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

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2146 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/addr.c between commit 496ad9aa8ef4 ("new helper: file_inode
(file)") from Linus' tree and commit 42f94fe2cf50 ("ceph: revert commit
22cddde104") from the ceph 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/ceph/addr.c
index a60ea97,068d2c8..0000000
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@@ -1066,29 -1104,10 +1104,10 @@@ static int ceph_write_begin(struct fil
  			    loff_t pos, unsigned len, unsigned flags,
  			    struct page **pagep, void **fsdata)
  {
 -	struct inode *inode = file->f_dentry->d_inode;
 +	struct inode *inode = file_inode(file);
- 	struct ceph_inode_info *ci = ceph_inode(inode);
- 	struct ceph_file_info *fi = file->private_data;
  	struct page *page;
  	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
- 	int r, want, got = 0;
- 
- 	if (fi->fmode & CEPH_FILE_MODE_LAZY)
- 		want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO;
- 	else
- 		want = CEPH_CAP_FILE_BUFFER;
- 
- 	dout("write_begin %p %llx.%llx %llu~%u getting caps. i_size %llu\n",
- 	     inode, ceph_vinop(inode), pos, len, inode->i_size);
- 	r = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, &got, pos+len);
- 	if (r < 0)
- 		return r;
- 	dout("write_begin %p %llx.%llx %llu~%u  got cap refs on %s\n",
- 	     inode, ceph_vinop(inode), pos, len, ceph_cap_string(got));
- 	if (!(got & (CEPH_CAP_FILE_BUFFER|CEPH_CAP_FILE_LAZYIO))) {
- 		ceph_put_cap_refs(ci, got);
- 		return -EAGAIN;
- 	}
+ 	int r;
  
  	do {
  		/* get a page */
@@@ -1124,8 -1134,7 +1134,7 @@@ static int ceph_write_end(struct file *
  			  loff_t pos, unsigned len, unsigned copied,
  			  struct page *page, void *fsdata)
  {
 -	struct inode *inode = file->f_dentry->d_inode;
 +	struct inode *inode = file_inode(file);
- 	struct ceph_inode_info *ci = ceph_inode(inode);
  	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
  	struct ceph_mds_client *mdsc = fsc->mdsc;
  	unsigned from = pos & (PAGE_CACHE_SIZE - 1);

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

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

[Sage Weil]

On Wed, 10 Jul 2013, Stephen Rothwell wrote:
> Hi Sage,
> 
> Today's linux-next merge of the ceph tree got conflicts in
> drivers/block/rbd.c and net/ceph/osd_client.c because the ceph tree was
> rebased before being sent to Linus and it looks like one patch
> was dropped and several more added.
> 
> I just used the upstream version of the cpeh tree for today - please
> clean up.

Fixed, thanks!

sage

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 403 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got conflicts in
drivers/block/rbd.c and net/ceph/osd_client.c because the ceph tree was
rebased before being sent to Linus and it looks like one patch
was dropped and several more added.

I just used the upstream version of the cpeh tree for today - please
clean up.

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

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

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1058 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
drivers/block/rbd.c between commit db2a144bedd5
("block_device_operations->release() should return void") from Linus'
tree and commit 87f11b7221f7 ("rbd: don't hold ctl_mutex to get/put
device") from the ceph tree.

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

BTW, Sage, there is no Signed-off-by for you in the ceph tree commit (and
several others - even though you committed them).
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc drivers/block/rbd.c
index 3063452,f24e241..0000000
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@@ -507,9 -505,9 +505,7 @@@ static void rbd_release(struct gendisk 
  	spin_unlock_irq(&rbd_dev->lock);
  	rbd_assert(open_count_before > 0);
  
- 	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
  	put_device(&rbd_dev->dev);
- 	mutex_unlock(&ctl_mutex);
 -
 -	return 0;
  }
  
  static const struct block_device_operations rbd_bd_ops = {

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

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 442 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
net/ceph/messenger.c between commit 5bdca4e0768d ("libceph: fix messenger
retry") from Linus' tree and commit ab166d5aa3bc ("libceph: separate
banner and connect writes") from the ceph tree.

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

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

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

[Stephen Rothwell]

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

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
net/ceph/messenger.c between commit 5bdca4e0768d ("libceph: fix messenger
retry") from Linus' tree and commit e22004235a90 ("libceph: rename
kvec_reset and kvec_add functions") from the ceph tree.

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

diff --cc net/ceph/messenger.c
index 10255e8,23073cf..0000000
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@@ -1450,11 -1530,11 +1530,11 @@@ static int process_connect(struct ceph_
  		 * If we sent a smaller connect_seq than the peer has, try
  		 * again with a larger value.
  		 */
 -		dout("process_connect got RETRY my seq = %u, peer_seq = %u\n",
 +		dout("process_connect got RETRY_SESSION my seq %u, peer %u\n",
  		     le32_to_cpu(con->out_connect.connect_seq),
 -		     le32_to_cpu(con->in_connect.connect_seq));
 -		con->connect_seq = le32_to_cpu(con->in_connect.connect_seq);
 +		     le32_to_cpu(con->in_reply.connect_seq));
 +		con->connect_seq = le32_to_cpu(con->in_reply.connect_seq);
- 		ceph_con_out_kvec_reset(con);
+ 		con_out_kvec_reset(con);
  		ret = prepare_write_connect(con);
  		if (ret < 0)
  			return ret;
@@@ -1468,10 -1548,10 +1548,10 @@@
  		 */
  		dout("process_connect got RETRY_GLOBAL my %u peer_gseq %u\n",
  		     con->peer_global_seq,
 -		     le32_to_cpu(con->in_connect.global_seq));
 +		     le32_to_cpu(con->in_reply.global_seq));
  		get_global_seq(con->msgr,
 -			       le32_to_cpu(con->in_connect.global_seq));
 +			       le32_to_cpu(con->in_reply.global_seq));
- 		ceph_con_out_kvec_reset(con);
+ 		con_out_kvec_reset(con);
  		ret = prepare_write_connect(con);
  		if (ret < 0)
  			return ret;

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

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

[Sage Weil]

On Tue, 22 May 2012, Stephen Rothwell wrote:
> Hi Sage,
> 
> Today's linux-next merge of the ceph tree got a conflict in
> net/ceph/messenger.c between commit 95c961747284 ("net: cleanup unsigned
> to unsigned int") from Linus' tree and commit 57dac9d16209 ("ceph:
> messenger: use read_partial() in read_partial_message()") from the ceph
> tree.
> 
> Just context changes. I fixed it up (see below) and can carry the fix as
> necessary.

Thanks, Stephen!  If there's an easier way to carry this in our tree, just 
let me know.

sage


> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> 
> diff --cc net/ceph/messenger.c
> index 36fa6bf,1a80907..0000000
> --- a/net/ceph/messenger.c
> +++ b/net/ceph/messenger.c
> @@@ -695,12 -698,12 +698,12 @@@ static void prepare_write_banner(struc
>   	set_bit(WRITE_PENDING, &con->state);
>   }
>   
> - static int prepare_write_connect(struct ceph_messenger *msgr,
> - 				 struct ceph_connection *con,
> - 				 int include_banner)
> + static int prepare_write_connect(struct ceph_connection *con)
>   {
>  -	unsigned global_seq = get_global_seq(con->msgr, 0);
>  +	unsigned int global_seq = get_global_seq(con->msgr, 0);
>   	int proto;
> + 	int auth_proto;
> + 	struct ceph_auth_handshake *auth;
>   
>   	switch (con->peer_name.type) {
>   	case CEPH_ENTITY_TYPE_MON:
> @@@ -1627,9 -1663,10 +1663,10 @@@ static int read_partial_message_bio(str
>   static int read_partial_message(struct ceph_connection *con)
>   {
>   	struct ceph_msg *m = con->in_msg;
> + 	int size;
> + 	int end;
>   	int ret;
> - 	int to, left;
>  -	unsigned front_len, middle_len, data_len;
>  +	unsigned int front_len, middle_len, data_len;
>   	bool do_datacrc = !con->msgr->nocrc;
>   	int skip;
>   	u64 seq;
> 

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1562 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
net/ceph/messenger.c between commit 95c961747284 ("net: cleanup unsigned
to unsigned int") from Linus' tree and commit 57dac9d16209 ("ceph:
messenger: use read_partial() in read_partial_message()") from the ceph
tree.

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

diff --cc net/ceph/messenger.c
index 36fa6bf,1a80907..0000000
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@@ -695,12 -698,12 +698,12 @@@ static void prepare_write_banner(struc
  	set_bit(WRITE_PENDING, &con->state);
  }
  
- static int prepare_write_connect(struct ceph_messenger *msgr,
- 				 struct ceph_connection *con,
- 				 int include_banner)
+ static int prepare_write_connect(struct ceph_connection *con)
  {
 -	unsigned global_seq = get_global_seq(con->msgr, 0);
 +	unsigned int global_seq = get_global_seq(con->msgr, 0);
  	int proto;
+ 	int auth_proto;
+ 	struct ceph_auth_handshake *auth;
  
  	switch (con->peer_name.type) {
  	case CEPH_ENTITY_TYPE_MON:
@@@ -1627,9 -1663,10 +1663,10 @@@ static int read_partial_message_bio(str
  static int read_partial_message(struct ceph_connection *con)
  {
  	struct ceph_msg *m = con->in_msg;
+ 	int size;
+ 	int end;
  	int ret;
- 	int to, left;
 -	unsigned front_len, middle_len, data_len;
 +	unsigned int front_len, middle_len, data_len;
  	bool do_datacrc = !con->msgr->nocrc;
  	int skip;
  	u64 seq;

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

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2497 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/xattr.c between commit 83eb26af0db7 ("ceph: ensure prealloc_blob
is in place when removing xattr") from Linus' tree and commit
15226d16c19c ("ceph: make ceph_setxattr() and ceph_removexattr() more
alike") from the ceph tree.

I used the resolution provided by Alex (thanks) (see below) and I can
carry the fix as necessary.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc fs/ceph/xattr.c
index a76f697,a3c1f9b..0000000
--- a/fs/ceph/xattr.c
+++ b/fs/ceph/xattr.c
@@@ -816,11 -883,10 +883,11 @@@ static int ceph_send_removexattr(struc
  int ceph_removexattr(struct dentry *dentry, const char *name)
  {
  	struct inode *inode = dentry->d_inode;
+ 	struct ceph_vxattr *vxattr;
  	struct ceph_inode_info *ci = ceph_inode(inode);
- 	struct ceph_vxattr_cb *vxattrs = ceph_inode_vxattrs(inode);
  	int issued;
  	int err;
 +	int required_blob_size;
  	int dirty;
  
  	if (ceph_snap(inode) != CEPH_NOSNAP)
@@@ -829,42 -895,19 +896,39 @@@
  	if (!ceph_is_valid_xattr(name))
  		return -EOPNOTSUPP;
  
- 	if (vxattrs) {
- 		struct ceph_vxattr_cb *vxattr =
- 			ceph_match_vxattr(vxattrs, name);
- 		if (vxattr && vxattr->readonly)
- 			return -EOPNOTSUPP;
- 	}
+ 	vxattr = ceph_match_vxattr(inode, name);
+ 	if (vxattr && vxattr->readonly)
+ 		return -EOPNOTSUPP;
  
 +	err = -ENOMEM;
  	spin_lock(&ci->i_ceph_lock);
- 	__build_xattrs(inode);
 +retry:
  	issued = __ceph_caps_issued(ci, NULL);
  	dout("removexattr %p issued %s\n", inode, ceph_cap_string(issued));
- 
  	if (!(issued & CEPH_CAP_XATTR_EXCL))
  		goto do_sync;
+ 	__build_xattrs(inode);
  
 +	required_blob_size = __get_required_blob_size(ci, 0, 0);
 +
 +	if (!ci->i_xattrs.prealloc_blob ||
 +	    required_blob_size > ci->i_xattrs.prealloc_blob->alloc_len) {
 +		struct ceph_buffer *blob;
 +
 +		spin_unlock(&ci->i_ceph_lock);
 +		dout(" preaallocating new blob size=%d\n", required_blob_size);
 +		blob = ceph_buffer_new(required_blob_size, GFP_NOFS);
 +		if (!blob)
 +			goto out;
 +		spin_lock(&ci->i_ceph_lock);
 +		if (ci->i_xattrs.prealloc_blob)
 +			ceph_buffer_put(ci->i_xattrs.prealloc_blob);
 +		ci->i_xattrs.prealloc_blob = blob;
 +		goto retry;
 +	}
 +
  	err = __remove_xattr_by_name(ceph_inode(inode), name);
+ 
  	dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_XATTR_EXCL);
  	ci->i_xattrs.dirty = true;
  	inode->i_ctime = CURRENT_TIME;

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

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 337 bytes --]

Hi Sage,

On Tue, 10 Jan 2012 17:35:36 -0800 (PST) Sage Weil <sage@newdream.net> wrote:
>
> BTW, should I let Linus resolve this, or merge his master back into my 
> tree, or rebase onto his master?

Let Linus resolve it (giving him a hint if necessary).

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

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

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

[Alex Elder]

On Wed, 2012-01-11 at 01:31 +0000, Al Viro wrote:
> On Wed, Jan 11, 2012 at 12:27:44PM +1100, Stephen Rothwell wrote:
> > Hi Sage,
> > 
> > Today's linux-next merge of the ceph tree got a conflict in
> > fs/ceph/super.c between commit 3c5184ef1216 ("ceph: d_alloc_root() may
> > fail") from Linus' tree and commit 26d913cdd955 ("ceph: always initialize
> > the dentry in open_root_dentry()") from the ceph tree.
> > 
> > I fixed it up (see below) and can carry the fix as necessary.
> 
> Ahem...  
> 
> >  +			root = d_obtain_alias(inode);
> >  +		}
> > + 		ceph_init_dentry(root);
> 
> What will happen if dentry returned by d_obtain_alias() had already existed?

I had *exactly* the same reaction, and maybe the code would
have been a little less surprising if the null check were
in the caller (here)  rather than in ceph_init_dentry().

					-Alex

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

[Al Viro]

On Tue, Jan 10, 2012 at 05:35:36PM -0800, Sage Weil wrote:
> On Wed, 11 Jan 2012, Al Viro wrote:
> > On Wed, Jan 11, 2012 at 12:27:44PM +1100, Stephen Rothwell wrote:
> > > Hi Sage,
> > > 
> > > Today's linux-next merge of the ceph tree got a conflict in
> > > fs/ceph/super.c between commit 3c5184ef1216 ("ceph: d_alloc_root() may
> > > fail") from Linus' tree and commit 26d913cdd955 ("ceph: always initialize
> > > the dentry in open_root_dentry()") from the ceph tree.
> > > 
> > > I fixed it up (see below) and can carry the fix as necessary.
> > 
> > Ahem...  
> > 
> > >  +			root = d_obtain_alias(inode);
> > >  +		}
> > > + 		ceph_init_dentry(root);
> > 
> > What will happen if dentry returned by d_obtain_alias() had already existed?
> 
> ceph_init_dentry() bails out immediately if d_fsdata is non-NULL.

Right.

> BTW, should I let Linus resolve this, or merge his master back into my 
> tree, or rebase onto his master?

My fault - I should've checked your tree for conflicts and feed that fix to
you since it turned out to have them.  I thought for some reason that ceph
pull had already happened...

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

[Sage Weil]

On Wed, 11 Jan 2012, Al Viro wrote:
> On Wed, Jan 11, 2012 at 12:27:44PM +1100, Stephen Rothwell wrote:
> > Hi Sage,
> > 
> > Today's linux-next merge of the ceph tree got a conflict in
> > fs/ceph/super.c between commit 3c5184ef1216 ("ceph: d_alloc_root() may
> > fail") from Linus' tree and commit 26d913cdd955 ("ceph: always initialize
> > the dentry in open_root_dentry()") from the ceph tree.
> > 
> > I fixed it up (see below) and can carry the fix as necessary.
> 
> Ahem...  
> 
> >  +			root = d_obtain_alias(inode);
> >  +		}
> > + 		ceph_init_dentry(root);
> 
> What will happen if dentry returned by d_obtain_alias() had already existed?

ceph_init_dentry() bails out immediately if d_fsdata is non-NULL.

BTW, should I let Linus resolve this, or merge his master back into my 
tree, or rebase onto his master?

sage

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

[Al Viro]

On Wed, Jan 11, 2012 at 12:27:44PM +1100, Stephen Rothwell wrote:
> Hi Sage,
> 
> Today's linux-next merge of the ceph tree got a conflict in
> fs/ceph/super.c between commit 3c5184ef1216 ("ceph: d_alloc_root() may
> fail") from Linus' tree and commit 26d913cdd955 ("ceph: always initialize
> the dentry in open_root_dentry()") from the ceph tree.
> 
> I fixed it up (see below) and can carry the fix as necessary.

Ahem...  

>  +			root = d_obtain_alias(inode);
>  +		}
> + 		ceph_init_dentry(root);

What will happen if dentry returned by d_obtain_alias() had already existed?

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1473 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/super.c between commit 3c5184ef1216 ("ceph: d_alloc_root() may
fail") from Linus' tree and commit 26d913cdd955 ("ceph: always initialize
the dentry in open_root_dentry()") from the ceph tree.

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

diff --cc fs/ceph/super.c
index 48f61a1,ec74313..0000000
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@@ -636,21 -636,14 +636,21 @@@ static struct dentry *open_root_dentry(
  	req->r_num_caps = 2;
  	err = ceph_mdsc_do_request(mdsc, NULL, req);
  	if (err == 0) {
 +		struct inode *inode = req->r_target_inode;
 +		req->r_target_inode = NULL;
  		dout("open_root_inode success\n");
 -		if (ceph_ino(req->r_target_inode) == CEPH_INO_ROOT &&
 -		    fsc->sb->s_root == NULL)
 -			root = d_alloc_root(req->r_target_inode);
 -		else
 -			root = d_obtain_alias(req->r_target_inode);
 +		if (ceph_ino(inode) == CEPH_INO_ROOT &&
 +		    fsc->sb->s_root == NULL) {
 +			root = d_alloc_root(inode);
 +			if (!root) {
 +				iput(inode);
 +				root = ERR_PTR(-ENOMEM);
 +				goto out;
 +			}
- 			ceph_init_dentry(root);
 +		} else {
 +			root = d_obtain_alias(inode);
 +		}
+ 		ceph_init_dentry(root);
 -		req->r_target_inode = NULL;
  		dout("open_root_inode success, root dentry is %p\n", root);
  	} else {
  		root = ERR_PTR(err);

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

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/mds_client.c between commit
25933abdd8c562182ca6dc9f8c4c2cc8265c3a80 ("ceph: Handle file locks in
replies from the MDS") from Linus' tree and commit
cba1a66d511d46c426ea02c322293b48f3b88cf2 ("ceph: implement DIRLAYOUTHASH
feature to get dir layout from MDS") from the ceph tree.

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

diff --cc fs/ceph/mds_client.c
index 38800ea,bb9ff98..0000000
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@@ -202,38 -212,6 +211,38 @@@ out_bad
  }
  
  /*
 + * parse fcntl F_GETLK results
 + */
 +static int parse_reply_info_filelock(void **p, void *end,
 +                struct ceph_mds_reply_info_parsed *info)
 +{
 +	if (*p + sizeof(*info->filelock_reply) > end)
 +		goto bad;
 +
 +	info->filelock_reply = *p;
 +	*p += sizeof(*info->filelock_reply);
 +
 +	if (unlikely(*p != end))
 +		goto bad;
 +	return 0;
 +
 +bad:
 +	return -EIO;
 +}
 +
 +/*
 + * parse extra results
 + */
 +static int parse_reply_info_extra(void **p, void *end,
-                 struct ceph_mds_reply_info_parsed *info)
++                struct ceph_mds_reply_info_parsed *info, int features)
 +{
 +	if (info->head->op == CEPH_MDS_OP_GETFILELOCK)
 +		return parse_reply_info_filelock(p, end, info);
 +	else
- 		return parse_reply_info_dir(p, end, info);
++		return parse_reply_info_dir(p, end, info, features);
 +}
 +
 +/*
   * parse entire mds reply
   */
  static int parse_reply_info(struct ceph_msg *msg,
@@@ -255,10 -234,10 +265,10 @@@
  			goto out_bad;
  	}
  
 -	/* dir content */
 +	/* extra */
  	ceph_decode_32_safe(&p, end, len, bad);
  	if (len > 0) {
- 		err = parse_reply_info_extra(&p, p+len, info);
 -		err = parse_reply_info_dir(&p, p+len, info, features);
++		err = parse_reply_info_extra(&p, p+len, info, features);
  		if (err < 0)
  			goto out_bad;
  	}

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

[Sage Weil]

Thanks, Stephen.

I've rebased my for-next on the latest upstream (now that lock_flocks() is 
in place), so these should all go away the next time around.

sage


On Thu, 23 Sep 2010, Stephen Rothwell wrote:

> Hi Sage,
> 
> Today's linux-next merge of the ceph tree got a conflict in
> fs/ceph/Kconfig between commit be4f104dfd3b5e3ae262bff607965cfc38027dec
> ("ceph: select CRYPTO") from Linus' tree and commit
> 49dde0c8bee346a349862c4a3eddcd58e0d0ae90 ("ceph: factor out libceph from
> Ceph file system") from the ceph tree.
> 
> Just overlapping additions.  I fixed it up (see below) and can carry the
> fix for a while.  (though the adding of "default n" is, I think,
> redundant as that is the default anyway).
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> 
> diff --cc fs/ceph/Kconfig
> index 0fcd264,73a7b31..0000000
> --- a/fs/ceph/Kconfig
> +++ b/fs/ceph/Kconfig
> @@@ -1,9 -1,10 +1,11 @@@
>   config CEPH_FS
>           tristate "Ceph distributed file system (EXPERIMENTAL)"
>   	depends on INET && EXPERIMENTAL
> + 	select CEPH_LIB
>   	select LIBCRC32C
>   	select CRYPTO_AES
>  +	select CRYPTO
> + 	default n
>   	help
>   	  Choose Y or M here to include support for mounting the
>   	  experimental Ceph distributed file system.  Ceph is an extremely
> 
> 

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 585 bytes --]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/pagelist.c between commit
3d4401d9d0aef5c40706350685ddea3df6708496 ("ceph: fix pagelist kunmap
tail") from Linus' tree and commit
49dde0c8bee346a349862c4a3eddcd58e0d0ae90 ("ceph: factor out libceph from
Ceph file system") from the ceph tree.

The latter moved the file to net/ceph/pagelist.c, and the former patch
has been applied there, so I just removed the fs/ceph version of the file.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/mds_client.c between commit
3612abbd5df6baa9ca3e0777c6c8646e202d3f66 ("ceph: fix reconnect encoding
for old servers") from Linus' tree and commit
20d5970378f075f8f567b92bd2c16373fbe09b72 ("ceph: preallocate flock state
without locks held") from the ceph tree.

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

diff --cc fs/ceph/mds_client.c
index fad95f8,9c648ed..0000000
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@@ -2361,21 -2365,35 +2365,37 @@@ static int encode_caps_cb(struct inode 
  
  	if (recon_state->flock) {
  		int num_fcntl_locks, num_flock_locks;
- 
- 		lock_kernel();
- 		ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks);
- 		rec.v2.flock_len = (2*sizeof(u32) +
- 				    (num_fcntl_locks+num_flock_locks) *
- 				    sizeof(struct ceph_filelock));
- 
- 		err = ceph_pagelist_append(pagelist, &rec, reclen);
- 		if (!err)
- 			err = ceph_encode_locks(inode, pagelist,
- 						num_fcntl_locks,
- 						num_flock_locks);
- 		unlock_kernel();
+ 		struct ceph_pagelist_cursor trunc_point;
+ 
+ 		ceph_pagelist_set_cursor(pagelist, &trunc_point);
+ 		do {
+ 			lock_kernel();
+ 			ceph_count_locks(inode, &num_fcntl_locks,
+ 					 &num_flock_locks);
+ 			rec.v2.flock_len = (2*sizeof(u32) +
+ 					    (num_fcntl_locks+num_flock_locks) *
+ 					    sizeof(struct ceph_filelock));
+ 			unlock_kernel();
+ 
+ 			/* pre-alloc pagelist */
+ 			ceph_pagelist_truncate(pagelist, &trunc_point);
+ 			err = ceph_pagelist_append(pagelist, &rec, reclen);
+ 			if (!err)
+ 				err = ceph_pagelist_reserve(pagelist,
+ 							    rec.v2.flock_len);
+ 
+ 			/* encode locks */
+ 			if (!err) {
+ 				lock_kernel();
+ 				err = ceph_encode_locks(inode,
+ 							pagelist,
+ 							num_fcntl_locks,
+ 							num_flock_locks);
+ 				unlock_kernel();
+ 			}
+ 		} while (err == -ENOSPC);
 +	} else {
 +		err = ceph_pagelist_append(pagelist, &rec, reclen);
  	}
  
  out_free:

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/Kconfig between commit be4f104dfd3b5e3ae262bff607965cfc38027dec
("ceph: select CRYPTO") from Linus' tree and commit
49dde0c8bee346a349862c4a3eddcd58e0d0ae90 ("ceph: factor out libceph from
Ceph file system") from the ceph tree.

Just overlapping additions.  I fixed it up (see below) and can carry the
fix for a while.  (though the adding of "default n" is, I think,
redundant as that is the default anyway).
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc fs/ceph/Kconfig
index 0fcd264,73a7b31..0000000
--- a/fs/ceph/Kconfig
+++ b/fs/ceph/Kconfig
@@@ -1,9 -1,10 +1,11 @@@
  config CEPH_FS
          tristate "Ceph distributed file system (EXPERIMENTAL)"
  	depends on INET && EXPERIMENTAL
+ 	select CEPH_LIB
  	select LIBCRC32C
  	select CRYPTO_AES
 +	select CRYPTO
+ 	default n
  	help
  	  Choose Y or M here to include support for mounting the
  	  experimental Ceph distributed file system.  Ceph is an extremely

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in net/Makefile
between commit 1a4240f4764ac78adbf4b0ebb49b3bd8c72ffa11 ("DNS: Separate
out CIFS DNS Resolver code") from Linus' tree and commit
28419a6b287727e879aa94f2f8662a3810d14182 ("ceph: factor out libceph from
Ceph file system") from the ceph tree.

Just overlapping additions.  I fixed it up (see below) and can carry the
fix for a while.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc net/Makefile
index ea60fbc,873c2ae..0000000
--- a/net/Makefile
+++ b/net/Makefile
@@@ -67,4 -67,4 +67,5 @@@ ifeq ($(CONFIG_NET),y
  obj-$(CONFIG_SYSCTL)		+= sysctl_net.o
  endif
  obj-$(CONFIG_WIMAX)		+= wimax/
 +obj-$(CONFIG_DNS_RESOLVER)	+= dns_resolver/
+ obj-$(CONFIG_CEPH_LIB)		+= ceph/

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 393 bytes --]

Hi Sage,

On Thu, 13 May 2010 21:25:45 -0700 (PDT) Sage Weil <sage@newdream.net> wrote:
>
> Aie, sorry...  I rebased everything queued for the next window yesterday 
> but forgot to update the for-next tag.  These conflicts should all 
> disappear next time around.

Thanks.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

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

[Sage Weil]

Hi Stephen,

On Fri, 14 May 2010, Stephen Rothwell wrote:
> Hi Sage,
> 
> Today's linux-next merge of the ceph tree got a conflict in
> fs/ceph/super.c between commit 5dfc589a8467470226feccdc50f1b32713318e7b
> ("ceph: unregister bdi before kill_anon_super releases device name") from
> Linus' tree and commits 9f2a2a1faa0fd990d4930605583a67d7dff28bff ("ceph:
> use ceph_sb_to_client instead of ceph_client") and
> c72b09ba224e6f2c00bcf810a0a62059fcf405f7 ("ceph: name bdi ceph-%d instead
> of major:minor") from the ceph tree.

Aie, sorry...  I rebased everything queued for the next window yesterday 
but forgot to update the for-next tag.  These conflicts should all 
disappear next time around.

Thanks!
sage

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/messenger.c between commit
45c6ceb547ad2d98215351974a4686bf8cb13e14 ("ceph: zero unused message
header, footer fields") from Linus' tree and commit
c46c4a323cad3fb2678a6b9ffd200b7640b0dc61 ("ceph: simplify ceph_msg_new")
from the ceph tree.

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

diff --cc fs/ceph/messenger.c
index cd4fadb,402d9ee..0000000
--- a/fs/ceph/messenger.c
+++ b/fs/ceph/messenger.c
@@@ -1406,15 -1380,12 +1408,13 @@@ static int read_partial_message(struct 
  			con->in_base_pos = -front_len - middle_len - data_len -
  				sizeof(m->footer);
  			con->in_tag = CEPH_MSGR_TAG_READY;
 +			con->in_seq++;
  			return 0;
  		}
- 		if (IS_ERR(con->in_msg)) {
- 			ret = PTR_ERR(con->in_msg);
- 			con->in_msg = NULL;
+ 		if (!con->in_msg) {
  			con->error_msg =
  				"error allocating memory for incoming message";
- 			return ret;
+ 			return -ENOMEM;
  		}
  		m = con->in_msg;
  		m->front.iov_len = 0;    /* haven't read it yet */
@@@ -2094,15 -2061,12 +2093,15 @@@ struct ceph_msg *ceph_msg_new(int type
  	kref_init(&m->kref);
  	INIT_LIST_HEAD(&m->list_head);
  
 +	m->hdr.tid = 0;
  	m->hdr.type = cpu_to_le16(type);
 +	m->hdr.priority = cpu_to_le16(CEPH_MSG_PRIO_DEFAULT);
 +	m->hdr.version = 0;
  	m->hdr.front_len = cpu_to_le32(front_len);
  	m->hdr.middle_len = 0;
- 	m->hdr.data_len = cpu_to_le32(page_len);
- 	m->hdr.data_off = cpu_to_le16(page_off);
+ 	m->hdr.data_len = 0;
+ 	m->hdr.data_off = 0;
 -	m->hdr.priority = cpu_to_le16(CEPH_MSG_PRIO_DEFAULT);
 +	m->hdr.reserved = 0;
  	m->footer.front_crc = 0;
  	m->footer.middle_crc = 0;
  	m->footer.data_crc = 0;

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/caps.c between commit f818a73674c5d197f66b636a46d7d578d7258129
("ceph: fix cap removal races") from Linus' tree and commit
9f2a2a1faa0fd990d4930605583a67d7dff28bff ("ceph: use ceph_sb_to_client
instead of ceph_client") from the ceph tree.

Just context changes.  I fixed it up (see below) and can carry the fix
for a while.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc fs/ceph/caps.c
index d940053,1747db0..0000000
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@@ -867,8 -864,8 +867,9 @@@ void __ceph_remove_cap(struct ceph_cap 
  {
  	struct ceph_mds_session *session = cap->session;
  	struct ceph_inode_info *ci = cap->ci;
- 	struct ceph_mds_client *mdsc = &ceph_client(ci->vfs_inode.i_sb)->mdsc;
+ 	struct ceph_mds_client *mdsc =
+ 		&ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
 +	int removed = 0;
  
  	dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);
  

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
fs/ceph/super.c between commit 5dfc589a8467470226feccdc50f1b32713318e7b
("ceph: unregister bdi before kill_anon_super releases device name") from
Linus' tree and commits 9f2a2a1faa0fd990d4930605583a67d7dff28bff ("ceph:
use ceph_sb_to_client instead of ceph_client") and
c72b09ba224e6f2c00bcf810a0a62059fcf405f7 ("ceph: name bdi ceph-%d instead
of major:minor") from the ceph tree.

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

diff --cc fs/ceph/super.c
index 110857b,30fa891..0000000
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@@ -8,14 -8,10 +8,11 @@@
  #include <linux/module.h>
  #include <linux/mount.h>
  #include <linux/parser.h>
- #include <linux/rwsem.h>
  #include <linux/sched.h>
  #include <linux/seq_file.h>
 +#include <linux/slab.h>
  #include <linux/statfs.h>
  #include <linux/string.h>
- #include <linux/version.h>
- #include <linux/vmalloc.h>
  
  #include "decode.h"
  #include "super.h"
@@@ -893,9 -910,8 +921,10 @@@ static int ceph_register_bdi(struct sup
  		client->backing_dev_info.ra_pages =
  			(client->mount_args->rsize + PAGE_CACHE_SIZE - 1)
  			>> PAGE_SHIFT;
- 	err = bdi_register_dev(&client->backing_dev_info, sb->s_dev);
+ 	err = bdi_register(&client->backing_dev_info, NULL, "ceph-%d",
+ 			   atomic_long_inc_return(&bdi_seq));
 +	if (!err)
 +		sb->s_bdi = &client->backing_dev_info;
  	return err;
  }
  

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

[Stephen Rothwell]

Hi Sage,

Today's linux-next merge of the ceph tree got a conflict in
Documentation/ioctl/ioctl-number.txt between commit
cb5a8b2c92febbed57126e1b8416dfd7607ff03d ("docs: large update to
ioctl-number.txt") from Linus' tree and commit
8f4e91dee2a245e4be6942f4a8d83a769e13a47d ("ceph: ioctls") from the ceph
tree.

Just overlapping additions - not a big deal.  I fixed it up (see below)
and can carry the fix as necessary.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc Documentation/ioctl/ioctl-number.txt
index 35cf64d,91cfdd7..0000000
--- a/Documentation/ioctl/ioctl-number.txt
+++ b/Documentation/ioctl/ioctl-number.txt
@@@ -289,9 -181,8 +289,10 @@@ Code  Seq#(hex)	Include File		Comment
  0x8C	00-3F				WiNRADiO driver
  					<http://www.proximity.com.au/~brian/winradio/>
  0x90	00	drivers/cdrom/sbpcd.h
 +0x92	00-0F	drivers/usb/mon/mon_bin.c
  0x93	60-7F	linux/auto_fs.h
 +0x94	all	fs/btrfs/ioctl.h
+ 0x97    00-7F   fs/ceph/ioctl.h         Ceph file system
  0x99	00-0F				537-Addinboard driver
  					<mailto:buk@buks.ipn.de>
  0xA0	all	linux/sdp/sdp.h		Industrial Device Project