All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Zhang, Chen" <chen.zhang@intel.com>
To: Lukas Straub <lukasstraub2@web.de>
Cc: Jason Wang <jasowang@redhat.com>,
	Markus Armbruster <armbru@redhat.com>,
	qemu-dev <qemu-devel@nongnu.org>,
	Zhang Chen <zhangckid@gmail.com>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>
Subject: RE: [PATCH 3/3] net/colo-compare: Add handler for passthrough connection
Date: Fri, 15 Jan 2021 09:08:24 +0000	[thread overview]
Message-ID: <7b08aa481627408fbe930382d7ef4af1@intel.com> (raw)
In-Reply-To: <20210114145037.706fe8c7@gecko.fritz.box>



> -----Original Message-----
> From: Lukas Straub <lukasstraub2@web.de>
> Sent: Thursday, January 14, 2021 9:51 PM
> To: Zhang, Chen <chen.zhang@intel.com>
> Cc: Jason Wang <jasowang@redhat.com>; qemu-dev <qemu-
> devel@nongnu.org>; Eric Blake <eblake@redhat.com>; Dr. David Alan
> Gilbert <dgilbert@redhat.com>; Markus Armbruster <armbru@redhat.com>;
> Zhang Chen <zhangckid@gmail.com>
> Subject: Re: [PATCH 3/3] net/colo-compare: Add handler for passthrough
> connection
> 
> On Thu, 24 Dec 2020 09:09:18 +0800
> Zhang Chen <chen.zhang@intel.com > wrote:
> 
> > From: Zhang Chen <chen.zhang@intel.com>
> >
> > Currently, we just use guest's TCP/UDP source port as the key to
> > bypass certain network traffic.
> >
> > Signed-off-by: Zhang Chen <chen.zhang@intel.com>
> > ---
> >  net/colo-compare.c | 49
> > ++++++++++++++++++++++++++++++++++++++++++++++
> >  net/colo-compare.h |  2 ++
> >  net/net.c          | 27 +++++++++++++++++++++++++
> >  3 files changed, 78 insertions(+)
> >
> > diff --git a/net/colo-compare.c b/net/colo-compare.c index
> > 337025b44f..11a32caa9b 100644
> > --- a/net/colo-compare.c
> > +++ b/net/colo-compare.c
> > @@ -46,6 +46,9 @@ static QTAILQ_HEAD(, CompareState) net_compares =
> > static NotifierList colo_compare_notifiers =
> >      NOTIFIER_LIST_INITIALIZER(colo_compare_notifiers);
> >
> > +static QLIST_HEAD(, PassthroughEntry) passthroughlist =
> > +    QLIST_HEAD_INITIALIZER(passthroughlist);
> > +
> >  #define COMPARE_READ_LEN_MAX NET_BUFSIZE  #define
> MAX_QUEUE_SIZE 1024
> >
> > @@ -103,6 +106,12 @@ typedef struct SendEntry {
> >      uint8_t *buf;
> >  } SendEntry;
> >
> > +typedef struct PassthroughEntry {
> > +    bool is_tcp;
> > +    uint16_t port;
> > +    QLIST_ENTRY(PassthroughEntry) node; } PassthroughEntry;
> > +
> >  struct CompareState {
> >      Object parent;
> >
> > @@ -247,6 +256,7 @@ static int packet_enqueue(CompareState *s, int
> mode, Connection **con)
> >      ConnectionKey key;
> >      Packet *pkt = NULL;
> >      Connection *conn;
> > +    PassthroughEntry *bypass, *next;
> >      int ret;
> >
> >      if (mode == PRIMARY_IN) {
> > @@ -264,8 +274,23 @@ static int packet_enqueue(CompareState *s, int
> mode, Connection **con)
> >          pkt = NULL;
> >          return -1;
> >      }
> > +
> >      fill_connection_key(pkt, &key);
> >
> > +    /* Check COLO passthrough connenction */
> > +    if (!QLIST_EMPTY(&passthroughlist)) {
> > +        QLIST_FOREACH_SAFE(bypass, &passthroughlist, node, next) {
> > +            if (((key.ip_proto == IPPROTO_TCP) && bypass->is_tcp) ||
> > +                ((key.ip_proto == IPPROTO_UDP) && !bypass->is_tcp)) {
> > +                if (bypass->port == key.src_port) {
> > +                    packet_destroy(pkt, NULL);
> > +                    pkt = NULL;
> > +                    return -1;
> > +                }
> > +            }
> > +        }
> > +    }
> > +
> >      conn = connection_get(s->connection_track_table,
> >                            &key,
> >                            &s->conn_list); @@ -1373,6 +1398,30 @@
> > static void colo_flush_packets(void *opaque, void *user_data)
> >      }
> >  }
> >
> > +void colo_compare_passthrough_add(bool is_tcp, const uint16_t port) {
> > +    PassthroughEntry *bypass = NULL;
> > +
> > +    bypass = g_new0(PassthroughEntry, 1);
> > +    bypass->is_tcp = is_tcp;
> > +    bypass->port = port;
> > +    QLIST_INSERT_HEAD(&passthroughlist, bypass, node); }
> > +
> > +void colo_compare_passthrough_del(bool is_tcp, const uint16_t port) {
> > +    PassthroughEntry *bypass = NULL, *next = NULL;
> > +
> > +    if (!QLIST_EMPTY(&passthroughlist)) {
> > +        QLIST_FOREACH_SAFE(bypass, &passthroughlist, node, next) {
> > +            if ((bypass->is_tcp == is_tcp) && (bypass->port == port)) {
> > +                QLIST_REMOVE(bypass, node);
> > +                g_free(bypass);
> > +            }
> > +        }
> > +    }
> > +}
> > +
> 
> Access to "passtroughlist" needs to be protected by a lock, as
> "packet_enqueue" is called from a different iothread.

OK, I will add the lock in next version.

Thanks
Chen

> 
> >  static void colo_compare_class_init(ObjectClass *oc, void *data)  {
> >      UserCreatableClass *ucc = USER_CREATABLE_CLASS(oc); diff --git
> > a/net/colo-compare.h b/net/colo-compare.h index
> 22ddd512e2..1fa026c85e
> > 100644
> > --- a/net/colo-compare.h
> > +++ b/net/colo-compare.h
> > @@ -20,5 +20,7 @@
> >  void colo_notify_compares_event(void *opaque, int event, Error
> > **errp);  void colo_compare_register_notifier(Notifier *notify);  void
> > colo_compare_unregister_notifier(Notifier *notify);
> > +void colo_compare_passthrough_add(bool is_tcp, const uint16_t port);
> > +void colo_compare_passthrough_del(bool is_tcp, const uint16_t port);
> >
> >  #endif /* QEMU_COLO_COMPARE_H */
> > diff --git a/net/net.c b/net/net.c
> > index eac7a92618..1f303e8309 100644
> > --- a/net/net.c
> > +++ b/net/net.c
> > @@ -55,6 +55,7 @@
> >  #include "sysemu/sysemu.h"
> >  #include "net/filter.h"
> >  #include "qapi/string-output-visitor.h"
> > +#include "net/colo-compare.h"
> >
> >  /* Net bridge is currently not supported for W32. */  #if
> > !defined(_WIN32) @@ -1155,12 +1156,38 @@ void
> > qmp_colo_passthrough_add(const char *prot, const uint32_t port,
> >                                Error **errp)  {
> >      /* Setup passthrough connection */
> > +    if (port > 65536) {
> > +        error_setg(errp, "COLO pass through get wrong port");
> > +        return;
> > +    }
> > +
> > +    if (!strcmp(prot, "tcp") || !strcmp(prot, "TCP")) {
> > +        colo_compare_passthrough_add(true, (uint16_t)port);
> > +    } else if (!strcmp(prot, "udp") || !strcmp(prot, "UDP")) {
> > +        colo_compare_passthrough_add(false, (uint16_t)port);
> > +    } else {
> > +        error_setg(errp, "COLO pass through just support tcp or udp
> protocol");
> > +        return;
> > +    }
> >  }
> >
> >  void qmp_colo_passthrough_del(const char *prot, const uint32_t port,
> >                                Error **errp)  {
> >      /* Delete passthrough connection */
> > +    if (port > 65536) {
> > +        error_setg(errp, "COLO pass through get wrong port");
> > +        return;
> > +    }
> > +
> > +    if (!strcmp(prot, "tcp") || !strcmp(prot, "TCP")) {
> > +        colo_compare_passthrough_del(true, (uint16_t)port);
> > +    } else if (!strcmp(prot, "udp") || !strcmp(prot, "UDP")) {
> > +        colo_compare_passthrough_del(false, (uint16_t)port);
> > +    } else {
> > +        error_setg(errp, "COLO pass through just support tcp or udp
> protocol");
> > +        return;
> > +    }
> >  }
> >
> >  static void netfilter_print_info(Monitor *mon, NetFilterState *nf)
> 
> 
> 
> --



  reply	other threads:[~2021-01-15  9:13 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-24  1:09 [PATCH 0/3] Bypass specific network traffic in COLO Zhang Chen
2020-12-24  1:09 ` [PATCH 1/3] qapi/net: Add new QMP command for COLO passthrough Zhang Chen
2020-12-25  6:20   ` Jason Wang
2020-12-28  0:38     ` Zhang, Chen
2020-12-28  7:11       ` Jason Wang
2020-12-29  2:56         ` Zhang, Chen
2020-12-30  3:56           ` Jason Wang
2021-01-05  3:28             ` Zhang, Chen
2021-01-05  4:17               ` Jason Wang
2021-01-05  6:29                 ` Zhang, Chen
2021-01-19 19:32   ` Eric Blake
2021-01-21  1:50     ` Zhang, Chen
2020-12-24  1:09 ` [PATCH 2/3] hmp-commands: Add new HMP " Zhang Chen
2020-12-24  1:09 ` [PATCH 3/3] net/colo-compare: Add handler for passthrough connection Zhang Chen
2021-01-14 13:45   ` Lukas Straub
2021-01-15  9:07     ` Zhang, Chen
2021-01-15 16:06       ` Lukas Straub
2021-01-14 13:50   ` Lukas Straub
2021-01-15  9:08     ` Zhang, Chen [this message]
2020-12-25  6:23 ` [PATCH 0/3] Bypass specific network traffic in COLO Jason Wang
2020-12-28  0:38   ` Zhang, Chen
2021-01-04 13:06 ` Dr. David Alan Gilbert
2021-01-05  3:28   ` Zhang, Chen

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=7b08aa481627408fbe930382d7ef4af1@intel.com \
    --to=chen.zhang@intel.com \
    --cc=armbru@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=jasowang@redhat.com \
    --cc=lukasstraub2@web.de \
    --cc=qemu-devel@nongnu.org \
    --cc=zhangckid@gmail.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.