Commit graph

4802 commits

Author SHA1 Message Date
cinap_lenrek 0dab8869ad kernel: ignore memory pages with singular kernel addresses
addresses va's of 0 and -BY2PG cause trouble with some memmove()/memset()
implementations and possibly other code because of the nil pointer
and end pointers wrapping to zero.
2015-06-18 12:15:33 +02:00
cinap_lenrek fd8597ac31 zynq: fix barriers
unlock()/iunlock():

we need to place the coherence() *before* "l->key = 0", so that any
stores that where done while holding the lock become observable
*before* other processors see the lock released.

cas()/tas():

place memory barrier before successfull return to prevent reordering.
2015-06-18 04:35:46 +02:00
cinap_lenrek 58dc03cec0 kernel: do not inherit Proc.dot (current working directory) in kproc()
making sure to close the dot in every kproc appears repetitive,
so instead stop inheriting the dot in kproc() as this is usually
never what you wanted in the first place.
2015-06-18 03:13:50 +02:00
cinap_lenrek b48078c12c kernel: do not inherit current directory channel (dot) to pager
kproc() inherits dot and slash, pager needs to drop these
channels, otherwise it will keep the files open preventing
say, ramfs to exit.
2015-06-18 22:58:56 +02:00
ftrvxmtrx 47f05d4d8b usps: cut the useless part 2015-06-17 18:07:32 +02:00
cinap_lenrek 921aa0a6c3 stats: add kernel malloc and kernel draw allocation size graphs 2015-06-16 08:08:42 +02:00
cinap_lenrek 45b79036be devcons: add current pool allocations to #c/swap 2015-06-16 08:05:33 +02:00
cinap_lenrek 6c99d2f028 kernel: remove waserror() arround newpage() in mntcache
newpage() does not raise error().
2015-06-16 06:05:12 +02:00
cinap_lenrek 64ed3658d2 kernel: add pagechaindone() to wakeup processes waiting for memory
we keep the details about palloc in page.c, providing pagechaindone()
for mmu code to be called after a series of pagechainhead() calls.
2015-06-15 17:40:47 +02:00
cinap_lenrek bce2696b17 dontkill: only use noswap flag to mark processes as not to be terminated on oom
changing procmode has the nasty side effect that notes cannot
be posted to them, prventing sysfatal() and threadexitsall()
to cleanup properly.

recent kernels also check the noswap flag to void killing them
in out of memory condition, so this is really all we need now.
2015-06-15 17:13:19 +02:00
aiju 5329fbf89e there's plenty of room in /lib 2015-06-15 16:27:27 +02:00
cinap_lenrek 8a3b388ffe kernel: implement separate wait queues for page allocation
give kernel processes and local disk file servers (procs
having noswap flag set) a clear advantage for page allocation
under starved condition by giving them ther own wait queue so
they get readied as soon as pages become available.
2015-06-15 16:05:00 +02:00
cinap_lenrek 9ab48ee370 hjfs: defer parent directory permission check for wstat
check parent directory permission *after* we determined
that the new name does not exist in the parent, so that
when the new name is the same as old name then no write
permission is required in the parent directory.
2015-06-15 08:28:10 +02:00
cinap_lenrek 13ec55922f hjfs: fix wstat
- rename to same name is ok
- only truncate when file size is different
- check for uid/gid *change*, not if it is specified or not
2015-06-15 01:57:54 +02:00
cinap_lenrek 5adf0606f1 nusb/serial: remove unused channel 2015-06-15 00:30:19 +02:00
cinap_lenrek 584caf6d1c zynq: remove unused PTE typedef
all the mmu code uses ulong, so get rid of the typeded.
2015-06-15 22:32:15 +02:00
cinap_lenrek 13a79fbb72 zynq: remove unused statistics fields from Mach structure 2015-06-15 22:13:55 +02:00
cinap_lenrek 65db705481 ndb/dns: cleanup forwarding code (redistrib())
instead of copying the whole packet, just save the
udp header and restore it aftwards. dont call redistrib()
when there are no forwards (this should be almost always
the case).
2015-06-14 21:31:44 +02:00
stanley lieber ffd294e0da theo: You are on your own. 2015-06-13 20:34:33 -04:00
ftrvxmtrx a55eed44c6 aux/vga: == is not strncmp 2015-06-13 21:28:33 +02:00
cinap_lenrek 7f4b037b0b htmlroff: fix wrong newline in \name handling (| -> ||) 2015-06-13 20:58:56 +02:00
ftrvxmtrx 453ca8db43 qc, 9/ppc, 9/mtx: fix wrong if statements 2015-06-13 20:21:26 +02:00
cinap_lenrek d6eb7cc71c kernel: dont use smalloc() to allocate pte array in ibrk()
when we'r out of kernel memory, it is probably better to
let that alloc fail instead of hanging while holding the
segment qlock.
2015-06-13 17:50:26 +02:00
cinap_lenrek 27fb90eb6e vblade, cifs, usbuhci: fix parenthesis 2015-06-13 15:03:56 +02:00
cinap_lenrek 9c1dff3fa9 lock: reopen lockfile after changing mode to make lock effective
lock is acquired on open/create when file has DMEXCL set in mode,
so we need to reopen the file after setting the bit with wstat.
2015-06-13 15:01:02 +02:00
cinap_lenrek 1473e5d437 lib9p: fix lib9p wstat qid.type/mode checks (fixes lock(1) for ramfs/hjfs) 2015-06-13 12:43:15 +02:00
cinap_lenrek f92057cc39 ndb/dns: use same buffer size of udp packet in redistrib() as dnudpserver() (fixes assert) 2015-06-13 22:37:26 +02:00
cinap_lenrek cc3abe39aa games/doom: remove Makefile 2015-06-12 17:50:43 +02:00
cinap_lenrek e4c3f92c16 games/doom: fix desyncing demo (thanks qwx)
the code used P_Random()-P_Random() in some places which has
undefined evaluation order resulting in the wrong pseudo random
numbers being returned causing demo playback to desync.

this change adds P_Random2() function which returns the right
delta-random number and uses it in place of P_Random()-P_Random()
expression.
2015-06-12 17:28:09 +02:00
cinap_lenrek 34ae4649cc kernel: fix accounttime() for HZ >= 1000
"milli-CPU's" is too low resolution for the decaying load average
calculation when HZ >= 1000.
2015-06-12 14:28:31 +02:00
cinap_lenrek 2112ad6793 zynq: increase kernel memory for graphics 2015-06-11 21:26:26 +02:00
cinap_lenrek d8bcd6c8e4 zynq: mouse acceleration 2015-06-11 16:44:05 +02:00
cinap_lenrek 36d2504adb zynq: make screen and cursor kprocs exit properly 2015-06-11 16:14:11 +02:00
cinap_lenrek 0dbbe79116 zynq: add support for hardware cursor
to enable hardware cursor, write "addr va" to /dev/mousectl where
va is the virtual address of the cursor position register.
2015-06-11 16:04:24 +02:00
ftrvxmtrx 98640950d7 merge 2015-06-11 22:14:24 +02:00
ftrvxmtrx a25758dce7 fix strncmp/memcmp used with a wrong number 2015-06-11 22:13:55 +02:00
cinap_lenrek 036ae2e397 cifsd: smbrename is still ok when file is readonly (fixed git) 2015-06-10 17:18:31 +02:00
cinap_lenrek c89b0b3c23 zynq: fix flushmemscreen() 2015-06-10 01:27:44 +02:00
ftrvxmtrx be36fab405 remove drawlsetrefresh and convM2Tnoenc: undefined functions 2015-06-09 18:25:09 +02:00
ftrvxmtrx e56e61fed1 addpt(2): put badrect in .SH NAME so that it gets into the searchindex 2015-06-09 12:32:38 +02:00
ftrvxmtrx 86acce7c4d esetcursor(2), alarm(2): fix prototypes 2015-06-09 12:29:57 +02:00
ftrvxmtrx 72a168b5fb event(2): fix estartfn prototype 2015-06-09 11:54:22 +02:00
ftrvxmtrx 3bd0d66686 authsrv.h: remove nonexistent prototypes 2015-06-09 11:46:37 +02:00
ftrvxmtrx a314302e64 libdraw: sync allocimage/allocwindow prototypes with man pages 2015-06-09 10:33:30 +02:00
cinap_lenrek 94333d83ab ip: fix wrong radix for iphash() (thanks yoann padioleau)
yoann padioleaus report on 9fans:

> I think I’ve found a bug in the network stack.
> in 9/ip/ip.h there is
> struct Ipht
> {
> 	Lock;
> 	Iphash	*tab[Nipht];
> };
>
> where Night is 521,
>
> but then in 9/ip/ipaux.c there is
>
> ulong
> iphash(uchar *sa, ushort sp, uchar *da, ushort dp)
> {
> 	return ((sa[IPaddrlen-1]<<24) ^ (sp << 16) ^ (da[IPaddrlen-1]<<8) ^ dp ) % Nhash;
> }
>
> where Nhash is just 64,
2015-06-09 10:04:04 +02:00
cinap_lenrek 90b4fe9cf6 zynq: add /dev/fbctl to attach framebuffer to devdraw 2015-06-09 09:17:55 +02:00
cinap_lenrek cda46731d8 devsegment: fix parsecmd() memory leak 2015-06-09 03:33:37 +02:00
cinap_lenrek e2a5d674d9 libdraw: consistent use of nil for pointers, error handling 2015-06-09 01:42:59 +02:00
ftrvxmtrx 0fc761bc84 unix is BUGGERED. remove it 2015-06-09 00:15:59 +02:00
ftrvxmtrx 7a95bef00e authsrv: remove unused opasstokey 2015-06-09 23:53:07 +02:00