linux-kernel - Re: [PATCH 1/6] nfsd: drop inode parameter from nfsd4_change

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a70e34202c13ef2ac69f04a153268950b485781a.camel@kernel.org>
Date: Tue, 15 Oct 2024 14:42:10 -0400
From: Jeff Layton <jlayton@...nel.org>
To: Chuck Lever <chuck.lever@...cle.com>
Cc: Neil Brown <neilb@...e.de>, Olga Kornievskaia <okorniev@...hat.com>, Dai
 Ngo <Dai.Ngo@...cle.com>, Tom Talpey <tom@...pey.com>, Jonathan Corbet
 <corbet@....net>,  Trond Myklebust <trondmy@...nel.org>, Anna Schumaker
 <anna@...nel.org>, Thomas Haynes <loghyr@...il.com>, 
 linux-nfs@...r.kernel.org, linux-kernel@...r.kernel.org, 
 linux-doc@...r.kernel.org
Subject: Re: [PATCH 1/6] nfsd: drop inode parameter from
 nfsd4_change_attribute()

On Tue, 2024-10-15 at 14:29 -0400, Chuck Lever wrote:
> Hey Jeff -
> 
> On Mon, Oct 14, 2024 at 03:26:49PM -0400, Jeff Layton wrote:
> > Fix up nfs4_delegation_stat() to fetch STATX_MODE,
> 
> The patch description isn't clear about why this change is needed.
> After reading through the other patches, I'm not sure I'm any more
> enlightened about it ;-)
> 

My original thinking was that this patch was just a cleanup and
simplification, but now that I look, I think the inode we were passing
to this function in nfs4_open_delegation was wrong and that could throw
off the result in some cases.

Maybe we should add a Fixes: tag for this:

    bf92e5008b17 nfsd: fix initial getattr on write delegation

...since that should have fixed this call as well.

> 
> > and then drop the
> > inode parameter from nfsd4_change_attribute(), since it's no longer
> > needed.
> 
> Since nfsd4_change_attribute() expects @stat to be filled in by the
> caller, it needs a kdoc-style comment that documents that part of
> the API contract.
> 

Agreed. It needs STATX_MODE and STATX_CTIME. It can also make use of
STATX_CHANGE_COOKIE if present.

> I can add one when applying this patch, unless you would like to
> resend this one or send me something to squash into this change.
> 

That sounds great. A kdoc comment over that is a good idea.

> 
> > Signed-off-by: Jeff Layton <jlayton@...nel.org>
> > ---
> >  fs/nfsd/nfs4state.c |  5 ++---
> >  fs/nfsd/nfs4xdr.c   |  2 +-
> >  fs/nfsd/nfsfh.c     | 11 ++++-------
> >  fs/nfsd/nfsfh.h     |  3 +--
> >  4 files changed, 8 insertions(+), 13 deletions(-)
> > 
> > diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> > index d753926db09eedf629fc3e0938f10b1a6fdb0245..2961a277a79c1f4cdb8c29a7c19abcb3305b61a1 100644
> > --- a/fs/nfsd/nfs4state.c
> > +++ b/fs/nfsd/nfs4state.c
> > @@ -5953,7 +5953,7 @@ nfs4_delegation_stat(struct nfs4_delegation *dp, struct svc_fh *currentfh,
> >  	path.dentry = file_dentry(nf->nf_file);
> >  
> >  	rc = vfs_getattr(&path, stat,
> > -			 (STATX_SIZE | STATX_CTIME | STATX_CHANGE_COOKIE),
> > +			 (STATX_MODE | STATX_SIZE | STATX_CTIME | STATX_CHANGE_COOKIE),
> >  			 AT_STATX_SYNC_AS_STAT);
> >  
> >  	nfsd_file_put(nf);
> > @@ -6037,8 +6037,7 @@ nfs4_open_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
> >  		}
> >  		open->op_delegate_type = NFS4_OPEN_DELEGATE_WRITE;
> >  		dp->dl_cb_fattr.ncf_cur_fsize = stat.size;
> > -		dp->dl_cb_fattr.ncf_initial_cinfo =
> > -			nfsd4_change_attribute(&stat, d_inode(currentfh->fh_dentry));
> > +		dp->dl_cb_fattr.ncf_initial_cinfo = nfsd4_change_attribute(&stat);
> >  		trace_nfsd_deleg_write(&dp->dl_stid.sc_stateid);
> >  	} else {
> >  		open->op_delegate_type = NFS4_OPEN_DELEGATE_READ;
> > diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
> > index 6286ad2afa069f5274ffa352209b7d3c8c577dac..da7ec663da7326ad5c68a9c738b12d09cfcdc65a 100644
> > --- a/fs/nfsd/nfs4xdr.c
> > +++ b/fs/nfsd/nfs4xdr.c
> > @@ -3621,7 +3621,7 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
> >  		args.change_attr = ncf->ncf_initial_cinfo;
> >  		nfs4_put_stid(&dp->dl_stid);
> >  	} else {
> > -		args.change_attr = nfsd4_change_attribute(&args.stat, d_inode(dentry));
> > +		args.change_attr = nfsd4_change_attribute(&args.stat);
> >  	}
> >  
> >  	if (err)
> > diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
> > index 4c5deea0e9535f2b197aa6ca1786d61730d53c44..453b7b52317d538971ce41f7e0492e5ab28b236d 100644
> > --- a/fs/nfsd/nfsfh.c
> > +++ b/fs/nfsd/nfsfh.c
> > @@ -670,20 +670,18 @@ fh_update(struct svc_fh *fhp)
> >  __be32 __must_check fh_fill_pre_attrs(struct svc_fh *fhp)
> >  {
> >  	bool v4 = (fhp->fh_maxsize == NFS4_FHSIZE);
> > -	struct inode *inode;
> >  	struct kstat stat;
> >  	__be32 err;
> >  
> >  	if (fhp->fh_no_wcc || fhp->fh_pre_saved)
> >  		return nfs_ok;
> >  
> > -	inode = d_inode(fhp->fh_dentry);
> >  	err = fh_getattr(fhp, &stat);
> >  	if (err)
> >  		return err;
> >  
> >  	if (v4)
> > -		fhp->fh_pre_change = nfsd4_change_attribute(&stat, inode);
> > +		fhp->fh_pre_change = nfsd4_change_attribute(&stat);
> >  
> >  	fhp->fh_pre_mtime = stat.mtime;
> >  	fhp->fh_pre_ctime = stat.ctime;
> > @@ -700,7 +698,6 @@ __be32 __must_check fh_fill_pre_attrs(struct svc_fh *fhp)
> >  __be32 fh_fill_post_attrs(struct svc_fh *fhp)
> >  {
> >  	bool v4 = (fhp->fh_maxsize == NFS4_FHSIZE);
> > -	struct inode *inode = d_inode(fhp->fh_dentry);
> >  	__be32 err;
> >  
> >  	if (fhp->fh_no_wcc)
> > @@ -716,7 +713,7 @@ __be32 fh_fill_post_attrs(struct svc_fh *fhp)
> >  	fhp->fh_post_saved = true;
> >  	if (v4)
> >  		fhp->fh_post_change =
> > -			nfsd4_change_attribute(&fhp->fh_post_attr, inode);
> > +			nfsd4_change_attribute(&fhp->fh_post_attr);
> >  	return nfs_ok;
> >  }
> >  
> > @@ -824,13 +821,13 @@ enum fsid_source fsid_source(const struct svc_fh *fhp)
> >   * assume that the new change attr is always logged to stable storage in some
> >   * fashion before the results can be seen.
> >   */
> > -u64 nfsd4_change_attribute(const struct kstat *stat, const struct inode *inode)
> > +u64 nfsd4_change_attribute(const struct kstat *stat)
> >  {
> >  	u64 chattr;
> >  
> >  	if (stat->result_mask & STATX_CHANGE_COOKIE) {
> >  		chattr = stat->change_cookie;
> > -		if (S_ISREG(inode->i_mode) &&
> > +		if (S_ISREG(stat->mode) &&
> >  		    !(stat->attributes & STATX_ATTR_CHANGE_MONOTONIC)) {
> >  			chattr += (u64)stat->ctime.tv_sec << 30;
> >  			chattr += stat->ctime.tv_nsec;
> > diff --git a/fs/nfsd/nfsfh.h b/fs/nfsd/nfsfh.h
> > index 5b7394801dc4270dbd5236f3e2f2237130c73dad..876152a91f122f83fb94ffdfb0eedf8fca56a20c 100644
> > --- a/fs/nfsd/nfsfh.h
> > +++ b/fs/nfsd/nfsfh.h
> > @@ -297,8 +297,7 @@ static inline void fh_clear_pre_post_attrs(struct svc_fh *fhp)
> >  	fhp->fh_pre_saved = false;
> >  }
> >  
> > -u64 nfsd4_change_attribute(const struct kstat *stat,
> > -			   const struct inode *inode);
> > +u64 nfsd4_change_attribute(const struct kstat *stat);
> >  __be32 __must_check fh_fill_pre_attrs(struct svc_fh *fhp);
> >  __be32 fh_fill_post_attrs(struct svc_fh *fhp);
> >  __be32 __must_check fh_fill_both_attrs(struct svc_fh *fhp);
> > 
> > -- 
> > 2.47.0
> > 
> 

-- 
Jeff Layton <jlayton@...nel.org>