Re: [RFC] sunrpc: Fix flood of warnings from iov_iter_kvec in linux-next

[Stephen Rothwell <sfr@canb.auug.org.au>]

[-- Attachment #1: Type: text/plain, Size: 4171 bytes --]

Hi all,

On Wed, 31 Oct 2018 12:49:52 +0000 Leonard Crestez <leonard.crestez@nxp.com> wrote:
>
> Since commit aa563d7bca6e ("iov_iter: Separate type from direction and
> use accessor functions") the iov_iter_kvec and iov_iter_bvec functions
> warn if the direction parameter contains anything other than READ/WRITE.
> 
> That commit also updated users of iov_iter_kvec/bvec but a new call was
> added on a different branch by commit 277e4ab7d530 ("SUNRPC: Simplify
> TCP receive code by switching to using iterators")
> 
> This causes a flood of warnings in linux-next like this:
> 
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 110 at ../lib/iov_iter.c:1082 iov_iter_kvec+0x4c/0x5c
> Modules linked in:
> CPU: 0 PID: 110 Comm: kworker/u3:2 Tainted: G        W         4.19.0-next-20181031 #157
> Hardware name: Freescale i.MX6 SoloLite (Device Tree)
> Workqueue: xprtiod xs_stream_data_receive_workfn
> [<c0112908>] (unwind_backtrace) from [<c010d238>] (show_stack+0x10/0x14)
> [<c010d238>] (show_stack) from [<c0b3894c>] (dump_stack+0xb0/0xe8)
> [<c0b3894c>] (dump_stack) from [<c0126e14>] (__warn+0xe0/0x10c)
> [<c0126e14>] (__warn) from [<c0126f54>] (warn_slowpath_null+0x3c/0x48)
> [<c0126f54>] (warn_slowpath_null) from [<c04b225c>] (iov_iter_kvec+0x4c/0x5c)
> [<c04b225c>] (iov_iter_kvec) from [<c0a2c650>] (xs_stream_data_receive_workfn+0x538/0x8e0)
> [<c0a2c650>] (xs_stream_data_receive_workfn) from [<c014609c>] (process_one_work+0x2ac/0x6fc)
> [<c014609c>] (process_one_work) from [<c014678c>] (worker_thread+0x2a0/0x574)
> [<c014678c>] (worker_thread) from [<c014c43c>] (kthread+0x134/0x14c)
> [<c014c43c>] (kthread) from [<c01010b4>] (ret_from_fork+0x14/0x20)
> Exception stack(0xec9fbfb0 to 0xec9fbff8)
> bfa0:                                     00000000 00000000 00000000 00000000
> bfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> bfe0: 00000000 00000000 00000000 00000000 00000013 00000000
> irq event stamp: 91225
> hardirqs last  enabled at (91233): [<c0186630>] console_unlock+0x3e4/0x5d0
> hardirqs last disabled at (91250): [<c01862c8>] console_unlock+0x7c/0x5d0
> softirqs last  enabled at (91266): [<c0102500>] __do_softirq+0x360/0x524
> softirqs last disabled at (91277): [<c012e66c>] irq_exit+0xf8/0x1a4
> ---[ end trace bc12311e869d672a ]---
> 
> This fix updates sunrpc code and make nfs boot cleanly.
> 
> Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
> ---
>  net/sunrpc/xprtsock.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> You can treat this as a bug report; I did a brief search and didn't see
> anyone else complain about this specific issue. Big ugly stacktrace was
> included so that it hopefully shows up in searches.
> 
> Here is a mention of a different conflict between iov_iter and sunrpc:
> 	https://lkml.org/lkml/2018/10/29/64
> 
> diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
> index 1b51e04d3566..ae77c71c1f64 100644
> --- a/net/sunrpc/xprtsock.c
> +++ b/net/sunrpc/xprtsock.c
> @@ -359,20 +359,20 @@ xs_sock_recvmsg(struct socket *sock, struct msghdr *msg, int flags, size_t seek)
>  
>  static ssize_t
>  xs_read_kvec(struct socket *sock, struct msghdr *msg, int flags,
>  		struct kvec *kvec, size_t count, size_t seek)
>  {
> -	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, kvec, 1, count);
> +	iov_iter_kvec(&msg->msg_iter, READ, kvec, 1, count);
>  	return xs_sock_recvmsg(sock, msg, flags, seek);
>  }
>  
>  static ssize_t
>  xs_read_bvec(struct socket *sock, struct msghdr *msg, int flags,
>  		struct bio_vec *bvec, unsigned long nr, size_t count,
>  		size_t seek)
>  {
> -	iov_iter_bvec(&msg->msg_iter, READ | ITER_BVEC, bvec, nr, count);
> +	iov_iter_bvec(&msg->msg_iter, READ, bvec, nr, count);
>  	return xs_sock_recvmsg(sock, msg, flags, seek);
>  }
>  
>  static ssize_t
>  xs_read_discard(struct socket *sock, struct msghdr *msg, int flags,
> -- 
> 2.17.1

I have added that as a merge resolution for the merge of the vfs tree
in linux-next today.  Please remember to tell Linus about it.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]