[GIT PULL] one more nfsd fix for 6.5-rc

[GIT PULL] one more nfsd fix for 6.5-rc

[Chuck Lever III]

The following changes since commit 2ccdd1b13c591d306f0401d98dedc4bdcd02b421:

  Linux 6.5-rc6 (2023-08-13 11:29:55 -0700)

are available in the Git repository at:

  https://git.kernel.org/pub/scm/linux/kernel/git/cel/linux.git tags/nfsd-6.5-4

for you to fetch changes up to c96e2a695e00bca5487824d84b85aab6aa2c1891:

  sunrpc: set the bv_offset of first bvec in svc_tcp_sendmsg (2023-08-14 15:02:25 -0400)

----------------------------------------------------------------
nfsd-6.5 fixes:
- Fix new MSG_SPLICE_PAGES support in server's TCP sendmsg helper

----------------------------------------------------------------
Jeff Layton (1):
      sunrpc: set the bv_offset of first bvec in svc_tcp_sendmsg

 net/sunrpc/svcsock.c | 3 +++
 1 file changed, 3 insertions(+)

--
Chuck Lever

Re: [GIT PULL] one more nfsd fix for 6.5-rc

[Linus Torvalds]

On Thu, 17 Aug 2023 at 16:11, Chuck Lever III <chuck.lever@oracle.com> wrote:
>
>   https://git.kernel.org/pub/scm/linux/kernel/git/cel/linux.git tags/nfsd-6.5-4

I've pulled this, but...

> Jeff Layton (1):
>       sunrpc: set the bv_offset of first bvec in svc_tcp_sendmsg

.. what an odd place to set bv_offset that is.

I'm sure it's right, but it really smells odd to set that initial
offset not when the bvec is created, but long afterwards, just before
it is used.

Is there some reason why 'bv_offset' isn't initialized when the bvec is created?

             Linus

Re: [GIT PULL] one more nfsd fix for 6.5-rc

[Chuck Lever III]

> On Aug 17, 2023, at 10:49 AM, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> 
> On Thu, 17 Aug 2023 at 16:11, Chuck Lever III <chuck.lever@oracle.com> wrote:
>> 
>>  https://git.kernel.org/pub/scm/linux/kernel/git/cel/linux.git tags/nfsd-6.5-4
> 
> I've pulled this, but...
> 
>> Jeff Layton (1):
>>      sunrpc: set the bv_offset of first bvec in svc_tcp_sendmsg
> 
> .. what an odd place to set bv_offset that is.
> 
> I'm sure it's right, but it really smells odd to set that initial
> offset not when the bvec is created, but long afterwards, just before
> it is used.
> 
> Is there some reason why 'bv_offset' isn't initialized when the bvec is created?

Yes:

https://lore.kernel.org/linux-nfs/7c9421cc4b92dee76cc7560c50a4a0ab3fb1ef0d.camel@kernel.org/T/#t

But also note that this fix will get replaced in v6.6:

https://git.kernel.org/pub/scm/linux/kernel/git/cel/linux.git/commit/?h=nfsd-next&id=383bc8bbc3e61d185ff8082cad2da831f158be4a

I just wanted to ensure that v6.5 wasn't broken. It's a little late in the
cycle to apply 383bc8bbc.

If you feel it's appropriate, perhaps you could add a link to the above
lore discussion to the fix... but as the fix is temporary, perhaps not
worth the bother.


--
Chuck Lever

Re: [GIT PULL] one more nfsd fix for 6.5-rc

[Linus Torvalds]

On Thu, 17 Aug 2023 at 16:55, Chuck Lever III <chuck.lever@oracle.com> wrote:
> > On Aug 17, 2023, at 10:49 AM, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> >
> > I'm sure it's right, but it really smells odd to set that initial
> > offset not when the bvec is created, but long afterwards, just before
> > it is used.
> >
> > Is there some reason why 'bv_offset' isn't initialized when the bvec is created?
>
> Yes:
>
> https://lore.kernel.org/linux-nfs/7c9421cc4b92dee76cc7560c50a4a0ab3fb1ef0d.camel@kernel.org/T/#t

Ugh, how ugly.

> But also note that this fix will get replaced in v6.6
>
> I just wanted to ensure that v6.5 wasn't broken. It's a little late in the
> cycle to apply 383bc8bbc.

Sounds good to me. Thanks for the explanation.

              Linus

Re: [GIT PULL] one more nfsd fix for 6.5-rc

[pr-tracker-bot]

The pull request you sent on Thu, 17 Aug 2023 14:11:20 +0000:

> https://git.kernel.org/pub/scm/linux/kernel/git/cel/linux.git tags/nfsd-6.5-4

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/16931859a6500d360b90aeacab3b505a3560a3ed

Thank you!

-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html