linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: trondmy@kernel.org
To: linux-nfs@vger.kernel.org
Subject: [PATCH v2 21/26] NFS: Use information about the change attribute to optimise updates
Date: Wed, 14 Apr 2021 09:43:48 -0400	[thread overview]
Message-ID: <20210414134353.11860-22-trondmy@kernel.org> (raw)
In-Reply-To: <20210414134353.11860-21-trondmy@kernel.org>

From: Trond Myklebust <trond.myklebust@hammerspace.com>

If the NFSv4.2 server supports the 'change_attr_type' attribute, then
allow the client to optimise its attribute cache update strategy.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 fs/nfs/inode.c    | 111 ++++++++++++++++++++++++++++++++++++++--------
 fs/nfs/nfs4proc.c |  20 +++++++--
 2 files changed, 110 insertions(+), 21 deletions(-)

diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index fc8e92794636..d218d164414f 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -1657,25 +1657,20 @@ EXPORT_SYMBOL_GPL(_nfs_display_fhandle);
 #endif
 
 /**
- * nfs_inode_attrs_need_update - check if the inode attributes need updating
- * @inode: pointer to inode
+ * nfs_inode_attrs_cmp_generic - compare attributes
  * @fattr: attributes
+ * @inode: pointer to inode
  *
  * Attempt to divine whether or not an RPC call reply carrying stale
  * attributes got scheduled after another call carrying updated ones.
- *
- * To do so, the function first assumes that a more recent ctime means
- * that the attributes in fattr are newer, however it also attempt to
- * catch the case where ctime either didn't change, or went backwards
- * (if someone reset the clock on the server) by looking at whether
- * or not this RPC call was started after the inode was last updated.
  * Note also the check for wraparound of 'attr_gencount'
  *
- * The function returns 'true' if it thinks the attributes in 'fattr' are
- * more recent than the ones cached in the inode.
- *
+ * The function returns '1' if it thinks the attributes in @fattr are
+ * more recent than the ones cached in @inode. Otherwise it returns
+ * the value '0'.
  */
-static int nfs_inode_attrs_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
+static int nfs_inode_attrs_cmp_generic(const struct nfs_fattr *fattr,
+				       const struct inode *inode)
 {
 	unsigned long attr_gencount = NFS_I(inode)->attr_gencount;
 
@@ -1683,15 +1678,93 @@ static int nfs_inode_attrs_need_update(const struct inode *inode, const struct n
 	       (long)(attr_gencount - nfs_read_attr_generation_counter()) > 0;
 }
 
-static int nfs_refresh_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
+/**
+ * nfs_inode_attrs_cmp_monotonic - compare attributes
+ * @fattr: attributes
+ * @inode: pointer to inode
+ *
+ * Attempt to divine whether or not an RPC call reply carrying stale
+ * attributes got scheduled after another call carrying updated ones.
+ *
+ * We assume that the server observes monotonic semantics for
+ * the change attribute, so a larger value means that the attributes in
+ * @fattr are more recent, in which case the function returns the
+ * value '1'.
+ * A return value of '0' indicates no measurable change
+ * A return value of '-1' means that the attributes in @inode are
+ * more recent.
+ */
+static int nfs_inode_attrs_cmp_monotonic(const struct nfs_fattr *fattr,
+					 const struct inode *inode)
 {
-	int ret;
+	s64 diff = fattr->change_attr - inode_peek_iversion_raw(inode);
+	if (diff > 0)
+		return 1;
+	return diff == 0 ? 0 : -1;
+}
+
+/**
+ * nfs_inode_attrs_cmp_strict_monotonic - compare attributes
+ * @fattr: attributes
+ * @inode: pointer to inode
+ *
+ * Attempt to divine whether or not an RPC call reply carrying stale
+ * attributes got scheduled after another call carrying updated ones.
+ *
+ * We assume that the server observes strictly monotonic semantics for
+ * the change attribute, so a larger value means that the attributes in
+ * @fattr are more recent, in which case the function returns the
+ * value '1'.
+ * A return value of '-1' means that the attributes in @inode are
+ * more recent or unchanged.
+ */
+static int nfs_inode_attrs_cmp_strict_monotonic(const struct nfs_fattr *fattr,
+						const struct inode *inode)
+{
+	return  nfs_inode_attrs_cmp_monotonic(fattr, inode) > 0 ? 1 : -1;
+}
+
+/**
+ * nfs_inode_attrs_cmp - compare attributes
+ * @fattr: attributes
+ * @inode: pointer to inode
+ *
+ * This function returns '1' if it thinks the attributes in @fattr are
+ * more recent than the ones cached in @inode. It returns '-1' if
+ * the attributes in @inode are more recent than the ones in @fattr,
+ * and it returns 0 if not sure.
+ */
+static int nfs_inode_attrs_cmp(const struct nfs_fattr *fattr,
+			       const struct inode *inode)
+{
+	if (nfs_inode_attrs_cmp_generic(fattr, inode) > 0)
+		return 1;
+	switch (NFS_SERVER(inode)->change_attr_type) {
+	case NFS4_CHANGE_TYPE_IS_UNDEFINED:
+		break;
+	case NFS4_CHANGE_TYPE_IS_TIME_METADATA:
+		if (!(fattr->valid & NFS_ATTR_FATTR_CHANGE))
+			break;
+		return nfs_inode_attrs_cmp_monotonic(fattr, inode);
+	default:
+		if (!(fattr->valid & NFS_ATTR_FATTR_CHANGE))
+			break;
+		return nfs_inode_attrs_cmp_strict_monotonic(fattr, inode);
+	}
+	return 0;
+}
+
+static int nfs_refresh_inode_locked(struct inode *inode,
+				    struct nfs_fattr *fattr)
+{
+	int attr_cmp = nfs_inode_attrs_cmp(fattr, inode);
+	int ret = 0;
 
 	trace_nfs_refresh_inode_enter(inode);
 
-	if (nfs_inode_attrs_need_update(inode, fattr))
+	if (attr_cmp > 0)
 		ret = nfs_update_inode(inode, fattr);
-	else
+	else if (attr_cmp == 0)
 		ret = nfs_check_inode_attributes(inode, fattr);
 
 	trace_nfs_refresh_inode_exit(inode, ret);
@@ -1776,11 +1849,13 @@ EXPORT_SYMBOL_GPL(nfs_post_op_update_inode);
  */
 int nfs_post_op_update_inode_force_wcc_locked(struct inode *inode, struct nfs_fattr *fattr)
 {
+	int attr_cmp = nfs_inode_attrs_cmp(fattr, inode);
 	int status;
 
 	/* Don't do a WCC update if these attributes are already stale */
-	if ((fattr->valid & NFS_ATTR_FATTR) == 0 ||
-			!nfs_inode_attrs_need_update(inode, fattr)) {
+	if (attr_cmp < 0)
+		return 0;
+	if ((fattr->valid & NFS_ATTR_FATTR) == 0 || !attr_cmp) {
 		fattr->valid &= ~(NFS_ATTR_FATTR_PRECHANGE
 				| NFS_ATTR_FATTR_PRESIZE
 				| NFS_ATTR_FATTR_PREMTIME
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 6992c88a25e7..0b0a48d78299 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -1186,10 +1186,23 @@ nfs4_update_changeattr_locked(struct inode *inode,
 		unsigned long timestamp, unsigned long cache_validity)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
+	u64 change_attr = inode_peek_iversion_raw(inode);
 
 	cache_validity |= NFS_INO_INVALID_CTIME | NFS_INO_INVALID_MTIME;
 
-	if (cinfo->atomic && cinfo->before == inode_peek_iversion_raw(inode)) {
+	switch (NFS_SERVER(inode)->change_attr_type) {
+	case NFS4_CHANGE_TYPE_IS_UNDEFINED:
+		break;
+	case NFS4_CHANGE_TYPE_IS_TIME_METADATA:
+		if ((s64)(change_attr - cinfo->after) > 0)
+			goto out;
+		break;
+	default:
+		if ((s64)(change_attr - cinfo->after) >= 0)
+			goto out;
+	}
+
+	if (cinfo->atomic && cinfo->before == change_attr) {
 		nfsi->attrtimeo_timestamp = jiffies;
 	} else {
 		if (S_ISDIR(inode->i_mode)) {
@@ -1201,7 +1214,7 @@ nfs4_update_changeattr_locked(struct inode *inode,
 				cache_validity |= NFS_INO_REVAL_PAGECACHE;
 		}
 
-		if (cinfo->before != inode_peek_iversion_raw(inode))
+		if (cinfo->before != change_attr)
 			cache_validity |= NFS_INO_INVALID_ACCESS |
 					  NFS_INO_INVALID_ACL |
 					  NFS_INO_INVALID_XATTR;
@@ -1209,8 +1222,9 @@ nfs4_update_changeattr_locked(struct inode *inode,
 	inode_set_iversion_raw(inode, cinfo->after);
 	nfsi->read_cache_jiffies = timestamp;
 	nfsi->attr_gencount = nfs_inc_attr_generation_counter();
-	nfs_set_cache_invalid(inode, cache_validity);
 	nfsi->cache_validity &= ~NFS_INO_INVALID_CHANGE;
+out:
+	nfs_set_cache_invalid(inode, cache_validity);
 }
 
 void
-- 
2.30.2


  reply	other threads:[~2021-04-14 13:44 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-14 13:43 [PATCH v2 00/26] Attribute revalidation updates trondmy
2021-04-14 13:43 ` [PATCH v2 01/26] NFS: Deal correctly with attribute generation counter overflow trondmy
2021-04-14 13:43   ` [PATCH v2 02/26] NFS: Fix up inode cache tracing trondmy
2021-04-14 13:43     ` [PATCH v2 03/26] NFS: Mask out unsupported attributes in nfs_getattr() trondmy
2021-04-14 13:43       ` [PATCH v2 04/26] NFS: NFS_INO_REVAL_PAGECACHE should mark the change attribute invalid trondmy
2021-04-14 13:43         ` [PATCH v2 05/26] NFS: Fix up revalidation of space used trondmy
2021-04-14 13:43           ` [PATCH v2 06/26] NFS: Don't revalidate attributes that are not being asked for trondmy
2021-04-14 13:43             ` [PATCH v2 07/26] NFS: Fix up statx() results trondmy
2021-04-14 13:43               ` [PATCH v2 08/26] NFS: nfs_setattr_update_inode() should clear the suid/sgid bits trondmy
2021-04-14 13:43                 ` [PATCH v2 09/26] NFS: Add a cache validity flag argument to nfs_revalidate_inode() trondmy
2021-04-14 13:43                   ` [PATCH v2 10/26] NFS: Replace use of NFS_INO_REVAL_PAGECACHE when checking cache validity trondmy
2021-04-14 13:43                     ` [PATCH v2 11/26] NFS: Don't set NFS_INO_REVAL_PAGECACHE in the inode " trondmy
2021-04-14 13:43                       ` [PATCH v2 12/26] NFSv4: Fix nfs4_bitmap_copy_adjust() trondmy
2021-04-14 13:43                         ` [PATCH v2 13/26] NFS: Separate tracking of file nlinks cache validity from the mode/uid/gid trondmy
2021-04-14 13:43                           ` [PATCH v2 14/26] NFS: Separate tracking of file mode cache validity from the uid/gid trondmy
2021-04-14 13:43                             ` [PATCH v2 15/26] NFS: Fix up handling of outstanding layoutcommit in nfs_update_inode() trondmy
2021-04-14 13:43                               ` [PATCH v2 16/26] NFS: Remove a line of code that has no effect " trondmy
2021-04-14 13:43                                 ` [PATCH v2 17/26] NFS: Simplify cache consistency in nfs_check_inode_attributes() trondmy
2021-04-14 13:43                                   ` [PATCH v2 18/26] NFSv4: Fix value of decode_fsinfo_maxsz trondmy
2021-04-14 13:43                                     ` [PATCH v2 19/26] NFSv4: Don't modify the change attribute cached in the inode trondmy
2021-04-14 13:43                                       ` [PATCH v2 20/26] NFSv4: Add support for the NFSv4.2 "change_attr_type" attribute trondmy
2021-04-14 13:43                                         ` trondmy [this message]
2021-04-14 13:43                                           ` [PATCH v2 22/26] NFS: Another inode revalidation improvement trondmy
2021-04-14 13:43                                             ` [PATCH v2 23/26] NFSv4: nfs4_inc/dec_nlink_locked should also invalidate ctime trondmy
2021-04-14 13:43                                               ` [PATCH v2 24/26] NFSv4: link must update the inode nlink trondmy
2021-04-14 13:43                                                 ` [PATCH v2 25/26] NFS: Don't store NFS_INO_REVAL_FORCED trondmy
2021-04-14 13:43                                                   ` [PATCH v2 26/26] NFS: Split attribute support out from the server capabilities trondmy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210414134353.11860-22-trondmy@kernel.org \
    --to=trondmy@kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).