Root/target/linux/generic-2.6/files/crypto/ocf/ChangeLog

12008-09-18 01:27 davidm
2
3    * Makefile, README, README.sglinux,
4    patches/linux-2.6.26-natt.patch, patches/linux-2.6.26-ocf.patch,
5    patches/openssl-0.9.8g.patch, patches/openssl-0.9.8i.patch:
6    
7    Updates for a new OCF release with openssl-0.9.8i and linux-2.6.26
8    support.
9
102008-09-18 00:19 davidm
11
12    * Config.in, Kconfig, Makefile, ep80579/Makefile,
13    ep80579/environment.mk, ep80579/icp_asym.c, ep80579/icp_common.c,
14    ep80579/icp_ocf.h, ep80579/icp_sym.c,
15    ep80579/linux_2.6_kernel_space.mk:
16    
17    A new driver from Intel for their Intel QuickAssist enabled EP80579
18    Integrated Processor Product Line.
19    
20    Adrian Hoban Brad Vrabete
21    
22
232008-07-25 01:01 gerg
24
25    * ocf-compat.h:
26    
27    From linux-2.6.26 onwards there is now a linux/fdtable.h include
28    that contains the file_fdtable() definition.
29
302008-07-05 01:20 davidm
31
32    * Makefile, patches/linux-2.6.25-natt.patch,
33    patches/linux-2.6.25-ocf.patch:
34    
35    A new ocf-linux release and some patches to send to the OS guys.
36
372008-07-03 21:21 davidm
38
39    * crypto.c:
40    
41    Clean up a some possible deadlock/busy wait issues with
42    locking/sleeping. This has greatly improved openswan 2.6.14
43    reliability ;-)
44    
45    Make more of our state available in /sys for debugging.
46
472008-06-23 20:38 davidm
48
49    * Config.in, Kconfig:
50    
51    Some settings should be bool's John Gumb
52
532008-05-28 03:43 philipc
54
55    * ixp4xx/ixp4xx.c: Revert previous checkin since the
56    callbacks are meant to complete the crypto request. Before
57    reverting this, CONFIG_DEBUG_SLAB gives: slab error in
58    verify_redzone_free(): cache `ixp4xx_q': double free detected when
59    trying to ssh to the device.
60
612008-04-27 19:31 davidm
62
63    * ixp4xx/ixp4xx.c:
64    
65    We were not completing a crypto request under some error
66    conditions.
67
682008-04-02 01:51 davidm
69
70    * talitos/talitos.c:
71    
72    This patch is for crypto/ocf/talitos for use on linux 2.6.23. It
73    is applied to the ocf-linux-20071215 release.
74    
75    Signed-off-by: Lee Nipper
76
772008-02-29 00:43 davidm
78
79    * crypto.c, ocf-compat.h, ixp4xx/ixp4xx.c:
80    
81    freshen up the 2.4 support, some recent OCF changes and openswan
82    changes are not that old-os friendly.
83    
84    Force OCF to select HW/SW, otherwise it may get stuck on the first
85    device. This change means we will favour HW over SW, but we will
86    use both as required. Passing in a crid of 0 to crypto_newsession
87    effectively meant we were stuck on the first device registered with
88    OCF, not good. This only applied to ipsec, cryptodev already did
89    the right thing.
90
912008-01-31 07:37 gerg
92
93    * hifn/hifn7751.c:
94    
95    The linux-2.6.24 modules build fails if the pci ID table doesn't
96    have a NULL entry at the end. So add one.
97
982008-01-29 09:16 gerg
99
100    * cryptosoft.c, ocf-compat.h:
101    
102    Added some compatability macros for scatterlist changes from 2.6.24
103    onwards.
104
1052007-12-16 07:31 davidm
106
107    * Makefile:
108    
109    missed an openssl patch name change
110
1112007-12-16 07:27 davidm
112
113    * Makefile, README, README.sglinux, patches/linux-2.6.23-ocf.patch,
114    patches/openssl-0.9.8e.patch, patches/openssl-0.9.8g.patch,
115    patches/ssl.patch:
116    
117    updates for a new ocf release and associated bits
118
1192007-12-16 06:36 davidm
120
121    * crypto.c:
122    
123    Be very careful what you do while potentially in an driver unload
124    state or we will call through NULL pointers.
125    
126    Reported by Nawang Chhetan .
127
1282007-12-14 22:32 davidm
129
130    * cryptodev.c:
131    
132    Add in an unlock_ioctl when available to help SMP systems a lot.
133    Otherwise all ioctls get a BKL :-(
134    
135    Problem found by Egor N. Martovetsky
136
1372007-12-14 18:29 davidm
138
139    * cryptosoft.c:
140    
141    reformat the alg table to make it easier to read.
142
1432007-12-14 18:29 davidm
144
145    * crypto.c:
146    
147    Fix more driver locking/sleeping bugs report by Nawang Chhetan
148    
149
1502007-12-12 21:36 davidm
151
152    * Config.in, Kconfig, Makefile, pasemi/Makefile, pasemi/pasemi.c,
153    pasemi/pasemi_fnu.h:
154    
155    Here I'm including my PA Semi driver patch to OCF. Please consider
156    it for inclusion into next OCF release.
157    
158    Egor N. Martovetsky
159
1602007-12-05 00:37 davidm
161
162    * patches/: linux-2.4.35-ocf.patch, linux-2.6.22-ocf.patch:
163    
164    More correct count setting if we get a signal
165    
166    Adrian Hoban
167
1682007-12-05 00:02 davidm
169
170    * random.c:
171    
172    OCF has a static array for holding random data. The random number
173    generator I have can write directly into physically contiguous
174    memory. Static memory comes from the heap and isn't physically
175    contiguous. I could use kmalloc'd memory and then copy into the OCF
176    static buf but I'd like to avoid a memory copy. The following patch
177    (Physically_Contig_Rand.patch) allows me to avoid a memory copy and
178    should not impact the other OCF drivers:
179    
180    Adrian Hoban
181
1822007-12-05 00:01 davidm
183
184    * Kconfig:
185    
186    Fix a typo in the Kconfig
187
1882007-11-23 19:15 davidm
189
190    * talitos/talitos_dev.h:
191    
192    fix the DPRINTF macro so that it actually compiles.
193
1942007-11-22 19:41 davidm
195
196    * ocf-compat.h, talitos/talitos.c, talitos/talitos_dev.h:
197    
198    Various updates to get talitos compiling and work on real-world (ie
199    ubuntu) kernels.
200
2012007-11-08 02:21 davidm
202
203    * crypto.c:
204    
205    keep and eye on us being completely blocked. If we have Q's to
206    process, but all the requests are blocked, sleep. We do not want
207    to busy loop until a driver unblocks as it uses valuable CPU
208    resources that could be doing something much more important ;-)
209
2102007-11-07 19:04 davidm
211
212    * hifn/hifn7751.c:
213    
214    hifn driver was failing to unblock itself under some "out of
215    resources" conditions. It would return ERESTART to signal it was
216    full but never call crypto_unblock to start things moving again.
217
2182007-11-06 02:09 davidm
219
220    * hifn/hifn7751.c:
221    
222    Remove some bogus trace left in the driver for the overflow (too
223    busy) case.
224
2252007-10-12 21:10 gerg
226
227    * crypto.c, ixp4xx/ixp4xx.c:
228    
229    Fix up use of kmem_cache_create() - it takes one less argument in
230    2.6.23 onwards.
231
2322007-10-03 02:41 gerg
233
234    * ixp4xx/Makefile:
235    
236    The directory locations for includes in CSR-2.4 is different. Need
237    to modify the CFLAGS accordingly if using CSR-2.4.
238
2392007-09-22 00:39 philipc
240
241    * ixp4xx/Makefile: linux 2.4 make dep was failing. This is
242    a quick fix to get it building, need to double check this.
243
2442007-09-19 00:13 mmccreat
245
246    * Config.in: Add config option CONFIG_OCF_IXP4XX_SHA1_MD5,
247    that enables SHA1 and MD5 hashing to be done by the IXP4xx crypto
248    accelerator (although it is much slower than using cryptosoft).
249
2502007-09-18 21:45 mmccreat
251
252    * Makefile, random.c: - Force the inclusion of autoconf.h,
253    which contains #defines for CONFIG_xxx options for OCF. -
254    Removing additional -D option, now that we are including the
255    CONFIG_xxx #defines.
256
2572007-09-18 21:44 mmccreat
258
259    * Kconfig: Add config option CONFIG_OCF_IXP4XX_SHA1_MD5,
260    that enables SHA1 and MD5 hashing to be done by the IXP4xx crypto
261    accelerator (although it is much slower than using cryptosoft).
262
2632007-09-18 21:37 mmccreat
264
265    * cryptodev.h: The CRYPTO_MAX_DATA_LEN limit should be
266    0xFFFF ie 64K - 1.
267
2682007-09-18 21:19 mmccreat
269
270    * ixp4xx/ixp4xx.c: - Rework the code so that the correct IXP
271    function, ixCryptoAccHashPerform(), is used to calculate SHA1 and
272    MD5 hashes. NB: The performance of using the IXP4xx hardware is
273    really, really poor compared to using cryptosoft (and the kernel
274    crypto). - Only support SHA1 and MD5 hashing if the
275    CONFIG_OCF_IXP4XX_SHA1_MD5 is enabled.
276
2772007-08-30 21:42 davidm
278
279    * Makefile:
280    
281    do not archive build files in the crypto-tools archive
282
2832007-08-22 19:19 mmccreat
284
285    * cryptodev.c: Fix up the checking for key lengths, when the
286    key can be of unlimited size.
287
2882007-08-16 01:50 davidm
289
290    * Makefile:
291    
292    Better 2.4 compat for "make dep" now working with fastdep.
293
2942007-07-28 08:25 davidm
295
296    * Makefile, README, README.sglinux, patches/crypto-tools.patch,
297    patches/linux-2.4.29-ocf.patch, patches/linux-2.4.35-ocf.patch,
298    patches/linux-2.6.11-ocf.patch, patches/linux-2.6.22-ocf.patch,
299    patches/ssl.patch:
300    
301    Update all the patches and put the patch making target back into
302    the Makefile.
303
3042007-07-28 08:25 davidm
305
306    * hifn/hifn7751.c:
307    
308    fix an unused variable warning when HARVESTING is disabled
309
3102007-07-27 21:33 davidm
311
312    * hifn/hifn7751.c, ixp4xx/ixp4xx.c, safe/safe.c, talitos/talitos.c:
313    
314    
315    Remove all the random code if OCF does not have radom harvesting
316    enabled.
317
3182007-07-26 00:36 davidm
319
320    * Kconfig, hifn/hifnHIPP.c, hifn/hifnHIPPvar.h:
321    
322    Changes to get the hifn HIPP stub driver to build.
323
3242007-07-25 21:25 davidm
325
326    * Makefile, hifn/Makefile, hifn/hifnHIPP.c, hifn/hifnHIPPreg.h,
327    hifn/hifnHIPPvar.h, ixp4xx/Makefile, ocfnull/Makefile,
328    safe/Makefile, talitos/Makefile:
329    
330    Bring in the hifnHIPP driver written by Xelerance. This is the
331    super hifn chip with full protocol offload.
332    
333    Switch to much more traditional Makefile/subdir building. The
334    Makefiles are nicer now, but still not beautiful, 2.6 and 2.4
335    capable builds result in a certain amount of uglyiness.
336
3372007-07-24 21:46 davidm
338
339    * cryptodev.c:
340    
341    Clean up all the driver id checking and session management so
342    adding/removing drivers all continues to run cleanly.
343
3442007-07-24 20:14 davidm
345
346    * talitos/talitos.c:
347    
348    From: Ahsan Kabir
349    
350    less than .1% packet corruption was detected using the talitos
351    driver. It turns out we don't need the cipher iv out len/ptr field
352    to do ESP IPsec. Therefore we set the len field as 0, which tells
353    the SEC not to do anything with this len/ptr field.
354    
355    Signed-off-by: Ahsan Kabir
356    Signed-off-by: Kim Phillips
357
3582007-07-24 08:25 davidm
359
360    * cryptosoft.c, ocf-bench.c, hifn/hifn7751.c, ixp4xx/ixp4xx.c,
361    safe/safe.c, talitos/talitos.c:
362    
363    Switch the remaining GFP_ATOMIC to the newer and not deprecated
364    SLAB_ATOMIC
365
3662007-07-23 22:16 mmccreat
367
368    * ixp4xx/ixp4xx.c: Add missing ";" from end of dprintk()
369    call!
370
3712007-07-21 01:16 davidm
372
373    * ocf-compat.h, rndtest.c, hifn/hifn7751.c, safe/safe.c:
374    
375    pci_register_driver is nothing like it is depending on the kernel,
376    so we need a compat function to fix it up for all kernels before
377    2.6.10, and in different ways for older and not so older versions.
378
3792007-07-20 21:54 davidm
380
381    * safe/safe.c:
382    
383    Make the debug macro ';' safe so you do not get compiler warnings
384
3852007-07-20 21:53 davidm
386
387    * talitos/: talitos.c, talitos_dev.h, talitos_soft.h:
388    
389    update to the latest FreeBSD driver structure and fix up the code
390    as required.
391
3922007-07-20 03:07 davidm
393
394    * rndtest.c:
395    
396    more headers needed to compile on 2.4
397
3982007-07-20 03:00 davidm
399
400    * cryptosoft.c:
401    
402    Put in the 2.4 stubs to support compression
403
4042007-07-20 02:53 davidm
405
406    * crypto.c, ocf-compat.h:
407    
408    move some more compat stuff into the compat header.
409
4102007-07-20 02:47 davidm
411
412    * talitos/talitos.c:
413    
414    support of_platform_driver for newer, ARCH=powerpc based kernels.
415    Signed-off-by: Kim Phillips
416
4172007-07-20 02:46 davidm
418
419    * talitos/talitos.c:
420    
421    From: Ahsan Kabir
422    
423    When Talitos completes job both the channel and execution unit are
424    capable of generating interrupts. Talitos used to take two
425    interrupts per request - one for channel completion and the other
426    for execution unit completion. This patch ensures that Talitos
427    takes interrupt only for channel completion. Execution unit will
428    generate interrupt only when there is error and the error
429    interrupts for execution units are not masked.
430    
431    Signed-off-by: Ahsan Kabir
432    Signed-off-by: Kim Phillips
433
4342007-07-20 02:37 davidm
435
436    * cryptodev.c:
437    
438    Unless the user specifies, select from both HW and SW.
439    
440    Clean up some debug to report the actual ioctl name.
441    
442    Compiler warning on newer compilers.
443
4442007-07-20 02:35 davidm
445
446    * cryptodev.h:
447    
448    moved dprintk to the compat code, seemed nicer in there.
449
4502007-07-20 02:35 davidm
451
452    * rndtest.c:
453    
454    Need to clean up some warnings etc, more includes
455
4562007-07-20 02:34 davidm
457
458    * ocf-compat.h, hifn/hifn7751.c, safe/safe.c:
459    
460    new shared IRQ flags for 2.6.22 and a safer version of the debug
461    macro
462
4632007-07-20 00:52 davidm
464
465    * cryptosoft.c:
466    
467    Implement compression based on the code from the openswan guys.
468
4692007-07-20 00:52 davidm
470
471    * criov.c:
472    
473    Fix compiler warning on non-value returning void func.
474
4752007-07-18 22:55 davidm
476
477    * hifn/hifn7751.c, safe/safe.c:
478    
479    Use pci_register_driver rather than pci_module_init.
480    pci_module_init has been dropped in 2.6.22 yet pci_register_driver
481    has always existed and used to do some crazy hotplug junk.
482
4832007-07-18 21:55 gerg
484
485    * ixp4xx/ixp4xx.c:
486    
487    Added a missing ";" at the end of the ixp_kproces() prototype. It
488    is IXP465 specific, only showed up when generateing for SG720.
489
4902007-07-17 00:37 davidm
491
492    * Makefile, cryptodev.c, random.c, rndtest.c, hifn/hifn7751.c,
493    ocfnull/ocfnull.c, safe/safe.c:
494    
495    Fixup all the debug support for 2.4 kernels, clean up the entropy
496    harvester to be far more robust.
497
4982007-07-14 02:19 davidm
499
500    * talitos/talitos.c:
501    
502    Old patch that had not been applied Kim Phillips
503    
504
5052007-07-14 01:12 davidm
506
507    * Config.in, Kconfig, Makefile, cryptodev.c, cryptodev.h, random.c,
508    syscall.h:
509    
510    Finally ditched all the syscall stuff. You can now enable/disable
511    the random harvestor. Pulled in most of random.c from openswan
512    project and fixed some obvious bugs (that were always there).
513
5142007-07-13 21:59 davidm
515
516    * ocf-compat.h:
517    
518    Better error printing and checking for drivers
519
5202007-07-13 21:56 davidm
521
522    * cryptosoft.c:
523    
524    Fix some incorrect debug (reporting wrong error type)
525
5262007-07-13 21:55 davidm
527
528    * hifn/hifn7751.c, safe/safe.c:
529    
530    Make the code more similar to Free-BSD by reverting to the same
531    debug macros
532
5332007-07-13 21:53 davidm
534
535    * ocfnull/ocfnull.c:
536    
537    Fix up the null driver to work again in the new framework.
538
5392007-07-06 23:54 mmccreat
540
541    * cryptodev.c, cryptodev.h, cryptosoft.c: - Update OCF to
542    work with new Crypto API introduced in 2.6.19 kerneli, and add
543    macros so it work with older kernels. - Add support for SHA256,
544    SHA384 and SHA512 HASH and HMAC algorithms. - Cryptosoft: Only
545    register algorithms that the kernel has implementations for.
546
5472007-07-03 19:52 davidm
548
549    * Kconfig, README, README.sglinux, criov.c, crypto.c, cryptodev.c,
550    cryptodev.h, cryptosoft.c, ocf-bench.c, ocf-compat.h, random.c,
551    rndtest.c, uio.h, hifn/hifn7751.c, hifn/hifn7751reg.h,
552    hifn/hifn7751var.h, ixp4xx/ixp4xx.c, ocfnull/ocfnull.c,
553    safe/safe.c, safe/safevar.h, talitos/talitos.c:
554    
555    Updated OCF to the lastest FreeBSD version.
556    
557    There was a lot of change in here, some of which will help FIP's,
558    some which won't.
559    
560    Did lots of cleaning and diff reduction against the freebsd code.
561    Still more cleaning to do.
562
5632007-06-01 21:58 gerg
564
565    * Config.in:
566    
567    Put the regular old Config.in back, needed for puclic releases.
568
5692007-06-01 21:58 gerg
570
571    * Kconfig:
572    
573    Change the CONFIG_OCF_IXP400 dependencies to be the same as they
574    where in the Config.in file.
575
5762007-05-28 21:40 gerg
577
578    * Config.in, Kconfig:
579    
580    Switch all module configuration over to new style Kconfigs.
581
5822007-05-24 18:49 davidm
583
584    * cryptodev.c, random.c:
585    
586    Work around some problems on redhat systems with errno redefinition
587
5882007-04-30 21:09 gerg
589
590    * cryptosoft.c:
591    
592    The CRYPTO_TFM_MODE_ family of defines no longer exists from 2.6.21
593    onwards. As far as I can tell you don't need to pass it to the
594    crypto_alloc_tfm() function anymore.
595    
596    So define it to be 0 if it doesn't exist.
597
5982007-04-03 02:13 gerg
599
600    * syscall.h:
601    
602    Added syscall macros for SH architecture. Just temporary, 'till
603    Dave fixes the OCF code to not use syscalls from the modules :-)
604
6052007-02-16 23:10 davidm
606
607    * syscall.h:
608    
609    ensure the temprary 2.6 fix doesn't break 2.4
610
6112007-02-07 22:23 gerg
612
613    * cryptodev.c, random.c, syscall.h:
614    
615    Temporary fix for new 2.6 kernels no longer defining in-kernel
616    system call functions. Define them locally for now until we fix
617    properly.
618
6192007-02-07 03:10 gerg
620
621    * ixp4xx/ixp4xx.c:
622    
623    Changes to support the different INIT_WORK() mechanism from kernels
624    2.6.20 onwards.
625
6262007-02-06 02:38 gerg
627
628    * crypto.c:
629    
630    Cleaned up use of kmem_cache_t and use of SLAB_ATOMIC.
631
6322006-12-05 20:50 cpascoe
633
634    * hifn/hifn7751.c, safe/safe.c, talitos/talitos.c:
635    
636    Remove pt_regs from OCF interrupt handlers for 2.6.19+
637
6382006-12-02 03:36 gerg
639
640    * criov.c, crypto.c, cryptodev.c, cryptosoft.c, ocf-bench.c,
641    random.c, rndtest.c, hifn/hifn7751.c, ixp4xx/ixp4xx.c,
642    ocfnull/ocfnull.c, safe/safe.c, talitos/talitos.c:
643    
644    Can no longer include linux/config.h as of 2.6.19 kernels. Need to
645    conditionally include it based on AUTOCONF_INCLUDED not being
646    defined.
647
6482006-10-13 21:52 cpascoe
649
650    * random.c:
651    
652    Remove another race condition that may result in us running more
653    than one random thread if modules are reloaded during heavy system
654    load.
655
6562006-10-13 21:18 cpascoe
657
658    * random.c:
659    
660    - Permit delivery of SIGKILL to the random thread. - Do not exit
661    prematurely if poll() is interrupted. - Improve exit conditions so
662    that we quit immediately, rather than loop infinitely, if the
663    last RNG source is removed while we are trying to fill our
664    buffer.
665
6662006-10-03 20:28 cpascoe
667
668    * crypto.c, crypto.c: AutoMerged
669    >
670    > Zero the empty half of "new driver" buffer, and not past the end
671    of the old
672    > buffer that we are about to free.
673
6742006-10-03 20:28 cpascoe
675
676    * crypto.c:
677    
678    Zero the empty half of "new driver" buffer, and not past the end of
679    the old buffer that we are about to free.
680
6812006-08-25 23:57 davidm
682
683    * cryptosoft.c:
684    
685    Do not print errors for failed ALG requests unless debugging
686
6872006-07-14 21:44 davidm
688
689    * cryptodev.h:
690    
691    2.6.11 and earlier did not have the files_fdtable macro.
692
6932006-06-21 21:26 gerg
694
695    * cryptodev.h, hifn/hifn7751.c, ocfnull/ocfnull.c, safe/safe.c,
696    talitos/talitos.c:
697    
698    Fixed up more occurrances of MODULE_PARM() needing to be converted
699    to module_param() (as of 2.6.17 and onwards).
700
7012006-06-21 00:28 gerg
702
703    * cryptosoft.c:
704    
705    Change use of MODULE_PARM to module_param() for 2.6.17+ kernels.
706
7072006-06-20 22:13 gerg
708
709    * crypto.c, cryptodev.c, cryptodev.h, ocf-bench.c, ixp4xx/ixp4xx.c:
710    
711    
712    As of 2.6.17 and onwards module_param is used in place of
713    MODULE_PARM.
714
7152006-06-06 00:31 gerg
716
717    * Makefile:
718    
719    Fix Makefile to find includes is using CSR-2.1.
720
7212006-05-31 01:44 gerg
722
723    * cryptodev.h:
724    
725    The vars "crypto_usercrypto", "crypto_userasymcrypto" are declared
726    as extern in the header file, but static in the c file. I guessed
727    that they should probably be truely static, so removed the extern
728    declarations from the header.
729
7302006-05-25 21:06 davidm
731
732    * talitos/talitos.c:
733    
734    This fixes a situation that I never provably experienced, where a
735    descriptor in memory may be reserved within the proper lock, and
736    freed immediately after, only for a few cycles, right outside the
737    lock. Kim Phillips
738
7392006-05-15 19:49 davidm
740
741    * criov.c, crypto.c, cryptodev.c, cryptodev.h, cryptosoft.c,
742    ocf-bench.c, random.c, rndtest.c, uio.h, ixp4xx/ixp4xx.c:
743    
744    Remove the "all rights reserved" from the Intel copyrights.
745
7462006-05-12 21:19 davidm
747
748    * Config.in, Kconfig, Makefile, cryptodev.c, ocf-bench.c,
749    ocfnull/ocfnull.c:
750    
751    Add in a null OCF driver that does nothing at all, useful for
752    measuring the cost of various parts of the ipsec stack.
753
7542006-05-12 21:17 davidm
755
756    * ixp4xx/ixp4xx.c:
757    
758    make sure we do not overwrite a correctly set error type.
759
7602006-05-12 06:52 davidm
761
762    * crypto.c:
763    
764    Fix a problem where a driver would return ERESTART (full) but then
765    unblock itself before the upper layer had marked it as blocked.
766    This caused the code to get stuck in crypto_proc and process no
767    more requests.
768
7692006-05-12 06:47 davidm
770
771    * cryptosoft.c:
772    
773    Implement CRD_F_KEY_EXPLICIT for cryptosoft so keys can be changed
774    on an active session.
775
7762006-05-10 20:09 davidm
777
778    * README, criov.c, crypto.c, cryptodev.h, cryptosoft.c,
779    ocf-bench.c, random.c, rndtest.c, uio.h, hifn/hifn7751.c,
780    ixp4xx/ixp4xx.c, safe/safe.c, safe/safevar.h:
781    
782    update email addresses and other house cleaning
783
7842006-05-10 20:08 davidm
785
786    * cryptodev.c:
787    
788    pull in better error checking from openswan modifications
789
7902006-05-10 19:11 davidm
791
792    * cryptosoft.c:
793    
794    Fix an unused variable warning when various options are disabled.
795
7962006-05-10 19:10 davidm
797
798    * cryptodev.h:
799    
800    Add support for 2.4 kernels for the new FD cloning operation
801
8022006-05-09 19:48 davidm
803
804    * hifn/hifn7751.c:
805    
806    remove the hifn 7855 support, this driver will never work with
807    that chip.
808
8092006-05-08 23:34 davidm
810
811    * hifn/hifn7751var.h:
812    
813    Contiguous buffer support so that ocf-bench can run properly.
814
8152006-05-05 23:21 davidm
816
817    * hifn/hifn7751.c:
818    
819    Add in contiguous buffer support so that ocf-bench and run on the
820    driver.
821
8222006-05-05 23:14 davidm
823
824    * ocf-bench.c:
825    
826    Our requests were out of order, need to do crypto then auth on
827    encrypt requests. Some drivers enforce this.
828
8292006-05-04 23:21 davidm
830
831    * crypto.c:
832    
833    Do not run "newsession" with lock, since newsession may sleep on
834    some targets. Handle the accounting so that things are not pulled
835    from underneath us.
836
8372006-05-04 23:20 davidm
838
839    * cryptodev.c:
840    
841    Switch to a less optimal (marginally) solution for creating a new
842    fd that appears to work in far more versions of the kernel
843    including 64bit versions.
844
8452006-05-04 18:54 davidm
846
847    * ocf-bench.c:
848    
849    Turn off the IXP access lib benchmarking by default as most people
850    don't have it.
851    
852    Paul Wouters
853
8542006-04-01 08:23 davidm
855
856    * Makefile:
857    
858    Remove more temp files when cleaning
859
8602006-04-01 08:12 davidm
861
862    * hifn/hifn7751reg.h:
863    
864    7855 PCI id's as yet untested
865
8662006-04-01 08:08 davidm
867
868    * hifn/hifn7751.c:
869    
870    add PCI id's for the 7855 and AES support, card is untested still
871    as it requires 128MB of PCI memory !
872
8732006-03-31 08:38 davidm
874
875    * README.sglinux:
876    
877    small update to instructions with corrected patch name
878
8792006-03-31 00:23 davidm
880
881    * Config.in:
882    
883    Add the Talitos driver to the 2.4 config, even though it probably
884    won't compile.
885
8862006-03-30 07:48 davidm
887
888    * Kconfig, Makefile, talitos/talitos.c, talitos/talitos_dev.h,
889    talitos/talitos_soft.h:
890    
891    Please find attached the freescale SEC driver for OCF. It's been
892    (most recently) tested on an SEC2.0 based MPC8541E
893    (cryptographically identical to the MPC8555E) under 2.6.15.2, with
894    openssl-0.9.8a and openswan-2.4.3 (2.4.5rc5 won't keep the security
895    association up for me for some reason).
896    
897    Please feel free to add it to your next release of OCF-Linux :-)
898    
899    Kim Phillips
900
9012006-03-20 19:34 davidm
902
903    * safe/: safe.c, safevar.h:
904    
905    Safenet 1141 v1.0 chips have a DMA lockup error if you access the
906    chip while DMA'ing. As a work around you need to limit your DMA to
907    256 byte chunks using scatter/gather descriptors. All the SG/SME
908    products have v1.0 chips and would lockup with more than two
909    outstanding packets at one time.
910    
911    Fix the KASSERT macro
912    
913    Add some more exhaustive initialisation.
914
9152006-03-15 21:58 davidm
916
917    * cryptodev.h, random.c:
918    
919    Switch random support to "ints" since that is what the kernel uses
920    and not using the same thing is 64bit wise a bad idea.
921    
922    Fix FIP's code to ensure correct amount of data is passed in.
923    
924    Add work around for broken 64bit OS RNG support (disable it)
925    
926    General code cleanups.
927
9282006-03-15 21:55 davidm
929
930    * hifn/hifn7751.c:
931    
932    Fixes for 64bit OS's, make sure PCI address are within bus space,
933    make sure we order writes to the bus so that chip functions
934    correctly. Some small cleanups.
935
9362006-03-15 21:48 davidm
937
938    * hifn/hifn7751var.h:
939    
940    Remove unused field from structure
941
9422006-03-15 21:47 davidm
943
944    * safe/safe.c:
945    
946    Make the code more 64bit OS compatible, force PCI address space
947    and so on.
948
9492006-03-09 20:42 davidm
950
951    * ixp4xx/ixp4xx.c:
952    
953    If we call ixpCryptoAccInit() and it fails, just assume that it
954    has already been called. This allows our "rc" scripts to be
955    openswan and freeswan compatible ore easily.
956
9572006-02-28 23:21 davidm
958
959    * README:
960    
961    generalise it a bit so rel-dates don't get in the way
962
9632006-02-28 01:52 davidm
964
965    * README, patches/ssh.patch, patches/ssl.patch:
966    
967    Updated the README and patches for a release
968
9692006-02-25 09:21 davidm
970
971    * README, README.sglinux:
972    
973    updates for a new release of OCF, ssl patches and so on.
974
9752006-02-25 08:44 davidm
976
977    * crypto.c:
978    
979    We were calling the "process"routines with interrupts disabled. I
980    can see no good reason for this and it provokes badness warnings in
981    2.6 under some conditions.
982    
983    I am going to run with the Q's unlocked for processing, and
984    hopefully it will allow the system to be more responsive. It
985    hasn't affected ipsec throughput in any way. Userland throughput
986    (multi threaded) may have improved significantly though, but it
987    needs more testing.
988
9892006-02-24 23:32 davidm
990
991    * cryptodev.c:
992    
993    Whoa, set the segments with uninitted values can't be good. Clean
994    out the rest of the old code that was accidently left in.
995
9962006-02-22 01:02 davidm
997
998    * cryptodev.c:
999    
1000    Still not sure about this one, but this is working for all the
1001    cases I can see so far. If it gets weird, I am going to switch to
1002    a simple clone and chain the sessions rather than a new fcr per fd.
1003
10042006-02-20 22:12 davidm
1005
1006    * cryptodev.c:
1007    
1008    Error handling case could free data that was not allocated
1009    
1010    Ronen Shitrit Marvell Semiconductor Israel
1011    Ltd
1012
10132006-02-20 21:57 davidm
1014
1015    * cryptosoft.c:
1016    
1017    Proper SHA/MD5 (non hmac) implementation, remove some retrictions
1018    for hashes (which only applied to crypto) and fix over zealous
1019    error checking.
1020    
1021    Ronen Shitrit Marvell Semiconductor Israel
1022    Ltd
1023
10242006-02-09 21:15 davidm
1025
1026    * cryptodev.c:
1027    
1028    The code wasn't quite right and needed some fixing for proper file
1029    accounting.
1030
10312006-01-24 20:08 davidm
1032
1033    * crypto.c, cryptodev.c, hifn/hifn7751.c:
1034    
1035    We implement our own open in the CRIOGET ioctl now which does
1036    pretty much the same thing as the BSD version, rather than use an
1037    open system call which gets caught out by chroot.
1038
10392006-01-06 00:42 gerg
1040
1041    * cryptosoft.c:
1042    
1043    Moved "{" block marker inside the set of
1044    "defined(CONFIG_CRYPTO_HMAC) || defined(CONFIG_CRYPTO_HMAC_MODULE)"
1045    code (in function swcr_process) so that it compiled if these where
1046    not defined.
1047
10482005-11-11 01:44 davidm
1049
1050    * Makefile, README:
1051    
1052    cleanups for a general release of OCF
1053
10542005-11-11 01:18 davidm
1055
1056    * patches/ssl.patch:
1057    
1058    Updated the ssl patch to the latest and greatest changed we have
1059    made.
1060
10612005-11-10 21:41 davidm
1062
1063    * Makefile, ixp4xx/ixp4xx.c:
1064    
1065    Better debug for bad input.
1066    
1067    Have make clean do more cleaning and less talking.
1068
10692005-11-03 20:53 davidm
1070
1071    * cryptosoft.c:
1072    
1073    clean up some compilation errors with various options on/off
1074
10752005-10-25 00:25 davidm
1076
1077    * Makefile:
1078    
1079    Add a patch target that generates full kernel patches to add OCF
1080    into either a 2.4 or 2.6 kernel as a single patch.
1081
10822005-10-25 00:24 davidm
1083
1084    * Kconfig:
1085    
1086    Make sure all OCF settings depend on OCF_OCF
1087
10882005-09-23 02:45 davidm
1089
1090    * README, README.sglinux:
1091    
1092    new crypto-tools archive to keep the tools up to date
1093
10942005-09-23 02:08 davidm
1095
1096    * Makefile, README, README.sglinux:
1097    
1098    updates for doing OCF releases
1099
11002005-09-23 01:59 davidm
1101
1102    * patches/: ssh.patch, ssl.patch:
1103    
1104    updated the patches for the latest in fixes etc to ssh/ssl for a
1105    new OCF release before the openswan 2.4.1 merge.
1106
11072005-09-21 00:57 davidm
1108
1109    * Makefile, cryptosoft.c, hifn/hifn7751.c, ixp4xx/ixp4xx.c:
1110    
1111    Fixes for building cleanly under 2.6
1112
11132005-09-13 23:11 davidm
1114
1115    * ocf-bench.c:
1116    
1117    Add an exit function for cleaner 2.6 support. Patch from Ronen
1118    Shitrit
1119
11202005-07-30 00:23 davidm
1121
1122    * cryptosoft.c:
1123    
1124    Add proper hooks for MD5/SHA and their HMAC countrparts processing
1125    from cryptodev.
1126
11272005-07-29 01:50 davidm
1128
1129    * cryptodev.c:
1130    
1131    cryptodev did not support MD5 and SHA1, only the HMAC versions
1132
11332005-07-29 01:05 davidm
1134
1135    * ixp4xx/ixp4xx.c:
1136    
1137    register for MD5, return 16 bytes for MD5 and 12 for MD5_HMAC,
1138    likewise for SHA1 (only 20 is not HMAC).
1139
11402005-07-28 21:52 davidm
1141
1142    * ixp4xx/ixp4xx.c:
1143    
1144    First pass fix of userland MD5 processing. We now do as well as
1145    the safenet does, I think.
1146
11472005-07-27 06:23 davidm
1148
1149    * cryptodev.c:
1150    
1151    udelay doesn't give up the current thread, thus the kernel will get
1152    locked if a process is killed but the hardware never completes the
1153    crypto request.
1154
11552005-07-22 02:07 davidm
1156
1157    * crypto.c, cryptodev.h:
1158    
1159    Implement queuing limits for input/output and OCF requests.
1160    
1161    Implement 2.6 style work queues instead of the 2.4 task_queues.
1162
11632005-07-21 20:42 davidm
1164
1165    * cryptodev.h, ocf-bench.c, ixp4xx/ixp4xx.c:
1166    
1167    Fix OCF to use work queues, add 2.4 version of work queues to
1168    cryptodev.h for use in ipsec and elsewhere.
1169    
1170    Problem and initial patch provided by David Vrabel
1171    . Cleaned up and 2.4 support added by me.
1172
11732005-07-21 19:08 davidm
1174
1175    * random.c:
1176    
1177    Fix spinlock initialisation, problem reported by Andy @
1178    windsorcarclub.co.uk.
1179
11802005-07-20 20:24 davidm
1181
1182    * cryptodev.c:
1183    
1184    fix a silly spelling mistake
1185
11862005-07-08 00:56 gerg
1187
1188    * Makefile:
1189    
1190    Only build ocf-bench when CONFIG_OCF_BENCH is acrually enabled.
1191
11922005-06-27 20:29 davidm
1193
1194    * ixp4xx/ixp4xx.c:
1195    
1196    Ok, we need a simple implentation here or we go too slow for UDP
1197    tests. For now, if the Q is full just ditch the packet, someone
1198    has to do it.
1199
12002005-06-25 01:13 davidm
1201
1202    * safe/safe.c:
1203    
1204    bytes swapping etc all seems wrong for safenet on BE, these fixes
1205    see both MD5 and SHA1 working with OpenSwan.
1206
12072005-06-22 23:10 davidm
1208
1209    * random.c:
1210    
1211    clean up some ifdef code a little
1212
12132005-06-22 21:28 davidm
1214
1215    * ixp4xx/ixp4xx.c:
1216    
1217    Make sure we do not call blocking functions from ISR callable
1218    routines. In this case we were calling ixCryptoAccCtxUnregister.
1219    
1220    Run all the random numbers through SHA1 process to ensure more
1221    uniform distribution of bits (NOTE: it is not more random in any
1222    way)
1223
12242005-06-21 00:11 davidm
1225
1226    * ixp4xx/ixp4xx.c:
1227    
1228    do not process requests from the register callback, gets the AES
1229    code all messed up.
1230    
1231    Align caches on HW cache boundaries ot improve speed.
1232    
1233    More tracking or potential errors.
1234
12352005-06-15 01:55 davidm
1236
1237    * ixp4xx/ixp4xx.c:
1238    
1239    technically ixCryptoAccCtxRegister cannot be called from IRQ
1240    context, so run it from the immediate BH.
1241
12422005-06-14 23:13 davidm
1243
1244    * ocf-bench.c:
1245    
1246    Fix some compile warnings/errors
1247
12482005-06-14 20:52 davidm
1249
1250    * Config.in, Kconfig, Makefile, ocf-bench.c:
1251    
1252    Add in kernel benchmark driver
1253    
1254    Support for building under CSR 1.4 and 2.0 is now complete
1255
12562005-06-14 20:51 davidm
1257
1258    * crypto.c:
1259    
1260    Do not use immediate processing for SW drivers otherwise we hold a
1261    lock for too long. Instead force BATCH processing.
1262    
1263    Problem found by David Vrabel
1264
12652005-06-14 20:46 davidm
1266
1267    * cryptodev.c:
1268    
1269    Fix up AES minimum key size
1270    
1271    Make some more variables static
1272
12732005-06-14 20:36 davidm
1274
1275    * ixp4xx/ixp4xx.c:
1276    
1277    Use kernel caches as they are more efficient and faster to obtain.
1278    
1279    Fix some spelling
1280    
1281    Tune PKE to only use the space required. Turn off go fast options
1282    to reduce any speed-related cracking.
1283    
1284    Only zero data that needs to be zeroed (save some cycles)
1285
12862005-06-02 21:42 davidm
1287
1288    * ixp4xx/ixp4xx.c:
1289    
1290    Added optimisations ideas from Intel the improve the PKE
1291    performance for 512 and 1024 bits operations.
1292
12932005-06-01 02:13 davidm
1294
1295    * ixp4xx/ixp4xx.c:
1296    
1297    Looks like I found the ixp bug. Using OSAL buffer routines on
1298    non-OSAL buffers is a very very bad thing to do. Must double check
1299    all the API's I am using (ie., PKE) just to be sure.
1300
13012005-05-31 21:38 davidm
1302
1303    * Config.in, Kconfig:
1304    
1305    Updated/Added the menu wrapper for the config options
1306
13072005-05-31 21:18 gerg
1308
1309    * Config.in:
1310    
1311    Reworked the config.in so that each sub-module has its own
1312    Config.in. That way it is easier to make release trees with some
1313    modules left in.
1314
13152005-05-30 19:46 davidm
1316
1317    * ixp4xx/ixp4xx.c:
1318    
1319    Not all the Pke code was appropriately ifdef'd
1320
13212005-05-28 01:49 davidm
1322
1323    * cryptosoft.c:
1324    
1325    We were not injecting data in the right places nor calling the
1326    cipher code the best was under all situations. We are now good
1327    enought to do ESP/AH processing with 3DES and SHA1-HMAC.
1328
13292005-05-28 01:42 davidm
1330
1331    * hifn/hifn7751.c:
1332    
1333    Non atomic kmallocs at IRQ time are bad and cause lockups
1334
13352005-05-21 08:31 davidm
1336
1337    * README, README.sglinux:
1338    
1339    Some small updates to email and patches that no longer exist
1340
13412005-05-21 08:25 davidm
1342
1343    * cryptosoft.c:
1344    
1345    better error message on failure, a lot of the error check we do
1346    sems to break apps like cryptotest, not sure who is right yet.
1347
13482005-05-21 00:55 davidm
1349
1350    * criov.c, crypto.c, cryptodev.c, cryptodev.h, cryptosoft.c,
1351    random.c, rndtest.c, uio.h, hifn/hifn7751.c, ixp4xx/ixp4xx.c,
1352    safe/safe.c, safe/safevar.h:
1353    
1354    Convert to CyberGuard email addresses for OCF files.
1355
13562005-05-21 00:28 davidm
1357
1358    * crypto.c, cryptodev.c, cryptosoft.c, hifn/hifn7751.c,
1359    ixp4xx/ixp4xx.c, safe/safe.c:
1360    
1361    final round of naming changes, hifn driver also confirmed working
1362    in SG710
1363
13642005-05-20 23:50 davidm
1365
1366    * cryptosoft.c, random.c, uio.h, ixp4xx/ixp4xx.c:
1367    
1368    more copyright/author updates etc for Intel
1369
13702005-05-20 23:48 davidm
1371
1372    * criov.c, crypto.c, cryptodev.c, cryptodev.h, cryptosoft.c,
1373    random.c, rndtest.c, uio.h, ixp4xx/ixp4xx.c, safe/safe.c,
1374    safe/safevar.h:
1375    
1376    updated copyrights to reflect Intels interest/investment in the
1377    project
1378
13792005-05-20 20:31 davidm
1380
1381    * hifn/: hifn7751.c, hifn7751var.h:
1382    
1383    Most of a working SKB implementation now
1384
13852005-05-20 20:30 davidm
1386
1387    * safe/: md5.c, md5.h, safe.c, sha1.c, sha1.h:
1388    
1389    Full AH/ESP processing now working, just added BSD sha/md5
1390    routines as needed for now.
1391
13922005-05-20 20:30 davidm
1393
1394    * random.c:
1395    
1396    use the new BSD rndtest code rather than the old GPL fips code
1397
13982005-05-20 20:28 davidm
1399
1400    * Makefile, fips.c, fips.h, rndtest.c, rndtest.h:
1401    
1402    Get rid of the GPL version of the fips test and use a BSD licensed
1403    version so no one can get upset and cry derivative :-)
1404
14052005-05-20 08:19 davidm
1406
1407    * ixp4xx/ixp4xx.c:
1408    
1409    High throughput improvements. Can now handle multiple simultaneous
1410    requests.
1411
14122005-05-20 00:55 davidm
1413
1414    * ixp4xx/ixp4xx.c:
1415    
1416    works with openswan/OCF for ipsec receive.
1417    
1418    fixes ESP/AUTH processing (esp. HMAC)
1419    
1420    This driver needs a rework for kernel IPSEC as it's current packet
1421    at a time processing is not up to the task.
1422
14232005-05-20 00:53 davidm
1424
1425    * cryptosoft.c:
1426    
1427    working ESP/AUTH code now impleneted. Can do RX processing for
1428    openswan KLIPS.
1429    
1430    Fixes numerous problems in skb processing
1431    
1432    Fixes broken HMAC code and IV processing
1433
14342005-05-14 01:44 davidm
1435
1436    * cryptodev.c, ixp4xx/ixp4xx.c:
1437    
1438    Cleanup IXP key processing to guarantee only a single outstanding
1439    request rather than relying on the intel driver to get it right.
1440    Stops us losing requests.
1441    
1442    Tighten up the cryptodev response to "no answer" so that we don't
1443    take all the CPU. Only happens if there is a driver bug.
1444
14452005-05-14 00:07 davidm
1446
1447    * Config.in, Kconfig, Makefile, crypto.c, fips.c, fips.h, random.c,
1448    ixp4xx/ixp4xx.c, tools/bench-ocf:
1449    
1450    PKE and RNG support running on the ixp driver, added PKE bench
1451    marking to script. Still some multi-thread problems in the PKE
1452    code.
1453    
1454    Added FIP RNG checking option to config and code.
1455
14562005-05-10 19:18 davidm
1457
1458    * Makefile, ixp4xx/ixp4xx.c:
1459    
1460    Get the OCF stuff building for the 465 and CSR-2.0
1461
14622005-04-27 19:18 davidm
1463
1464    * cryptodev.h, random.c, hifn/hifn7751.c, safe/safe.c:
1465    cleanup the random number interface some more
1466
14672005-04-27 00:57 davidm
1468
1469    * cryptodev.h, cryptosoft.c, random.c, hifn/hifn7751.c,
1470    ixp4xx/ixp4xx.c, patches/linux-2.4.29-add_true_randomness.patch,
1471    patches/linux-2.6.11-add_true_randomness.patch, safe/safe.c,
1472    safe/safevar.h:
1473    
1474    Switch to a more "user" like random number handling. Drivers no
1475    longer poll for RNG data, we pull it as needed to fill
1476    /dev/random's entropy.
1477    
1478    Implement sk_buff handling within the OCF framework.
1479    
1480    fixup IV handling in cryptosoft.
1481
14822005-04-27 00:41 davidm
1483
1484    * crypto.c:
1485    
1486    Fix a race condition with the starting of kernel threads. The
1487    threads were running before the pid assignment in the parent.
1488
14892005-03-24 23:57 davidm
1490
1491    * Makefile:
1492    include crypto-tools.patch in the release file
1493    
1494    
1495
14962005-03-24 20:14 davidm
1497
1498    * safe/safe.c:
1499    Fixup compile time warnings due to some left over BSDisms
1500    
1501
15022005-03-24 00:53 davidm
1503
1504    * Makefile, README, README.sglinux, patches/linux-2.4.29-ocf.patch,
1505    patches/linux-2.6.11-ocf.patch, patches/ocf-linux-2.4.29.patch,
1506    patches/ocf-linux-2.6.11.patch:
1507    added cleaner patch names and a tarball target to aid releases
1508    
1509
15102005-03-24 00:28 davidm
1511
1512    * patches/crypto-tools.patch,
1513    patches/linux-2.4.29-add_true_randomness.patch,
1514    patches/linux-2.6.11-add_true_randomness.patch,
1515    patches/ocf-linux-2.4.29.patch, patches/ocf-linux-2.6.11.patch,
1516    patches/ssh.patch, patches/ssl.patch, tools/bench-ocf:
1517    move all the release file patches into CVS for simplicity
1518    
1519
15202005-03-23 20:37 davidm
1521
1522    * safe/safe.c:
1523    remove excessive debug from RNG routines so that you can turn on
1524    debug and live through it
1525    
1526    
1527
15282005-03-23 02:23 davidm
1529
1530    * safe/safe.c:
1531    fix memory corruption for mod_exp and the safenet, we were copying
1532    back more than the space available.
1533    
1534    
1535
15362005-03-22 21:45 davidm
1537
1538    * crypto.c, cryptodev.c, cryptodev.h, safe/safe.c:
1539    fixup a major sync issues with key processing (callback called
1540    before sleeping). Improve its performance while we are there with
1541    a CBIMM (callback immediate) option.
1542    
1543    
1544
15452005-03-19 00:33 davidm
1546
1547    * random.c:
1548    A new randomness function for both 2.4 and 2.6 that replaces out
1549    previous old solution for the hifn driver with more generic code
1550    that works on both kernels.
1551    
1552        add_true_randomness(u_int32_t *buf, int nwords);
1553    
1554    
1555
15562005-03-18 21:01 davidm
1557
1558    * Makefile, cryptodev.h, random.c, hifn/hifn7751.c, safe/safe.c:
1559    RNG support in both the safenet and the hifn plus the required
1560    kernel support.
1561    
1562    "hd /dev/random" runs much much faster now :-)
1563    
1564    
1565
15662005-03-17 23:29 toby
1567
1568    * cryptodev.c: Make sure the CIOCASYMFEAT ioctl on
1569    /dev/crypto copies out the capable features.
1570
15712005-03-17 01:19 davidm
1572
1573    * safe/: safe.c, safereg.h, safevar.h:
1574    hardware PK acceleration on the safenet (CRK_MOD_EXP only)
1575    
1576
15772005-03-16 04:28 davidm
1578
1579    * criov.c, crypto.c, cryptodev.c, safe/safe.c:
1580    fixup the FBSD id stuff to compile :-)
1581
15822005-03-16 04:02 davidm
1583
1584    * README, README.sglinux, TODO:
1585    Updated with versions, removed tabs, new kernel versions, web
1586    site etc
1587
15882005-03-16 03:45 davidm
1589
1590    * criov.c, crypto.c, cryptodev.c, cryptodev.h:
1591    more license formatting and version id's to help keep in touch with
1592    FreeBSD
1593
15942005-03-16 03:16 davidm
1595
1596    * safe/safe.c:
1597    up to date with FreeBSD versioning, no changes to merge
1598
15992005-03-16 03:15 davidm
1600
1601    * safe/safe.c:
1602    include FreeBSD version info so I can track changes better
1603    
1604
16052005-03-16 03:11 davidm
1606
1607    * hifn/: hifn7751.c, hifn7751reg.h, hifn7751var.h:
1608    incorporate latest updates from FreeBSD:
1609    
1610    | Update support for 795x parts: | o rework pll setup code to
1611    follow h/w specification | o add hint.hifn.X.pllconfig to specify
1612    reference clock setup | requirements; default is pci66 which
1613    means the clock is | derived from the PCI bus clock and the card
1614    resides in a | 66MHz slot | | Tested on 7955 and 7956 cards;
1615    support for 7954 cards not enabled | since we have no cards to test
1616    against.
1617    
1618    
1619
16202005-03-16 02:50 davidm
1621
1622    * Config.in, Kconfig, README, crypto.c, cryptodev.c, cryptodev.h,
1623    hifn/hifn7751.c, hifn/hifn7751var.h, safe/safe.c, safe/safevar.h:
1624    cleanups to build and run on 2.6.11 and 2.4.29 for a public release
1625    
1626    included configs locally to reduce patch to kernel and required
1627    user effort.
1628    
1629    pci_dma_sync_single fix from Michele Baldessari
1630    , with modifications by me to work on all
1631    kernels.
1632    
1633    
1634
16352005-03-11 23:30 davidm
1636
1637    * Makefile:
1638    fix up the compiling again, I had outstanding commits for this one
1639    :-)
1640
16412005-03-11 21:56 danield
1642
1643    * Makefile: Get ocf modules building again.
1644
16452004-12-25 07:12 davidm
1646
1647    * TODO:
1648    safenet is all good now AKAIK
1649
16502004-12-25 07:02 davidm
1651
1652    * Makefile, crypto.c, cryptodev.c, cryptosoft.c, uio.h:
1653    2.6 port of the OCF framework, tested on Xscale and compiled for
1654    x86
1655    
1656    
1657
16582004-12-10 02:41 davidm
1659
1660    * hifn/hifn7751.c:
1661    Fix compilation as l_flags was not defined for LOCK in pci_remove
1662    
1663
16642004-12-06 19:15 davidm
1665
1666    * hifn/hifn7751.c, safe/safe.c:
1667    * Put locking into hifn_remove to ensure interrupts are not running
1668    while we remove the driver
1669    
1670      Use del_timer_sync (need to ensure timer is not running on
1671    another CPU
1672    * when we delete it).
1673    
1674    Improvements suggested by Evgeniy Polyakov <johnpol@2ka.mipt.ru>
1675    
1676
16772004-12-02 09:16 davidm
1678
1679    * README, README.sglinux:
1680    * Fix up all tabs to be spaces * explain how to update the ocf
1681    support in the patch to the current version.
1682    
1683    
1684
16852004-12-02 09:11 davidm
1686
1687    * README:
1688    * fix some spelling/grammar * add more info on the ssl.patch file
1689    and what it contains
1690
16912004-12-02 09:08 davidm
1692
1693    * README.sglinux:
1694    give some more instructions on the shar archive and extracting the
1695    release.
1696
16972004-12-02 09:03 davidm
1698
1699    * README.sglinux:
1700    how to include crypto-tools into 3.1.6
1701    
1702    
1703
17042004-12-02 08:48 davidm
1705
1706    * README.sglinux:
1707    updated for a new crypto patch for openssl-0.9.7e
1708
17092004-12-02 08:04 davidm
1710
1711    * README, README.sglinux:
1712    Clean up the README's to use the new SG Linux alpha and a specific
1713    2.4.28 patch.
1714    
1715    
1716
17172004-12-01 23:11 davidm
1718
1719    * Makefile, criov.c, crypto.c, cryptodev.c, cryptosoft.c,
1720    hifn/Makefile, hifn/hifn7751.c, ixp4xx/Makefile, ixp4xx/ixp4xx.c,
1721    safe/Makefile, safe/safe.c:
1722    Re-worked the Makefiles so that including the code into a standard
1723    non-SG kernel is really easy.
1724    
1725    Fixed a non-initialised bug that was introduce into cryptosoft.c
1726    with the additional error checking.
1727    
1728    
1729
17302004-12-01 01:54 davidm
1731
1732    * TODO:
1733    updates based on things that have been fixed.
1734
17352004-12-01 01:53 davidm
1736
1737    * safe/: safe.c, safevar.h:
1738    Cleanup some old hacks based on the much cleaner port of the hifn
1739    driver.
1740
17412004-12-01 01:53 davidm
1742
1743    * cryptosoft.c:
1744    fix some crashes due to bad buffer sizes etc.
1745
17462004-11-27 09:16 davidm
1747
1748    * crypto.c, cryptodev.c, ixp4xx/ixp4xx.c:
1749    Fixed some nasty problems that were stopping reliable behaviour:
1750    
1751    1) we were not initialising some of our lists/wait queues, which
1752     meant they appears to have things in them when in fact they did
1753    not. This actually didn't seems to cause any problems but is
1754    extremely bogus.
1755    
1756    2) While a process was waiting for the crypto engine, if
1757    itreceived a signal we would lose sync with the engine and end
1758    up allowing out-of-band actions that were invalid (ie.,
1759    unregistering a context that was still active in a crypto
1760    operation).
1761    
1762       Of source the CryptACC should probably deal with this as well
1763    ;-)
1764    
1765    
1766
17672004-11-26 01:33 davidm
1768
1769    * TODO:
1770    We should also hook in the random number generators to linux
1771    
1772
17732004-11-26 01:33 davidm
1774
1775    * hifn/hifn7751.c:
1776    Changed all the accesses to DMS descriptors to not set the valid
1777    bit until after everything else was set. This got the driver
1778    running smoothly, along with a fixup to the pci_map_uio which
1779    wasn't settings lengths correctly.
1780    
1781    
1782
17832004-11-25 21:15 davidm
1784
1785    * ixp4xx/ixp4xx.c:
1786    document why using a new context for each packet is actually a
1787    better idea than allocating one (actually two) per session.
1788    
1789    
1790
17912004-11-25 08:48 davidm
1792
1793    * README, README.sglinux:
1794    updates from email with Intel to fix a few little things
1795    
1796
17972004-11-25 00:02 davidm
1798
1799    * README, README.sglinux, safe/safe.c:
1800    Some cleanups of doc and so on for Intel/General use
1801
18022004-11-23 07:58 davidm
1803
1804    * TODO:
1805    safenet is working on big endian machines now
1806    
1807    
1808
18092004-11-23 07:56 davidm
1810
1811    * ixp4xx/: Makefile, ixp4xx.c:
1812    Reference the Intel library, cleanup the IV sizes and turn on AES.
1813     Fix Makefile for new config options.
1814    
1815    
1816
18172004-11-23 07:52 davidm
1818
1819    * Makefile:
1820    Complete the changes of ixp to ixp4xx (IXP4xx) as appropriate
1821    
1822
18232004-11-23 00:33 davidm
1824
1825    * Makefile, README, ixp4xx/Makefile:
1826    Changes all references to ixp to be ixp4xx at Intels request
1827    
1828
18292004-11-20 01:07 davidm
1830
1831    * safe/: safe.c, safereg.h:
1832    fully working safenet on BE machines
1833
18342004-11-19 01:03 davidm
1835
1836    * Makefile, hifn/Makefile, hifn/hifn7751.c, hifn/hifn7751reg.h,
1837    hifn/hifn7751var.h:
1838    hifn driver code complete and compiling, needs a test now ;-)
1839    
1840
18412004-11-18 21:45 davidm
1842
1843    * hifn/: Makefile, hifn7751.c, hifn7751var.h:
1844    Makefile for building the hifn driver
1845
18462004-11-18 21:44 davidm
1847
1848    * hifn/: hifn7751.c, hifn7751reg.h, hifn7751var.h:
1849    Checkin the orginal freebsd source for the hifn driver as a
1850    reference in the future.
1851    
1852    
1853
18542004-11-18 10:05 davidm
1855
1856    * Makefile, README, TODO, criov.c, crypto.c, cryptodev.c,
1857    cryptodev.h, cryptosoft.c, uio.h, ixp4xx/Makefile, ixp4xx/ixp4xx.c,
1858    safe/Makefile, safe/safe.c, safe/safevar.h:
1859    Clean up license and copyright info to be more acceptable (and
1860    present) Clean up debug and trace Fixup memory freeing etc on
1861    safenet Fix compiler warnings (some were bugs)
1862    
1863    
1864
18652004-11-17 02:23 davidm
1866
1867    * safe/safe.c:
1868    working on Xscale (big endian) now but packet data is getting
1869    stuffed up due to endian problems (at least now we are talking to
1870    the chip correctly for BE). Good enough to test packet throughput,
1871     no good for testing scp.
1872    
1873    
1874
18752004-11-17 02:21 davidm
1876
1877    * criov.c:
1878    Make sure public symbols are exported by including the correct
1879    header files
1880
18812004-11-17 02:15 davidm
1882
1883    * crypto.c:
1884    even better cleanup of kernel threads on exit
1885
18862004-11-17 02:15 davidm
1887
1888    * cryptosoft.c:
1889    return some trace to debug so it doesn't appear
1890
18912004-11-17 02:14 davidm
1892
1893    * Makefile:
1894    Make sure all drivers get built, not just IXP
1895
18962004-11-16 21:31 davidm
1897
1898    * crypto.c:
1899    Fix problem with reboots and driver not unloading cleanly, we were
1900    not handling signals correctly in the kernel threads,
1901
19022004-11-10 10:46 davidm
1903
1904    * ixp4xx/ixp4xx.c:
1905    fix serious context leak, itturns out the context is still
1906    considered busy while it is calling the perform callback, so we
1907    cleanup on closing the session and on allocating the next context
1908    to work around this.
1909    
1910    
1911
19122004-11-10 05:26 davidm
1913
1914    * crypto.c, cryptodev.c, ixp4xx/ixp4xx.c, safe/safe.c:
1915    cleaned out some debug, found MAX tunnels bug, traced it various
1916    other cleanups.
1917    
1918    
1919
19202004-11-10 04:02 davidm
1921
1922    * cryptodev.c, cryptosoft.c, ixp4xx/ixp4xx.c, safe/Makefile,
1923    safe/safe.c:
1924    Fix up kmalloc usage to always zero resulting buffer everywhere
1925    (stops crashes in ixp)
1926    
1927    Add some function debug to ixp so you can see it working
1928    
1929    Fix safe driver to build and install in a real tree.
1930    
1931    
1932
19332004-11-10 02:27 davidm
1934
1935    * Makefile, criov.c, crypto.c, cryptosoft.c, ixp4xx/Makefile,
1936    ixp4xx/ixp4xx.c:
1937    Compiling OCF modules for the IXP crypto, needs testing now
1938    
1939
19402004-11-09 19:16 davidm
1941
1942    * criov.c, crypto.c, cryptodev.c, cryptodev.h, cryptosoft.c,
1943    cryptosoft.h, uio.h, safe/safe.c, safe/safevar.h:
1944    The linux port of OCF with working safenet and software modules.
1945    Still some bugs with mutiple crypto threads using the safenet
1946    driver.
1947    
1948    
1949
19502004-11-09 18:49 davidm
1951
1952    * Makefile, criov.c, crypto.c, cryptodev.c, cryptodev.h,
1953    cryptosoft.c, cryptosoft.h, uio.h, safe/Makefile, safe/safe.c,
1954    safe/safereg.h, safe/safevar.h:
1955    Check in the orginal free-bsd sources for the OCF support. This
1956    allows us to diff against it later to see if we botched anything
1957    major league.
1958    
1959    
1960
1961

Archive Download this file

Branches:
backfire
history
master
release_2010-11-17
release_2010-12-14
release_2011-02-23
release_2011-05-28
release_2011-08-27
release_2011-11-13
release_2012-03-18
release_2012-04-09
release_2012-10-11
uclibc_nptl



interactive