linux-hardening - Re: [PATCH][next] nfsd: Replace one-element array with flexible-array member

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]

Date:   Tue, 23 May 2023 21:31:04 -0400
From:   Chuck Lever <cel@...nel.org>
To:     "Gustavo A. R. Silva" <gustavo@...eddedor.com>
Cc:     "Gustavo A. R. Silva" <gustavoars@...nel.org>,
        Chuck Lever <chuck.lever@...cle.com>,
        Jeff Layton <jlayton@...nel.org>, linux-nfs@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-hardening@...r.kernel.org,
        dai.ngo@...cle.com
Subject: Re: [PATCH][next] nfsd: Replace one-element array with
 flexible-array member

On Tue, May 23, 2023 at 07:11:37PM -0600, Gustavo A. R. Silva wrote:
> 
> On 5/23/23 19:01, Chuck Lever wrote:
> > On Tue, May 23, 2023 at 06:44:23PM -0600, Gustavo A. R. Silva wrote:
> > > One-element arrays are deprecated, and we are replacing them with
> > > flexible array members instead. So, replace a one-element array
> > > with a flexible-arrayº member in struct vbi_anc_data and refactor
> > 
> > I don't know what "struct vbi_anc_data" is. Is the patch description
> > correct?
> 
> Oops, copy/paste error. I'll fix it up. :)
> 
> > 
> > 
> > > the rest of the code, accordingly.
> > > 
> > > This results in no differences in binary output.
> > > 
> > > Link: https://github.com/KSPP/linux/issues/79
> > > Link: https://github.com/KSPP/linux/issues/298
> > > Signed-off-by: Gustavo A. R. Silva <gustavoars@...nel.org>
> > 
> > > ---
> > >   fs/nfsd/nfs4callback.c | 2 +-
> > >   fs/nfsd/xdr4.h         | 2 +-
> > >   2 files changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
> > > index 4039ffcf90ba..2c688d51135d 100644
> > > --- a/fs/nfsd/nfs4callback.c
> > > +++ b/fs/nfsd/nfs4callback.c
> > > @@ -353,7 +353,7 @@ encode_cb_recallany4args(struct xdr_stream *xdr,
> > >   {
> > >   	encode_nfs_cb_opnum4(xdr, OP_CB_RECALL_ANY);
> > >   	encode_uint32(xdr, ra->ra_keep);
> > > -	encode_bitmap4(xdr, ra->ra_bmval, ARRAY_SIZE(ra->ra_bmval));
> > > +	encode_bitmap4(xdr, ra->ra_bmval, 1);
> > 
> > I find the new code less self-documenting.
> > 
> > 
> > >   	hdr->nops++;
> > >   }
> > > diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h
> > > index 510978e602da..68072170eac8 100644
> > > --- a/fs/nfsd/xdr4.h
> > > +++ b/fs/nfsd/xdr4.h
> > > @@ -899,7 +899,7 @@ struct nfsd4_operation {
> > >   struct nfsd4_cb_recall_any {
> > >   	struct nfsd4_callback	ra_cb;
> > >   	u32			ra_keep;
> > > -	u32			ra_bmval[1];
> > > +	u32			ra_bmval[];
> > 
> > This is not a placeholder for "1 or more elements". We actually want
> > just a single u32 element in this array. Doesn't this change the
> > sizeof(struct nfsd4_cb_recall_any) ?
> 
> I see. Yes, it does change the size. Can we replace it with a simple
> object of type u32? or do you actually need this to stay an array?

It's not impossible to make it a scalar u32, however...

In this area of code, *_bmval is always a bitmap -- an array of u32s.
Helpers like encode_bitmap4() assume an array. I think it would be
less confusing overall to human readers if it remained an array.

In this case, it is a single element array because CB_RECALL_ANY
doesn't happen to use bits above the first 32-bit word of the
bitmap.

We could make it a 2-element array, I think, without harm. Send a
patch for that, and Dai can test it to make sure there are no
unexpected interoperability consequences.

I hope that would avoid suspicious-looking array definitions.


> > >   };
> > >   #endif
> > > -- 
> > > 2.34.1
> > >