docs/design/notification-alarm-evaluator.rst


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248

.. This work is licensed under a Creative Commons Attribution 4.0 International License.
.. http://creativecommons.org/licenses/by/4.0

============================
Notification Alarm Evaluator
============================

.. NOTE::
   This is spec draft of blueprint for OpenStack Ceilomter Liberty.
   To see current version: https://review.openstack.org/172893
   To track development activity:
   https://blueprints.launchpad.net/ceilometer/+spec/notification-alarm-evaluator

https://blueprints.launchpad.net/ceilometer/+spec/notification-alarm-evaluator

This blueprint proposes to add a new alarm evaluator for handling alarms on
events passed from other OpenStack services, that provides event-driven alarm
evaluation which makes new sequence in Ceilometer instead of the polling-based
approach of the existing Alarm Evaluator, and realizes immediate alarm
notification to end users.

Problem description
===================

As an end user, I need to receive alarm notification immediately once
Ceilometer captured an event which would make alarm fired, so that I can
perform recovery actions promptly to shorten downtime of my service.
The typical use case is that an end user set alarm on "compute.instance.update"
in order to trigger recovery actions once the instance status has changed to
'shutdown' or 'error'. It should be nice that an end user can receive
notification within 1 second after fault observed as the same as other helth-
check mechanisms can do in some cases.

The existing Alarm Evaluator is periodically querying/polling the databases
in order to check all alarms independently from other processes. This is good
approach for evaluating an alarm on samples stored in a certain period.
However, this is not efficient to evaluate an alarm on events which are emitted
by other OpenStack servers once in a while.

The periodical evaluation leads delay on sending alarm notification to users.
The default period of evaluation cycle is 60 seconds. It is recommended that
an operator set longer interval than configured pipeline interval for
underlying metrics, and also longer enough to evaluate all defined alarms
in certain period while taking into account the number of resources, users and
alarms.

Proposed change
===============

The proposal is to add a new event-driven alarm evaluator which receives
messages from Notification Agent and finds related Alarms, then evaluates each
alarms;

* New alarm evaluator could receive event notification from Notification Agent
  by which adding a dedicated notifier as a publisher in pipeline.yaml
  (e.g. notifier://?topic=event_eval).

* When new alarm evaluator received event notification, it queries alarm
  database by Project ID and Resource ID written in the event notification.

* Found alarms are evaluated by referring event notification.

* Depending on the result of evaluation, those alarms would be fired through
  Alarm Notifier as the same as existing Alarm Evaluator does.

This proposal also adds new alarm type "notification" and "notification_rule".
This enables users to create alarms on events. The separation from other alarm
types (such as "threshold" type) is intended to show different timing of
evaluation and different format of condition, since the new evaluator will
check each event notification once it received whereas "threshold" alarm can
evaluate average of values in certain period calculated from multiple samples.

The new alarm evaluator handles Notification type alarms, so we have to change
existing alarm evaluator to exclude "notification" type alarms from evaluation
targets.

Alternatives
------------

There was similar blueprint proposal "Alarm type based on notification", but
the approach is different. The old proposal was to adding new step (alarm
evaluations) in Notification Agent every time it received event from other
OpenStack services, whereas this proposal intends to execute alarm evaluation
in another component which can minimize impact to existing pipeline processing.

Another approach is enhancement of existing alarm evaluator by adding
notification listener. However, there are two issues; 1) this approach could
cause stall of periodical evaluations when it receives bulk of notifications,
and 2) this could break the alarm portioning i.e. when alarm evaluator received
notification, it might have to evaluate some alarms which are not assign to it.

Data model impact
-----------------

Resource ID will be added to Alarm model as an optional attribute.
This would help the new alarm evaluator to filter out non-related alarms
while querying alarms, otherwise it have to evaluate all alarms in the project.

REST API impact
---------------

Alarm API will be extended as follows;

* Add "notification" type into alarm type list
* Add "resource_id" to "alarm"
* Add "notification_rule" to "alarm"

Sample data of Notification-type alarm::

  {
      "alarm_actions": [
          "http://site:8000/alarm"
      ],
      "alarm_id": null,
      "description": "An alarm",
      "enabled": true,
      "insufficient_data_actions": [
          "http://site:8000/nodata"
      ],
      "name": "InstanceStatusAlarm",
      "notification_rule": {
          "event_type": "compute.instance.update",
          "query" : [
              {
                  "field" : "traits.state",
                  "type" : "string",
                  "value" : "error",
                  "op" : "eq",
              },
          ]
      },
      "ok_actions": [],
      "project_id": "c96c887c216949acbdfbd8b494863567",
      "repeat_actions": false,
      "resource_id": "153462d0-a9b8-4b5b-8175-9e4b05e9b856",
      "severity": "moderate",
      "state": "ok",
      "state_timestamp": "2015-04-03T17:49:38.406845",
      "timestamp": "2015-04-03T17:49:38.406839",
      "type": "notification",
      "user_id": "c96c887c216949acbdfbd8b494863567"
  }

"resource_id" will be refered to query alarm and will not be check permission
and belonging of project.

Security impact
---------------

None

Pipeline impact
---------------

None

Other end user impact
---------------------

None

Performance/Scalability Impacts
-------------------------------

When Ceilomter received a number of events from other OpenStack services in
short period, this alarm evaluator can keep working since events are queued in
a messaging queue system, but it can cause delay of alarm notification to users
and increase the number of read and write access to alarm database.

"resource_id" can be optional, but restricting it to mandatory could be reduce
performance impact. If user create "notification" alarm without "resource_id",
those alarms will be evaluated every time event occurred in the project.
That may lead new evaluator heavy.

Other deployer impact
---------------------

New service process have to be run.

Developer impact
----------------

Developers should be aware that events could be notified to end users and avoid
passing raw infra information to end users, while defining events and traits.

Implementation
==============

Assignee(s)
-----------

Primary assignee:
  r-mibu

Other contributors:
  None

Ongoing maintainer:
  None

Work Items
----------

* New event-driven alarm evaluator

* Add new alarm type "notification" as well as AlarmNotificationRule

* Add "resource_id" to Alarm model

* Modify existing alarm evaluator to filter out "notification" alarms

* Add new config parameter for alarm request check whether accepting alarms
  without specifying "resource_id" or not

Future lifecycle
================

This proposal is key feature to provide information of cloud resources to end
users in real-time that enables efficient integration with user-side manager
or Orchestrator, whereas currently those information are considered to be
consumed by admin side tool or service.
Based on this change, we will seek orchestrating scenarios including fault
recovery and add useful event definition as well as additional traits.

Dependencies
============

None

Testing
=======

New unit/scenario tests are required for this change.

Documentation Impact
====================

* Proposed evaluator will be described in the developer document.

* New alarm type and how to use will be explained in user guide.

References
==========

* OPNFV Doctor project: https://wiki.opnfv.org/doctor

* Blueprint "Alarm type based on notification":
  https://blueprints.launchpad.net/ceilometer/+spec/alarm-on-notification
/*****************************************************************
 *
 * Filename:		donauboe.c
 * Version: 		2.17
 * Description:   Driver for the Toshiba OBOE (or type-O or 701)
 *                FIR Chipset, also supports the DONAUOBOE (type-DO
 *                or d01) FIR chipset which as far as I know is
 *                register compatible.
 * Documentation: http://libxg.free.fr/irda/lib-irda.html
 * Status:        Experimental.
 * Author:        James McKenzie <james@fishsoup.dhs.org>
 * Created at:    Sat May 8  12:35:27 1999
 * Modified:      Paul Bristow <paul.bristow@technologist.com>
 * Modified:      Mon Nov 11 19:10:05 1999
 * Modified:      James McKenzie <james@fishsoup.dhs.org>
 * Modified:      Thu Mar 16 12:49:00 2000 (Substantial rewrite)
 * Modified:      Sat Apr 29 00:23:03 2000 (Added DONAUOBOE support)
 * Modified:      Wed May 24 23:45:02 2000 (Fixed chipio_t structure)
 * Modified: 2.13 Christian Gennerat <christian.gennerat@polytechnique.org>
 * Modified: 2.13 dim jan 07 21:57:39 2001 (tested with kernel 2.4 & irnet/ppp)
 * Modified: 2.14 Christian Gennerat <christian.gennerat@polytechnique.org>
 * Modified: 2.14 lun fev 05 17:55:59 2001 (adapted to patch-2.4.1-pre8-irda1)
 * Modified: 2.15 Martin Lucina <mato@kotelna.sk>
 * Modified: 2.15 Fri Jun 21 20:40:59 2002 (sync with 2.4.18, substantial fixes)
 * Modified: 2.16 Martin Lucina <mato@kotelna.sk>
 * Modified: 2.16 Sat Jun 22 18:54:29 2002 (fix freeregion, default to verbose)
 * Modified: 2.17 Christian Gennerat <christian.gennerat@polytechnique.org>
 * Modified: 2.17 jeu sep 12 08:50:20 2002 (save_flags();cli(); replaced by spinlocks)
 * Modified: 2.18 Christian Gennerat <christian.gennerat@polytechnique.org>
 * Modified: 2.18 ven jan 10 03:14:16 2003 Change probe default options
 *
 *     Copyright (c) 1999 James McKenzie, All Rights Reserved.
 *
 *     This program is free software; you can redistribute it and/or
 *     modify it under the terms of the GNU General Public License as
 *     published by the Free Software Foundation; either version 2 of
 *     the License, or (at your option) any later version.
 *
 *     Neither James McKenzie nor Cambridge University admit liability nor
 *     provide warranty for any of this software. This material is
 *     provided "AS-IS" and at no charge.
 *
 *     Applicable Models : Libretto 100/110CT and many more.
 *     Toshiba refers to this chip as the type-O IR port,
 *     or the type-DO IR port.
 *
 ********************************************************************/

/* Look at toshoboe.h (currently in include/net/irda) for details of */
/* Where to get documentation on the chip         */

/* See below for a description of the logic in this driver */

/* User servicable parts */
/* USE_PROBE Create the code which probes the chip and does a few tests */
/* do_probe module parameter Enable this code */
/* Probe code is very useful for understanding how the hardware works */
/* Use it with various combinations of TT_LEN, RX_LEN */
/* Strongly recommended, disable if the probe fails on your machine */
/* and send me <james@fishsoup.dhs.org> the output of dmesg */
#define USE_PROBE 1
#undef  USE_PROBE

/* Trace Transmit ring, interrupts, Receive ring or not ? */
#define PROBE_VERBOSE 1

/* Debug option, examine sent and received raw data */
/* Irdadump is better, but does not see all packets. enable it if you want. */
#undef DUMP_PACKETS

/* MIR mode has not been tested. Some behaviour is different */
/* Seems to work against an Ericsson R520 for me. -Martin */
#define USE_MIR

/* Schedule back to back hardware transmits wherever possible, otherwise */
/* we need an interrupt for every frame, unset if oboe works for a bit and */
/* then hangs */
#define OPTIMIZE_TX

/* Set the number of slots in the rings */
/* If you get rx/tx fifo overflows at high bitrates, you can try increasing */
/* these */

#define RING_SIZE (OBOE_RING_SIZE_RX8 | OBOE_RING_SIZE_TX8)
#define TX_SLOTS    8
#define RX_SLOTS    8


/* Less user servicable parts below here */

/* Test, Transmit and receive buffer sizes, adjust at your peril */
/* remarks: nfs usually needs 1k blocks */
/* remarks: in SIR mode, CRC is received, -> RX_LEN=TX_LEN+2 */
/* remarks: test accepts large blocks. Standard is 0x80 */
/* When TT_LEN > RX_LEN (SIR mode) data is stored in successive slots. */
/* When 3 or more slots are needed for each test packet, */
/* data received in the first slots is overwritten, even */
/* if OBOE_CTL_RX_HW_OWNS is not set, without any error! */
#define TT_LEN      0x80
#define TX_LEN      0xc00
#define RX_LEN      0xc04
/* Real transmitted length (SIR mode) is about 14+(2%*TX_LEN) more */
/* long than user-defined length (see async_wrap_skb) and is less then 4K */
/* Real received length is (max RX_LEN) differs from user-defined */
/* length only b the CRC (2 or 4 bytes) */
#define BUF_SAFETY  0x7a
#define RX_BUF_SZ   (RX_LEN)
#define TX_BUF_SZ   (TX_LEN+BUF_SAFETY)


/* Logic of the netdev part of this driver                             */

/* The RX ring is filled with buffers, when a packet arrives           */
/* it is DMA'd into the buffer which is marked used and RxDone called  */
/* RxDone forms an skb (and checks the CRC if in SIR mode) and ships   */
/* the packet off upstairs */

/* The transmitter on the oboe chip can work in one of two modes       */
/* for each ring->tx[] the transmitter can either                      */
/* a) transmit the packet, leave the trasmitter enabled and proceed to */
/*    the next ring                                                    */
/* OR                                                                  */
/* b) transmit the packet, switch off the transmitter and issue TxDone */

/* All packets are entered into the ring in mode b), if the ring was   */
/* empty the transmitter is started.                                   */

/* If OPTIMIZE_TX is defined then in TxDone if the ring contains       */
/* more than one packet, all but the last are set to mode a) [HOWEVER  */
/* the hardware may not notice this, this is why we start in mode b) ] */
/* then restart the transmitter                                        */

/* If OPTIMIZE_TX is not defined then we just restart the transmitter  */
/* if the ring isn't empty */

/* Speed changes are delayed until the TxRing is empty                 */
/* mtt is handled by generating packets with bad CRCs, before the data */

/* TODO: */
/* check the mtt works ok      */
/* finish the watchdog         */

/* No user servicable parts below here */

#include <linux/module.h>

#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/ioport.h>
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/pci.h>
#include <linux/rtnetlink.h>

#include <asm/io.h>

#include <net/irda/wrapper.h>
#include <net/irda/irda.h>
//#include <net/irda/irmod.h>
//#include <net/irda/irlap_frame.h>
#include <net/irda/irda_device.h>
#include <net/irda/crc.h>

#include "donauboe.h"

#define INB(port)       inb_p(port)
#define OUTB(val,port)  outb_p(val,port)
#define OUTBP(val,port) outb_p(val,port)

#define PROMPT  OUTB(OBOE_PROMPT_BIT,OBOE_PROMPT);

#if PROBE_VERBOSE
#define PROBE_DEBUG(args...) (printk (args))
#else
#define PROBE_DEBUG(args...) ;
#endif

/* Set the DMA to be byte at a time */
#define CONFIG0H_DMA_OFF OBOE_CONFIG0H_RCVANY
#define CONFIG0H_DMA_ON_NORX CONFIG0H_DMA_OFF| OBOE_CONFIG0H_ENDMAC
#define CONFIG0H_DMA_ON CONFIG0H_DMA_ON_NORX | OBOE_CONFIG0H_ENRX

static const struct pci_device_id toshoboe_pci_tbl[] = {
	{ PCI_VENDOR_ID_TOSHIBA, PCI_DEVICE_ID_FIR701, PCI_ANY_ID, PCI_ANY_ID, },
	{ PCI_VENDOR_ID_TOSHIBA, PCI_DEVICE_ID_FIRD01, PCI_ANY_ID, PCI_ANY_ID, },
	{ }			/* Terminating entry */
};
MODULE_DEVICE_TABLE(pci, toshoboe_pci_tbl);

#define DRIVER_NAME "toshoboe"
static char *driver_name = DRIVER_NAME;

static int max_baud = 4000000;
#ifdef USE_PROBE
static bool do_probe = false;
#endif


/**********************************************************************/
static int
toshoboe_checkfcs (unsigned char *buf, int len)
{
  int i;
  union
  {
    __u16 value;
    __u8 bytes[2];
  }
  fcs;

  fcs.value = INIT_FCS;

  for (i = 0; i < len; ++i)
    fcs.value = irda_fcs (fcs.value, *(buf++));

  return fcs.value == GOOD_FCS;
}

/***********************************************************************/
/* Generic chip handling code */
#ifdef DUMP_PACKETS
static unsigned char dump[50];
static void
_dumpbufs (unsigned char *data, int len, char tete)
{
int i,j;
char head=tete;
for (i=0;i<len;i+=16) {
    for (j=0;j<16 && i+j<len;j++) { sprintf(&dump[3*j],"%02x.",data[i+j]); }
    dump [3*j]=0;
    pr_debug("%c%s\n", head, dump);
    head='+';
    }
}
#endif

#ifdef USE_PROBE
/* Dump the registers */
static void
toshoboe_dumpregs (struct toshoboe_cb *self)
{
  __u32 ringbase;

  ringbase = INB (OBOE_RING_BASE0) << 10;
  ringbase |= INB (OBOE_RING_BASE1) << 18;
  ringbase |= INB (OBOE_RING_BASE2) << 26;

  printk (KERN_ERR DRIVER_NAME ": Register dump:\n");
  printk (KERN_ERR "Interrupts: Tx:%d Rx:%d TxUnder:%d RxOver:%d Sip:%d\n",
          self->int_tx, self->int_rx, self->int_txunder, self->int_rxover,
          self->int_sip);
  printk (KERN_ERR "RX %02x TX %02x RingBase %08x\n",
          INB (OBOE_RXSLOT), INB (OBOE_TXSLOT), ringbase);
  printk (KERN_ERR "RING_SIZE %02x IER %02x ISR %02x\n",
          INB (OBOE_RING_SIZE), INB (OBOE_IER), INB (OBOE_ISR));
  printk (KERN_ERR "CONFIG1 %02x STATUS %02x\n",
          INB (OBOE_CONFIG1), INB (OBOE_STATUS));
  printk (KERN_ERR "CONFIG0 %02x%02x ENABLE %02x%02x\n",
          INB (OBOE_CONFIG0H), INB (OBOE_CONFIG0L),
          INB (OBOE_ENABLEH), INB (OBOE_ENABLEL));
  printk (KERN_ERR "NEW_PCONFIG %02x%02x CURR_PCONFIG %02x%02x\n",
          INB (OBOE_NEW_PCONFIGH), INB (OBOE_NEW_PCONFIGL),
          INB (OBOE_CURR_PCONFIGH), INB (OBOE_CURR_PCONFIGL));
  printk (KERN_ERR "MAXLEN %02x%02x RXCOUNT %02x%02x\n",
          INB (OBOE_MAXLENH), INB (OBOE_MAXLENL),
          INB (OBOE_RXCOUNTL), INB (OBOE_RXCOUNTH));

  if (self->ring)
    {
      int i;
      ringbase = virt_to_bus (self->ring);
      printk (KERN_ERR "Ring at %08x:\n", ringbase);
      printk (KERN_ERR "RX:");
      for (i = 0; i < RX_SLOTS; ++i)
        printk (" (%d,%02x)",self->ring->rx[i].len,self->ring->rx[i].control);
      printk ("\n");
      printk (KERN_ERR "TX:");
      for (i = 0; i < RX_SLOTS; ++i)
        printk (" (%d,%02x)",self->ring->tx[i].len,self->ring->tx[i].control);
      printk ("\n");
    }
}
#endif

/*Don't let the chip look at memory */
static void
toshoboe_disablebm (struct toshoboe_cb *self)
{
  __u8 command;
  pci_read_config_byte (self->pdev, PCI_COMMAND, &command);
  command &= ~PCI_COMMAND_MASTER;
  pci_write_config_byte (self->pdev, PCI_COMMAND, command);

}

/* Shutdown the chip and point the taskfile reg somewhere else */
static void
toshoboe_stopchip (struct toshoboe_cb *self)
{
  /*Disable interrupts */
  OUTB (0x0, OBOE_IER);
  /*Disable DMA, Disable Rx, Disable Tx */
  OUTB (CONFIG0H_DMA_OFF, OBOE_CONFIG0H);
  /*Disable SIR MIR FIR, Tx and Rx */
  OUTB (0x00, OBOE_ENABLEH);
  /*Point the ring somewhere safe */
  OUTB (0x3f, OBOE_RING_BASE2);
  OUTB (0xff, OBOE_RING_BASE1);
  OUTB (0xff, OBOE_RING_BASE0);

  OUTB (RX_LEN >> 8, OBOE_MAXLENH);
  OUTB (RX_LEN & 0xff, OBOE_MAXLENL);

  /*Acknoledge any pending interrupts */
  OUTB (0xff, OBOE_ISR);

  /*Why */
  OUTB (OBOE_ENABLEH_PHYANDCLOCK, OBOE_ENABLEH);

  /*switch it off */
  OUTB (OBOE_CONFIG1_OFF, OBOE_CONFIG1);

  toshoboe_disablebm (self);
}

/* Transmitter initialization */
static void
toshoboe_start_DMA (struct toshoboe_cb *self, int opts)
{
  OUTB (0x0, OBOE_ENABLEH);
  OUTB (CONFIG0H_DMA_ON | opts,  OBOE_CONFIG0H);
  OUTB (OBOE_ENABLEH_PHYANDCLOCK, OBOE_ENABLEH);
  PROMPT;
}

/*Set the baud rate */
static void
toshoboe_setbaud (struct toshoboe_cb *self)
{
  __u16 pconfig = 0;
  __u8 config0l = 0;

  pr_debug("%s(%d/%d)\n", __func__, self->speed, self->io.speed);

  switch (self->speed)
    {
    case 2400:
    case 4800:
    case 9600:
    case 19200:
    case 38400:
    case 57600:
    case 115200:
#ifdef USE_MIR
    case 1152000:
#endif
    case 4000000:
      break;
    default:

      printk (KERN_ERR DRIVER_NAME ": switch to unsupported baudrate %d\n",
              self->speed);
      return;
    }

  switch (self->speed)
    {
      /* For SIR the preamble is done by adding XBOFs */
      /* to the packet */
      /* set to filtered SIR mode, filter looks for BOF and EOF */
    case 2400:
      pconfig |= 47 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    case 4800:
      pconfig |= 23 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    case 9600:
      pconfig |= 11 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    case 19200:
      pconfig |= 5 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    case 38400:
      pconfig |= 2 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    case 57600:
      pconfig |= 1 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    case 115200:
      pconfig |= 0 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 25 << OBOE_PCONFIG_WIDTHSHIFT;
      break;
    default:
      /*Set to packet based reception */
      OUTB (RX_LEN >> 8, OBOE_MAXLENH);
      OUTB (RX_LEN & 0xff, OBOE_MAXLENL);
      break;
    }

  switch (self->speed)
    {
    case 2400:
    case 4800:
    case 9600:
    case 19200:
    case 38400:
    case 57600:
    case 115200:
      config0l = OBOE_CONFIG0L_ENSIR;
      if (self->async)
        {
          /*Set to character based reception */
          /*System will lock if MAXLEN=0 */
          /*so have to be careful */
          OUTB (0x01, OBOE_MAXLENH);
          OUTB (0x01, OBOE_MAXLENL);
          OUTB (0x00, OBOE_MAXLENH);
        }
      else
        {
          /*Set to packet based reception */
          config0l |= OBOE_CONFIG0L_ENSIRF;
          OUTB (RX_LEN >> 8, OBOE_MAXLENH);
          OUTB (RX_LEN & 0xff, OBOE_MAXLENL);
        }
      break;

#ifdef USE_MIR
      /* MIR mode */
      /* Set for 16 bit CRC and enable MIR */
      /* Preamble now handled by the chip */
    case 1152000:
      pconfig |= 0 << OBOE_PCONFIG_BAUDSHIFT;
      pconfig |= 8 << OBOE_PCONFIG_WIDTHSHIFT;
      pconfig |= 1 << OBOE_PCONFIG_PREAMBLESHIFT;
      config0l = OBOE_CONFIG0L_CRC16 | OBOE_CONFIG0L_ENMIR;
      break;
#endif
      /* FIR mode */
      /* Set for 32 bit CRC and enable FIR */
      /* Preamble handled by the chip */
    case 4000000:
      pconfig |= 0 << OBOE_PCONFIG_BAUDSHIFT;
      /* Documentation says 14, but toshiba use 15 in their drivers */
      pconfig |= 15 << OBOE_PCONFIG_PREAMBLESHIFT;
      config0l = OBOE_CONFIG0L_ENFIR;
      break;
    }

  /* Copy into new PHY config buffer */
  OUTBP (pconfig >> 8, OBOE_NEW_PCONFIGH);
  OUTB (pconfig & 0xff, OBOE_NEW_PCONFIGL);
  OUTB (config0l, OBOE_CONFIG0L);

  /* Now make OBOE copy from new PHY to current PHY */
  OUTB (0x0, OBOE_ENABLEH);
  OUTB (OBOE_ENABLEH_PHYANDCLOCK, OBOE_ENABLEH);
  PROMPT;

  /* speed change executed */
  self->new_speed = 0;
  self->io.speed = self->speed;
}

/*Let the chip look at memory */
static void
toshoboe_enablebm (struct toshoboe_cb *self)
{
  pci_set_master (self->pdev);
}

/*setup the ring */
static void
toshoboe_initring (struct toshoboe_cb *self)
{
  int i;

  for (i = 0; i < TX_SLOTS; ++i)
    {
      self->ring->tx[i].len = 0;
      self->ring->tx[i].control = 0x00;
      self->ring->tx[i].address = virt_to_bus (self->tx_bufs[i]);
    }

  for (i = 0; i < RX_SLOTS; ++i)
    {
      self->ring->rx[i].len = RX_LEN;
      self->ring->rx[i].len = 0;
      self->ring->rx[i].address = virt_to_bus (self->rx_bufs[i]);
      self->ring->rx[i].control = OBOE_CTL_RX_HW_OWNS;
    }
}

static void
toshoboe_resetptrs (struct toshoboe_cb *self)
{
  /* Can reset pointers by twidling DMA */
  OUTB (0x0, OBOE_ENABLEH);
  OUTBP (CONFIG0H_DMA_OFF, OBOE_CONFIG0H);
  OUTB (OBOE_ENABLEH_PHYANDCLOCK, OBOE_ENABLEH);

  self->rxs = inb_p (OBOE_RXSLOT) & OBOE_SLOT_MASK;
  self->txs = inb_p (OBOE_TXSLOT) & OBOE_SLOT_MASK;
}

/* Called in locked state */
static void
toshoboe_initptrs (struct toshoboe_cb *self)
{

  /* spin_lock_irqsave(self->spinlock, flags); */
  /* save_flags (flags); */

  /* Can reset pointers by twidling DMA */
  toshoboe_resetptrs (self);

  OUTB (0x0, OBOE_ENABLEH);
  OUTB (CONFIG0H_DMA_ON, OBOE_CONFIG0H);
  OUTB (OBOE_ENABLEH_PHYANDCLOCK, OBOE_ENABLEH);

  self->txpending = 0;

  /* spin_unlock_irqrestore(self->spinlock, flags); */
  /* restore_flags (flags); */
}

/* Wake the chip up and get it looking at the rings */
/* Called in locked state */
static void
toshoboe_startchip (struct toshoboe_cb *self)
{
  __u32 physaddr;

  toshoboe_initring (self);
  toshoboe_enablebm (self);
  OUTBP (OBOE_CONFIG1_RESET, OBOE_CONFIG1);
  OUTBP (OBOE_CONFIG1_ON, OBOE_CONFIG1);

  /* Stop the clocks */
  OUTB (0, OBOE_ENABLEH);

  /*Set size of rings */
  OUTB (RING_SIZE, OBOE_RING_SIZE);

  /*Acknoledge any pending interrupts */
  OUTB (0xff, OBOE_ISR);

  /*Enable ints */
  OUTB (OBOE_INT_TXDONE  | OBOE_INT_RXDONE |
        OBOE_INT_TXUNDER | OBOE_INT_RXOVER | OBOE_INT_SIP , OBOE_IER);

  /*Acknoledge any pending interrupts */
  OUTB (0xff, OBOE_ISR);

  /*Set the maximum packet length to 0xfff (4095) */
  OUTB (RX_LEN >> 8, OBOE_MAXLENH);
  OUTB (RX_LEN & 0xff, OBOE_MAXLENL);

  /*Shutdown DMA */
  OUTB (CONFIG0H_DMA_OFF, OBOE_CONFIG0H);

  /*Find out where the rings live */
  physaddr = virt_to_bus (self->ring);

  IRDA_ASSERT ((physaddr & 0x3ff) == 0,
	       printk (KERN_ERR DRIVER_NAME "ring not correctly aligned\n");
	       return;);

  OUTB ((physaddr >> 10) & 0xff, OBOE_RING_BASE0);
  OUTB ((physaddr >> 18) & 0xff, OBOE_RING_BASE1);
  OUTB ((physaddr >> 26) & 0x3f, OBOE_RING_BASE2);

  /*Enable DMA controller in byte mode and RX */
  OUTB (CONFIG0H_DMA_ON, OBOE_CONFIG0H);

  /* Start up the clocks */
  OUTB (OBOE_ENABLEH_PHYANDCLOCK, OBOE_ENABLEH);

  /*set to sensible speed */
  self->speed = 9600;
  toshoboe_setbaud (self);
  toshoboe_initptrs (self);
}

static void
toshoboe_isntstuck (struct toshoboe_cb *self)
{
}

static void
toshoboe_checkstuck (struct toshoboe_cb *self)
{
  unsigned long flags;

  if (0)
    {
      spin_lock_irqsave(&self->spinlock, flags);

      /* This will reset the chip completely */
      printk (KERN_ERR DRIVER_NAME ": Resetting chip\n");

      toshoboe_stopchip (self);
      toshoboe_startchip (self);
      spin_unlock_irqrestore(&self->spinlock, flags);
    }
}

/*Generate packet of about mtt us long */
static int
toshoboe_makemttpacket (struct toshoboe_cb *self, void *buf, int mtt)
{
  int xbofs;

  xbofs = ((int) (mtt/100)) * (int) (self->speed);
  xbofs=xbofs/80000; /*Eight bits per byte, and mtt is in us*/
  xbofs++;

  pr_debug(DRIVER_NAME ": generated mtt of %d bytes for %d us at %d baud\n",
	   xbofs, mtt, self->speed);

  if (xbofs > TX_LEN)
    {
      printk (KERN_ERR DRIVER_NAME ": wanted %d bytes MTT but TX_LEN is %d\n",
              xbofs, TX_LEN);
      xbofs = TX_LEN;
    }

  /*xbofs will do for SIR, MIR and FIR,SIR mode doesn't generate a checksum anyway */
  memset (buf, XBOF, xbofs);

  return xbofs;
}

#ifdef USE_PROBE
/***********************************************************************/
/* Probe code */

static void
toshoboe_dumptx (struct toshoboe_cb *self)
{
  int i;
  PROBE_DEBUG(KERN_WARNING "TX:");
  for (i = 0; i < RX_SLOTS; ++i)
    PROBE_DEBUG(" (%d,%02x)",self->ring->tx[i].len,self->ring->tx[i].control);
  PROBE_DEBUG(" [%d]\n",self->speed);
}

static void
toshoboe_dumprx (struct toshoboe_cb *self, int score)
{
  int i;
  PROBE_DEBUG(" %d\nRX:",score);
  for (i = 0; i < RX_SLOTS; ++i)
    PROBE_DEBUG(" (%d,%02x)",self->ring->rx[i].len,self->ring->rx[i].control);
  PROBE_DEBUG("\n");
}

static inline int
stuff_byte (__u8 byte, __u8 * buf)
{
  switch (byte)
    {
    case BOF:                  /* FALLTHROUGH */
    case EOF:                  /* FALLTHROUGH */
    case CE:
      /* Insert transparently coded */
      buf[0] = CE;              /* Send link escape */
      buf[1] = byte ^ IRDA_TRANS; /* Complement bit 5 */
      return 2;
      /* break; */
    default:
      /* Non-special value, no transparency required */
      buf[0] = byte;
      return 1;
      /* break; */
    }
}

static irqreturn_t
toshoboe_probeinterrupt (int irq, void *dev_id)
{
  struct toshoboe_cb *self = dev_id;
  __u8 irqstat;

  irqstat = INB (OBOE_ISR);

/* was it us */
  if (!(irqstat & OBOE_INT_MASK))
    return IRQ_NONE;

/* Ack all the interrupts */
  OUTB (irqstat, OBOE_ISR);

  if (irqstat & OBOE_INT_TXDONE)
    {
      int txp;

      self->int_tx++;
      PROBE_DEBUG("T");

      txp = INB (OBOE_TXSLOT) & OBOE_SLOT_MASK;
      if (self->ring->tx[txp].control & OBOE_CTL_TX_HW_OWNS)
        {
          self->int_tx+=100;
          PROBE_DEBUG("S");
          toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX | OBOE_CONFIG0H_LOOP);
        }
    }

  if (irqstat & OBOE_INT_RXDONE) {
    self->int_rx++;
    PROBE_DEBUG("R"); }
  if (irqstat & OBOE_INT_TXUNDER) {
    self->int_txunder++;
    PROBE_DEBUG("U"); }
  if (irqstat & OBOE_INT_RXOVER) {
    self->int_rxover++;
    PROBE_DEBUG("O"); }
  if (irqstat & OBOE_INT_SIP) {
    self->int_sip++;
    PROBE_DEBUG("I"); }
  return IRQ_HANDLED;
}

static int
toshoboe_maketestpacket (unsigned char *buf, int badcrc, int fir)
{
  int i;
  int len = 0;
  union
  {
    __u16 value;
    __u8 bytes[2];
  }
  fcs;

  if (fir)
    {
      memset (buf, 0, TT_LEN);
      return TT_LEN;
    }

  fcs.value = INIT_FCS;

  memset (buf, XBOF, 10);
  len += 10;
  buf[len++] = BOF;

  for (i = 0; i < TT_LEN; ++i)
    {
      len += stuff_byte (i, buf + len);
      fcs.value = irda_fcs (fcs.value, i);
    }

  len += stuff_byte (fcs.bytes[0] ^ badcrc, buf + len);
  len += stuff_byte (fcs.bytes[1] ^ badcrc, buf + len);
  buf[len++] = EOF;
  len++;
  return len;
}

static int
toshoboe_probefail (struct toshoboe_cb *self, char *msg)
{
  printk (KERN_ERR DRIVER_NAME "probe(%d) failed %s\n",self-> speed, msg);
  toshoboe_dumpregs (self);
  toshoboe_stopchip (self);
  free_irq (self->io.irq, (void *) self);
  return 0;
}

static int
toshoboe_numvalidrcvs (struct toshoboe_cb *self)
{
  int i, ret = 0;
  for (i = 0; i < RX_SLOTS; ++i)
    if ((self->ring->rx[i].control & 0xe0) == 0)
      ret++;

  return ret;
}

static int
toshoboe_numrcvs (struct toshoboe_cb *self)
{
  int i, ret = 0;
  for (i = 0; i < RX_SLOTS; ++i)
    if (!(self->ring->rx[i].control & OBOE_CTL_RX_HW_OWNS))
      ret++;

  return ret;
}

static int
toshoboe_probe (struct toshoboe_cb *self)
{
  int i, j, n;
#ifdef USE_MIR
  static const int bauds[] = { 9600, 115200, 4000000, 1152000 };
#else
  static const int bauds[] = { 9600, 115200, 4000000 };
#endif
  unsigned long flags;

  if (request_irq (self->io.irq, toshoboe_probeinterrupt,
                   self->io.irqflags, "toshoboe", (void *) self))
    {
      printk (KERN_ERR DRIVER_NAME ": probe failed to allocate irq %d\n",
              self->io.irq);
      return 0;
    }

  /* test 1: SIR filter and back to back */

  for (j = 0; j < ARRAY_SIZE(bauds); ++j)
    {
      int fir = (j > 1);
      toshoboe_stopchip (self);


      spin_lock_irqsave(&self->spinlock, flags);
      /*Address is already setup */
      toshoboe_startchip (self);
      self->int_rx = self->int_tx = 0;
      self->speed = bauds[j];
      toshoboe_setbaud (self);
      toshoboe_initptrs (self);
      spin_unlock_irqrestore(&self->spinlock, flags);

      self->ring->tx[self->txs].control =
/*   (FIR only) OBOE_CTL_TX_SIP needed for switching to next slot */
/*    MIR: all received data is stored in one slot */
        (fir) ? OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX
              : OBOE_CTL_TX_HW_OWNS ;
      self->ring->tx[self->txs].len =
        toshoboe_maketestpacket (self->tx_bufs[self->txs], 0, fir);
      self->txs++;
      self->txs %= TX_SLOTS;

      self->ring->tx[self->txs].control =
        (fir) ? OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_SIP
              : OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX ;
      self->ring->tx[self->txs].len =
        toshoboe_maketestpacket (self->tx_bufs[self->txs], 0, fir);
      self->txs++;
      self->txs %= TX_SLOTS;

      self->ring->tx[self->txs].control =
        (fir) ? OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX
              : OBOE_CTL_TX_HW_OWNS ;
      self->ring->tx[self->txs].len =
        toshoboe_maketestpacket (self->tx_bufs[self->txs], 0, fir);
      self->txs++;
      self->txs %= TX_SLOTS;

      self->ring->tx[self->txs].control =
        (fir) ? OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX
              | OBOE_CTL_TX_SIP     | OBOE_CTL_TX_BAD_CRC
              : OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX ;
      self->ring->tx[self->txs].len =
        toshoboe_maketestpacket (self->tx_bufs[self->txs], 0, fir);
      self->txs++;
      self->txs %= TX_SLOTS;

      toshoboe_dumptx (self);
      /* Turn on TX and RX and loopback */
      toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX | OBOE_CONFIG0H_LOOP);

      i = 0;
      n = fir ? 1 : 4;
      while (toshoboe_numvalidrcvs (self) != n)
        {
          if (i > 4800)
              return toshoboe_probefail (self, "filter test");
          udelay ((9600*(TT_LEN+16))/self->speed);
          i++;
        }

      n = fir ? 203 : 102;
      while ((toshoboe_numrcvs(self) != self->int_rx) || (self->int_tx != n))
        {
          if (i > 4800)
              return toshoboe_probefail (self, "interrupt test");
          udelay ((9600*(TT_LEN+16))/self->speed);
          i++;
        }
     toshoboe_dumprx (self,i);

     }

  /* test 2: SIR in char at a time */

  toshoboe_stopchip (self);
  self->int_rx = self->int_tx = 0;

  spin_lock_irqsave(&self->spinlock, flags);
  toshoboe_startchip (self);
  spin_unlock_irqrestore(&self->spinlock, flags);

  self->async = 1;
  self->speed = 115200;
  toshoboe_setbaud (self);
  self->ring->tx[self->txs].control =
    OBOE_CTL_TX_RTCENTX | OBOE_CTL_TX_HW_OWNS;
  self->ring->tx[self->txs].len = 4;

  ((unsigned char *) self->tx_bufs[self->txs])[0] = 'f';
  ((unsigned char *) self->tx_bufs[self->txs])[1] = 'i';
  ((unsigned char *) self->tx_bufs[self->txs])[2] = 's';
  ((unsigned char *) self->tx_bufs[self->txs])[3] = 'h';
  toshoboe_dumptx (self);
  toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX | OBOE_CONFIG0H_LOOP);

  i = 0;
  while (toshoboe_numvalidrcvs (self) != 4)
    {
      if (i > 100)
          return toshoboe_probefail (self, "Async test");
      udelay (100);
      i++;
    }

  while ((toshoboe_numrcvs (self) != self->int_rx) || (self->int_tx != 1))
    {
      if (i > 100)
          return toshoboe_probefail (self, "Async interrupt test");
      udelay (100);
      i++;
    }
  toshoboe_dumprx (self,i);

  self->async = 0;
  self->speed = 9600;
  toshoboe_setbaud (self);
  toshoboe_stopchip (self);

  free_irq (self->io.irq, (void *) self);

  printk (KERN_WARNING DRIVER_NAME ": Self test passed ok\n");

  return 1;
}
#endif

/******************************************************************/
/* Netdev style code */

/* Transmit something */
static netdev_tx_t
toshoboe_hard_xmit (struct sk_buff *skb, struct net_device *dev)
{
  struct toshoboe_cb *self;
  __s32 speed;
  int mtt, len, ctl;
  unsigned long flags;
  struct irda_skb_cb *cb = (struct irda_skb_cb *) skb->cb;

  self = netdev_priv(dev);

  IRDA_ASSERT (self != NULL, return NETDEV_TX_OK; );

  pr_debug("%s.tx:%x(%x)%x\n",
	   __func__, skb->len, self->txpending, INB(OBOE_ENABLEH));
  if (!cb->magic) {
	  pr_debug("%s.Not IrLAP:%x\n", __func__, cb->magic);
#ifdef DUMP_PACKETS
      _dumpbufs(skb->data,skb->len,'>');
#endif
    }

  /* change speed pending, wait for its execution */
  if (self->new_speed)
      return NETDEV_TX_BUSY;

  /* device stopped (apm) wait for restart */
  if (self->stopped)
      return NETDEV_TX_BUSY;

  toshoboe_checkstuck (self);

 /* Check if we need to change the speed */
  /* But not now. Wait after transmission if mtt not required */
  speed=irda_get_next_speed(skb);
  if ((speed != self->io.speed) && (speed != -1))
    {
      spin_lock_irqsave(&self->spinlock, flags);

      if (self->txpending || skb->len)
        {
          self->new_speed = speed;
	  pr_debug("%s: Queued TxDone scheduled speed change %d\n" ,
		   __func__, speed);
          /* if no data, that's all! */
          if (!skb->len)
            {
	      spin_unlock_irqrestore(&self->spinlock, flags);
              dev_kfree_skb (skb);
              return NETDEV_TX_OK;
            }
          /* True packet, go on, but */
          /* do not accept anything before change speed execution */
          netif_stop_queue(dev);
          /* ready to process TxDone interrupt */
	  spin_unlock_irqrestore(&self->spinlock, flags);
        }
      else
        {
          /* idle and no data, change speed now */
          self->speed = speed;
          toshoboe_setbaud (self);
	  spin_unlock_irqrestore(&self->spinlock, flags);
          dev_kfree_skb (skb);
          return NETDEV_TX_OK;
        }

    }

  if ((mtt = irda_get_mtt(skb)))
    {
      /* This is fair since the queue should be empty anyway */
      spin_lock_irqsave(&self->spinlock, flags);

      if (self->txpending)
        {
	  spin_unlock_irqrestore(&self->spinlock, flags);
          return NETDEV_TX_BUSY;
        }

      /* If in SIR mode we need to generate a string of XBOFs */
      /* In MIR and FIR we need to generate a string of data */
      /* which we will add a wrong checksum to */

      mtt = toshoboe_makemttpacket (self, self->tx_bufs[self->txs], mtt);
      pr_debug("%s.mtt:%x(%x)%d\n", __func__, skb->len, mtt, self->txpending);
      if (mtt)
        {
          self->ring->tx[self->txs].len = mtt & 0xfff;

          ctl = OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX;
          if (INB (OBOE_ENABLEH) & OBOE_ENABLEH_FIRON)
            {
              ctl |= OBOE_CTL_TX_BAD_CRC | OBOE_CTL_TX_SIP ;
            }
#ifdef USE_MIR
          else if (INB (OBOE_ENABLEH) & OBOE_ENABLEH_MIRON)
            {
              ctl |= OBOE_CTL_TX_BAD_CRC;
            }
#endif
          self->ring->tx[self->txs].control = ctl;

          OUTB (0x0, OBOE_ENABLEH);
          /* It is only a timer. Do not send mtt packet outside! */
          toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX | OBOE_CONFIG0H_LOOP);

          self->txpending++;

          self->txs++;
          self->txs %= TX_SLOTS;

        }
      else
        {
          printk(KERN_ERR DRIVER_NAME ": problem with mtt packet - ignored\n");
        }
      spin_unlock_irqrestore(&self->spinlock, flags);
    }

#ifdef DUMP_PACKETS
dumpbufs(skb->data,skb->len,'>');
#endif

  spin_lock_irqsave(&self->spinlock, flags);

  if (self->ring->tx[self->txs].control & OBOE_CTL_TX_HW_OWNS)
    {
	    pr_debug("%s.ful:%x(%x)%x\n",
		     __func__, skb->len, self->ring->tx[self->txs].control,
		     self->txpending);
      toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX);
      spin_unlock_irqrestore(&self->spinlock, flags);
      return NETDEV_TX_BUSY;
    }

  if (INB (OBOE_ENABLEH) & OBOE_ENABLEH_SIRON)
    {
      len = async_wrap_skb (skb, self->tx_bufs[self->txs], TX_BUF_SZ);
    }
  else
    {
      len = skb->len;
      skb_copy_from_linear_data(skb, self->tx_bufs[self->txs], len);
    }
  self->ring->tx[self->txs].len = len & 0x0fff;

  /*Sometimes the HW doesn't see us assert RTCENTX in the interrupt code */
  /*later this plays safe, we garuntee the last packet to be transmitted */
  /*has RTCENTX set */

  ctl = OBOE_CTL_TX_HW_OWNS | OBOE_CTL_TX_RTCENTX;
  if (INB (OBOE_ENABLEH) & OBOE_ENABLEH_FIRON)
    {
      ctl |= OBOE_CTL_TX_SIP ;
    }
  self->ring->tx[self->txs].control = ctl;

  /* If transmitter is idle start in one-shot mode */

  if (!self->txpending)
      toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX);

  self->txpending++;

  self->txs++;
  self->txs %= TX_SLOTS;

  spin_unlock_irqrestore(&self->spinlock, flags);
  dev_kfree_skb (skb);

  return NETDEV_TX_OK;
}

/*interrupt handler */
static irqreturn_t
toshoboe_interrupt (int irq, void *dev_id)
{
  struct toshoboe_cb *self = dev_id;
  __u8 irqstat;
  struct sk_buff *skb = NULL;

  irqstat = INB (OBOE_ISR);

/* was it us */
  if (!(irqstat & OBOE_INT_MASK))
      return IRQ_NONE;

/* Ack all the interrupts */
  OUTB (irqstat, OBOE_ISR);

  toshoboe_isntstuck (self);

/* Txdone */
  if (irqstat & OBOE_INT_TXDONE)
    {
      int txp, txpc;
      int i;

      txp = self->txpending;
      self->txpending = 0;

      for (i = 0; i < TX_SLOTS; ++i)
        {
          if (self->ring->tx[i].control & OBOE_CTL_TX_HW_OWNS)
              self->txpending++;
        }
      pr_debug("%s.txd(%x)%x/%x\n", __func__, irqstat, txp, self->txpending);

      txp = INB (OBOE_TXSLOT) & OBOE_SLOT_MASK;

      /* Got anything queued ? start it together */
      if (self->ring->tx[txp].control & OBOE_CTL_TX_HW_OWNS)
        {
          txpc = txp;
#ifdef OPTIMIZE_TX
          while (self->ring->tx[txpc].control & OBOE_CTL_TX_HW_OWNS)
            {
              txp = txpc;
              txpc++;
              txpc %= TX_SLOTS;
              self->netdev->stats.tx_packets++;
              if (self->ring->tx[txpc].control & OBOE_CTL_TX_HW_OWNS)
                  self->ring->tx[txp].control &= ~OBOE_CTL_TX_RTCENTX;
            }
          self->netdev->stats.tx_packets--;
#else
          self->netdev->stats.tx_packets++;
#endif
          toshoboe_start_DMA(self, OBOE_CONFIG0H_ENTX);
        }

      if ((!self->txpending) && (self->new_speed))
        {
          self->speed = self->new_speed;
	  pr_debug("%s: Executed TxDone scheduled speed change %d\n",
		   __func__, self->speed);
          toshoboe_setbaud (self);
        }

      /* Tell network layer that we want more frames */
      if (!self->new_speed)
          netif_wake_queue(self->netdev);
    }

  if (irqstat & OBOE_INT_RXDONE)
    {
      while (!(self->ring->rx[self->rxs].control & OBOE_CTL_RX_HW_OWNS))
        {
          int len = self->ring->rx[self->rxs].len;
          skb = NULL;
	  pr_debug("%s.rcv:%x(%x)\n", __func__
		   , len, self->ring->rx[self->rxs].control);

#ifdef DUMP_PACKETS
dumpbufs(self->rx_bufs[self->rxs],len,'<');
#endif

          if (self->ring->rx[self->rxs].control == 0)
            {
              __u8 enable = INB (OBOE_ENABLEH);

              /* In SIR mode we need to check the CRC as this */
              /* hasn't been done by the hardware */
              if (enable & OBOE_ENABLEH_SIRON)
                {
                  if (!toshoboe_checkfcs (self->rx_bufs[self->rxs], len))
                      len = 0;
                  /*Trim off the CRC */
                  if (len > 1)
                      len -= 2;
                  else
                      len = 0;
                  pr_debug("%s.SIR:%x(%x)\n", __func__, len, enable);
                }

#ifdef USE_MIR
              else if (enable & OBOE_ENABLEH_MIRON)
                {
                  if (len > 1)
                      len -= 2;
                  else
                      len = 0;
                  pr_debug("%s.MIR:%x(%x)\n", __func__, len, enable);
                }
#endif
              else if (enable & OBOE_ENABLEH_FIRON)
                {
                  if (len > 3)
                      len -= 4;   /*FIXME: check this */
                  else
                      len = 0;
                  pr_debug("%s.FIR:%x(%x)\n", __func__, len, enable);
                }
              else
		      pr_debug("%s.?IR:%x(%x)\n", __func__, len, enable);

              if (len)
                {
                  skb = dev_alloc_skb (len + 1);
                  if (skb)
                    {
                      skb_reserve (skb, 1);

                      skb_put (skb, len);
                      skb_copy_to_linear_data(skb, self->rx_bufs[self->rxs],
					      len);
                      self->netdev->stats.rx_packets++;
                      skb->dev = self->netdev;
                      skb_reset_mac_header(skb);
                      skb->protocol = htons (ETH_P_IRDA);
                    }
                  else
                    {
                      printk (KERN_INFO
                              "%s(), memory squeeze, dropping frame.\n",
			      __func__);
                    }
                }
            }
          else
            {
            /* TODO: =========================================== */
            /*  if OBOE_CTL_RX_LENGTH, our buffers are too small */
            /* (MIR or FIR) data is lost. */
            /* (SIR) data is splitted in several slots. */
            /* we have to join all the received buffers received */
            /*in a large buffer before checking CRC. */
		    pr_debug("%s.err:%x(%x)\n", __func__
			     , len, self->ring->rx[self->rxs].control);
            }

          self->ring->rx[self->rxs].len = 0x0;
          self->ring->rx[self->rxs].control = OBOE_CTL_RX_HW_OWNS;

          self->rxs++;
          self->rxs %= RX_SLOTS;

          if (skb)
              netif_rx (skb);

        }
    }

  if (irqstat & OBOE_INT_TXUNDER)
    {
      printk (KERN_WARNING DRIVER_NAME ": tx fifo underflow\n");
    }
  if (irqstat & OBOE_INT_RXOVER)
    {
      printk (KERN_WARNING DRIVER_NAME ": rx fifo overflow\n");
    }
/* This must be useful for something... */
  if (irqstat & OBOE_INT_SIP)
    {
      self->int_sip++;
      pr_debug("%s.sip:%x(%x)%x\n",
	       __func__, self->int_sip, irqstat, self->txpending);
    }
  return IRQ_HANDLED;
}


static int
toshoboe_net_open (struct net_device *dev)
{
  struct toshoboe_cb *self;
  unsigned long flags;
  int rc;

  self = netdev_priv(dev);

  if (self->async)
    return -EBUSY;

  if (self->stopped)
    return 0;

  rc = request_irq (self->io.irq, toshoboe_interrupt,
                    IRQF_SHARED, dev->name, self);
  if (rc)
  	return rc;

  spin_lock_irqsave(&self->spinlock, flags);
  toshoboe_startchip (self);
  spin_unlock_irqrestore(&self->spinlock, flags);

  /* Ready to play! */
  netif_start_queue(dev);

  /*
   * Open new IrLAP layer instance, now that everything should be
   * initialized properly
   */
  self->irlap = irlap_open (dev, &self->qos, driver_name);

  self->irdad = 1;

  return 0;
}

static int
toshoboe_net_close (struct net_device *dev)
{
  struct toshoboe_cb *self;

  IRDA_ASSERT (dev != NULL, return -1; );
  self = netdev_priv(dev);

  /* Stop device */
  netif_stop_queue(dev);

  /* Stop and remove instance of IrLAP */
  if (self->irlap)
    irlap_close (self->irlap);
  self->irlap = NULL;

  self->irdad = 0;

  free_irq (self->io.irq, (void *) self);

  if (!self->stopped)
    {
      toshoboe_stopchip (self);
    }

  return 0;
}

/*
 * Function toshoboe_net_ioctl (dev, rq, cmd)
 *
 *    Process IOCTL commands for this device
 *
 */
static int
toshoboe_net_ioctl (struct net_device *dev, struct ifreq *rq, int cmd)
{
  struct if_irda_req *irq = (struct if_irda_req *) rq;
  struct toshoboe_cb *self;
  unsigned long flags;
  int ret = 0;

  IRDA_ASSERT (dev != NULL, return -1; );

  self = netdev_priv(dev);

  IRDA_ASSERT (self != NULL, return -1; );

  pr_debug("%s(), %s, (cmd=0x%X)\n", __func__, dev->name, cmd);

  /* Disable interrupts & save flags */
  spin_lock_irqsave(&self->spinlock, flags);

  switch (cmd)
    {
    case SIOCSBANDWIDTH:       /* Set bandwidth */
      /* This function will also be used by IrLAP to change the
       * speed, so we still must allow for speed change within
       * interrupt context.
       */
	    pr_debug("%s(BANDWIDTH), %s, (%X/%ld\n",
		     __func__, dev->name, INB(OBOE_STATUS), irq->ifr_baudrate);
      if (!in_interrupt () && !capable (CAP_NET_ADMIN)) {
	ret = -EPERM;
	goto out;
      }

      /* self->speed=irq->ifr_baudrate; */
      /* toshoboe_setbaud(self); */
      /* Just change speed once - inserted by Paul Bristow */
      self->new_speed = irq->ifr_baudrate;
      break;
    case SIOCSMEDIABUSY:       /* Set media busy */
	    pr_debug("%s(MEDIABUSY), %s, (%X/%x)\n",
		     __func__, dev->name,
		     INB(OBOE_STATUS), capable(CAP_NET_ADMIN));
      if (!capable (CAP_NET_ADMIN)) {
	ret = -EPERM;
	goto out;
      }
      irda_device_set_media_busy (self->netdev, TRUE);
      break;
    case SIOCGRECEIVING:       /* Check if we are receiving right now */
      irq->ifr_receiving = (INB (OBOE_STATUS) & OBOE_STATUS_RXBUSY) ? 1 : 0;
	    pr_debug("%s(RECEIVING), %s, (%X/%x)\n",
		     __func__, dev->name, INB(OBOE_STATUS), irq->ifr_receiving);
      break;
    default:
	    pr_debug("%s(?), %s, (cmd=0x%X)\n", __func__, dev->name, cmd);
      ret = -EOPNOTSUPP;
    }
out:
  spin_unlock_irqrestore(&self->spinlock, flags);
  return ret;

}

MODULE_DESCRIPTION("Toshiba OBOE IrDA Device Driver");
MODULE_AUTHOR("James McKenzie <james@fishsoup.dhs.org>");
MODULE_LICENSE("GPL");

module_param (max_baud, int, 0);
MODULE_PARM_DESC(max_baud, "Maximum baud rate");

#ifdef USE_PROBE
module_param (do_probe, bool, 0);
MODULE_PARM_DESC(do_probe, "Enable/disable chip probing and self-test");
#endif

static void
toshoboe_close (struct pci_dev *pci_dev)
{
  int i;
  struct toshoboe_cb *self = pci_get_drvdata(pci_dev);

  IRDA_ASSERT (self != NULL, return; );

  if (!self->stopped)
    {
      toshoboe_stopchip (self);
    }

  release_region (self->io.fir_base, self->io.fir_ext);

  for (i = 0; i < TX_SLOTS; ++i)
    {
      kfree (self->tx_bufs[i]);
      self->tx_bufs[i] = NULL;
    }

  for (i = 0; i < RX_SLOTS; ++i)
    {
      kfree (self->rx_bufs[i]);
      self->rx_bufs[i] = NULL;
    }

  unregister_netdev(self->netdev);

  kfree (self->ringbuf);
  self->ringbuf = NULL;
  self->ring = NULL;

  free_netdev(self->netdev);
}

static const struct net_device_ops toshoboe_netdev_ops = {
	.ndo_open	= toshoboe_net_open,
	.ndo_stop	= toshoboe_net_close,
	.ndo_start_xmit	= toshoboe_hard_xmit,
	.ndo_do_ioctl	= toshoboe_net_ioctl,
};

static int
toshoboe_open (struct pci_dev *pci_dev, const struct pci_device_id *pdid)
{
  struct toshoboe_cb *self;
  struct net_device *dev;
  int i = 0;
  int ok = 0;
  int err;

  if ((err=pci_enable_device(pci_dev)))
    return err;

  dev = alloc_irdadev(sizeof (struct toshoboe_cb));
  if (dev == NULL)
    {
      printk (KERN_ERR DRIVER_NAME ": can't allocate memory for "
              "IrDA control block\n");
      return -ENOMEM;
    }

  self = netdev_priv(dev);
  self->netdev = dev;
  self->pdev = pci_dev;
  self->base = pci_resource_start(pci_dev,0);

  self->io.fir_base = self->base;
  self->io.fir_ext = OBOE_IO_EXTENT;
  self->io.irq = pci_dev->irq;
  self->io.irqflags = IRQF_SHARED;

  self->speed = self->io.speed = 9600;
  self->async = 0;

  /* Lock the port that we need */
  if (NULL==request_region (self->io.fir_base, self->io.fir_ext, driver_name))
    {
      printk (KERN_ERR DRIVER_NAME ": can't get iobase of 0x%03x\n"
	      ,self->io.fir_base);
      err = -EBUSY;
      goto freeself;
    }

  spin_lock_init(&self->spinlock);

  irda_init_max_qos_capabilies (&self->qos);
  self->qos.baud_rate.bits = 0;

  if (max_baud >= 2400)
    self->qos.baud_rate.bits |= IR_2400;
  /*if (max_baud>=4800) idev->qos.baud_rate.bits|=IR_4800; */
  if (max_baud >= 9600)
    self->qos.baud_rate.bits |= IR_9600;
  if (max_baud >= 19200)
    self->qos.baud_rate.bits |= IR_19200;
  if (max_baud >= 115200)
    self->qos.baud_rate.bits |= IR_115200;
#ifdef USE_MIR
  if (max_baud >= 1152000)
    {
      self->qos.baud_rate.bits |= IR_1152000;
    }
#endif
  if (max_baud >= 4000000)
    {
      self->qos.baud_rate.bits |= (IR_4000000 << 8);
    }

  /*FIXME: work this out... */
  self->qos.min_turn_time.bits = 0xff;

  irda_qos_bits_to_value (&self->qos);

  /* Allocate twice the size to guarantee alignment */
  self->ringbuf = kmalloc(OBOE_RING_LEN << 1, GFP_KERNEL);
  if (!self->ringbuf)
    {
      err = -ENOMEM;
      goto freeregion;
    }

#if (BITS_PER_LONG == 64)
#error broken on 64-bit:  casts pointer to 32-bit, and then back to pointer.
#endif

  /*We need to align the taskfile on a taskfile size boundary */
  {
    unsigned long addr;

    addr = (__u32) self->ringbuf;
    addr &= ~(OBOE_RING_LEN - 1);
    addr += OBOE_RING_LEN;
    self->ring = (struct OboeRing *) addr;
  }

  memset (self->ring, 0, OBOE_RING_LEN);
  self->io.mem_base = (__u32) self->ring;

  ok = 1;
  for (i = 0; i < TX_SLOTS; ++i)
    {
      self->tx_bufs[i] = kmalloc (TX_BUF_SZ, GFP_KERNEL);
      if (!self->tx_bufs[i])
        ok = 0;
    }

  for (i = 0; i < RX_SLOTS; ++i)
    {
      self->rx_bufs[i] = kmalloc (RX_BUF_SZ, GFP_KERNEL);
      if (!self->rx_bufs[i])
        ok = 0;
    }

  if (!ok)
    {
      err = -ENOMEM;
      goto freebufs;
    }


#ifdef USE_PROBE
  if (do_probe)
    if (!toshoboe_probe (self))
      {
        err = -ENODEV;
        goto freebufs;
      }
#endif

  SET_NETDEV_DEV(dev, &pci_dev->dev);
  dev->netdev_ops = &toshoboe_netdev_ops;

  err = register_netdev(dev);
  if (err)
    {
      printk (KERN_ERR DRIVER_NAME ": register_netdev() failed\n");
      err = -ENOMEM;
      goto freebufs;
    }
  printk (KERN_INFO "IrDA: Registered device %s\n", dev->name);

  pci_set_drvdata(pci_dev,self);

  printk (KERN_INFO DRIVER_NAME ": Using multiple tasks\n");

  return 0;

freebufs:
  for (i = 0; i < TX_SLOTS; ++i)
    kfree (self->tx_bufs[i]);
  for (i = 0; i < RX_SLOTS; ++i)
    kfree (self->rx_bufs[i]);
  kfree(self->ringbuf);

freeregion:
  release_region (self->io.fir_base, self->io.fir_ext);

freeself:
  free_netdev(dev);

  return err;
}

static int
toshoboe_gotosleep (struct pci_dev *pci_dev, pm_message_t crap)
{
  struct toshoboe_cb *self = pci_get_drvdata(pci_dev);
  unsigned long flags;
  int i = 10;

  if (!self || self->stopped)
    return 0;

  if ((!self->irdad) && (!self->async))
    return 0;

/* Flush all packets */
  while ((i--) && (self->txpending))
    msleep(10);

  spin_lock_irqsave(&self->spinlock, flags);

  toshoboe_stopchip (self);
  self->stopped = 1;
  self->txpending = 0;

  spin_unlock_irqrestore(&self->spinlock, flags);
  return 0;
}

static int
toshoboe_wakeup (struct pci_dev *pci_dev)
{
  struct toshoboe_cb *self = pci_get_drvdata(pci_dev);
  unsigned long flags;

  if (!self || !self->stopped)
    return 0;

  if ((!self->irdad) && (!self->async))
    return 0;

  spin_lock_irqsave(&self->spinlock, flags);

  toshoboe_startchip (self);
  self->stopped = 0;

  netif_wake_queue(self->netdev);
  spin_unlock_irqrestore(&self->spinlock, flags);
  return 0;
}

static struct pci_driver donauboe_pci_driver = {
	.name		= "donauboe",
	.id_table	= toshoboe_pci_tbl,
	.probe		= toshoboe_open,
	.remove		= toshoboe_close,
	.suspend	= toshoboe_gotosleep,
	.resume		= toshoboe_wakeup 
};

module_pci_driver(donauboe_pci_driver);