From: Chuck Lever <chuck.lever@oracle.com>
To: Sergey Bashirov <sergeybashirov@gmail.com>,
Dan Carpenter <dan.carpenter@linaro.org>,
Jeff Layton <jlayton@kernel.org>, NeilBrown <neil@brown.name>,
Olga Kornievskaia <okorniev@redhat.com>,
Dai Ngo <Dai.Ngo@oracle.com>, Tom Talpey <tom@talpey.com>
Cc: linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] NFSD: Change ret code of xdr_stream_decode_opaque_fixed
Date: Thu, 17 Jul 2025 16:08:07 -0400 [thread overview]
Message-ID: <a7f6e186-ad9c-47eb-88d7-64605d7a7854@oracle.com> (raw)
In-Reply-To: <20250717194838.69200-1-sergeybashirov@gmail.com>
On 7/17/25 3:48 PM, Sergey Bashirov wrote:
> Since the XDR field is fixed in size, the caller already knows how many
> bytes were decoded, on success. Thus, xdr_stream_decode_opaque_fixed()
> doesn't need to return that value. And, xdr_stream_decode_u32 and _u64
> both return zero on success.
>
> This patch also simplifies the error checking to avoid potential integer
> promotion issues.
>
> Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com>
> ---
> fs/nfsd/blocklayoutxdr.c | 2 +-
> include/linux/nfs_xdr.h | 2 +-
> include/linux/sunrpc/xdr.h | 4 ++--
> .../xdrgen/templates/C/pointer/decoder/fixed_length_opaque.j2 | 2 +-
> .../xdrgen/templates/C/struct/decoder/fixed_length_opaque.j2 | 2 +-
> .../xdrgen/templates/C/typedef/decoder/fixed_length_opaque.j2 | 2 +-
> 6 files changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/fs/nfsd/blocklayoutxdr.c b/fs/nfsd/blocklayoutxdr.c
> index 266b2737882e1..2a4dcb428d823 100644
> --- a/fs/nfsd/blocklayoutxdr.c
> +++ b/fs/nfsd/blocklayoutxdr.c
> @@ -157,7 +157,7 @@ nfsd4_block_decode_layoutupdate(struct xdr_stream *xdr, struct iomap **iomapp,
>
> ret = xdr_stream_decode_opaque_fixed(xdr,
> &bex.vol_id, sizeof(bex.vol_id));
> - if (ret < sizeof(bex.vol_id)) {
> + if (ret) {
I've dropped "nfsd: Implement large extent array support in pNFS" from
nfsd-next because of the number of (minor) issues that have been found
so far.
Also, I think this patch needs to come before "Implement large extent",
and so does the forthcoming patch that moves nfsd4_decode_deviceid4().
So, please rebase this one today's nfsd-next.
> nfserr = nfserr_bad_xdr;
> goto fail;
> }
> diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
> index 67f6632f723b4..dd80163e0140c 100644
> --- a/include/linux/nfs_xdr.h
> +++ b/include/linux/nfs_xdr.h
> @@ -1863,7 +1863,7 @@ static inline int decode_opaque_fixed(struct xdr_stream *xdr,
> void *buf, size_t len)
> {
> ssize_t ret = xdr_stream_decode_opaque_fixed(xdr, buf, len);
> - if (unlikely(ret < 0))
> + if (unlikely(ret))
Unless I've misunderstood Dan's bug report, I don't think these
call sites need to change ... The compiler should promote the
naked integer to the proper integer type so that the less-than
comparison is still valid.
> return -EIO;
> return 0;
> }
> diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
> index e3358c630ba18..ffb699a02b17d 100644
> --- a/include/linux/sunrpc/xdr.h
> +++ b/include/linux/sunrpc/xdr.h
> @@ -730,7 +730,7 @@ xdr_stream_decode_u64(struct xdr_stream *xdr, __u64 *ptr)
> * @len: size of buffer pointed to by @ptr
> *
> * Return values:
> - * On success, returns size of object stored in @ptr
> + * %0 on success
> * %-EBADMSG on XDR buffer overflow
> */
> static inline ssize_t
> @@ -741,7 +741,7 @@ xdr_stream_decode_opaque_fixed(struct xdr_stream *xdr, void *ptr, size_t len)
> if (unlikely(!p))
> return -EBADMSG;
> xdr_decode_opaque_fixed(p, ptr, len);
> - return len;
> + return 0;
> }
>
> /**
> diff --git a/tools/net/sunrpc/xdrgen/templates/C/pointer/decoder/fixed_length_opaque.j2 b/tools/net/sunrpc/xdrgen/templates/C/pointer/decoder/fixed_length_opaque.j2
> index b4695ece1884b..ea9295f91aecc 100644
> --- a/tools/net/sunrpc/xdrgen/templates/C/pointer/decoder/fixed_length_opaque.j2
> +++ b/tools/net/sunrpc/xdrgen/templates/C/pointer/decoder/fixed_length_opaque.j2
> @@ -2,5 +2,5 @@
> {% if annotate %}
> /* member {{ name }} (fixed-length opaque) */
> {% endif %}
> - if (xdr_stream_decode_opaque_fixed(xdr, ptr->{{ name }}, {{ size }}) < 0)
> + if (xdr_stream_decode_opaque_fixed(xdr, ptr->{{ name }}, {{ size }}) != 0)
Ditto.
> return false;
> diff --git a/tools/net/sunrpc/xdrgen/templates/C/struct/decoder/fixed_length_opaque.j2 b/tools/net/sunrpc/xdrgen/templates/C/struct/decoder/fixed_length_opaque.j2
> index b4695ece1884b..ea9295f91aecc 100644
> --- a/tools/net/sunrpc/xdrgen/templates/C/struct/decoder/fixed_length_opaque.j2
> +++ b/tools/net/sunrpc/xdrgen/templates/C/struct/decoder/fixed_length_opaque.j2
> @@ -2,5 +2,5 @@
> {% if annotate %}
> /* member {{ name }} (fixed-length opaque) */
> {% endif %}
> - if (xdr_stream_decode_opaque_fixed(xdr, ptr->{{ name }}, {{ size }}) < 0)
> + if (xdr_stream_decode_opaque_fixed(xdr, ptr->{{ name }}, {{ size }}) != 0)
> return false;
Ditto.
> diff --git a/tools/net/sunrpc/xdrgen/templates/C/typedef/decoder/fixed_length_opaque.j2 b/tools/net/sunrpc/xdrgen/templates/C/typedef/decoder/fixed_length_opaque.j2
> index 8b4ff08c49e5e..bdc7bd24ffb13 100644
> --- a/tools/net/sunrpc/xdrgen/templates/C/typedef/decoder/fixed_length_opaque.j2
> +++ b/tools/net/sunrpc/xdrgen/templates/C/typedef/decoder/fixed_length_opaque.j2
> @@ -13,5 +13,5 @@ xdrgen_decode_{{ name }}(struct xdr_stream *xdr, {{ classifier }}{{ name }} *ptr
> {% if annotate %}
> /* (fixed-length opaque) */
> {% endif %}
> - return xdr_stream_decode_opaque_fixed(xdr, ptr, {{ size }}) >= 0;
> + return xdr_stream_decode_opaque_fixed(xdr, ptr, {{ size }}) == 0;
> };
The existing comparison will work fine. But "== 0" will match the new
return values better. LGTM.
--
Chuck Lever
next prev parent reply other threads:[~2025-07-17 20:08 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-17 19:48 Sergey Bashirov
2025-07-17 20:08 ` Chuck Lever [this message]
2025-07-17 20:42 ` Dan Carpenter
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=a7f6e186-ad9c-47eb-88d7-64605d7a7854@oracle.com \
--to=chuck.lever@oracle.com \
--cc=Dai.Ngo@oracle.com \
--cc=dan.carpenter@linaro.org \
--cc=jlayton@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nfs@vger.kernel.org \
--cc=neil@brown.name \
--cc=okorniev@redhat.com \
--cc=sergeybashirov@gmail.com \
--cc=tom@talpey.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
all inboxes | Powered by JetHome®