500 > node_cnt = { > counter = 1086911 > }, > " > > As reported by Xiuhong, there will be a huge number of extent nodes > in extent tree, it may potentially cause: > - slab memory fragments > - extreme long time shrink on extent tree > - low mapping efficiency > > Let's add a sysfs node to limit max read extent count for each inode, > by default, value of this threshold is 10240, it can be updated > according to user's requirement. > > Reported-by: Xiuhong Wang <xiuhong.wang@unisoc.com> > Closes: https://lore.kernel.org/linux-f2fs-devel/20241112110627.1314632-1-xiuhong.wang@unisoc.com/ > Signed-off-by: Xiuhong Wang <xiuhong.wang@unisoc.com> > Signed-off-by: Zhiguo Niu <zhiguo.niu@unisoc.com> > Signed-off-by: Chao Yu <chao@kernel.org> > --- > Documentation/ABI/testing/sysfs-fs-f2fs | 6 ++++++ > fs/f2fs/extent_cache.c | 5 ++++- > fs/f2fs/f2fs.h | 4 ++++ > fs/f2fs/sysfs.c | 7 +++++++ > 4 files changed, 21 insertions(+), 1 deletion(-) > > diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs > index 513296bb6f29..3e1630c70d8a 100644 > --- a/Documentation/ABI/testing/sysfs-fs-f2fs > +++ b/Documentation/ABI/testing/sysfs-fs-f2fs > @@ -822,3 +822,9 @@ Description: It controls the valid block ratio threshold not to trigger excessiv > for zoned deivces. The initial value of it is 95(%). F2FS will stop the > background GC thread from intiating GC for sections having valid blocks > exceeding the ratio. > + > +What: /sys/fs/f2fs/<disk>/max_read_extent_count > +Date: November 2024 > +Contact: "Chao Yu" <chao@kernel.org> > +Description: It controls max read extent count for per-inode, the value of threshold > + is 10240 by default. > diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c > index b7a6817b44b0..347b3b647834 100644 > --- a/fs/f2fs/extent_cache.c > +++ b/fs/f2fs/extent_cache.c > @@ -717,7 +717,9 @@ static void __update_extent_tree_range(struct inode *inode, > } > > if (end < org_end && (type != EX_READ || > - org_end - end >= F2FS_MIN_EXTENT_LEN)) { > + (org_end - end >= F2FS_MIN_EXTENT_LEN && > + atomic_read(&et->node_cnt) < > + sbi->max_read_extent_count))) { > if (parts) { > __set_extent_info(&ei, > end, org_end - end, > @@ -1212,6 +1214,7 @@ void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi) > sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD; > sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD; > sbi->last_age_weight = LAST_AGE_WEIGHT; > + sbi->max_read_extent_count = DEF_MAX_READ_EXTENT_COUNT; > } > > int __init f2fs_create_extent_cache(void) > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h > index b65b023a588a..6f2cbf4c5740 100644 > --- a/fs/f2fs/f2fs.h > +++ b/fs/f2fs/f2fs.h > @@ -635,6 +635,9 @@ enum { > #define DEF_HOT_DATA_AGE_THRESHOLD 262144 > #define DEF_WARM_DATA_AGE_THRESHOLD 2621440 > > +/* default max read extent count per inode */ > +#define DEF_MAX_READ_EXTENT_COUNT 10240 > + > /* extent cache type */ > enum extent_type { > EX_READ, > @@ -1619,6 +1622,7 @@ struct f2fs_sb_info { > /* for extent tree cache */ > struct extent_tree_info extent_tree[NR_EXTENT_CACHES]; > atomic64_t allocated_data_blocks; /* for block age extent_cache */ > + unsigned int max_read_extent_count; /* max read extent count per inode */ > > /* The threshold used for hot and warm data seperation*/ > unsigned int hot_data_age_threshold; > diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c > index bdbf24db667b..d1356c656cac 100644 > --- a/fs/f2fs/sysfs.c > +++ b/fs/f2fs/sysfs.c > @@ -787,6 +787,13 @@ static ssize_t __sbi_store(struct f2fs_attr *a, > return count; > } > > + if (!strcmp(a->attr.name, "max_read_extent_count")) { > + if (t > UINT_MAX) > + return -EINVAL; > + *ui = (unsigned int)t; > + return count; > + } > + Hi Chao, from Xiuhong remider, it seems miss codes of adding sysfs node "max_read_extent_count"? F2FS_SBI_GENERAL_RW_ATTR(max_read_extent_count); ATTR_LIST(max_read_extent_count); thanks！ > if (!strcmp(a->attr.name, "ipu_policy")) { > if (t >= BIT(F2FS_IPU_MAX)) > return -EINVAL; > -- > 2.40.1 > > > > _______________________________________________ > Linux-f2fs-devel mailing list > Linux-f2fs-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
On 2024/11/22 13:23, Zhiguo Niu wrote: > Chao Yu via Linux-f2fs-devel <linux-f2fs-devel@lists.sourceforge.net> > 于2024年11月21日周四 10:01写道： >> >> Quoted: >> "at this time, there are still 1086911 extent nodes in this zombie >> extent tree that need to be cleaned up. >> >> crash_arm64_sprd_v8.0.3++> extent_tree.node_cnt ffffff80896cc500 >> node_cnt = { >> counter = 1086911 >> }, >> " >> >> As reported by Xiuhong, there will be a huge number of extent nodes >> in extent tree, it may potentially cause: >> - slab memory fragments >> - extreme long time shrink on extent tree >> - low mapping efficiency >> >> Let's add a sysfs node to limit max read extent count for each inode, >> by default, value of this threshold is 10240, it can be updated >> according to user's requirement. >> >> Reported-by: Xiuhong Wang <xiuhong.wang@unisoc.com> >> Closes: https://lore.kernel.org/linux-f2fs-devel/20241112110627.1314632-1-xiuhong.wang@unisoc.com/ >> Signed-off-by: Xiuhong Wang <xiuhong.wang@unisoc.com> >> Signed-off-by: Zhiguo Niu <zhiguo.niu@unisoc.com> >> Signed-off-by: Chao Yu <chao@kernel.org> >> --- >> Documentation/ABI/testing/sysfs-fs-f2fs | 6 ++++++ >> fs/f2fs/extent_cache.c | 5 ++++- >> fs/f2fs/f2fs.h | 4 ++++ >> fs/f2fs/sysfs.c | 7 +++++++ >> 4 files changed, 21 insertions(+), 1 deletion(-) >> >> diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs >> index 513296bb6f29..3e1630c70d8a 100644 >> --- a/Documentation/ABI/testing/sysfs-fs-f2fs >> +++ b/Documentation/ABI/testing/sysfs-fs-f2fs >> @@ -822,3 +822,9 @@ Description: It controls the valid block ratio threshold not to trigger excessiv >> for zoned deivces. The initial value of it is 95(%). F2FS will stop the >> background GC thread from intiating GC for sections having valid blocks >> exceeding the ratio. >> + >> +What: /sys/fs/f2fs/<disk>/max_read_extent_count >> +Date: November 2024 >> +Contact: "Chao Yu" <chao@kernel.org> >> +Description: It controls max read extent count for per-inode, the value of threshold >> + is 10240 by default. >> diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c >> index b7a6817b44b0..347b3b647834 100644 >> --- a/fs/f2fs/extent_cache.c >> +++ b/fs/f2fs/extent_cache.c >> @@ -717,7 +717,9 @@ static void __update_extent_tree_range(struct inode *inode, >> } >> >> if (end < org_end && (type != EX_READ || >> - org_end - end >= F2FS_MIN_EXTENT_LEN)) { >> + (org_end - end >= F2FS_MIN_EXTENT_LEN && >> + atomic_read(&et->node_cnt) < >> + sbi->max_read_extent_count))) { >> if (parts) { >> __set_extent_info(&ei, >> end, org_end - end, >> @@ -1212,6 +1214,7 @@ void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi) >> sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD; >> sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD; >> sbi->last_age_weight = LAST_AGE_WEIGHT; >> + sbi->max_read_extent_count = DEF_MAX_READ_EXTENT_COUNT; >> } >> >> int __init f2fs_create_extent_cache(void) >> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h >> index b65b023a588a..6f2cbf4c5740 100644 >> --- a/fs/f2fs/f2fs.h >> +++ b/fs/f2fs/f2fs.h >> @@ -635,6 +635,9 @@ enum { >> #define DEF_HOT_DATA_AGE_THRESHOLD 262144 >> #define DEF_WARM_DATA_AGE_THRESHOLD 2621440 >> >> +/* default max read extent count per inode */ >> +#define DEF_MAX_READ_EXTENT_COUNT 10240 >> + >> /* extent cache type */ >> enum extent_type { >> EX_READ, >> @@ -1619,6 +1622,7 @@ struct f2fs_sb_info { >> /* for extent tree cache */ >> struct extent_tree_info extent_tree[NR_EXTENT_CACHES]; >> atomic64_t allocated_data_blocks; /* for block age extent_cache */ >> + unsigned int max_read_extent_count; /* max read extent count per inode */ >> >> /* The threshold used for hot and warm data seperation*/ >> unsigned int hot_data_age_threshold; >> diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c >> index bdbf24db667b..d1356c656cac 100644 >> --- a/fs/f2fs/sysfs.c >> +++ b/fs/f2fs/sysfs.c >> @@ -787,6 +787,13 @@ static ssize_t __sbi_store(struct f2fs_attr *a, >> return count; >> } >> >> + if (!strcmp(a->attr.name, "max_read_extent_count")) { >> + if (t > UINT_MAX) >> + return -EINVAL; >> + *ui = (unsigned int)t; >> + return count; >> + } >> + > Hi Chao, > from Xiuhong remider, it seems miss codes of adding sysfs node > "max_read_extent_count"? > F2FS_SBI_GENERAL_RW_ATTR(max_read_extent_count); > ATTR_LIST(max_read_extent_count); Oops, let me fix this soon. Thanks, > thanks！ >> if (!strcmp(a->attr.name, "ipu_policy")) { >> if (t >= BIT(F2FS_IPU_MAX)) >> return -EINVAL; >> -- >> 2.40.1 >> >> >> >> _______________________________________________ >> Linux-f2fs-devel mailing list >> Linux-f2fs-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index 513296bb6f29..3e1630c70d8a 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -822,3 +822,9 @@ Description: It controls the valid block ratio threshold not to trigger excessiv for zoned deivces. The initial value of it is 95(%). F2FS will stop the background GC thread from intiating GC for sections having valid blocks exceeding the ratio. + +What: /sys/fs/f2fs/<disk>/max_read_extent_count +Date: November 2024 +Contact: "Chao Yu" <chao@kernel.org> +Description: It controls max read extent count for per-inode, the value of threshold + is 10240 by default. diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c index b7a6817b44b0..347b3b647834 100644 --- a/fs/f2fs/extent_cache.c +++ b/fs/f2fs/extent_cache.c @@ -717,7 +717,9 @@ static void __update_extent_tree_range(struct inode *inode, } if (end < org_end && (type != EX_READ || - org_end - end >= F2FS_MIN_EXTENT_LEN)) { + (org_end - end >= F2FS_MIN_EXTENT_LEN && + atomic_read(&et->node_cnt) < + sbi->max_read_extent_count))) { if (parts) { __set_extent_info(&ei, end, org_end - end, @@ -1212,6 +1214,7 @@ void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi) sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD; sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD; sbi->last_age_weight = LAST_AGE_WEIGHT; + sbi->max_read_extent_count = DEF_MAX_READ_EXTENT_COUNT; } int __init f2fs_create_extent_cache(void) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index b65b023a588a..6f2cbf4c5740 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -635,6 +635,9 @@ enum { #define DEF_HOT_DATA_AGE_THRESHOLD 262144 #define DEF_WARM_DATA_AGE_THRESHOLD 2621440 +/* default max read extent count per inode */ +#define DEF_MAX_READ_EXTENT_COUNT 10240 + /* extent cache type */ enum extent_type { EX_READ, @@ -1619,6 +1622,7 @@ struct f2fs_sb_info { /* for extent tree cache */ struct extent_tree_info extent_tree[NR_EXTENT_CACHES]; atomic64_t allocated_data_blocks; /* for block age extent_cache */ + unsigned int max_read_extent_count; /* max read extent count per inode */ /* The threshold used for hot and warm data seperation*/ unsigned int hot_data_age_threshold; diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index bdbf24db667b..d1356c656cac 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -787,6 +787,13 @@ static ssize_t __sbi_store(struct f2fs_attr *a, return count; } + if (!strcmp(a->attr.name, "max_read_extent_count")) { + if (t > UINT_MAX) + return -EINVAL; + *ui = (unsigned int)t; + return count; + } + if (!strcmp(a->attr.name, "ipu_policy")) { if (t >= BIT(F2FS_IPU_MAX)) return -EINVAL;
Quoted: "at this time, there are still 1086911 extent nodes in this zombie extent tree that need to be cleaned up. crash_arm64_sprd_v8.0.3++> extent_tree.node_cnt ffffff80896cc500 node_cnt = { counter = 1086911 }, " As reported by Xiuhong, there will be a huge number of extent nodes in extent tree, it may potentially cause: - slab memory fragments - extreme long time shrink on extent tree - low mapping efficiency Let's add a sysfs node to limit max read extent count for each inode, by default, value of this threshold is 10240, it can be updated according to user's requirement. Reported-by: Xiuhong Wang <xiuhong.wang@unisoc.com> Closes: https://lore.kernel.org/linux-f2fs-devel/20241112110627.1314632-1-xiuhong.wang@unisoc.com/ Signed-off-by: Xiuhong Wang <xiuhong.wang@unisoc.com> Signed-off-by: Zhiguo Niu <zhiguo.niu@unisoc.com> Signed-off-by: Chao Yu <chao@kernel.org> --- Documentation/ABI/testing/sysfs-fs-f2fs | 6 ++++++ fs/f2fs/extent_cache.c | 5 ++++- fs/f2fs/f2fs.h | 4 ++++ fs/f2fs/sysfs.c | 7 +++++++ 4 files changed, 21 insertions(+), 1 deletion(-)