Commit graph

5498 commits

Author SHA1 Message Date
cinap_lenrek 1848f4e946 kernel: tsemacquire() use MACHP(0)->ticks for time delta
we might wake up on a different cpu after the sleep so
delta from machX->ticks - machY->ticks can become negative
giving spurious timeouts. to avoid this always use the
same mach 0 tick counter for the delta.
2016-09-07 23:36:04 +02:00
cinap_lenrek bfd8098b8d devcap: timeout capabilities after a minute, fix memory leak, paranoia
the manpage states that capabilities time out after a minute,
so we add ticks field into the Caphash struct and record the
time when the capability was inserted. freeing old capabilities
is handled in trimcaps(), which makes room for one extra cap
and frees timed out ones.

we also limit the capuse write size to less than 1024 bytes to
prevent denial of service as we have to copy the user buffer.
(memory exhaustion).

we have to check the from user *before* attempting to remove
the capability! the wrong user shouldnt be able to change any
state. this fixes the memory leak of the caphash.

do the hash comparsion with tsmemcmp(), avoiding timing
side channels.

allocate the capabilities in secret memory pool to prevent
debugger access.
2016-09-07 21:14:23 +02:00
cinap_lenrek 76daf9f863 merge 2016-09-06 22:29:40 +02:00
cinap_lenrek cf78fd37cb devproc: do unsigned subtraction to get MACHP(0)->ticks - up->times[TReal] delta 2016-09-06 22:27:26 +02:00
aiju 6d96b35fad pc(1): bugfix: allow setting output base to 0 (thanks, deuteron) 2016-09-05 15:48:46 +00:00
cinap_lenrek 9e926eeb4e 5c: do shift propagation for rotate right (ROR) 2016-09-03 19:05:28 +02:00
cinap_lenrek a5d8cd30bb 5c: format assembly constant right shift encoding 0 as >>32 2016-09-03 18:18:28 +02:00
cinap_lenrek b46a0e97ea 5a: assemble constant >>0 right shifts as <<0 (no shift), allow >>32
previously, right shift >>0 resulted in >>32 being emited. this
is especially problematic when the shift count comes from a macro
expansion.

we now handle constant shift >>0 as <<0 (no shift) and allow
shift by 32 be specified.

this applies to logical right shift (>>) arithmetic right shift (->)
and right rotate (@>).
2016-09-03 17:11:38 +02:00
cinap_lenrek 607f3bc55c 5l: format assembly constant right shift encoding 0 as >>32 2016-09-03 17:01:56 +02:00
cinap_lenrek 2e120a29db libmach: fix RORREG, right shift with shift count 0 means >>32 2016-09-03 16:47:51 +02:00
aiju 1b4f432eae pc(1): manpage: restore previous size rather than set to 10 2016-09-01 11:46:44 +00:00
aiju 183f700a20 pc(1): manpage formatting 2016-09-01 11:38:33 +00:00
aiju 30c9e34c0d pc: add cat() function 2016-09-01 10:55:12 +00:00
aiju d552fed385 mptrunc: don't write to r->p[r->top] 2016-09-01 10:46:37 +00:00
aiju 560ea5474c pc: add rev function 2016-08-31 13:00:50 +00:00
aiju e95082f66c pc: add gcd, rand and minv; set base of logical operation results to 0 2016-08-29 09:57:15 +02:00
cinap_lenrek a08bf6831f libmp: remove unused mpeuclid.c 2016-08-29 02:09:34 +02:00
cinap_lenrek 5852f8a144 libmp: allow passing nil to v,x,y results of mpextendedgcd(), simplify mpinvert() 2016-08-29 02:07:52 +02:00
cinap_lenrek 33862ff793 libmp: mpnrand(), what was i *THINKING*
the prior implementation was unneccesarily complicated for
no good reason due to me misunderstanding how libc's nrand()
works. in contrast to libc, we already generate the *closest*
power-of-2 random number with mprand() in the sampling loop.
2016-08-29 00:45:16 +02:00
cinap_lenrek b80684a1d2 python: remove automatic compiled module loading (.pyc files) 2016-08-28 17:39:18 +02:00
cinap_lenrek 1f9bdb6f5d libmp: mpdiv(): fix divisor==quotient case (again) 2016-08-28 16:46:32 +02:00
cinap_lenrek bdc87e17bd libmp: timingsafe sign flip for small power-of-two negative divisor for mpdiv() 2016-08-28 16:33:52 +02:00
aiju b65f409d54 mp.h: add mpasr 2016-08-28 16:05:22 +02:00
aiju c3da573eb7 pc(1): SYNOPSIS... 2016-08-28 13:59:16 +02:00
aiju d9d2925fa0 pc(1): document -n option in manpage 2016-08-28 13:49:26 +02:00
aiju a931ad737a add pc(1) 2016-08-28 13:40:01 +02:00
aiju 43bb71c8cc mpdiv: negative divisor has to flip sign of quotient 2016-08-28 12:00:25 +02:00
aiju 344ff48512 mp: fix mptov and mptouv 2016-08-28 11:38:29 +02:00
aiju cf2f2a8841 mp: fix mpnot and add mpasr 2016-08-28 10:49:41 +02:00
cinap_lenrek 193daffafb mercurial: use new d_stat from dirent structure in osutil.listdir 2016-08-28 03:41:25 +02:00
cinap_lenrek 5debddf8af ape: add d_stat struct in dirent struct allowing the avoidance of stats 2016-08-28 03:39:49 +02:00
aiju a8ffbcabea timmy: manpage 2016-08-28 01:23:51 +02:00
aiju 913ce62cfc strtomp: update the returned char* even if there were no characters parsed 2016-08-28 01:08:26 +02:00
aiju 261ec9fe0e strtomp: fix mpbits() call in octal code 2016-08-28 01:07:34 +02:00
cinap_lenrek 75990fc719 fix wrong manpage section index generation for rc-httpd(8) 2016-08-27 22:37:08 +02:00
cinap_lenrek f2519d701c rio(4): document text truncation 2016-08-27 21:28:21 +02:00
cinap_lenrek b59aa67922 rand(2), cons(3): clarify /dev/random behaviour 2016-08-27 21:27:52 +02:00
cinap_lenrek f777743b72 merge 2016-08-27 20:50:55 +02:00
cinap_lenrek 0a5f81a442 kernel: switch to fast portable chacha based seed-once random number generator 2016-08-27 20:42:31 +02:00
cinap_lenrek 71ac88392f devsdp: keep cipher states in secret memory 2016-08-27 20:39:36 +02:00
cinap_lenrek 7f16c92762 ip/esp: allocate cipher states in secret memory 2016-08-27 20:38:33 +02:00
cinap_lenrek 2967f942ea devtls: allocate cipher states in secret memory 2016-08-27 20:37:31 +02:00
cinap_lenrek 7250c438bb devssl: allocate cipher states in secret memory 2016-08-27 20:37:14 +02:00
cinap_lenrek 0ac260b18a wifi: update wifi.h header 2016-08-27 20:36:08 +02:00
cinap_lenrek 0b8851ddb6 wifi: allocate cipher states in secret memory, do AESstate key setup once 2016-08-27 20:35:39 +02:00
cinap_lenrek 0f97eb3a60 kernel: add secalloc() and secfree() functions for secret memory allocation
The kernel needs to keep cryptographic keys and cipher states
confidential. secalloc() allocates memory from the secret pool
which is protected from debuggers reading the memory thru devproc.
secfree() releases the memory, overriding the data with garbage.
2016-08-27 20:33:03 +02:00
cinap_lenrek 8a73650874 libc: add poolisoverlap() and definitions for Pool *secrmem 2016-08-27 20:23:55 +02:00
aiju 762e98d47e Bfn: set +x 2016-08-26 21:23:13 +02:00
aiju 1bd3ace9e7 document Bfn 2016-08-25 23:39:07 +02:00
aiju 81b1fc5a12 fix Bfn botch 2016-08-25 23:03:54 +02:00