|
1 /* Name: usbdrvasm165.inc |
|
2 * Project: V-USB, virtual USB port for Atmel's(r) AVR(r) microcontrollers |
|
3 * Author: Christian Starkjohann |
|
4 * Creation Date: 2007-04-22 |
|
5 * Tabsize: 4 |
|
6 * Copyright: (c) 2007 by OBJECTIVE DEVELOPMENT Software GmbH |
|
7 * License: GNU GPL v2 (see License.txt), GNU GPL v3 or proprietary (CommercialLicense.txt) |
|
8 */ |
|
9 |
|
10 /* Do not link this file! Link usbdrvasm.S instead, which includes the |
|
11 * appropriate implementation! |
|
12 */ |
|
13 |
|
14 /* |
|
15 General Description: |
|
16 This file is the 16.5 MHz version of the USB driver. It is intended for the |
|
17 ATTiny45 and similar controllers running on 16.5 MHz internal RC oscillator. |
|
18 This version contains a phase locked loop in the receiver routine to cope with |
|
19 slight clock rate deviations of up to +/- 1%. |
|
20 |
|
21 See usbdrv.h for a description of the entire driver. |
|
22 |
|
23 Since almost all of this code is timing critical, don't change unless you |
|
24 really know what you are doing! Many parts require not only a maximum number |
|
25 of CPU cycles, but even an exact number of cycles! |
|
26 */ |
|
27 |
|
28 ;Software-receiver engine. Strict timing! Don't change unless you can preserve timing! |
|
29 ;interrupt response time: 4 cycles + insn running = 7 max if interrupts always enabled |
|
30 ;max allowable interrupt latency: 59 cycles -> max 52 cycles interrupt disable |
|
31 ;max stack usage: [ret(2), r0, SREG, YL, YH, shift, x1, x2, x3, x4, cnt] = 12 bytes |
|
32 ;nominal frequency: 16.5 MHz -> 11 cycles per bit |
|
33 ; 16.3125 MHz < F_CPU < 16.6875 MHz (+/- 1.1%) |
|
34 ; Numbers in brackets are clocks counted from center of last sync bit |
|
35 ; when instruction starts |
|
36 |
|
37 |
|
38 USB_INTR_VECTOR: |
|
39 ;order of registers pushed: YL, SREG [sofError], r0, YH, shift, x1, x2, x3, x4, cnt |
|
40 push YL ;[-23] push only what is necessary to sync with edge ASAP |
|
41 in YL, SREG ;[-21] |
|
42 push YL ;[-20] |
|
43 ;---------------------------------------------------------------------------- |
|
44 ; Synchronize with sync pattern: |
|
45 ;---------------------------------------------------------------------------- |
|
46 ;sync byte (D-) pattern LSb to MSb: 01010100 [1 = idle = J, 0 = K] |
|
47 ;sync up with J to K edge during sync pattern -- use fastest possible loops |
|
48 ;The first part waits at most 1 bit long since we must be in sync pattern. |
|
49 ;YL is guarenteed to be < 0x80 because I flag is clear. When we jump to |
|
50 ;waitForJ, ensure that this prerequisite is met. |
|
51 waitForJ: |
|
52 inc YL |
|
53 sbis USBIN, USBMINUS |
|
54 brne waitForJ ; just make sure we have ANY timeout |
|
55 waitForK: |
|
56 ;The following code results in a sampling window of < 1/4 bit which meets the spec. |
|
57 sbis USBIN, USBMINUS ;[-15] |
|
58 rjmp foundK ;[-14] |
|
59 sbis USBIN, USBMINUS |
|
60 rjmp foundK |
|
61 sbis USBIN, USBMINUS |
|
62 rjmp foundK |
|
63 sbis USBIN, USBMINUS |
|
64 rjmp foundK |
|
65 sbis USBIN, USBMINUS |
|
66 rjmp foundK |
|
67 sbis USBIN, USBMINUS |
|
68 rjmp foundK |
|
69 #if USB_COUNT_SOF |
|
70 lds YL, usbSofCount |
|
71 inc YL |
|
72 sts usbSofCount, YL |
|
73 #endif /* USB_COUNT_SOF */ |
|
74 #ifdef USB_SOF_HOOK |
|
75 USB_SOF_HOOK |
|
76 #endif |
|
77 rjmp sofError |
|
78 foundK: ;[-12] |
|
79 ;{3, 5} after falling D- edge, average delay: 4 cycles [we want 5 for center sampling] |
|
80 ;we have 1 bit time for setup purposes, then sample again. Numbers in brackets |
|
81 ;are cycles from center of first sync (double K) bit after the instruction |
|
82 push r0 ;[-12] |
|
83 ; [---] ;[-11] |
|
84 push YH ;[-10] |
|
85 ; [---] ;[-9] |
|
86 lds YL, usbInputBufOffset;[-8] |
|
87 ; [---] ;[-7] |
|
88 clr YH ;[-6] |
|
89 subi YL, lo8(-(usbRxBuf));[-5] [rx loop init] |
|
90 sbci YH, hi8(-(usbRxBuf));[-4] [rx loop init] |
|
91 mov r0, x2 ;[-3] [rx loop init] |
|
92 sbis USBIN, USBMINUS ;[-2] we want two bits K (sample 2 cycles too early) |
|
93 rjmp haveTwoBitsK ;[-1] |
|
94 pop YH ;[0] undo the pushes from before |
|
95 pop r0 ;[2] |
|
96 rjmp waitForK ;[4] this was not the end of sync, retry |
|
97 ; The entire loop from waitForK until rjmp waitForK above must not exceed two |
|
98 ; bit times (= 22 cycles). |
|
99 |
|
100 ;---------------------------------------------------------------------------- |
|
101 ; push more registers and initialize values while we sample the first bits: |
|
102 ;---------------------------------------------------------------------------- |
|
103 haveTwoBitsK: ;[1] |
|
104 push shift ;[1] |
|
105 push x1 ;[3] |
|
106 push x2 ;[5] |
|
107 push x3 ;[7] |
|
108 ldi shift, 0xff ;[9] [rx loop init] |
|
109 ori x3, 0xff ;[10] [rx loop init] == ser x3, clear zero flag |
|
110 |
|
111 in x1, USBIN ;[11] <-- sample bit 0 |
|
112 bst x1, USBMINUS ;[12] |
|
113 bld shift, 0 ;[13] |
|
114 push x4 ;[14] == phase |
|
115 ; [---] ;[15] |
|
116 push cnt ;[16] |
|
117 ; [---] ;[17] |
|
118 ldi phase, 0 ;[18] [rx loop init] |
|
119 ldi cnt, USB_BUFSIZE;[19] [rx loop init] |
|
120 rjmp rxbit1 ;[20] |
|
121 ; [---] ;[21] |
|
122 |
|
123 ;---------------------------------------------------------------------------- |
|
124 ; Receiver loop (numbers in brackets are cycles within byte after instr) |
|
125 ;---------------------------------------------------------------------------- |
|
126 /* |
|
127 byte oriented operations done during loop: |
|
128 bit 0: store data |
|
129 bit 1: SE0 check |
|
130 bit 2: overflow check |
|
131 bit 3: catch up |
|
132 bit 4: rjmp to achieve conditional jump range |
|
133 bit 5: PLL |
|
134 bit 6: catch up |
|
135 bit 7: jump, fixup bitstuff |
|
136 ; 87 [+ 2] cycles |
|
137 ------------------------------------------------------------------ |
|
138 */ |
|
139 continueWithBit5: |
|
140 in x2, USBIN ;[055] <-- bit 5 |
|
141 eor r0, x2 ;[056] |
|
142 or phase, r0 ;[057] |
|
143 sbrc phase, USBMINUS ;[058] |
|
144 lpm ;[059] optional nop3; modifies r0 |
|
145 in phase, USBIN ;[060] <-- phase |
|
146 eor x1, x2 ;[061] |
|
147 bst x1, USBMINUS ;[062] |
|
148 bld shift, 5 ;[063] |
|
149 andi shift, 0x3f ;[064] |
|
150 in x1, USBIN ;[065] <-- bit 6 |
|
151 breq unstuff5 ;[066] *** unstuff escape |
|
152 eor phase, x1 ;[067] |
|
153 eor x2, x1 ;[068] |
|
154 bst x2, USBMINUS ;[069] |
|
155 bld shift, 6 ;[070] |
|
156 didUnstuff6: ;[ ] |
|
157 in r0, USBIN ;[071] <-- phase |
|
158 cpi shift, 0x02 ;[072] |
|
159 brlo unstuff6 ;[073] *** unstuff escape |
|
160 didUnstuff5: ;[ ] |
|
161 nop2 ;[074] |
|
162 ; [---] ;[075] |
|
163 in x2, USBIN ;[076] <-- bit 7 |
|
164 eor x1, x2 ;[077] |
|
165 bst x1, USBMINUS ;[078] |
|
166 bld shift, 7 ;[079] |
|
167 didUnstuff7: ;[ ] |
|
168 eor r0, x2 ;[080] |
|
169 or phase, r0 ;[081] |
|
170 in r0, USBIN ;[082] <-- phase |
|
171 cpi shift, 0x04 ;[083] |
|
172 brsh rxLoop ;[084] |
|
173 ; [---] ;[085] |
|
174 unstuff7: ;[ ] |
|
175 andi x3, ~0x80 ;[085] |
|
176 ori shift, 0x80 ;[086] |
|
177 in x2, USBIN ;[087] <-- sample stuffed bit 7 |
|
178 nop ;[088] |
|
179 rjmp didUnstuff7 ;[089] |
|
180 ; [---] ;[090] |
|
181 ;[080] |
|
182 |
|
183 unstuff5: ;[067] |
|
184 eor phase, x1 ;[068] |
|
185 andi x3, ~0x20 ;[069] |
|
186 ori shift, 0x20 ;[070] |
|
187 in r0, USBIN ;[071] <-- phase |
|
188 mov x2, x1 ;[072] |
|
189 nop ;[073] |
|
190 nop2 ;[074] |
|
191 ; [---] ;[075] |
|
192 in x1, USBIN ;[076] <-- bit 6 |
|
193 eor r0, x1 ;[077] |
|
194 or phase, r0 ;[078] |
|
195 eor x2, x1 ;[079] |
|
196 bst x2, USBMINUS ;[080] |
|
197 bld shift, 6 ;[081] no need to check bitstuffing, we just had one |
|
198 in r0, USBIN ;[082] <-- phase |
|
199 rjmp didUnstuff5 ;[083] |
|
200 ; [---] ;[084] |
|
201 ;[074] |
|
202 |
|
203 unstuff6: ;[074] |
|
204 andi x3, ~0x40 ;[075] |
|
205 in x1, USBIN ;[076] <-- bit 6 again |
|
206 ori shift, 0x40 ;[077] |
|
207 nop2 ;[078] |
|
208 ; [---] ;[079] |
|
209 rjmp didUnstuff6 ;[080] |
|
210 ; [---] ;[081] |
|
211 ;[071] |
|
212 |
|
213 unstuff0: ;[013] |
|
214 eor r0, x2 ;[014] |
|
215 or phase, r0 ;[015] |
|
216 andi x2, USBMASK ;[016] check for SE0 |
|
217 in r0, USBIN ;[017] <-- phase |
|
218 breq didUnstuff0 ;[018] direct jump to se0 would be too long |
|
219 andi x3, ~0x01 ;[019] |
|
220 ori shift, 0x01 ;[020] |
|
221 mov x1, x2 ;[021] mov existing sample |
|
222 in x2, USBIN ;[022] <-- bit 1 again |
|
223 rjmp didUnstuff0 ;[023] |
|
224 ; [---] ;[024] |
|
225 ;[014] |
|
226 |
|
227 unstuff1: ;[024] |
|
228 eor r0, x1 ;[025] |
|
229 or phase, r0 ;[026] |
|
230 andi x3, ~0x02 ;[027] |
|
231 in r0, USBIN ;[028] <-- phase |
|
232 ori shift, 0x02 ;[029] |
|
233 mov x2, x1 ;[030] |
|
234 rjmp didUnstuff1 ;[031] |
|
235 ; [---] ;[032] |
|
236 ;[022] |
|
237 |
|
238 unstuff2: ;[035] |
|
239 eor r0, x2 ;[036] |
|
240 or phase, r0 ;[037] |
|
241 andi x3, ~0x04 ;[038] |
|
242 in r0, USBIN ;[039] <-- phase |
|
243 ori shift, 0x04 ;[040] |
|
244 mov x1, x2 ;[041] |
|
245 rjmp didUnstuff2 ;[042] |
|
246 ; [---] ;[043] |
|
247 ;[033] |
|
248 |
|
249 unstuff3: ;[043] |
|
250 in x2, USBIN ;[044] <-- bit 3 again |
|
251 eor r0, x2 ;[045] |
|
252 or phase, r0 ;[046] |
|
253 andi x3, ~0x08 ;[047] |
|
254 ori shift, 0x08 ;[048] |
|
255 nop ;[049] |
|
256 in r0, USBIN ;[050] <-- phase |
|
257 rjmp didUnstuff3 ;[051] |
|
258 ; [---] ;[052] |
|
259 ;[042] |
|
260 |
|
261 unstuff4: ;[053] |
|
262 andi x3, ~0x10 ;[054] |
|
263 in x1, USBIN ;[055] <-- bit 4 again |
|
264 ori shift, 0x10 ;[056] |
|
265 rjmp didUnstuff4 ;[057] |
|
266 ; [---] ;[058] |
|
267 ;[048] |
|
268 |
|
269 rxLoop: ;[085] |
|
270 eor x3, shift ;[086] reconstruct: x3 is 0 at bit locations we changed, 1 at others |
|
271 in x1, USBIN ;[000] <-- bit 0 |
|
272 st y+, x3 ;[001] |
|
273 ; [---] ;[002] |
|
274 eor r0, x1 ;[003] |
|
275 or phase, r0 ;[004] |
|
276 eor x2, x1 ;[005] |
|
277 in r0, USBIN ;[006] <-- phase |
|
278 ser x3 ;[007] |
|
279 bst x2, USBMINUS ;[008] |
|
280 bld shift, 0 ;[009] |
|
281 andi shift, 0xf9 ;[010] |
|
282 rxbit1: ;[ ] |
|
283 in x2, USBIN ;[011] <-- bit 1 |
|
284 breq unstuff0 ;[012] *** unstuff escape |
|
285 andi x2, USBMASK ;[013] SE0 check for bit 1 |
|
286 didUnstuff0: ;[ ] Z only set if we detected SE0 in bitstuff |
|
287 breq se0 ;[014] |
|
288 eor r0, x2 ;[015] |
|
289 or phase, r0 ;[016] |
|
290 in r0, USBIN ;[017] <-- phase |
|
291 eor x1, x2 ;[018] |
|
292 bst x1, USBMINUS ;[019] |
|
293 bld shift, 1 ;[020] |
|
294 andi shift, 0xf3 ;[021] |
|
295 didUnstuff1: ;[ ] |
|
296 in x1, USBIN ;[022] <-- bit 2 |
|
297 breq unstuff1 ;[023] *** unstuff escape |
|
298 eor r0, x1 ;[024] |
|
299 or phase, r0 ;[025] |
|
300 subi cnt, 1 ;[026] overflow check |
|
301 brcs overflow ;[027] |
|
302 in r0, USBIN ;[028] <-- phase |
|
303 eor x2, x1 ;[029] |
|
304 bst x2, USBMINUS ;[030] |
|
305 bld shift, 2 ;[031] |
|
306 andi shift, 0xe7 ;[032] |
|
307 didUnstuff2: ;[ ] |
|
308 in x2, USBIN ;[033] <-- bit 3 |
|
309 breq unstuff2 ;[034] *** unstuff escape |
|
310 eor r0, x2 ;[035] |
|
311 or phase, r0 ;[036] |
|
312 eor x1, x2 ;[037] |
|
313 bst x1, USBMINUS ;[038] |
|
314 in r0, USBIN ;[039] <-- phase |
|
315 bld shift, 3 ;[040] |
|
316 andi shift, 0xcf ;[041] |
|
317 didUnstuff3: ;[ ] |
|
318 breq unstuff3 ;[042] *** unstuff escape |
|
319 nop ;[043] |
|
320 in x1, USBIN ;[044] <-- bit 4 |
|
321 eor x2, x1 ;[045] |
|
322 bst x2, USBMINUS ;[046] |
|
323 bld shift, 4 ;[047] |
|
324 didUnstuff4: ;[ ] |
|
325 eor r0, x1 ;[048] |
|
326 or phase, r0 ;[049] |
|
327 in r0, USBIN ;[050] <-- phase |
|
328 andi shift, 0x9f ;[051] |
|
329 breq unstuff4 ;[052] *** unstuff escape |
|
330 rjmp continueWithBit5;[053] |
|
331 ; [---] ;[054] |
|
332 |
|
333 macro POP_STANDARD ; 16 cycles |
|
334 pop cnt |
|
335 pop x4 |
|
336 pop x3 |
|
337 pop x2 |
|
338 pop x1 |
|
339 pop shift |
|
340 pop YH |
|
341 pop r0 |
|
342 endm |
|
343 macro POP_RETI ; 5 cycles |
|
344 pop YL |
|
345 out SREG, YL |
|
346 pop YL |
|
347 endm |
|
348 |
|
349 #include "asmcommon.inc" |
|
350 |
|
351 |
|
352 ; USB spec says: |
|
353 ; idle = J |
|
354 ; J = (D+ = 0), (D- = 1) |
|
355 ; K = (D+ = 1), (D- = 0) |
|
356 ; Spec allows 7.5 bit times from EOP to SOP for replies |
|
357 |
|
358 bitstuff7: |
|
359 eor x1, x4 ;[4] |
|
360 ldi x2, 0 ;[5] |
|
361 nop2 ;[6] C is zero (brcc) |
|
362 rjmp didStuff7 ;[8] |
|
363 |
|
364 bitstuffN: |
|
365 eor x1, x4 ;[5] |
|
366 ldi x2, 0 ;[6] |
|
367 lpm ;[7] 3 cycle NOP, modifies r0 |
|
368 out USBOUT, x1 ;[10] <-- out |
|
369 rjmp didStuffN ;[0] |
|
370 |
|
371 #define bitStatus x3 |
|
372 |
|
373 sendNakAndReti: |
|
374 ldi cnt, USBPID_NAK ;[-19] |
|
375 rjmp sendCntAndReti ;[-18] |
|
376 sendAckAndReti: |
|
377 ldi cnt, USBPID_ACK ;[-17] |
|
378 sendCntAndReti: |
|
379 mov r0, cnt ;[-16] |
|
380 ldi YL, 0 ;[-15] R0 address is 0 |
|
381 ldi YH, 0 ;[-14] |
|
382 ldi cnt, 2 ;[-13] |
|
383 ; rjmp usbSendAndReti fallthrough |
|
384 |
|
385 ;usbSend: |
|
386 ;pointer to data in 'Y' |
|
387 ;number of bytes in 'cnt' -- including sync byte [range 2 ... 12] |
|
388 ;uses: x1...x4, shift, cnt, Y |
|
389 ;Numbers in brackets are time since first bit of sync pattern is sent |
|
390 usbSendAndReti: ; 12 cycles until SOP |
|
391 in x2, USBDDR ;[-12] |
|
392 ori x2, USBMASK ;[-11] |
|
393 sbi USBOUT, USBMINUS;[-10] prepare idle state; D+ and D- must have been 0 (no pullups) |
|
394 in x1, USBOUT ;[-8] port mirror for tx loop |
|
395 out USBDDR, x2 ;[-7] <- acquire bus |
|
396 ; need not init x2 (bitstuff history) because sync starts with 0 |
|
397 ldi x4, USBMASK ;[-6] exor mask |
|
398 ldi shift, 0x80 ;[-5] sync byte is first byte sent |
|
399 ldi bitStatus, 0xff ;[-4] init bit loop counter, works for up to 12 bytes |
|
400 byteloop: |
|
401 bitloop: |
|
402 sbrs shift, 0 ;[8] [-3] |
|
403 eor x1, x4 ;[9] [-2] |
|
404 out USBOUT, x1 ;[10] [-1] <-- out |
|
405 ror shift ;[0] |
|
406 ror x2 ;[1] |
|
407 didStuffN: |
|
408 cpi x2, 0xfc ;[2] |
|
409 brcc bitstuffN ;[3] |
|
410 nop ;[4] |
|
411 subi bitStatus, 37 ;[5] 256 / 7 ~=~ 37 |
|
412 brcc bitloop ;[6] when we leave the loop, bitStatus has almost the initial value |
|
413 sbrs shift, 0 ;[7] |
|
414 eor x1, x4 ;[8] |
|
415 ror shift ;[9] |
|
416 didStuff7: |
|
417 out USBOUT, x1 ;[10] <-- out |
|
418 ror x2 ;[0] |
|
419 cpi x2, 0xfc ;[1] |
|
420 brcc bitstuff7 ;[2] |
|
421 ld shift, y+ ;[3] |
|
422 dec cnt ;[5] |
|
423 brne byteloop ;[6] |
|
424 ;make SE0: |
|
425 cbr x1, USBMASK ;[7] prepare SE0 [spec says EOP may be 21 to 25 cycles] |
|
426 lds x2, usbNewDeviceAddr;[8] |
|
427 lsl x2 ;[10] we compare with left shifted address |
|
428 out USBOUT, x1 ;[11] <-- out SE0 -- from now 2 bits = 22 cycles until bus idle |
|
429 ;2006-03-06: moved transfer of new address to usbDeviceAddr from C-Code to asm: |
|
430 ;set address only after data packet was sent, not after handshake |
|
431 subi YL, 2 ;[0] Only assign address on data packets, not ACK/NAK in r0 |
|
432 sbci YH, 0 ;[1] |
|
433 breq skipAddrAssign ;[2] |
|
434 sts usbDeviceAddr, x2; if not skipped: SE0 is one cycle longer |
|
435 skipAddrAssign: |
|
436 ;end of usbDeviceAddress transfer |
|
437 ldi x2, 1<<USB_INTR_PENDING_BIT;[4] int0 occurred during TX -- clear pending flag |
|
438 USB_STORE_PENDING(x2) ;[5] |
|
439 ori x1, USBIDLE ;[6] |
|
440 in x2, USBDDR ;[7] |
|
441 cbr x2, USBMASK ;[8] set both pins to input |
|
442 mov x3, x1 ;[9] |
|
443 cbr x3, USBMASK ;[10] configure no pullup on both pins |
|
444 ldi x4, 4 ;[11] |
|
445 se0Delay: |
|
446 dec x4 ;[12] [15] [18] [21] |
|
447 brne se0Delay ;[13] [16] [19] [22] |
|
448 out USBOUT, x1 ;[23] <-- out J (idle) -- end of SE0 (EOP signal) |
|
449 out USBDDR, x2 ;[24] <-- release bus now |
|
450 out USBOUT, x3 ;[25] <-- ensure no pull-up resistors are active |
|
451 rjmp doReturn |
|
452 |