1 Linux for S/390 and zSeries
3 Common Device Support (CDS)
4 Device Driver I/O Support Routines
9 Copyright, IBM Corp. 1999-2002
13 This document describes the common device support routines for Linux/390.
14 Different than other hardware architectures, ESA/390 has defined a unified
15 I/O access method. This gives relief to the device drivers as they don't
16 have to deal with different bus types, polling versus interrupt
17 processing, shared versus non-shared interrupt processing, DMA versus port
18 I/O (PIO), and other hardware features more. However, this implies that
19 either every single device driver needs to implement the hardware I/O
20 attachment functionality itself, or the operating system provides for a
21 unified method to access the hardware, providing all the functionality that
22 every single device driver would have to provide itself.
24 The document does not intend to explain the ESA/390 hardware architecture in
25 every detail.This information can be obtained from the ESA/390 Principles of
26 Operation manual (IBM Form. No. SA22-7201).
28 In order to build common device support for ESA/390 I/O interfaces, a
29 functional layer was introduced that provides generic I/O access methods to
32 The common device support layer comprises the I/O support routines defined
33 below. Some of them implement common Linux device driver interfaces, while
34 some of them are ESA/390 platform specific.
37 In order to write a driver for S/390, you also need to look into the interface
38 described in Documentation/s390/driver-model.txt.
40 Note for porting drivers from 2.4:
41 The major changes are:
42 * The functions use a ccw_device instead of an irq (subchannel).
43 * All drivers must define a ccw_driver (see driver-model.txt) and the associated
45 * request_irq() and free_irq() are no longer done by the driver.
46 * The oper_handler is (kindof) replaced by the probe() and set_online() functions
48 * The not_oper_handler is (kindof) replaced by the remove() and set_offline()
49 functions of the ccw_driver.
50 * The channel device layer is gone.
51 * The interrupt handlers must be adapted to use a ccw_device as argument.
52 Moreover, they don't return a devstat, but an irb.
53 * Before initiating an io, the options must be set via ccw_device_set_options().
54 * Instead of calling read_dev_chars()/read_conf_data(), the driver issues
55 the channel program and handles the interrupt itself.
58 get commands from extended sense data.
61 ccw_device_start_timeout()
62 ccw_device_start_key()
63 ccw_device_start_key_timeout()
64 initiate an I/O request.
67 resume channel program execution.
70 terminate the current I/O request processed on the device.
73 generic interrupt routine. This function is called by the interrupt entry
74 routine whenever an I/O interrupt is presented to the system. The do_IRQ()
75 routine determines the interrupt status and calls the device specific
76 interrupt handler according to the rules (flags) defined during I/O request
77 initiation with do_IO().
79 The next chapters describe the functions other than do_IRQ() in more details.
80 The do_IRQ() interface is not described, as it is called from the Linux/390
81 first level interrupt handler only and does not comprise a device driver
82 callable interface. Instead, the functional description of do_IO() also
83 describes the input to the device specific interrupt handler.
85 Note: All explanations apply also to the 64 bit architecture s390x.
88 Common Device Support (CDS) for Linux/390 Device Drivers
92 The following chapters describe the I/O related interface routines the
93 Linux/390 common device support (CDS) provides to allow for device specific
94 driver implementations on the IBM ESA/390 hardware platform. Those interfaces
95 intend to provide the functionality required by every device driver
96 implementation to allow to drive a specific hardware device on the ESA/390
97 platform. Some of the interface routines are specific to Linux/390 and some
98 of them can be found on other Linux platforms implementations too.
99 Miscellaneous function prototypes, data declarations, and macro definitions
100 can be found in the architecture specific C header file
101 linux/include/asm-s390/irq.h.
103 Overview of CDS interface concepts
105 Different to other hardware platforms, the ESA/390 architecture doesn't define
106 interrupt lines managed by a specific interrupt controller and bus systems
107 that may or may not allow for shared interrupts, DMA processing, etc.. Instead,
108 the ESA/390 architecture has implemented a so called channel subsystem, that
109 provides a unified view of the devices physically attached to the systems.
110 Though the ESA/390 hardware platform knows about a huge variety of different
111 peripheral attachments like disk devices (aka. DASDs), tapes, communication
112 controllers, etc. they can all be accessed by a well defined access method and
113 they are presenting I/O completion a unified way : I/O interruptions. Every
114 single device is uniquely identified to the system by a so called subchannel,
115 where the ESA/390 architecture allows for 64k devices be attached.
117 Linux, however, was first built on the Intel PC architecture, with its two
118 cascaded 8259 programmable interrupt controllers (PICs), that allow for a
119 maximum of 15 different interrupt lines. All devices attached to such a system
120 share those 15 interrupt levels. Devices attached to the ISA bus system must
121 not share interrupt levels (aka. IRQs), as the ISA bus bases on edge triggered
122 interrupts. MCA, EISA, PCI and other bus systems base on level triggered
123 interrupts, and therewith allow for shared IRQs. However, if multiple devices
124 present their hardware status by the same (shared) IRQ, the operating system
125 has to call every single device driver registered on this IRQ in order to
126 determine the device driver owning the device that raised the interrupt.
128 Up to kernel 2.4, Linux/390 used to provide interfaces via the IRQ (subchannel).
129 For internal use of the common I/O layer, these are still there. However,
130 device drivers should use the new calling interface via the ccw_device only.
132 During its startup the Linux/390 system checks for peripheral devices. Each
133 of those devices is uniquely defined by a so called subchannel by the ESA/390
134 channel subsystem. While the subchannel numbers are system generated, each
135 subchannel also takes a user defined attribute, the so called device number.
136 Both subchannel number and device number cannot exceed 65535. During driverfs
137 initialisation, the information about control unit type and device types that
138 imply specific I/O commands (channel command words - CCWs) in order to operate
139 the device are gathered. Device drivers can retrieve this set of hardware
140 information during their initialization step to recognize the devices they
141 support using the information saved in the struct ccw_device given to them.
142 This methods implies that Linux/390 doesn't require to probe for free (not
143 armed) interrupt request lines (IRQs) to drive its devices with. Where
144 applicable, the device drivers can use issue the READ DEVICE CHARACTERISTICS
145 ccw to retrieve device characteristics in its online routine.
147 In order to allow for easy I/O initiation the CDS layer provides a
148 ccw_device_start() interface that takes a device specific channel program (one
149 or more CCWs) as input sets up the required architecture specific control blocks
150 and initiates an I/O request on behalf of the device driver. The
151 ccw_device_start() routine allows to specify whether it expects the CDS layer
152 to notify the device driver for every interrupt it observes, or with final status
153 only. See ccw_device_start() for more details. A device driver must never issue
154 ESA/390 I/O commands itself, but must use the Linux/390 CDS interfaces instead.
156 For long running I/O request to be canceled, the CDS layer provides the
157 ccw_device_halt() function. Some devices require to initially issue a HALT
158 SUBCHANNEL (HSCH) command without having pending I/O requests. This function is
159 also covered by ccw_device_halt().
162 get_ciw() - get command information word
164 This call enables a device driver to get information about supported commands
165 from the extended SenseID data.
168 ccw_device_get_ciw(struct ccw_device *cdev, __u32 cmd);
170 cdev - The ccw_device for which the command is to be retrieved.
171 cmd - The command type to be retrieved.
173 ccw_device_get_ciw() returns:
174 NULL - No extended data available, invalid device or command not found.
175 !NULL - The command requested.
178 ccw_device_start() - Initiate I/O Request
180 The ccw_device_start() routines is the I/O request front-end processor. All
181 device driver I/O requests must be issued using this routine. A device driver
182 must not issue ESA/390 I/O commands itself. Instead the ccw_device_start()
183 routine provides all interfaces required to drive arbitrary devices.
185 This description also covers the status information passed to the device
186 driver's interrupt handler as this is related to the rules (flags) defined
187 with the associated I/O request when calling ccw_device_start().
189 int ccw_device_start(struct ccw_device *cdev,
191 unsigned long intparm,
193 unsigned long flags);
194 int ccw_device_start_timeout(struct ccw_device *cdev,
196 unsigned long intparm,
200 int ccw_device_start_key(struct ccw_device *cdev,
202 unsigned long intparm,
205 unsigned long flags);
206 int ccw_device_start_key_timeout(struct ccw_device *cdev,
208 unsigned long intparm,
214 cdev : ccw_device the I/O is destined for
215 cpa : logical start address of channel program
216 user_intparm : user specific interrupt information; will be presented
217 back to the device driver's interrupt handler. Allows a
218 device driver to associate the interrupt with a
219 particular I/O request.
220 lpm : defines the channel path to be used for a specific I/O
221 request. A value of 0 will make cio use the opm.
222 key : the storage key to use for the I/O (useful for operating on a
223 storage with a storage key != default key)
224 flag : defines the action to be performed for I/O processing
225 expires : timeout value in jiffies. The common I/O layer will terminate
226 the running program after this and call the interrupt handler
227 with ERR_PTR(-ETIMEDOUT) as irb.
229 Possible flag values are :
231 DOIO_ALLOW_SUSPEND - channel program may become suspended
232 DOIO_DENY_PREFETCH - don't allow for CCW prefetch; usually
233 this implies the channel program might
235 DOIO_SUPPRESS_INTER - don't call the handler on intermediate status
237 The cpa parameter points to the first format 1 CCW of a channel program :
240 __u8 cmd_code;/* command code */
241 __u8 flags; /* flags, like IDA addressing, etc. */
242 __u16 count; /* byte count */
243 __u32 cda; /* data address */
244 } __attribute__ ((packed,aligned(8)));
246 with the following CCW flags values defined :
248 CCW_FLAG_DC - data chaining
249 CCW_FLAG_CC - command chaining
250 CCW_FLAG_SLI - suppress incorrect length
253 CCW_FLAG_IDA - indirect addressing
254 CCW_FLAG_SUSPEND - suspend
257 Via ccw_device_set_options(), the device driver may specify the following
258 options for the device:
260 DOIO_EARLY_NOTIFICATION - allow for early interrupt notification
261 DOIO_REPORT_ALL - report all interrupt conditions
264 The ccw_device_start() function returns :
266 0 - successful completion or request successfully initiated
267 -EBUSY - The device is currently processing a previous I/O request, or there is
268 a status pending at the device.
269 -ENODEV - cdev is invalid, the device is not operational or the ccw_device is
272 When the I/O request completes, the CDS first level interrupt handler will
273 accumulate the status in a struct irb and then call the device interrupt handler.
274 The intparm field will contain the value the device driver has associated with a
275 particular I/O request. If a pending device status was recognized,
276 intparm will be set to 0 (zero). This may happen during I/O initiation or delayed
277 by an alert status notification. In any case this status is not related to the
278 current (last) I/O request. In case of a delayed status notification no special
279 interrupt will be presented to indicate I/O completion as the I/O request was
280 never started, even though ccw_device_start() returned with successful completion.
282 The irb may contain an error value, and the device driver should check for this
285 -ETIMEDOUT: the common I/O layer terminated the request after the specified
287 -EIO: the common I/O layer terminated the request due to an error state
289 If the concurrent sense flag in the extended status word in the irb is set, the
290 field irb->scsw.count describes the number of device specific sense bytes
291 available in the extended control word irb->scsw.ecw[0]. No device sensing by
292 the device driver itself is required.
294 The device interrupt handler can use the following definitions to investigate
295 the primary unit check source coded in sense byte 0 :
298 SNS0_INTERVENTION_REQ 0x40
299 SNS0_BUS_OUT_CHECK 0x20
300 SNS0_EQUIPMENT_CHECK 0x10
303 SNS0_INCOMPL_DOMAIN 0x01
305 Depending on the device status, multiple of those values may be set together.
306 Please refer to the device specific documentation for details.
308 The irb->scsw.cstat field provides the (accumulated) subchannel status :
310 SCHN_STAT_PCI - program controlled interrupt
311 SCHN_STAT_INCORR_LEN - incorrect length
312 SCHN_STAT_PROG_CHECK - program check
313 SCHN_STAT_PROT_CHECK - protection check
314 SCHN_STAT_CHN_DATA_CHK - channel data check
315 SCHN_STAT_CHN_CTRL_CHK - channel control check
316 SCHN_STAT_INTF_CTRL_CHK - interface control check
317 SCHN_STAT_CHAIN_CHECK - chaining check
319 The irb->scsw.dstat field provides the (accumulated) device status :
321 DEV_STAT_ATTENTION - attention
322 DEV_STAT_STAT_MOD - status modifier
323 DEV_STAT_CU_END - control unit end
325 DEV_STAT_CHN_END - channel end
326 DEV_STAT_DEV_END - device end
327 DEV_STAT_UNIT_CHECK - unit check
328 DEV_STAT_UNIT_EXCEP - unit exception
330 Please see the ESA/390 Principles of Operation manual for details on the
331 individual flag meanings.
335 ccw_device_start() must be called disabled and with the ccw device lock held.
337 The device driver is allowed to issue the next ccw_device_start() call from
338 within its interrupt handler already. It is not required to schedule a
339 bottom-half, unless a non deterministically long running error recovery procedure
340 or similar needs to be scheduled. During I/O processing the Linux/390 generic
341 I/O device driver support has already obtained the IRQ lock, i.e. the handler
342 must not try to obtain it again when calling ccw_device_start() or we end in a
345 If a device driver relies on an I/O request to be completed prior to start the
346 next it can reduce I/O processing overhead by chaining a NoOp I/O command
347 CCW_CMD_NOOP to the end of the submitted CCW chain. This will force Channel-End
348 and Device-End status to be presented together, with a single interrupt.
349 However, this should be used with care as it implies the channel will remain
350 busy, not being able to process I/O requests for other devices on the same
351 channel. Therefore e.g. read commands should never use this technique, as the
352 result will be presented by a single interrupt anyway.
354 In order to minimize I/O overhead, a device driver should use the
355 DOIO_REPORT_ALL only if the device can report intermediate interrupt
356 information prior to device-end the device driver urgently relies on. In this
357 case all I/O interruptions are presented to the device driver until final
358 status is recognized.
360 If a device is able to recover from asynchronously presented I/O errors, it can
361 perform overlapping I/O using the DOIO_EARLY_NOTIFICATION flag. While some
362 devices always report channel-end and device-end together, with a single
363 interrupt, others present primary status (channel-end) when the channel is
364 ready for the next I/O request and secondary status (device-end) when the data
365 transmission has been completed at the device.
367 Above flag allows to exploit this feature, e.g. for communication devices that
368 can handle lost data on the network to allow for enhanced I/O processing.
370 Unless the channel subsystem at any time presents a secondary status interrupt,
371 exploiting this feature will cause only primary status interrupts to be
372 presented to the device driver while overlapping I/O is performed. When a
373 secondary status without error (alert status) is presented, this indicates
374 successful completion for all overlapping ccw_device_start() requests that have
375 been issued since the last secondary (final) status.
377 Channel programs that intend to set the suspend flag on a channel command word
378 (CCW) must start the I/O operation with the DOIO_ALLOW_SUSPEND option or the
379 suspend flag will cause a channel program check. At the time the channel program
380 becomes suspended an intermediate interrupt will be generated by the channel
383 ccw_device_resume() - Resume Channel Program Execution
385 If a device driver chooses to suspend the current channel program execution by
386 setting the CCW suspend flag on a particular CCW, the channel program execution
387 is suspended. In order to resume channel program execution the CIO layer
388 provides the ccw_device_resume() routine.
390 int ccw_device_resume(struct ccw_device *cdev);
392 cdev - ccw_device the resume operation is requested for
394 The ccw_device_resume() function returns:
396 0 - suspended channel program is resumed
397 -EBUSY - status pending
398 -ENODEV - cdev invalid or not-operational subchannel
399 -EINVAL - resume function not applicable
400 -ENOTCONN - there is no I/O request pending for completion
403 Please have a look at the ccw_device_start() usage notes for more details on
404 suspended channel programs.
406 ccw_device_halt() - Halt I/O Request Processing
408 Sometimes a device driver might need a possibility to stop the processing of
409 a long-running channel program or the device might require to initially issue
410 a halt subchannel (HSCH) I/O command. For those purposes the ccw_device_halt()
413 ccw_device_halt() must be called disabled and with the ccw device lock held.
415 int ccw_device_halt(struct ccw_device *cdev,
416 unsigned long intparm);
418 cdev : ccw_device the halt operation is requested for
419 intparm : interruption parameter; value is only used if no I/O
420 is outstanding, otherwise the intparm associated with
421 the I/O request is returned
423 The ccw_device_halt() function returns :
425 0 - request successfully initiated
426 -EBUSY - the device is currently busy, or status pending.
427 -ENODEV - cdev invalid.
428 -EINVAL - The device is not operational or the ccw device is not online.
432 A device driver may write a never-ending channel program by writing a channel
433 program that at its end loops back to its beginning by means of a transfer in
434 channel (TIC) command (CCW_CMD_TIC). Usually this is performed by network
435 device drivers by setting the PCI CCW flag (CCW_FLAG_PCI). Once this CCW is
436 executed a program controlled interrupt (PCI) is generated. The device driver
437 can then perform an appropriate action. Prior to interrupt of an outstanding
438 read to a network device (with or without PCI flag) a ccw_device_halt()
439 is required to end the pending operation.
441 ccw_device_clear() - Terminage I/O Request Processing
443 In order to terminate all I/O processing at the subchannel, the clear subchannel
444 (CSCH) command is used. It can be issued via ccw_device_clear().
446 ccw_device_clear() must be called disabled and with the ccw device lock held.
448 int ccw_device_clear(struct ccw_device *cdev, unsigned long intparm);
450 cdev: ccw_device the clear operation is requested for
451 intparm: interruption parameter (see ccw_device_halt())
453 The ccw_device_clear() function returns:
455 0 - request successfully initiated
456 -ENODEV - cdev invalid
457 -EINVAL - The device is not operational or the ccw device is not online.
459 Miscellaneous Support Routines
461 This chapter describes various routines to be used in a Linux/390 device
462 driver programming environment.
466 Get the address of the device specific lock. This is then used in
467 spin_lock() / spin_unlock() calls.
470 __u8 ccw_device_get_path_mask(struct ccw_device *cdev);
472 Get the mask of the path currently available for cdev.