From ce60ab3964782df9ba34f0a64c0bc766dd508bde Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 29 May 2025 06:45:45 -0400 Subject: Expand the type of nfs_fattr->valid We need to be able to track more than 32 attributes per inode. Signed-off-by: Trond Myklebust Signed-off-by: Lance Shelton Signed-off-by: Benjamin Coddington Reviewed-by: Jeff Layton Link: https://lore.kernel.org/r/1e3405fca54efd0be7c91c1da77917b94f5dfcc4.1748515333.git.bcodding@redhat.com Signed-off-by: Trond Myklebust --- include/linux/nfs_fs_sb.h | 2 +- include/linux/nfs_xdr.h | 54 +++++++++++++++++++++++------------------------ 2 files changed, 28 insertions(+), 28 deletions(-) (limited to 'include') diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 63141320c2a8..d7895eeccea3 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -172,8 +172,8 @@ struct nfs_server { #define NFS_MOUNT_FORCE_RDIRPLUS 0x20000000 #define NFS_MOUNT_NETUNREACH_FATAL 0x40000000 - unsigned int fattr_valid; /* Valid attributes */ unsigned int caps; /* server capabilities */ + __u64 fattr_valid; /* Valid attributes */ unsigned int rsize; /* read size */ unsigned int rpages; /* read size (in pages) */ unsigned int wsize; /* write size */ diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 67f6632f723b..9cacbbd14787 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -45,7 +45,7 @@ struct nfs4_threshold { }; struct nfs_fattr { - unsigned int valid; /* which fields are valid */ + __u64 valid; /* which fields are valid */ umode_t mode; __u32 nlink; kuid_t uid; @@ -80,32 +80,32 @@ struct nfs_fattr { struct nfs4_label *label; }; -#define NFS_ATTR_FATTR_TYPE (1U << 0) -#define NFS_ATTR_FATTR_MODE (1U << 1) -#define NFS_ATTR_FATTR_NLINK (1U << 2) -#define NFS_ATTR_FATTR_OWNER (1U << 3) -#define NFS_ATTR_FATTR_GROUP (1U << 4) -#define NFS_ATTR_FATTR_RDEV (1U << 5) -#define NFS_ATTR_FATTR_SIZE (1U << 6) -#define NFS_ATTR_FATTR_PRESIZE (1U << 7) -#define NFS_ATTR_FATTR_BLOCKS_USED (1U << 8) -#define NFS_ATTR_FATTR_SPACE_USED (1U << 9) -#define NFS_ATTR_FATTR_FSID (1U << 10) -#define NFS_ATTR_FATTR_FILEID (1U << 11) -#define NFS_ATTR_FATTR_ATIME (1U << 12) -#define NFS_ATTR_FATTR_MTIME (1U << 13) -#define NFS_ATTR_FATTR_CTIME (1U << 14) -#define NFS_ATTR_FATTR_PREMTIME (1U << 15) -#define NFS_ATTR_FATTR_PRECTIME (1U << 16) -#define NFS_ATTR_FATTR_CHANGE (1U << 17) -#define NFS_ATTR_FATTR_PRECHANGE (1U << 18) -#define NFS_ATTR_FATTR_V4_LOCATIONS (1U << 19) -#define NFS_ATTR_FATTR_V4_REFERRAL (1U << 20) -#define NFS_ATTR_FATTR_MOUNTPOINT (1U << 21) -#define NFS_ATTR_FATTR_MOUNTED_ON_FILEID (1U << 22) -#define NFS_ATTR_FATTR_OWNER_NAME (1U << 23) -#define NFS_ATTR_FATTR_GROUP_NAME (1U << 24) -#define NFS_ATTR_FATTR_V4_SECURITY_LABEL (1U << 25) +#define NFS_ATTR_FATTR_TYPE BIT_ULL(0) +#define NFS_ATTR_FATTR_MODE BIT_ULL(1) +#define NFS_ATTR_FATTR_NLINK BIT_ULL(2) +#define NFS_ATTR_FATTR_OWNER BIT_ULL(3) +#define NFS_ATTR_FATTR_GROUP BIT_ULL(4) +#define NFS_ATTR_FATTR_RDEV BIT_ULL(5) +#define NFS_ATTR_FATTR_SIZE BIT_ULL(6) +#define NFS_ATTR_FATTR_PRESIZE BIT_ULL(7) +#define NFS_ATTR_FATTR_BLOCKS_USED BIT_ULL(8) +#define NFS_ATTR_FATTR_SPACE_USED BIT_ULL(9) +#define NFS_ATTR_FATTR_FSID BIT_ULL(10) +#define NFS_ATTR_FATTR_FILEID BIT_ULL(11) +#define NFS_ATTR_FATTR_ATIME BIT_ULL(12) +#define NFS_ATTR_FATTR_MTIME BIT_ULL(13) +#define NFS_ATTR_FATTR_CTIME BIT_ULL(14) +#define NFS_ATTR_FATTR_PREMTIME BIT_ULL(15) +#define NFS_ATTR_FATTR_PRECTIME BIT_ULL(16) +#define NFS_ATTR_FATTR_CHANGE BIT_ULL(17) +#define NFS_ATTR_FATTR_PRECHANGE BIT_ULL(18) +#define NFS_ATTR_FATTR_V4_LOCATIONS BIT_ULL(19) +#define NFS_ATTR_FATTR_V4_REFERRAL BIT_ULL(20) +#define NFS_ATTR_FATTR_MOUNTPOINT BIT_ULL(21) +#define NFS_ATTR_FATTR_MOUNTED_ON_FILEID BIT_ULL(22) +#define NFS_ATTR_FATTR_OWNER_NAME BIT_ULL(23) +#define NFS_ATTR_FATTR_GROUP_NAME BIT_ULL(24) +#define NFS_ATTR_FATTR_V4_SECURITY_LABEL BIT_ULL(25) #define NFS_ATTR_FATTR (NFS_ATTR_FATTR_TYPE \ | NFS_ATTR_FATTR_MODE \ -- cgit v1.2.3 From 1c7ae2dd3f0e6d07ec0a5a348f2561f2171b9c81 Mon Sep 17 00:00:00 2001 From: Anne Marie Merritt Date: Thu, 29 May 2025 06:45:46 -0400 Subject: nfs: Add timecreate to nfs inode Add tracking of the create time (a.k.a. btime) along with corresponding bitfields, request, and decode xdr routines. Signed-off-by: Anne Marie Merritt Signed-off-by: Lance Shelton Signed-off-by: Benjamin Coddington Reviewed-by: Jeff Layton Link: https://lore.kernel.org/r/1e3677b0655fa2bbaba0817b41d111d94a06e5ee.1748515333.git.bcodding@redhat.com Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 17 +++++++++++++++-- fs/nfs/nfs4proc.c | 14 +++++++++++++- fs/nfs/nfs4xdr.c | 24 ++++++++++++++++++++++++ fs/nfs/nfstrace.h | 3 ++- include/linux/nfs_fs.h | 8 ++++++++ include/linux/nfs_xdr.h | 3 +++ 6 files changed, 65 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 17f5dcda2a00..c5462aed6bf5 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -197,6 +197,7 @@ void nfs_set_cache_invalid(struct inode *inode, unsigned long flags) if (!(flags & NFS_INO_REVAL_FORCED)) flags &= ~(NFS_INO_INVALID_MODE | NFS_INO_INVALID_OTHER | + NFS_INO_INVALID_BTIME | NFS_INO_INVALID_XATTR); flags &= ~(NFS_INO_INVALID_CHANGE | NFS_INO_INVALID_SIZE); } @@ -522,6 +523,7 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr) inode_set_atime(inode, 0, 0); inode_set_mtime(inode, 0, 0); inode_set_ctime(inode, 0, 0); + memset(&nfsi->btime, 0, sizeof(nfsi->btime)); inode_set_iversion_raw(inode, 0); inode->i_size = 0; clear_nlink(inode); @@ -545,6 +547,10 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr) inode_set_ctime_to_ts(inode, fattr->ctime); else if (fattr_supported & NFS_ATTR_FATTR_CTIME) nfs_set_cache_invalid(inode, NFS_INO_INVALID_CTIME); + if (fattr->valid & NFS_ATTR_FATTR_BTIME) + nfsi->btime = fattr->btime; + else if (fattr_supported & NFS_ATTR_FATTR_BTIME) + nfs_set_cache_invalid(inode, NFS_INO_INVALID_BTIME); if (fattr->valid & NFS_ATTR_FATTR_CHANGE) inode_set_iversion_raw(inode, fattr->change_attr); else @@ -1943,7 +1949,7 @@ static int nfs_inode_finish_partial_attr_update(const struct nfs_fattr *fattr, NFS_INO_INVALID_ATIME | NFS_INO_INVALID_CTIME | NFS_INO_INVALID_MTIME | NFS_INO_INVALID_SIZE | NFS_INO_INVALID_BLOCKS | NFS_INO_INVALID_OTHER | - NFS_INO_INVALID_NLINK; + NFS_INO_INVALID_NLINK | NFS_INO_INVALID_BTIME; unsigned long cache_validity = NFS_I(inode)->cache_validity; enum nfs4_change_attr_type ctype = NFS_SERVER(inode)->change_attr_type; @@ -2304,7 +2310,8 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) | NFS_INO_INVALID_BLOCKS | NFS_INO_INVALID_NLINK | NFS_INO_INVALID_MODE - | NFS_INO_INVALID_OTHER; + | NFS_INO_INVALID_OTHER + | NFS_INO_INVALID_BTIME; if (S_ISDIR(inode->i_mode)) nfs_force_lookup_revalidate(inode); attr_changed = true; @@ -2338,6 +2345,12 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) nfsi->cache_validity |= save_cache_validity & NFS_INO_INVALID_CTIME; + if (fattr->valid & NFS_ATTR_FATTR_BTIME) + nfsi->btime = fattr->btime; + else if (fattr_supported & NFS_ATTR_FATTR_BTIME) + nfsi->cache_validity |= + save_cache_validity & NFS_INO_INVALID_BTIME; + /* Check if our cached file size is stale */ if (fattr->valid & NFS_ATTR_FATTR_SIZE) { new_isize = nfs_size_to_loff_t(fattr->size); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 341740fa293d..92f1b2601b67 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -222,6 +222,7 @@ const u32 nfs4_fattr_bitmap[3] = { | FATTR4_WORD1_RAWDEV | FATTR4_WORD1_SPACE_USED | FATTR4_WORD1_TIME_ACCESS + | FATTR4_WORD1_TIME_CREATE | FATTR4_WORD1_TIME_METADATA | FATTR4_WORD1_TIME_MODIFY | FATTR4_WORD1_MOUNTED_ON_FILEID, @@ -243,6 +244,7 @@ static const u32 nfs4_pnfs_open_bitmap[3] = { | FATTR4_WORD1_RAWDEV | FATTR4_WORD1_SPACE_USED | FATTR4_WORD1_TIME_ACCESS + | FATTR4_WORD1_TIME_CREATE | FATTR4_WORD1_TIME_METADATA | FATTR4_WORD1_TIME_MODIFY, FATTR4_WORD2_MDSTHRESHOLD @@ -323,6 +325,9 @@ static void nfs4_bitmap_copy_adjust(__u32 *dst, const __u32 *src, if (!(cache_validity & NFS_INO_INVALID_OTHER)) dst[1] &= ~(FATTR4_WORD1_OWNER | FATTR4_WORD1_OWNER_GROUP); + if (!(cache_validity & NFS_INO_INVALID_BTIME)) + dst[1] &= ~FATTR4_WORD1_TIME_CREATE; + if (nfs_have_delegated_mtime(inode)) { if (!(cache_validity & NFS_INO_INVALID_ATIME)) dst[1] &= ~(FATTR4_WORD1_TIME_ACCESS|FATTR4_WORD1_TIME_ACCESS_SET); @@ -1307,7 +1312,8 @@ nfs4_update_changeattr_locked(struct inode *inode, NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL | NFS_INO_INVALID_SIZE | NFS_INO_INVALID_OTHER | NFS_INO_INVALID_BLOCKS | NFS_INO_INVALID_NLINK | - NFS_INO_INVALID_MODE | NFS_INO_INVALID_XATTR; + NFS_INO_INVALID_MODE | NFS_INO_INVALID_BTIME | + NFS_INO_INVALID_XATTR; nfsi->attrtimeo = NFS_MINATTRTIMEO(inode); } nfsi->attrtimeo_timestamp = jiffies; @@ -4047,6 +4053,10 @@ static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *f server->fattr_valid &= ~NFS_ATTR_FATTR_CTIME; if (!(res.attr_bitmask[1] & FATTR4_WORD1_TIME_MODIFY)) server->fattr_valid &= ~NFS_ATTR_FATTR_MTIME; + if (!(res.attr_bitmask[1] & FATTR4_WORD1_TIME_MODIFY)) + server->fattr_valid &= ~NFS_ATTR_FATTR_MTIME; + if (!(res.attr_bitmask[1] & FATTR4_WORD1_TIME_CREATE)) + server->fattr_valid &= ~NFS_ATTR_FATTR_BTIME; memcpy(server->attr_bitmask_nl, res.attr_bitmask, sizeof(server->attr_bitmask)); server->attr_bitmask_nl[2] &= ~FATTR4_WORD2_SECURITY_LABEL; @@ -5781,6 +5791,8 @@ void nfs4_bitmask_set(__u32 bitmask[], const __u32 src[], bitmask[1] |= FATTR4_WORD1_TIME_MODIFY; if (cache_validity & NFS_INO_INVALID_BLOCKS) bitmask[1] |= FATTR4_WORD1_SPACE_USED; + if (cache_validity & NFS_INO_INVALID_BTIME) + bitmask[1] |= FATTR4_WORD1_TIME_CREATE; if (cache_validity & NFS_INO_INVALID_SIZE) bitmask[0] |= FATTR4_WORD0_SIZE; diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 318afde38057..49ff98571fa5 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1623,6 +1623,7 @@ static void encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg | FATTR4_WORD1_RAWDEV | FATTR4_WORD1_SPACE_USED | FATTR4_WORD1_TIME_ACCESS + | FATTR4_WORD1_TIME_CREATE | FATTR4_WORD1_TIME_METADATA | FATTR4_WORD1_TIME_MODIFY; attrs[2] |= FATTR4_WORD2_SECURITY_LABEL; @@ -4207,6 +4208,24 @@ static int decode_attr_time_access(struct xdr_stream *xdr, uint32_t *bitmap, str return status; } +static int decode_attr_time_create(struct xdr_stream *xdr, uint32_t *bitmap, struct timespec64 *time) +{ + int status = 0; + + time->tv_sec = 0; + time->tv_nsec = 0; + if (unlikely(bitmap[1] & (FATTR4_WORD1_TIME_CREATE - 1U))) + return -EIO; + if (likely(bitmap[1] & FATTR4_WORD1_TIME_CREATE)) { + status = decode_attr_time(xdr, time); + if (status == 0) + status = NFS_ATTR_FATTR_BTIME; + bitmap[1] &= ~FATTR4_WORD1_TIME_CREATE; + } + dprintk("%s: btime=%lld\n", __func__, time->tv_sec); + return status; +} + static int decode_attr_time_metadata(struct xdr_stream *xdr, uint32_t *bitmap, struct timespec64 *time) { int status = 0; @@ -4781,6 +4800,11 @@ static int decode_getfattr_attrs(struct xdr_stream *xdr, uint32_t *bitmap, goto xdr_error; fattr->valid |= status; + status = decode_attr_time_create(xdr, bitmap, &fattr->btime); + if (status < 0) + goto xdr_error; + fattr->valid |= status; + status = decode_attr_time_metadata(xdr, bitmap, &fattr->ctime); if (status < 0) goto xdr_error; diff --git a/fs/nfs/nfstrace.h b/fs/nfs/nfstrace.h index 7a058bd8c566..f49f064c5ee5 100644 --- a/fs/nfs/nfstrace.h +++ b/fs/nfs/nfstrace.h @@ -32,7 +32,8 @@ { NFS_INO_INVALID_BLOCKS, "INVALID_BLOCKS" }, \ { NFS_INO_INVALID_XATTR, "INVALID_XATTR" }, \ { NFS_INO_INVALID_NLINK, "INVALID_NLINK" }, \ - { NFS_INO_INVALID_MODE, "INVALID_MODE" }) + { NFS_INO_INVALID_MODE, "INVALID_MODE" }, \ + { NFS_INO_INVALID_BTIME, "INVALID_BTIME" }) #define nfs_show_nfsi_flags(v) \ __print_flags(v, "|", \ diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 67ae2c3f41d2..c585939b6cd6 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -160,6 +160,12 @@ struct nfs_inode { unsigned long flags; /* atomic bit ops */ unsigned long cache_validity; /* bit mask */ + /* + * NFS Attributes not included in struct inode + */ + + struct timespec64 btime; + /* * read_cache_jiffies is when we started read-caching this inode. * attrtimeo is for how long the cached information is assumed @@ -316,10 +322,12 @@ struct nfs4_copy_state { #define NFS_INO_INVALID_XATTR BIT(15) /* xattrs are invalid */ #define NFS_INO_INVALID_NLINK BIT(16) /* cached nlinks is invalid */ #define NFS_INO_INVALID_MODE BIT(17) /* cached mode is invalid */ +#define NFS_INO_INVALID_BTIME BIT(18) /* cached btime is invalid */ #define NFS_INO_INVALID_ATTR (NFS_INO_INVALID_CHANGE \ | NFS_INO_INVALID_CTIME \ | NFS_INO_INVALID_MTIME \ + | NFS_INO_INVALID_BTIME \ | NFS_INO_INVALID_SIZE \ | NFS_INO_INVALID_NLINK \ | NFS_INO_INVALID_MODE \ diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 9cacbbd14787..ac4bff6e9913 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -67,6 +67,7 @@ struct nfs_fattr { struct timespec64 atime; struct timespec64 mtime; struct timespec64 ctime; + struct timespec64 btime; __u64 change_attr; /* NFSv4 change attribute */ __u64 pre_change_attr;/* pre-op NFSv4 change attribute */ __u64 pre_size; /* pre_op_attr.size */ @@ -106,6 +107,7 @@ struct nfs_fattr { #define NFS_ATTR_FATTR_OWNER_NAME BIT_ULL(23) #define NFS_ATTR_FATTR_GROUP_NAME BIT_ULL(24) #define NFS_ATTR_FATTR_V4_SECURITY_LABEL BIT_ULL(25) +#define NFS_ATTR_FATTR_BTIME BIT_ULL(26) #define NFS_ATTR_FATTR (NFS_ATTR_FATTR_TYPE \ | NFS_ATTR_FATTR_MODE \ @@ -126,6 +128,7 @@ struct nfs_fattr { | NFS_ATTR_FATTR_SPACE_USED) #define NFS_ATTR_FATTR_V4 (NFS_ATTR_FATTR \ | NFS_ATTR_FATTR_SPACE_USED \ + | NFS_ATTR_FATTR_BTIME \ | NFS_ATTR_FATTR_V4_SECURITY_LABEL) /* -- cgit v1.2.3 From 0715a72ee9a38461eac4b34388b772914f269119 Mon Sep 17 00:00:00 2001 From: Anthony Iliopoulos Date: Fri, 13 Jun 2025 11:44:37 +0200 Subject: NFS: remove unused wpages field from struct nfs_server The wpages field is not serving any purpose since commit c63c7b051395 ("NFS: Fix a race when doing NFS write coalescing") which was merged in v2.6.22-rc1. Remove it completely. Signed-off-by: Anthony Iliopoulos Link: https://lore.kernel.org/r/20250613094439.82338-2-ailiop@suse.com Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 1 - include/linux/nfs_fs_sb.h | 1 - 2 files changed, 2 deletions(-) (limited to 'include') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index cf35ad3f818a..23dafc590476 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -814,7 +814,6 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, server->wsize = max_rpc_payload; if (server->wsize > NFS_MAX_FILE_IO_SIZE) server->wsize = NFS_MAX_FILE_IO_SIZE; - server->wpages = (server->wsize + PAGE_SIZE - 1) >> PAGE_SHIFT; server->wtmult = nfs_block_bits(fsinfo->wtmult, NULL); diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index d7895eeccea3..7048f9b867ab 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -177,7 +177,6 @@ struct nfs_server { unsigned int rsize; /* read size */ unsigned int rpages; /* read size (in pages) */ unsigned int wsize; /* write size */ - unsigned int wpages; /* write size (in pages) */ unsigned int wtmult; /* server disk block size */ unsigned int dtsize; /* readdir size */ unsigned short port; /* "port=" setting */ -- cgit v1.2.3 From 74a33326cfe8e62ebe0a65ba01ea8a8bceb532f8 Mon Sep 17 00:00:00 2001 From: Anthony Iliopoulos Date: Fri, 13 Jun 2025 11:44:38 +0200 Subject: NFS: remove unused time_delta field from struct nfs_server The last code that was using this was removed via commit ca0daa277aca ("NFS: Cache aggressively when file is open for writing") which was merged in v4.8-rc1, so it can be removed completely. Signed-off-by: Anthony Iliopoulos Link: https://lore.kernel.org/r/20250613094439.82338-3-ailiop@suse.com Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 1 - include/linux/nfs_fs_sb.h | 1 - 2 files changed, 2 deletions(-) (limited to 'include') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 23dafc590476..47258dc3af70 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -830,7 +830,6 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, server->maxfilesize = fsinfo->maxfilesize; - server->time_delta = fsinfo->time_delta; server->change_attr_type = fsinfo->change_attr_type; server->clone_blksize = fsinfo->clone_blksize; diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 7048f9b867ab..e1b2cf57e765 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -202,7 +202,6 @@ struct nfs_server { struct nfs_fsid fsid; int s_sysfs_id; /* sysfs dentry index */ __u64 maxfilesize; /* maximum file size */ - struct timespec64 time_delta; /* smallest time granularity */ unsigned long mount_time; /* when this fs was mounted */ struct super_block *super; /* VFS super block */ dev_t s_dev; /* superblock dev numbers */ -- cgit v1.2.3 From 2c665d91c2a2d8b5bdf1374d1253b3c89fca4ede Mon Sep 17 00:00:00 2001 From: Anthony Iliopoulos Date: Fri, 13 Jun 2025 11:44:39 +0200 Subject: NFS: remove unused pnfs_ld_data field from struct nfs_server The last code that was using this was removed via commit 20d655d6197d ("pnfs/blocklayout: use the device id cache") which was merged in v3.18-rc1, so it can be removed completely. Signed-off-by: Anthony Iliopoulos Link: https://lore.kernel.org/r/20250613094439.82338-4-ailiop@suse.com Signed-off-by: Trond Myklebust --- include/linux/nfs_fs_sb.h | 1 - 1 file changed, 1 deletion(-) (limited to 'include') diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index e1b2cf57e765..d2d36711a119 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -246,7 +246,6 @@ struct nfs_server { filesystem */ struct pnfs_layoutdriver_type *pnfs_curr_ld; /* Active layout driver */ struct rpc_wait_queue roc_rpcwaitq; - void *pnfs_ld_data; /* per mount point data */ /* the following fields are protected by nfs_client->cl_lock */ struct rb_root state_owners; -- cgit v1.2.3 From 48693d119b2114f8eaf8b8f972b29e05ae581ad4 Mon Sep 17 00:00:00 2001 From: "Dr. David Alan Gilbert" Date: Sun, 13 Jul 2025 00:30:06 +0100 Subject: SUNRPC: Remove unused xdr functions Remove a bunch of unused xdr_*decode* functions: The last use of xdr_decode_netobj() was removed in 2021 by: commit 7cf96b6d0104 ("lockd: Update the NLMv4 SHARE arguments decoder to use struct xdr_stream") The last use of xdr_decode_string_inplace() was removed in 2021 by: commit 3049e974a7c7 ("lockd: Update the NLMv4 FREE_ALL arguments decoder to use struct xdr_stream") The last use of xdr_stream_decode_opaque() was removed in 2024 by: commit fed8a17c61ff ("xdrgen: typedefs should use the built-in string and opaque functions") The functions xdr_stream_decode_string() and xdr_stream_decode_opaque_dup() were both added in 2018 by the commit 0e779aa70308 ("SUNRPC: Add helpers for decoding opaque and string types") but never used. Remove them. Signed-off-by: Dr. David Alan Gilbert Link: https://lore.kernel.org/r/20250712233006.403226-1-linux@treblig.org Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xdr.h | 9 ---- net/sunrpc/xdr.c | 110 --------------------------------------------- 2 files changed, 119 deletions(-) (limited to 'include') diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h index a2ab813a9800..e370886632b0 100644 --- a/include/linux/sunrpc/xdr.h +++ b/include/linux/sunrpc/xdr.h @@ -128,10 +128,7 @@ xdr_buf_init(struct xdr_buf *buf, void *start, size_t len) __be32 *xdr_encode_opaque_fixed(__be32 *p, const void *ptr, unsigned int len); __be32 *xdr_encode_opaque(__be32 *p, const void *ptr, unsigned int len); __be32 *xdr_encode_string(__be32 *p, const char *s); -__be32 *xdr_decode_string_inplace(__be32 *p, char **sp, unsigned int *lenp, - unsigned int maxlen); __be32 *xdr_encode_netobj(__be32 *p, const struct xdr_netobj *); -__be32 *xdr_decode_netobj(__be32 *p, struct xdr_netobj *); void xdr_inline_pages(struct xdr_buf *, unsigned int, struct page **, unsigned int, unsigned int); @@ -341,12 +338,6 @@ xdr_stream_remaining(const struct xdr_stream *xdr) return xdr->nwords << 2; } -ssize_t xdr_stream_decode_opaque(struct xdr_stream *xdr, void *ptr, - size_t size); -ssize_t xdr_stream_decode_opaque_dup(struct xdr_stream *xdr, void **ptr, - size_t maxlen, gfp_t gfp_flags); -ssize_t xdr_stream_decode_string(struct xdr_stream *xdr, char *str, - size_t size); ssize_t xdr_stream_decode_string_dup(struct xdr_stream *xdr, char **str, size_t maxlen, gfp_t gfp_flags); ssize_t xdr_stream_decode_opaque_auth(struct xdr_stream *xdr, u32 *flavor, diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c index 2ea00e354ba6..a0aae1144212 100644 --- a/net/sunrpc/xdr.c +++ b/net/sunrpc/xdr.c @@ -37,19 +37,6 @@ xdr_encode_netobj(__be32 *p, const struct xdr_netobj *obj) } EXPORT_SYMBOL_GPL(xdr_encode_netobj); -__be32 * -xdr_decode_netobj(__be32 *p, struct xdr_netobj *obj) -{ - unsigned int len; - - if ((len = be32_to_cpu(*p++)) > XDR_MAX_NETOBJ) - return NULL; - obj->len = len; - obj->data = (u8 *) p; - return p + XDR_QUADLEN(len); -} -EXPORT_SYMBOL_GPL(xdr_decode_netobj); - /** * xdr_encode_opaque_fixed - Encode fixed length opaque data * @p: pointer to current position in XDR buffer. @@ -102,21 +89,6 @@ xdr_encode_string(__be32 *p, const char *string) } EXPORT_SYMBOL_GPL(xdr_encode_string); -__be32 * -xdr_decode_string_inplace(__be32 *p, char **sp, - unsigned int *lenp, unsigned int maxlen) -{ - u32 len; - - len = be32_to_cpu(*p++); - if (len > maxlen) - return NULL; - *lenp = len; - *sp = (char *) p; - return p + XDR_QUADLEN(len); -} -EXPORT_SYMBOL_GPL(xdr_decode_string_inplace); - /** * xdr_terminate_string - '\0'-terminate a string residing in an xdr_buf * @buf: XDR buffer where string resides @@ -2247,88 +2219,6 @@ out: } EXPORT_SYMBOL_GPL(xdr_process_buf); -/** - * xdr_stream_decode_opaque - Decode variable length opaque - * @xdr: pointer to xdr_stream - * @ptr: location to store opaque data - * @size: size of storage buffer @ptr - * - * Return values: - * On success, returns size of object stored in *@ptr - * %-EBADMSG on XDR buffer overflow - * %-EMSGSIZE on overflow of storage buffer @ptr - */ -ssize_t xdr_stream_decode_opaque(struct xdr_stream *xdr, void *ptr, size_t size) -{ - ssize_t ret; - void *p; - - ret = xdr_stream_decode_opaque_inline(xdr, &p, size); - if (ret <= 0) - return ret; - memcpy(ptr, p, ret); - return ret; -} -EXPORT_SYMBOL_GPL(xdr_stream_decode_opaque); - -/** - * xdr_stream_decode_opaque_dup - Decode and duplicate variable length opaque - * @xdr: pointer to xdr_stream - * @ptr: location to store pointer to opaque data - * @maxlen: maximum acceptable object size - * @gfp_flags: GFP mask to use - * - * Return values: - * On success, returns size of object stored in *@ptr - * %-EBADMSG on XDR buffer overflow - * %-EMSGSIZE if the size of the object would exceed @maxlen - * %-ENOMEM on memory allocation failure - */ -ssize_t xdr_stream_decode_opaque_dup(struct xdr_stream *xdr, void **ptr, - size_t maxlen, gfp_t gfp_flags) -{ - ssize_t ret; - void *p; - - ret = xdr_stream_decode_opaque_inline(xdr, &p, maxlen); - if (ret > 0) { - *ptr = kmemdup(p, ret, gfp_flags); - if (*ptr != NULL) - return ret; - ret = -ENOMEM; - } - *ptr = NULL; - return ret; -} -EXPORT_SYMBOL_GPL(xdr_stream_decode_opaque_dup); - -/** - * xdr_stream_decode_string - Decode variable length string - * @xdr: pointer to xdr_stream - * @str: location to store string - * @size: size of storage buffer @str - * - * Return values: - * On success, returns length of NUL-terminated string stored in *@str - * %-EBADMSG on XDR buffer overflow - * %-EMSGSIZE on overflow of storage buffer @str - */ -ssize_t xdr_stream_decode_string(struct xdr_stream *xdr, char *str, size_t size) -{ - ssize_t ret; - void *p; - - ret = xdr_stream_decode_opaque_inline(xdr, &p, size); - if (ret > 0) { - memcpy(str, p, ret); - str[ret] = '\0'; - return strlen(str); - } - *str = '\0'; - return ret; -} -EXPORT_SYMBOL_GPL(xdr_stream_decode_string); - /** * xdr_stream_decode_string_dup - Decode and duplicate variable length string * @xdr: pointer to xdr_stream -- cgit v1.2.3 From 2fb4af5ea3c735a205d97de10f044f809b20af51 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 18 Jul 2025 10:14:49 +0200 Subject: NFS: track active delegations per-server The active delegation watermark was added to avoid overloading servers. Track the active delegation per-server instead of globally so that clients talking to multiple servers aren't limited by the global limit. Signed-off-by: Christoph Hellwig Reviewed-by: Jeff Layton Link: https://lore.kernel.org/r/20250718081509.2607553-5-hch@lst.de Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 1 + fs/nfs/delegation.c | 35 +++++++++++++++++++---------------- include/linux/nfs_fs_sb.h | 1 + 3 files changed, 21 insertions(+), 16 deletions(-) (limited to 'include') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 47258dc3af70..e13eb429b8b5 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -1005,6 +1005,7 @@ struct nfs_server *nfs_alloc_server(void) INIT_LIST_HEAD(&server->ss_src_copies); atomic_set(&server->active, 0); + atomic_long_set(&server->nr_active_delegations, 0); server->io_stats = nfs_alloc_iostats(); if (!server->io_stats) { diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index 5f85966d7709..ea96f77e38c2 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -27,7 +27,6 @@ #define NFS_DEFAULT_DELEGATION_WATERMARK (5000U) -static atomic_long_t nfs_active_delegations; static unsigned nfs_delegation_watermark = NFS_DEFAULT_DELEGATION_WATERMARK; module_param_named(delegation_watermark, nfs_delegation_watermark, uint, 0644); @@ -38,11 +37,12 @@ static void __nfs_free_delegation(struct nfs_delegation *delegation) kfree_rcu(delegation, rcu); } -static void nfs_mark_delegation_revoked(struct nfs_delegation *delegation) +static void nfs_mark_delegation_revoked(struct nfs_server *server, + struct nfs_delegation *delegation) { if (!test_and_set_bit(NFS_DELEGATION_REVOKED, &delegation->flags)) { delegation->stateid.type = NFS4_INVALID_STATEID_TYPE; - atomic_long_dec(&nfs_active_delegations); + atomic_long_dec(&server->nr_active_delegations); if (!test_bit(NFS_DELEGATION_RETURNING, &delegation->flags)) nfs_clear_verifier_delegated(delegation->inode); } @@ -60,9 +60,10 @@ static void nfs_put_delegation(struct nfs_delegation *delegation) __nfs_free_delegation(delegation); } -static void nfs_free_delegation(struct nfs_delegation *delegation) +static void nfs_free_delegation(struct nfs_server *server, + struct nfs_delegation *delegation) { - nfs_mark_delegation_revoked(delegation); + nfs_mark_delegation_revoked(server, delegation); nfs_put_delegation(delegation); } @@ -261,7 +262,7 @@ void nfs_inode_reclaim_delegation(struct inode *inode, const struct cred *cred, } clear_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags); if (test_and_clear_bit(NFS_DELEGATION_REVOKED, &delegation->flags)) - atomic_long_inc(&nfs_active_delegations); + atomic_long_inc(&NFS_SERVER(inode)->nr_active_delegations); spin_unlock(&delegation->lock); rcu_read_unlock(); put_cred(oldcred); @@ -413,7 +414,8 @@ nfs_update_delegation_cred(struct nfs_delegation *delegation, } static void -nfs_update_inplace_delegation(struct nfs_delegation *delegation, +nfs_update_inplace_delegation(struct nfs_server *server, + struct nfs_delegation *delegation, const struct nfs_delegation *update) { if (nfs4_stateid_is_newer(&update->stateid, &delegation->stateid)) { @@ -426,7 +428,7 @@ nfs_update_inplace_delegation(struct nfs_delegation *delegation, nfs_update_delegation_cred(delegation, update->cred); /* smp_mb__before_atomic() is implicit due to xchg() */ clear_bit(NFS_DELEGATION_REVOKED, &delegation->flags); - atomic_long_inc(&nfs_active_delegations); + atomic_long_inc(&server->nr_active_delegations); } } } @@ -481,7 +483,7 @@ int nfs_inode_set_delegation(struct inode *inode, const struct cred *cred, if (nfs4_stateid_match_other(&old_delegation->stateid, &delegation->stateid)) { spin_lock(&old_delegation->lock); - nfs_update_inplace_delegation(old_delegation, + nfs_update_inplace_delegation(server, old_delegation, delegation); spin_unlock(&old_delegation->lock); goto out; @@ -530,7 +532,7 @@ add_new: rcu_assign_pointer(nfsi->delegation, delegation); delegation = NULL; - atomic_long_inc(&nfs_active_delegations); + atomic_long_inc(&server->nr_active_delegations); trace_nfs4_set_delegation(inode, type); @@ -544,7 +546,7 @@ out: __nfs_free_delegation(delegation); if (freeme != NULL) { nfs_do_return_delegation(inode, freeme, 0); - nfs_free_delegation(freeme); + nfs_free_delegation(server, freeme); } return status; } @@ -756,7 +758,7 @@ void nfs_inode_evict_delegation(struct inode *inode) set_bit(NFS_DELEGATION_RETURNING, &delegation->flags); set_bit(NFS_DELEGATION_INODE_FREEING, &delegation->flags); nfs_do_return_delegation(inode, delegation, 1); - nfs_free_delegation(delegation); + nfs_free_delegation(NFS_SERVER(inode), delegation); } } @@ -842,7 +844,8 @@ void nfs4_inode_return_delegation_on_close(struct inode *inode) if (!delegation) goto out; if (test_bit(NFS_DELEGATION_RETURN_IF_CLOSED, &delegation->flags) || - atomic_long_read(&nfs_active_delegations) >= nfs_delegation_watermark) { + atomic_long_read(&NFS_SERVER(inode)->nr_active_delegations) >= + nfs_delegation_watermark) { spin_lock(&delegation->lock); if (delegation->inode && list_empty(&NFS_I(inode)->open_files) && @@ -1018,7 +1021,7 @@ static void nfs_revoke_delegation(struct inode *inode, } spin_unlock(&delegation->lock); } - nfs_mark_delegation_revoked(delegation); + nfs_mark_delegation_revoked(NFS_SERVER(inode), delegation); ret = true; out: rcu_read_unlock(); @@ -1050,7 +1053,7 @@ void nfs_delegation_mark_returned(struct inode *inode, delegation->stateid.seqid = stateid->seqid; } - nfs_mark_delegation_revoked(delegation); + nfs_mark_delegation_revoked(NFS_SERVER(inode), delegation); clear_bit(NFS_DELEGATION_RETURNING, &delegation->flags); spin_unlock(&delegation->lock); if (nfs_detach_delegation(NFS_I(inode), delegation, NFS_SERVER(inode))) @@ -1270,7 +1273,7 @@ restart: if (delegation != NULL) { if (nfs_detach_delegation(NFS_I(inode), delegation, server) != NULL) - nfs_free_delegation(delegation); + nfs_free_delegation(server, delegation); /* Match nfs_start_delegation_return_locked */ nfs_put_delegation(delegation); } diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index d2d36711a119..a9b44f12623f 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -254,6 +254,7 @@ struct nfs_server { struct list_head state_owners_lru; struct list_head layouts; struct list_head delegations; + atomic_long_t nr_active_delegations; struct list_head ss_copies; struct list_head ss_src_copies; -- cgit v1.2.3 From f5b3108e6a14418b120a3c38ca589b8d6cf87627 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 18 Jul 2025 10:14:50 +0200 Subject: NFS: use a hash table for delegation lookup nfs_delegation_find_inode currently has to walk the entire list of delegations per inode, which can become pretty large, and can become even larger when increasing the delegation watermark. Add a hash table to speed up the delegation lookup, sized as a fraction of the delegation watermark. Signed-off-by: Christoph Hellwig Link: https://lore.kernel.org/r/20250718081509.2607553-6-hch@lst.de Signed-off-by: Trond Myklebust --- fs/nfs/delegation.c | 28 +++++++++++++++++++++++++++- fs/nfs/delegation.h | 3 +++ fs/nfs/nfs4client.c | 5 +++++ fs/nfs/nfs4proc.c | 22 +++++++++++++++++++++- include/linux/nfs_fs_sb.h | 2 ++ 5 files changed, 58 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index ea96f77e38c2..9d3a5f29f17f 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -30,6 +30,13 @@ static unsigned nfs_delegation_watermark = NFS_DEFAULT_DELEGATION_WATERMARK; module_param_named(delegation_watermark, nfs_delegation_watermark, uint, 0644); +static struct hlist_head *nfs_delegation_hash(struct nfs_server *server, + const struct nfs_fh *fhandle) +{ + return server->delegation_hash_table + + (nfs_fhandle_hash(fhandle) & server->delegation_hash_mask); +} + static void __nfs_free_delegation(struct nfs_delegation *delegation) { put_cred(delegation->cred); @@ -367,6 +374,7 @@ nfs_detach_delegation_locked(struct nfs_inode *nfsi, spin_unlock(&delegation->lock); return NULL; } + hlist_del_init_rcu(&delegation->hash); list_del_rcu(&delegation->super_list); delegation->inode = NULL; rcu_assign_pointer(nfsi->delegation, NULL); @@ -529,6 +537,8 @@ add_new: spin_unlock(&inode->i_lock); list_add_tail_rcu(&delegation->super_list, &server->delegations); + hlist_add_head_rcu(&delegation->hash, + nfs_delegation_hash(server, &NFS_I(inode)->fh)); rcu_assign_pointer(nfsi->delegation, delegation); delegation = NULL; @@ -1166,11 +1176,12 @@ static struct inode * nfs_delegation_find_inode_server(struct nfs_server *server, const struct nfs_fh *fhandle) { + struct hlist_head *head = nfs_delegation_hash(server, fhandle); struct nfs_delegation *delegation; struct super_block *freeme = NULL; struct inode *res = NULL; - list_for_each_entry_rcu(delegation, &server->delegations, super_list) { + hlist_for_each_entry_rcu(delegation, head, hash) { spin_lock(&delegation->lock); if (delegation->inode != NULL && !test_bit(NFS_DELEGATION_REVOKED, &delegation->flags) && @@ -1577,3 +1588,18 @@ out: rcu_read_unlock(); return ret; } + +int nfs4_delegation_hash_alloc(struct nfs_server *server) +{ + int delegation_buckets, i; + + delegation_buckets = roundup_pow_of_two(nfs_delegation_watermark / 16); + server->delegation_hash_mask = delegation_buckets - 1; + server->delegation_hash_table = kmalloc_array(delegation_buckets, + sizeof(*server->delegation_hash_table), GFP_KERNEL); + if (!server->delegation_hash_table) + return -ENOMEM; + for (i = 0; i < delegation_buckets; i++) + INIT_HLIST_HEAD(&server->delegation_hash_table[i]); + return 0; +} diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h index 8ff5ab9c5c25..08ec2e9c68a4 100644 --- a/fs/nfs/delegation.h +++ b/fs/nfs/delegation.h @@ -14,6 +14,7 @@ * NFSv4 delegation */ struct nfs_delegation { + struct hlist_node hash; struct list_head super_list; const struct cred *cred; struct inode *inode; @@ -123,4 +124,6 @@ static inline int nfs_have_delegated_mtime(struct inode *inode) NFS_DELEGATION_FLAG_TIME); } +int nfs4_delegation_hash_alloc(struct nfs_server *server); + #endif diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index 5943a192f36b..2ea98f1f116f 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -802,6 +802,7 @@ static void nfs4_destroy_server(struct nfs_server *server) unset_pnfs_layoutdriver(server); nfs4_purge_state_owners(server, &freeme); nfs4_free_state_owners(&freeme); + kfree(server->delegation_hash_table); } /* @@ -1096,6 +1097,10 @@ static int nfs4_server_common_setup(struct nfs_server *server, { int error; + error = nfs4_delegation_hash_alloc(server); + if (error) + return error; + /* data servers support only a subset of NFSv4.1 */ if (is_ds_only_client(server->nfs_client)) return -EPROTONOSUPPORT; diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ef2077e185b6..d8bebd757af3 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -10967,6 +10967,26 @@ static const struct inode_operations nfs4_file_inode_operations = { .listxattr = nfs4_listxattr, }; +static struct nfs_server *nfs4_clone_server(struct nfs_server *source, + struct nfs_fh *fh, struct nfs_fattr *fattr, + rpc_authflavor_t flavor) +{ + struct nfs_server *server; + int error; + + server = nfs_clone_server(source, fh, fattr, flavor); + if (IS_ERR(server)) + return server; + + error = nfs4_delegation_hash_alloc(server); + if (error) { + nfs_free_server(server); + return ERR_PTR(error); + } + + return server; +} + const struct nfs_rpc_ops nfs_v4_clientops = { .version = 4, /* protocol version */ .dentry_ops = &nfs4_dentry_operations, @@ -11019,7 +11039,7 @@ const struct nfs_rpc_ops nfs_v4_clientops = { .init_client = nfs4_init_client, .free_client = nfs4_free_client, .create_server = nfs4_create_server, - .clone_server = nfs_clone_server, + .clone_server = nfs4_clone_server, .discover_trunking = nfs4_discover_trunking, .enable_swap = nfs4_enable_swap, .disable_swap = nfs4_disable_swap, diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index a9b44f12623f..d30c0245031c 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -255,6 +255,8 @@ struct nfs_server { struct list_head layouts; struct list_head delegations; atomic_long_t nr_active_delegations; + unsigned int delegation_hash_mask; + struct hlist_head *delegation_hash_table; struct list_head ss_copies; struct list_head ss_src_copies; -- cgit v1.2.3