Readdir improvements
Subsystems (2)
[email protected]
maintainerv9 cover Readdir improvements
From: Trond Myklebust <[email protected]>
The current NFS readdir code will always try to maximise the amount of readahead it performs on the assumption that we can cache anything that isn't immediately read by the process.
There are several cases where this assumption breaks down, including when the 'ls -l' heuristic kicks in to try to force use of readdirplus as a batch replacement for lookup/getattr.
This series also implement Ben's page cache filter to ensure that we can improve the ability to share cached data between processes that are reading the same directory at the same time, and to avoid live-locks when the directory is simultaneously changing.
--
v2: Remove reset of dtsize when NFS_INO_FORCE_READDIR is set
v3: Avoid excessive window shrinking in uncached_readdir case
v4: Track 'ls -l' cache hit/miss statistics
Improved algorithm for falling back to uncached readdir Skip readdirplus when files are being written to
v5: bugfixes
Skip readdirplus when the acdirmax/acregmax values are low Request a full XDR buffer when doing READDIRPLUS
v6: Add tracing
Don't have lookup request readdirplus when it won't help
v7: Implement Ben's page cache filter
Reduce the use of uncached readdir
Change indexing of the page cache to improve seekdir() performance.
v8: Reduce the page cache overhead by shrinking the cookie hashvalue size
Incorporate other feedback from Anna, Ben and Neil
Fix nfs2/3_decode_dirent return values
Fix the change attribute value set in nfs_readdir_page_get_next()
v9: Address bugs that were hit when testing with large directories
Misc cleanups
Trond Myklebust (27):
NFS: Return valid errors from nfs2/3_decode_dirent()
NFS: constify nfs_server_capable() and nfs_have_writebacks()
NFS: Trace lookup revalidation failure
NFS: Initialise the readdir verifier as best we can in nfs_opendir()
NFS: Use kzalloc() to avoid initialising the nfs_open_dir_context
NFS: Calculate page offsets algorithmically
NFS: Store the change attribute in the directory page cache
NFS: Don't re-read the entire page cache to find the next cookie
NFS: Don't advance the page pointer unless the page is full
NFS: Adjust the amount of readahead performed by NFS readdir
NFS: If the cookie verifier changes, we must invalidate the page cache
NFS: Simplify nfs_readdir_xdr_to_array()
NFS: Reduce use of uncached readdir
NFS: Improve heuristic for readdirplus
NFS: Don't ask for readdirplus unless it can help nfs_getattr()
NFSv4: Ask for a full XDR buffer of readdir goodness
NFS: Readdirplus can't help lookup for case insensitive filesystems
NFS: Don't request readdirplus when revalidation was forced
NFS: Add basic readdir tracing
NFS: Trace effects of readdirplus on the dcache
NFS: Trace effects of the readdirplus heuristic
NFS: Clean up page array initialisation/free
NFS: Convert readdir page cache to use a cookie based index
NFS: Fix up forced readdirplus
NFS: Remove unnecessary cache invalidations for directories
NFS: Optimise away the previous cookie field
NFS: Cache all entries in the readdirplus reply
fs/nfs/Kconfig | 4 + fs/nfs/dir.c | 602 ++++++++++++++++++++++++---------------- fs/nfs/inode.c | 46 ++- fs/nfs/internal.h | 4 +- fs/nfs/nfs2xdr.c | 3 +- fs/nfs/nfs3xdr.c | 29 +- fs/nfs/nfs4proc.c | 2 - fs/nfs/nfs4xdr.c | 7 +- fs/nfs/nfstrace.h | 123 +++++++- include/linux/nfs_fs.h | 19 +- include/linux/nfs_xdr.h | 3 +- 11 files changed, 532 insertions(+), 310 deletions(-)
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 1/27 NFS: Return valid errors from nfs2/3_decode_dirent()
From: Trond Myklebust <[email protected]>
Valid return values for decode_dirent() callback functions are:
0: Success
-EBADCOOKIE: End of directory
-EAGAIN: End of xdr_stream
All errors need to map into one of those three values.
--- fs/nfs/nfs2xdr.c | 2 +- fs/nfs/nfs3xdr.c | 21 ++++++--------------- 2 files changed, 7 insertions(+), 16 deletions(-)
fs/nfs/nfs2xdr.c +1 -1
@@ -949,7 +949,7 @@ int nfs2_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry, error =decode_filename_inline (xdr, &entry->name, &entry->len);if (unlikely (error)) -return error; +return -EAGAIN ;/* * The type (size and byte order) of nfscookie isn't defined in
fs/nfs/nfs3xdr.c +6 -15
@@ -1967,7 +1967,6 @@ int nfs3_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry, bool plus) {struct user_namespace *userns =rpc_userns (entry->server->client); -struct nfs_entry old = *entry;__be32 *p;int error;u64 new_cookie;
@@ -1987,15 +1986,15 @@ int nfs3_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry, error =decode_fileid3 (xdr, &entry->ino);if (unlikely (error)) -return error; +return -EAGAIN ; error =decode_inline_filename3 (xdr, &entry->name, &entry->len);if (unlikely (error)) -return error; +return -EAGAIN ; error =decode_cookie3 (xdr, &new_cookie);if (unlikely (error)) -return error; +return -EAGAIN ; entry->d_type =DT_UNKNOWN ;
@@ -2003,7 +2002,7 @@ int nfs3_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry, entry->fattr->valid =0 ; error =decode_post_op_attr (xdr, entry->fattr, userns);if (unlikely (error)) -return error; +return -EAGAIN ;if (entry->fattr->valid &NFS_ATTR_FATTR_V3 ) entry->d_type =nfs_umode_to_dtype (entry->fattr->mode);
@@ -2018,11 +2017,8 @@ int nfs3_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry,return -EAGAIN ;if (*p != xdr_zero) { error =decode_nfs_fh3 (xdr, entry->fh); -if (unlikely (error)) { -if (error == -E2BIG ) - goto out_truncated; -return error; - } +if (unlikely (error)) +return -EAGAIN ; }else zero_nfs_fh3 (entry->fh); }
@@ -2031,11 +2027,6 @@ int nfs3_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry, entry->cookie = new_cookie;return 0 ; - -out_truncated: -dprintk ("NFS: directory entry contains invalid file handle\n" ); - *entry = old; -return -EAGAIN ; }/*
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 2/27 NFS: constify nfs_server_capable() and nfs_have_writebacks()
From: Trond Myklebust <[email protected]>
--- include/linux/nfs_fs.h | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)
include/linux/nfs_fs.h +3 -4
@@ -363,7 +363,7 @@ static inline void nfs_mark_for_revalidate(struct inode *inode)spin_unlock (&inode->i_lock); } -static inline int nfs_server_capable (struct inode *inode,int cap) +static inline int nfs_server_capable (const struct inode *inode,int cap) {return NFS_SERVER (inode)->caps & cap; }
@@ -587,12 +587,11 @@ extern struct nfs_commit_data *nfs_commitdata_alloc(bool never_fail);extern void nfs_commit_free (struct nfs_commit_data *data);bool nfs_commit_end (struct nfs_mds_commit_info *cinfo); -static inline int -nfs_have_writebacks (struct inode *inode) +static inline bool nfs_have_writebacks (const struct inode *inode) {if (S_ISREG (inode->i_mode))return atomic_long_read (&NFS_I (inode)->nrequests) !=0 ; -return 0 ; +return false; }/*
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 4/27 NFS: Initialise the readdir verifier as best we can in nfs_opendir()
In reply to [email protected]From: Trond Myklebust <[email protected]>
For the purpose of ensuring that opendir() followed by seekdir() work as correctly as possible, try to initialise the readdir verifier in nfs_opendir().
--- fs/nfs/dir.c | 1 + 1 file changed, 1 insertion(+)
fs/nfs/dir.c +1 -0
@@ -89,6 +89,7 @@ static struct nfs_open_dir_context *alloc_nfs_open_dir_context(struct inode *dir NFS_INO_REVAL_FORCED);list_add (&ctx->list, &nfsi->open_files);clear_bit (NFS_INO_FORCE_READDIR , &nfsi->flags); +memcpy (ctx->verf, nfsi->cookieverf,sizeof (ctx->verf));spin_unlock (&dir->i_lock);return ctx; }
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 7/27 NFS: Store the change attribute in the directory page cache
In reply to [email protected]From: Trond Myklebust <[email protected]>
Use the change attribute and the first cookie in a directory page cache entry to validate that the page is up to date.
--- fs/nfs/dir.c | 68 ++++++++++++++++++++++++++++------------------------ 1 file changed, 37 insertions(+), 31 deletions(-)
fs/nfs/dir.c +37 -31
@@ -140,6 +140,7 @@ struct nfs_cache_array_entry { };struct nfs_cache_array { +u64 change_attr;u64 last_cookie;unsigned int size; unsigned char page_full : 1,
@@ -176,12 +177,14 @@ static void nfs_readdir_array_init(struct nfs_cache_array *array)memset (array,0 ,sizeof (struct nfs_cache_array )); } -static void nfs_readdir_page_init_array (struct page *page,u64 last_cookie) +static void nfs_readdir_page_init_array (struct page *page,u64 last_cookie, +u64 change_attr) {struct nfs_cache_array *array; array =kmap_atomic (page);nfs_readdir_array_init (array); + array->change_attr = change_attr; array->last_cookie = last_cookie; array->cookies_are_ordered =1 ;kunmap_atomic (array);
@@ -208,7 +211,7 @@ nfs_readdir_page_array_alloc(u64 last_cookie, gfp_t gfp_flags) {struct page *page =alloc_page (gfp_flags);if (page) -nfs_readdir_page_init_array (page, last_cookie); +nfs_readdir_page_init_array (page, last_cookie,0 );return page; }
@@ -305,19 +308,43 @@ int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page)return ret; } +static bool nfs_readdir_page_validate (struct page *page,u64 last_cookie, +u64 change_attr) +{ +struct nfs_cache_array *array =kmap_atomic (page); +int ret = true; + +if (array->change_attr != change_attr) + ret = false; +if (array->size >0 && array->array[0 ].cookie != last_cookie) + ret = false; +kunmap_atomic (array); +return ret; +} + +static void nfs_readdir_page_unlock_and_put (struct page *page) +{ +unlock_page (page); +put_page (page); +} +static struct page *nfs_readdir_page_get_locked (struct address_space *mapping,pgoff_t index,u64 last_cookie) {struct page *page; +u64 change_attr; page =grab_cache_page (mapping, index); -if (page && !PageUptodate (page)) { -nfs_readdir_page_init_array (page, last_cookie); -if (invalidate_inode_pages2_range (mapping, index +1 ,-1 ) <0 ) -nfs_zap_mapping (mapping->host, mapping); -SetPageUptodate (page); +if (!page) +return NULL ; + change_attr =inode_peek_iversion_raw (mapping->host); +if (PageUptodate (page)) { +if (nfs_readdir_page_validate (page, last_cookie, change_attr)) +return page; +nfs_readdir_clear_array (page); } - +nfs_readdir_page_init_array (page, last_cookie, change_attr); +SetPageUptodate (page);return page; }
@@ -357,12 +384,6 @@ static void nfs_readdir_page_set_eof(struct page *page)kunmap_atomic (array); } -static void nfs_readdir_page_unlock_and_put (struct page *page) -{ -unlock_page (page); -put_page (page); -} -static struct page *nfs_readdir_page_get_next (struct address_space *mapping,pgoff_t index,u64 cookie) {
@@ -419,16 +440,6 @@ static int nfs_readdir_search_for_pos(struct nfs_cache_array *array,return -EBADCOOKIE ; } -static bool -nfs_readdir_inode_mapping_valid (struct nfs_inode *nfsi) -{ -if (nfsi->cache_validity & (NFS_INO_INVALID_CHANGE | -NFS_INO_INVALID_DATA )) -return false; -smp_rmb (); -return !test_bit (NFS_INO_INVALIDATING , &nfsi->flags); -} -static bool nfs_readdir_array_cookie_in_range (struct nfs_cache_array *array,u64 cookie) {
@@ -457,8 +468,7 @@ static int nfs_readdir_search_for_cookie(struct nfs_cache_array *array,struct nfs_inode *nfsi =NFS_I (file_inode (desc->file)); new_pos =nfs_readdir_page_offset (desc->page) + i; -if (desc->attr_gencount != nfsi->attr_gencount || - !nfs_readdir_inode_mapping_valid (nfsi)) { +if (desc->attr_gencount != nfsi->attr_gencount) { desc->duped =0 ; desc->attr_gencount = nfsi->attr_gencount; }else if (new_pos < desc->prev_index) {
@@ -1095,11 +1105,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)* to either find the entry with the appropriate number or * revalidate the cookie. */ -if (ctx->pos ==0 ||nfs_attribute_cache_expired (inode)) { - res =nfs_revalidate_mapping (inode, file->f_mapping); -if (res <0 ) - goto out; - } +nfs_revalidate_inode (inode,NFS_INO_INVALID_CHANGE ); res = -ENOMEM ; desc =kzalloc (sizeof (*desc),GFP_KERNEL );
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 5/27 NFS: Use kzalloc() to avoid initialising the nfs_open_dir_context
In reply to [email protected]From: Trond Myklebust <[email protected]>
--- fs/nfs/dir.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-)
fs/nfs/dir.c +4 -7
@@ -69,18 +69,15 @@ const struct address_space_operations nfs_dir_aops = { .freepage = nfs_readdir_clear_array, }; -static struct nfs_open_dir_context *alloc_nfs_open_dir_context (struct inode *dir) +static struct nfs_open_dir_context * +alloc_nfs_open_dir_context (struct inode *dir) {struct nfs_inode *nfsi =NFS_I (dir);struct nfs_open_dir_context *ctx; - ctx =kmalloc (sizeof (*ctx),GFP_KERNEL_ACCOUNT ); + + ctx =kzalloc (sizeof (*ctx),GFP_KERNEL_ACCOUNT );if (ctx !=NULL ) { - ctx->duped =0 ; ctx->attr_gencount = nfsi->attr_gencount; - ctx->dir_cookie =0 ; - ctx->dup_cookie =0 ; - ctx->page_index =0 ; - ctx->eof = false;spin_lock (&dir->i_lock);if (list_empty (&nfsi->open_files) && (nfsi->cache_validity &NFS_INO_DATA_INVAL_DEFER ))
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 6/27 NFS: Calculate page offsets algorithmically
From: Trond Myklebust <[email protected]>
Instead of relying on counting the page offsets as we walk through the page cache, switch to calculating them algorithmically.
--- fs/nfs/dir.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-)
fs/nfs/dir.c +13 -5
@@ -249,17 +249,20 @@ static const char *nfs_readdir_copy_name(const char *name, unsigned int len)return ret; } +static size_t nfs_readdir_array_maxentries (void ) +{ +return (PAGE_SIZE -sizeof (struct nfs_cache_array )) / +sizeof (struct nfs_cache_array_entry ); +} +/* * Check that the next array entry lies entirely within the page bounds */ static int nfs_readdir_array_can_expand (struct nfs_cache_array *array) { -struct nfs_cache_array_entry *cache_entry; -if (array->page_full)return -ENOSPC ; - cache_entry = &array->array[array->size +1 ]; -if ((char *)cache_entry - (char *)array >PAGE_SIZE ) { +if (array->size ==nfs_readdir_array_maxentries ()) { array->page_full =1 ;return -ENOSPC ; }
@@ -318,6 +321,11 @@ static struct page *nfs_readdir_page_get_locked(struct address_space *mapping,return page; } +static loff_t nfs_readdir_page_offset (struct page *page) +{ +return (loff_t )page->index * (loff_t )nfs_readdir_array_maxentries (); +} +static u64 nfs_readdir_page_last_cookie (struct page *page) {struct nfs_cache_array *array;
@@ -448,7 +456,7 @@ static int nfs_readdir_search_for_cookie(struct nfs_cache_array *array,if (array->array[i].cookie == desc->dir_cookie) {struct nfs_inode *nfsi =NFS_I (file_inode (desc->file)); - new_pos = desc->current_index + i; + new_pos =nfs_readdir_page_offset (desc->page) + i;if (desc->attr_gencount != nfsi->attr_gencount || !nfs_readdir_inode_mapping_valid (nfsi)) { desc->duped =0 ;
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 8/27 NFS: Don't re-read the entire page cache to find the next cookie
In reply to [email protected]From: Trond Myklebust <[email protected]>
If the page cache entry that was last read gets invalidated for some reason, then make sure we can re-create it on the next call to readdir. This, combined with the cache page validation, allows us to reuse the cached value of page-index on successive calls to nfs_readdir.
Credit is due to Benjamin Coddington for showing that the concept works, and that it allows for improved cache sharing between processes even in the case where pages are lost due to LRU or active invalidation.
--- fs/nfs/dir.c | 10 +++++++--- include/linux/nfs_fs.h | 1 + 2 files changed, 8 insertions(+), 3 deletions(-)
fs/nfs/dir.c +7 -3
@@ -1120,6 +1120,8 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) desc->dup_cookie = dir_ctx->dup_cookie; desc->duped = dir_ctx->duped; page_index = dir_ctx->page_index; + desc->page_index = page_index; + desc->last_cookie = dir_ctx->last_cookie; desc->attr_gencount = dir_ctx->attr_gencount; desc->eof = dir_ctx->eof;memcpy (desc->verf, dir_ctx->verf,sizeof (desc->verf));
@@ -1168,6 +1170,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)spin_lock (&file->f_lock); dir_ctx->dir_cookie = desc->dir_cookie; dir_ctx->dup_cookie = desc->dup_cookie; + dir_ctx->last_cookie = desc->last_cookie; dir_ctx->duped = desc->duped; dir_ctx->attr_gencount = desc->attr_gencount; dir_ctx->page_index = desc->page_index;
@@ -1209,10 +1212,11 @@ static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence) }if (offset != filp->f_pos) { filp->f_pos = offset; -if (nfs_readdir_use_cookie (filp)) - dir_ctx->dir_cookie = offset; -else +if (!nfs_readdir_use_cookie (filp)) { dir_ctx->dir_cookie =0 ; + dir_ctx->page_index =0 ; + }else + dir_ctx->dir_cookie = offset;if (offset ==0 )memset (dir_ctx->verf,0 ,sizeof (dir_ctx->verf)); dir_ctx->duped =0 ;
include/linux/nfs_fs.h +1 -0
@@ -105,6 +105,7 @@ struct nfs_open_dir_context {__be32 verf[NFS_DIR_VERIFIER_SIZE ];__u64 dir_cookie;__u64 dup_cookie; +__u64 last_cookie;pgoff_t page_index;signed char duped;bool eof;
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 3/27 NFS: Trace lookup revalidation failure
In reply to [email protected]From: Trond Myklebust <[email protected]>
Enable tracing of lookup revalidation failures.
--- fs/nfs/dir.c | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-)
fs/nfs/dir.c +5 -12
@@ -1474,9 +1474,7 @@ nfs_lookup_revalidate_done(struct inode *dir, struct dentry *dentry, {switch (error) {case 1 : -dfprintk (LOOKUPCACHE ,"NFS: %s(%pd2) is valid\n" , - __func__, dentry); -return 1 ; +break ;case 0 :/* * We can't d_drop the root of a disconnected tree:
@@ -1485,13 +1483,10 @@ nfs_lookup_revalidate_done(struct inode *dir, struct dentry *dentry,* inodes on unmount and further oopses. */ if (inode &&IS_ROOT (dentry)) -return 1 ; -dfprintk (LOOKUPCACHE ,"NFS: %s(%pd2) is invalid\n" , - __func__, dentry); -return 0 ; + error =1 ; +break ; } -dfprintk (LOOKUPCACHE ,"NFS: %s(%pd2) lookup returned error %d\n" , - __func__, dentry, error); +trace_nfs_lookup_revalidate_exit (dir, dentry,0 , error);return error; }
@@ -1623,9 +1618,7 @@ nfs_do_lookup_revalidate(struct inode *dir, struct dentry *dentry, goto out_bad;trace_nfs_lookup_revalidate_enter (dir, dentry, flags); - error =nfs_lookup_revalidate_dentry (dir, dentry, inode); -trace_nfs_lookup_revalidate_exit (dir, dentry, flags, error); -return error; +return nfs_lookup_revalidate_dentry (dir, dentry, inode); out_valid:return nfs_lookup_revalidate_done (dir, dentry, inode,1 ); out_bad:
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 10/27 NFS: Adjust the amount of readahead performed by NFS readdir
In reply to [email protected]From: Trond Myklebust <[email protected]>
The current NFS readdir code will always try to maximise the amount of readahead it performs on the assumption that we can cache anything that isn't immediately read by the process.
There are several cases where this assumption breaks down, including when the 'ls -l' heuristic kicks in to try to force use of readdirplus as a batch replacement for lookup/getattr.
This patch therefore tries to tone down the amount of readahead we perform, and adjust it to try to match the amount of data being requested by user space.
--- fs/nfs/dir.c | 53 +++++++++++++++++++++++++++++++++++++++++- include/linux/nfs_fs.h | 1 + 2 files changed, 53 insertions(+), 1 deletion(-)
fs/nfs/dir.c +52 -1
@@ -69,6 +69,8 @@ const struct address_space_operations nfs_dir_aops = { .freepage = nfs_readdir_clear_array, }; +#define NFS_INIT_DTSIZE PAGE_SIZE +static struct nfs_open_dir_context *alloc_nfs_open_dir_context (struct inode *dir) {
@@ -78,6 +80,7 @@ alloc_nfs_open_dir_context(struct inode *dir) ctx =kzalloc (sizeof (*ctx),GFP_KERNEL_ACCOUNT );if (ctx !=NULL ) { ctx->attr_gencount = nfsi->attr_gencount; + ctx->dtsize =NFS_INIT_DTSIZE ;spin_lock (&dir->i_lock);if (list_empty (&nfsi->open_files) && (nfsi->cache_validity &NFS_INO_DATA_INVAL_DEFER ))
@@ -154,6 +157,7 @@ struct nfs_readdir_descriptor {struct page *page;struct dir_context *ctx;pgoff_t page_index; +pgoff_t page_index_max;u64 dir_cookie;u64 last_cookie;u64 dup_cookie;
@@ -166,12 +170,36 @@ struct nfs_readdir_descriptor {unsigned long gencount;unsigned long attr_gencount;unsigned int cache_entry_index; +unsigned int buffer_fills; +unsigned int dtsize;signed char duped;bool plus;bool eob;bool eof; }; +static void nfs_set_dtsize (struct nfs_readdir_descriptor *desc,unsigned int sz) +{ +struct nfs_server *server =NFS_SERVER (file_inode (desc->file)); +unsigned int maxsize = server->dtsize; + +if (sz > maxsize) + sz = maxsize; +if (sz <NFS_MIN_FILE_IO_SIZE ) + sz =NFS_MIN_FILE_IO_SIZE ; + desc->dtsize = sz; +} + +static void nfs_shrink_dtsize (struct nfs_readdir_descriptor *desc) +{ +nfs_set_dtsize (desc, desc->dtsize >>1 ); +} + +static void nfs_grow_dtsize (struct nfs_readdir_descriptor *desc) +{ +nfs_set_dtsize (desc, desc->dtsize <<1 ); +} +static void nfs_readdir_array_init (struct nfs_cache_array *array) {memset (array,0 ,sizeof (struct nfs_cache_array ));
@@ -784,6 +812,7 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc,break ; arrays++; *arrays = page = new; + desc->page_index_max++; }else { new = nfs_readdir_page_get_next(mapping, page->index + 1,
@@ -793,6 +822,7 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc,if (page != *arrays)nfs_readdir_page_unlock_and_put (page); page = new; + desc->page_index_max = new->index; } status =nfs_readdir_add_to_array (entry, page); }while (!status && !entry->eof);
@@ -858,7 +888,7 @@ static int nfs_readdir_xdr_to_array(struct nfs_readdir_descriptor *desc,struct nfs_entry *entry;size_t array_size;struct inode *inode =file_inode (desc->file); -size_t dtsize =NFS_SERVER (inode)->dtsize; +unsigned int dtsize = desc->dtsize;int status = -ENOMEM ; entry =kzalloc (sizeof (*entry),GFP_KERNEL );
@@ -894,6 +924,7 @@ static int nfs_readdir_xdr_to_array(struct nfs_readdir_descriptor *desc, status =nfs_readdir_page_filler (desc, entry, pages, pglen, arrays, narrays); + desc->buffer_fills++; }while (!status &&nfs_readdir_page_needs_filling (page) &&page_mapping (page));
@@ -941,6 +972,10 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)if (!desc->page)return -ENOMEM ;if (nfs_readdir_page_needs_filling (desc->page)) { +/* Grow the dtsize if we had to go back for more pages */ +if (desc->page_index == desc->page_index_max) +nfs_grow_dtsize (desc); + desc->page_index_max = desc->page_index; res =nfs_readdir_xdr_to_array (desc, nfsi->cookieverf, verf, &desc->page,1 );if (res <0 ) {
@@ -1075,6 +1110,7 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc) desc->cache_entry_index =0 ; desc->last_cookie = desc->dir_cookie; desc->duped =0 ; + desc->page_index_max =0 ; status =nfs_readdir_xdr_to_array (desc, desc->verf, verf, arrays, sz);
@@ -1084,10 +1120,22 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc) } desc->page =NULL ; +/* +* Grow the dtsize if we have to go back for more pages, +* or shrink it if we're reading too many. +*/ +if (!desc->eof) { +if (!desc->eob) +nfs_grow_dtsize (desc); +else if (desc->buffer_fills ==1 && + i < (desc->page_index_max >>1 )) +nfs_shrink_dtsize (desc); + }for (i =0 ; i < sz && arrays[i]; i++)nfs_readdir_page_array_free (arrays[i]); out: + desc->page_index_max =-1 ;kfree (arrays);dfprintk (DIRCACHE ,"NFS: %s: returns %d\n" , __func__, status);return status;
@@ -1126,6 +1174,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) desc->file = file; desc->ctx = ctx; desc->plus =nfs_use_readdirplus (inode, ctx); + desc->page_index_max =-1 ;spin_lock (&file->f_lock); desc->dir_cookie = dir_ctx->dir_cookie;
@@ -1136,6 +1185,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) desc->last_cookie = dir_ctx->last_cookie; desc->attr_gencount = dir_ctx->attr_gencount; desc->eof = dir_ctx->eof; +nfs_set_dtsize (desc, dir_ctx->dtsize);memcpy (desc->verf, dir_ctx->verf,sizeof (desc->verf));spin_unlock (&file->f_lock);
@@ -1187,6 +1237,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) dir_ctx->attr_gencount = desc->attr_gencount; dir_ctx->page_index = desc->page_index; dir_ctx->eof = desc->eof; + dir_ctx->dtsize = desc->dtsize;memcpy (dir_ctx->verf, desc->verf,sizeof (dir_ctx->verf));spin_unlock (&file->f_lock); out_free:
include/linux/nfs_fs.h +1 -0
@@ -107,6 +107,7 @@ struct nfs_open_dir_context {__u64 dup_cookie;__u64 last_cookie;pgoff_t page_index; +unsigned int dtsize;signed char duped;bool eof; };
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 12/27 NFS: Simplify nfs_readdir_xdr_to_array()
In reply to [email protected]From: Trond Myklebust <[email protected]>
Recent changes to readdir mean that we can cope with partially filled page cache entries, so we no longer need to rely on looping in nfs_readdir_xdr_to_array().
--- fs/nfs/dir.c | 29 +++++++++++------------------ 1 file changed, 11 insertions(+), 18 deletions(-)
fs/nfs/dir.c +11 -18
@@ -889,6 +889,7 @@ static int nfs_readdir_xdr_to_array(struct nfs_readdir_descriptor *desc,size_t array_size;struct inode *inode =file_inode (desc->file);unsigned int dtsize = desc->dtsize; +unsigned int pglen;int status = -ENOMEM ; entry =kzalloc (sizeof (*entry),GFP_KERNEL );
@@ -906,28 +907,20 @@ static int nfs_readdir_xdr_to_array(struct nfs_readdir_descriptor *desc,if (!pages) goto out; -do { -unsigned int pglen; - status =nfs_readdir_xdr_filler (desc, verf_arg, entry->cookie, - pages, dtsize, - verf_res); -if (status <0 ) -break ; - - pglen = status; -if (pglen ==0 ) { -nfs_readdir_page_set_eof (page); -break ; - } - - verf_arg = verf_res; + status =nfs_readdir_xdr_filler (desc, verf_arg, entry->cookie, pages, + dtsize, verf_res); +if (status <0 ) + goto free_pages; + pglen = status; +if (pglen !=0 ) status =nfs_readdir_page_filler (desc, entry, pages, pglen, arrays, narrays); - desc->buffer_fills++; - }while (!status &&nfs_readdir_page_needs_filling (page) && -page_mapping (page)); +else +nfs_readdir_page_set_eof (page); + desc->buffer_fills++; +free_pages:nfs_readdir_free_pages (pages, array_size); out:nfs_free_fattr (entry->fattr);
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 9/27 NFS: Don't advance the page pointer unless the page is full
In reply to [email protected]From: Trond Myklebust <[email protected]>
When we hit the end of the data in the readdir page, we don't want to start filling a new page, unless this one is full.
--- fs/nfs/dir.c | 32 ++++++++++++++++++++++---------- 1 file changed, 22 insertions(+), 10 deletions(-)
fs/nfs/dir.c +22 -10
@@ -417,6 +417,18 @@ bool nfs_readdir_use_cookie(const struct file *filp)return true; } +static void nfs_readdir_seek_next_array (struct nfs_cache_array *array, +struct nfs_readdir_descriptor *desc) +{ +if (array->page_full) { + desc->last_cookie = array->last_cookie; + desc->current_index += array->size; + desc->cache_entry_index =0 ; + desc->page_index++; + }else + desc->last_cookie = array->array[0 ].cookie; +} +static int nfs_readdir_search_for_pos (struct nfs_cache_array *array,struct nfs_readdir_descriptor *desc) {
@@ -428,6 +440,7 @@ static int nfs_readdir_search_for_pos(struct nfs_cache_array *array,if (diff >= array->size) {if (array->page_is_eof) goto out_eof; +nfs_readdir_seek_next_array (array, desc);return -EAGAIN ; }
@@ -500,7 +513,8 @@ static int nfs_readdir_search_for_cookie(struct nfs_cache_array *array, status = -EBADCOOKIE ;if (desc->dir_cookie == array->last_cookie) desc->eof = true; - } + }else +nfs_readdir_seek_next_array (array, desc); out:return status; }
@@ -517,11 +531,6 @@ static int nfs_readdir_search_array(struct nfs_readdir_descriptor *desc)else status =nfs_readdir_search_for_cookie (array, desc); -if (status == -EAGAIN ) { - desc->last_cookie = array->last_cookie; - desc->current_index += array->size; - desc->page_index++; - }kunmap_atomic (array);return status; }
@@ -998,7 +1007,7 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc, {struct file *file = desc->file;struct nfs_cache_array *array; -unsigned int i =0 ; +unsigned int i; array =kmap (desc->page);for (i = desc->cache_entry_index; i < array->size; i++) {
@@ -1011,10 +1020,13 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc,break ; }memcpy (desc->verf, verf,sizeof (desc->verf)); -if (i < (array->size-1 )) - desc->dir_cookie = array->array[i+1 ].cookie; -else +if (i == array->size -1 ) { desc->dir_cookie = array->last_cookie; +nfs_readdir_seek_next_array (array, desc); + }else { + desc->dir_cookie = array->array[i +1 ].cookie; + desc->last_cookie = array->array[0 ].cookie; + }if (nfs_readdir_use_cookie (file)) desc->ctx->pos = desc->dir_cookie;else
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 11/27 NFS: If the cookie verifier changes, we must invalidate the page cache
In reply to [email protected]From: Trond Myklebust <[email protected]>
Ensure that if the cookie verifier changes when we use the zero-valued cookie, then we invalidate any cached pages.
--- fs/nfs/dir.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
fs/nfs/dir.c +6 -1
@@ -990,9 +990,14 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)/* * Set the cookie verifier if the page cache was empty */ -if (desc->page_index ==0 ) +if (desc->last_cookie ==0 && +memcmp (nfsi->cookieverf, verf,sizeof (nfsi->cookieverf))) {memcpy (nfsi->cookieverf, verf,sizeof (nfsi->cookieverf)); +invalidate_inode_pages2_range (desc->file->f_mapping, + desc->page_index_max +1 , +-1 ); + } } res =nfs_readdir_search_array (desc);if (res ==0 )
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 14/27 NFS: Improve heuristic for readdirplus
In reply to [email protected]From: Trond Myklebust <[email protected]>
The heuristic for readdirplus is designed to try to detect 'ls -l' and similar patterns. It does so by looking for cache hit/miss patterns in both the attribute cache and in the dcache of the files in a given directory, and then sets a flag for the readdirplus code to interpret.
The problem with this approach is that a single attribute or dcache miss can cause the NFS code to force a refresh of the attributes for the entire set of files contained in the directory.
To be able to make a more nuanced decision, let's sample the number of hits and misses in the set of open directory descriptors. That allows us to set thresholds at which we start preferring READDIRPLUS over regular READDIR, or at which we start to force a re-read of the remaining readdir cache using READDIRPLUS.
--- fs/nfs/dir.c | 82 ++++++++++++++++++++++++++---------------- fs/nfs/inode.c | 4 +-- fs/nfs/internal.h | 4 +-- fs/nfs/nfstrace.h | 1 - include/linux/nfs_fs.h | 5 +-- 5 files changed, 58 insertions(+), 38 deletions(-)
fs/nfs/dir.c +51 -31
@@ -87,8 +87,7 @@ alloc_nfs_open_dir_context(struct inode *dir)nfs_set_cache_invalid (dir,NFS_INO_INVALID_DATA |NFS_INO_REVAL_FORCED ); -list_add (&ctx->list, &nfsi->open_files); -clear_bit (NFS_INO_FORCE_READDIR , &nfsi->flags); +list_add_tail_rcu (&ctx->list, &nfsi->open_files);memcpy (ctx->verf, nfsi->cookieverf,sizeof (ctx->verf));spin_unlock (&dir->i_lock);return ctx;
@@ -99,9 +98,9 @@ alloc_nfs_open_dir_context(struct inode *dir)static void put_nfs_open_dir_context (struct inode *dir,struct nfs_open_dir_context *ctx) {spin_lock (&dir->i_lock); -list_del (&ctx->list); +list_del_rcu (&ctx->list);spin_unlock (&dir->i_lock); -kfree (ctx); +kfree_rcu (ctx, rcu_head); }/*
@@ -594,7 +593,6 @@ static int nfs_readdir_xdr_filler(struct nfs_readdir_descriptor *desc,/* We requested READDIRPLUS, but the server doesn't grok it */ if (error == -ENOTSUPP && desc->plus) {NFS_SERVER (inode)->caps &= ~NFS_CAP_READDIRPLUS ; -clear_bit (NFS_INO_ADVISE_RDPLUS , &NFS_I (inode)->flags); desc->plus = arg.plus = false; goto again; }
@@ -644,51 +642,61 @@ int nfs_same_file(struct dentry *dentry, struct nfs_entry *entry)return 1 ; } -static -bool nfs_use_readdirplus (struct inode *dir,struct dir_context *ctx) +#define NFS_READDIR_CACHE_USAGE_THRESHOLD (8UL) + +static bool nfs_use_readdirplus (struct inode *dir,struct dir_context *ctx, +unsigned int cache_hits, +unsigned int cache_misses) {if (!nfs_server_capable (dir,NFS_CAP_READDIRPLUS ))return false; -if (test_and_clear_bit (NFS_INO_ADVISE_RDPLUS , &NFS_I (dir)->flags)) -return true; -if (ctx->pos ==0 ) +if (ctx->pos ==0 || + cache_hits + cache_misses >NFS_READDIR_CACHE_USAGE_THRESHOLD )return true;return false; }/* -* This function is called by the lookup and getattr code to request the +* This function is called by the getattr code to request the * use of readdirplus to accelerate any future lookups in the same * directory. */ -void nfs_advise_use_readdirplus (struct inode *dir) +void nfs_readdir_record_entry_cache_hit (struct inode *dir) {struct nfs_inode *nfsi =NFS_I (dir); +struct nfs_open_dir_context *ctx; -if (nfs_server_capable (dir,NFS_CAP_READDIRPLUS ) && - !list_empty (&nfsi->open_files)) -set_bit (NFS_INO_ADVISE_RDPLUS , &nfsi->flags); +if (nfs_server_capable (dir,NFS_CAP_READDIRPLUS )) { +rcu_read_lock (); + list_for_each_entry_rcu (ctx, &nfsi->open_files, list) +atomic_inc (&ctx->cache_hits); +rcu_read_unlock (); + } }/* * This function is mainly for use by nfs_getattr(). * * If this is an 'ls -l', we want to force use of readdirplus. -* Do this by checking if there is an active file descriptor -* and calling nfs_advise_use_readdirplus, then forcing a -* cache flush. */ -void nfs_force_use_readdirplus (struct inode *dir) +void nfs_readdir_record_entry_cache_miss (struct inode *dir) {struct nfs_inode *nfsi =NFS_I (dir); +struct nfs_open_dir_context *ctx; -if (nfs_server_capable (dir,NFS_CAP_READDIRPLUS ) && - !list_empty (&nfsi->open_files)) { -set_bit (NFS_INO_ADVISE_RDPLUS , &nfsi->flags); -set_bit (NFS_INO_FORCE_READDIR , &nfsi->flags); +if (nfs_server_capable (dir,NFS_CAP_READDIRPLUS )) { +rcu_read_lock (); + list_for_each_entry_rcu (ctx, &nfsi->open_files, list) +atomic_inc (&ctx->cache_misses); +rcu_read_unlock (); } } +static void nfs_lookup_advise_force_readdirplus (struct inode *dir) +{ +nfs_readdir_record_entry_cache_miss (dir); +} +static void nfs_prime_dcache (struct dentry *parent,struct nfs_entry *entry,unsigned long dir_verifier)
@@ -1122,6 +1130,19 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc)return status; } +#define NFS_READDIR_CACHE_MISS_THRESHOLD (16UL) + +static void nfs_readdir_handle_cache_misses (struct inode *inode, +struct nfs_readdir_descriptor *desc, +pgoff_t page_index, +unsigned int cache_misses) +{ +if (desc->ctx->pos ==0 || + cache_misses <=NFS_READDIR_CACHE_MISS_THRESHOLD ) +return ; +invalidate_mapping_pages (inode->i_mapping, page_index +1 ,-1 ); +} +/* The file offset position represents the dirent entry number. A last cookie cache takes care of the common case of reading the whole directory.
@@ -1133,6 +1154,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)struct nfs_inode *nfsi =NFS_I (inode);struct nfs_open_dir_context *dir_ctx = file->private_data;struct nfs_readdir_descriptor *desc; +unsigned int cache_hits, cache_misses;pgoff_t page_index;int res;
@@ -1154,7 +1176,6 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) goto out; desc->file = file; desc->ctx = ctx; - desc->plus =nfs_use_readdirplus (inode, ctx); desc->page_index_max =-1 ;spin_lock (&file->f_lock);
@@ -1168,6 +1189,8 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) desc->eof = dir_ctx->eof;nfs_set_dtsize (desc, dir_ctx->dtsize);memcpy (desc->verf, dir_ctx->verf,sizeof (desc->verf)); + cache_hits =atomic_xchg (&dir_ctx->cache_hits,0 ); + cache_misses =atomic_xchg (&dir_ctx->cache_misses,0 );spin_unlock (&file->f_lock);if (desc->eof) {
@@ -1175,9 +1198,8 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) goto out_free; } -if (test_and_clear_bit (NFS_INO_FORCE_READDIR , &nfsi->flags) && -list_is_singular (&nfsi->open_files)) -invalidate_mapping_pages (inode->i_mapping, page_index +1 ,-1 ); + desc->plus =nfs_use_readdirplus (inode, ctx, cache_hits, cache_misses); +nfs_readdir_handle_cache_misses (inode, desc, page_index, cache_misses); do { res = readdir_search_pagecache(desc);
@@ -1196,7 +1218,6 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)break ; }if (res == -ETOOSMALL && desc->plus) { -clear_bit (NFS_INO_ADVISE_RDPLUS , &nfsi->flags);nfs_zap_caches (inode); desc->page_index =0 ; desc->plus = false;
@@ -1610,7 +1631,7 @@ nfs_lookup_revalidate_dentry(struct inode *dir, struct dentry *dentry,nfs_set_verifier (dentry, dir_verifier);/* set a readdirplus hint that we had a cache miss */ -nfs_force_use_readdirplus (dir); +nfs_lookup_advise_force_readdirplus (dir); ret =1 ; out:nfs_free_fattr (fattr);
@@ -1667,7 +1688,6 @@ nfs_do_lookup_revalidate(struct inode *dir, struct dentry *dentry,nfs_mark_dir_for_revalidate (dir); goto out_bad; } -nfs_advise_use_readdirplus (dir); goto out_valid; }
@@ -1872,7 +1892,7 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in goto out;/* Notify readdir to use READDIRPLUS */ -nfs_force_use_readdirplus (dir); +nfs_lookup_advise_force_readdirplus (dir); no_entry: res =d_splice_alias (inode, dentry);
fs/nfs/inode.c +2 -2
@@ -787,7 +787,7 @@ static void nfs_readdirplus_parent_cache_miss(struct dentry *dentry)if (!nfs_server_capable (d_inode (dentry),NFS_CAP_READDIRPLUS ))return ; parent =dget_parent (dentry); -nfs_force_use_readdirplus (d_inode (parent)); +nfs_readdir_record_entry_cache_miss (d_inode (parent));dput (parent); }
@@ -798,7 +798,7 @@ static void nfs_readdirplus_parent_cache_hit(struct dentry *dentry)if (!nfs_server_capable (d_inode (dentry),NFS_CAP_READDIRPLUS ))return ; parent =dget_parent (dentry); -nfs_advise_use_readdirplus (d_inode (parent)); +nfs_readdir_record_entry_cache_hit (d_inode (parent));dput (parent); }
fs/nfs/internal.h +2 -2
@@ -366,8 +366,8 @@ extern struct nfs_client *nfs_init_client(struct nfs_client *clp, const struct nfs_client_initdata *);/* dir.c */ -extern void nfs_advise_use_readdirplus (struct inode *dir); -extern void nfs_force_use_readdirplus (struct inode *dir); +extern void nfs_readdir_record_entry_cache_hit (struct inode *dir); +extern void nfs_readdir_record_entry_cache_miss (struct inode *dir);extern unsigned long nfs_access_cache_count (struct shrinker *shrink,struct shrink_control *sc); extern unsigned long nfs_access_cache_scan(struct shrinker *shrink,
fs/nfs/nfstrace.h +0 -1
@@ -36,7 +36,6 @@#define nfs_show_nfsi_flags (v) \ __print_flags(v, "|", \ - { BIT(NFS_INO_ADVISE_RDPLUS), "ADVISE_RDPLUS" }, \ { BIT(NFS_INO_STALE), "STALE" }, \ { BIT(NFS_INO_ACL_LRU_SET), "ACL_LRU_SET" }, \ { BIT(NFS_INO_INVALIDATING), "INVALIDATING" }, \
include/linux/nfs_fs.h +3 -2
@@ -101,6 +101,8 @@ struct nfs_open_context {struct nfs_open_dir_context {struct list_head list; +atomic_t cache_hits; +atomic_t cache_misses;unsigned long attr_gencount;__be32 verf[NFS_DIR_VERIFIER_SIZE ];__u64 dir_cookie;
@@ -110,6 +112,7 @@ struct nfs_open_dir_context {unsigned int dtsize;signed char duped;bool eof; +struct rcu_head rcu_head; };/*
@@ -274,13 +277,11 @@ struct nfs4_copy_state {/* * Bit offsets in flags field */ -#define NFS_INO_ADVISE_RDPLUS (0)/* advise readdirplus */ #define NFS_INO_STALE (1)/* possible stale inode */ #define NFS_INO_ACL_LRU_SET (2)/* Inode is on the LRU list */ #define NFS_INO_INVALIDATING (3)/* inode is being invalidated */ #define NFS_INO_PRESERVE_UNLINKED (4)/* preserve file if removed while open */ #define NFS_INO_FSCACHE (5)/* inode can be cached by FS-Cache */ -#define NFS_INO_FORCE_READDIR (7)/* force readdirplus */ #define NFS_INO_LAYOUTCOMMIT (9)/* layoutcommit required */ #define NFS_INO_LAYOUTCOMMITTING (10)/* layoutcommit inflight */ #define NFS_INO_LAYOUTSTATS (11)/* layoutstats inflight */
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 15/27 NFS: Don't ask for readdirplus unless it can help nfs_getattr()
From: Trond Myklebust <[email protected]>
If attribute caching is turned off, then use of readdirplus is not going to help stat() performance.
Readdirplus also doesn't help if a file is being written to, since we will have to flush those writes in order to sync the mtime/ctime.
--- fs/nfs/inode.c | 33 +++++++++++++++++---------------- 1 file changed, 17 insertions(+), 16 deletions(-)
fs/nfs/inode.c +17 -16
@@ -780,24 +780,26 @@ void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr, }EXPORT_SYMBOL_GPL (nfs_setattr_update_inode); -static void nfs_readdirplus_parent_cache_miss (struct dentry *dentry) +/* +* Don't request help from readdirplus if the file is being written to, +* or if attribute caching is turned off +*/ +static bool nfs_getattr_readdirplus_enable (const struct inode *inode) { -struct dentry *parent; +return nfs_server_capable (inode,NFS_CAP_READDIRPLUS ) && + !nfs_have_writebacks (inode) &&NFS_MAXATTRTIMEO (inode) >5 *HZ ; +} -if (!nfs_server_capable (d_inode (dentry),NFS_CAP_READDIRPLUS )) -return ; - parent =dget_parent (dentry); +static void nfs_readdirplus_parent_cache_miss (struct dentry *dentry) +{ +struct dentry *parent =dget_parent (dentry);nfs_readdir_record_entry_cache_miss (d_inode (parent));dput (parent); }static void nfs_readdirplus_parent_cache_hit (struct dentry *dentry) { -struct dentry *parent; - -if (!nfs_server_capable (d_inode (dentry),NFS_CAP_READDIRPLUS )) -return ; - parent =dget_parent (dentry); +struct dentry *parent =dget_parent (dentry);nfs_readdir_record_entry_cache_hit (d_inode (parent));dput (parent); }
@@ -835,6 +837,7 @@ int nfs_getattr(struct user_namespace *mnt_userns, const struct path *path,int err =0 ;bool force_sync = query_flags &AT_STATX_FORCE_SYNC ;bool do_update = false; +bool readdirplus_enabled =nfs_getattr_readdirplus_enable (inode);trace_nfs_getattr_enter (inode);
@@ -843,7 +846,8 @@ int nfs_getattr(struct user_namespace *mnt_userns, const struct path *path,STATX_INO |STATX_SIZE |STATX_BLOCKS ;if ((query_flags &AT_STATX_DONT_SYNC ) && !force_sync) { -nfs_readdirplus_parent_cache_hit (path->dentry); +if (readdirplus_enabled) +nfs_readdirplus_parent_cache_hit (path->dentry); goto out_no_revalidate; }
@@ -893,15 +897,12 @@ int nfs_getattr(struct user_namespace *mnt_userns, const struct path *path, do_update |= cache_validity &NFS_INO_INVALID_BLOCKS ;if (do_update) { -/* Update the attribute cache */ -if (!(server->flags &NFS_MOUNT_NOAC )) +if (readdirplus_enabled)nfs_readdirplus_parent_cache_miss (path->dentry); -else -nfs_readdirplus_parent_cache_hit (path->dentry); err =__nfs_revalidate_inode (server, inode);if (err) goto out; - }else + }else if (readdirplus_enabled)nfs_readdirplus_parent_cache_hit (path->dentry); out_no_revalidate:/* Only return attributes that were revalidated. */
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 16/27 NFSv4: Ask for a full XDR buffer of readdir goodness
From: Trond Myklebust <[email protected]>
Instead of pretending that we know the ratio of directory info vs readdirplus attribute info, just set the 'dircount' field to the same value as the 'maxcount' field.
--- fs/nfs/nfs3xdr.c | 7 ++++--- fs/nfs/nfs4xdr.c | 6 +++--- 2 files changed, 7 insertions(+), 6 deletions(-)
fs/nfs/nfs3xdr.c +4 -3
@@ -1261,6 +1261,8 @@ static void nfs3_xdr_enc_readdir3args(struct rpc_rqst *req,static void encode_readdirplus3args (struct xdr_stream *xdr,const struct nfs3_readdirargs *args) { +uint32_t dircount = args->count; +uint32_t maxcount = args->count;__be32 *p;encode_nfs_fh3 (xdr, args->fh);
@@ -1273,9 +1275,8 @@ static void encode_readdirplus3args(struct xdr_stream *xdr,* readdirplus: need dircount + buffer size. * We just make sure we make dircount big enough */ - *p++ =cpu_to_be32 (args->count >>3 ); - - *p =cpu_to_be32 (args->count); + *p++ =cpu_to_be32 (dircount); + *p =cpu_to_be32 (maxcount); } static void nfs3_xdr_enc_readdirplus3args(struct rpc_rqst *req,
fs/nfs/nfs4xdr.c +3 -3
@@ -1605,7 +1605,8 @@ static void encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_argFATTR4_WORD0_RDATTR_ERROR , FATTR4_WORD1_MOUNTED_ON_FILEID, }; -uint32_t dircount = readdir->count >>1 ; +uint32_t dircount = readdir->count; +uint32_t maxcount = readdir->count;__be32 *p, verf[2 ];uint32_t attrlen =0 ;unsigned int i;
@@ -1618,7 +1619,6 @@ static void encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_argFATTR4_WORD1_SPACE_USED |FATTR4_WORD1_TIME_ACCESS |FATTR4_WORD1_TIME_METADATA |FATTR4_WORD1_TIME_MODIFY ; attrs[2 ] |=FATTR4_WORD2_SECURITY_LABEL ; - dircount >>=1 ; }/* Use mounted_on_fileid only if the server supports it */ if (!(readdir->bitmask[1 ] &FATTR4_WORD1_MOUNTED_ON_FILEID ))
@@ -1634,7 +1634,7 @@ static void encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_argencode_nfs4_verifier (xdr, &readdir->verifier); p =reserve_space (xdr,12 + (attrlen <<2 )); *p++ =cpu_to_be32 (dircount); - *p++ =cpu_to_be32 (readdir->count); + *p++ =cpu_to_be32 (maxcount); *p++ =cpu_to_be32 (attrlen);for (i =0 ; i < attrlen; i++) *p++ =cpu_to_be32 (attrs[i]);
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 17/27 NFS: Readdirplus can't help lookup for case insensitive filesystems
From: Trond Myklebust <[email protected]>
If the filesystem is case insensitive, then readdirplus can't help with cache misses, since it won't return case folded variants of the filename.
--- fs/nfs/dir.c | 2 ++ 1 file changed, 2 insertions(+)
fs/nfs/dir.c +2 -0
@@ -694,6 +694,8 @@ void nfs_readdir_record_entry_cache_miss(struct inode *dir)static void nfs_lookup_advise_force_readdirplus (struct inode *dir) { +if (nfs_server_capable (dir,NFS_CAP_CASE_INSENSITIVE )) +return ;nfs_readdir_record_entry_cache_miss (dir); }
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 18/27 NFS: Don't request readdirplus when revalidation was forced
From: Trond Myklebust <[email protected]>
If the revalidation was forced, due to the presence of a LOOKUP_EXCL or a LOOKUP_REVAL flag, then readdirplus won't help. It also can't help when we're doing a path component lookup.
--- fs/nfs/dir.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-)
fs/nfs/dir.c +16 -10
@@ -692,10 +692,13 @@ void nfs_readdir_record_entry_cache_miss(struct inode *dir) } } -static void nfs_lookup_advise_force_readdirplus (struct inode *dir) +static void nfs_lookup_advise_force_readdirplus (struct inode *dir, +unsigned int flags) {if (nfs_server_capable (dir,NFS_CAP_CASE_INSENSITIVE ))return ; +if (flags & (LOOKUP_EXCL |LOOKUP_PARENT |LOOKUP_REVAL )) +return ;nfs_readdir_record_entry_cache_miss (dir); }
@@ -1594,15 +1597,17 @@ nfs_lookup_revalidate_delegated(struct inode *dir, struct dentry *dentry,return nfs_lookup_revalidate_done (dir, dentry, inode,1 ); } -static int -nfs_lookup_revalidate_dentry (struct inode *dir,struct dentry *dentry, -struct inode *inode) +static int nfs_lookup_revalidate_dentry (struct inode *dir, +struct dentry *dentry, +struct inode *inode,unsigned int flags) {struct nfs_fh *fhandle;struct nfs_fattr *fattr;unsigned long dir_verifier;int ret; +trace_nfs_lookup_revalidate_enter (dir, dentry, flags); + ret = -ENOMEM ; fhandle =nfs_alloc_fhandle (); fattr =nfs_alloc_fattr_with_label (NFS_SERVER (inode));
@@ -1623,6 +1628,10 @@ nfs_lookup_revalidate_dentry(struct inode *dir, struct dentry *dentry, } goto out; } + +/* Request help from readdirplus */ +nfs_lookup_advise_force_readdirplus (dir, flags); + ret =0 ;if (nfs_compare_fh (NFS_FH (inode), fhandle)) goto out;
@@ -1632,8 +1641,6 @@ nfs_lookup_revalidate_dentry(struct inode *dir, struct dentry *dentry,nfs_setsecurity (inode, fattr);nfs_set_verifier (dentry, dir_verifier); -/* set a readdirplus hint that we had a cache miss */ -nfs_lookup_advise_force_readdirplus (dir); ret =1 ; out:nfs_free_fattr (fattr);
@@ -1699,8 +1706,7 @@ nfs_do_lookup_revalidate(struct inode *dir, struct dentry *dentry,if (NFS_STALE (inode)) goto out_bad; -trace_nfs_lookup_revalidate_enter (dir, dentry, flags); -return nfs_lookup_revalidate_dentry (dir, dentry, inode); +return nfs_lookup_revalidate_dentry (dir, dentry, inode, flags); out_valid:return nfs_lookup_revalidate_done (dir, dentry, inode,1 ); out_bad:
@@ -1894,7 +1900,7 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in goto out;/* Notify readdir to use READDIRPLUS */ -nfs_lookup_advise_force_readdirplus (dir); +nfs_lookup_advise_force_readdirplus (dir, flags); no_entry: res =d_splice_alias (inode, dentry);
@@ -2157,7 +2163,7 @@ nfs4_do_lookup_revalidate(struct inode *dir, struct dentry *dentry, reval_dentry:if (flags &LOOKUP_RCU )return -ECHILD ; -return nfs_lookup_revalidate_dentry (dir, dentry, inode); +return nfs_lookup_revalidate_dentry (dir, dentry, inode, flags); full_reval:return nfs_do_lookup_revalidate (dir, dentry, flags);
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 13/27 NFS: Reduce use of uncached readdir
In reply to [email protected]From: Trond Myklebust <[email protected]>
When reading a very large directory, we want to try to keep the page cache up to date if doing so is inexpensive. With the change to allow readdir to continue reading even when the cache is incomplete, we no longer need to fall back to uncached readdir in order to scale to large directories.
--- fs/nfs/dir.c | 23 +++-------------------- 1 file changed, 3 insertions(+), 20 deletions(-)
fs/nfs/dir.c +3 -20
@@ -999,28 +999,11 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)return res; } -static bool nfs_readdir_dont_search_cache (struct nfs_readdir_descriptor *desc) -{ -struct address_space *mapping = desc->file->f_mapping; -struct inode *dir =file_inode (desc->file); -unsigned int dtsize =NFS_SERVER (dir)->dtsize; -loff_t size =i_size_read (dir); - -/* -* Default to uncached readdir if the page cache is empty, and -* we're looking for a non-zero cookie in a large directory. -*/ -return desc->dir_cookie !=0 && mapping->nrpages ==0 && size > dtsize; -} -/* Search for desc->dir_cookie from the beginning of the page cache */ static int readdir_search_pagecache (struct nfs_readdir_descriptor *desc) {int res; -if (nfs_readdir_dont_search_cache (desc)) -return -EBADCOOKIE ; - do { if (desc->page_index == 0) { desc->current_index = 0;
@@ -1273,10 +1256,10 @@ static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence) }if (offset != filp->f_pos) { filp->f_pos = offset; -if (!nfs_readdir_use_cookie (filp)) { + dir_ctx->page_index =0 ; +if (!nfs_readdir_use_cookie (filp)) dir_ctx->dir_cookie =0 ; - dir_ctx->page_index =0 ; - }else +else dir_ctx->dir_cookie = offset;if (offset ==0 )memset (dir_ctx->verf,0 ,sizeof (dir_ctx->verf));
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 19/27 NFS: Add basic readdir tracing
In reply to [email protected]From: Trond Myklebust <[email protected]>
Add tracing to track how often the client goes to the server for updated readdir information.
--- fs/nfs/dir.c | 13 ++++++++- fs/nfs/nfstrace.h | 68 +++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 80 insertions(+), 1 deletion(-)
fs/nfs/dir.c +12 -1
@@ -982,10 +982,14 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)if (desc->page_index == desc->page_index_max)nfs_grow_dtsize (desc); desc->page_index_max = desc->page_index; +trace_nfs_readdir_cache_fill (desc->file, nfsi->cookieverf, + desc->last_cookie, + desc->page->index, desc->dtsize); res =nfs_readdir_xdr_to_array (desc, nfsi->cookieverf, verf, &desc->page,1 );if (res <0 ) {nfs_readdir_page_unlock_and_put_cached (desc); +trace_nfs_readdir_cache_fill_done (inode, res);if (res == -EBADCOOKIE || res == -ENOTSYNC ) {invalidate_inode_pages2 (desc->file->f_mapping); desc->page_index =0 ;
@@ -1106,7 +1110,14 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc) desc->duped =0 ; desc->page_index_max =0 ; +trace_nfs_readdir_uncached (desc->file, desc->verf, desc->last_cookie, +-1 , desc->dtsize); + status =nfs_readdir_xdr_to_array (desc, desc->verf, verf, arrays, sz); +if (status <0 ) { +trace_nfs_readdir_uncached_done (file_inode (desc->file), status); + goto out_free; + }for (i =0 ; !desc->eob && i < sz && arrays[i]; i++) { desc->page = arrays[i];
@@ -1125,7 +1136,7 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc) i < (desc->page_index_max >> 1))nfs_shrink_dtsize (desc); } - +out_free:for (i =0 ; i < sz && arrays[i]; i++)nfs_readdir_page_array_free (arrays[i]); out:
fs/nfs/nfstrace.h +68 -0
80 lines not on this page · raw message
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 21/27 NFS: Trace effects of the readdirplus heuristic
In reply to [email protected]From: Trond Myklebust <[email protected]>
Enable tracking of when the readdirplus heuristic causes a page cache invalidation.
--- fs/nfs/dir.c | 11 ++++++++++- fs/nfs/nfstrace.h | 50 +++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 1 deletion(-)
fs/nfs/dir.c +10 -1
@@ -998,6 +998,8 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)if (res == -EBADCOOKIE || res == -ENOTSYNC ) {invalidate_inode_pages2 (desc->file->f_mapping); desc->page_index =0 ; +trace_nfs_readdir_invalidate_cache_range ( + inode,0 ,MAX_LFS_FILESIZE );return -EAGAIN ; }return res;
@@ -1012,6 +1014,9 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)invalidate_inode_pages2_range (desc->file->f_mapping, desc->page_index_max +1 ,-1 ); +trace_nfs_readdir_invalidate_cache_range ( + inode, desc->page_index_max +1 , +MAX_LFS_FILESIZE ); } } res =nfs_readdir_search_array (desc);
@@ -1161,7 +1166,11 @@ static void nfs_readdir_handle_cache_misses(struct inode *inode,if (desc->ctx->pos ==0 || cache_misses <=NFS_READDIR_CACHE_MISS_THRESHOLD )return ; -invalidate_mapping_pages (inode->i_mapping, page_index +1 ,-1 ); +if (invalidate_mapping_pages (inode->i_mapping, page_index +1 ,-1 ) ==0 ) +return ; +trace_nfs_readdir_invalidate_cache_range ( + inode, (loff_t)(page_index +1 ) <<PAGE_SHIFT , +MAX_LFS_FILESIZE ); }/* The file offset position represents the dirent entry number. A
fs/nfs/nfstrace.h +50 -0
56 lines not on this page · raw message
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 22/27 NFS: Clean up page array initialisation/free
From: Trond Myklebust <[email protected]>
--- fs/nfs/dir.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-)
fs/nfs/dir.c +6 -10
@@ -199,20 +199,17 @@ static void nfs_grow_dtsize(struct nfs_readdir_descriptor *desc)nfs_set_dtsize (desc, desc->dtsize <<1 ); } -static void nfs_readdir_array_init (struct nfs_cache_array *array) -{ -memset (array,0 ,sizeof (struct nfs_cache_array )); -} -static void nfs_readdir_page_init_array (struct page *page,u64 last_cookie,u64 change_attr) {struct nfs_cache_array *array; array =kmap_atomic (page); -nfs_readdir_array_init (array); array->change_attr = change_attr; array->last_cookie = last_cookie; + array->size =0 ; + array->page_full =0 ; + array->page_is_eof =0 ; array->cookies_are_ordered =1 ;kunmap_atomic (array); }
@@ -220,16 +217,15 @@ static void nfs_readdir_page_init_array(struct page *page, u64 last_cookie,/* * we are freeing strings created by nfs_add_to_readdir_array() */ -static -void nfs_readdir_clear_array (struct page *page) +static void nfs_readdir_clear_array (struct page *page) {struct nfs_cache_array *array; -int i; +unsigned int i; array =kmap_atomic (page);for (i =0 ; i < array->size; i++)kfree (array->array[i].name); -nfs_readdir_array_init (array); + array->size =0 ;kunmap_atomic (array); }
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 20/27 NFS: Trace effects of readdirplus on the dcache
In reply to [email protected]From: Trond Myklebust <[email protected]>
Trace the effects of readdirplus on attribute and dentry revalidation.
--- fs/nfs/dir.c | 5 +++++ fs/nfs/nfstrace.h | 3 +++ 2 files changed, 8 insertions(+)
fs/nfs/dir.c +5 -0
@@ -752,8 +752,12 @@ void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry, status =nfs_refresh_inode (d_inode (dentry), entry->fattr);if (!status)nfs_setsecurity (d_inode (dentry), entry->fattr); +trace_nfs_readdir_lookup_revalidate (d_inode (parent), + dentry,0 , status); goto out; }else { +trace_nfs_readdir_lookup_revalidate_failed ( +d_inode (parent), dentry,0 );d_invalidate (dentry);dput (dentry); dentry =NULL ;
@@ -775,6 +779,7 @@ void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry, dentry = alias; }nfs_set_verifier (dentry, dir_verifier); +trace_nfs_readdir_lookup (d_inode (parent), dentry,0 ); out:dput (dentry); }
fs/nfs/nfstrace.h +3 -0
@@ -432,6 +432,9 @@ DEFINE_NFS_LOOKUP_EVENT(nfs_lookup_enter);DEFINE_NFS_LOOKUP_EVENT_DONE (nfs_lookup_exit);DEFINE_NFS_LOOKUP_EVENT (nfs_lookup_revalidate_enter);DEFINE_NFS_LOOKUP_EVENT_DONE (nfs_lookup_revalidate_exit); +DEFINE_NFS_LOOKUP_EVENT (nfs_readdir_lookup); +DEFINE_NFS_LOOKUP_EVENT (nfs_readdir_lookup_revalidate_failed); +DEFINE_NFS_LOOKUP_EVENT_DONE (nfs_readdir_lookup_revalidate); TRACE_EVENT(nfs_atomic_open_enter, TP_PROTO(
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 23/27 NFS: Convert readdir page cache to use a cookie based index
In reply to [email protected]From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of the first entry, since that is what we use to match the page anyway.
This allows us to avoid re-reading the entire cache on a seekdir() type of operation. The latter is very common when re-exporting NFS, and is a major performance drain.
The change does affect our duplicate cookie detection, since we can no longer rely on the page index as a linear offset for detecting whether we looped backwards. However since we no longer do a linear search through all the pages on each call to nfs_readdir(), this is less of a concern than it was previously.
The other downside is that invalidate_mapping_pages() no longer can use the page index to avoid clearing pages that have been read. A subsequent patch will restore the functionality this provides to the 'ls -l' heuristic.
--- fs/nfs/Kconfig | 4 ++ fs/nfs/dir.c | 149 ++++++++++++++++++----------------------- include/linux/nfs_fs.h | 2 - 3 files changed, 69 insertions(+), 86 deletions(-)
fs/nfs/Kconfig +4 -0
@@ -4,6 +4,10 @@ config NFS_FS depends on INET && FILE_LOCKING && MULTIUSER select LOCKD select SUNRPC + select CRYPTO + select CRYPTO_HASH + select XXHASH + select CRYPTO_XXHASH select NFS_ACL_SUPPORT if NFS_V3_ACL help Choose Y here if you want to access files residing on other
fs/nfs/dir.c +65 -84
360 lines not on this page · raw message
include/linux/nfs_fs.h +0 -2
@@ -106,11 +106,9 @@ struct nfs_open_dir_context {unsigned long attr_gencount;__be32 verf[NFS_DIR_VERIFIER_SIZE ];__u64 dir_cookie; -__u64 dup_cookie;__u64 last_cookie;pgoff_t page_index;unsigned int dtsize; -signed char duped;bool eof;struct rcu_head rcu_head; };
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 25/27 NFS: Remove unnecessary cache invalidations for directories
In reply to [email protected]From: Trond Myklebust <[email protected]>
Now that the directory page cache entries police themselves, don't bother with marking the page cache for invalidation.
--- fs/nfs/dir.c | 5 ----- fs/nfs/inode.c | 9 +++------ fs/nfs/nfs4proc.c | 2 -- include/linux/nfs_fs.h | 2 -- 4 files changed, 3 insertions(+), 15 deletions(-)
fs/nfs/dir.c +0 -5
@@ -83,11 +83,6 @@ alloc_nfs_open_dir_context(struct inode *dir) ctx->attr_gencount = nfsi->attr_gencount; ctx->dtsize =NFS_INIT_DTSIZE ;spin_lock (&dir->i_lock); -if (list_empty (&nfsi->open_files) && - (nfsi->cache_validity &NFS_INO_DATA_INVAL_DEFER )) -nfs_set_cache_invalid (dir, -NFS_INO_INVALID_DATA | -NFS_INO_REVAL_FORCED );list_add_tail_rcu (&ctx->list, &nfsi->open_files);memcpy (ctx->verf, nfsi->cookieverf,sizeof (ctx->verf));spin_unlock (&dir->i_lock);
fs/nfs/inode.c +3 -6
@@ -210,6 +210,8 @@ void nfs_set_cache_invalid(struct inode *inode, unsigned long flags)if (flags &NFS_INO_INVALID_DATA )nfs_fscache_invalidate (inode,0 ); flags &= ~NFS_INO_REVAL_FORCED ; +if (S_ISDIR (inode->i_mode)) + flags &= ~(NFS_INO_INVALID_DATA |NFS_INO_DATA_INVAL_DEFER ); nfsi->cache_validity |= flags;
@@ -1429,10 +1431,7 @@ static void nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr) && (fattr->valid & NFS_ATTR_FATTR_CHANGE) && inode_eq_iversion_raw(inode, fattr->pre_change_attr)) {inode_set_iversion_raw (inode, fattr->change_attr); -if (S_ISDIR (inode->i_mode)) -nfs_set_cache_invalid (inode,NFS_INO_INVALID_DATA ); -else if (nfs_server_capable (inode,NFS_CAP_XATTR )) -nfs_set_cache_invalid (inode,NFS_INO_INVALID_XATTR ); +nfs_set_cache_invalid (inode,NFS_INO_INVALID_XATTR ); }/* If we have atomic WCC data, we may update some attributes */ ts = inode->i_ctime;
@@ -1851,8 +1850,6 @@ EXPORT_SYMBOL_GPL(nfs_refresh_inode);static int nfs_post_op_update_inode_locked (struct inode *inode,struct nfs_fattr *fattr,unsigned int invalid) { -if (S_ISDIR (inode->i_mode)) - invalid |=NFS_INO_INVALID_DATA ;nfs_set_cache_invalid (inode, invalid);if ((fattr->valid &NFS_ATTR_FATTR ) ==0 )return 0 ;
fs/nfs/nfs4proc.c +0 -2
@@ -1206,8 +1206,6 @@ nfs4_update_changeattr_locked(struct inode *inode,u64 change_attr =inode_peek_iversion_raw (inode); cache_validity |=NFS_INO_INVALID_CTIME |NFS_INO_INVALID_MTIME ; -if (S_ISDIR (inode->i_mode)) - cache_validity |=NFS_INO_INVALID_DATA ;switch (NFS_SERVER (inode)->change_attr_type) {case NFS4_CHANGE_TYPE_IS_UNDEFINED :
include/linux/nfs_fs.h +0 -2
@@ -360,8 +360,6 @@ static inline void nfs_mark_for_revalidate(struct inode *inode) nfsi->cache_validity |=NFS_INO_INVALID_ACCESS |NFS_INO_INVALID_ACL |NFS_INO_INVALID_CHANGE |NFS_INO_INVALID_CTIME |NFS_INO_INVALID_SIZE ; -if (S_ISDIR (inode->i_mode)) - nfsi->cache_validity |=NFS_INO_INVALID_DATA ;spin_unlock (&inode->i_lock); }
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 24/27 NFS: Fix up forced readdirplus
In reply to [email protected]From: Trond Myklebust <[email protected]>
Avoid clearing the entire readdir page cache if we're just doing forced readdirplus for the 'ls -l' heuristic.
--- fs/nfs/dir.c | 56 +++++++++++++++++++++++++++++------------- fs/nfs/nfstrace.h | 1 + include/linux/nfs_fs.h | 1 + 3 files changed, 41 insertions(+), 17 deletions(-)
fs/nfs/dir.c +39 -17
@@ -170,6 +170,7 @@ struct nfs_readdir_descriptor {unsigned int cache_entry_index;unsigned int buffer_fills;unsigned int dtsize; +bool clear_cache;bool plus;bool eob;bool eof;
@@ -227,6 +228,13 @@ static void nfs_readdir_clear_array(struct page *page)kunmap_atomic (array); } +static void nfs_readdir_page_reinit_array (struct page *page,u64 last_cookie, +u64 change_attr) +{ +nfs_readdir_clear_array (page); +nfs_readdir_page_init_array (page, last_cookie, change_attr); +} +static struct page *nfs_readdir_page_array_alloc (u64 last_cookie,gfp_t gfp_flags) {
@@ -428,12 +436,11 @@ static struct page *nfs_readdir_page_get_next(struct address_space *mapping,struct page *page; page =nfs_readdir_page_get_locked (mapping, cookie, change_attr); -if (page) { -if (nfs_readdir_page_last_cookie (page) == cookie) -return page; -nfs_readdir_page_unlock_and_put (page); - } -return NULL ; +if (!page) +return NULL ; +if (nfs_readdir_page_last_cookie (page) != cookie) +nfs_readdir_page_reinit_array (page, cookie, change_attr); +return page; } static inline
@@ -958,9 +965,15 @@ nfs_readdir_page_get_cached(struct nfs_readdir_descriptor *desc) {struct address_space *mapping = desc->file->f_mapping;u64 change_attr =inode_peek_iversion_raw (mapping->host); +u64 cookie = desc->last_cookie; +struct page *page; -return nfs_readdir_page_get_locked (mapping, desc->last_cookie, - change_attr); + page =nfs_readdir_page_get_locked (mapping, cookie, change_attr); +if (!page) +return NULL ; +if (desc->clear_cache && !nfs_readdir_page_needs_filling (page)) +nfs_readdir_page_reinit_array (page, cookie, change_attr); +return page; }/*
@@ -1011,6 +1024,7 @@ static int find_and_lock_cache_page(struct nfs_readdir_descriptor *desc)trace_nfs_readdir_invalidate_cache_range ( inode,1 ,MAX_LFS_FILESIZE ); } + desc->clear_cache = false; } res =nfs_readdir_search_array (desc);if (res ==0 )
@@ -1145,16 +1159,17 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc)#define NFS_READDIR_CACHE_MISS_THRESHOLD (16UL) -static void nfs_readdir_handle_cache_misses (struct inode *inode, +static bool nfs_readdir_handle_cache_misses (struct inode *inode,struct nfs_readdir_descriptor *desc, -unsigned int cache_misses) +unsigned int cache_misses, +bool force_clear) { -if (desc->ctx->pos ==0 || - cache_misses <=NFS_READDIR_CACHE_MISS_THRESHOLD ) -return ; -if (invalidate_mapping_pages (inode->i_mapping,0 ,-1 ) ==0 ) -return ; -trace_nfs_readdir_invalidate_cache_range (inode,0 ,MAX_LFS_FILESIZE ); +if (desc->ctx->pos ==0 || !desc->plus) +return false; +if (cache_misses <=NFS_READDIR_CACHE_MISS_THRESHOLD && !force_clear) +return false; +trace_nfs_readdir_force_readdirplus (inode); +return true; }/* The file offset position represents the dirent entry number. A
@@ -1169,6 +1184,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)struct nfs_open_dir_context *dir_ctx = file->private_data;struct nfs_readdir_descriptor *desc;unsigned int cache_hits, cache_misses; +bool force_clear;int res; dfprintk(FILE, "NFS: readdir(%pD2) starting at cookie %llu\n",
@@ -1201,6 +1217,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)memcpy (desc->verf, dir_ctx->verf,sizeof (desc->verf)); cache_hits =atomic_xchg (&dir_ctx->cache_hits,0 ); cache_misses =atomic_xchg (&dir_ctx->cache_misses,0 ); + force_clear = dir_ctx->force_clear;spin_unlock (&file->f_lock);if (desc->eof) {
@@ -1209,7 +1226,9 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) } desc->plus =nfs_use_readdirplus (inode, ctx, cache_hits, cache_misses); -nfs_readdir_handle_cache_misses (inode, desc, cache_misses); + force_clear =nfs_readdir_handle_cache_misses (inode, desc, cache_misses, + force_clear); + desc->clear_cache = force_clear; do { res = readdir_search_pagecache(desc);
@@ -1238,6 +1257,8 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)nfs_do_filldir (desc, nfsi->cookieverf);nfs_readdir_page_unlock_and_put_cached (desc); +if (desc->page_index == desc->page_index_max) + desc->clear_cache = force_clear; }while (!desc->eob && !desc->eof);spin_lock (&file->f_lock);
@@ -1245,6 +1266,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx) dir_ctx->last_cookie = desc->last_cookie; dir_ctx->attr_gencount = desc->attr_gencount; dir_ctx->page_index = desc->page_index; + dir_ctx->force_clear = force_clear; dir_ctx->eof = desc->eof; dir_ctx->dtsize = desc->dtsize;memcpy (dir_ctx->verf, desc->verf,sizeof (dir_ctx->verf));
fs/nfs/nfstrace.h +1 -0
@@ -160,6 +160,7 @@ DEFINE_NFS_INODE_EVENT(nfs_fsync_enter);DEFINE_NFS_INODE_EVENT_DONE (nfs_fsync_exit);DEFINE_NFS_INODE_EVENT (nfs_access_enter);DEFINE_NFS_INODE_EVENT_DONE (nfs_set_cache_invalid); +DEFINE_NFS_INODE_EVENT (nfs_readdir_force_readdirplus);DEFINE_NFS_INODE_EVENT_DONE (nfs_readdir_cache_fill_done);DEFINE_NFS_INODE_EVENT_DONE (nfs_readdir_uncached_done);
include/linux/nfs_fs.h +1 -0
@@ -109,6 +109,7 @@ struct nfs_open_dir_context {__u64 last_cookie;pgoff_t page_index;unsigned int dtsize; +bool force_clear;bool eof;struct rcu_head rcu_head; };
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 27/27 NFS: Cache all entries in the readdirplus reply
In reply to [email protected]From: Trond Myklebust <[email protected]>
Even if we're not able to cache all the entries in the readdir buffer, let's ensure that we do prime the dcache.
--- fs/nfs/dir.c | 40 ++++++++++++++++++++++++++-------------- 1 file changed, 26 insertions(+), 14 deletions(-)
fs/nfs/dir.c +26 -14
@@ -782,6 +782,21 @@ void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry,dput (dentry); } +static int nfs_readdir_entry_decode (struct nfs_readdir_descriptor *desc, +struct nfs_entry *entry, +struct xdr_stream *stream) +{ +int ret; + +if (entry->fattr->label) + entry->fattr->label->len =NFS4_MAXLABELLEN ; + ret =xdr_decode (desc, entry, stream); +if (ret || !desc->plus) +return ret; +nfs_prime_dcache (file_dentry (desc->file), entry, desc->dir_verifier); +return 0 ; +} +/* Perform conversion from xdr to cache array */ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc, struct nfs_entry *entry,
@@ -804,17 +819,10 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc,xdr_set_scratch_page (&stream, scratch); do { - if (entry->fattr->label) - entry->fattr->label->len = NFS4_MAXLABELLEN; - - status = xdr_decode(desc, entry, &stream); + status = nfs_readdir_entry_decode(desc, entry, &stream); if (status != 0) break; - if (desc->plus) - nfs_prime_dcache(file_dentry(desc->file), entry, - desc->dir_verifier); - status = nfs_readdir_page_array_append(page, entry, &cookie); if (status != -ENOSPC) continue;
@@ -842,15 +850,19 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc,switch (status) {case -EBADCOOKIE : -if (entry->eof) { -nfs_readdir_page_set_eof (page); - status =0 ; - } -break ; -case -ENOSPC : +if (!entry->eof) +break ; +nfs_readdir_page_set_eof (page); + fallthrough;case -EAGAIN : status =0 ;break ; +case -ENOSPC : + status =0 ; +if (!desc->plus) +break ; +while (!nfs_readdir_entry_decode (desc, entry, &stream)) + ; }if (page != *arrays)
Signature · 1 line
2.35.1
[email protected]
maintainerPATCH v9 26/27 NFS: Optimise away the previous cookie field
In reply to [email protected]From: Trond Myklebust <[email protected]>
Replace the 'previous cookie' field in struct nfs_entry with the array->last_cookie.
--- fs/nfs/dir.c | 26 ++++++++++++++------------ fs/nfs/nfs2xdr.c | 1 - fs/nfs/nfs3xdr.c | 1 - fs/nfs/nfs4xdr.c | 1 - include/linux/nfs_xdr.h | 3 +-- 5 files changed, 15 insertions(+), 17 deletions(-)
fs/nfs/dir.c +14 -12
@@ -296,19 +296,20 @@ static int nfs_readdir_array_can_expand(struct nfs_cache_array *array)return 0 ; } -static -int nfs_readdir_add_to_array (struct nfs_entry *entry,struct page *page) +static int nfs_readdir_page_array_append (struct page *page, +const struct nfs_entry *entry, +u64 *cookie) {struct nfs_cache_array *array;struct nfs_cache_array_entry *cache_entry;const char *name; -int ret; +int ret = -ENOMEM ; name =nfs_readdir_copy_name (entry->name, entry->len); -if (!name) -return -ENOMEM ; array =kmap_atomic (page); +if (!name) + goto out; ret =nfs_readdir_array_can_expand (array);if (ret) {kfree (name);
@@ -316,7 +317,7 @@ int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page) } cache_entry = &array->array[array->size]; - cache_entry->cookie = entry->prev_cookie; + cache_entry->cookie = array->last_cookie; cache_entry->ino = entry->ino; cache_entry->d_type = entry->d_type; cache_entry->name_len = entry->len;
@@ -328,6 +329,7 @@ int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page)if (entry->eof !=0 )nfs_readdir_array_set_eof (array); out: + *cookie = array->last_cookie;kunmap_atomic (array);return ret; }
@@ -791,6 +793,7 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc,struct xdr_stream stream;struct xdr_buf buf;struct page *scratch, *new, *page = *arrays; +u64 cookie;int status; scratch =alloc_page (GFP_KERNEL );
@@ -812,22 +815,21 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc,nfs_prime_dcache (file_dentry (desc->file), entry, desc->dir_verifier); - status =nfs_readdir_add_to_array (entry, page); + status =nfs_readdir_page_array_append (page, entry, &cookie);if (status != -ENOSPC )continue ;if (page->mapping != mapping) {if (!--narrays)break ; - new =nfs_readdir_page_array_alloc (entry->prev_cookie, -GFP_KERNEL ); + new =nfs_readdir_page_array_alloc (cookie,GFP_KERNEL );if (!new)break ; arrays++; *arrays = page = new; }else { - new =nfs_readdir_page_get_next ( - mapping, entry->prev_cookie, change_attr); + new =nfs_readdir_page_get_next (mapping, cookie, + change_attr);if (!new)break ;if (page != *arrays)
@@ -835,7 +837,7 @@ static int nfs_readdir_page_filler(struct nfs_readdir_descriptor *desc, page = new; } desc->page_index_max++; - status =nfs_readdir_add_to_array (entry, page); + status =nfs_readdir_page_array_append (page, entry, &cookie); }while (!status && !entry->eof);switch (status) {
fs/nfs/nfs2xdr.c +0 -1
@@ -955,7 +955,6 @@ int nfs2_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry,* The type (size and byte order) of nfscookie isn't defined in * RFC 1094. This implementation assumes that it's an XDR uint32. */ - entry->prev_cookie = entry->cookie; p =xdr_inline_decode (xdr,4 );if (unlikely (!p))return -EAGAIN ;
fs/nfs/nfs3xdr.c +0 -1
@@ -2024,7 +2024,6 @@ int nfs3_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry,zero_nfs_fh3 (entry->fh); } - entry->prev_cookie = entry->cookie; entry->cookie = new_cookie;return 0 ;
fs/nfs/nfs4xdr.c +0 -1
@@ -7508,7 +7508,6 @@ int nfs4_decode_dirent(struct xdr_stream *xdr, struct nfs_entry *entry,if (entry->fattr->valid &NFS_ATTR_FATTR_TYPE ) entry->d_type =nfs_umode_to_dtype (entry->fattr->mode); - entry->prev_cookie = entry->cookie; entry->cookie = new_cookie;return 0 ;
include/linux/nfs_xdr.h +1 -2
@@ -745,8 +745,7 @@ struct nfs_auth_info {*/ struct nfs_entry {__u64 ino; -__u64 cookie, - prev_cookie; +__u64 cookie;const char * name;unsigned int len;int eof;
Signature · 1 line
2.35.1
Re: [PATCH v9 07/27] NFS: Store the change attribute in the directory page cache
In reply to [email protected]Hi Trond,
Quoted 5 lines · On Mon, Feb 28, 2022 at 5:51 AM <[email protected]> wrote:
From: Trond Myklebust <[email protected]>
Use the change attribute and the first cookie in a directory page cache entry to validate that the page is up to date.
Starting with this patch I'm seeing cthon basic tests fail on NFS v3:
Tue Mar 1 14:08:39 EST 2022
./server -b -o tcp,v3,sec=sys -m /mnt/nfsv3tcp -p /srv/test/anna/nfsv3tcp server ./server -b -o proto=tcp,sec=sys,v4.0 -m /mnt/nfsv4tcp -p /srv/test/anna/nfsv4tcp server
./server -b -o proto=tcp,sec=sys,v4.1 -m /mnt/nfsv41tcp -p /srv/test/anna/nfsv41tcp server
./server -b -o proto=tcp,sec=sys,v4.2 -m /mnt/nfsv42tcp -p /srv/test/anna/nfsv42tcp server
Waiting for 'b' to finish...
The '-b' test using '-o tcp,v3,sec=sys' args to server: Failed!!
Done: 14:08:41
Anna
Quoted 150 lines
Suggested-by Benjamin Coddington <[email protected]>Signed-off-by Trond Myklebust <[email protected]>--- fs/nfs/dir.c | 68 ++++++++++++++++++++++++++++------------------------ 1 file changed, 37 insertions(+), 31 deletions(-)fs/nfs/dir.c +37
-32@@ -140,6 +140,7 @@ struct nfs_cache_array_entry { };struct nfs_cache_array { +u64 change_attr;u64 last_cookie;unsigned int size; unsigned char page_full : 1,@@ -176,12 +177,14 @@ static void nfs_readdir_array_init(struct nfs_cache_array *array)memset (array,0 ,sizeof (struct nfs_cache_array )); } -static void nfs_readdir_page_init_array (struct page *page,u64 last_cookie) +static void nfs_readdir_page_init_array (struct page *page,u64 last_cookie, +u64 change_attr) {struct nfs_cache_array *array; array =kmap_atomic (page);nfs_readdir_array_init (array); + array->change_attr = change_attr; array->last_cookie = last_cookie; array->cookies_are_ordered =1 ;kunmap_atomic (array);@@ -208,7 +211,7 @@ nfs_readdir_page_array_alloc(u64 last_cookie, gfp_t gfp_flags) {struct page *page =alloc_page (gfp_flags);if (page) -nfs_readdir_page_init_array (page, last_cookie); +nfs_readdir_page_init_array (page, last_cookie,0 );return page; }@@ -305,19 +308,43 @@ int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page)return ret; } +static bool nfs_readdir_page_validate (struct page *page,u64 last_cookie, +u64 change_attr) +{ +struct nfs_cache_array *array =kmap_atomic (page); +int ret = true; + +if (array->change_attr != change_attr) + ret = false; +if (array->size >0 && array->array[0 ].cookie != last_cookie) + ret = false; +kunmap_atomic (array); +return ret; +} + +static void nfs_readdir_page_unlock_and_put (struct page *page) +{ +unlock_page (page); +put_page (page); +} +static struct page *nfs_readdir_page_get_locked (struct address_space *mapping,pgoff_t index,u64 last_cookie) {struct page *page; +u64 change_attr; page =grab_cache_page (mapping, index); -if (page && !PageUptodate (page)) { -nfs_readdir_page_init_array (page, last_cookie); -if (invalidate_inode_pages2_range (mapping, index +1 ,-1 ) <0 ) -nfs_zap_mapping (mapping->host, mapping); -SetPageUptodate (page); +if (!page) +return NULL ; + change_attr =inode_peek_iversion_raw (mapping->host); +if (PageUptodate (page)) { +if (nfs_readdir_page_validate (page, last_cookie, change_attr)) +return page; +nfs_readdir_clear_array (page); } - +nfs_readdir_page_init_array (page, last_cookie, change_attr); +SetPageUptodate (page);return page; }@@ -357,12 +384,6 @@ static void nfs_readdir_page_set_eof(struct page *page)kunmap_atomic (array); } -static void nfs_readdir_page_unlock_and_put (struct page *page) -{ -unlock_page (page); -put_page (page); -} -static struct page *nfs_readdir_page_get_next (struct address_space *mapping,pgoff_t index,u64 cookie) {@@ -419,16 +440,6 @@ static int nfs_readdir_search_for_pos(struct nfs_cache_array *array,return -EBADCOOKIE ; } -static bool -nfs_readdir_inode_mapping_valid (struct nfs_inode *nfsi) -{ -if (nfsi->cache_validity & (NFS_INO_INVALID_CHANGE | -NFS_INO_INVALID_DATA )) -return false; -smp_rmb (); -return !test_bit (NFS_INO_INVALIDATING , &nfsi->flags); -} -static bool nfs_readdir_array_cookie_in_range (struct nfs_cache_array *array,u64 cookie) {@@ -457,8 +468,7 @@ static int nfs_readdir_search_for_cookie(struct nfs_cache_array *array,struct nfs_inode *nfsi =NFS_I (file_inode (desc->file)); new_pos =nfs_readdir_page_offset (desc->page) + i; -if (desc->attr_gencount != nfsi->attr_gencount || - !nfs_readdir_inode_mapping_valid (nfsi)) { +if (desc->attr_gencount != nfsi->attr_gencount) { desc->duped =0 ; desc->attr_gencount = nfsi->attr_gencount; }else if (new_pos < desc->prev_index) {@@ -1095,11 +1105,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)* to either find the entry with the appropriate number or * revalidate the cookie. */ -if (ctx->pos ==0 ||nfs_attribute_cache_expired (inode)) { - res =nfs_revalidate_mapping (inode, file->f_mapping); -if (res <0 ) - goto out; - } +nfs_revalidate_inode (inode,NFS_INO_INVALID_CHANGE ); res = -ENOMEM ; desc =kzalloc (sizeof (*desc),GFP_KERNEL ); --2.35.1
Re: [PATCH v9 07/27] NFS: Store the change attribute in the directory page cache
Quoted 12 lines · Anna Schumaker
Hi Trond,
Quoted 6 lines · On Mon, Feb 28, 2022 at 5:51 AM <[email protected]> wrote:
From: Trond Myklebust <[email protected]>
Use the change attribute and the first cookie in a directory page cache
entry to validate that the page is up to date.Starting with this patch I'm seeing cthon basic tests fail on NFS v3:
Quoted 12 lines
Tue Mar 1 14:08:39 EST 2022
./server -b -o tcp,v3,sec=sys -m/mnt/nfsv3tcp-p/srv/test/anna/nfsv3tcpserver
./server -b -o proto=tcp,sec=sys,v4.0 -m/mnt/nfsv4tcp-p/srv/test/anna/nfsv4tcpserver
./server -b -o proto=tcp,sec=sys,v4.1 -m/mnt/nfsv41tcp-p/srv/test/anna/nfsv41tcpserver
./server -b -o proto=tcp,sec=sys,v4.2 -m/mnt/nfsv42tcp-p/srv/test/anna/nfsv42tcpserver
Waiting for 'b' to finish...
The '-b' test using '-o tcp,v3,sec=sys' args to server: Failed!!
Done: 14:08:41
Which tests are failing, and what is the server configuration that you're testing against?
I've not been seeing issues with either connectathon or xfstests on the platforms I've tested.
Quoted 179 lines
Anna
Quoted 177 lines
Suggested-by Benjamin Coddington <[email protected]>Signed-off-by Trond Myklebust <[email protected]>---
fs/nfs/dir.c | 68 ++++++++++++++++++++++++++++--------------------
----1 file changed, 37 insertions(+), 31 deletions(-)fs/nfs/dir.c +4
-1@@ -140,6 +140,7 @@ struct nfs_cache_array_entry { }; struct nfs_cache_array { + u64 change_attr; u64 last_cookie; unsigned int size; unsigned char page_full : 1,@@ -176,12 +177,14 @@ static void nfs_readdir_array_init(struct nfs_cache_array *array) memset(array, 0, sizeof(struct nfs_cache_array)); } -static void nfs_readdir_page_init_array (struct page *page,u64 last_cookie) +static void nfs_readdir_page_init_array(struct page *page, u64 last_cookie, + u64 change_attr) { struct nfs_cache_array *array; array = kmap_atomic(page); nfs_readdir_array_init(array); + array->change_attr = change_attr;
array->last_cookie=last_cookie;
array->cookies_are_ordered= 1;
kunmap_atomic(array);fs/nfs/dir.c +21
-1@@ -208,7 +211,7 @@ nfs_readdir_page_array_alloc(u64 last_cookie, gfp_t gfp_flags) { struct page *page = alloc_page(gfp_flags); if (page) - nfs_readdir_page_init_array(page, last_cookie); + nfs_readdir_page_init_array(page, last_cookie, 0); return page; }@@ -305,19 +308,43 @@ int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page) return ret; } +static bool nfs_readdir_page_validate (struct page *page,u64 last_cookie, + u64 change_attr) +{ + struct nfs_cache_array *array = kmap_atomic(page); + int ret = true; + + if (array->change_attr != change_attr) + ret = false; + if (array->size > 0 && array->array[0].cookie != last_cookie) + ret = false; + kunmap_atomic(array); + return ret; +} + +static void nfs_readdir_page_unlock_and_put (struct page *page) +{ + unlock_page(page); + put_page(page); +} +static struct page *nfs_readdir_page_get_locked(struct
address_space*mapping,
pgoff_tindex, u64last_cookie)
{
struct page *page;
+ u64change_attr;page =
grab_cache_page(mapping, index);
- if (page && !PageUptodate(page)) {
-nfs_readdir_page_init_array(page, last_cookie); - if (invalidate_inode_pages2_range(mapping, index + 1, -1)< 0)
-nfs_zap_mapping(mapping->host, mapping); - SetPageUptodate(page);
+ if (!page)
+ return NULL;
+change_attr=inode_peek_iversion_raw(mapping->host); + if (PageUptodate(page)) {
+ if (nfs_readdir_page_validate(page, last_cookie, change_attr))
+ return page;
+nfs_readdir_clear_array(page);
}
-
+nfs_readdir_page_init_array(page, last_cookie, change_attr);
+ SetPageUptodate(page);
return page;
}fs/nfs/dir.c +0
-6@@ -357,12 +384,6 @@ static void nfs_readdir_page_set_eof(struct page *page) kunmap_atomic(array); } -static void nfs_readdir_page_unlock_and_put (struct page *page) -{ - unlock_page(page); - put_page(page); -} -static struct page *nfs_readdir_page_get_next(struct
address_space*mapping,
pgoff_tindex, u64 cookie)
{fs/nfs/dir.c +0
-10@@ -419,16 +440,6 @@ static int nfs_readdir_search_for_pos(structnfs_cache_array *array, return -EBADCOOKIE; } -static bool -nfs_readdir_inode_mapping_valid (struct nfs_inode *nfsi) -{ - if (nfsi->cache_validity & (NFS_INO_INVALID_CHANGE | - NFS_INO_INVALID_DATA)) - return false; - smp_rmb(); - return !test_bit(NFS_INO_INVALIDATING, &nfsi->flags); -} -static bool nfs_readdir_array_cookie_in_range(struct
nfs_cache_array*array,
u64 cookie) {fs/nfs/dir.c +0
-0@@ -457,8 +468,7 @@ static int nfs_readdir_search_for_cookie(struct nfs_cache_array *array,struct
nfs_inode*nfsi =
NFS_I(file_inode(desc->file));
new_pos= nfs_readdir_page_offset(desc-Quoted 1 line
page) + i;
- if (
desc->attr_gencount!= nfsi-Quoted 1 line
attr_gencount||- !nfs_readdir_inode_mapping_valid(nfsi)) {
+ if (desc->attr_gencount!= nfsi-Quoted 1 line
attr_gencount) {
desc->duped= 0;desc->attr_gencount= nfsi-Quoted 1 line
attr_gencount;} else if (
new_pos<desc->prev_index) {fs/nfs/dir.c +1
-5@@ -1095,11 +1105,7 @@ static int nfs_readdir(struct file *file,struct dir_context *ctx) * to either find the entry with the appropriate number or * revalidate the cookie. */ - if (ctx->pos == 0 || nfs_attribute_cache_expired(inode)) { - res = nfs_revalidate_mapping(inode, file- >f_mapping); - if (res < 0) - goto out; - } + nfs_revalidate_inode(inode, NFS_INO_INVALID_CHANGE);res =
-ENOMEM;
desc =kzalloc(sizeof(*desc), GFP_KERNEL); --
2.35.1
Signature · 6 lines
Trond Myklebust CTO, Hammerspace Inc 4984 El Camino Real, Suite 208 Los Altos, CA 94022 www.hammer.space
Re: [PATCH v9 03/27] NFS: Trace lookup revalidation failure
In reply to [email protected]Quoted 7 lines · Trond Myklebust
Quoted 2 lines · Benjamin Coddington
There's a path through
nfs4_lookup_revalidatethat will now only produce this exit tracepoint. Does it need the_entertracepoint added?You're thinking about the
nfs_lookup_revalidate_delegated()path? The_enter()tracepoint doesn't provide any useful information that isn't already provided by the_exit(), AFAICS.
No, the path through nfs4_do_lookup_revalidate(), reval_dentry: jump. But I agree there's not much value in the _enter() tracepoint. Maybe we can remove it, and _exit more like _done.
I am thinking about hearing back from folks about mis-matched _enter() and _exit() results, but also realize this is nit-picking.
Ben
Re: [PATCH v9 03/27] NFS: Trace lookup revalidation failure
Quoted 13 lines · Benjamin Coddington
Quoted 7 lines · Trond Myklebust
Quoted 2 lines · Benjamin Coddington
There's a path through
nfs4_lookup_revalidatethat will now only produce this exit tracepoint. Does it need the_entertracepoint added?You're thinking about the
nfs_lookup_revalidate_delegated()path? The_enter()tracepoint doesn't provide any useful information that isn't already provided by the_exit(), AFAICS.
Quoted 6 lines
No, the path through
nfs4_do_lookup_revalidate(),reval_dentry: jump. But I agree there's not much value in the_enter()tracepoint. Maybe we can remove it, and_exitmore like_done.I am thinking about hearing back from folks about mis-matched
_enter()and_exit()results, but also realize this is nit-picking.
I think the _enter / _exit trace points simply replaced dprintk call sites which did much the same reporting. Maybe we should consider replacing some of these because we can rely on function call tracing instead.
But generally we like to see trace points that report exceptional events rather than "I made it to this point". The latter category of trace points are interesting while code is under development but often loses its value once the code is in the field.
--
Chuck Lever
Re: [PATCH v9 14/27] NFS: Improve heuristic for readdirplus
In reply to [email protected]Quoted 11 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
The heuristic for readdirplus is designed to try to detect 'ls -l' and similar patterns. It does so by looking for cache hit/miss patterns in both the attribute cache and in the dcache of the files in a given directory, and then sets a flag for the readdirplus code to interpret.
The problem with this approach is that a single attribute or dcache miss can cause the NFS code to force a refresh of the attributes for the entire set of files contained in the directory.
Quoted 5 lines
To be able to make a more nuanced decision, let's sample the number of hits and misses in the set of open directory descriptors. That allows us to set thresholds at which we start preferring READDIRPLUS over regular READDIR, or at which we start to force a re-read of the remaining readdir cache using READDIRPLUS.
I like this patch very much.
The heuristic doesn't kick-in until "ls -l" makes its second call into nfs_readdir(), and for my filenames with 8 chars, that means that there are about 5800 GETATTRs generated before we clean the cache to do more READDIRPLUS. That's a large number to compound on connection latency.
We've already got some complaints that folk's 2nd "ls -l" takes "so much longer" after 1a34c8c9a49e.
Can we possibly limit our first pass through nfs_readdir() so that the heuristic takes effect sooner?
Ben
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
In reply to [email protected]Quoted 9 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of the first entry, since that is what we use to match the page anyway.
This allows us to avoid re-reading the entire cache on a
seekdir()type of operation. The latter is very common when re-exporting NFS, and is a major performance drain.
Quoted 9 lines
The change does affect our duplicate cookie detection, since we can no longer rely on the page index as a linear offset for detecting whether we looped backwards. However since we no longer do a linear search through all the pages on each call to
nfs_readdir(), this is less of a concern than it was previously.
The other downside is thatinvalidate_mapping_pages()no longer can use the page index to avoid clearing pages that have been read. A subsequent patch will restore the functionality this provides to the 'ls -l' heuristic.
I didn't realize the approach was to also hash out the linearly-cached entries. I thought we'd do something like flag the context for hashed page indexes after a seekdir event, and if there are collisions with the linear entries, they'll get fixed up when found.
Doesn't that mean that with this approach seekdir() only hits the same pages when the entry offset is page-aligned? That's 1 in 127 odds.
It also means we're amplifying the pagecache's useage for slightly changing directories - rather than re-using the same pages we're scattering our usage across the index. Eh, maybe not a big deal if we just expect the page cache's LRU to do the work.
Ben
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
Quoted 26 lines · Benjamin Coddington
Quoted 9 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of the first entry, since that is what we use to match the page anyway.
This allows us to avoid re-reading the entire cache on a
seekdir()type of operation. The latter is very common when re-exporting NFS, and is a major performance drain.Quoted 9 lines
The change does affect our duplicate cookie detection, since we can no longer rely on the page index as a linear offset for detecting whether we looped backwards. However since we no longer do a linear search through all the pages on each call to
nfs_readdir(), this is less of a concern than it was previously.
The other downside is thatinvalidate_mapping_pages()no longer can use the page index to avoid clearing pages that have been read. A subsequent patch will restore the functionality this provides to the 'ls -l' heuristic.I didn't realize the approach was to also hash out the linearly-cached entries. I thought we'd do something like flag the context for hashed page indexes after a seekdir event, and if there are collisions with the linear entries, they'll get fixed up when found.
Quoted 7 lines
Doesn't that mean that with this approach
seekdir()only hits the same pages when the entry offset is page-aligned? That's 1 in 127 odds.It also means we're amplifying the pagecache's useage for slightly changing directories - rather than re-using the same pages we're scattering our usage across the index. Eh, maybe not a big deal if we just expect the page cache's LRU to do the work.
I don't have a better idea, though.. have you tested this performance?
..
maybe.. the hash divided the u64 cookie space into 262144 buckets, each being
a page the cookie could fall into. So cookies 1 - 70368744177663 map into
page 1.. bah. That wont work.
I was worried that I was wrong about this, but this program shows the problem by requiring a full READDIR for each entry if we walk the entries one-by-one with lseek(). I don't understand how the re-export seekdir() case is helped by this unless you're hitting the exact same offsets all the time.
I think that a hash of the page index for seekdir is no better than picking an arbitrary offset, or just using the lowest pages in the cache.
Ben
#define _GNU_SOURCE
#include <stdio.h>
#include <unistd.h>
#include <fcntl.h>
#include <sched.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/syscall.h>
#include <string.h>
#define NFSDIR "/mnt/fedora/127_dentries"
int main(int argc, char **argv)
{
int i, dir_fd, bpos, total = 0;
size_t nread;
struct linux_dirent {
long d_ino;
off_t d_off;
unsigned short d_reclen;
char d_name[];
};
struct linux_dirent *d;
int buf_size = sizeof(struct linux_dirent) + sizeof("file_000");
char buf[buf_size];
/* create files: */
for (i = 0; i < 127; i++) {
sprintf(buf, NFSDIR "/file_%03d", i);
close(open(buf, O_CREAT, 666));
}
dir_fd = open(NFSDIR, O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC);
if (dir_fd < 0) {
perror("cannot open dir");
return 1;
}
/* no cache pls */
posix_fadvise(dir_fd, 0, 0, POSIX_FADV_DONTNEED);
while (1) {
nread = syscall(SYS_getdents, dir_fd, buf, buf_size);
if (nread == 0 || nread == -1)
break;
for (bpos = 0; bpos < nread;) {
d = (struct linux_dirent *) (buf + bpos);
printf("%s offset %lu\n", d->d_name, d->d_off);
lseek(dir_fd, 0, SEEK_SET);
lseek(dir_fd, d->d_off, SEEK_SET);
total++;
bpos += d->d_reclen;
}
}
printf("Listing 1: %d total dirents\n", total);
close(dir_fd);
return 0;
}Quoted 26 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
The current NFS readdir code will always try to maximise the amount of readahead it performs on the assumption that we can cache anything that isn't immediately read by the process.
There are several cases where this assumption breaks down, including when the 'ls -l' heuristic kicks in to try to force use of readdirplus as a batch replacement for lookup/getattr.This series also implement Ben's page cache filter to ensure that we can improve the ability to share cached data between processes that are reading the same directory at the same time, and to avoid live-locks when the directory is simultaneously changing.
--
v2: Remove reset of dtsize whenNFS_INO_FORCE_READDIRis set
v3: Avoid excessive window shrinking inuncached_readdircase
v4: Track 'ls -l' cache hit/miss statistics
Improved algorithm for falling back to uncached readdir Skip readdirplus when files are being written to
v5: bugfixes
Skip readdirplus when the acdirmax/acregmax values are low Request a full XDR buffer when doing READDIRPLUS
v6: Add tracing
Don't have lookup request readdirplus when it won't help
v7: Implement Ben's page cache filter
Quoted 8 lines
Reduce the use of uncached readdir
Change indexing of the page cache to improveseekdir()performance.
v8: Reduce the page cache overhead by shrinking the cookie hashvalue size
Incorporate other feedback from Anna, Ben and Neil
Fix nfs2/3_decode_dirent return values
Fix the change attribute value set innfs_readdir_page_get_next()
v9: Address bugs that were hit when testing with large directories
Misc cleanups
Hi Trond, thanks for all this work! I went through these from your testing branch (612896ec5a4e) rather than the posting.
If it pleases, these can all get marked with my
and/or
.. except for 25/27, which is missing from the testing branch.
As I replied to 23/27, I don't understand how the page index hashing is going to help out the re-export seekdir case, I think it might make it worse, and I think its unnecessary extra complication.
I did test extensively total directory listing correctness, and it appears to me that you are correct, we are not regressing. We're in a similar place as before. I think we can be even more correct with directory verifiers or post-op updates with GETATTR in the READDIR compound for very little cost, but I've already made those arguments a few weeks ago.
Thanks again,
Ben
Re: [PATCH v9 03/27] NFS: Trace lookup revalidation failure
In reply to Chuck Lever IIIQuoted 24 lines · Chuck Lever III
Quoted 11 lines · Benjamin Coddington
Quoted 7 lines · Trond Myklebust
Quoted 2 lines · Benjamin Coddington
There's a path through
nfs4_lookup_revalidatethat will now only produce this exit tracepoint. Does it need the_entertracepoint added?You're thinking about the
nfs_lookup_revalidate_delegated()path? The_enter()tracepoint doesn't provide any useful information that isn't already provided by the_exit(), AFAICS.Quoted 7 lines
No, the path through
nfs4_do_lookup_revalidate(),reval_dentry: jump.
But I
agree there's not much value in the_enter()tracepoint. Maybe we can remove it, and_exitmore like_done.I am thinking about hearing back from folks about mis-matched
_enter()and_exit()results, but also realize this is nit-picking.I think the
_enter/_exittrace points simply replaced dprintk call sites which did much the same reporting. Maybe we should consider replacing some of these because we can rely on function call tracing instead.
Quoted 5 lines
But generally we like to see trace points that report exceptional events rather than "I made it to this point". The latter category of trace points are interesting while code is under development but often loses its value once the code is in the field.
Instead of "hearing back from folks" I should have said "I worry our QE team is going to discover and possibly report as a bug". :P Thanks for filling in Chuck!
Ben
PATCH 1/1 NFS: Trigger "ls -l" readdir heuristic sooner
In reply to Benjamin Coddington.. Something like this does the trick in my testing, but yes will have an impact on regular workloads:
8<------------------------------------------------------------------------
Since commit 1a34c8c9a49e ("NFS: Support larger readdir buffers") has updated dtsize and recent improvements to the READDIRPLUS helper heuristic, the heuristic may not trigger until many dentries are emitted to userspace, which may cause many thousands of GETATTR calls for "ls -l" when the directory's pagecache has already been populated. This typically manifests as a much slower total runtime for a _second_ invocation of "ls -l" within the directory attribute cache timeouts.
Fix this by emitting only 17 entries for any first pass through the NFS directory's ->iterate_shared(), which will allow userpace to prime the counters for the heuristic.
--- fs/nfs/dir.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)
fs/nfs/dir.c +7 -2
@@ -1060,6 +1060,8 @@ static int readdir_search_pagecache(struct nfs_readdir_descriptor *desc)return res; } +#define NFS_READDIR_CACHE_MISS_THRESHOLD (16UL) +/* * Once we've found the start of the dirent within a page: fill 'er up... */
@@ -1069,6 +1071,7 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc,struct file *file = desc->file;struct nfs_cache_array *array;unsigned int i; +bool first_emit = !desc->dir_cookie; array =kmap (desc->page);for (i = desc->cache_entry_index; i < array->size; i++) {
@@ -1092,6 +1095,10 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc, desc->ctx->pos = desc->dir_cookie; else desc->ctx->pos++; +if (first_emit && i >NFS_READDIR_CACHE_MISS_THRESHOLD +1 ) { + desc->eob = true; +break ; + } }if (array->page_is_eof) desc->eof = !desc->eob;
@@ -1173,8 +1180,6 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc)return status; } -#define NFS_READDIR_CACHE_MISS_THRESHOLD (16UL) - static bool nfs_readdir_handle_cache_misses(struct inode *inode, struct nfs_readdir_descriptor *desc, unsigned int cache_misses,
Signature · 1 line
2.31.1
Re: [PATCH v9 14/27] NFS: Improve heuristic for readdirplus
In reply to Benjamin CoddingtonQuoted 36 lines · Benjamin Coddington
Quoted 15 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
The heuristic for readdirplus is designed to try to detect 'ls -l' and
similar patterns. It does so by looking for cache hit/miss patterns in
both the attribute cache and in the dcache of the files in a given directory, and then sets a flag for the readdirplus code to interpret.The problem with this approach is that a single attribute or dcache miss
can cause the NFS code to force a refresh of the attributes for the entire set of files contained in the directory.Quoted 8 lines
To be able to make a more nuanced decision, let's sample the number of
hits and misses in the set of open directory descriptors. That allows us
to set thresholds at which we start preferring READDIRPLUS over regular
READDIR, or at which we start to force a re-read of the remaining readdir cache using READDIRPLUS.I like this patch very much.
The heuristic doesn't kick-in until "ls -l" makes its second call into
nfs_readdir(), and for my filenames with 8 chars, that means that there are
about 5800 GETATTRs generated before we clean the cache to do more READDIRPLUS. That's a large number to compound on connection latency.
Quoted 8 lines
We've already got some complaints that folk's 2nd "ls -l" takes "so much
longer" after1a34c8c9a49e.Can we possibly limit our first pass through
nfs_readdir()so that the
heuristic takes effect sooner?
The problem is really that 'ls' (or possibly glibc) is passing in a pretty huge buffer to the getdents() system call.
On my setup, that buffer appears to be 80K in size. So what happens is that we get that first getdents() call, and so we fill the 80K buffer with as many files as will fit. That can quickly run into several thousand entries, if the filenames are relatively short.
Then 'ls' goes through the contents and does a stat() (or a statx()) on each entry, and so we record the statistics. However that means those first several thousand entries are indeed going to use cached data, or force GETATTR to go on the wire. We only start using forced readdirplus on the second pass.
Yes, I suppose we could limit getdents() to ignore the buffer size, and just return fewer entries, however what's the "right" size in that case?
More to the point, how much pain are we going to accept before we give up trying these assorted heuristics, and just define a readdirplus() system call modelled on statx()?
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
In reply to Benjamin CoddingtonQuoted 30 lines · Benjamin Coddington
Quoted 13 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of
the first entry, since that is what we use to match the page anyway.This allows us to avoid re-reading the entire cache on a
seekdir()type
of operation. The latter is very common when re-exporting NFS, and is a
major performance drain.Quoted 14 lines
The change does affect our duplicate cookie detection, since we can no
longer rely on the page index as a linear offset for detecting whether
we looped backwards. However since we no longer do a linear search through all the pages on each call tonfs_readdir(), this is less of a
concern than it was previously.
The other downside is thatinvalidate_mapping_pages()no longer can use
the page index to avoid clearing pages that have been read. A subsequent
patch will restore the functionality this provides to the 'ls -l' heuristic.
Quoted 7 lines
I didn't realize the approach was to also hash out the linearly- cached
entries. I thought we'd do something like flag the context for hashed page
indexes after a seekdir event, and if there are collisions with the linear
entries, they'll get fixed up when found.
Why? What's the point of using 2 models where 1 will do?
Quoted 4 lines
Doesn't that mean that with this approach
seekdir()only hits the same pages
when the entry offset is page-aligned? That's 1 in 127 odds.
The point is not to stomp all over the pages that contain aligned data when the application does call seekdir().
IOW: we always optimise for the case where we do a linear read of the directory, but we support random seekdir() + read too.
Quoted 9 lines
It also means we're amplifying the pagecache's useage for slightly changing
directories - rather than re-using the same pages we're scattering our usage
across the index. Eh, maybe not a big deal if we just expect the page
cache's LRU to do the work.
I don't understand your point about 'not reusing'. If the user seeks to the same cookie, we reuse the page. However I don't know how you would go about setting up a schema that allows you to seek to an arbitrary cookie without doing a linear search.
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH v9 14/27] NFS: Improve heuristic for readdirplus
In reply to Trond MyklebustQuoted 15 lines · Trond Myklebust
The problem is really that 'ls' (or possibly glibc) is passing in a pretty huge buffer to the
getdents()system call.On my setup, that buffer appears to be 80K in size. So what happens is that we get that first
getdents()call, and so we fill the 80K buffer with as many files as will fit. That can quickly run into several thousand entries, if the filenames are relatively short.Then 'ls' goes through the contents and does a
stat()(or astatx()) on each entry, and so we record the statistics. However that means those first several thousand entries are indeed going to use cached data, or force GETATTR to go on the wire. We only start using forced readdirplus on the second pass.
Quoted 3 lines
Yes, I suppose we could limit
getdents()to ignore the buffer size, and just return fewer entries, however what's the "right" size in that case?
We can return fewer entries on the first call, so for the first pass the right size is NFS_READDIR_CACHE_MISS_THRESHOLD + 1. I sent a patch.
Quoted 3 lines
More to the point, how much pain are we going to accept before we give up trying these assorted heuristics, and just define a
readdirplus()system call modelled onstatx()?
We cursed ourselves by creating the heuristic, and now we've had to maintain it and try to make everyone happy. The pain for us is when the behavior keeps changing after sites have come to rely on previous performance.
I hope you can take a look at the patch.
Ben
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
In reply to Trond MyklebustQuoted 39 lines · Trond Myklebust
Quoted 30 lines · Benjamin Coddington
Quoted 13 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of
the first entry, since that is what we use to match the page anyway.This allows us to avoid re-reading the entire cache on a
seekdir()type
of operation. The latter is very common when re-exporting NFS, and is a
major performance drain.Quoted 14 lines
The change does affect our duplicate cookie detection, since we can no
longer rely on the page index as a linear offset for detecting whether
we looped backwards. However since we no longer do a linear search through all the pages on each call tonfs_readdir(), this is less of a
concern than it was previously.
The other downside is thatinvalidate_mapping_pages()no longer can use
the page index to avoid clearing pages that have been read. A subsequent
patch will restore the functionality this provides to the 'ls -l' heuristic.Quoted 7 lines
I didn't realize the approach was to also hash out the linearly- cached
entries. I thought we'd do something like flag the context for hashed page
indexes after a seekdir event, and if there are collisions with the linear
entries, they'll get fixed up when found.
Quoted 1 line
Why? What's the point of using 2 models where 1 will do?
I don't think the hashed model is quite as simple and efficient overall, and may produce impacts to a system beyond NFS.
Quoted 10 lines
Quoted 4 lines
Doesn't that mean that with this approach
seekdir()only hits the same pages
when the entry offset is page-aligned? That's 1 in 127 odds.The point is not to stomp all over the pages that contain aligned data when the application does call
seekdir().IOW: we always optimise for the case where we do a linear read of the directory, but we support random
seekdir()+ read too.
And that could be done just by bumping the seekdir users to some constant offset (index 262144 ?), or something else equally dead-nuts simple. That keeps tightly clustered page indexes, so walking the cache is faster. That reduces the "buckshot" effect the hashing has of eating up pagecache pages they'll never use again. That doesn't cap our caching ability at 33 million entries.
Its weird to me that we're doing exactly what XArray says not to do, hash the index, when we don't have to.
Quoted 9 lines
Quoted 8 lines
It also means we're amplifying the pagecache's useage for slightly changing
directories - rather than re-using the same pages we're scattering our usage
across the index. Eh, maybe not a big deal if we just expect the page
cache's LRU to do the work.
Quoted 4 lines
I don't understand your point about 'not reusing'. If the user seeks to the same cookie, we reuse the page. However I don't know how you would go about setting up a schema that allows you to seek to an arbitrary cookie without doing a linear search.
So when I was taking about 'reusing' a page, that's about re-filling the same pages rather than constantly conjuring new ones, which requires less of the pagecache's resources in total. Maybe the pagecache can handle that without it negatively impacting other users of the cache that /will/ re-use their cached pages, but I worry it might be irresponsible of us to fill the pagecache with pages we know we're never going to find again.
Ben
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
Quoted 68 lines · Benjamin Coddington
Quoted 47 lines · Trond Myklebust
Quoted 37 lines · Benjamin Coddington
Quoted 26 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of
the first entry, since that is what we use to match the page anyway.This allows us to avoid re-reading the entire cache on a
seekdir()
type
of operation. The latter is very common when re-exporting NFS, and
is a
major performance drain.The change does affect our duplicate cookie detection, since we can
no
longer rely on the page index as a linear offset for detecting whether
we looped backwards. However since we no longer do a linear search
through all the pages on each call tonfs_readdir(), this is less
of a
concern than it was previously.Quoted 8 lines
The other downside is that
invalidate_mapping_pages()no longer can
use
the page index to avoid clearing pages that have been read. A subsequent
patch will restore the functionality this provides to the 'ls - l'
heuristic.Quoted 8 lines
I didn't realize the approach was to also hash out the linearly- cached
entries. I thought we'd do something like flag the context for hashed page
indexes after a seekdir event, and if there are collisions with the
linear
entries, they'll get fixed up when found.Quoted 1 line
Why? What's the point of using 2 models where 1 will do?
I don't think the hashed model is quite as simple and efficient overall, and
may produce impacts to a system beyond NFS.Quoted 5 lines
Quoted 4 lines
Doesn't that mean that with this approach
seekdir()only hits the same pages
when the entry offset is page-aligned? That's 1 in 127 odds.Quoted 7 lines
The point is not to stomp all over the pages that contain aligned data
when the application does callseekdir().IOW: we always optimise for the case where we do a linear read of the
directory, but we support randomseekdir()+ read too.
Quoted 12 lines
And that could be done just by bumping the seekdir users to some constant
offset (index 262144 ?), or something else equally dead-nuts simple. That
keeps tightly clustered page indexes, so walking the cache is faster. That
reduces the "buckshot" effect the hashing has of eating up pagecache pages
they'll never use again. That doesn't cap our caching ability at 33 million
entries.
What you say would make sense if readdir cookies truly were offsets, but in general they're not. Cookies are unstructured data, and should be treated as unstructured data.
Let's say I do cache more than 33 million entries and I have to find a cookie. I have to linearly read through at least 1GB of cached data before I can give up and start a new readdir. Either that, or I need to have a heuristic that tells me when to stop searching, and then another heuristic that tells me where to store the data in a way that doesn't trash the page cache.
With the hashing, I seek to the page matching the hash, and I either immediately find what I need, or I immediately know to start a readdir. There is no need for any additional heuristic.
Quoted 23 lines
Its weird to me that we're doing exactly what XArray says not to do, hash
the index, when we don't have to.Quoted 11 lines
Quoted 10 lines
It also means we're amplifying the pagecache's useage for slightly
changing
directories - rather than re-using the same pages we're scattering
our usage
across the index. Eh, maybe not a big deal if we just expect the page
cache's LRU to do the work.Quoted 7 lines
I don't understand your point about 'not reusing'. If the user seeks to
the same cookie, we reuse the page. However I don't know how you would
go about setting up a schema that allows you to seek to an arbitrary
cookie without doing a linear search.
Quoted 12 lines
So when I was taking about 'reusing' a page, that's about re-filling the
same pages rather than constantly conjuring new ones, which requires less of
the pagecache's resources in total. Maybe the pagecache can handle that
without it negatively impacting other users of the cache that /will/ re-use
their cached pages, but I worry it might be irresponsible of us to fill the
pagecache with pages we know we're never going to find again.
In the case where the processes are reading linearly through a directory that is not changing (or at least where the beginning of the directory is not changing), we will reuse the cached data, because just like in the linearly indexed case, each process ends up reading the exact same sequence of cookies, and looking up the exact same sequence of hashes.
The sequences start to diverge only if they hit a part of the directory that is being modified. At that point, we're going to be invalidating page cache entries anyway with the last reader being more likely to be following the new sequence of cookies.
The hashed indexing does come with a cost, thanks to XArray but that cost is limited to a max of 8MB with the current scheme.
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
Quoted 93 lines · Trond Myklebust
Quoted 68 lines · Benjamin Coddington
Quoted 47 lines · Trond Myklebust
Quoted 37 lines · Benjamin Coddington
Quoted 26 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the cookie of
the first entry, since that is what we use to match the page anyway.This allows us to avoid re-reading the entire cache on a
seekdir()
type
of operation. The latter is very common when re-exporting NFS, and
is a
major performance drain.The change does affect our duplicate cookie detection, since we can
no
longer rely on the page index as a linear offset for detecting whether
we looped backwards. However since we no longer do a linear search
through all the pages on each call tonfs_readdir(), this is less
of a
concern than it was previously.Quoted 8 lines
The other downside is that
invalidate_mapping_pages()no longer can
use
the page index to avoid clearing pages that have been read. A subsequent
patch will restore the functionality this provides to the 'ls - l'
heuristic.Quoted 8 lines
I didn't realize the approach was to also hash out the linearly- cached
entries. I thought we'd do something like flag the context for hashed page
indexes after a seekdir event, and if there are collisions with the
linear
entries, they'll get fixed up when found.Quoted 1 line
Why? What's the point of using 2 models where 1 will do?
I don't think the hashed model is quite as simple and efficient overall, and
may produce impacts to a system beyond NFS.Quoted 5 lines
Quoted 4 lines
Doesn't that mean that with this approach
seekdir()only hits the same pages
when the entry offset is page-aligned? That's 1 in 127 odds.Quoted 7 lines
The point is not to stomp all over the pages that contain aligned data
when the application does callseekdir().IOW: we always optimise for the case where we do a linear read of the
directory, but we support randomseekdir()+ read too.Quoted 12 lines
And that could be done just by bumping the seekdir users to some constant
offset (index 262144 ?), or something else equally dead-nuts simple. That
keeps tightly clustered page indexes, so walking the cache is faster. That
reduces the "buckshot" effect the hashing has of eating up pagecache pages
they'll never use again. That doesn't cap our caching ability at 33 million
entries.What you say would make sense if readdir cookies truly were offsets, but in general they're not. Cookies are unstructured data, and should be treated as unstructured data.
Let's say I do cache more than 33 million entries and I have to find a cookie. I have to linearly read through at least 1GB of cached data before I can give up and start a new readdir. Either that, or I need to have a heuristic that tells me when to stop searching, and then another heuristic that tells me where to store the data in a way that doesn't trash the page cache.
Quoted 3 lines
With the hashing, I seek to the page matching the hash, and I either immediately find what I need, or I immediately know to start a readdir. There is no need for any additional heuristic.
The scenario where we want to find a cookie while not doing a linear pass through the directory will be the seekdir() case. In a linear walk, we have the cached page index to help. So in the seekdir case, the chances of having someone already fill a page and also having the cookie be the 1 in 127 that are page-aligned (and so match an already cached page) are small, I think. Unless your use-case will often hit the exact same offsets over and over.
So with the hashing and seekdir case, I think that the cache will be pretty heavily filled with the same duplicated data at various offsets and rarely useful. That's why I wondered if you'd tested your use-case for it and found it to be advantageous. I think what we've got is going to work fine, but I wonder if you've seen it to work well.
The major pain point most of our users complain about is not being able to perform a complete walk in linear time with respect to size with invalidations at play. This series fixes that, and is a huge bonus. Other smaller performance improvements are pale in comparison for us, and might just get us forever chasing one or two minor optimizations that have trade-offs.
There's a lot of variables at play. For some client/server setups (like some low-latency RDMA), and very large directories and cache sizes, it might be more performant to just do the READDIR every time, walking local caches be damned.
Quoted 43 lines
Quoted 23 lines
Its weird to me that we're doing exactly what XArray says not to do, hash
the index, when we don't have to.Quoted 11 lines
Quoted 10 lines
It also means we're amplifying the pagecache's useage for slightly
changing
directories - rather than re-using the same pages we're scattering
our usage
across the index. Eh, maybe not a big deal if we just expect the page
cache's LRU to do the work.Quoted 7 lines
I don't understand your point about 'not reusing'. If the user seeks to
the same cookie, we reuse the page. However I don't know how you would
go about setting up a schema that allows you to seek to an arbitrary
cookie without doing a linear search.Quoted 12 lines
So when I was taking about 'reusing' a page, that's about re-filling the
same pages rather than constantly conjuring new ones, which requires less of
the pagecache's resources in total. Maybe the pagecache can handle that
without it negatively impacting other users of the cache that /will/ re-use
their cached pages, but I worry it might be irresponsible of us to fill the
pagecache with pages we know we're never going to find again.In the case where the processes are reading linearly through a directory that is not changing (or at least where the beginning of the directory is not changing), we will reuse the cached data, because just like in the linearly indexed case, each process ends up reading the exact same sequence of cookies, and looking up the exact same sequence of hashes.
Quoted 4 lines
The sequences start to diverge only if they hit a part of the directory that is being modified. At that point, we're going to be invalidating page cache entries anyway with the last reader being more likely to be following the new sequence of cookies.
I don't think we clean up behind ourselves anymore. Now that we are going to validate each page before using it, we don't invalidate the whole cache at any point. That means that a divergence duplicates the pagecache usage beyond the divergence.
Ben
Re: [PATCH v9 23/27] NFS: Convert readdir page cache to use a cookie based index
Quoted 125 lines · Benjamin Coddington
Quoted 117 lines · Trond Myklebust
Quoted 83 lines · Benjamin Coddington
Quoted 59 lines · Trond Myklebust
Quoted 46 lines · Benjamin Coddington
Quoted 35 lines · On 27 Feb 2022, at 18:12, [email protected] wrote:
From: Trond Myklebust <[email protected]>
Instead of using a linear index to address the pages, use the
cookie of
the first entry, since that is what we use to match the page
anyway.This allows us to avoid re-reading the entire cache on a
seekdir()
type
of operation. The latter is very common when re-exporting NFS,
and
is a
major performance drain.The change does affect our duplicate cookie detection, since we
can
no
longer rely on the page index as a linear offset for detecting
whether
we looped backwards. However since we no longer do a linear search
through all the pages on each call tonfs_readdir(), this is
less
of a
concern than it was previously.
The other downside is thatinvalidate_mapping_pages()no longer
canQuoted 8 lines
use
the page index to avoid clearing pages that have been read. A
subsequent
patch will restore the functionality this provides to the 'ls -
l'
heuristic.Quoted 11 lines
I didn't realize the approach was to also hash out the linearly-
cached
entries. I thought we'd do something like flag the context for
hashed page
indexes after a seekdir event, and if there are collisions with
the
linear
entries, they'll get fixed up when found.Quoted 1 line
Why? What's the point of using 2 models where 1 will do?
I don't think the hashed model is quite as simple and efficient overall, and
may produce impacts to a system beyond NFS.Quoted 11 lines
Quoted 5 lines
Doesn't that mean that with this approach
seekdir()only hits the
same pages
when the entry offset is page-aligned? That's 1 in 127 odds.The point is not to stomp all over the pages that contain aligned
data
when the application does callseekdir().Quoted 4 lines
IOW: we always optimise for the case where we do a linear read of
the
directory, but we support randomseekdir()+ read too.Quoted 15 lines
And that could be done just by bumping the seekdir users to some constant
offset (index 262144 ?), or something else equally dead-nuts simple.
That
keeps tightly clustered page indexes, so walking the cache is faster. That
reduces the "buckshot" effect the hashing has of eating up pagecache
pages
they'll never use again. That doesn't cap our caching ability at 33
million
entries.What you say would make sense if readdir cookies truly were offsets,
but in general they're not. Cookies are unstructured data, and should
be treated as unstructured data.Let's say I do cache more than 33 million entries and I have to find a
cookie. I have to linearly read through at least 1GB of cached data before I can give up and start a new readdir. Either that, or I need to
have a heuristic that tells me when to stop searching, and then another
heuristic that tells me where to store the data in a way that doesn't
trash the page cache.Quoted 5 lines
With the hashing, I seek to the page matching the hash, and I either
immediately find what I need, or I immediately know to start a readdir.
There is no need for any additional heuristic.
Quoted 13 lines
The scenario where we want to find a cookie while not doing a linear pass
through the directory will be theseekdir()case. In a linear walk, we have
the cached page index to help. So in the seekdir case, the chances of
having someone already fill a page and also having the cookie be the 1 in
127 that are page-aligned (and so match an already cached page) are small, I
think. Unless your use-case will often hit the exact same offsets over and
over.
For the use case where we are reexporting NFS, it can definitely happen.
Firstly, the clients usually are reading the reexported directory linearly, so they will tend to follow the same cookie request patterns. Secondly, we're not going to replay the readdir from the duplicate reply cache if the client resends the request. So even if you only have one client, there can be a benefit.
Quoted 21 lines
So with the hashing and seekdir case, I think that the cache will be pretty
heavily filled with the same duplicated data at various offsets and rarely
useful. That's why I wondered if you'd tested your use-case for it and found
it to be advantageous. I think what we've got is going to work fine, but I
wonder if you've seen it to work well.The major pain point most of our users complain about is not being able to
perform a complete walk in linear time with respect to size with invalidations at play. This series fixes that, and is a huge bonus. Other
smaller performance improvements are pale in comparison for us, and might
just get us forever chasing one or two minor optimizations that have trade-offs.
Quoted 8 lines
There's a lot of variables at play. For some client/server setups (like
some low-latency RDMA), and very large directories and cache sizes, it might
be more performant to just do the READDIR every time, walking local caches
be damned.
Sure, so a dedicated readdirplus() system call could help by providing the same kind of guidance that statx() does today.
Quoted 62 lines
Quoted 54 lines
Quoted 26 lines
Its weird to me that we're doing exactly what XArray says not to do,
hash
the index, when we don't have to.Quoted 12 lines
Quoted 11 lines
It also means we're amplifying the pagecache's useage for slightly
changing
directories - rather than re-using the same pages we're scattering
our usage
across the index. Eh, maybe not a big deal if we just expect the
page
cache's LRU to do the work.Quoted 8 lines
I don't understand your point about 'not reusing'. If the user seeks to
the same cookie, we reuse the page. However I don't know how you
would
go about setting up a schema that allows you to seek to an arbitrary
cookie without doing a linear search.Quoted 17 lines
So when I was taking about 'reusing' a page, that's about re- filling
the
same pages rather than constantly conjuring new ones, which requires
less of
the pagecache's resources in total. Maybe the pagecache can handle
that
without it negatively impacting other users of the cache that /will/
re-use
their cached pages, but I worry it might be irresponsible of us to
fill the
pagecache with pages we know we're never going to find again.In the case where the processes are reading linearly through a directory that is not changing (or at least where the beginning of the
directory is not changing), we will reuse the cached data, because just
like in the linearly indexed case, each process ends up reading the exact same sequence of cookies, and looking up the exact same sequence
of hashes.Quoted 7 lines
The sequences start to diverge only if they hit a part of the directory
that is being modified. At that point, we're going to be invalidating
page cache entries anyway with the last reader being more likely to be
following the new sequence of cookies.
Quoted 8 lines
I don't think we clean up behind ourselves anymore. Now that we are going
to validate each page before using it, we don't invalidate the whole cache
at any point. That means that a divergence duplicates the pagecache usage
beyond the divergence.
No. I reinstated the call to nfs_revalidate_mapping() in the linux-next branch after Olga demonstrated that NFSv3 is still troubled with crappy mtime/ctime resolutions on the server causing directory changes to not be reflected in the readdir cache.
Signature · 3 lines
Trond Myklebust Linux NFS client maintainer, Hammerspace [email protected]
Re: [PATCH] NFS: Trigger "ls -l" readdir heuristic sooner
In reply to Benjamin CoddingtonQuoted 14 lines · Benjamin Coddington
.. Something like this does the trick in my testing, but yes will have an impact on regular workloads:
8<------------------------------------------------------------------------
Since commit
1a34c8c9a49e("NFS: Support larger readdir buffers") has updated dtsize and recent improvements to the READDIRPLUS helper heuristic, the heuristic may not trigger until many dentries are emitted to userspace, which may cause many thousands of GETATTR calls for "ls -l" when the directory's pagecache has already been populated. This typically manifests as a much slower total runtime for a _second_ invocation of "ls -l" within the directory attribute cache timeouts.
Quoted 3 lines
Fix this by emitting only 17 entries for any first pass through the NFS directory's ->iterate_shared(), which will allow userpace to prime the counters for the heuristic.
Here's for what it's worth. An experiment between linux to linux where the linux server had a "small" directory structure of 57274 directories, 5727390 files in total where each directory had ~100 files each.
With this patch:
date; time tree vol1 > tree.out && date; time tree vol1 > tree.out Wed Mar 16 12:21:30 EDT 2022
real 11m7.923s user 0m20.507s sys 0m39.683s
Wed Mar 16 12:32:38 EDT 2022
real 40m1.751s user 0m23.477s sys 0m45.663s
Without the patch:
date; time tree vol1 > tree.out && date; time tree vol1 > tree.out Wed Mar 16 13:49:12 EDT 2022
real 10m52.909s user 0m21.342s sys 0m39.198s
Wed Mar 16 14:00:05 EDT 2022
real 222m56.990s user 0m30.392s sys 2m25.202s
Quoted 50 lines
Signed-off-by Benjamin Coddington <[email protected]>--- fs/nfs/dir.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)fs/nfs/dir.c +7
-3@@ -1060,6 +1060,8 @@ static int readdir_search_pagecache(struct nfs_readdir_descriptor *desc)return res; } +#define NFS_READDIR_CACHE_MISS_THRESHOLD (16UL) +/* * Once we've found the start of the dirent within a page: fill 'er up... */ @@ -1069,6 +1071,7 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc,struct file *file = desc->file;struct nfs_cache_array *array;unsigned int i; +bool first_emit = !desc->dir_cookie; array =kmap (desc->page);for (i = desc->cache_entry_index; i < array->size; i++) {@@ -1092,6 +1095,10 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc, desc->ctx->pos = desc->dir_cookie; else desc->ctx->pos++; +if (first_emit && i >NFS_READDIR_CACHE_MISS_THRESHOLD +1 ) { + desc->eob = true; +break ; + } }if (array->page_is_eof) desc->eof = !desc->eob;@@ -1173,8 +1180,6 @@ static int uncached_readdir(struct nfs_readdir_descriptor *desc)return status; } -#define NFS_READDIR_CACHE_MISS_THRESHOLD (16UL) - static bool nfs_readdir_handle_cache_misses(struct inode *inode, struct nfs_readdir_descriptor *desc, unsigned int cache_misses, --2.31.1