Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
184,859 Commits 606 Branches 0 Tags 3.1 GiB
f92e474cf5
Commit Graph

7133 Commits

Author SHA1 Message Date
rmind
554a0142dc Initialise struct emul members by name (it is readable now and one can search 
them in the tree).
 2009-10-25 01:14:03 +00:00
rmind
33963b1448 Avoid #ifndef __NO_CPU_LWP_FREE, only ia64 is missing cpu_lwp_free 
routines and it can/should provide stubs.
 2009-10-22 22:28:57 +00:00
rmind
30d0b02e57 Make lwp_park_sobj and lwp_park_tab static. 
Wrap long lines while here.
 2009-10-22 13:12:47 +00:00
rmind
40cf6f3659 Remove uarea swap-out functionality: 
- Addresses the issue described in PR/38828.
- Some simplification in threading and sleepq subsystems.
- Eliminates pmap_collect() and, as a side note, allows pmap optimisations.
- Eliminates XS_CTL_DATA_ONSTACK in scsipi code.
- Avoids few scans on LWP list and thus potentially long holds of proc_lock.
- Cuts ~1.5k lines of code.  Reduces amd64 kernel size by ~4k.
- Removes __SWAP_BROKEN cases.

Tested on x86, mips, acorn32 (thanks <mpumford>) and partly tested on
acorn26 (thanks to <bjh21>).

Discussed on <tech-kern>, reviewed by <ad>.
 2009-10-21 21:11:57 +00:00
jym
de3d6f78cf Fix a bug where on MP systems, pool_cache_invalidate(9) could be called 
early during boot, just after CPUs are attached but before they are marked
as running.

This will result in a list of CPUs without the SPCF_RUNNING flag set, and
will trigger the 'KASSERT(xc_tailp < xc_headp)' in xc_lowpri() as no cross
call is issued.

Bug reported and patch tested by tron@.

See also http://mail-index.netbsd.org/tech-kern/2009/10/19/msg006293.html
 2009-10-20 17:24:22 +00:00
snj
07ce40632e Follow upstream's lead and remove third and fourth clauses (except on 
from usr.sbin/mopd/common/pf.c, where only the ad clause is removed,
because it has a shared UCB copyright) on Mats O Jansson's files.

thorpej OK'd usr.sbin/rpc.yppasswdd/yppasswdd_mkpw.c, where he shares
copyright.
 2009-10-20 00:51:13 +00:00
snj
4968c04d96 Move Eduardo Horvath's license to 2 clause. OK eeh@. 2009-10-19 18:12:37 +00:00
jnemeth
30d0592bd3 allow passing a NULL proplib dictionary to modctl(MODCTL_LOAD, ...) 2009-10-16 00:27:07 +00:00
thorpej
1f59a448f4 - pool_cache_invalidate(): broadcast a cross-call to drain the per-CPU 
caches before draining the global cache.
- pool_cache_invalidate_local(): remove.
 2009-10-15 20:50:12 +00:00
pooka
624234c0c5 Generate scheduling points around rump vnode operations. 2009-10-15 00:29:40 +00:00
dsl
931ac5949a Error out of ptcread() if the uio length supplied is zero before the code 
has a chance to panic in ureadc().
 2009-10-14 19:25:39 +00:00
pooka
ddc943db02 regen: fix rump varargs syscalls prototypes 2009-10-13 21:57:52 +00:00
pooka
0d8bdf6131 For varargs syscalls, create rump prototypes which match the regular 
system call counterparts, e.g.:
open(const char *, int, mode_t) -> open(const char *, int, ...)
 2009-10-13 21:54:29 +00:00
yamt
e894729250 sys___aio_suspend50, sys_lio_listio: 
- fix the buffer sizes.
	- use kmem_alloc instead of kmem_zalloc for buffers which we will
	  overwrite soon.
 2009-10-12 23:43:13 +00:00
yamt
29e552b036 wrap long lines. no functional changes. 2009-10-12 23:38:08 +00:00
yamt
b8562be527 make aio_worker static. 2009-10-12 23:36:56 +00:00
yamt
5873138145 constify 2009-10-12 23:36:02 +00:00
yamt
28bf72b353 fix KMEM_SIZE vs KMEM_GUARD 2009-10-12 23:35:09 +00:00
yamt
de25ce6a4c remove no longer necessary include of drvctl.h 2009-10-12 23:33:02 +00:00
yamt
199e4526f3 aio_suspend1: fix a double free bug. 2009-10-12 23:31:59 +00:00
dsl
65dd100015 Check for zero length read here - and return zero. 
Most times we've come through spec_read() which has already done the test,
but not always (eg pty with ptsfs mounted).
Without this there is a simple local-user panic in ureadc().
Noted Matthew Mondor on tech-kern.
 2009-10-11 17:20:48 +00:00
dsl
270307174b Fix locking when collecting pt_read and pt_ucntl. 2009-10-11 08:08:32 +00:00
jym
31629a1342 Add pool_cache_invalidate_local() to the pool_cache(9) API, to permit 
per-CPU objects invalidation when cached in the pool cache.

See http://mail-index.netbsd.org/tech-kern/2009/10/05/msg006206.html .

Reviewed by bouyer@. Thanks!
 2009-10-08 21:54:45 +00:00
elad
2cb56be586 Add a (weak aliased) machdep_init() as a place to do machdep initialization 
that can't happen as early as the other init functions as called from
cpu_startup() -- for example, register kauth(9) listeners.

Put unprivileged policy in the x86 code; used by i386, amd64, and xen.
 2009-10-06 21:07:05 +00:00
elad
756638cf95 Factor out a block of code that appears in three places (Veriexec, keylock, 
and securelevel) so that others can use it as well.
 2009-10-06 04:28:10 +00:00
rmind
c9a5a18df3 mq_timedsend/mq_timedreceive: timeout value is absolute, not relative. 
While here, drop unecessary (since fdesc API changes) lwp_t arguments.

Bug reported by Stathis Kamperis, thanks!
 2009-10-05 23:49:46 +00:00
rmind
5503429772 shmexit: simplify a lot by avoiding unnecessary memory allocations, since 
it is a last reference, just re-lock and check mapping list again.  Often
there wont be re-locks at all, moreover, shm_lock is not contended at all.
 2009-10-05 23:47:04 +00:00
rmind
c3a98b4c87 semu_alloc: simplify a little. 2009-10-05 23:46:02 +00:00
rmind
ac8f63538a Convert cpu_number(), which can be sparse, to cpu_index(), which is MI. 2009-10-05 23:39:27 +00:00
elad
4c9fcb77c3 - Add usermount_common_policy() that implements some common (everything 
but access control) user mounting policies: enforced MNT_NOSUID and
  MNT_NODEV, no MNT_EXPORT, MNT_EXEC propagation. This can be useful for
  secmodels that are interested in simply adding finer grained user mount
  support.

- Add a mount subsystem listener for KAUTH_REQ_SYSTEM_MOUNT_GET.
 2009-10-05 04:20:13 +00:00
elad
fa69dc186a Install floppies (haha) don't get built with ktrace/ptrace, so they don't 
include kern/sys_process.c. Move proc_uidmatch() to kern/kern_proc.c which
always gets built instead.

Pointed out by Kurt Schreiner on current-users@:

    http://mail-index.netbsd.org/current-users/2009/10/03/msg010745.html
 2009-10-04 03:15:08 +00:00
elad
b2f3768346 - Move sched_listener and co. from kern_synch.c to sys_sched.c, where it 
really belongs (suggested by rmind@),

- Rename sched_init() to synch_init(), and introduce a new sched_init()
  in sys_sched.c where we (a) initialize the sysctl node (no more
  link-set) and (b) listen on the process scope with sched_listener.

Reviewed by and okay rmind@.
 2009-10-03 22:32:56 +00:00
elad
458410e7b5 Oops, forgot to make sched_listener static. Pointed out by rmind@, thansk! 2009-10-03 21:21:56 +00:00
elad
54d08ac134 Update a comment. No functional change. 2009-10-03 21:03:55 +00:00
elad
a39251ecc2 Introduce time_wraps() to check if setting the time will wrap it (or 
close to it). Useful for secmodels.

Replace open-coded form with it in secmodel code (securelevel, keylock).

Note: I need to find a way to make secmodel_keylock.c ~<100 lines.
 2009-10-03 20:48:42 +00:00
elad
7f720ad562 KAUTH_GENERIC_CANSEE -> KAUTH_REQ_NETWORK_SOCKET_CANSEE. 
Not quite the same semantics but it's okay. Once our sockets have
credentials (and they will) it's all the same.
 2009-10-03 20:24:39 +00:00
elad
5b3a96a24d Move KAUTH_NETWORK_BIND::KAUTH_REQ_NETWORK_BIND_PORT policy back to the 
subsystem (or close to it).

Note: Revisit KAUTH_REQ_NETWORK_BIND_PRIVPORT.
 2009-10-03 03:59:39 +00:00
elad
82ce55ed44 Move policies for KAUTH_PROCESS_{CANSEE,CORENAME,STOPFLAG,FORK} back to 
the subsystem.

Note: Consider killing the signal listener and sticking
      KAUTH_PROCESS_SIGNAL here as well.
 2009-10-03 03:38:31 +00:00
elad
111de3833c Finish moving socket policy to the subsystem. 2009-10-03 01:41:39 +00:00
elad
452ced03bd Move sched policy back to the subsystem. 2009-10-03 01:30:25 +00:00
elad
212f5fa214 Move kevent policy back to the subsystem. 2009-10-03 00:14:07 +00:00
elad
abc7a4290b Put module loading policy back in the subsystem. 
Revisit: consider moving kauth_init() above module_init() in main().
 2009-10-03 00:06:37 +00:00
elad
1f98cab201 Put the tty opening policy back in the subsystem. 
Remove include we don't need from the secmodel code.
 2009-10-02 23:58:53 +00:00
elad
510083464f Move some of the socket policy back to the subsystem. 
Remove include we don't need in the secmodel code.
 2009-10-02 23:50:16 +00:00
elad
8751f894d8 Put signal delivery policy back in the subsystem. 2009-10-02 23:24:15 +00:00
elad
09f3ac9e2f Stick nice policy in its own subsystem and call the listener "resource" 
rather than "rlimit"...
 2009-10-02 22:46:18 +00:00
elad
bcc5014bd0 Move rlimit policy back to the subsystem. 
For this we needed proc_uidmatch() exposed, which makes a lot of sense,
so put it back in sys_process.c for use in other places as well.
 2009-10-02 22:38:45 +00:00
elad
2ae3a70827 Move ptrace's security policy back to the subsystem itself. 
Add a ptrace_init() so we have a place to register the listener; called
next to ktrinit().
 2009-10-02 22:18:56 +00:00
elad
40cc528a28 Move psets security policy back to the subsystem and keep suser logic only 
in the suser secmodel code.
 2009-10-02 21:56:28 +00:00
elad
932cd15f91 Move ktrace's subsystem security policy to the subsystem itself, and keep 
just the suser-related logic in the suser secmodel.
 2009-10-02 21:47:35 +00:00