public inbox for linux-nfs@vger.kernel.org
 help / color / mirror / Atom feed
From: NeilBrown <neilb@suse.com>
To: "J. Bruce Fields" <bfields@fieldses.org>,
	Chuck Lever <chuck.lever@oracle.com>,
	Jeff Layton <jlayton@kernel.org>,
	Trond Myklebust <trond.myklebust@hammerspace.com>,
	Anna Schumaker <anna.schumaker@netapp.com>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH 00/23 - V4] NFS: Remove generic RPC credentials.
Date: Fri, 30 Nov 2018 10:19:05 +1100	[thread overview]
Message-ID: <87sgzjo4k6.fsf@notabene.neil.brown.name> (raw)
In-Reply-To: <154156285766.24086.14262073575778354276.stgit@noble>

[-- Attachment #1: Type: text/plain, Size: 5956 bytes --]


Hi,
 has anyone else had a chance to look at this yet?

 There is a small conflict now due to
        SUNRPC: Fix a bogus get/put in generic_key_to_expire()
 Should I resend with that fixed up?

Thanks,
NeilBrown


On Wed, Nov 07 2018, NeilBrown wrote:

> This is an updated version of a series I sent in Feb of this year.
> Since then there have only been minor improvement and updates to sync
> with the changing kernel.
>
> There doesn't seem to be a maintainer for the 'cred' code, so I don't
> know who to ask to approve the first 4 patches.  Maybe if the NFS
> team like them, they can just go to Linus with a note for him to look
> at them if he wants to.
>
> The original motivation for this was performance.  In some
> circumstances the cred caches can get big and particularly can get
> long chains.  The hash function has been changed at least once to
> improve the hashing and it still isn't perfect.
> Rather than improving pruning of the cache, or resizing the hashtable
> etc, it is easiest to just get rid of it.
>
> As well as discarding generic credentials completely (using 'struct
> cred' instead), we also stop storing AUTH_UNIX credentials in a hash
> table - that brings no value.  Just allocate as needed and discard
> when finished with.
> So the only hash table will still have is for AUTH_GSS.
> One of the main triggers for hashtable problems was users changing
> groups a lot, so there would be many entries for the one user, each
> with a different set of groups.  That doesn't apply for
> AUTH_GSS as the groupids on the client are ignored.
>
> That was the original motivation, but as I worked on it, I realized
> that it was making a log of code simpler.
>
>  44 files changed, 550 insertions(+), 925 deletions(-)
>
> That is sufficient motivation in itself I think.
>
> Review comments most welcome.
>
> Thanks,
> NeilBrown
>
>
> ---
>
> NeilBrown (23):
>       cred: add cred_fscmp() for comparing creds.
>       cred: add get_cred_rcu()
>       cred: export get_task_cred().
>       cred: allow get_cred() and put_cred() to be given NULL.
>       SUNRPC: add 'struct cred *' to auth_cred and rpc_cred
>       SUNRPC: remove groupinfo from struct auth_cred.
>       SUNRPC: remove uid and gid from struct auth_cred
>       SUNRPC: remove machine_cred field from struct auth_cred
>       NFSv4: add cl_root_cred for use when machine cred is not available.
>       NFSv4: don't require lock for get_renew_cred or get_machine_cred
>       SUNRPC: discard RPC_DO_ROOTOVERRIDE()
>       NFS/SUNRPC: don't lookup machine credential until rpcauth_bindcred().
>       SUNRPC: introduce RPC_TASK_NULLCREDS to request auth_none
>       SUNRPC: add side channel to use non-generic cred for rpc call.
>       NFS: move credential expiry tracking out of SUNRPC into NFS.
>       SUNRPC: remove RPCAUTH_AUTH_NO_CRKEY_TIMEOUT
>       NFS: change access cache to use 'struct cred'.
>       NFS: struct nfs_open_dir_context: convert rpc_cred pointer to cred.
>       NFS/NFSD/SUNRPC: replace generic creds with 'struct cred'.
>       SUNRPC: remove generic cred code.
>       SUNRPC: remove crbind rpc_cred operation
>       SUNRPC: simplify auth_unix.
>       SUNRPC discard cr_uid from struct rpc_cred.
>
>
>  fs/lockd/clntproc.c                       |    6 -
>  fs/nfs/blocklayout/blocklayout.c          |    2 
>  fs/nfs/client.c                           |    9 -
>  fs/nfs/delegation.c                       |   28 +--
>  fs/nfs/delegation.h                       |   10 -
>  fs/nfs/dir.c                              |   59 ++----
>  fs/nfs/flexfilelayout/flexfilelayout.c    |   64 +++---
>  fs/nfs/flexfilelayout/flexfilelayout.h    |    8 -
>  fs/nfs/flexfilelayout/flexfilelayoutdev.c |   16 +-
>  fs/nfs/inode.c                            |   13 +
>  fs/nfs/internal.h                         |    8 -
>  fs/nfs/nfs3proc.c                         |    4 
>  fs/nfs/nfs4_fs.h                          |   65 +++---
>  fs/nfs/nfs4client.c                       |    4 
>  fs/nfs/nfs4proc.c                         |  150 +++++++--------
>  fs/nfs/nfs4renewd.c                       |    9 -
>  fs/nfs/nfs4session.c                      |    5 
>  fs/nfs/nfs4state.c                        |  129 ++++++-------
>  fs/nfs/pagelist.c                         |    2 
>  fs/nfs/pnfs.c                             |   14 +
>  fs/nfs/pnfs.h                             |   10 -
>  fs/nfs/pnfs_dev.c                         |    4 
>  fs/nfs/pnfs_nfs.c                         |    2 
>  fs/nfs/proc.c                             |    2 
>  fs/nfs/unlink.c                           |   15 -
>  fs/nfs/write.c                            |   24 ++
>  fs/nfsd/nfs4callback.c                    |   31 +--
>  fs/nfsd/state.h                           |    2 
>  include/linux/cred.h                      |   26 ++-
>  include/linux/nfs_fs.h                    |   13 +
>  include/linux/nfs_fs_sb.h                 |    2 
>  include/linux/nfs_xdr.h                   |   16 +-
>  include/linux/sunrpc/auth.h               |   51 -----
>  include/linux/sunrpc/clnt.h               |    1 
>  include/linux/sunrpc/sched.h              |    6 -
>  kernel/cred.c                             |   58 ++++++
>  net/sunrpc/Makefile                       |    2 
>  net/sunrpc/auth.c                         |  116 ++++++-----
>  net/sunrpc/auth_generic.c                 |  299 -----------------------------
>  net/sunrpc/auth_gss/auth_gss.c            |   45 +---
>  net/sunrpc/auth_null.c                    |    4 
>  net/sunrpc/auth_unix.c                    |  110 +++--------
>  net/sunrpc/clnt.c                         |   26 +--
>  net/sunrpc/sched.c                        |    5 
>  44 files changed, 550 insertions(+), 925 deletions(-)
>  delete mode 100644 net/sunrpc/auth_generic.c
>
> --
> Signature

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

  parent reply	other threads:[~2018-11-29 23:19 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-07  4:12 [PATCH 00/23 - V4] NFS: Remove generic RPC credentials NeilBrown
2018-11-07  4:12 ` [PATCH 01/23] cred: add cred_fscmp() for comparing creds NeilBrown
2018-11-07  4:12 ` [PATCH 02/23] cred: add get_cred_rcu() NeilBrown
2018-11-07  4:12 ` [PATCH 05/23] SUNRPC: add 'struct cred *' to auth_cred and rpc_cred NeilBrown
2018-11-07  4:12 ` [PATCH 06/23] SUNRPC: remove groupinfo from struct auth_cred NeilBrown
2018-11-07  4:12 ` [PATCH 03/23] cred: export get_task_cred() NeilBrown
2018-11-07  4:12 ` [PATCH 04/23] cred: allow get_cred() and put_cred() to be given NULL NeilBrown
2018-11-07  4:12 ` [PATCH 14/23] SUNRPC: add side channel to use non-generic cred for rpc call NeilBrown
2018-11-07  4:12 ` [PATCH 08/23] SUNRPC: remove machine_cred field from struct auth_cred NeilBrown
2018-11-07  4:12 ` [PATCH 10/23] NFSv4: don't require lock for get_renew_cred or get_machine_cred NeilBrown
2018-11-07  4:12 ` [PATCH 22/23] SUNRPC: simplify auth_unix NeilBrown
2018-11-07 15:19   ` Chuck Lever
2018-11-08  1:41     ` NeilBrown
2018-11-08 15:54       ` Chuck Lever
2018-11-09  0:45         ` NeilBrown
2018-11-07  4:12 ` [PATCH 20/23] SUNRPC: remove generic cred code NeilBrown
2018-11-07  4:12 ` [PATCH 16/23] SUNRPC: remove RPCAUTH_AUTH_NO_CRKEY_TIMEOUT NeilBrown
2018-11-07  4:12 ` [PATCH 18/23] NFS: struct nfs_open_dir_context: convert rpc_cred pointer to cred NeilBrown
2018-11-07  4:12 ` [PATCH 19/23] NFS/NFSD/SUNRPC: replace generic creds with 'struct cred' NeilBrown
2018-11-07  4:12 ` [PATCH 07/23] SUNRPC: remove uid and gid from struct auth_cred NeilBrown
2018-11-07  4:12 ` [PATCH 17/23] NFS: change access cache to use 'struct cred' NeilBrown
2018-11-07  4:12 ` [PATCH 09/23] NFSv4: add cl_root_cred for use when machine cred is not available NeilBrown
2018-11-07  4:12 ` [PATCH 15/23] NFS: move credential expiry tracking out of SUNRPC into NFS NeilBrown
2018-11-07  4:12 ` [PATCH 13/23] SUNRPC: introduce RPC_TASK_NULLCREDS to request auth_none NeilBrown
2018-11-07  4:12 ` [PATCH 23/23] SUNRPC discard cr_uid from struct rpc_cred NeilBrown
2018-11-07  4:12 ` [PATCH 21/23] SUNRPC: remove crbind rpc_cred operation NeilBrown
2018-11-07  4:12 ` [PATCH 12/23] NFS/SUNRPC: don't lookup machine credential until rpcauth_bindcred() NeilBrown
2018-11-07  4:12 ` [PATCH 11/23] SUNRPC: discard RPC_DO_ROOTOVERRIDE() NeilBrown
2018-11-29 23:19 ` NeilBrown [this message]
2018-11-30 19:39   ` [PATCH 00/23 - V4] NFS: Remove generic RPC credentials Schumaker, Anna

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=87sgzjo4k6.fsf@notabene.neil.brown.name \
    --to=neilb@suse.com \
    --cc=anna.schumaker@netapp.com \
    --cc=bfields@fieldses.org \
    --cc=chuck.lever@oracle.com \
    --cc=jlayton@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=trond.myklebust@hammerspace.com \
    /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