[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180417072946.GD15310@kroah.com>
Date: Tue, 17 Apr 2018 09:29:46 +0200
From: Greg KH <gregkh@...uxfoundation.org>
To: longli@...rosoft.com
Cc: Steve French <sfrench@...ba.org>, linux-cifs@...r.kernel.org,
samba-technical@...ts.samba.org, linux-kernel@...r.kernel.org,
linux-rdma@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH 4/6] cifs: smbd: Don't use RDMA read/write when signing
is used
On Mon, Apr 16, 2018 at 05:49:16PM -0700, Long Li wrote:
> From: Long Li <longli@...rosoft.com>
>
> SMB server will not sign data transferred through RDMA read/write. When
> signing is used, it's a good idea to have all the data signed.
>
> In this case, use RDMA send/recv for all data transfers. This will degrade
> performance as this is not generally configured in RDMA environemnt. So
> warn the user on signing and RDMA send/recv.
>
> Signed-off-by: Long Li <longli@...rosoft.com>
> ---
> fs/cifs/cifssmb.c | 3 +++
> fs/cifs/smb2ops.c | 18 ++++++++++++++----
> fs/cifs/smb2pdu.c | 4 ++--
> 3 files changed, 19 insertions(+), 6 deletions(-)
>
> diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
> index 6d3e40d..1529a08 100644
> --- a/fs/cifs/cifssmb.c
> +++ b/fs/cifs/cifssmb.c
> @@ -455,6 +455,9 @@ cifs_enable_signing(struct TCP_Server_Info *server, bool mnt_sign_required)
> server->sign = true;
> }
>
> + if (cifs_rdma_enabled(server) && server->sign)
> + cifs_dbg(VFS, "Signing is enabled, and RDMA read/write will be disabled");
> +
> return 0;
> }
>
> diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
> index 38ebf3f..b76b858 100644
> --- a/fs/cifs/smb2ops.c
> +++ b/fs/cifs/smb2ops.c
> @@ -252,9 +252,14 @@ smb2_negotiate_wsize(struct cifs_tcon *tcon, struct smb_vol *volume_info)
> wsize = volume_info->wsize ? volume_info->wsize : CIFS_DEFAULT_IOSIZE;
> wsize = min_t(unsigned int, wsize, server->max_write);
> #ifdef CONFIG_CIFS_SMB_DIRECT
> - if (server->rdma)
> - wsize = min_t(unsigned int,
> + if (server->rdma) {
> + if (server->sign)
> + wsize = min_t(unsigned int,
> + wsize, server->smbd_conn->max_fragmented_send_size);
> + else
> + wsize = min_t(unsigned int,
> wsize, server->smbd_conn->max_readwrite_size);
> + }
> #endif
> if (!(server->capabilities & SMB2_GLOBAL_CAP_LARGE_MTU))
> wsize = min_t(unsigned int, wsize, SMB2_MAX_BUFFER_SIZE);
> @@ -272,9 +277,14 @@ smb2_negotiate_rsize(struct cifs_tcon *tcon, struct smb_vol *volume_info)
> rsize = volume_info->rsize ? volume_info->rsize : CIFS_DEFAULT_IOSIZE;
> rsize = min_t(unsigned int, rsize, server->max_read);
> #ifdef CONFIG_CIFS_SMB_DIRECT
> - if (server->rdma)
> - rsize = min_t(unsigned int,
> + if (server->rdma) {
> + if (server->sign)
> + rsize = min_t(unsigned int,
> + rsize, server->smbd_conn->max_fragmented_recv_size);
> + else
> + rsize = min_t(unsigned int,
> rsize, server->smbd_conn->max_readwrite_size);
> + }
> #endif
>
> if (!(server->capabilities & SMB2_GLOBAL_CAP_LARGE_MTU))
> diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c
> index abbefe2..6759035 100644
> --- a/fs/cifs/smb2pdu.c
> +++ b/fs/cifs/smb2pdu.c
> @@ -2596,7 +2596,7 @@ smb2_new_read_req(void **buf, unsigned int *total_len,
> * If we want to do a RDMA write, fill in and append
> * smbd_buffer_descriptor_v1 to the end of read request
> */
> - if (server->rdma && rdata &&
> + if (server->rdma && rdata && !server->sign &&
> rdata->bytes >= server->smbd_conn->rdma_readwrite_threshold) {
>
> struct smbd_buffer_descriptor_v1 *v1;
> @@ -2974,7 +2974,7 @@ smb2_async_writev(struct cifs_writedata *wdata,
> * If we want to do a server RDMA read, fill in and append
> * smbd_buffer_descriptor_v1 to the end of write request
> */
> - if (server->rdma && wdata->bytes >=
> + if (server->rdma && !server->sign && wdata->bytes >=
> server->smbd_conn->rdma_readwrite_threshold) {
>
> struct smbd_buffer_descriptor_v1 *v1;
> --
> 2.7.4
<formletter>
This is not the correct way to submit patches for inclusion in the
stable kernel tree. Please read:
https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
for how to do this properly.
</formletter>
Powered by blists - more mailing lists