linux-stable.git/fs/nfs/flexfilelayout, branch linux-5.0.y

NFS/NFSD/SUNRPC: replace generic creds with 'struct cred'.

2018-12-19T18:52:46+00:00

SUNRPC has two sorts of credentials, both of which appear as
"struct rpc_cred".
There are "generic credentials" which are supplied by clients
such as NFS and passed in 'struct rpc_message' to indicate
which user should be used to authorize the request, and there
are low-level credentials such as AUTH_NULL, AUTH_UNIX, AUTH_GSS
which describe the credential to be sent over the wires.

This patch replaces all the generic credentials by 'struct cred'
pointers - the credential structure used throughout Linux.

For machine credentials, there is a special 'struct cred *' pointer
which is statically allocated and recognized where needed as
having a special meaning.  A look-up of a low-level cred will
map this to a machine credential.

Signed-off-by: NeilBrown 
Acked-by: J. Bruce Fields 
Signed-off-by: Anna Schumaker

SUNRPC: remove uid and gid from struct auth_cred

2018-12-19T18:52:45+00:00

Use cred->fsuid and cred->fsgid instead.

Signed-off-by: NeilBrown 
Signed-off-by: Anna Schumaker

SUNRPC: remove groupinfo from struct auth_cred.

2018-12-19T18:52:45+00:00

We can use cred->groupinfo (from the 'struct cred') instead.

Signed-off-by: NeilBrown 
Signed-off-by: Anna Schumaker

SUNRPC: add 'struct cred *' to auth_cred and rpc_cred

2018-12-19T18:52:44+00:00

The SUNRPC credential framework was put together before
Linux has 'struct cred'.  Now that we have it, it makes sense to
use it.
This first step just includes a suitable 'struct cred *' pointer
in every 'struct auth_cred' and almost every 'struct rpc_cred'.

The rpc_cred used for auth_null has a NULL 'struct cred *' as nothing
else really makes sense.

For rpc_cred, the pointer is reference counted.
For auth_cred it isn't.  struct auth_cred are either allocated on
the stack, in which case the thread owns a reference to the auth,
or are part of 'struct generic_cred' in which case gc_base owns the
reference, and "acred" shares it.

Signed-off-by: NeilBrown 
Signed-off-by: Anna Schumaker

flexfiles: enforce per-mirror stateid only for v4 DSes

2018-12-02T14:43:56+00:00

Since commit bb21ce0ad227 we always enforce per-mirror stateid.
However, this makes sense only for v4+ servers.

Signed-off-by: Tigran Mkrtchyan 
Signed-off-by: Trond Myklebust

flexfiles: use per-mirror specified stateid for IO

2018-11-22T19:04:55+00:00

rfc8435 says:

  For tight coupling, ffds_stateid provides the stateid to be used by
  the client to access the file.

However current implementation replaces per-mirror provided stateid with
by open or lock stateid.

Ensure that per-mirror stateid is used by ff_layout_write_prepare_v4 and
nfs4_ff_layout_prepare_ds.

Signed-off-by: Tigran Mkrtchyan 
Signed-off-by: Rick Macklem 
Signed-off-by: Trond Myklebust

nfs4: flex_file: ignore synthetic uid/gid for tightly coupled DSes

2018-09-30T19:35:17+00:00

for tightly coupled DSes client must ignore provided synthetic uid and
gid as stated in draft-ietf-nfsv4-flex-files-19#section-5.1.

Signed-off-by: Tigran Mkrtchyan 
Signed-off-by: Trond Myklebust

pNFS: Don't allocate more pages than we need to fit a layoutget response

2018-09-30T19:35:16+00:00

For the 'files' and 'flexfiles' layout types, we do not expect the reply
to be any larger than 4k. The block and scsi layout types are a little more
greedy, so we keep allocating the maximum response size for now.

Signed-off-by: Trond Myklebust

pNFS/flexfiles: ff_layout_pg_init_read should exit on error

2018-08-21T17:39:05+00:00

If we get an error while retrieving the layout, then we should
report it rather than falling back to I/O through the MDS.

Signed-off-by: Trond Myklebust 
Signed-off-by: Anna Schumaker

pNFS/flexfiles: Ensure we always return a layout if it has layoutstats

2018-07-26T20:25:25+00:00

If a layout segment is carrying layoutstats or layout error information,
then we always want to return it rather than using a forgetful model.

Signed-off-by: Trond Myklebust