5506 Commits

Author SHA1 Message Date
John Baldwin
5715307f74 - Move p_cpulimit to struct proc from struct plimit and protect it with
sched_lock.  This means that we no longer access p_limit in mi_switch()
  and the p_limit pointer can be protected by the proc lock.
- Remove PRS_ZOMBIE check from CPU limit test in mi_switch().  PRS_ZOMBIE
  processes don't call mi_switch(), and even if they did there is no longer
  the danger of p_limit being NULL (which is what the original zombie check
  was added for).
- When we bump the current processes soft CPU limit in ast(), just bump the
  private p_cpulimit instead of the shared rlimit.  This fixes an XXX for
  some value of fix.  There is still a (probably benign) bug in that this
  code doesn't check that the new soft limit exceeds the hard limit.

Inspired by:	bde (2)
2002-10-09 17:17:24 +00:00
Julian Elischer
48bfcddd94 Round out the facilty for a 'bound' thread to loan out its KSE
in specific situations. The owner thread must be blocked, and the
borrower can not proceed back to user space with the borrowed KSE.
The borrower will return the KSE on the next context switch where
teh owner wants it back. This removes a lot of possible
race conditions and deadlocks. It is consceivable that the
borrower should inherit the priority of the owner too.
that's another discussion and would be simple to do.

Also, as part of this, the "preallocatd spare thread" is attached to the
thread doing a syscall rather than the KSE. This removes the need to lock
the scheduler when we want to access it, as it's now "at hand".

DDB now shows a lot mor info for threaded proceses though it may need
some optimisation to squeeze it all back into 80 chars again.
(possible JKH project)

Upcalls are now "bound" threads, but "KSE Lending" now means that
other completing syscalls can be completed using that KSE before the upcall
finally makes it back to the UTS. (getting threads OUT OF THE KERNEL is
one of the highest priorities in the KSE system.) The upcall when it happens
will present all the completed syscalls to the KSE for selection.
2002-10-09 02:33:36 +00:00
Warner Losh
0b294f891d Introducing /dev/devctl. This device reports events in the
configuration device hierarchy.  Device arrival, departure and not
matched are presently reported.  This will be the basis for devd, which
I still need to polish a little more before I commit it.  If you don't
use /dev/devctl, it will be a noop.
2002-10-07 23:17:44 +00:00
Warner Losh
c17fdbe3a9 Two minor bugfixes:
o Allow the bus_debug variable to be set via the bus.debug tunable.
o Return pnpinfo and location info via the devinfo interface to userland.
  devinfo(8) needs to be updated to print it.
2002-10-07 23:15:40 +00:00
Ian Dowse
197b023b1b Add back a fdrop() call at the end of kern_open() that got lost in
revision 1.218. This bug caused a "struct file" reference to be
leaked if VOP_ADVLOCK(), vn_start_write(), or mac_check_vnode_write()
failed during the open operation.

PR:		kern/43739
Reported by:	Arne Woerner <woerner@mediabase-gmbh.de>
2002-10-07 20:49:22 +00:00
Warner Losh
0a1d3ef9b8 Add wrappers around the newly created bus_child_pnpinfo_str and
bus_child_location_str.
2002-10-07 07:08:00 +00:00
Warner Losh
d71dec96bf Minor string handling cleanup that I've had in my tree for a while:
Don't use snprintf where strlcpy() will do the job.
Also, a NUL is '\0' not 0 in our style (C doesn't care), so spell it like.
Remove useless {} and () in the general area of this change.
2002-10-07 06:50:35 +00:00
Warner Losh
da7b83f9ea Don't need to NUL terminate after snprintf 2002-10-07 06:26:17 +00:00
Warner Losh
3d9841b4eb Add two interfaces to allow for busses to report the pnpinfo for
devices as well as their location on the bus.
2002-10-07 05:06:38 +00:00
Alfred Perlstein
c814aa3fdb disable debug output by default. 2002-10-07 04:13:21 +00:00
Robert Watson
b371c939ce Integrate mac_check_socket_send() and mac_check_socket_receive()
checks from the MAC tree: allow policies to perform access control
for the ability of a process to send and receive data via a socket.
At some point, we might also pass in additional address information
if an explicit address is requested on send.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-06 14:39:15 +00:00
Robert Watson
e183f80e54 Sync from MAC tree: break out the single mmap entry point into
seperate entry points for each occasion:

mac_check_vnode_mmap()		Check at initial mapping
mac_check_vnode_mprotect()	Check at mapping protection change
mac_check_vnode_mmap_downgrade()	Determine if a mapping downgrade
					should take place following
					subject relabel.

Implement mmap() and mprotect() entry points for labeled vnode
policies.  These entry points are currently not hooked up to the
VM system in the base tree.  These changes improve the consistency
of the access control interface and offer more flexibility regarding
limiting access to vnode mmaping.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
Robert Watson
83985c267e Modify label allocation semantics for sockets: pass in soalloc's malloc
flags so that we can call malloc with M_NOWAIT if necessary, avoiding
potential sleeps while holding mutexes in the TCP syncache code.
Similar to the existing support for mbuf label allocation: if we can't
allocate all the necessary label store in each policy, we back out
the label allocation and fail the socket creation.  Sync from MAC tree.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
Robert Watson
b497ca81d6 Make sure that the accounting credential is saved along with the vp
when accounting is suspended--otherwise when accounting is restored,
we may incorrectly assume the credential is valid.

Panics experienced by:	juli
2002-10-05 20:05:23 +00:00
Robert Watson
74e62b1b75 Integrate a devfs/MAC fix from the MAC tree: avoid a race condition during
devfs VOP symlink creation by introducing a new entry point to determine
the label of the devfs_dirent prior to allocation of a vnode for the
symlink.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 18:40:10 +00:00
Robert Watson
0a69419678 Merge support for mac_check_vnode_link(), a MAC framework/policy entry
point that instruments the creation of hard links.  Policy implementations
to follow.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 18:11:36 +00:00
Robert Watson
56c1541237 While the MAC API has supported the ability to handle M_NOWAIT passed
to mbuf label initialization, that functionality was never merged to
the main tree.  Go ahead and merge that functionality now.  Note that
this requires policy modules to accept the case where the label
element may be destroyed even if init has not succeeded on it (in
the event that policy failed the init).  This will shortly also
apply to sockets.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 17:44:49 +00:00
Robert Watson
87807196f8 Rearrange object and label init/destroy functions to match the
order used in mac_policy.h and elsewhere.  Sort order is basically
"by operation category", then "alphabetically by object". Sync to
MAC tree.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
Robert Watson
a931e345a9 Sync to MAC tree: use 'flag' instead of 'how' for mac_init_mbuf();
remove a slightly less than useful comment.
2002-10-05 17:18:43 +00:00
Brian Feldman
dab3d85fd7 Don't allow dev_stdclone(9) to accept minors larger than the system is
able to handle (0xffffff).
2002-10-05 17:10:28 +00:00
Robert Watson
69bbb5b1c7 Another big diff, little functional change: move label internalization,
externalization, and cred label life cycle events to entirely above
devfs and vnode events.  Sync from MAC tree.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 16:57:16 +00:00
Robert Watson
08bcdc586e Move all object label init/destroy routines to the head of the
entry points to better match the entry point ordering in mac_policy.h.
Big diff, no functional change; merge from the MAC tree.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Robert Watson
ea599aa018 Synch from TrustedBSD MAC tree:
- If a policy isn't registered when a policy module unloads, silently
  succeed.

- Hold the policy list lock across more of the validity tests to avoid
  races.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
Poul-Henning Kamp
3bd6561289 NB: This commit does *NOT* make GEOM the default in FreeBSD
NB: But it will enable it in all kernels not having options "NO_GEOM"

Put the GEOM related options into the intended order.

Add "options NO_GEOM" to all kernel configs apart from NOTES.

In some order of controlled fashion, the NO_GEOM options will be
removed, architecture by architecture in the coming days.

There are currently three known issues which may force people to
need the NO_GEOM option:

boot0cfg/fdisk:
        Tries to update the MBR while it is being used to control
        slices.  GEOM does not allow this as a direct operation.

SCSI floppy drives:
        Appearantly the scsi-da driver return "EBUSY" if no media
        is inserted.  This is wrong, it should return ENXIO.

PC98:
        It is unclear if GEOM correctly recognizes all variants of
        PC98 disklabels.  (Help Wanted!  I have neither docs nor HW)

These issues are all being worked.

Sponsored by:	DARPA & NAI Labs.
2002-10-05 16:35:33 +00:00
Robert Watson
226b96fb6d Cosmetic line wrap synchronization. 2002-10-05 16:33:46 +00:00
Robert Watson
b2f0927ad6 Push the debugging obect label counters into security.mac.debug.counters
rather than directly under security.mac.debug.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
Robert Watson
96adb90996 Begin another merge from the TrustedBSD MAC branch:
- Change mpo_init_foo(obj, label) and mpo_destroy_foo(obj, label) policy
  entry points to mpo_init_foo_label(label) and
  mpo_destroy_foo_label(label).  This will permit the use of the same
  entry points for holding temporary type-specific label during
  internalization and externalization, as well as for caching purposes.
- Because of this, break out mpo_{init,destroy}_socket() and
  mpo_{init,destroy}_mount() into seperate entry points for socket
  main/peer labels and mount main/fs labels.
- Since the prototype for label initialization is the same across almost
  all entry points, implement these entry points using common
  implementations for Biba, MLS, and Test, reducing the number of
  almost identical looking functions.

This simplifies policy implementation, as well as preparing us for the
merge of the new flexible userland API for managing labels on objects.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-05 15:10:00 +00:00
Maxim Sobolev
790a8088d0 Fix problem introduced in rev.1.406, which can cause already unlocked
mutex being unlocked again causing system panic.
2002-10-05 12:56:10 +00:00
Brian Somers
52ae0b7fb5 If dsgetlabel() returns a label with a size of zero in diskdumpconf(),
treat it as an invalid partition.

This fixes a bug where ``dumpon <device>'' will configure the dump
device at a random offset on the disk if <device> isn't a valid
partition.

Reviewed by: phk
2002-10-05 11:24:21 +00:00
Juli Mallett
0d29446006 Put an easy-to-miss assignment into the proper place. It was stray in the
middle of a block of code, with no clear assignment.  While here, move one
nearby assignment out of declaration.
2002-10-05 04:49:46 +00:00
Juli Mallett
ecafb24b41 Remove bogus duplicate assignment of local variables. 2002-10-05 04:35:59 +00:00
Poul-Henning Kamp
c5f9218b48 Add the new function "sbuf_done()" which returns non-zero if the sbuf is
finished.

This allows sbufs to be used for request/response scenarioes without
needing additional communication flags.

Sponsored by:	DARPA & NAI Labs.
2002-10-04 09:58:17 +00:00
Peter Wemm
c281972e61 Add some unspeakable hackery to the tree under #ifdef __ia64__ to work
around limitations in the ia64 kernel stack handling code.  Basically
preallocate a bunch of threads (and hence kstacks) while contigmalloc()
still works, and never free them back to the general memory pool.  After
the system has been running for a while, contigmalloc() eventually fails
at a critical momemt and panics the system.
2002-10-04 01:31:39 +00:00
Don Lewis
cb81d3ca4d hashinit() calls MALLOC(), so release the filedesc lock in knote_attach()
before calling hashinit() and relock afterwards, taking care to see that
we don't lose a race.
2002-10-03 06:03:26 +00:00
Juli Mallett
a723033a4d XXX Add a check for p->p_limit being NULL before dereferencing it. This is
totally bogus but will hide the occurances of access of 0xbc(NULL) which
people have run into lately.  This is not a proper fix, just a bandaid, until
the cause of this happening is tracked down and fixed.

Reviewed by:	rwatson
2002-10-03 04:09:00 +00:00
Don Lewis
91e97a8266 In an SMP environment post-Giant it is no longer safe to blindly
dereference the struct sigio pointer without any locking.  Change
fgetown() to take a reference to the pointer instead of a copy of the
pointer and call SIGIO_LOCK() before copying the pointer and
dereferencing it.

Reviewed by:	rwatson
2002-10-03 02:13:00 +00:00
David Xu
5da2b58aeb set ke_bound to NULL when kse owner thread becomes runnable.
Reviewed by: julian (mentor)
2002-10-03 01:22:05 +00:00
Julian Elischer
4162f2fe92 Whitespace fix only 2002-10-02 23:12:01 +00:00
John Baldwin
551cf4e150 Rename the mutex thread and process states to use a more generic 'LOCK'
name instead.  (e.g., SLOCK instead of SMTX, TD_ON_LOCK() instead of
TD_ON_MUTEX())  Eventually a turnstile abstraction will be added that
will be shared with mutexes and other types of locks.  SLOCK/TDI_LOCK will
be used internally by the turnstile code and will not be specific to
mutexes.  Making the change now ensures that turnstiles can be dropped
in at a later date without affecting the ABI of userland applications.
2002-10-02 20:31:47 +00:00
Juli Mallett
289e1e23d1 Access td->td_kse inside sched_lock.
Submitted by:	julian
2002-10-02 18:25:09 +00:00
Archie Cobbs
36a8dac10d Let kse_wakeup() take a KSE mailbox pointer argument.
Reviewed by:	julian
2002-10-02 16:48:16 +00:00
Juli Mallett
bc7b9f1dba De-obfuscate local use of members of 'struct thread', for which we have
local variables, and group assignment.
2002-10-02 16:39:39 +00:00
Poul-Henning Kamp
c56c20f13d Absorb <sys/bus_private.h> into kern/subr_bus.c to prevent misunderstandings.
Suggested by:	bde
Approved by:	dfr
2002-10-02 09:34:29 +00:00
Poul-Henning Kamp
8c5d013757 Fix mis-indentation.
Spotted by:	FlexeLint
2002-10-02 09:09:25 +00:00
Scott Long
316ec49abd Some kernel threads try to do significant work, and the default KSTACK_PAGES
doesn't give them enough stack to do much before blowing away the pcb.
This adds MI and MD code to allow the allocation of an alternate kstack
who's size can be speficied when calling kthread_create.  Passing the
value 0 prevents the alternate kstack from being created.  Note that the
ia64 MD code is missing for now, and PowerPC was only partially written
due to the pmap.c being incomplete there.
Though this patch does not modify anything to make use of the alternate
kstack, acpi and usb are good candidates.

Reviewed by:	jake, peter, jhb
2002-10-02 07:44:29 +00:00
Robert Watson
92dbb82a47 Add a new MAC entry point, mac_thread_userret(td), which permits policy
modules to perform MAC-related events when a thread returns to user
space.  This is required for policies that have floating process labels,
as it's not always possible to acquire the process lock at arbitrary
points in the stack during system call processing; process labels might
represent traditional authentication data, process history information,
or other data.

LOMAC will use this entry point to perform the process label update
prior to the thread returning to userspace, when plugged into the MAC
framework.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
2002-10-02 02:42:38 +00:00
Juli Mallett
1d9c56964d Back our kernel support for reliable signal queues.
Requested by:	rwatson, phk, and many others
2002-10-01 17:15:53 +00:00
John Baldwin
feb2449610 Minor style nits in a comment. 2002-10-01 15:49:32 +00:00
Poul-Henning Kamp
8d3574c7a4 Fix some harmless mis-indents.
Spotted by:	FlexeLint
2002-10-01 15:48:31 +00:00
Poul-Henning Kamp
328048bc56 Remember to include "opt_devfs.h" so we get any relevant changes
to NDEVFSINO before we include devfs.h.

Spotted by:	FlexeLint
2002-10-01 15:24:35 +00:00