All of lore.kernel.org
 help / color / mirror / Atom feed
From: Anna Schumaker <Anna.Schumaker@netapp.com>
To: <andros@netapp.com>
Cc: <bfieldses.org@netapp.com>, <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH Version 3 06/16] SUNRPC AUTH_GSS gss3 reply verifier
Date: Wed, 4 Jan 2017 16:46:54 -0500	[thread overview]
Message-ID: <12835369-fb6b-d55a-124f-c70e58e529c1@Netapp.com> (raw)
In-Reply-To: <1482509068-24516-7-git-send-email-andros@netapp.com>

Hi Andy,

On 12/23/2016 11:04 AM, andros@netapp.com wrote:
> From: Andy Adamson <andros@netapp.com>
> 
> The new GSS Version 3 reply  verifier is taken over the same data as
> the call verifier, caveat REPLY direction
> 
> Verifier Data
> 
>    xid          tk_rqstp->rq_xid
>    direction    REPLY (always a 1) RPC_REPLY
>    rpcvers      RPC_VERSION
>    prog         clnt->cl_prog
>    vers         clnt->cl_vers
>    proc         tk_msg.rpc_proc->p_proc
>    credential
>          flavor       RPC_AUTH_GSS
>          length       cred_len is in gss_marshal (new gv_crlen)
>          gss version  ctx->gc_v
>          gss proc     ctx->gv_proc
>          gss seq      tk_rqstp->rq_seqno
>          gss svc      gss_cred->gc_service
>          gss ctx len  ctx->gc_wire_ctx
>          gss ctx data ctx->gc_wire_ctx
> 
> Signed-off-by: Andy Adamson <andros@netapp.com>
> ---
>  net/sunrpc/auth_gss/auth_gss.c | 60 ++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 58 insertions(+), 2 deletions(-)
> 
> diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
> index 9288cc2..d11f421 100644
> --- a/net/sunrpc/auth_gss/auth_gss.c
> +++ b/net/sunrpc/auth_gss/auth_gss.c
> @@ -1624,6 +1624,53 @@ static int gss_cred_is_negative_entry(struct rpc_cred *cred)
>  	return 0;
>  }
>  
> +/**
> + * gss3_reply_verifier: The new gssv3 verifier uses same data as call
> + * caveat REPLY direction - see rpc_encode_header
> + */
> +static  void *
> +gss3_reply_verifier(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
> +		    struct rpc_task *task, __be32 *seq, struct kvec *iov)
> +{
> +	struct gss_cred *g_cred = container_of(cred, struct gss_cred, gc_base);
> +	void	*gss3_buf = NULL;
> +	__be32 *crlen, *ptr = NULL;
> +	int len;
> +
> +	/* freed in gss_validate */
> +	len = (13 * 4) + ctx->gc_wire_ctx.len;
> +	gss3_buf = kmalloc(len, GFP_NOFS);
> +	if (!gss3_buf) {
> +		gss3_buf = ERR_PTR(-EIO);
> +		goto out;
> +	}
> +	ptr = (__be32 *)gss3_buf;
> +
> +	*ptr++ = htonl(task->tk_rqstp->rq_xid);
> +	*ptr++ = htonl(RPC_REPLY);
> +	*ptr++ = htonl(RPC_VERSION);
> +	*ptr++ = htonl(task->tk_client->cl_prog);
> +	*ptr++ = htonl(task->tk_client->cl_vers);
> +	*ptr++ = htonl(task->tk_msg.rpc_proc->p_proc);
> +	*ptr++ = htonl(RPC_AUTH_GSS);
> +
> +	/* credential */
> +	crlen = ptr++;
> +	*ptr++ = htonl(ctx->gc_v);
> +	*ptr++ = htonl(ctx->gc_proc);
> +	*ptr++ = *seq;
> +	*ptr++ = htonl(g_cred->gc_service);
> +	ptr = xdr_encode_netobj(ptr, &ctx->gc_wire_ctx);
> +
> +	/* backfill cred length */
> +	*crlen = htonl((ptr - (crlen + 1)) << 2);
> +
> +	iov->iov_base = gss3_buf;
> +	iov->iov_len = (ptr - (__be32 *)gss3_buf) << 2;
> +out:
> +	return gss3_buf;
> +}
> +
>  static __be32 *
>  gss_validate(struct rpc_task *task, __be32 *p)
>  {
> @@ -1633,6 +1680,7 @@ static int gss_cred_is_negative_entry(struct rpc_cred *cred)
>  	struct kvec	iov;
>  	struct xdr_buf	verf_buf;
>  	struct xdr_netobj mic;
> +	void	*g3_buf = NULL;
>  	u32		flav,len;
>  	u32		maj_stat;
>  	__be32		*ret = ERR_PTR(-EIO);
> @@ -1648,14 +1696,22 @@ static int gss_cred_is_negative_entry(struct rpc_cred *cred)
>  	if (!seq)
>  		goto out_bad;
>  	*seq = htonl(task->tk_rqstp->rq_seqno);
> -	iov.iov_base = seq;
> -	iov.iov_len = 4;
> +	if (ctx->gc_v == RPC_GSS_VERSION) {
> +		iov.iov_base = seq;
> +		iov.iov_len = 4;
> +	}
> +	if (ctx->gc_v == RPC_GSS3_VERSION) {

Can this be written as an else-if instead?  I don't think it's likely for gc_v to have two values at once :)

Thanks,
Anna

> +		g3_buf = gss3_reply_verifier(cred, ctx, task, seq, &iov);
> +		if (IS_ERR(g3_buf))
> +			goto out_bad;
> +	}
>  	xdr_buf_from_iov(&iov, &verf_buf);
>  	mic.data = (u8 *)p;
>  	mic.len = len;
>  
>  	ret = ERR_PTR(-EACCES);
>  	maj_stat = gss_verify_mic(ctx->gc_gss_ctx, &verf_buf, &mic);
> +	kfree(g3_buf);
>  	if (maj_stat == GSS_S_CONTEXT_EXPIRED)
>  		clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
>  	if (maj_stat) {
> 

  reply	other threads:[~2017-01-04 21:47 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-12-23 16:04 [PATCH Version 3 00/16] RFC: RPCSEC_GSS Version 3 prototype: Full Mode MAC andros
2016-12-23 16:04 ` [PATCH Version 3 01/16] SUNRPC handle unsupported RPC_GSS_SVC_CHANNEL_PROT andros
2016-12-23 18:31   ` kbuild test robot
2017-01-04 21:11   ` Anna Schumaker
2016-12-23 16:04 ` [PATCH Version 3 02/16] SUNRPC: add a null call with payload GSSv3 andros
2016-12-23 16:04 ` [PATCH Version 3 03/16] SELINUX export security_current_sid_to_context andros
2016-12-23 16:04 ` [PATCH Version 3 04/16] SUNRPC GSSv3: base definitions andros
2016-12-23 16:04 ` [PATCH Version 3 05/16] SUNRPC AUTH_GSS get RPCSEC_GSS version from gssd downcall andros
2016-12-23 16:04 ` [PATCH Version 3 06/16] SUNRPC AUTH_GSS gss3 reply verifier andros
2017-01-04 21:46   ` Anna Schumaker [this message]
2016-12-23 16:04 ` [PATCH Version 3 07/16] SUNRPC AUTH_GSS RPCSEC_GSS_CREATE with label payload andros
2016-12-23 18:01   ` kbuild test robot
2017-01-04 21:51   ` Anna Schumaker
2016-12-23 16:04 ` [PATCH Version 3 08/16] SUNRPC AUTH_GSS store and use gss3 label assertion andros
2016-12-23 16:04 ` [PATCH Version 3 09/16] SUNRPC AUTH_GSS free assertions andros
2016-12-23 16:04 ` [PATCH Version 3 10/16] SUNRPC: AUTH_GSS add RPC_GSS_PROC_CREATE case for wrap and unwrap andros
2016-12-23 16:04 ` [PATCH Version 3 11/16] SUNRPC SVCAUTH_GSS reap the rsc cache entry on RPC_GSS_PROC_DESTROY andros
2016-12-23 16:04 ` [PATCH Version 3 12/16] SUNRPC SVCAUTH_GSS allow RPCSEC_GSS version 1 or 3 andros
2016-12-23 16:04 ` [PATCH Version 3 13/16] SUNRPC SVCAUTH_GSS gss3 reply verifier andros
2016-12-23 16:04 ` [PATCH Version 3 14/16] SUNRPC SVCAUTH_GSS gss3 create label andros
2016-12-23 16:04 ` [PATCH Version 3 15/16] SUNRPC SVCAUTH_GSS set gss3 label on nfsd thread andros
2016-12-23 16:04 ` [PATCH Version 3 16/16] SUNRPC SVCAUTH_gss store gss3 child handles in parent rsc andros

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=12835369-fb6b-d55a-124f-c70e58e529c1@Netapp.com \
    --to=anna.schumaker@netapp.com \
    --cc=andros@netapp.com \
    --cc=bfieldses.org@netapp.com \
    --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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.