ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev_iouring.c
(Generate patch)

Comparing libev/ev_iouring.c (file contents):
Revision 1.17 by root, Sat Dec 28 07:39:18 2019 UTC vs.
Revision 1.23 by root, Sun Jul 26 11:10:45 2020 UTC

1/* 1/*
2 * libev linux io_uring fd activity backend 2 * libev linux io_uring fd activity backend
3 * 3 *
4 * Copyright (c) 2019 Marc Alexander Lehmann <libev@schmorp.de> 4 * Copyright (c) 2019-2020 Marc Alexander Lehmann <libev@schmorp.de>
5 * All rights reserved. 5 * All rights reserved.
6 * 6 *
7 * Redistribution and use in source and binary forms, with or without modifica- 7 * Redistribution and use in source and binary forms, with or without modifica-
8 * tion, are permitted provided that the following conditions are met: 8 * tion, are permitted provided that the following conditions are met:
9 * 9 *
… …
44 * b) best is not necessarily very good. 44 * b) best is not necessarily very good.
45 * c) it's better than the aio mess, doesn't suffer from the fork problems 45 * c) it's better than the aio mess, doesn't suffer from the fork problems
46 * of linux aio or epoll and so on and so on. and you could do event stuff 46 * of linux aio or epoll and so on and so on. and you could do event stuff
47 * without any syscalls. what's not to like? 47 * without any syscalls. what's not to like?
48 * d) ok, it's vastly more complex, but that's ok, really. 48 * d) ok, it's vastly more complex, but that's ok, really.
49 * e) why 3 mmaps instead of one? one would be more space-efficient, 49 * e) why two mmaps instead of one? one would be more space-efficient,
50 * and I can't see what benefit three would have (other than being 50 * and I can't see what benefit two would have (other than being
51 * somehow resizable/relocatable, but that's apparently not possible). 51 * somehow resizable/relocatable, but that's apparently not possible).
52 * (FIXME: newer kernels can use 2 mmaps only, need to look into this).
53 * f) hmm, it's practiclaly undebuggable (gdb can't access the memory, and 52 * f) hmm, it's practically undebuggable (gdb can't access the memory, and
54 * the bizarre way structure offsets are communicated makes it hard to 53 * the bizarre way structure offsets are communicated makes it hard to
55 * just print the ring buffer heads, even *iff* the memory were visible 54 * just print the ring buffer heads, even *iff* the memory were visible
56 * in gdb. but then, that's also ok, really. 55 * in gdb. but then, that's also ok, really.
57 * g) well, you cannot specify a timeout when waiting for events. no, 56 * g) well, you cannot specify a timeout when waiting for events. no,
58 * seriously, the interface doesn't support a timeout. never seen _that_ 57 * seriously, the interface doesn't support a timeout. never seen _that_
59 * before. sure, you can use a timerfd, but that's another syscall 58 * before. sure, you can use a timerfd, but that's another syscall
60 * you could have avoided. overall, this bizarre omission smells 59 * you could have avoided. overall, this bizarre omission smells
61 * like a µ-optimisation by the io_uring author for his personal 60 * like a µ-optimisation by the io_uring author for his personal
62 * applications, to the detriment of everybody else who just wants 61 * applications, to the detriment of everybody else who just wants
63 * an event loop. but, umm, ok, if that's all, it could be worse. 62 * an event loop. but, umm, ok, if that's all, it could be worse.
64 * (FIXME: jens mentioned timeout commands, need to investigate) 63 * (from what I gather from the author Jens Axboe, it simply didn't
64 * occur to him, and he made good on it by adding an unlimited number
65 * of timeouts later :).
65 * h) there is a hardcoded limit of 4096 outstanding events. okay, 66 * h) initially there was a hardcoded limit of 4096 outstanding events.
66 * at least there is no arbitrary low system-wide limit... 67 * later versions not only bump this to 32k, but also can handle
67 * (FIXME: apparently, this was increased to 32768 in later kernels( 68 * an unlimited amount of events, so this only affects the batch size.
68 * i) unlike linux aio, you *can* register more then the limit 69 * i) unlike linux aio, you *can* register more then the limit
69 * of fd events, and the kernel will "gracefully" signal an 70 * of fd events. while early verisons of io_uring signalled an overflow
70 * overflow, after which you could destroy and recreate the kernel 71 * and you ended up getting wet. 5.5+ does not do this anymore.
71 * state, a bit bigger, or fall back to e.g. poll. thats not
72 * totally insane, but kind of questions the point a high
73 * performance I/O framework when it doesn't really work
74 * under stress.
75 * (FIXME: iouring should no longer drop events, need to investigate)
76 * j) but, oh my! is has exactly the same bugs as the linux aio backend, 72 * j) but, oh my! it had exactly the same bugs as the linux aio backend,
77 * where some undocumented poll combinations just fail. 73 * where some undocumented poll combinations just fail. fortunately,
78 * so we need epoll AGAIN as a fallback. AGAIN! epoll!! and of course, 74 * after finally reaching the author, he was more than willing to fix
79 * this is completely undocumented, have I mantioned this already? 75 * this probably in 5.6+.
80 * k) overall, the *API* itself is, I dare to say, not a total trainwreck. 76 * k) overall, the *API* itself is, I dare to say, not a total trainwreck.
81 * the big isuess with it are the bugs requiring epoll, which might 77 * once the bugs ae fixed (probably in 5.6+), it will be without
82 * or might not get fixed (do I hold my breath?). 78 * competition.
83 */ 79 */
84 80
85/* TODO: use internal TIMEOUT */ 81/* TODO: use internal TIMEOUT */
86/* TODO: take advantage of single mmap, NODROP etc. */ 82/* TODO: take advantage of single mmap, NODROP etc. */
87/* TODO: resize cq/sq size independently */ 83/* TODO: resize cq/sq size independently */
… …
122 __u32 timeout_flags; 118 __u32 timeout_flags;
123 __u32 accept_flags; 119 __u32 accept_flags;
124 __u32 cancel_flags; 120 __u32 cancel_flags;
125 __u32 open_flags; 121 __u32 open_flags;
126 __u32 statx_flags; 122 __u32 statx_flags;
123 __u32 fadvise_advice;
127 }; 124 };
128 __u64 user_data; 125 __u64 user_data;
129 union { 126 union {
130 __u16 buf_index; 127 __u16 buf_index;
128 __u16 personality;
131 __u64 __pad2[3]; 129 __u64 __pad2[3];
132 }; 130 };
133}; 131};
134 132
135struct io_uring_cqe 133struct io_uring_cqe
… …
174 __u32 resv[4]; 172 __u32 resv[4];
175 struct io_sqring_offsets sq_off; 173 struct io_sqring_offsets sq_off;
176 struct io_cqring_offsets cq_off; 174 struct io_cqring_offsets cq_off;
177}; 175};
178 176
177#define IORING_FEAT_SINGLE_MMAP 0x00000001
178#define IORING_FEAT_NODROP 0x00000002
179#define IORING_FEAT_SUBMIT_STABLE 0x00000004
180
179#define IORING_SETUP_CQSIZE 0x00000008 181#define IORING_SETUP_CQSIZE 0x00000008
182#define IORING_SETUP_CLAMP 0x00000010
180 183
181#define IORING_OP_POLL_ADD 6 184#define IORING_OP_POLL_ADD 6
182#define IORING_OP_POLL_REMOVE 7 185#define IORING_OP_POLL_REMOVE 7
183#define IORING_OP_TIMEOUT 11 186#define IORING_OP_TIMEOUT 11
184#define IORING_OP_TIMEOUT_REMOVE 12 187#define IORING_OP_TIMEOUT_REMOVE 12
185 188
189#define IORING_REGISTER_EVENTFD 4
190#define IORING_REGISTER_EVENTFD_ASYNC 7
191#define IORING_REGISTER_PROBE 8
192
193#define IO_URING_OP_SUPPORTED 1
194
195struct io_uring_probe_op {
196 __u8 op;
197 __u8 resv;
198 __u16 flags;
199 __u32 resv2;
200};
201
202struct io_uring_probe
203{
204 __u8 last_op;
205 __u8 ops_len;
206 __u16 resv;
207 __u32 resv2[3];
208 struct io_uring_probe_op ops[0];
209};
210
186/* relative or absolute, reference clock is CLOCK_MONOTONIC */ 211/* relative or absolute, reference clock is CLOCK_MONOTONIC */
187struct iouring_kernel_timespec 212struct iouring_kernel_timespec
188{ 213{
189 int64_t tv_sec; 214 int64_t tv_sec;
190 long long tv_nsec; 215 long long tv_nsec;
… …
193#define IORING_TIMEOUT_ABS 0x00000001 218#define IORING_TIMEOUT_ABS 0x00000001
194 219
195#define IORING_ENTER_GETEVENTS 0x01 220#define IORING_ENTER_GETEVENTS 0x01
196 221
197#define IORING_OFF_SQ_RING 0x00000000ULL 222#define IORING_OFF_SQ_RING 0x00000000ULL
198#define IORING_OFF_CQ_RING 0x08000000ULL
199#define IORING_OFF_SQES 0x10000000ULL 223#define IORING_OFF_SQES 0x10000000ULL
200 224
201#define IORING_FEAT_SINGLE_MMAP 0x00000001 225#define IORING_FEAT_SINGLE_MMAP 0x00000001
202#define IORING_FEAT_NODROP 0x00000002 226#define IORING_FEAT_NODROP 0x00000002
203#define IORING_FEAT_SUBMIT_STABLE 0x00000004 227#define IORING_FEAT_SUBMIT_STABLE 0x00000004
… …
214evsys_io_uring_enter (int fd, unsigned to_submit, unsigned min_complete, unsigned flags, const sigset_t *sig, size_t sigsz) 238evsys_io_uring_enter (int fd, unsigned to_submit, unsigned min_complete, unsigned flags, const sigset_t *sig, size_t sigsz)
215{ 239{
216 return ev_syscall6 (SYS_io_uring_enter, fd, to_submit, min_complete, flags, sig, sigsz); 240 return ev_syscall6 (SYS_io_uring_enter, fd, to_submit, min_complete, flags, sig, sigsz);
217} 241}
218 242
243inline_size
244int
245evsys_io_uring_register (unsigned int fd, unsigned int opcode, void *arg, unsigned int nr_args)
246{
247 return ev_syscall4 (SYS_io_uring_register, fd, opcode, arg, nr_args);
248}
249
219/*****************************************************************************/ 250/*****************************************************************************/
220/* actual backed implementation */ 251/* actual backend implementation */
221 252
222/* we hope that volatile will make the compiler access this variables only once */ 253/* we hope that volatile will make the compiler access this variables only once */
223#define EV_SQ_VAR(name) *(volatile unsigned *)((char *)iouring_sq_ring + iouring_sq_ ## name)
224#define EV_CQ_VAR(name) *(volatile unsigned *)((char *)iouring_cq_ring + iouring_cq_ ## name) 254#define EV_SQ_VAR(name) *(volatile unsigned *)((char *)iouring_ring + iouring_sq_ ## name)
255#define EV_CQ_VAR(name) *(volatile unsigned *)((char *)iouring_ring + iouring_cq_ ## name)
225 256
226/* the index array */ 257/* the index array */
227#define EV_SQ_ARRAY ((unsigned *)((char *)iouring_sq_ring + iouring_sq_array)) 258#define EV_SQ_ARRAY ((unsigned *)((char *)iouring_ring + iouring_sq_array))
228 259
229/* the submit/completion queue entries */ 260/* the submit/completion queue entries */
230#define EV_SQES ((struct io_uring_sqe *) iouring_sqes) 261#define EV_SQES ((struct io_uring_sqe *) iouring_sqes)
231#define EV_CQES ((struct io_uring_cqe *)((char *)iouring_cq_ring + iouring_cq_cqes)) 262#define EV_CQES ((struct io_uring_cqe *)((char *)iouring_ring + iouring_cq_cqes))
232 263
233inline_speed 264inline_speed
234int 265int
235iouring_enter (EV_P_ ev_tstamp timeout) 266iouring_enter (EV_P_ ev_tstamp timeout)
236{ 267{
… …
293inline_size 324inline_size
294struct io_uring_sqe * 325struct io_uring_sqe *
295iouring_sqe_submit (EV_P_ struct io_uring_sqe *sqe) 326iouring_sqe_submit (EV_P_ struct io_uring_sqe *sqe)
296{ 327{
297 unsigned idx = sqe - EV_SQES; 328 unsigned idx = sqe - EV_SQES;
329
330 printf ("submit idx %d, op %d, fd %d, us5r %p, poll %d\n", idx, sqe->opcode, sqe->fd, sqe->user_data, sqe->poll_events);
298 331
299 EV_SQ_ARRAY [idx] = idx; 332 EV_SQ_ARRAY [idx] = idx;
300 ECB_MEMORY_FENCE_RELEASE; 333 ECB_MEMORY_FENCE_RELEASE;
301 ++EV_SQ_VAR (tail); 334 ++EV_SQ_VAR (tail);
302 /*ECB_MEMORY_FENCE_RELEASE; /* for the time being we assume this is not needed */ 335 /*ECB_MEMORY_FENCE_RELEASE; /* for the time being we assume this is not needed */
… …
321iouring_internal_destroy (EV_P) 354iouring_internal_destroy (EV_P)
322{ 355{
323 close (iouring_tfd); 356 close (iouring_tfd);
324 close (iouring_fd); 357 close (iouring_fd);
325 358
326 if (iouring_sq_ring != MAP_FAILED) munmap (iouring_sq_ring, iouring_sq_ring_size); 359 if (iouring_ring != MAP_FAILED) munmap (iouring_ring, iouring_ring_size);
327 if (iouring_cq_ring != MAP_FAILED) munmap (iouring_cq_ring, iouring_cq_ring_size);
328 if (iouring_sqes != MAP_FAILED) munmap (iouring_sqes , iouring_sqes_size ); 360 if (iouring_sqes != MAP_FAILED) munmap (iouring_sqes, iouring_sqes_size);
329 361
330 if (ev_is_active (&iouring_tfd_w)) 362 if (ev_is_active (&iouring_tfd_w))
331 { 363 {
332 ev_ref (EV_A); 364 ev_ref (EV_A);
333 ev_io_stop (EV_A_ &iouring_tfd_w); 365 ev_io_stop (EV_A_ &iouring_tfd_w);
… …
337ecb_cold 369ecb_cold
338static int 370static int
339iouring_internal_init (EV_P) 371iouring_internal_init (EV_P)
340{ 372{
341 struct io_uring_params params = { 0 }; 373 struct io_uring_params params = { 0 };
374 uint32_t sq_size, cq_size;
375
376 params.flags = IORING_SETUP_CLAMP;
342 377
343 iouring_to_submit = 0; 378 iouring_to_submit = 0;
344 379
345 iouring_tfd = -1; 380 iouring_tfd = -1;
346 iouring_sq_ring = MAP_FAILED; 381 iouring_ring = MAP_FAILED;
347 iouring_cq_ring = MAP_FAILED;
348 iouring_sqes = MAP_FAILED; 382 iouring_sqes = MAP_FAILED;
349 383
350 if (!have_monotonic) /* cannot really happen, but what if11 */ 384 if (!have_monotonic) /* cannot really happen, but what if11 */
351 return -1; 385 return -1;
352 386
353 for (;;)
354 {
355 iouring_fd = evsys_io_uring_setup (iouring_entries, &params); 387 iouring_fd = evsys_io_uring_setup (iouring_entries, &params);
356 388
357 if (iouring_fd >= 0) 389 if (iouring_fd < 0)
358 break; /* yippie */ 390 return -1;
359 391
360 if (errno != EINVAL)
361 return -1; /* we failed */
362
363#if TODO
364 if ((~params.features) & (IORING_FEAT_NODROP | IORING_FEATURE_SINGLE_MMAP | IORING_FEAT_SUBMIT_STABLE)) 392 if ((~params.features) & (IORING_FEAT_NODROP | IORING_FEAT_SINGLE_MMAP | IORING_FEAT_SUBMIT_STABLE))
365 return -1; /* we require the above features */ 393 return -1; /* we require the above features */
366#endif
367 394
368 /* EINVAL: lots of possible reasons, but maybe 395 /* TODO: remember somehow whether our queue size has been clamped */
369 * it is because we hit the unqueryable hardcoded size limit
370 */
371 396
372 /* we hit the limit already, give up */
373 if (iouring_max_entries)
374 return -1;
375
376 /* first time we hit EINVAL? assume we hit the limit, so go back and retry */
377 iouring_entries >>= 1;
378 iouring_max_entries = iouring_entries;
379 }
380
381 iouring_sq_ring_size = params.sq_off.array + params.sq_entries * sizeof (unsigned); 397 sq_size = params.sq_off.array + params.sq_entries * sizeof (unsigned);
382 iouring_cq_ring_size = params.cq_off.cqes + params.cq_entries * sizeof (struct io_uring_cqe); 398 cq_size = params.cq_off.cqes + params.cq_entries * sizeof (struct io_uring_cqe);
399
400 iouring_ring_size = sq_size > cq_size ? sq_size : cq_size;
383 iouring_sqes_size = params.sq_entries * sizeof (struct io_uring_sqe); 401 iouring_sqes_size = params.sq_entries * sizeof (struct io_uring_sqe);
384 402
385 iouring_sq_ring = mmap (0, iouring_sq_ring_size, PROT_READ | PROT_WRITE, 403 iouring_ring = mmap (0, iouring_ring_size, PROT_READ | PROT_WRITE,
386 MAP_SHARED | MAP_POPULATE, iouring_fd, IORING_OFF_SQ_RING); 404 MAP_SHARED | MAP_POPULATE, iouring_fd, IORING_OFF_SQ_RING);
387 iouring_cq_ring = mmap (0, iouring_cq_ring_size, PROT_READ | PROT_WRITE,
388 MAP_SHARED | MAP_POPULATE, iouring_fd, IORING_OFF_CQ_RING);
389 iouring_sqes = mmap (0, iouring_sqes_size, PROT_READ | PROT_WRITE, 405 iouring_sqes = mmap (0, iouring_sqes_size, PROT_READ | PROT_WRITE,
390 MAP_SHARED | MAP_POPULATE, iouring_fd, IORING_OFF_SQES); 406 MAP_SHARED | MAP_POPULATE, iouring_fd, IORING_OFF_SQES);
391 407
392 if (iouring_sq_ring == MAP_FAILED || iouring_cq_ring == MAP_FAILED || iouring_sqes == MAP_FAILED) 408 if (iouring_ring == MAP_FAILED || iouring_sqes == MAP_FAILED)
393 return -1; 409 return -1;
394 410
395 iouring_sq_head = params.sq_off.head; 411 iouring_sq_head = params.sq_off.head;
396 iouring_sq_tail = params.sq_off.tail; 412 iouring_sq_tail = params.sq_off.tail;
397 iouring_sq_ring_mask = params.sq_off.ring_mask; 413 iouring_sq_ring_mask = params.sq_off.ring_mask;
… …
405 iouring_cq_ring_mask = params.cq_off.ring_mask; 421 iouring_cq_ring_mask = params.cq_off.ring_mask;
406 iouring_cq_ring_entries = params.cq_off.ring_entries; 422 iouring_cq_ring_entries = params.cq_off.ring_entries;
407 iouring_cq_overflow = params.cq_off.overflow; 423 iouring_cq_overflow = params.cq_off.overflow;
408 iouring_cq_cqes = params.cq_off.cqes; 424 iouring_cq_cqes = params.cq_off.cqes;
409 425
426 iouring_tfd_to = EV_TSTAMP_HUGE;
427
410 iouring_tfd = timerfd_create (CLOCK_MONOTONIC, TFD_CLOEXEC); 428 iouring_tfd = timerfd_create (CLOCK_MONOTONIC, TFD_CLOEXEC);
411 429
412 if (iouring_tfd < 0) 430 if (iouring_tfd < 0)
413 return iouring_tfd; 431 return -1;
414
415 iouring_tfd_to = EV_TSTAMP_HUGE;
416 432
417 return 0; 433 return 0;
418} 434}
419 435
420ecb_cold 436ecb_cold
… …
460 if (nev) 476 if (nev)
461 { 477 {
462 struct io_uring_sqe *sqe = iouring_sqe_get (EV_A); 478 struct io_uring_sqe *sqe = iouring_sqe_get (EV_A);
463 sqe->opcode = IORING_OP_POLL_ADD; 479 sqe->opcode = IORING_OP_POLL_ADD;
464 sqe->fd = fd; 480 sqe->fd = fd;
481 sqe->addr = 0;
465 sqe->user_data = (uint32_t)fd | ((__u64)(uint32_t)anfds [fd].egen << 32); 482 sqe->user_data = (uint32_t)fd | ((__u64)(uint32_t)anfds [fd].egen << 32);
466 sqe->poll_events = 483 sqe->poll_events =
467 (nev & EV_READ ? POLLIN : 0) 484 (nev & EV_READ ? POLLIN : 0)
468 | (nev & EV_WRITE ? POLLOUT : 0); 485 | (nev & EV_WRITE ? POLLOUT : 0);
469 iouring_sqe_submit (EV_A_ sqe); 486 iouring_sqe_submit (EV_A_ sqe);
… …
520 return; 537 return;
521 538
522 if (ecb_expect_false (res < 0)) 539 if (ecb_expect_false (res < 0))
523 { 540 {
524 /*TODO: EINVAL handling (was something failed with this fd)*/ 541 /*TODO: EINVAL handling (was something failed with this fd)*/
525 /*TODO: EBUSY happens when?*/
526 542
527 if (res == -EBADF) 543 if (res == -EBADF)
528 { 544 {
529 assert (("libev: event loop rejected bad fd", res != -EBADF)); 545 assert (("libev: event loop rejected bad fd", res != -EBADF));
530 fd_kill (EV_A_ fd); 546 fd_kill (EV_A_ fd);

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines