From: "Wu, Jiaxin" <jiaxin.wu@intel.com>
To: "Fu, Siyuan" <siyuan.fu@intel.com>,
"edk2-devel@lists.01.org" <edk2-devel@lists.01.org>
Cc: Andrey Tepin <atepin@kraftway.ru>, "Ye, Ting" <ting.ye@intel.com>
Subject: Re: [Patch V2] NetworkPkg: Addressing TCP Window Retraction when window scale factor is used.
Date: Fri, 5 May 2017 05:12:26 +0000 [thread overview]
Message-ID: <895558F6EA4E3B41AC93A00D163B7274162C7A50@SHSMSX103.ccr.corp.intel.com> (raw)
In-Reply-To: <1493949365-105904-2-git-send-email-siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
> -----Original Message-----
> From: Fu, Siyuan
> Sent: Friday, May 5, 2017 9:56 AM
> To: edk2-devel@lists.01.org
> Cc: Wu, Jiaxin <jiaxin.wu@intel.com>; Andrey Tepin <atepin@kraftway.ru>; Ye,
> Ting <ting.ye@intel.com>
> Subject: [Patch V2] NetworkPkg: Addressing TCP Window Retraction when
> window scale factor is used.
>
> The RFC1323 which defines the TCP window scale option has been obsoleted by
> RFC7323.
> This patch is to follow the RFC7323 to address the TCP window retraction
> problem
> when a non-zero scale factor is used.
> The changes has been test in high packet loss rate network by using HTTP boot
> and
> iSCSI file read/write.
>
> Cc: Wu Jiaxin <jiaxin.wu@intel.com>
> Cc: Andrey Tepin <atepin@kraftway.ru>
> Cc: Ye Ting <ting.ye@intel.com>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
> ---
> NetworkPkg/TcpDxe/TcpMisc.c | 3 ++-
> NetworkPkg/TcpDxe/TcpOutput.c | 33 ++++++++++++++++++++++++++++-----
> NetworkPkg/TcpDxe/TcpProto.h | 8 +++++++-
> 3 files changed, 37 insertions(+), 7 deletions(-)
>
> diff --git a/NetworkPkg/TcpDxe/TcpMisc.c b/NetworkPkg/TcpDxe/TcpMisc.c
> index a8592c9..4435036 100644
> --- a/NetworkPkg/TcpDxe/TcpMisc.c
> +++ b/NetworkPkg/TcpDxe/TcpMisc.c
> @@ -2,7 +2,7 @@
> Misc support routines for TCP driver.
>
> (C) Copyright 2014 Hewlett-Packard Development Company, L.P.<BR>
> - Copyright (c) 2009 - 2016, Intel Corporation. All rights reserved.<BR>
> + Copyright (c) 2009 - 2017, Intel Corporation. All rights reserved.<BR>
>
> This program and the accompanying materials
> are licensed and made available under the terms and conditions of the BSD
> License
> @@ -86,6 +86,7 @@ TcpInitTcbLocal (
> // First window size is never scaled
> //
> Tcb->RcvWndScale = 0;
> + Tcb->RetxmitSeqMax = 0;
>
> Tcb->ProbeTimerOn = FALSE;
> }
> diff --git a/NetworkPkg/TcpDxe/TcpOutput.c
> b/NetworkPkg/TcpDxe/TcpOutput.c
> index a46cb60..f3dacf3 100644
> --- a/NetworkPkg/TcpDxe/TcpOutput.c
> +++ b/NetworkPkg/TcpDxe/TcpOutput.c
> @@ -1,7 +1,7 @@
> /** @file
> TCP output process routines.
>
> - Copyright (c) 2009 - 2016, Intel Corporation. All rights reserved.<BR>
> + Copyright (c) 2009 - 2017, Intel Corporation. All rights reserved.<BR>
>
> This program and the accompanying materials
> are licensed and made available under the terms and conditions of the BSD
> License
> @@ -664,7 +664,27 @@ TcpRetransmit (
> // 2. Must in the current send window
> // 3. Will not change the boundaries of queued segments.
> //
> - if (TCP_SEQ_LT (Tcb->SndWl2 + Tcb->SndWnd, Seq)) {
> +
> + if ((Tcb->SndWndScale != 0) &&
> + (TCP_SEQ_GT (Seq, Tcb->RetxmitSeqMax) || TCP_SEQ_GT (Tcb->SndWl2 +
> Tcb->SndWnd + (1 << Tcb->SndWndScale), Seq))) {
> + //
> + // Handle the Window Retraction if TCP window scale is enabled according
> to RFC7323:
> + // On first retransmission, or if the sequence number is out of
> + // window by less than 2^Rcv.Wind.Shift, then do normal
> + // retransmission(s) without regard to the receiver window as long
> + // as the original segment was in window when it was sent.
> + //
> + Len = TCP_SUB_SEQ (Tcb->SndNxt, Seq);
> + DEBUG (
> + (EFI_D_WARN,
> + "TcpRetransmit: retransmission without regard to the receiver window for
> TCB %p\n",
> + Tcb)
> + );
> +
> + } else if (TCP_SEQ_GEQ (Tcb->SndWl2 + Tcb->SndWnd, Seq)) {
> + Len = TCP_SUB_SEQ (Tcb->SndWl2 + Tcb->SndWnd, Seq);
> +
> + } else {
> DEBUG (
> (EFI_D_WARN,
> "TcpRetransmit: retransmission cancelled because send window too small
> for TCB %p\n",
> @@ -674,10 +694,9 @@ TcpRetransmit (
> return 0;
> }
>
> - Len = TCP_SUB_SEQ (Tcb->SndWl2 + Tcb->SndWnd, Seq);
> - Len = MIN (Len, Tcb->SndMss);
> + Len = MIN (Len, Tcb->SndMss);
>
> - Nbuf = TcpGetSegmentSndQue (Tcb, Seq, Len);
> + Nbuf = TcpGetSegmentSndQue (Tcb, Seq, Len);
> if (Nbuf == NULL) {
> return -1;
> }
> @@ -688,6 +707,10 @@ TcpRetransmit (
> goto OnError;
> }
>
> + if (TCP_SEQ_GT (Seq, Tcb->RetxmitSeqMax)) {
> + Tcb->RetxmitSeqMax = Seq;
> + }
> +
> //
> // The retransmitted buffer may be on the SndQue,
> // trim TCP head because all the buffers on SndQue
> diff --git a/NetworkPkg/TcpDxe/TcpProto.h b/NetworkPkg/TcpDxe/TcpProto.h
> index ee35134..81397d7 100644
> --- a/NetworkPkg/TcpDxe/TcpProto.h
> +++ b/NetworkPkg/TcpDxe/TcpProto.h
> @@ -1,7 +1,7 @@
> /** @file
> TCP protocol header file.
>
> - Copyright (c) 2009 - 2012, Intel Corporation. All rights reserved.<BR>
> + Copyright (c) 2009 - 2017, Intel Corporation. All rights reserved.<BR>
>
> This program and the accompanying materials
> are licensed and made available under the terms and conditions of the BSD
> License
> @@ -316,6 +316,12 @@ struct _TCP_CONTROL_BLOCK {
> TCP_SEQNO LossRecover; ///< Recover point for retxmit.
>
> //
> + // RFC7323
> + // Addressing Window Retraction for TCP Window Scale Option.
> + //
> + TCP_SEQNO RetxmitSeqMax; ///< Max Seq number in previous
> retransmission.
> +
> + //
> // configuration parameters, for EFI_TCP4_PROTOCOL specification
> //
> UINT32 KeepAliveIdle; ///< Idle time before sending first probe.
> --
> 1.9.5.msysgit.1
prev parent reply other threads:[~2017-05-05 5:12 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-05 1:56 [Patch V2] MdeModulePkg: Addressing TCP Window Retraction when window scale factor is used Fu Siyuan
2017-05-05 1:56 ` [Patch V2] NetworkPkg: " Fu Siyuan
2017-05-05 5:03 ` Ye, Ting
2017-05-05 5:12 ` Wu, Jiaxin [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-list 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=895558F6EA4E3B41AC93A00D163B7274162C7A50@SHSMSX103.ccr.corp.intel.com \
--to=devel@edk2.groups.io \
/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