linux-next: manual merge of the nfsd tree with Linus' tree

linux-next: manual merge of the nfsd tree with Linus' tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1111 bytes --]

Hi,

Today's linux-next merge of the nfsd tree got a conflict in
fs/nfsd/nfscache.c between commit a517b608fa3d ("nfsd: only unhash DRC
entries that are in the hashtable") from Linus' tree and commit
559aa350587e ("nfsd: track memory utilization in the DRC") from the nfsd
tree.

I fixed it up (see below) and can carry the fix as necessary (no action
is required).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc fs/nfsd/nfscache.c
index ca05f6d,3af1767..0000000
--- a/fs/nfsd/nfscache.c
+++ b/fs/nfsd/nfscache.c
@@@ -100,12 -130,14 +130,15 @@@ nfsd_reply_cache_alloc(void
  static void
  nfsd_reply_cache_free_locked(struct svc_cacherep *rp)
  {
- 	if (rp->c_type == RC_REPLBUFF)
+ 	if (rp->c_type == RC_REPLBUFF && rp->c_replvec.iov_base) {
+ 		drc_mem_usage -= rp->c_replvec.iov_len;
  		kfree(rp->c_replvec.iov_base);
+ 	}
 -	hlist_del(&rp->c_hash);
 +	if (!hlist_unhashed(&rp->c_hash))
 +		hlist_del(&rp->c_hash);
  	list_del(&rp->c_lru);
  	--num_drc_entries;
+ 	drc_mem_usage -= sizeof(*rp);
  	kmem_cache_free(drc_slab, rp);
  }
  

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: linux-next: manual merge of the nfsd tree with Linus' tree

[J. Bruce Fields]

On Wed, Jul 29, 2015 at 09:35:11PM -0400, Trond Myklebust wrote:
> Hi Stephen,
> 
> On Wed, Jul 29, 2015 at 9:23 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> > Hi all,
> >
> > Today's linux-next merge of the nfsd tree got a conflict in:
> >
> >   fs/nfs/nfs42proc.c
> >
> > between commit:
> >
> >   bdcc2cd14e4e ("NFSv4.2: handle NFS-specific llseek errors")
> >
> > from Linus' tree and commit:
> >
> >   0183ae17c741 ("NFSv4.2: handle NFS-specific llseek errors")
> >
> > from the nfsd tree.
> >
> > The only difference here is that _nfs42_proc_llseek is static in the
> > former, so I used that.

Whoops, thanks, I shouldn't have even had that one in my tree....

> Yes, I snuck that declaration into the patch since it was obvious that
> we would never want to export _nfs42_proc_llseek(), and because
> "sparse" complained. Apologies if that caused a conflict...

I actually noticed that, then noticed a bunch of other stuff there had
the same problem, then started to make a patch to fix all those in one
fell swoop, then decided I was being annoying and dropped it.

Um.

--b.

Re: linux-next: manual merge of the nfsd tree with Linus' tree

[Trond Myklebust]

Hi Stephen,

On Wed, Jul 29, 2015 at 9:23 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> Hi all,
>
> Today's linux-next merge of the nfsd tree got a conflict in:
>
>   fs/nfs/nfs42proc.c
>
> between commit:
>
>   bdcc2cd14e4e ("NFSv4.2: handle NFS-specific llseek errors")
>
> from Linus' tree and commit:
>
>   0183ae17c741 ("NFSv4.2: handle NFS-specific llseek errors")
>
> from the nfsd tree.
>
> The only difference here is that _nfs42_proc_llseek is static in the
> former, so I used that.
>

Yes, I snuck that declaration into the patch since it was obvious that
we would never want to export _nfs42_proc_llseek(), and because
"sparse" complained. Apologies if that caused a conflict...

Cheers
  Trond

linux-next: manual merge of the nfsd tree with Linus' tree

[Stephen Rothwell]

Hi all,

Today's linux-next merge of the nfsd tree got a conflict in:

  fs/nfs/nfs42proc.c

between commit:

  bdcc2cd14e4e ("NFSv4.2: handle NFS-specific llseek errors")

from Linus' tree and commit:

  0183ae17c741 ("NFSv4.2: handle NFS-specific llseek errors")

from the nfsd tree.

The only difference here is that _nfs42_proc_llseek is static in the
former, so I used that.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

linux-next: manual merge of the nfsd tree with Linus' tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1828 bytes --]

Hi all,

Today's linux-next merge of the nfsd tree got a conflict in
fs/nfsd/nfs4acl.c between commit 4ac7249ea5a0 ("nfsd: use get_acl and
->set_acl") from Linus' tree and commit 3554116d3aae ("nfsd4: simplify
xdr encoding of nfsv4 names") from the nfsd tree.

I fixed it up (see below) and can carry the fix as necessary (no action
is required).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc fs/nfsd/nfs4acl.c
index 649ad7cf2204,eea24c9a561d..000000000000
--- a/fs/nfsd/nfs4acl.c
+++ b/fs/nfsd/nfs4acl.c
@@@ -37,13 -37,9 +37,14 @@@
  #include <linux/slab.h>
  #include <linux/nfs_fs.h>
  #include <linux/export.h>
 +#include "nfsfh.h"
+ #include "nfsd.h"
  #include "acl.h"
 +#include "vfs.h"
  
 +#define NFS4_ACL_TYPE_DEFAULT	0x01
 +#define NFS4_ACL_DIR		0x02
 +#define NFS4_ACL_OWNER		0x04
  
  /* mode bit translations: */
  #define NFS4_READ_MODE (NFS4_ACE_READ_DATA)
@@@ -916,17 -849,26 +917,22 @@@ nfs4_acl_get_whotype(char *p, u32 len
  	return NFS4_ACL_WHO_NAMED;
  }
  
- int
- nfs4_acl_write_who(int who, char *p)
+ __be32 nfs4_acl_write_who(int who, __be32 **p, int *len)
  {
  	int i;
+ 	int bytes;
  
  	for (i = 0; i < ARRAY_SIZE(s2t_map); i++) {
- 		if (s2t_map[i].type == who) {
- 			memcpy(p, s2t_map[i].string, s2t_map[i].stringlen);
- 			return s2t_map[i].stringlen;
- 		}
+ 		if (s2t_map[i].type != who)
+ 			continue;
+ 		bytes = 4 + (XDR_QUADLEN(s2t_map[i].stringlen) << 2);
+ 		if (bytes > *len)
+ 			return nfserr_resource;
+ 		*p = xdr_encode_opaque(*p, s2t_map[i].string,
+ 					s2t_map[i].stringlen);
+ 		*len -= bytes;
+ 		return 0;
  	}
- 	BUG();
+ 	WARN_ON_ONCE(1);
  	return -1;
  }
 -
 -EXPORT_SYMBOL(nfs4_acl_new);
 -EXPORT_SYMBOL(nfs4_acl_get_whotype);
 -EXPORT_SYMBOL(nfs4_acl_write_who);

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

linux-next: manual merge of the nfsd tree with Linus' tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 4683 bytes --]

Hi all,

Today's linux-next merge of the nfsd tree got a conflict in
fs/nfsd/nfscache.c between commit b67bfe0d42ca ("hlist: drop the node
parameter from iterators") from Linus' tree and commit a4a3ec329124
("nfsd: break out hashtable search into separate function") from the nfsd
tree.

I fixed it up (see below) and can carry the fix as necessary (no action
is required).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc fs/nfsd/nfscache.c
index da3dbd0,ca43664..0000000
--- a/fs/nfsd/nfscache.c
+++ b/fs/nfsd/nfscache.c
@@@ -112,10 -182,130 +182,129 @@@ hash_refile(struct svc_cacherep *rp
  	hlist_add_head(&rp->c_hash, cache_hash + request_hash(rp->c_xid));
  }
  
+ static inline bool
+ nfsd_cache_entry_expired(struct svc_cacherep *rp)
+ {
+ 	return rp->c_state != RC_INPROG &&
+ 	       time_after(jiffies, rp->c_timestamp + RC_EXPIRE);
+ }
+ 
+ /*
+  * Walk the LRU list and prune off entries that are older than RC_EXPIRE.
+  * Also prune the oldest ones when the total exceeds the max number of entries.
+  */
+ static void
+ prune_cache_entries(void)
+ {
+ 	struct svc_cacherep *rp, *tmp;
+ 
+ 	list_for_each_entry_safe(rp, tmp, &lru_head, c_lru) {
+ 		if (!nfsd_cache_entry_expired(rp) &&
+ 		    num_drc_entries <= max_drc_entries)
+ 			break;
+ 		nfsd_reply_cache_free_locked(rp);
+ 	}
+ 
+ 	/*
+ 	 * Conditionally rearm the job. If we cleaned out the list, then
+ 	 * cancel any pending run (since there won't be any work to do).
+ 	 * Otherwise, we rearm the job or modify the existing one to run in
+ 	 * RC_EXPIRE since we just ran the pruner.
+ 	 */
+ 	if (list_empty(&lru_head))
+ 		cancel_delayed_work(&cache_cleaner);
+ 	else
+ 		mod_delayed_work(system_wq, &cache_cleaner, RC_EXPIRE);
+ }
+ 
+ static void
+ cache_cleaner_func(struct work_struct *unused)
+ {
+ 	spin_lock(&cache_lock);
+ 	prune_cache_entries();
+ 	spin_unlock(&cache_lock);
+ }
+ 
+ static int
+ nfsd_reply_cache_shrink(struct shrinker *shrink, struct shrink_control *sc)
+ {
+ 	unsigned int num;
+ 
+ 	spin_lock(&cache_lock);
+ 	if (sc->nr_to_scan)
+ 		prune_cache_entries();
+ 	num = num_drc_entries;
+ 	spin_unlock(&cache_lock);
+ 
+ 	return num;
+ }
+ 
+ /*
+  * Walk an xdr_buf and get a CRC for at most the first RC_CSUMLEN bytes
+  */
+ static __wsum
+ nfsd_cache_csum(struct svc_rqst *rqstp)
+ {
+ 	int idx;
+ 	unsigned int base;
+ 	__wsum csum;
+ 	struct xdr_buf *buf = &rqstp->rq_arg;
+ 	const unsigned char *p = buf->head[0].iov_base;
+ 	size_t csum_len = min_t(size_t, buf->head[0].iov_len + buf->page_len,
+ 				RC_CSUMLEN);
+ 	size_t len = min(buf->head[0].iov_len, csum_len);
+ 
+ 	/* rq_arg.head first */
+ 	csum = csum_partial(p, len, 0);
+ 	csum_len -= len;
+ 
+ 	/* Continue into page array */
+ 	idx = buf->page_base / PAGE_SIZE;
+ 	base = buf->page_base & ~PAGE_MASK;
+ 	while (csum_len) {
+ 		p = page_address(buf->pages[idx]) + base;
+ 		len = min_t(size_t, PAGE_SIZE - base, csum_len);
+ 		csum = csum_partial(p, len, csum);
+ 		csum_len -= len;
+ 		base = 0;
+ 		++idx;
+ 	}
+ 	return csum;
+ }
+ 
+ /*
+  * Search the request hash for an entry that matches the given rqstp.
+  * Must be called with cache_lock held. Returns the found entry or
+  * NULL on failure.
+  */
+ static struct svc_cacherep *
+ nfsd_cache_search(struct svc_rqst *rqstp, __wsum csum)
+ {
+ 	struct svc_cacherep	*rp;
 -	struct hlist_node	*hn;
+ 	struct hlist_head 	*rh;
+ 	__be32			xid = rqstp->rq_xid;
+ 	u32			proto =  rqstp->rq_prot,
+ 				vers = rqstp->rq_vers,
+ 				proc = rqstp->rq_proc;
+ 
+ 	rh = &cache_hash[request_hash(xid)];
 -	hlist_for_each_entry(rp, hn, rh, c_hash) {
++	hlist_for_each_entry(rp, rh, c_hash) {
+ 		if (xid == rp->c_xid && proc == rp->c_proc &&
+ 		    proto == rp->c_prot && vers == rp->c_vers &&
+ 		    rqstp->rq_arg.len == rp->c_len && csum == rp->c_csum &&
+ 		    rpc_cmp_addr(svc_addr(rqstp), (struct sockaddr *)&rp->c_addr) &&
+ 		    rpc_get_port(svc_addr(rqstp)) == rpc_get_port((struct sockaddr *)&rp->c_addr))
+ 			return rp;
+ 	}
+ 	return NULL;
+ }
+ 
  /*
   * Try to find an entry matching the current call in the cache. When none
-  * is found, we grab the oldest unlocked entry off the LRU list.
-  * Note that no operation within the loop may sleep.
+  * is found, we try to grab the oldest expired entry off the LRU list. If
+  * a suitable one isn't there, then drop the cache_lock and allocate a
+  * new one, then search again in case one got inserted while this thread
+  * didn't hold the lock.
   */
  int
  nfsd_cache_lookup(struct svc_rqst *rqstp)

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

linux-next: manual merge of the nfsd tree with Linus' tree

[Stephen Rothwell]

Hi all,

Today's linux-next merge of the nfsd tree got a conflict in
Documentation/filesystems/Locking between commit
b83be6f20a0e468f715b14225c9f897538dfe5ad ("update
Documentation/filesystems/Locking") from Linus' tree and commit
e63eb9375089f9d2041305d04c3f33a194e0e014 ("nfsd4: eliminate lease delete
callback") from the nfsd tree.

I fixed it up (see below) and can carry the fix as necessary.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc Documentation/filesystems/Locking
index 33fa3e5,075be12..0000000
--- a/Documentation/filesystems/Locking
+++ b/Documentation/filesystems/Locking
@@@ -337,22 -327,17 +337,20 @@@ fl_release_private:	maybe		n
  prototypes:
  	int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
  	void (*fl_notify)(struct file_lock *);  /* unblock callback */
 +	int (*fl_grant)(struct file_lock *, struct file_lock *, int);
- 	void (*fl_release_private)(struct file_lock *);
  	void (*fl_break)(struct file_lock *); /* break_lease callback */
 +	int (*fl_mylease)(struct file_lock *, struct file_lock *);
 +	int (*fl_change)(struct file_lock **, int);
  
  locking rules:
 -			BKL	may block
 -fl_compare_owner:	yes	no
 -fl_notify:		yes	no
 -fl_break:		yes	no
 -
 -	Currently only NFSD and NLM provide instances of this class. None of the
 -them block. If you have out-of-tree instances - please, show up. Locking
 -in that area will change.
 +			file_lock_lock	may block
 +fl_compare_owner:	yes		no
 +fl_notify:		yes		no
 +fl_grant:		no		no
- fl_release_private:	maybe		no
 +fl_break:		yes		no
 +fl_mylease:		yes		no
 +fl_change		yes		no
 +
  --------------------------- buffer_head -----------------------------------
  prototypes:
  	void (*b_end_io)(struct buffer_head *bh, int uptodate);