Linux NFS development
 help / color / mirror / Atom feed
From: Chuck Lever III <chuck.lever@oracle.com>
To: Jeff Layton <jlayton@kernel.org>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH v1 00/25] Server-side RPC call header parsing overhaul
Date: Tue, 3 Jan 2023 16:33:20 +0000	[thread overview]
Message-ID: <14A17DBB-CD3D-41A4-B2C2-D46DEBB1243C@oracle.com> (raw)
In-Reply-To: <159dd1253c7b49b6654cb4373477f69cc7387b1e.camel@kernel.org>



> On Jan 3, 2023, at 9:52 AM, Jeff Layton <jlayton@kernel.org> wrote:
> 
> On Mon, 2023-01-02 at 12:05 -0500, Chuck Lever wrote:
>> Happy new year, campers.
>> 
>> The following series has been percolating for quite a while, thanks
>> to the extended 6.1-rc cycle. I'd like to get this work reviewed
>> for possible inclusion in v6.3, so I'm starting early.
>> 
>> The purpose of this series is to replace the svc_get* macros in the
>> Linux kernel server's RPC call header parsing code with xdr_stream
>> helpers. I've measured no change in CPU utilization after the
>> overhaul; svc_recv() and friends remain the highest CPU consumers by
>> an order of magnitude.
>> 
>> Memory safety: Buffer bounds checking after decoding each XDR item
>> is more memory-safe than the current decoding mechanism. Subsequent
>> memory safety improvements to the xdr_stream helpers will benefit
>> all who use them.
>> 
>> Audit friendliness: The new code has lots of comments and other
>> clean-up to help align it with the relevant RPC specifications. The
>> use of common helpers also makes the decoders easier to audit.
>> 
>> I've split the full series in half to make it easier to review. The
>> patches posted here handle RPC call header decoding. Remaining
>> patches, to be posted later, deal with RPC reply header encoding.
>> 
>> Yes, there are a lot of patches, but they are each small, easily
>> chewed mechanical changes.
>> 
>> ---
>> 
>> Chuck Lever (25):
>>      SUNRPC: Push svcxdr_init_decode() into svc_process_common()
>>      SUNRPC: Move svcxdr_init_decode() into ->accept methods
>>      SUNRPC: Add an XDR decoding helper for struct opaque_auth
>>      SUNRPC: Convert svcauth_null_accept() to use xdr_stream
>>      SUNRPC: Convert svcauth_unix_accept() to use xdr_stream
>>      SUNRPC: Convert svcauth_tls_accept() to use xdr_stream
>>      SUNRPC: Move the server-side GSS upcall to a noinline function
>>      SUNRPC: Hoist common verifier decoding code into svcauth_gss_proc_init()
>>      SUNRPC: Remove gss_read_common_verf()
>>      SUNRPC: Remove gss_read_verf()
>>      SUNRPC: Convert server-side GSS upcall helpers to use xdr_stream
>>      SUNRPC: Replace read_u32_from_xdr_buf() with existing XDR helper
>>      SUNRPC: Rename automatic variables in unwrap_integ_data()
>>      SUNRPC: Convert unwrap_integ_data() to use xdr_stream
>>      SUNRPC: Rename automatic variables in unwrap_priv_data()
>>      SUNRPC: Convert unwrap_priv_data() to use xdr_stream
>>      SUNRPC: Convert gss_verify_header() to use xdr_stream
>>      SUNRPC: Clean up svcauth_gss_accept's NULL procedure check
>>      SUNRPC: Convert the svcauth_gss_accept() pre-amble to use xdr_stream
>>      SUNRPC: Hoist init_decode out of svc_authenticate()
>>      SUNRPC: Re-order construction of the first reply fields
>>      SUNRPC: Eliminate unneeded variable
>>      SUNRPC: Decode most of RPC header with xdr_stream
>>      SUNRPC: Remove svc_process_common's argv parameter
>>      SUNRPC: Hoist svcxdr_init_decode() into svc_process()
>> 
>> 
>> fs/lockd/svc.c                    |   1 -
>> fs/nfs/callback_xdr.c             |   1 -
>> fs/nfsd/nfssvc.c                  |   1 -
>> include/linux/sunrpc/msg_prot.h   |   5 +
>> include/linux/sunrpc/xdr.h        |   5 +-
>> net/sunrpc/auth_gss/svcauth_gss.c | 512 ++++++++++++++++--------------
>> net/sunrpc/svc.c                  |  69 ++--
>> net/sunrpc/svc_xprt.c             |   1 -
>> net/sunrpc/svcauth.c              |  13 +-
>> net/sunrpc/svcauth_unix.c         | 132 +++++---
>> net/sunrpc/xdr.c                  |  50 ++-
>> 11 files changed, 468 insertions(+), 322 deletions(-)
>> 
>> --
>> Chuck Lever
>> 
> 
> I went through the whole set and it's all a mostly logical set of
> methodical changes. You can add:
> 
> Reviewed-by: Jeff Layton <jlayton@kernel.org>

Thanks Jeff! I've applied these and pushed them to nfsd's for-next,
but comments are still open, and testing results are welcome.


--
Chuck Lever




      reply	other threads:[~2023-01-03 16:33 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-02 17:05 [PATCH v1 00/25] Server-side RPC call header parsing overhaul Chuck Lever
2023-01-02 17:05 ` [PATCH v1 01/25] SUNRPC: Push svcxdr_init_decode() into svc_process_common() Chuck Lever
2023-01-02 17:05 ` [PATCH v1 02/25] SUNRPC: Move svcxdr_init_decode() into ->accept methods Chuck Lever
2023-01-02 17:05 ` [PATCH v1 03/25] SUNRPC: Add an XDR decoding helper for struct opaque_auth Chuck Lever
2023-01-02 17:05 ` [PATCH v1 04/25] SUNRPC: Convert svcauth_null_accept() to use xdr_stream Chuck Lever
2023-01-02 17:05 ` [PATCH v1 05/25] SUNRPC: Convert svcauth_unix_accept() " Chuck Lever
2023-01-02 17:06 ` [PATCH v1 06/25] SUNRPC: Convert svcauth_tls_accept() " Chuck Lever
2023-01-02 17:06 ` [PATCH v1 07/25] SUNRPC: Move the server-side GSS upcall to a noinline function Chuck Lever
2023-01-02 17:06 ` [PATCH v1 08/25] SUNRPC: Hoist common verifier decoding code into svcauth_gss_proc_init() Chuck Lever
2023-01-02 17:06 ` [PATCH v1 09/25] SUNRPC: Remove gss_read_common_verf() Chuck Lever
2023-01-02 17:06 ` [PATCH v1 10/25] SUNRPC: Remove gss_read_verf() Chuck Lever
2023-01-02 17:06 ` [PATCH v1 11/25] SUNRPC: Convert server-side GSS upcall helpers to use xdr_stream Chuck Lever
2023-01-02 17:06 ` [PATCH v1 12/25] SUNRPC: Replace read_u32_from_xdr_buf() with existing XDR helper Chuck Lever
2023-01-02 17:06 ` [PATCH v1 13/25] SUNRPC: Rename automatic variables in unwrap_integ_data() Chuck Lever
2023-01-02 17:06 ` [PATCH v1 14/25] SUNRPC: Convert unwrap_integ_data() to use xdr_stream Chuck Lever
2023-01-02 17:07 ` [PATCH v1 15/25] SUNRPC: Rename automatic variables in unwrap_priv_data() Chuck Lever
2023-01-02 17:07 ` [PATCH v1 16/25] SUNRPC: Convert unwrap_priv_data() to use xdr_stream Chuck Lever
2023-01-02 17:07 ` [PATCH v1 17/25] SUNRPC: Convert gss_verify_header() " Chuck Lever
2023-01-02 17:07 ` [PATCH v1 18/25] SUNRPC: Clean up svcauth_gss_accept's NULL procedure check Chuck Lever
2023-01-02 17:07 ` [PATCH v1 19/25] SUNRPC: Convert the svcauth_gss_accept() pre-amble to use xdr_stream Chuck Lever
2023-01-02 17:07 ` [PATCH v1 20/25] SUNRPC: Hoist init_decode out of svc_authenticate() Chuck Lever
2023-01-02 17:07 ` [PATCH v1 21/25] SUNRPC: Re-order construction of the first reply fields Chuck Lever
2023-01-02 17:07 ` [PATCH v1 22/25] SUNRPC: Eliminate unneeded variable Chuck Lever
2023-01-02 17:07 ` [PATCH v1 23/25] SUNRPC: Decode most of RPC header with xdr_stream Chuck Lever
2023-01-02 17:07 ` [PATCH v1 24/25] SUNRPC: Remove svc_process_common's argv parameter Chuck Lever
2023-01-02 17:08 ` [PATCH v1 25/25] SUNRPC: Hoist svcxdr_init_decode() into svc_process() Chuck Lever
2023-01-03 14:52 ` [PATCH v1 00/25] Server-side RPC call header parsing overhaul Jeff Layton
2023-01-03 16:33   ` Chuck Lever III [this message]

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=14A17DBB-CD3D-41A4-B2C2-D46DEBB1243C@oracle.com \
    --to=chuck.lever@oracle.com \
    --cc=jlayton@kernel.org \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox