All of lore.kernel.org
 help / color / mirror / Atom feed
From: Denis Kenzior <denkenz@gmail.com>
To: ofono@ofono.org
Subject: Re: [PATCH 2/4] Add GAtServer basic parsing support
Date: Thu, 14 Jan 2010 11:24:30 -0600	[thread overview]
Message-ID: <201001141124.30514.denkenz@gmail.com> (raw)
In-Reply-To: <1263207230-22036-2-git-send-email-zhenhua.zhang@intel.com>

[-- Attachment #1: Type: text/plain, Size: 6692 bytes --]

Hi Zhenhua,

> +
> +#include <glib.h>
> +
> +#include "ringbuffer.h"
> +#include "gatresult.h"

Is this include really necessary?

> +#include "gatserver.h"
> +
> +#define DBG(fmt, arg...) g_print("%s:%s() " fmt, __FILE__, __FUNCTION__ ,
>  ## arg) +

Move this to gat.h

> +struct _GAtServer {
> +	gint ref_count;				/* Ref count */
> +	struct v250_settings *v250;		/* V.250 command setting */

This one doesn't need to be a pointer.

> +	GIOChannel *server_io;			/* Server IO */
> +	int server_watch;			/* Watch for server IO */
> +	char *modem_path;			/* Emulated modem path */

Get rid of this, there's no purpose for it.

> +static int at_server_parse(GAtServer *server, char *buf)
> +{
> +	int res = G_AT_SERVER_RESULT_ERROR;
> +	struct v250_settings *v250 = server->v250;
> +	gsize i = 0;
> +	char c;
> +
> +	/* skip space after "AT" or previous command */
> +	i = skip_space(buf, i);
> +
> +	c = buf[i];
> +	/* skip semicolon */
> +	if (c == ';')
> +		c = buf[++i];
> +
> +	if (g_ascii_isalpha(c) || c == '&')
> +		res = parse_v250_settings(server, buf + i);

This part makes no sense

> +	else if (c == v250->s3)
> +		res = G_AT_SERVER_RESULT_OK;
> +
> +	return res;
> +}
> +
> +static void parse_buffer(GAtServer *server, unsigned int len)
> +{
> +	int res = G_AT_SERVER_RESULT_ERROR;
> +	gsize i = 0;
> +	char *buf = NULL;
> +
> +	g_at_server_ref(server);

Lets get rid of this part for now.  It was a necessary evil inside GAtChat 
since the client can close the channel in the command result callback.  I 
don't think this is relevant for GAtServer.  If it ever becomes relevant we 
can add this back in.

> +
> +	buf = g_try_new0(char, len+1);

You're leaking the buf memory in this function.

> +
> +	if (!buf)
> +		goto out;
> +
> +	if (-1 == ring_buffer_read(server->buf, buf, len))
> +		goto out;
> +
> +	buf[len] = '\0';
> +
> +	DBG("received %s\n", buf);
> +
> +	/* skip header space */
> +	buf += skip_space(buf, i);
> +
> +	/* Make sure the command line prefix is "AT" or "at" */
> +	if (g_str_has_prefix(buf, "AT") ||
> +				g_str_has_prefix(buf, "at"))
> +		res = at_server_parse(server, (char *) buf + 2);
> +
> +	g_at_server_send_result_code(server, res);
> +
> +out:
> +	/* We're overflowing the buffer, shutdown the socket */
> +	if (server->buf && ring_buffer_avail(server->buf) == 0)
> +		g_at_server_shutdown(server);
> +
> +	g_at_server_unref(server);

Same here, see above.

> +}
> +
> +static gboolean received_data(GIOChannel *chan, GIOCondition cond,
> +				gpointer data)
> +{
> +	GAtServer *server = data;
> +	struct v250_settings *v250 = server->v250;
> +	GIOError err;
> +	gsize rbytes;
> +	gsize total_read = 0;
> +	unsigned char *total_buf = ring_buffer_write_ptr(server->buf);

You're totally abusing the ring buffer concept here, the fact that this works 
is pure luck.

> +	static gsize total_len;
> +
> +	if (cond & G_IO_NVAL)
> +		return FALSE;
> +
> +	if (cond & (G_IO_HUP | G_IO_ERR)) {
> +		g_at_server_shutdown(server);
> +
> +		return FALSE;
> +	}
> +
> +	/* Regardless of condition, try to read all the data available */
> +	do {
> +		unsigned char *buf;
> +		gsize toread;
> +
> +		rbytes = 0;
> +
> +		toread = ring_buffer_avail_no_wrap(server->buf);
> +
> +		if (toread == 0)
> +			break;
> +
> +		buf = ring_buffer_write_ptr(server->buf);
> +
> +		err = g_io_channel_read(chan, (char *) buf, toread, &rbytes);
> +
> +		total_read += rbytes;
> +
> +		if (rbytes > 0)
> +			ring_buffer_write_advance(server->buf, rbytes);
> +
> +	} while (err == G_IO_ERROR_NONE && rbytes > 0);
> +
> +	g_at_util_debug_chat(server->debugf, TRUE, (char *) total_buf,
> +						total_read, server->debug_data);
> +
> +	if (total_read == 0) {
> +		g_at_server_shutdown(server);

You should be checking rbytes == 0 && err != EAGAIN here.  Your check below is 
also redundant.

> +
> +		return FALSE;
> +	}
> +
> +	total_len += total_read;
> +
> +	/* Add '\0' to perform strchr */
> +	total_buf[total_read] = '\0';
> +
> +	/* Parse buffer until we meet the terminator so that
> +	 * all preceding characters will be processed
> +	 */
> +	if (strchr((char *) total_buf, v250->s3)) {
> +		parse_buffer(server, total_len);
> +
> +		total_len = 0;
> +	}

Should this be inside a while loop? Several commands can come in at once.

> +
> +	if (err != G_IO_ERROR_NONE && err != G_IO_ERROR_AGAIN)
> +		return FALSE;
> +
> +	return TRUE;
> +}
> +

> +GAtServer *g_at_server_new(GIOChannel *io, const char *modem_path)

Get rid of modem_path

> +{
> +	GAtServer *server;
> +
> +	if (!io || !modem_path)
> +		return NULL;
> +
> +	server = g_try_new0(GAtServer, 1);
> +	if (!server)
> +		return NULL;
> +
> +	server->ref_count = 1;
> +	server->v250 = v250_settings_create();
> +	server->server_io = io;
> +	server->modem_path = g_strdup(modem_path);

And here

> +	server->user_disconnect = NULL;
> +	server->user_disconnect_data = NULL;
> +	server->debugf = NULL;
> +	server->debug_data = NULL;
> +	server->buf = ring_buffer_new(4096);
> +
> +	if (!server->v250)
> +		goto error;
> +
> +	if (!server->buf)
> +		goto error;
> +
> +	if (!g_at_util_set_io(server->server_io))
> +		goto error;
> +
> +	server->server_watch = g_io_add_watch_full(io,
> +				G_PRIORITY_DEFAULT,
> +				G_IO_IN | G_IO_HUP | G_IO_ERR | G_IO_NVAL,
> +				received_data, server, NULL);
> +
> +	return server;
> +
> +error:
> +	if (server->buf)
> +		ring_buffer_free(server->buf);
> +
> +	if (server->modem_path)
> +		g_free(server->modem_path);

And here

> +
> +	if (server->v250)
> +		g_free(server->v250);
> +
> +	if (server)
> +		g_free(server);
> +
> +	return NULL;
> +}
> +
> +gboolean g_at_server_shutdown(GAtServer *server)
> +{
> +	if (!server)
> +		return FALSE;
> +
> +	if (server->modem_path)
> +		g_free(server->modem_path);
> +
> +	if (server->v250)
> +		g_free(server->v250);
> +
> +	ring_buffer_free(server->buf);
> +	server->buf = NULL;
> +
> +	if (server->server_watch) {
> +		g_source_remove(server->server_watch);
> +		server->server_watch = 0;
> +	}
> +
> +	if (server->server_io) {
> +		g_io_channel_shutdown(server->server_io, FALSE, NULL);
> +		g_io_channel_unref(server->server_io);
> +		server->server_io = NULL;
> +	}
> +
> +	if (server->user_disconnect)
> +		server->user_disconnect(server->user_disconnect_data);

We should not trigger a user_disconnect if we're shutting down normally.  The 
purpose is to detect when the remote side has disconnected the connection.

Regards,
-Denis

  parent reply	other threads:[~2010-01-14 17:24 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-11 10:53 [PATCH 1/4] Add gatutil.c to share common APIs with GAtServer Zhenhua Zhang
2010-01-11 10:53 ` [PATCH 2/4] Add GAtServer basic parsing support Zhenhua Zhang
2010-01-11 10:53   ` [PATCH 3/4] Add AT command parsing support and interface Zhenhua Zhang
2010-01-11 10:53     ` [PATCH 4/4] Add test-server test application for GAtServer Zhenhua Zhang
2010-01-14 17:24   ` Denis Kenzior [this message]
2010-01-14 21:53     ` [PATCH 2/4] Add GAtServer basic parsing support Marcel Holtmann
2010-01-14 20:57       ` Denis Kenzior
2010-01-14 15:49 ` [PATCH 1/4] Add gatutil.c to share common APIs with GAtServer Denis Kenzior
2010-01-15  1:24   ` Zhang, Zhenhua
2010-01-20 20:23     ` Denis Kenzior

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=201001141124.30514.denkenz@gmail.com \
    --to=denkenz@gmail.com \
    --cc=ofono@ofono.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 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.