blob: 9e2078b2a615c332f227bc4bd841416678f2b4f7 [file] [log] [blame]
James Bottomley2908d772006-08-29 09:22:51 -05001SAS Layer
2---------
3
4The SAS Layer is a management infrastructure which manages
5SAS LLDDs. It sits between SCSI Core and SAS LLDDs. The
6layout is as follows: while SCSI Core is concerned with
7SAM/SPC issues, and a SAS LLDD+sequencer is concerned with
8phy/OOB/link management, the SAS layer is concerned with:
9
10 * SAS Phy/Port/HA event management (LLDD generates,
11 SAS Layer processes),
12 * SAS Port management (creation/destruction),
13 * SAS Domain discovery and revalidation,
14 * SAS Domain device management,
15 * SCSI Host registration/unregistration,
16 * Device registration with SCSI Core (SAS) or libata
17 (SATA), and
18 * Expander management and exporting expander control
19 to user space.
20
21A SAS LLDD is a PCI device driver. It is concerned with
22phy/OOB management, and vendor specific tasks and generates
23events to the SAS layer.
24
25The SAS Layer does most SAS tasks as outlined in the SAS 1.1
26spec.
27
28The sas_ha_struct describes the SAS LLDD to the SAS layer.
29Most of it is used by the SAS Layer but a few fields need to
30be initialized by the LLDDs.
31
32After initializing your hardware, from the probe() function
33you call sas_register_ha(). It will register your LLDD with
34the SCSI subsystem, creating a SCSI host and it will
35register your SAS driver with the sysfs SAS tree it creates.
36It will then return. Then you enable your phys to actually
37start OOB (at which point your driver will start calling the
38notify_* event callbacks).
39
40Structure descriptions:
41
42struct sas_phy --------------------
43Normally this is statically embedded to your driver's
44phy structure:
45 struct my_phy {
46 blah;
47 struct sas_phy sas_phy;
48 bleh;
49 };
50And then all the phys are an array of my_phy in your HA
51struct (shown below).
52
53Then as you go along and initialize your phys you also
54initialize the sas_phy struct, along with your own
55phy structure.
56
57In general, the phys are managed by the LLDD and the ports
58are managed by the SAS layer. So the phys are initialized
59and updated by the LLDD and the ports are initialized and
60updated by the SAS layer.
61
62There is a scheme where the LLDD can RW certain fields,
63and the SAS layer can only read such ones, and vice versa.
64The idea is to avoid unnecessary locking.
65
66enabled -- must be set (0/1)
67id -- must be set [0,MAX_PHYS)
68class, proto, type, role, oob_mode, linkrate -- must be set
69oob_mode -- you set this when OOB has finished and then notify
70the SAS Layer.
71
72sas_addr -- this normally points to an array holding the sas
73address of the phy, possibly somewhere in your my_phy
74struct.
75
76attached_sas_addr -- set this when you (LLDD) receive an
77IDENTIFY frame or a FIS frame, _before_ notifying the SAS
78layer. The idea is that sometimes the LLDD may want to fake
79or provide a different SAS address on that phy/port and this
80allows it to do this. At best you should copy the sas
81address from the IDENTIFY frame or maybe generate a SAS
82address for SATA directly attached devices. The Discover
83process may later change this.
84
85frame_rcvd -- this is where you copy the IDENTIFY/FIS frame
86when you get it; you lock, copy, set frame_rcvd_size and
87unlock the lock, and then call the event. It is a pointer
88since there's no way to know your hw frame size _exactly_,
89so you define the actual array in your phy struct and let
90this pointer point to it. You copy the frame from your
91DMAable memory to that area holding the lock.
92
93sas_prim -- this is where primitives go when they're
94received. See sas.h. Grab the lock, set the primitive,
95release the lock, notify.
96
97port -- this points to the sas_port if the phy belongs
98to a port -- the LLDD only reads this. It points to the
99sas_port this phy is part of. Set by the SAS Layer.
100
101ha -- may be set; the SAS layer sets it anyway.
102
103lldd_phy -- you should set this to point to your phy so you
104can find your way around faster when the SAS layer calls one
105of your callbacks and passes you a phy. If the sas_phy is
106embedded you can also use container_of -- whatever you
107prefer.
108
109
110struct sas_port --------------------
111The LLDD doesn't set any fields of this struct -- it only
112reads them. They should be self explanatory.
113
114phy_mask is 32 bit, this should be enough for now, as I
115haven't heard of a HA having more than 8 phys.
116
117lldd_port -- I haven't found use for that -- maybe other
118LLDD who wish to have internal port representation can make
119use of this.
120
121
122struct sas_ha_struct --------------------
123It normally is statically declared in your own LLDD
124structure describing your adapter:
125struct my_sas_ha {
126 blah;
127 struct sas_ha_struct sas_ha;
128 struct my_phy phys[MAX_PHYS];
129 struct sas_port sas_ports[MAX_PHYS]; /* (1) */
130 bleh;
131};
132
133(1) If your LLDD doesn't have its own port representation.
134
135What needs to be initialized (sample function given below).
136
137pcidev
138sas_addr -- since the SAS layer doesn't want to mess with
139 memory allocation, etc, this points to statically
140 allocated array somewhere (say in your host adapter
141 structure) and holds the SAS address of the host
142 adapter as given by you or the manufacturer, etc.
143sas_port
144sas_phy -- an array of pointers to structures. (see
145 note above on sas_addr).
146 These must be set. See more notes below.
147num_phys -- the number of phys present in the sas_phy array,
148 and the number of ports present in the sas_port
149 array. There can be a maximum num_phys ports (one per
150 port) so we drop the num_ports, and only use
151 num_phys.
152
153The event interface:
154
155 /* LLDD calls these to notify the class of an event. */
156 void (*notify_ha_event)(struct sas_ha_struct *, enum ha_event);
157 void (*notify_port_event)(struct sas_phy *, enum port_event);
158 void (*notify_phy_event)(struct sas_phy *, enum phy_event);
159
160When sas_register_ha() returns, those are set and can be
161called by the LLDD to notify the SAS layer of such events
162the SAS layer.
163
164The port notification:
165
166 /* The class calls these to notify the LLDD of an event. */
167 void (*lldd_port_formed)(struct sas_phy *);
168 void (*lldd_port_deformed)(struct sas_phy *);
169
170If the LLDD wants notification when a port has been formed
171or deformed it sets those to a function satisfying the type.
172
173A SAS LLDD should also implement at least one of the Task
174Management Functions (TMFs) described in SAM:
175
176 /* Task Management Functions. Must be called from process context. */
177 int (*lldd_abort_task)(struct sas_task *);
178 int (*lldd_abort_task_set)(struct domain_device *, u8 *lun);
179 int (*lldd_clear_aca)(struct domain_device *, u8 *lun);
180 int (*lldd_clear_task_set)(struct domain_device *, u8 *lun);
181 int (*lldd_I_T_nexus_reset)(struct domain_device *);
182 int (*lldd_lu_reset)(struct domain_device *, u8 *lun);
183 int (*lldd_query_task)(struct sas_task *);
184
185For more information please read SAM from T10.org.
186
187Port and Adapter management:
188
189 /* Port and Adapter management */
190 int (*lldd_clear_nexus_port)(struct sas_port *);
191 int (*lldd_clear_nexus_ha)(struct sas_ha_struct *);
192
193A SAS LLDD should implement at least one of those.
194
195Phy management:
196
197 /* Phy management */
198 int (*lldd_control_phy)(struct sas_phy *, enum phy_func);
199
200lldd_ha -- set this to point to your HA struct. You can also
201use container_of if you embedded it as shown above.
202
203A sample initialization and registration function
204can look like this (called last thing from probe())
205*but* before you enable the phys to do OOB:
206
207static int register_sas_ha(struct my_sas_ha *my_ha)
208{
209 int i;
210 static struct sas_phy *sas_phys[MAX_PHYS];
211 static struct sas_port *sas_ports[MAX_PHYS];
212
213 my_ha->sas_ha.sas_addr = &my_ha->sas_addr[0];
214
215 for (i = 0; i < MAX_PHYS; i++) {
216 sas_phys[i] = &my_ha->phys[i].sas_phy;
217 sas_ports[i] = &my_ha->sas_ports[i];
218 }
219
220 my_ha->sas_ha.sas_phy = sas_phys;
221 my_ha->sas_ha.sas_port = sas_ports;
222 my_ha->sas_ha.num_phys = MAX_PHYS;
223
224 my_ha->sas_ha.lldd_port_formed = my_port_formed;
225
226 my_ha->sas_ha.lldd_dev_found = my_dev_found;
227 my_ha->sas_ha.lldd_dev_gone = my_dev_gone;
228
229 my_ha->sas_ha.lldd_max_execute_num = lldd_max_execute_num; (1)
230
231 my_ha->sas_ha.lldd_queue_size = ha_can_queue;
232 my_ha->sas_ha.lldd_execute_task = my_execute_task;
233
234 my_ha->sas_ha.lldd_abort_task = my_abort_task;
235 my_ha->sas_ha.lldd_abort_task_set = my_abort_task_set;
236 my_ha->sas_ha.lldd_clear_aca = my_clear_aca;
237 my_ha->sas_ha.lldd_clear_task_set = my_clear_task_set;
238 my_ha->sas_ha.lldd_I_T_nexus_reset= NULL; (2)
239 my_ha->sas_ha.lldd_lu_reset = my_lu_reset;
240 my_ha->sas_ha.lldd_query_task = my_query_task;
241
242 my_ha->sas_ha.lldd_clear_nexus_port = my_clear_nexus_port;
243 my_ha->sas_ha.lldd_clear_nexus_ha = my_clear_nexus_ha;
244
245 my_ha->sas_ha.lldd_control_phy = my_control_phy;
246
247 return sas_register_ha(&my_ha->sas_ha);
248}
249
250(1) This is normally a LLDD parameter, something of the
251lines of a task collector. What it tells the SAS Layer is
252whether the SAS layer should run in Direct Mode (default:
253value 0 or 1) or Task Collector Mode (value greater than 1).
254
255In Direct Mode, the SAS Layer calls Execute Task as soon as
256it has a command to send to the SDS, _and_ this is a single
257command, i.e. not linked.
258
259Some hardware (e.g. aic94xx) has the capability to DMA more
260than one task at a time (interrupt) from host memory. Task
261Collector Mode is an optional feature for HAs which support
262this in their hardware. (Again, it is completely optional
263even if your hardware supports it.)
264
265In Task Collector Mode, the SAS Layer would do _natural_
266coalescing of tasks and at the appropriate moment it would
267call your driver to DMA more than one task in a single HA
268interrupt. DMBS may want to use this by insmod/modprobe
269setting the lldd_max_execute_num to something greater than
2701.
271
272(2) SAS 1.1 does not define I_T Nexus Reset TMF.
273
274Events
275------
276
277Events are _the only way_ a SAS LLDD notifies the SAS layer
278of anything. There is no other method or way a LLDD to tell
279the SAS layer of anything happening internally or in the SAS
280domain.
281
282Phy events:
283 PHYE_LOSS_OF_SIGNAL, (C)
284 PHYE_OOB_DONE,
285 PHYE_OOB_ERROR, (C)
286 PHYE_SPINUP_HOLD.
287
288Port events, passed on a _phy_:
289 PORTE_BYTES_DMAED, (M)
290 PORTE_BROADCAST_RCVD, (E)
291 PORTE_LINK_RESET_ERR, (C)
292 PORTE_TIMER_EVENT, (C)
293 PORTE_HARD_RESET.
294
295Host Adapter event:
296 HAE_RESET
297
298A SAS LLDD should be able to generate
299 - at least one event from group C (choice),
300 - events marked M (mandatory) are mandatory (only one),
301 - events marked E (expander) if it wants the SAS layer
302 to handle domain revalidation (only one such).
303 - Unmarked events are optional.
304
305Meaning:
306
307HAE_RESET -- when your HA got internal error and was reset.
308
309PORTE_BYTES_DMAED -- on receiving an IDENTIFY/FIS frame
310PORTE_BROADCAST_RCVD -- on receiving a primitive
311PORTE_LINK_RESET_ERR -- timer expired, loss of signal, loss
312of DWS, etc. (*)
313PORTE_TIMER_EVENT -- DWS reset timeout timer expired (*)
314PORTE_HARD_RESET -- Hard Reset primitive received.
315
316PHYE_LOSS_OF_SIGNAL -- the device is gone (*)
317PHYE_OOB_DONE -- OOB went fine and oob_mode is valid
318PHYE_OOB_ERROR -- Error while doing OOB, the device probably
319got disconnected. (*)
320PHYE_SPINUP_HOLD -- SATA is present, COMWAKE not sent.
321
322(*) should set/clear the appropriate fields in the phy,
323 or alternatively call the inlined sas_phy_disconnected()
324 which is just a helper, from their tasklet.
325
326The Execute Command SCSI RPC:
327
328 int (*lldd_execute_task)(struct sas_task *, int num,
329 unsigned long gfp_flags);
330
331Used to queue a task to the SAS LLDD. @task is the tasks to
332be executed. @num should be the number of tasks being
333queued at this function call (they are linked listed via
334task::list), @gfp_mask should be the gfp_mask defining the
335context of the caller.
336
337This function should implement the Execute Command SCSI RPC,
338or if you're sending a SCSI Task as linked commands, you
339should also use this function.
340
341That is, when lldd_execute_task() is called, the command(s)
342go out on the transport *immediately*. There is *no*
343queuing of any sort and at any level in a SAS LLDD.
344
345The use of task::list is two-fold, one for linked commands,
346the other discussed below.
347
348It is possible to queue up more than one task at a time, by
349initializing the list element of struct sas_task, and
350passing the number of tasks enlisted in this manner in num.
351
352Returns: -SAS_QUEUE_FULL, -ENOMEM, nothing was queued;
353 0, the task(s) were queued.
354
355If you want to pass num > 1, then either
356A) you're the only caller of this function and keep track
357 of what you've queued to the LLDD, or
358B) you know what you're doing and have a strategy of
359 retrying.
360
361As opposed to queuing one task at a time (function call),
362batch queuing of tasks, by having num > 1, greatly
363simplifies LLDD code, sequencer code, and _hardware design_,
364and has some performance advantages in certain situations
365(DBMS).
366
367The LLDD advertises if it can take more than one command at
368a time at lldd_execute_task(), by setting the
369lldd_max_execute_num parameter (controlled by "collector"
370module parameter in aic94xx SAS LLDD).
371
372You should leave this to the default 1, unless you know what
373you're doing.
374
375This is a function of the LLDD, to which the SAS layer can
376cater to.
377
378int lldd_queue_size
379 The host adapter's queue size. This is the maximum
380number of commands the lldd can have pending to domain
381devices on behalf of all upper layers submitting through
382lldd_execute_task().
383
384You really want to set this to something (much) larger than
3851.
386
387This _really_ has absolutely nothing to do with queuing.
388There is no queuing in SAS LLDDs.
389
390struct sas_task {
391 dev -- the device this task is destined to
392 list -- must be initialized (INIT_LIST_HEAD)
393 task_proto -- _one_ of enum sas_proto
394 scatter -- pointer to scatter gather list array
395 num_scatter -- number of elements in scatter
396 total_xfer_len -- total number of bytes expected to be transfered
397 data_dir -- PCI_DMA_...
398 task_done -- callback when the task has finished execution
399};
400
401When an external entity, entity other than the LLDD or the
402SAS Layer, wants to work with a struct domain_device, it
403_must_ call kobject_get() when getting a handle on the
404device and kobject_put() when it is done with the device.
405
406This does two things:
407 A) implements proper kfree() for the device;
408 B) increments/decrements the kref for all players:
409 domain_device
410 all domain_device's ... (if past an expander)
411 port
412 host adapter
413 pci device
414 and up the ladder, etc.
415
416DISCOVERY
417---------
418
419The sysfs tree has the following purposes:
420 a) It shows you the physical layout of the SAS domain at
421 the current time, i.e. how the domain looks in the
422 physical world right now.
423 b) Shows some device parameters _at_discovery_time_.
424
425This is a link to the tree(1) program, very useful in
426viewing the SAS domain:
427ftp://mama.indstate.edu/linux/tree/
428I expect user space applications to actually create a
429graphical interface of this.
430
431That is, the sysfs domain tree doesn't show or keep state if
432you e.g., change the meaning of the READY LED MEANING
433setting, but it does show you the current connection status
434of the domain device.
435
436Keeping internal device state changes is responsibility of
437upper layers (Command set drivers) and user space.
438
439When a device or devices are unplugged from the domain, this
440is reflected in the sysfs tree immediately, and the device(s)
441removed from the system.
442
443The structure domain_device describes any device in the SAS
444domain. It is completely managed by the SAS layer. A task
445points to a domain device, this is how the SAS LLDD knows
446where to send the task(s) to. A SAS LLDD only reads the
447contents of the domain_device structure, but it never creates
448or destroys one.
449
450Expander management from User Space
451-----------------------------------
452
453In each expander directory in sysfs, there is a file called
454"smp_portal". It is a binary sysfs attribute file, which
455implements an SMP portal (Note: this is *NOT* an SMP port),
456to which user space applications can send SMP requests and
457receive SMP responses.
458
459Functionality is deceptively simple:
460
4611. Build the SMP frame you want to send. The format and layout
462 is described in the SAS spec. Leave the CRC field equal 0.
463open(2)
4642. Open the expander's SMP portal sysfs file in RW mode.
465write(2)
4663. Write the frame you built in 1.
467read(2)
4684. Read the amount of data you expect to receive for the frame you built.
469 If you receive different amount of data you expected to receive,
470 then there was some kind of error.
471close(2)
472All this process is shown in detail in the function do_smp_func()
473and its callers, in the file "expander_conf.c".
474
475The kernel functionality is implemented in the file
476"sas_expander.c".
477
478The program "expander_conf.c" implements this. It takes one
479argument, the sysfs file name of the SMP portal to the
480expander, and gives expander information, including routing
481tables.
482
483The SMP portal gives you complete control of the expander,
484so please be careful.