qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Adam Litke <agl@us.ibm.com>
To: Michael Roth <mdroth@linux.vnet.ibm.com>
Cc: abeekhof@redhat.com, agl@linux.vnet.ibm.com,
	qemu-devel@nongnu.org, aliguori@linux.vnet.ibm.com
Subject: [Qemu-devel] Re: [RFC][RESEND][PATCH v1 13/15] virtproxy: add read handler for proxied connections
Date: Wed, 03 Nov 2010 20:21:22 -0500	[thread overview]
Message-ID: <1288833682.2846.99.camel@aglitke> (raw)
In-Reply-To: <1288798090-7127-14-git-send-email-mdroth@linux.vnet.ibm.com>

On Wed, 2010-11-03 at 10:28 -0500, Michael Roth wrote:
> reads data from client/server connections as they become readable, then
> sends the data over the channel
> 
> Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
> ---
>  virtproxy.c |   80 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 files changed, 80 insertions(+), 0 deletions(-)
> 
> diff --git a/virtproxy.c b/virtproxy.c
> index 86a8e5b..f3f7f46 100644
> --- a/virtproxy.c
> +++ b/virtproxy.c
> @@ -200,6 +200,86 @@ static VPIForward *get_iforward(const VPDriver *drv, const char *service_id)
>      return NULL;
>  }
> 
> +/* read handler for proxied connections */
> +static void vp_conn_read(void *opaque)
> +{
> +    VPConn *conn = opaque;
> +    VPDriver *drv = conn->drv;
> +    VPPacket pkt;
> +    char buf[VP_CONN_DATA_LEN];
> +    int fd, count, ret;
> +    bool client;
> +
> +    TRACE("called with opaque: %p, drv: %p", opaque, drv);
> +
> +    if (conn->state != VP_STATE_CONNECTED) {
> +        LOG("invalid connection state");
> +        return;
> +    }
> +
> +    if (conn->type != VP_CONN_CLIENT && conn->type != VP_CONN_SERVER) {
> +        LOG("invalid connection type");
> +        return;
> +    }

When either of these happen (or some of the other errors below), would
it make sense to close and clean up this connection since you're now in
an invalid state?

> +    /* TODO: all fields should be explicitly set so we shouldn't
> +     * need to memset. this might hurt if we beef up VPPacket size
> +     */
> +    memset(&pkt, 0, sizeof(VPPacket));
> +    pkt.magic = VP_MAGIC;
> +
> +    if (conn->type == VP_CONN_CLIENT) {
> +        client = true;
> +        fd = conn->client_fd;
> +    } else {
> +        client = false;
> +        fd = conn->server_fd;
> +    }
> +
> +    count = read(fd, buf, VP_CONN_DATA_LEN);
> +    if (count == -1) {
> +        LOG("read() failed: %s", strerror(errno));
> +        return;
> +    } else if (count == 0) {
> +        /* connection closed, tell remote end to clean up */
> +        TRACE("connection closed");
> +        pkt.type = VP_PKT_CONTROL;
> +        pkt.payload.msg.type = VP_CONTROL_CLOSE;
> +        if (client) {
> +            /* we're closing the client, have remote close the server conn */
> +            TRACE("closing connection for client fd %d", conn->client_fd);
> +            pkt.payload.msg.args.close.client_fd = -1;
> +            pkt.payload.msg.args.close.server_fd = conn->server_fd;
> +        } else {
> +            TRACE("closing connection for server fd %d", conn->server_fd);
> +            pkt.payload.msg.args.close.server_fd = -1;
> +            pkt.payload.msg.args.close.client_fd = conn->client_fd;;
> +        }
> +        /* clean up things on our end */
> +        closesocket(fd);
> +        vp_set_fd_handler(fd, NULL, NULL, NULL);
> +        QLIST_REMOVE(conn, next);
> +        qemu_free(conn);
> +    } else {
> +        TRACE("data read");
> +        pkt.type = client ? VP_PKT_CLIENT : VP_PKT_SERVER;
> +        pkt.payload.proxied.client_fd = conn->client_fd;
> +        pkt.payload.proxied.server_fd = conn->server_fd;
> +        memcpy(pkt.payload.proxied.data, buf, count);
> +        pkt.payload.proxied.bytes = count;
> +    }
> +
> +    ret = vp_send_all(drv->channel_fd, (void *)&pkt, sizeof(VPPacket));
> +    if (ret == -1) {
> +        LOG("error sending data over channel");
> +        return;
> +    }
> +    if (ret != sizeof(VPPacket)) {
> +        TRACE("buffer full?");
> +        return;
> +    }
> +}
> +
>  /* accept handler for communication channel
>   *
>   * accept()s connection to communication channel (for sockets), and sets

-- 
Thanks,
Adam

  reply	other threads:[~2010-11-04  1:33 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-11-03 15:27 [Qemu-devel] [RFC][RESEND][PATCH v1 00/15] virtproxy: host/guest communication layer Michael Roth
2010-11-03 15:27 ` [Qemu-devel] [RFC][RESEND][PATCH v1 01/15] virtproxy: base data structures and constants Michael Roth
2010-11-03 22:33   ` [Qemu-devel] " Adam Litke
2010-11-03 15:27 ` [Qemu-devel] [RFC][RESEND][PATCH v1 02/15] virtproxy: qemu-vp, standalone daemon skeleton Michael Roth
2010-11-03 22:47   ` [Qemu-devel] " Adam Litke
2010-11-04 13:57     ` Michael Roth
2010-11-05 13:32       ` Adam Litke
2010-11-09 10:45         ` Amit Shah
2010-11-10  2:51           ` Michael Roth
2010-11-03 15:27 ` [Qemu-devel] [RFC][RESEND][PATCH v1 03/15] virtproxy: add debug functions for virtproxy core Michael Roth
2010-11-03 22:51   ` [Qemu-devel] " Adam Litke
2010-11-03 15:27 ` [Qemu-devel] [RFC][RESEND][PATCH v1 04/15] virtproxy: list look-up functions conns/oforwards/iforwards Michael Roth
2010-11-03 22:56   ` [Qemu-devel] " Adam Litke
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 05/15] virtproxy: add accept handler for communication channel Michael Roth
2010-11-03 23:02   ` [Qemu-devel] " Adam Litke
2010-11-04 16:17     ` Michael Roth
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 06/15] virtproxy: add read " Michael Roth
2010-11-03 23:38   ` [Qemu-devel] " Adam Litke
2010-11-04 17:00     ` Michael Roth
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 07/15] virtproxy: add vp_new() VPDriver constructor Michael Roth
2010-11-03 23:45   ` [Qemu-devel] " Adam Litke
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 08/15] virtproxy: interfaces to set/remove/handle VPOForwards Michael Roth
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 09/15] virtproxy: add handler for data packets Michael Roth
2010-11-04  0:46   ` [Qemu-devel] " Adam Litke
2010-11-04 18:23     ` Michael Roth
2010-11-04  1:48   ` Adam Litke
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 10/15] virtproxy: add handler for control packet Michael Roth
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 11/15] virtproxy: add vp_handle_packet() Michael Roth
2010-11-04  1:13   ` [Qemu-devel] " Adam Litke
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 12/15] virtproxy: interfaces to set/remove VPIForwards Michael Roth
2010-11-04  1:12   ` [Qemu-devel] " Adam Litke
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 13/15] virtproxy: add read handler for proxied connections Michael Roth
2010-11-04  1:21   ` Adam Litke [this message]
2010-11-04 18:26     ` [Qemu-devel] " Michael Roth
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 14/15] virtproxy: Makefile/configure changes to build qemu-vp Michael Roth
2010-11-03 15:28 ` [Qemu-devel] [RFC][RESEND][PATCH v1 15/15] virtproxy: qemu-vp, main logic Michael Roth
2010-11-03 23:44 ` [Qemu-devel] Re: [RFC][RESEND][PATCH v1 00/15] virtproxy: host/guest communication layer Adam Litke
2010-11-04 18:46   ` Michael Roth

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=1288833682.2846.99.camel@aglitke \
    --to=agl@us.ibm.com \
    --cc=abeekhof@redhat.com \
    --cc=agl@linux.vnet.ibm.com \
    --cc=aliguori@linux.vnet.ibm.com \
    --cc=mdroth@linux.vnet.ibm.com \
    --cc=qemu-devel@nongnu.org \
    /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;
as well as URLs for NNTP newsgroup(s).