1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 ========================================= 3 =========================================
4 IAA Compression Accelerator Crypto Driver 4 IAA Compression Accelerator Crypto Driver
5 ========================================= 5 =========================================
6 6
7 Tom Zanussi <tom.zanussi@linux.intel.com> 7 Tom Zanussi <tom.zanussi@linux.intel.com>
8 8
9 The IAA crypto driver supports compression/dec 9 The IAA crypto driver supports compression/decompression compatible
10 with the DEFLATE compression standard describe 10 with the DEFLATE compression standard described in RFC 1951, which is
11 the compression/decompression algorithm export 11 the compression/decompression algorithm exported by this module.
12 12
13 The IAA hardware spec can be found here: 13 The IAA hardware spec can be found here:
14 14
15 https://cdrdv2.intel.com/v1/dl/getContent/72 15 https://cdrdv2.intel.com/v1/dl/getContent/721858
16 16
17 The iaa_crypto driver is designed to work as a 17 The iaa_crypto driver is designed to work as a layer underneath
18 higher-level compression devices such as zswap 18 higher-level compression devices such as zswap.
19 19
20 Users can select IAA compress/decompress accel 20 Users can select IAA compress/decompress acceleration by specifying
21 one of the supported IAA compression algorithm 21 one of the supported IAA compression algorithms in whatever facility
22 allows compression algorithms to be selected. 22 allows compression algorithms to be selected.
23 23
24 For example, a zswap device can select the IAA 24 For example, a zswap device can select the IAA 'fixed' mode
25 represented by selecting the 'deflate-iaa' cry 25 represented by selecting the 'deflate-iaa' crypto compression
26 algorithm:: 26 algorithm::
27 27
28 # echo deflate-iaa > /sys/module/zswap/param 28 # echo deflate-iaa > /sys/module/zswap/parameters/compressor
29 29
30 This will tell zswap to use the IAA 'fixed' co 30 This will tell zswap to use the IAA 'fixed' compression mode for all
31 compresses and decompresses. 31 compresses and decompresses.
32 32
33 Currently, there is only one compression modes 33 Currently, there is only one compression modes available, 'fixed'
34 mode. 34 mode.
35 35
36 The 'fixed' compression mode implements the co 36 The 'fixed' compression mode implements the compression scheme
37 specified by RFC 1951 and is given the crypto 37 specified by RFC 1951 and is given the crypto algorithm name
38 'deflate-iaa'. (Because the IAA hardware has 38 'deflate-iaa'. (Because the IAA hardware has a 4k history-window
39 limitation, only buffers <= 4k, or that have b 39 limitation, only buffers <= 4k, or that have been compressed using a
40 <= 4k history window, are technically complian 40 <= 4k history window, are technically compliant with the deflate spec,
41 which allows for a window of up to 32k. Becau 41 which allows for a window of up to 32k. Because of this limitation,
42 the IAA fixed mode deflate algorithm is given 42 the IAA fixed mode deflate algorithm is given its own algorithm name
43 rather than simply 'deflate'). 43 rather than simply 'deflate').
44 44
45 45
46 Config options and other setup 46 Config options and other setup
47 ============================== 47 ==============================
48 48
49 The IAA crypto driver is available via menucon 49 The IAA crypto driver is available via menuconfig using the following
50 path:: 50 path::
51 51
52 Cryptographic API -> Hardware crypto devices 52 Cryptographic API -> Hardware crypto devices -> Support for Intel(R) IAA Compression Accelerator
53 53
54 In the configuration file the option called CO 54 In the configuration file the option called CONFIG_CRYPTO_DEV_IAA_CRYPTO.
55 55
56 The IAA crypto driver also supports statistics 56 The IAA crypto driver also supports statistics, which are available
57 via menuconfig using the following path:: 57 via menuconfig using the following path::
58 58
59 Cryptographic API -> Hardware crypto devices 59 Cryptographic API -> Hardware crypto devices -> Support for Intel(R) IAA Compression -> Enable Intel(R) IAA Compression Accelerator Statistics
60 60
61 In the configuration file the option called CO 61 In the configuration file the option called CONFIG_CRYPTO_DEV_IAA_CRYPTO_STATS.
62 62
63 The following config options should also be en 63 The following config options should also be enabled::
64 64
65 CONFIG_IRQ_REMAP=y 65 CONFIG_IRQ_REMAP=y
66 CONFIG_INTEL_IOMMU=y 66 CONFIG_INTEL_IOMMU=y
67 CONFIG_INTEL_IOMMU_SVM=y 67 CONFIG_INTEL_IOMMU_SVM=y
68 CONFIG_PCI_ATS=y 68 CONFIG_PCI_ATS=y
69 CONFIG_PCI_PRI=y 69 CONFIG_PCI_PRI=y
70 CONFIG_PCI_PASID=y 70 CONFIG_PCI_PASID=y
71 CONFIG_INTEL_IDXD=m 71 CONFIG_INTEL_IDXD=m
72 CONFIG_INTEL_IDXD_SVM=y 72 CONFIG_INTEL_IDXD_SVM=y
73 73
74 IAA is one of the first Intel accelerator IPs 74 IAA is one of the first Intel accelerator IPs that can work in
75 conjunction with the Intel IOMMU. There are m 75 conjunction with the Intel IOMMU. There are multiple modes that exist
76 for testing. Based on IOMMU configuration, the 76 for testing. Based on IOMMU configuration, there are 3 modes::
77 77
78 - Scalable 78 - Scalable
79 - Legacy 79 - Legacy
80 - No IOMMU 80 - No IOMMU
81 81
82 82
83 Scalable mode 83 Scalable mode
84 ------------- 84 -------------
85 85
86 Scalable mode supports Shared Virtual Memory ( 86 Scalable mode supports Shared Virtual Memory (SVM or SVA). It is
87 entered when using the kernel boot commandline 87 entered when using the kernel boot commandline::
88 88
89 intel_iommu=on,sm_on 89 intel_iommu=on,sm_on
90 90
91 with VT-d turned on in BIOS. 91 with VT-d turned on in BIOS.
92 92
93 With scalable mode, both shared and dedicated 93 With scalable mode, both shared and dedicated workqueues are available
94 for use. 94 for use.
95 95
96 For scalable mode, the following BIOS settings 96 For scalable mode, the following BIOS settings should be enabled::
97 97
98 Socket Configuration > IIO Configuration > I 98 Socket Configuration > IIO Configuration > Intel VT for Directed I/O (VT-d) > Intel VT for Directed I/O
99 99
100 Socket Configuration > IIO Configuration > P 100 Socket Configuration > IIO Configuration > PCIe ENQCMD > ENQCMDS
101 101
102 102
103 Legacy mode 103 Legacy mode
104 ----------- 104 -----------
105 105
106 Legacy mode is entered when using the kernel b 106 Legacy mode is entered when using the kernel boot commandline::
107 107
108 intel_iommu=off 108 intel_iommu=off
109 109
110 or VT-d is not turned on in BIOS. 110 or VT-d is not turned on in BIOS.
111 111
112 If you have booted into Linux and not sure if 112 If you have booted into Linux and not sure if VT-d is on, do a "dmesg
113 | grep -i dmar". If you don't see a number of 113 | grep -i dmar". If you don't see a number of DMAR devices enumerated,
114 most likely VT-d is not on. 114 most likely VT-d is not on.
115 115
116 With legacy mode, only dedicated workqueues ar 116 With legacy mode, only dedicated workqueues are available for use.
117 117
118 118
119 No IOMMU mode 119 No IOMMU mode
120 ------------- 120 -------------
121 121
122 No IOMMU mode is entered when using the kernel 122 No IOMMU mode is entered when using the kernel boot commandline::
123 123
124 iommu=off. 124 iommu=off.
125 125
126 With no IOMMU mode, only dedicated workqueues 126 With no IOMMU mode, only dedicated workqueues are available for use.
127 127
128 128
129 Usage 129 Usage
130 ===== 130 =====
131 131
132 accel-config 132 accel-config
133 ------------ 133 ------------
134 134
135 When loaded, the iaa_crypto driver automatical 135 When loaded, the iaa_crypto driver automatically creates a default
136 configuration and enables it, and assigns defa 136 configuration and enables it, and assigns default driver attributes.
137 If a different configuration or set of driver 137 If a different configuration or set of driver attributes is required,
138 the user must first disable the IAA devices an 138 the user must first disable the IAA devices and workqueues, reset the
139 configuration, and then re-register the deflat 139 configuration, and then re-register the deflate-iaa algorithm with the
140 crypto subsystem by removing and reinserting t 140 crypto subsystem by removing and reinserting the iaa_crypto module.
141 141
142 The :ref:`iaa_disable_script` in the 'Use Case 142 The :ref:`iaa_disable_script` in the 'Use Cases'
143 section below can be used to disable the defau 143 section below can be used to disable the default configuration.
144 144
145 See :ref:`iaa_default_config` below for detail 145 See :ref:`iaa_default_config` below for details of the default
146 configuration. 146 configuration.
147 147
148 More likely than not, however, and because of 148 More likely than not, however, and because of the complexity and
149 configurability of the accelerator devices, th 149 configurability of the accelerator devices, the user will want to
150 configure the device and manually enable the d 150 configure the device and manually enable the desired devices and
151 workqueues. 151 workqueues.
152 152
153 The userspace tool to help doing that is calle 153 The userspace tool to help doing that is called accel-config. Using
154 accel-config to configure device or loading a 154 accel-config to configure device or loading a previously saved config
155 is highly recommended. The device can be cont 155 is highly recommended. The device can be controlled via sysfs
156 directly but comes with the warning that you s 156 directly but comes with the warning that you should do this ONLY if
157 you know exactly what you are doing. The foll 157 you know exactly what you are doing. The following sections will not
158 cover the sysfs interface but assumes you will 158 cover the sysfs interface but assumes you will be using accel-config.
159 159
160 The :ref:`iaa_sysfs_config` section in the app 160 The :ref:`iaa_sysfs_config` section in the appendix below can be
161 consulted for the sysfs interface details if i 161 consulted for the sysfs interface details if interested.
162 162
163 The accel-config tool along with instructions 163 The accel-config tool along with instructions for building it can be
164 found here: 164 found here:
165 165
166 https://github.com/intel/idxd-config/#readme 166 https://github.com/intel/idxd-config/#readme
167 167
168 Typical usage 168 Typical usage
169 ------------- 169 -------------
170 170
171 In order for the iaa_crypto module to actually 171 In order for the iaa_crypto module to actually do any
172 compression/decompression work on behalf of a 172 compression/decompression work on behalf of a facility, one or more
173 IAA workqueues need to be bound to the iaa_cry 173 IAA workqueues need to be bound to the iaa_crypto driver.
174 174
175 For instance, here's an example of configuring 175 For instance, here's an example of configuring an IAA workqueue and
176 binding it to the iaa_crypto driver (note that 176 binding it to the iaa_crypto driver (note that device names are
177 specified as 'iax' rather than 'iaa' - this is 177 specified as 'iax' rather than 'iaa' - this is because upstream still
178 has the old 'iax' device naming in place) :: 178 has the old 'iax' device naming in place) ::
179 179
180 # configure wq1.0 180 # configure wq1.0
181 181
182 accel-config config-wq --group-id=0 --mode=d 182 accel-config config-wq --group-id=0 --mode=dedicated --type=kernel --priority=10 --name="iaa_crypto" --driver-name="crypto" iax1/wq1.0
183 183
184 accel-config config-engine iax1/engine1.0 -- 184 accel-config config-engine iax1/engine1.0 --group-id=0
185 185
186 # enable IAA device iax1 186 # enable IAA device iax1
187 187
188 accel-config enable-device iax1 188 accel-config enable-device iax1
189 189
190 # enable wq1.0 on IAX device iax1 190 # enable wq1.0 on IAX device iax1
191 191
192 accel-config enable-wq iax1/wq1.0 192 accel-config enable-wq iax1/wq1.0
193 193
194 Whenever a new workqueue is bound to or unboun 194 Whenever a new workqueue is bound to or unbound from the iaa_crypto
195 driver, the available workqueues are 'rebalanc 195 driver, the available workqueues are 'rebalanced' such that work
196 submitted from a particular CPU is given to th 196 submitted from a particular CPU is given to the most appropriate
197 workqueue available. Current best practice is 197 workqueue available. Current best practice is to configure and bind
198 at least one workqueue for each IAA device, bu 198 at least one workqueue for each IAA device, but as long as there is at
199 least one workqueue configured and bound to an 199 least one workqueue configured and bound to any IAA device in the
200 system, the iaa_crypto driver will work, albei 200 system, the iaa_crypto driver will work, albeit most likely not as
201 efficiently. 201 efficiently.
202 202
203 The IAA crypto algorigthms is operational and 203 The IAA crypto algorigthms is operational and compression and
204 decompression operations are fully enabled fol 204 decompression operations are fully enabled following the successful
205 binding of the first IAA workqueue to the iaa_ 205 binding of the first IAA workqueue to the iaa_crypto driver.
206 206
207 Similarly, the IAA crypto algorithm is not ope 207 Similarly, the IAA crypto algorithm is not operational and compression
208 and decompression operations are disabled foll 208 and decompression operations are disabled following the unbinding of
209 the last IAA worqueue to the iaa_crypto driver 209 the last IAA worqueue to the iaa_crypto driver.
210 210
211 As a result, the IAA crypto algorithms and thu 211 As a result, the IAA crypto algorithms and thus the IAA hardware are
212 only available when one or more workques are b 212 only available when one or more workques are bound to the iaa_crypto
213 driver. 213 driver.
214 214
215 When there are no IAA workqueues bound to the 215 When there are no IAA workqueues bound to the driver, the IAA crypto
216 algorithms can be unregistered by removing the 216 algorithms can be unregistered by removing the module.
217 217
218 218
219 Driver attributes 219 Driver attributes
220 ----------------- 220 -----------------
221 221
222 There are a couple user-configurable driver at 222 There are a couple user-configurable driver attributes that can be
223 used to configure various modes of operation. 223 used to configure various modes of operation. They're listed below,
224 along with their default values. To set any o 224 along with their default values. To set any of these attributes, echo
225 the appropriate values to the attribute file l 225 the appropriate values to the attribute file located under
226 /sys/bus/dsa/drivers/crypto/ 226 /sys/bus/dsa/drivers/crypto/
227 227
228 The attribute settings at the time the IAA alg 228 The attribute settings at the time the IAA algorithms are registered
229 are captured in each algorithm's crypto_ctx an 229 are captured in each algorithm's crypto_ctx and used for all compresses
230 and decompresses when using that algorithm. 230 and decompresses when using that algorithm.
231 231
232 The available attributes are: 232 The available attributes are:
233 233
234 - verify_compress 234 - verify_compress
235 235
236 Toggle compression verification. If set, 236 Toggle compression verification. If set, each compress will be
237 internally decompressed and the contents v 237 internally decompressed and the contents verified, returning error
238 codes if unsuccessful. This can be toggle 238 codes if unsuccessful. This can be toggled with 0/1::
239 239
240 echo 0 > /sys/bus/dsa/drivers/crypto/ver 240 echo 0 > /sys/bus/dsa/drivers/crypto/verify_compress
241 241
242 The default setting is '1' - verify all co 242 The default setting is '1' - verify all compresses.
243 243
244 - sync_mode 244 - sync_mode
245 245
246 Select mode to be used to wait for complet 246 Select mode to be used to wait for completion of each compresses
247 and decompress operation. 247 and decompress operation.
248 248
249 The crypto async interface support impleme 249 The crypto async interface support implemented by iaa_crypto
250 provides an implementation that satisfies 250 provides an implementation that satisfies the interface but does
251 so in a synchronous manner - it fills and 251 so in a synchronous manner - it fills and submits the IDXD
252 descriptor and then loops around waiting f 252 descriptor and then loops around waiting for it to complete before
253 returning. This isn't a problem at the mo 253 returning. This isn't a problem at the moment, since all existing
254 callers (e.g. zswap) wrap any asynchronous 254 callers (e.g. zswap) wrap any asynchronous callees in a
255 synchronous wrapper anyway. 255 synchronous wrapper anyway.
256 256
257 The iaa_crypto driver does however provide 257 The iaa_crypto driver does however provide true asynchronous
258 support for callers that can make use of i 258 support for callers that can make use of it. In this mode, it
259 fills and submits the IDXD descriptor, the 259 fills and submits the IDXD descriptor, then returns immediately
260 with -EINPROGRESS. The caller can then ei 260 with -EINPROGRESS. The caller can then either poll for completion
261 itself, which requires specific code in th 261 itself, which requires specific code in the caller which currently
262 nothing in the upstream kernel implements, 262 nothing in the upstream kernel implements, or go to sleep and wait
263 for an interrupt signaling completion. Th 263 for an interrupt signaling completion. This latter mode is
264 supported by current users in the kernel s 264 supported by current users in the kernel such as zswap via
265 synchronous wrappers. Although it is supp 265 synchronous wrappers. Although it is supported this mode is
266 significantly slower than the synchronous 266 significantly slower than the synchronous mode that does the
267 polling in the iaa_crypto driver previousl 267 polling in the iaa_crypto driver previously mentioned.
268 268
269 This mode can be enabled by writing 'async 269 This mode can be enabled by writing 'async_irq' to the sync_mode
270 iaa_crypto driver attribute:: 270 iaa_crypto driver attribute::
271 271
272 echo async_irq > /sys/bus/dsa/drivers/cr 272 echo async_irq > /sys/bus/dsa/drivers/crypto/sync_mode
273 273
274 Async mode without interrupts (caller must 274 Async mode without interrupts (caller must poll) can be enabled by
275 writing 'async' to it:: 275 writing 'async' to it::
276 276
277 echo async > /sys/bus/dsa/drivers/crypto 277 echo async > /sys/bus/dsa/drivers/crypto/sync_mode
278 278
279 The mode that does the polling in the iaa_ 279 The mode that does the polling in the iaa_crypto driver can be
280 enabled by writing 'sync' to it:: 280 enabled by writing 'sync' to it::
281 281
282 echo sync > /sys/bus/dsa/drivers/crypto/ 282 echo sync > /sys/bus/dsa/drivers/crypto/sync_mode
283 283
284 The default mode is 'sync'. 284 The default mode is 'sync'.
285 285
286 .. _iaa_default_config: 286 .. _iaa_default_config:
287 287
288 IAA Default Configuration 288 IAA Default Configuration
289 ------------------------- 289 -------------------------
290 290
291 When the iaa_crypto driver is loaded, each IAA 291 When the iaa_crypto driver is loaded, each IAA device has a single
292 work queue configured for it, with the followi 292 work queue configured for it, with the following attributes::
293 293
294 mode "dedicated" 294 mode "dedicated"
295 threshold 0 295 threshold 0
296 size Total WQ Size from 296 size Total WQ Size from WQCAP
297 priority 10 297 priority 10
298 type IDXD_WQT_KERNEL 298 type IDXD_WQT_KERNEL
299 group 0 299 group 0
300 name "iaa_crypto" 300 name "iaa_crypto"
301 driver_name "crypto" 301 driver_name "crypto"
302 302
303 The devices and workqueues are also enabled an 303 The devices and workqueues are also enabled and therefore the driver
304 is ready to be used without any additional con 304 is ready to be used without any additional configuration.
305 305
306 The default driver attributes in effect when t 306 The default driver attributes in effect when the driver is loaded are::
307 307
308 sync_mode "sync" 308 sync_mode "sync"
309 verify_compress 1 309 verify_compress 1
310 310
311 In order to change either the device/work queu 311 In order to change either the device/work queue or driver attributes,
312 the enabled devices and workqueues must first 312 the enabled devices and workqueues must first be disabled. In order
313 to have the new configuration applied to the d 313 to have the new configuration applied to the deflate-iaa crypto
314 algorithm, it needs to be re-registered by rem 314 algorithm, it needs to be re-registered by removing and reinserting
315 the iaa_crypto module. The :ref:`iaa_disable_ 315 the iaa_crypto module. The :ref:`iaa_disable_script` in the 'Use
316 Cases' section below can be used to disable th 316 Cases' section below can be used to disable the default configuration.
317 317
318 Statistics 318 Statistics
319 ========== 319 ==========
320 320
321 If the optional debugfs statistics support is 321 If the optional debugfs statistics support is enabled, the IAA crypto
322 driver will generate statistics which can be a 322 driver will generate statistics which can be accessed in debugfs at::
323 323
324 # ls -al /sys/kernel/debug/iaa-crypto/ 324 # ls -al /sys/kernel/debug/iaa-crypto/
325 total 0 325 total 0
326 drwxr-xr-x 2 root root 0 Mar 3 07:55 . 326 drwxr-xr-x 2 root root 0 Mar 3 07:55 .
327 drwx------ 53 root root 0 Mar 3 07:55 .. 327 drwx------ 53 root root 0 Mar 3 07:55 ..
328 -rw-r--r-- 1 root root 0 Mar 3 07:55 globa 328 -rw-r--r-- 1 root root 0 Mar 3 07:55 global_stats
329 -rw-r--r-- 1 root root 0 Mar 3 07:55 stats 329 -rw-r--r-- 1 root root 0 Mar 3 07:55 stats_reset
330 -rw-r--r-- 1 root root 0 Mar 3 07:55 wq_st 330 -rw-r--r-- 1 root root 0 Mar 3 07:55 wq_stats
331 331
332 The global_stats file shows a set of global st 332 The global_stats file shows a set of global statistics collected since
333 the driver has been loaded or reset:: 333 the driver has been loaded or reset::
334 334
335 # cat global_stats 335 # cat global_stats
336 global stats: 336 global stats:
337 total_comp_calls: 4300 337 total_comp_calls: 4300
338 total_decomp_calls: 4164 338 total_decomp_calls: 4164
339 total_sw_decomp_calls: 0 339 total_sw_decomp_calls: 0
340 total_comp_bytes_out: 5993989 340 total_comp_bytes_out: 5993989
341 total_decomp_bytes_in: 5993989 341 total_decomp_bytes_in: 5993989
342 total_completion_einval_errors: 0 342 total_completion_einval_errors: 0
343 total_completion_timeout_errors: 0 343 total_completion_timeout_errors: 0
344 total_completion_comp_buf_overflow_errors: 344 total_completion_comp_buf_overflow_errors: 136
345 345
346 The wq_stats file shows per-wq stats, a set fo 346 The wq_stats file shows per-wq stats, a set for each iaa device and wq
347 in addition to some global stats:: 347 in addition to some global stats::
348 348
349 # cat wq_stats 349 # cat wq_stats
350 iaa device: 350 iaa device:
351 id: 1 351 id: 1
352 n_wqs: 1 352 n_wqs: 1
353 comp_calls: 0 353 comp_calls: 0
354 comp_bytes: 0 354 comp_bytes: 0
355 decomp_calls: 0 355 decomp_calls: 0
356 decomp_bytes: 0 356 decomp_bytes: 0
357 wqs: 357 wqs:
358 name: iaa_crypto 358 name: iaa_crypto
359 comp_calls: 0 359 comp_calls: 0
360 comp_bytes: 0 360 comp_bytes: 0
361 decomp_calls: 0 361 decomp_calls: 0
362 decomp_bytes: 0 362 decomp_bytes: 0
363 363
364 iaa device: 364 iaa device:
365 id: 3 365 id: 3
366 n_wqs: 1 366 n_wqs: 1
367 comp_calls: 0 367 comp_calls: 0
368 comp_bytes: 0 368 comp_bytes: 0
369 decomp_calls: 0 369 decomp_calls: 0
370 decomp_bytes: 0 370 decomp_bytes: 0
371 wqs: 371 wqs:
372 name: iaa_crypto 372 name: iaa_crypto
373 comp_calls: 0 373 comp_calls: 0
374 comp_bytes: 0 374 comp_bytes: 0
375 decomp_calls: 0 375 decomp_calls: 0
376 decomp_bytes: 0 376 decomp_bytes: 0
377 377
378 iaa device: 378 iaa device:
379 id: 5 379 id: 5
380 n_wqs: 1 380 n_wqs: 1
381 comp_calls: 1360 381 comp_calls: 1360
382 comp_bytes: 1999776 382 comp_bytes: 1999776
383 decomp_calls: 0 383 decomp_calls: 0
384 decomp_bytes: 0 384 decomp_bytes: 0
385 wqs: 385 wqs:
386 name: iaa_crypto 386 name: iaa_crypto
387 comp_calls: 1360 387 comp_calls: 1360
388 comp_bytes: 1999776 388 comp_bytes: 1999776
389 decomp_calls: 0 389 decomp_calls: 0
390 decomp_bytes: 0 390 decomp_bytes: 0
391 391
392 iaa device: 392 iaa device:
393 id: 7 393 id: 7
394 n_wqs: 1 394 n_wqs: 1
395 comp_calls: 2940 395 comp_calls: 2940
396 comp_bytes: 3994213 396 comp_bytes: 3994213
397 decomp_calls: 4164 397 decomp_calls: 4164
398 decomp_bytes: 5993989 398 decomp_bytes: 5993989
399 wqs: 399 wqs:
400 name: iaa_crypto 400 name: iaa_crypto
401 comp_calls: 2940 401 comp_calls: 2940
402 comp_bytes: 3994213 402 comp_bytes: 3994213
403 decomp_calls: 4164 403 decomp_calls: 4164
404 decomp_bytes: 5993989 404 decomp_bytes: 5993989
405 ... 405 ...
406 406
407 Writing to 'stats_reset' resets all the stats, 407 Writing to 'stats_reset' resets all the stats, including the
408 per-device and per-wq stats:: 408 per-device and per-wq stats::
409 409
410 # echo 1 > stats_reset 410 # echo 1 > stats_reset
411 # cat wq_stats 411 # cat wq_stats
412 global stats: 412 global stats:
413 total_comp_calls: 0 413 total_comp_calls: 0
414 total_decomp_calls: 0 414 total_decomp_calls: 0
415 total_comp_bytes_out: 0 415 total_comp_bytes_out: 0
416 total_decomp_bytes_in: 0 416 total_decomp_bytes_in: 0
417 total_completion_einval_errors: 0 417 total_completion_einval_errors: 0
418 total_completion_timeout_errors: 0 418 total_completion_timeout_errors: 0
419 total_completion_comp_buf_overflow_errors: 419 total_completion_comp_buf_overflow_errors: 0
420 ... 420 ...
421 421
422 422
423 Use cases 423 Use cases
424 ========= 424 =========
425 425
426 Simple zswap test 426 Simple zswap test
427 ----------------- 427 -----------------
428 428
429 For this example, the kernel should be configu 429 For this example, the kernel should be configured according to the
430 dedicated mode options described above, and zs 430 dedicated mode options described above, and zswap should be enabled as
431 well:: 431 well::
432 432
433 CONFIG_ZSWAP=y 433 CONFIG_ZSWAP=y
434 434
435 This is a simple test that uses iaa_compress a 435 This is a simple test that uses iaa_compress as the compressor for a
436 swap (zswap) device. It sets up the zswap dev 436 swap (zswap) device. It sets up the zswap device and then uses the
437 memory_memadvise program listed below to forci 437 memory_memadvise program listed below to forcibly swap out and in a
438 specified number of pages, demonstrating both 438 specified number of pages, demonstrating both compress and decompress.
439 439
440 The zswap test expects the work queues for eac 440 The zswap test expects the work queues for each IAA device on the
441 system to be configured properly as a kernel w 441 system to be configured properly as a kernel workqueue with a
442 workqueue driver_name of "crypto". 442 workqueue driver_name of "crypto".
443 443
444 The first step is to make sure the iaa_crypto 444 The first step is to make sure the iaa_crypto module is loaded::
445 445
446 modprobe iaa_crypto 446 modprobe iaa_crypto
447 447
448 If the IAA devices and workqueues haven't prev 448 If the IAA devices and workqueues haven't previously been disabled and
449 reconfigured, then the default configuration s 449 reconfigured, then the default configuration should be in place and no
450 further IAA configuration is necessary. See : 450 further IAA configuration is necessary. See :ref:`iaa_default_config`
451 below for details of the default configuration 451 below for details of the default configuration.
452 452
453 If the default configuration is in place, you 453 If the default configuration is in place, you should see the iaa
454 devices and wq0s enabled:: 454 devices and wq0s enabled::
455 455
456 # cat /sys/bus/dsa/devices/iax1/state 456 # cat /sys/bus/dsa/devices/iax1/state
457 enabled 457 enabled
458 # cat /sys/bus/dsa/devices/iax1/wq1.0/state 458 # cat /sys/bus/dsa/devices/iax1/wq1.0/state
459 enabled 459 enabled
460 460
461 To demonstrate that the following steps work a 461 To demonstrate that the following steps work as expected, these
462 commands can be used to enable debug output:: 462 commands can be used to enable debug output::
463 463
464 # echo -n 'module iaa_crypto +p' > /sys/kern 464 # echo -n 'module iaa_crypto +p' > /sys/kernel/debug/dynamic_debug/control
465 # echo -n 'module idxd +p' > /sys/kernel/deb 465 # echo -n 'module idxd +p' > /sys/kernel/debug/dynamic_debug/control
466 466
467 Use the following commands to enable zswap:: 467 Use the following commands to enable zswap::
468 468
469 # echo 0 > /sys/module/zswap/parameters/enab 469 # echo 0 > /sys/module/zswap/parameters/enabled
470 # echo 50 > /sys/module/zswap/parameters/max 470 # echo 50 > /sys/module/zswap/parameters/max_pool_percent
471 # echo deflate-iaa > /sys/module/zswap/param 471 # echo deflate-iaa > /sys/module/zswap/parameters/compressor
472 # echo zsmalloc > /sys/module/zswap/paramete 472 # echo zsmalloc > /sys/module/zswap/parameters/zpool
473 # echo 1 > /sys/module/zswap/parameters/enab 473 # echo 1 > /sys/module/zswap/parameters/enabled
474 # echo 100 > /proc/sys/vm/swappiness 474 # echo 100 > /proc/sys/vm/swappiness
475 # echo never > /sys/kernel/mm/transparent_hu 475 # echo never > /sys/kernel/mm/transparent_hugepage/enabled
476 # echo 1 > /proc/sys/vm/overcommit_memory 476 # echo 1 > /proc/sys/vm/overcommit_memory
477 477
478 Now you can now run the zswap workload you wan 478 Now you can now run the zswap workload you want to measure. For
479 example, using the memory_memadvise code below 479 example, using the memory_memadvise code below, the following command
480 will swap in and out 100 pages:: 480 will swap in and out 100 pages::
481 481
482 ./memory_madvise 100 482 ./memory_madvise 100
483 483
484 Allocating 100 pages to swap in/out 484 Allocating 100 pages to swap in/out
485 Swapping out 100 pages 485 Swapping out 100 pages
486 Swapping in 100 pages 486 Swapping in 100 pages
487 Swapped out and in 100 pages 487 Swapped out and in 100 pages
488 488
489 You should see something like the following in 489 You should see something like the following in the dmesg output::
490 490
491 [ 404.202972] idxd 0000:e7:02.0: iaa_comp_a 491 [ 404.202972] idxd 0000:e7:02.0: iaa_comp_acompress: dma_map_sg, src_addr 223925c000, nr_sgs 1, req->src 00000000ee7cb5e6, req->slen 4096, sg_dma_len(sg) 4096
492 [ 404.202973] idxd 0000:e7:02.0: iaa_comp_a 492 [ 404.202973] idxd 0000:e7:02.0: iaa_comp_acompress: dma_map_sg, dst_addr 21dadf8000, nr_sgs 1, req->dst 000000008d6acea8, req->dlen 4096, sg_dma_len(sg) 8192
493 [ 404.202975] idxd 0000:e7:02.0: iaa_compre 493 [ 404.202975] idxd 0000:e7:02.0: iaa_compress: desc->src1_addr 223925c000, desc->src1_size 4096, desc->dst_addr 21dadf8000, desc->max_dst_size 4096, desc->src2_addr 2203543000, desc->src2_size 1568
494 [ 404.202981] idxd 0000:e7:02.0: iaa_compre 494 [ 404.202981] idxd 0000:e7:02.0: iaa_compress_verify: (verify) desc->src1_addr 21dadf8000, desc->src1_size 228, desc->dst_addr 223925c000, desc->max_dst_size 4096, desc->src2_addr 0, desc->src2_size 0
495 ... 495 ...
496 496
497 Now that basic functionality has been demonstr 497 Now that basic functionality has been demonstrated, the defaults can
498 be erased and replaced with a different config 498 be erased and replaced with a different configuration. To do that,
499 first disable zswap:: 499 first disable zswap::
500 500
501 # echo lzo > /sys/module/zswap/parameters/co 501 # echo lzo > /sys/module/zswap/parameters/compressor
502 # swapoff -a 502 # swapoff -a
503 # echo 0 > /sys/module/zswap/parameters/acce 503 # echo 0 > /sys/module/zswap/parameters/accept_threshold_percent
504 # echo 0 > /sys/module/zswap/parameters/max_ 504 # echo 0 > /sys/module/zswap/parameters/max_pool_percent
505 # echo 0 > /sys/module/zswap/parameters/enab 505 # echo 0 > /sys/module/zswap/parameters/enabled
506 # echo 0 > /sys/module/zswap/parameters/enab 506 # echo 0 > /sys/module/zswap/parameters/enabled
507 507
508 Then run the :ref:`iaa_disable_script` in the 508 Then run the :ref:`iaa_disable_script` in the 'Use Cases' section
509 below to disable the default configuration. 509 below to disable the default configuration.
510 510
511 Finally turn swap back on:: 511 Finally turn swap back on::
512 512
513 # swapon -a 513 # swapon -a
514 514
515 Following all that the IAA device(s) can now b 515 Following all that the IAA device(s) can now be re-configured and
516 enabled as desired for further testing. Below 516 enabled as desired for further testing. Below is one example.
517 517
518 The zswap test expects the work queues for eac 518 The zswap test expects the work queues for each IAA device on the
519 system to be configured properly as a kernel w 519 system to be configured properly as a kernel workqueue with a
520 workqueue driver_name of "crypto". 520 workqueue driver_name of "crypto".
521 521
522 The below script automatically does that:: 522 The below script automatically does that::
523 523
524 #!/bin/bash 524 #!/bin/bash
525 525
526 echo "IAA devices:" 526 echo "IAA devices:"
527 lspci -d:0cfe 527 lspci -d:0cfe
528 echo "# IAA devices:" 528 echo "# IAA devices:"
529 lspci -d:0cfe | wc -l 529 lspci -d:0cfe | wc -l
530 530
531 # 531 #
532 # count iaa instances 532 # count iaa instances
533 # 533 #
534 iaa_dev_id="0cfe" 534 iaa_dev_id="0cfe"
535 num_iaa=$(lspci -d:${iaa_dev_id} | wc -l) 535 num_iaa=$(lspci -d:${iaa_dev_id} | wc -l)
536 echo "Found ${num_iaa} IAA instances" 536 echo "Found ${num_iaa} IAA instances"
537 537
538 # 538 #
539 # disable iaa wqs and devices 539 # disable iaa wqs and devices
540 # 540 #
541 echo "Disable IAA" 541 echo "Disable IAA"
542 542
543 for ((i = 1; i < ${num_iaa} * 2; i += 2)); d 543 for ((i = 1; i < ${num_iaa} * 2; i += 2)); do
544 echo disable wq iax${i}/wq${i}.0 544 echo disable wq iax${i}/wq${i}.0
545 accel-config disable-wq iax${i}/wq${i}.0 545 accel-config disable-wq iax${i}/wq${i}.0
546 echo disable iaa iax${i} 546 echo disable iaa iax${i}
547 accel-config disable-device iax${i} 547 accel-config disable-device iax${i}
548 done 548 done
549 549
550 echo "End Disable IAA" 550 echo "End Disable IAA"
551 551
552 echo "Reload iaa_crypto module" 552 echo "Reload iaa_crypto module"
553 553
554 rmmod iaa_crypto 554 rmmod iaa_crypto
555 modprobe iaa_crypto 555 modprobe iaa_crypto
556 556
557 echo "End Reload iaa_crypto module" 557 echo "End Reload iaa_crypto module"
558 558
559 # 559 #
560 # configure iaa wqs and devices 560 # configure iaa wqs and devices
561 # 561 #
562 echo "Configure IAA" 562 echo "Configure IAA"
563 for ((i = 1; i < ${num_iaa} * 2; i += 2)); d 563 for ((i = 1; i < ${num_iaa} * 2; i += 2)); do
564 accel-config config-wq --group-id=0 --mo 564 accel-config config-wq --group-id=0 --mode=dedicated --wq-size=128 --priority=10 --type=kernel --name="iaa_crypto" --driver-name="crypto" iax${i}/wq${i}.0
565 accel-config config-engine iax${i}/engin 565 accel-config config-engine iax${i}/engine${i}.0 --group-id=0
566 done 566 done
567 567
568 echo "End Configure IAA" 568 echo "End Configure IAA"
569 569
570 # 570 #
571 # enable iaa wqs and devices 571 # enable iaa wqs and devices
572 # 572 #
573 echo "Enable IAA" 573 echo "Enable IAA"
574 574
575 for ((i = 1; i < ${num_iaa} * 2; i += 2)); d 575 for ((i = 1; i < ${num_iaa} * 2; i += 2)); do
576 echo enable iaa iax${i} 576 echo enable iaa iax${i}
577 accel-config enable-device iax${i} 577 accel-config enable-device iax${i}
578 echo enable wq iax${i}/wq${i}.0 578 echo enable wq iax${i}/wq${i}.0
579 accel-config enable-wq iax${i}/wq${i}.0 579 accel-config enable-wq iax${i}/wq${i}.0
580 done 580 done
581 581
582 echo "End Enable IAA" 582 echo "End Enable IAA"
583 583
584 When the workqueues are bound to the iaa_crypt 584 When the workqueues are bound to the iaa_crypto driver, you should
585 see something similar to the following in dmes 585 see something similar to the following in dmesg output if you've
586 enabled debug output (echo -n 'module iaa_cryp 586 enabled debug output (echo -n 'module iaa_crypto +p' >
587 /sys/kernel/debug/dynamic_debug/control):: 587 /sys/kernel/debug/dynamic_debug/control)::
588 588
589 [ 60.752344] idxd 0000:f6:02.0: add_iaa_wq 589 [ 60.752344] idxd 0000:f6:02.0: add_iaa_wq: added wq 000000004068d14d to iaa 00000000c9585ba2, n_wq 1
590 [ 60.752346] iaa_crypto: rebalance_wq_tabl 590 [ 60.752346] iaa_crypto: rebalance_wq_table: nr_nodes=2, nr_cpus 160, nr_iaa 8, cpus_per_iaa 20
591 [ 60.752347] iaa_crypto: rebalance_wq_tabl 591 [ 60.752347] iaa_crypto: rebalance_wq_table: iaa=0
592 [ 60.752349] idxd 0000:6a:02.0: request_ia 592 [ 60.752349] idxd 0000:6a:02.0: request_iaa_wq: getting wq from iaa_device 0000000042d7bc52 (0)
593 [ 60.752350] idxd 0000:6a:02.0: request_ia 593 [ 60.752350] idxd 0000:6a:02.0: request_iaa_wq: returning unused wq 00000000c8bb4452 (0) from iaa device 0000000042d7bc52 (0)
594 [ 60.752352] iaa_crypto: rebalance_wq_tabl 594 [ 60.752352] iaa_crypto: rebalance_wq_table: assigned wq for cpu=0, node=0 = wq 00000000c8bb4452
595 [ 60.752354] iaa_crypto: rebalance_wq_tabl 595 [ 60.752354] iaa_crypto: rebalance_wq_table: iaa=0
596 [ 60.752355] idxd 0000:6a:02.0: request_ia 596 [ 60.752355] idxd 0000:6a:02.0: request_iaa_wq: getting wq from iaa_device 0000000042d7bc52 (0)
597 [ 60.752356] idxd 0000:6a:02.0: request_ia 597 [ 60.752356] idxd 0000:6a:02.0: request_iaa_wq: returning unused wq 00000000c8bb4452 (0) from iaa device 0000000042d7bc52 (0)
598 [ 60.752358] iaa_crypto: rebalance_wq_tabl 598 [ 60.752358] iaa_crypto: rebalance_wq_table: assigned wq for cpu=1, node=0 = wq 00000000c8bb4452
599 [ 60.752359] iaa_crypto: rebalance_wq_tabl 599 [ 60.752359] iaa_crypto: rebalance_wq_table: iaa=0
600 [ 60.752360] idxd 0000:6a:02.0: request_ia 600 [ 60.752360] idxd 0000:6a:02.0: request_iaa_wq: getting wq from iaa_device 0000000042d7bc52 (0)
601 [ 60.752361] idxd 0000:6a:02.0: request_ia 601 [ 60.752361] idxd 0000:6a:02.0: request_iaa_wq: returning unused wq 00000000c8bb4452 (0) from iaa device 0000000042d7bc52 (0)
602 [ 60.752362] iaa_crypto: rebalance_wq_tabl 602 [ 60.752362] iaa_crypto: rebalance_wq_table: assigned wq for cpu=2, node=0 = wq 00000000c8bb4452
603 [ 60.752364] iaa_crypto: rebalance_wq_tabl 603 [ 60.752364] iaa_crypto: rebalance_wq_table: iaa=0
604 . 604 .
605 . 605 .
606 . 606 .
607 607
608 Once the workqueues and devices have been enab 608 Once the workqueues and devices have been enabled, the IAA crypto
609 algorithms are enabled and available. When th 609 algorithms are enabled and available. When the IAA crypto algorithms
610 have been successfully enabled, you should see 610 have been successfully enabled, you should see the following dmesg
611 output:: 611 output::
612 612
613 [ 64.893759] iaa_crypto: iaa_crypto_enable 613 [ 64.893759] iaa_crypto: iaa_crypto_enable: iaa_crypto now ENABLED
614 614
615 Now run the following zswap-specific setup com 615 Now run the following zswap-specific setup commands to have zswap use
616 the 'fixed' compression mode:: 616 the 'fixed' compression mode::
617 617
618 echo 0 > /sys/module/zswap/parameters/enable 618 echo 0 > /sys/module/zswap/parameters/enabled
619 echo 50 > /sys/module/zswap/parameters/max_p 619 echo 50 > /sys/module/zswap/parameters/max_pool_percent
620 echo deflate-iaa > /sys/module/zswap/paramet 620 echo deflate-iaa > /sys/module/zswap/parameters/compressor
621 echo zsmalloc > /sys/module/zswap/parameters 621 echo zsmalloc > /sys/module/zswap/parameters/zpool
622 echo 1 > /sys/module/zswap/parameters/enable 622 echo 1 > /sys/module/zswap/parameters/enabled
623 623
624 echo 100 > /proc/sys/vm/swappiness 624 echo 100 > /proc/sys/vm/swappiness
625 echo never > /sys/kernel/mm/transparent_huge 625 echo never > /sys/kernel/mm/transparent_hugepage/enabled
626 echo 1 > /proc/sys/vm/overcommit_memory 626 echo 1 > /proc/sys/vm/overcommit_memory
627 627
628 Finally, you can now run the zswap workload yo 628 Finally, you can now run the zswap workload you want to measure. For
629 example, using the code below, the following c 629 example, using the code below, the following command will swap in and
630 out 100 pages:: 630 out 100 pages::
631 631
632 ./memory_madvise 100 632 ./memory_madvise 100
633 633
634 Allocating 100 pages to swap in/out 634 Allocating 100 pages to swap in/out
635 Swapping out 100 pages 635 Swapping out 100 pages
636 Swapping in 100 pages 636 Swapping in 100 pages
637 Swapped out and in 100 pages 637 Swapped out and in 100 pages
638 638
639 You should see something like the following in 639 You should see something like the following in the dmesg output if
640 you've enabled debug output (echo -n 'module i 640 you've enabled debug output (echo -n 'module iaa_crypto +p' >
641 /sys/kernel/debug/dynamic_debug/control):: 641 /sys/kernel/debug/dynamic_debug/control)::
642 642
643 [ 404.202972] idxd 0000:e7:02.0: iaa_comp_a 643 [ 404.202972] idxd 0000:e7:02.0: iaa_comp_acompress: dma_map_sg, src_addr 223925c000, nr_sgs 1, req->src 00000000ee7cb5e6, req->slen 4096, sg_dma_len(sg) 4096
644 [ 404.202973] idxd 0000:e7:02.0: iaa_comp_a 644 [ 404.202973] idxd 0000:e7:02.0: iaa_comp_acompress: dma_map_sg, dst_addr 21dadf8000, nr_sgs 1, req->dst 000000008d6acea8, req->dlen 4096, sg_dma_len(sg) 8192
645 [ 404.202975] idxd 0000:e7:02.0: iaa_compre 645 [ 404.202975] idxd 0000:e7:02.0: iaa_compress: desc->src1_addr 223925c000, desc->src1_size 4096, desc->dst_addr 21dadf8000, desc->max_dst_size 4096, desc->src2_addr 2203543000, desc->src2_size 1568
646 [ 404.202981] idxd 0000:e7:02.0: iaa_compre 646 [ 404.202981] idxd 0000:e7:02.0: iaa_compress_verify: (verify) desc->src1_addr 21dadf8000, desc->src1_size 228, desc->dst_addr 223925c000, desc->max_dst_size 4096, desc->src2_addr 0, desc->src2_size 0
647 [ 409.203227] idxd 0000:e7:02.0: iaa_comp_a 647 [ 409.203227] idxd 0000:e7:02.0: iaa_comp_adecompress: dma_map_sg, src_addr 21ddd8b100, nr_sgs 1, req->src 0000000084adab64, req->slen 228, sg_dma_len(sg) 228
648 [ 409.203235] idxd 0000:e7:02.0: iaa_comp_a 648 [ 409.203235] idxd 0000:e7:02.0: iaa_comp_adecompress: dma_map_sg, dst_addr 21ee3dc000, nr_sgs 1, req->dst 000000004e2990d0, req->dlen 4096, sg_dma_len(sg) 4096
649 [ 409.203239] idxd 0000:e7:02.0: iaa_decomp 649 [ 409.203239] idxd 0000:e7:02.0: iaa_decompress: desc->src1_addr 21ddd8b100, desc->src1_size 228, desc->dst_addr 21ee3dc000, desc->max_dst_size 4096, desc->src2_addr 0, desc->src2_size 0
650 [ 409.203254] idxd 0000:e7:02.0: iaa_comp_a 650 [ 409.203254] idxd 0000:e7:02.0: iaa_comp_adecompress: dma_map_sg, src_addr 21ddd8b100, nr_sgs 1, req->src 0000000084adab64, req->slen 228, sg_dma_len(sg) 228
651 [ 409.203256] idxd 0000:e7:02.0: iaa_comp_a 651 [ 409.203256] idxd 0000:e7:02.0: iaa_comp_adecompress: dma_map_sg, dst_addr 21f1551000, nr_sgs 1, req->dst 000000004e2990d0, req->dlen 4096, sg_dma_len(sg) 4096
652 [ 409.203257] idxd 0000:e7:02.0: iaa_decomp 652 [ 409.203257] idxd 0000:e7:02.0: iaa_decompress: desc->src1_addr 21ddd8b100, desc->src1_size 228, desc->dst_addr 21f1551000, desc->max_dst_size 4096, desc->src2_addr 0, desc->src2_size 0
653 653
654 In order to unregister the IAA crypto algorith 654 In order to unregister the IAA crypto algorithms, and register new
655 ones using different parameters, any users of 655 ones using different parameters, any users of the current algorithm
656 should be stopped and the IAA workqueues and d 656 should be stopped and the IAA workqueues and devices disabled.
657 657
658 In the case of zswap, remove the IAA crypto al 658 In the case of zswap, remove the IAA crypto algorithm as the
659 compressor and turn off swap (to remove all re 659 compressor and turn off swap (to remove all references to
660 iaa_crypto):: 660 iaa_crypto)::
661 661
662 echo lzo > /sys/module/zswap/parameters/comp 662 echo lzo > /sys/module/zswap/parameters/compressor
663 swapoff -a 663 swapoff -a
664 664
665 echo 0 > /sys/module/zswap/parameters/accept 665 echo 0 > /sys/module/zswap/parameters/accept_threshold_percent
666 echo 0 > /sys/module/zswap/parameters/max_po 666 echo 0 > /sys/module/zswap/parameters/max_pool_percent
667 echo 0 > /sys/module/zswap/parameters/enable 667 echo 0 > /sys/module/zswap/parameters/enabled
668 668
669 Once zswap is disabled and no longer using iaa 669 Once zswap is disabled and no longer using iaa_crypto, the IAA wqs and
670 devices can be disabled. 670 devices can be disabled.
671 671
672 .. _iaa_disable_script: 672 .. _iaa_disable_script:
673 673
674 IAA disable script 674 IAA disable script
675 ------------------ 675 ------------------
676 676
677 The below script automatically does that:: 677 The below script automatically does that::
678 678
679 #!/bin/bash 679 #!/bin/bash
680 680
681 echo "IAA devices:" 681 echo "IAA devices:"
682 lspci -d:0cfe 682 lspci -d:0cfe
683 echo "# IAA devices:" 683 echo "# IAA devices:"
684 lspci -d:0cfe | wc -l 684 lspci -d:0cfe | wc -l
685 685
686 # 686 #
687 # count iaa instances 687 # count iaa instances
688 # 688 #
689 iaa_dev_id="0cfe" 689 iaa_dev_id="0cfe"
690 num_iaa=$(lspci -d:${iaa_dev_id} | wc -l) 690 num_iaa=$(lspci -d:${iaa_dev_id} | wc -l)
691 echo "Found ${num_iaa} IAA instances" 691 echo "Found ${num_iaa} IAA instances"
692 692
693 # 693 #
694 # disable iaa wqs and devices 694 # disable iaa wqs and devices
695 # 695 #
696 echo "Disable IAA" 696 echo "Disable IAA"
697 697
698 for ((i = 1; i < ${num_iaa} * 2; i += 2)); d 698 for ((i = 1; i < ${num_iaa} * 2; i += 2)); do
699 echo disable wq iax${i}/wq${i}.0 699 echo disable wq iax${i}/wq${i}.0
700 accel-config disable-wq iax${i}/wq${i}.0 700 accel-config disable-wq iax${i}/wq${i}.0
701 echo disable iaa iax${i} 701 echo disable iaa iax${i}
702 accel-config disable-device iax${i} 702 accel-config disable-device iax${i}
703 done 703 done
704 704
705 echo "End Disable IAA" 705 echo "End Disable IAA"
706 706
707 Finally, at this point the iaa_crypto module c 707 Finally, at this point the iaa_crypto module can be removed, which
708 will unregister the current IAA crypto algorit 708 will unregister the current IAA crypto algorithms::
709 709
710 rmmod iaa_crypto 710 rmmod iaa_crypto
711 711
712 712
713 memory_madvise.c (gcc -o memory_memadvise memo 713 memory_madvise.c (gcc -o memory_memadvise memory_madvise.c)::
714 714
715 #include <stdio.h> 715 #include <stdio.h>
716 #include <stdlib.h> 716 #include <stdlib.h>
717 #include <string.h> 717 #include <string.h>
718 #include <unistd.h> 718 #include <unistd.h>
719 #include <sys/mman.h> 719 #include <sys/mman.h>
720 #include <linux/mman.h> 720 #include <linux/mman.h>
721 721
722 #ifndef MADV_PAGEOUT 722 #ifndef MADV_PAGEOUT
723 #define MADV_PAGEOUT 21 /* force pag 723 #define MADV_PAGEOUT 21 /* force pages out immediately */
724 #endif 724 #endif
725 725
726 #define PG_SZ 4096 726 #define PG_SZ 4096
727 727
728 int main(int argc, char **argv) 728 int main(int argc, char **argv)
729 { 729 {
730 int i, nr_pages = 1; 730 int i, nr_pages = 1;
731 int64_t *dump_ptr; 731 int64_t *dump_ptr;
732 char *addr, *a; 732 char *addr, *a;
733 int loop = 1; 733 int loop = 1;
734 734
735 if (argc > 1) 735 if (argc > 1)
736 nr_pages = atoi(argv[1]); 736 nr_pages = atoi(argv[1]);
737 737
738 printf("Allocating %d pages to swap in 738 printf("Allocating %d pages to swap in/out\n", nr_pages);
739 739
740 /* allocate pages */ 740 /* allocate pages */
741 addr = mmap(NULL, nr_pages * PG_SZ, PR 741 addr = mmap(NULL, nr_pages * PG_SZ, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_ANONYMOUS, -1, 0);
742 *addr = 1; 742 *addr = 1;
743 743
744 /* initialize data in page to all '*' 744 /* initialize data in page to all '*' chars */
745 memset(addr, '*', nr_pages * PG_SZ); 745 memset(addr, '*', nr_pages * PG_SZ);
746 746
747 printf("Swapping out %d pages\n", nr_ 747 printf("Swapping out %d pages\n", nr_pages);
748 748
749 /* Tell kernel to swap it out */ 749 /* Tell kernel to swap it out */
750 madvise(addr, nr_pages * PG_SZ, MADV_P 750 madvise(addr, nr_pages * PG_SZ, MADV_PAGEOUT);
751 751
752 while (loop > 0) { 752 while (loop > 0) {
753 /* Wait for swap out to finish 753 /* Wait for swap out to finish */
754 sleep(5); 754 sleep(5);
755 755
756 a = addr; 756 a = addr;
757 757
758 printf("Swapping in %d pages\n 758 printf("Swapping in %d pages\n", nr_pages);
759 759
760 /* Access the page ... this wi 760 /* Access the page ... this will swap it back in again */
761 for (i = 0; i < nr_pages; i++) 761 for (i = 0; i < nr_pages; i++) {
762 if (a[0] != '*') { 762 if (a[0] != '*') {
763 printf("Bad da 763 printf("Bad data from decompress!!!!!\n");
764 764
765 dump_ptr = (in 765 dump_ptr = (int64_t *)a;
766 for (int j = 766 for (int j = 0; j < 100; j++) {
767 printf 767 printf(" page %d data: %#llx\n", i, *dump_ptr);
768 dump_p 768 dump_ptr++;
769 } 769 }
770 } 770 }
771 771
772 a += PG_SZ; 772 a += PG_SZ;
773 } 773 }
774 774
775 loop --; 775 loop --;
776 } 776 }
777 777
778 printf("Swapped out and in %d pages\n", 778 printf("Swapped out and in %d pages\n", nr_pages);
779 779
780 Appendix 780 Appendix
781 ======== 781 ========
782 782
783 .. _iaa_sysfs_config: 783 .. _iaa_sysfs_config:
784 784
785 IAA sysfs config interface 785 IAA sysfs config interface
786 -------------------------- 786 --------------------------
787 787
788 Below is a description of the IAA sysfs interf 788 Below is a description of the IAA sysfs interface, which as mentioned
789 in the main document, should only be used if y 789 in the main document, should only be used if you know exactly what you
790 are doing. Even then, there's no compelling r 790 are doing. Even then, there's no compelling reason to use it directly
791 since accel-config can do everything the sysfs 791 since accel-config can do everything the sysfs interface can and in
792 fact accel-config is based on it under the cov 792 fact accel-config is based on it under the covers.
793 793
794 The 'IAA config path' is /sys/bus/dsa/devices 794 The 'IAA config path' is /sys/bus/dsa/devices and contains
795 subdirectories representing each IAA device, w 795 subdirectories representing each IAA device, workqueue, engine, and
796 group. Note that in the sysfs interface, the 796 group. Note that in the sysfs interface, the IAA devices are actually
797 named using iax e.g. iax1, iax3, etc. (Note th 797 named using iax e.g. iax1, iax3, etc. (Note that IAA devices are the
798 odd-numbered devices; the even-numbered device 798 odd-numbered devices; the even-numbered devices are DSA devices and
799 can be ignored for IAA). 799 can be ignored for IAA).
800 800
801 The 'IAA device bind path' is /sys/bus/dsa/dri 801 The 'IAA device bind path' is /sys/bus/dsa/drivers/idxd/bind and is
802 the file that is written to enable an IAA devi 802 the file that is written to enable an IAA device.
803 803
804 The 'IAA workqueue bind path' is /sys/bus/dsa/ 804 The 'IAA workqueue bind path' is /sys/bus/dsa/drivers/crypto/bind and
805 is the file that is written to enable an IAA w 805 is the file that is written to enable an IAA workqueue.
806 806
807 Similarly /sys/bus/dsa/drivers/idxd/unbind and 807 Similarly /sys/bus/dsa/drivers/idxd/unbind and
808 /sys/bus/dsa/drivers/crypto/unbind are used to 808 /sys/bus/dsa/drivers/crypto/unbind are used to disable IAA devices and
809 workqueues. 809 workqueues.
810 810
811 The basic sequence of commands needed to set u 811 The basic sequence of commands needed to set up the IAA devices and
812 workqueues is: 812 workqueues is:
813 813
814 For each device:: 814 For each device::
815 1) Disable any workqueues enabled on the dev 815 1) Disable any workqueues enabled on the device. For example to
816 disable workques 0 and 1 on IAA device 3: 816 disable workques 0 and 1 on IAA device 3::
817 817
818 # echo wq3.0 > /sys/bus/dsa/drivers/cry 818 # echo wq3.0 > /sys/bus/dsa/drivers/crypto/unbind
819 # echo wq3.1 > /sys/bus/dsa/drivers/cry 819 # echo wq3.1 > /sys/bus/dsa/drivers/crypto/unbind
820 820
821 2) Disable the device. For example to disabl 821 2) Disable the device. For example to disable IAA device 3::
822 822
823 # echo iax3 > /sys/bus/dsa/drivers/idxd 823 # echo iax3 > /sys/bus/dsa/drivers/idxd/unbind
824 824
825 3) configure the desired workqueues. For ex 825 3) configure the desired workqueues. For example, to configure
826 workqueue 3 on IAA device 3:: 826 workqueue 3 on IAA device 3::
827 827
828 # echo dedicated > /sys/bus/dsa/devices 828 # echo dedicated > /sys/bus/dsa/devices/iax3/wq3.3/mode
829 # echo 128 > /sys/bus/dsa/devices/iax3/ 829 # echo 128 > /sys/bus/dsa/devices/iax3/wq3.3/size
830 # echo 0 > /sys/bus/dsa/devices/iax3/wq 830 # echo 0 > /sys/bus/dsa/devices/iax3/wq3.3/group_id
831 # echo 10 > /sys/bus/dsa/devices/iax3/w 831 # echo 10 > /sys/bus/dsa/devices/iax3/wq3.3/priority
832 # echo "kernel" > /sys/bus/dsa/devices/ 832 # echo "kernel" > /sys/bus/dsa/devices/iax3/wq3.3/type
833 # echo "iaa_crypto" > /sys/bus/dsa/devi 833 # echo "iaa_crypto" > /sys/bus/dsa/devices/iax3/wq3.3/name
834 # echo "crypto" > /sys/bus/dsa/devices/ 834 # echo "crypto" > /sys/bus/dsa/devices/iax3/wq3.3/driver_name
835 835
836 4) Enable the device. For example to enable 836 4) Enable the device. For example to enable IAA device 3::
837 837
838 # echo iax3 > /sys/bus/dsa/drivers/idxd 838 # echo iax3 > /sys/bus/dsa/drivers/idxd/bind
839 839
840 5) Enable the desired workqueues on the devi 840 5) Enable the desired workqueues on the device. For example to
841 enable workques 0 and 1 on IAA device 3:: 841 enable workques 0 and 1 on IAA device 3::
842 842
843 # echo wq3.0 > /sys/bus/dsa/drivers/cry 843 # echo wq3.0 > /sys/bus/dsa/drivers/crypto/bind
844 # echo wq3.1 > /sys/bus/dsa/drivers/cry 844 # echo wq3.1 > /sys/bus/dsa/drivers/crypto/bind
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.