Commit Graph

2730 Commits

Author SHA1 Message Date
John Baldwin
26f9f5c7c7 - Use TRAPF_PC() on the alpha to acess the PC in the trap frame.
- Don't hold sched_lock around addupc_task() as this apparently breaks
  profiling badly due to sched_lock being held across copyin().

Reported by:	bde (2)
2001-02-22 16:23:12 +00:00
John Baldwin
0246af0995 GC unused and now obsolete assertion macros. 2001-02-22 15:45:49 +00:00
John Baldwin
3e5da75445 - Add a new ithread_schedule() function to do the bulk of the work of
scheduling an interrupt thread to run when needed.  This has the side
  effect of enabling support for entropy gathering from interrupts on
  all architectures.
- Change the software interrupt and x86 and alpha hardware interrupt code
  to use ithread_schedule() for most of their processing when scheduling
  an interrupt to run.
- Remove the pesky Warning message about interrupt threads having entropy
  enabled.  I'm not sure why I put that in there in the first place.
- Add more error checking for parameters and change some cases that
  returned EINVAL to panic on failure instead via KASSERT().
- Instead of doing a documented evil hack of setting the P_NOLOAD flag
  on every interrupt thread whose pri was SWI_CLOCK, set the flag
  explicity for clk_ithd's proc during start_softintr().
2001-02-20 10:25:29 +00:00
John Baldwin
5813dc03bd - Don't call clear_resched() in userret(), instead, clear the resched flag
in mi_switch() just before calling cpu_switch() so that the first switch
  after a resched request will satisfy the request.
- While I'm at it, move a few things into mi_switch() and out of
  cpu_switch(), specifically set the p_oncpu and p_lastcpu members of
  proc in mi_switch(), and handle the sched_lock state change across a
  context switch in mi_switch().
- Since cpu_switch() no longer handles the sched_lock state change, we
  have to setup an initial state for sched_lock in fork_exit() before we
  release it.
2001-02-20 05:26:15 +00:00
Bruce Evans
0ad74739ac Removed all traces of T_ASTFLT (except for gaps where it was). It became
unused except in dead code when ast() was split off from trap().
2001-02-19 15:47:38 +00:00
Bruce Evans
866546105a Changed the aston() family to operate on a specified process instead of
always on curproc.  This is needed to implement signal delivery properly
(see a future log message for kern_sig.c).

Debogotified the definition of aston().  aston() was defined in terms
of signotify() (perhaps because only the latter already operated on
a specified process), but aston() is the primitive.

Similar changes are needed in the ia64 versions of cpu.h and trap.c.
I didn't make them because the ia64 is missing the prerequisite changes
to make astpending and need_resched per-process and those changes are
too large to make without testing.
2001-02-19 04:15:59 +00:00
Bruce Evans
12a586bbda Fixed style bugs in clock.c rev.1.164 and cpu.h rev.1.52-1.53 -- declare
tsc_present in the right places (together with other variables of the
same linkage), and don't use messy ifdefs just to avoid exporting it in
some cases.
2001-02-19 03:00:34 +00:00
Mark Murray
2564fe499d Allow the superuser to prefent all interrupt harvesting on
her system.
2001-02-18 17:47:55 +00:00
Jake Burkholder
d5a08a6065 Implement a unified run queue and adjust priority levels accordingly.
- All processes go into the same array of queues, with different
  scheduling classes using different portions of the array.  This
  allows user processes to have their priorities propogated up into
  interrupt thread range if need be.
- I chose 64 run queues as an arbitrary number that is greater than
  32.  We used to have 4 separate arrays of 32 queues each, so this
  may not be optimal.  The new run queue code was written with this
  in mind; changing the number of run queues only requires changing
  constants in runq.h and adjusting the priority levels.
- The new run queue code takes the run queue as a parameter.  This
  is intended to be used to create per-cpu run queues.  Implement
  wrappers for compatibility with the old interface which pass in
  the global run queue structure.
- Group the priority level, user priority, native priority (before
  propogation) and the scheduling class into a struct priority.
- Change any hard coded priority levels that I found to use
  symbolic constants (TTIPRI and TTOPRI).
- Remove the curpriority global variable and use that of curproc.
  This was used to detect when a process' priority had lowered and
  it should yield.  We now effectively yield on every interrupt.
- Activate propogate_priority().  It should now have the desired
  effect without needing to also propogate the scheduling class.
- Temporarily comment out the call to vm_page_zero_idle() in the
  idle loop.  It interfered with propogate_priority() because
  the idle process needed to do a non-blocking acquire of Giant
  and then other processes would try to propogate their priority
  onto it.  The idle process should not do anything except idle.
  vm_page_zero_idle() will return in the form of an idle priority
  kernel thread which is woken up at apprioriate times by the vm
  system.
- Update struct kinfo_proc to the new priority interface.  Deliberately
  change its size by adjusting the spare fields.  It remained the same
  size, but the layout has changed, so userland processes that use it
  would parse the data incorrectly.  The size constraint should really
  be changed to an arbitrary version number.  Also add a debug.sizeof
  sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
Mark Murray
d888fc4e73 RIP <machine/lock.h>.
Some things needed bits of <i386/include/lock.h> - cy.c now has its
own (only) copy of the COM_(UN)LOCK() macros, and IMASK_(UN)LOCK()
has been moved to <i386/include/apic.h> (AKA <machine/apic.h>).
Reviewed by:	jhb
2001-02-11 10:44:09 +00:00
Jake Burkholder
3cbe75a414 Clear the reschedule flag after finding it set in userret(). This
used to be in cpu_switch(), but I don't see any difference between
doing it here.
2001-02-10 20:33:35 +00:00
John Baldwin
e990501c21 Re-enable preemption on interrupts. My last commit accidentally reverted
it as I was playing with some other ways of doing kernel preemption.
2001-02-10 02:41:50 +00:00
John Baldwin
142ba5f3d7 - Make astpending and need_resched process attributes rather than CPU
attributes.  This is needed for AST's to be properly posted in a preemptive
  kernel.  They are backed by two new flags in p_sflag: PS_ASTPENDING and
  PS_NEEDRESCHED.  They are still accesssed by their old macros:
  aston(), astoff(), etc.  For completeness, an astpending() macro has been
  added to check for a pending AST, and clear_resched() has been added to
  clear need_resched().
- Rename syscall2() on the x86 back to syscall() to be consistent with
  other architectures.
2001-02-10 02:20:34 +00:00
John Baldwin
e109e2b4cd Add a macro mtx_intr_enable() to alter a spin lock such that interrupts
will be enabled when it is released.
2001-02-10 02:15:18 +00:00
John Baldwin
5781f5419e Catch up to changes to inthand_add(). 2001-02-09 17:48:33 +00:00
John Baldwin
2e0c76cd20 Use the MI ithread helper functions in the x86 interrupt code. 2001-02-09 17:47:44 +00:00
John Baldwin
062d8ff5a0 - Catch up to the new swi API changes:
- Use swi_* function names.
  - Use void * to hold cookies to handlers instead of struct intrhand *.
- In sio.c, use 'driver_name' instead of "sio" as the name of the driver
  lock to minimize diffs with cy(4).
2001-02-09 17:46:35 +00:00
John Baldwin
929604ec9b Move the initailization of the proc lock for proc0 very early into the MD
startup code.
2001-02-09 16:25:16 +00:00
John Baldwin
a91fe908db Woops, remove an obsolete reference to gd_cpu_lockid. 2001-02-09 16:13:57 +00:00
John Baldwin
40bcb55186 Remove unused forward_irq counters. 2001-02-09 14:30:03 +00:00
John Baldwin
7f918b0230 Axe gd_cpu_lockid as it is no longer used. 2001-02-09 14:25:22 +00:00
Bosko Milekic
9ed346bab0 Change and clean the mutex lock interface.
mtx_enter(lock, type) becomes:

mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks)
mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized)

similarily, for releasing a lock, we now have:

mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN.
We change the caller interface for the two different types of locks
because the semantics are entirely different for each case, and this
makes it explicitly clear and, at the same time, it rids us of the
extra `type' argument.

The enter->lock and exit->unlock change has been made with the idea
that we're "locking data" and not "entering locked code" in mind.

Further, remove all additional "flags" previously passed to the
lock acquire/release routines with the exception of two:

MTX_QUIET and MTX_NOSWITCH

The functionality of these flags is preserved and they can be passed
to the lock/unlock routines by calling the corresponding wrappers:

mtx_{lock, unlock}_flags(lock, flag(s)) and
mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN
locks, respectively.

Re-inline some lock acq/rel code; in the sleep lock case, we only
inline the _obtain_lock()s in order to ensure that the inlined code
fits into a cache line. In the spin lock case, we inline recursion and
actually only perform a function call if we need to spin. This change
has been made with the idea that we generally tend to avoid spin locks
and that also the spin locks that we do have and are heavily used
(i.e. sched_lock) do recurse, and therefore in an effort to reduce
function call overhead for some architectures (such as alpha), we
inline recursion for this case.

Create a new malloc type for the witness code and retire from using
the M_DEV type. The new type is called M_WITNESS and is only declared
if WITNESS is enabled.

Begin cleaning up some machdep/mutex.h code - specifically updated the
"optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN
and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently
need those.

Finally, caught up to the interface changes in all sys code.

Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Mike Smith
6c9eb5f3a6 Free the memory we get from devclass_get_devices and device_get_children.
Submitted by:	wpaul
2001-02-08 20:44:49 +00:00
John Baldwin
297c46b68c Don't enable interrupts for a kernel breakpoint or trace trap. Otherwise,
this negates the explicit disabling of interrupts when entering the
debugger in Debugger().
2001-02-08 00:10:07 +00:00
John Baldwin
67b00ca875 When SMPng was first committed, we removed 'cpl' from the interrupt
frame.  Teach ddb about this as there is one less word for it to skip
over when finding a trapframe on the interrupt frame stack.
2001-02-07 22:41:47 +00:00
Jeroen Ruigrok van der Werven
1a6e52d0e9 Fix typo: seperate -> separate.
Seperate does not exist in the english language.
2001-02-06 11:21:58 +00:00
Peter Wemm
bcf77694d1 Clean up some leftovers from the root mount cleanup that was done some
time ago.  FFS_ROOT and CD9660_ROOT are obsolete.
2001-02-04 15:35:10 +00:00
Matthew Dillon
4e71e795a1 This commit represents work mainly submitted by Tor and slightly modified
by myself.  It solves a serious vm_map corruption problem that can occur
with the buffer cache when block sizes > 64K are used.  This code has been
heavily tested in -stable but only tested somewhat on -current.  An MFC
will occur in a few days.  My additions include the vm_map_simplify_entry()
and minor buffer cache boundry case fix.

Make the buffer cache use a system map for buffer cache KVM rather then a
normal map.

Ensure that VM objects are not allocated for system maps.  There were cases
where a buffer map could wind up with a backing VM object -- normally
harmless, but this could also result in the buffer cache blocking in places
where it assumes no blocking will occur, possibly resulting in corrupted
maps.

Fix a minor boundry case in the buffer cache size limit is reached that
could result in non-optimal code.

Add vm_map_simplify_entry() calls to prevent 'creeping proliferation'
of vm_map_entry's in the buffer cache's vm_map.  Previously only a simple
linear optimization was made.  (The buffer vm_map typically has only a
handful of vm_map_entry's.  This stabilizes it at that level permanently).

PR: 20609
Submitted by: (Tor Egge) tegge
2001-02-04 06:19:28 +00:00
Jake Burkholder
05f6ee66ea Implement preemptive scheduling of hardware interrupt threads.
- If possible, context switch to the thread directly in sched_ithd(),
  rather than triggering a delayed ast reschedule.

- Disable interrupts while restoring fpu state in the trap handler,
  in order to ensure that we are not preempted in the middle, which
  could cause migration to another cpu.

Reviewed by:	peter
Tested by:	peter (alpha)
2001-02-01 03:34:20 +00:00
Peter Wemm
8ab109d131 Remove count for NSIO. The only places it was used it were incorrect.
(alpha-gdbstub.c got sync'ed up a bit with the i386 version)
2001-01-31 10:54:45 +00:00
Peter Wemm
aa0b4c590f Remove some leftovers from the CMAP* stuff in globaldata and the
BSP and AP startup.
2001-01-30 04:02:28 +00:00
Peter Wemm
d87ad35b4a Remove unused GD_CPU_LOCKID, GD_OTHER_CPUS, PS_IDLESTACK and
PS_IDLESTACK_TOP
2001-01-30 04:00:11 +00:00
John Baldwin
7ad1d369bb Remove unnecessary locking to protect the p_upages_obj and p_addr
pointers.
2001-01-30 00:35:35 +00:00
Peter Wemm
f444a0efe9 Convert mca (microchannel bus support) from something that we count
(bogus) to something that we test for the presence of.
2001-01-29 11:57:27 +00:00
Peter Wemm
03927d3c33 Send "#if NISA > 0" to the bit-bucket and replace it with an option.
These were compile-time "is the isa code present?" tests and not
'how many isa busses' tests.
2001-01-29 09:38:39 +00:00
Peter Wemm
5022186bc6 Remove stray #include "isa.h" 2001-01-29 08:33:55 +00:00
Jake Burkholder
07ce8269b1 Clear intr_nesting_level when an interrupt thread has no more
handlers and wants to exit, so it doesn't panic in exit1()
which malloc()s with M_WAITOK.

Reported by:	Bob Bishop <rb@gid.co.uk>
2001-01-28 17:20:11 +00:00
Bosko Milekic
84e11fbc2e Move the setting of curproc to idleproc up earlier in ap_init(). The
problem is that a mutex lock, prior to this change, is acquired before
the curproc is set to idleproc, so we mess ourselves up by calling
the mutex lock routine with curproc == NULL.

Moving it up after the aps_ready spin-wait has us hopefully setting it
after idleproc is setup.

Solved by: jake (the allmighty) :-)
2001-01-28 03:41:01 +00:00
Tor Egge
48bed92485 Defer assignment of low level interrupt handlers for PCI interrupts
described in the MP table until something asks for the interrupt number
later on.
2001-01-28 01:07:54 +00:00
Poul-Henning Kamp
b0254d109c Turn DEVFS on by default.
You may need to turn this off if you you vinum.  Apart from that I know of
no reason not to run with DEVFS.
2001-01-27 08:18:36 +00:00
Jake Burkholder
28df158b49 Push Giant down into the trap handlers that need it, instead of
acquiring it unconditionally.

Reviewed by:	jhb
2001-01-26 04:16:16 +00:00
John Baldwin
b00df9c9ba Whitespace fix: convert code indented 6 spaces to use tabs instead. 2001-01-25 19:37:26 +00:00
John Baldwin
2a36ec35ae - Change fork_exit() to take a pointer to a trapframe as its 3rd argument
instead of a trapframe directly.  (Requested by bde.)
- Convert the alpha switch_trampoline to call fork_exit() and use the MI
  fork_return() instead of child_return().
- Axe child_return().
2001-01-24 21:59:25 +00:00
Jason Evans
1b367556b5 Convert all simplelocks to mutexes and remove the simplelock implementations. 2001-01-24 12:35:55 +00:00
John Baldwin
bb4b2d8b2a Remove the Xforward_irq IPI. 2001-01-24 10:01:13 +00:00
John Baldwin
cd60bc00a7 - Remove all the #if 0'd code that used to implement IRQ forwarding.
- Remove #if 0'd lazy interrupt mask.
2001-01-24 10:00:07 +00:00
John Baldwin
5b8c74d59a - Proc locking.
- P_OWEUPC -> PS_OWEUPC.
- Remove obsolete prototype for MD fork_return().
2001-01-24 09:56:49 +00:00
John Baldwin
0d54b9d12c Setup the return values for a child process in the trapframe when we setup
the rest of the trapframe instead of doing it in fork_return().
2001-01-24 09:54:44 +00:00
John Baldwin
625c76db3a - Kill the have_giant parameter to userret() along with all instances of
that name as a variable.  Use mtx_owned(&Giant) where appropriate
  instead.
- Proc locking.
- P_FOO -> PS_FOO.
- Update comments about enable interrupts during trap and why this may be
  bad if we trap while holding a spin mutex.
- Don't bother resetting p to curproc in syscall() in case we are the child
  returning from fork.  The child hasn't returned from fork through syscall
  in a while.
- Remove fork_return() as it has been superseded by the MI version.
2001-01-24 09:53:49 +00:00
John Baldwin
16fdce5377 - Proc locking.
- P_INMEM -> PS_INMEM.
2001-01-24 09:49:49 +00:00
John Baldwin
4cbdef8448 - Relocate portions of this file to get it into an order closer to that of
the alpha mp_machdep.c.
- Proc locking.
- Catch up to the P_FOO -> PS_FOO proc flags changes.
- Stick ap_init()'s prototype with the other prototypes.
- Remove the Xforwardirq IPI.
- Remove unused simplelocks.
- Don't try to psignal() from forward_statclock(), but set the appropriate
  signal pending flag in p_sflag instead.
- Add in KTR_SMP tracepoints for various SMP functions.   (Brought over
  from the alpha port)
2001-01-24 09:48:52 +00:00
John Baldwin
52070bf2e6 - Proc locking.
- Setup proc0.p_heldmtx, proc0.contested, and curproc earlier so that we
  can use mutexes.
- Initialize sched_lock and Giant earlier and enter Giant during init386.
- Use suser(9) instead of checking cr_uid directly.
2001-01-24 09:45:20 +00:00
John Baldwin
89cb8fc912 Call fork_exit() now instead of futzing around in assembly during a fork
return.
2001-01-24 09:43:26 +00:00
Jason Evans
0cde2e34af Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex
inline functions non-inlined.  Hide parts of the mutex implementation that
should not be exposed.

Make sure that WITNESS code is not executed during boot until the mutexes
are fully initialized by SI_SUB_MUTEX (the original motivation for this
commit).

Submitted by:	peter
2001-01-21 22:34:43 +00:00
Dag-Erling Smørgrav
a3ea6d41b9 First step towards an MP-safe zone allocator:
- have zalloc() and zfree() always lock the vm_zone.
 - remove zalloci() and zfreei(), which are now redundant.

Reviewed by:	bmilekic, jasone
2001-01-21 22:23:11 +00:00
Jake Burkholder
a448b62ac9 Make intr_nesting_level per-process, rather than per-cpu. Setup
interrupt threads to run with it always >= 1, so that malloc can
detect M_WAITOK from "interrupt" context.  This is also necessary
in order to context switch from sched_ithd() directly.

Reviewed By:	peter
2001-01-21 19:25:07 +00:00
Peter Wemm
654c30a008 Remove APIC_INTR_DIAGNOSTIC - this has been disabled for some time now.
Remove some leftovers of removed SMP options.
2001-01-21 07:54:10 +00:00
Jason Evans
d1c1b8413e Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex
initialization until after malloc() is safe to call, then iterate through
all mutexes and complete their initialization.

This change is necessary in order to avoid some circular bootstrapping
dependencies.
2001-01-21 07:52:20 +00:00
Jake Burkholder
3e899e1063 Remove the per-cpu pages used for copy and zero-ing pages of memory
for SMP; just use the same ones as UP.  These weren't used without
holding Giant anyway, and the routines that use them would have to
be protected from pre-emption to avoid migrating cpus.
2001-01-21 06:50:03 +00:00
Jake Burkholder
7dd2de5bc2 Rename the ASSYM MTX_RECURSE to MTX_RECURSECNT in order to not conflict
with the flag of the same name.
2001-01-20 06:07:36 +00:00
Jake Burkholder
87dce36809 Simplify the i386 asm MTX_{ENTER,EXIT} macros to just call the
appropriate function, rather than doing a horse-and-buggy
acquire.  They now take the mutex type as an arg and can be
used with sleep as well as spin mutexes.
2001-01-20 04:14:25 +00:00
Jake Burkholder
c1ef8aac9e - Make npx_intr INTR_MPSAFE and move acquiring Giant into the
function itself.
- Remove a hack to allow acquiring Giant from the npx asm trap
  vector.
2001-01-20 02:30:58 +00:00
Peter Wemm
1467a651ab Convert apm from a bogus 'count' into a plain option. Clean out some
other cruft from the files.alpha and files.ia64 that were related to this.
2001-01-19 14:09:54 +00:00
Peter Wemm
aaa69bf4d0 Zap unused #include "apm.h" 2001-01-19 13:56:31 +00:00
Peter Wemm
558226eae7 Use #ifdef DEV_NPX from opt_npx.h instead of #if NNPX > 0 from npx.h 2001-01-19 13:19:02 +00:00
Peter Wemm
d008d67eff At great personal risk to my sanity, turn off COMPAT_OLDISA and the
two drivers that depend on it - ie and le.  The compat code has not been
disabled.
2001-01-19 13:06:37 +00:00
John Baldwin
8339bca6db Add in a space that got lost in the previous commit in some debugging code
so that '&' becomes a binary operator and not a unary operator.
2001-01-19 11:43:13 +00:00
Peter Wemm
95bca00f46 EEK! I missed a couple of places with the 24->32 interrupt change. 2001-01-19 10:55:13 +00:00
Peter Wemm
031b6a102e Remove reference to splz_unpend - it is long gone. 2001-01-19 10:51:13 +00:00
Peter Wemm
07a79932e3 Catch a few alternative names for the syscall entry frame, eg: post-ELF
and int $0x80 entry methods.
2001-01-19 10:49:20 +00:00
Peter Wemm
f7b6e45d5b apic_itrace_splz[] is unused 2001-01-19 10:48:35 +00:00
Peter Wemm
47c6b72621 Fix a warning due to missing prototype. 2001-01-19 09:10:14 +00:00
Peter Wemm
dfb1071f3c Fix a warning (the prototypes probably shouldn't be so over-zealously
#ifdef'ed though)
2001-01-19 09:07:16 +00:00
Bosko Milekic
08812b3925 Implement MTX_RECURSE flag for mtx_init().
All calls to mtx_init() for mutexes that recurse must now include
the MTX_RECURSE bit in the flag argument variable. This change is in
preparation for an upcoming (further) mutex API cleanup.
The witness code will call panic() if a lock is found to recurse but
the MTX_RECURSE bit was not set during the lock's initialization.

The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to
MTX_RECURSED, which is more appropriate given its meaning.

The following locks have been made "recursive," thus far:
eventhandler, Giant, callout, sched_lock, possibly some others declared
in the architecture-specific code, all of the network card driver locks
in pci/, as well as some other locks in dev/ stuff that I've found to
be recursive.

Reviewed by: jhb
2001-01-19 01:59:14 +00:00
John Baldwin
dcfc09d931 Protect p_stat and p_oncpu with sched_lock in forward_signal(). 2001-01-18 08:19:25 +00:00
John Baldwin
324fffaed1 - Sort of lie and say that %eax is an output only and not an input for the
non-386 atomic_load_acq().  %eax is an input since its value is used in
  the cmpxchg instruction, but we don't care what value it is, so setting
  it to a specific value is just wasteful.  Thus, it is being used without
  being initialized as the warning stated, but it is ok for it to be used
  because its value isn't important.  Thus, we are only sort of lying when
  we say it is an output only operand.
- Add "cc" to the clobber list for atomic_load_acq() since the cmpxchgl
  changes ZF.
2001-01-17 02:15:11 +00:00
Peter Wemm
e44a0ea311 Stop doing runtime checking on i386 cpus for cpu class. The cpu is
slow enough as it is, without having to constantly check that it really
is an i386 still.  It was possible to compile out the conditionals for
faster cpus by leaving out 'I386_CPU', but it was not possible to
unconditionally compile for the i386.  You got the runtime checking whether
you wanted it or not.  This makes I386_CPU mutually exclusive with the
other cpu types, and tidies things up a little in the process.

Reviewed by:  alfred, markm, phk, benno, jlemon, jhb, jake, grog, msmith,
              jasone, dcs, des (and a bunch more people who encouraged it)
2001-01-16 09:10:34 +00:00
John Baldwin
248c2e72aa Argh, disable the micro-ops again. I didn't test these adequately and
managed to lock up one of my machines in world again.

Pointy-hat to:	me
2001-01-16 04:48:38 +00:00
John Baldwin
2ccd992403 - Use "+a" instead of "=&a" for several constraints. This should fix
compiling errors where gcc would run out of registers.
- Add "cc" to the list of clobbers for micro-ops where we perform
  instructions that alter %eflags.
- Use xchgl instead of cmpxchgl to release a spin lock.  This could allow
  for more efficient register allocation as we no longer mandate that %eax
  be used.
- Reenable the optimized mutex micro-ops in the non-i386 case.
2001-01-16 03:45:54 +00:00
John Baldwin
10fd583277 Free the intrhand name when free'ing a intrhand.
Submitted by:	bde
2001-01-16 02:17:51 +00:00
John Baldwin
8a6b1c8f54 - Fix atomic_load_* and atomic_store_* to generate functions for atomic.c
that modules can call.
- Remove the old gcc <= 2.8 versions of the atomic ops.
- Resort the order of some things in the file so that there is only
  one #ifdef for KLD_MODULE, and so that all WANT_FUNCTIONS stuff is
  moved to the bottom of the file.
- Remove ATOMIC_ACQ_REL() and just use explicit macros instead.
2001-01-16 00:18:36 +00:00
Peter Wemm
194b08ffc3 Implement an optimization for INTREN/INTRDIS that bde pointed out last
time I tinkered around here.  Since INTREN is called from the interrupt
critical path now, it should not be too expensive.  In this case, we
look at the bits being changed to decide which 8 bit IO port to write to
rather than unconditionally writing to both.  I could probably have gone
further and only done the write if the bits actually changed, but that
seemed overkill for the usual case in interrupt threads.

[an outb is rather expensive when it has to cross the ISA bus]
2001-01-15 04:18:58 +00:00
Mark Murray
b79ad7e642 Remove NOBLOCKRANDOM as a compile-time option. Instead, provide
exactly the same functionality via a sysctl, making this feature
a run-time option.

The default is 1(ON), which means that /dev/random device will
NOT block at startup.

setting kern.random.sys.seeded to 0(OFF) will cause /dev/random
to block until the next reseed, at which stage the sysctl
will be changed back to 1(ON).

While I'm here, clean up the sysctls, and make them dynamic.
Reviewed by:		des
Tested on Alpha by:	obrien
2001-01-14 17:50:15 +00:00
John Baldwin
4408a4e8ff Argh, remove a local customization that snuck in here.
Noticed by:	jasone
2001-01-14 10:19:42 +00:00
John Baldwin
decc707afc Remove I386_CPU from GENERIC. Support for the 386 seriously pessimizes
performance on other x86 processors.  Custom kernels can still be built
that will run on the 386.
2001-01-14 10:11:10 +00:00
John Baldwin
285e3ac733 Revert the previous revision now that atomic_store_rel_ptr() actually
works.
2001-01-14 09:56:35 +00:00
John Baldwin
9d979d8912 Fix the atomic_load_acq() and atomic_store_rel() functions to properly
implement memory fences for the 486+.  The 386 still uses versions w/o
memory fences as all operations on the 386 are not program ordered.
The 386 versions are not MP safe.
2001-01-14 09:55:21 +00:00
John Baldwin
6fe65214f2 Work around the broken atomic_store_rel_ptr() on the i386 arch by just
using atomic_cmpset_rel_ptr() instead for _release_lock_quick().  When
atomic_store_rel_ptr() is functional and MP safe, then this can be
reverted.
2001-01-14 00:16:17 +00:00
Jake Burkholder
063415120b Change return ??? to return -1 in some #if 0'ed code. 2001-01-12 08:24:25 +00:00
Bosko Milekic
6a47d852b9 Remove declaration of airq variable from outer block. There were two
declarations of a variable of the same name. The one in the outer block
was unused and probably just slipped in at one point or another. This
silences a compiler warning.
2001-01-12 07:49:29 +00:00
Jake Burkholder
7586909279 Remove unused per-cpu variables inside_intr and ss_eflags. 2001-01-12 07:47:54 +00:00
Bosko Milekic
00e5e98182 Remove useless include of sys/mbuf.h (no longer useful since the
mbuf subsystem init was moved to a better place).
2001-01-12 07:46:28 +00:00
Jake Burkholder
df729d6f00 - Remove compatibility macros for accessing per-cpu variables.
__FreeBSD_version 500015 can be used to detect their disappearance.
- Move the symbols for SMP_prvspace and lapic from globals.s to
  locore.s.
- Remove globals.s with extreme prejudice.
2001-01-11 14:46:26 +00:00
Jake Burkholder
ef73ae4b0c Use PCPU_GET, PCPU_PTR and PCPU_SET to access all per-cpu variables
other then curproc.
2001-01-10 04:43:51 +00:00
Jake Burkholder
bb5c0622b7 Fix a warning. The type of globaldata.gd_prvspace has changed. 2001-01-08 15:25:45 +00:00
Jake Burkholder
f8761e53a7 Implement accessors for per-cpu variables which don't depend on the
symbols in globals.s.

	PCPU_GET(name) returns the value of the per-cpu variable
	PCPU_PTR(name) returns a pointer to the per-cpu variable
	PCPU_SET(name, val) sets the value of the per-cpu variable

In general these are not yet used, compatibility macros remain.

Unifdef SMP struct globaldata, this makes variables such as cpuid
available for UP as well.

Rebuilding modules is probably a good idea, but I believe old
modules will still work, as most of the old infrastructure
remains.
2001-01-06 19:55:42 +00:00
Jake Burkholder
41ed17bfec Use %fs to access per-cpu variables in uni-processor kernels the same
as multi-processor kernels.  The old way made it difficult for kernel
modules to be portable between uni-processor and multi-processor
kernels.  It is no longer necessary to jump through hoops.

- always load %fs with the private segment on entry to the kernel
- change the type of the self referntial pointer from struct privatespace
  to struct globaldata
- make the globaldata symbol have value 0 in all cases, so the symbols
  in globals.s are always offsets, not aliases for fields in globaldata
- define the globaldata space used for uniprocessor kernels in C, rather
  than assembler
- change the assmebly language accessors to use %fs, add a macro
  PCPU_ADDR(member, reg), which loads the register reg with the address
  of the per-cpu variable member
2001-01-06 17:40:04 +00:00
Jake Burkholder
98f03f9030 Protect proc.p_pptr and proc.p_children/p_sibling with the
proctree_lock.

linprocfs not locked pending response from informal maintainer.

Reviewed by:	jhb, -smp@
2000-12-23 19:43:10 +00:00
Paul Richards
3164c5878a Re-enable the lnc driver in GENERIC. 2000-12-20 11:24:47 +00:00
John Baldwin
5f38ca907c Remove the "machine dependent" KTR trace buffer ddb commands. The code was
exactly the same on all platforms.
2000-12-14 23:57:30 +00:00
Jake Burkholder
7d8e3aa013 Use _lapic+offset to access the local apic from assembly language
files, rather than the symbols in globals.s.  The offsets are
generated by genassym.
2000-12-14 04:16:16 +00:00
John Baldwin
05f9877c15 If we fail to emulate a vm86 trap in kernel mode, then we use
vm86_trap() to return to the calling program directly.  vm86_trap()
doesn't return, thus it was never returning to trap() to release
Giant.  Thus, release Giant before calling vm86_trap().
2000-12-13 18:57:15 +00:00
Seigo Tanimura
21cd6e6232 - If swap metadata does not fit into the KVM, reduce the number of
struct swblock entries by dividing the number of the entries by 2
until the swap metadata fits.

- Reject swapon(2) upon failure of swap_zone allocation.

This is just a temporary fix. Better solutions include:
(suggested by:	dillon)

o reserving swap in SWAP_META_PAGES chunks, and
o swapping the swblock structures themselves.

Reviewed by:	alfred, dillon
2000-12-13 10:01:00 +00:00
Jake Burkholder
6d43764a10 Introduce a new potientially cleaner interface for accessing per-cpu
variables from i386 assembly language.  The syntax is PCPU(member)
where member is the capitalized name of the per-cpu variable, without
the gd_ prefix.  Example: movl %eax,PCPU(CURPROC).  The capitalization
is due to using the offsets generated by genassym rather than the symbols
provided by linking with globals.o.  asmacros.h is the wrong place for
this but it seemed as good a place as any for now.  The old implementation
in asnames.h has not been removed because it is still used to de-mangle
the symbols used by the C variables for the UP case.
2000-12-13 09:23:53 +00:00
Mike Smith
e8d5a72218 Remove the COMPAT_OLDPCI option, it's going away.
Turn 'lnc' off in GENERIC for the moment, pending its update to newbus.
2000-12-13 01:11:34 +00:00
Jake Burkholder
c0c2557090 - Change the allproc_lock to use a macro, ALLPROC_LOCK(how), instead
of explicit calls to lockmgr.  Also provides macros for the flags
  pased to specify shared, exclusive or release which map to the
  lockmgr flags.  This is so that the use of lockmgr can be easily
  replaced with optimized reader-writer locks.
- Add some locking that I missed the first time.
2000-12-13 00:17:05 +00:00
John Baldwin
5057d21069 Add in symbols needed in the WITNESS_ENTER and WITNESS_EXIT macros in
i386/include/mutex.h.
2000-12-12 16:40:20 +00:00
John Baldwin
18f619abce Fix the assembly mutex macros to call the appropriate witness functions if
the witness code is compiled in.  Without this, the witness code doesn't
notice that sched_lock is released by fork_trampoline() and thus gets all
confused about spin lock order later on.
2000-12-12 03:49:58 +00:00
Jake Burkholder
92cf772d8d - Add code to detect if a system call returns with locks other than Giant
held and panic if so (conditional on witness).
- Change witness_list to return the number of locks held so this is easier.
- Add kern/syscalls.c to the kernel build if witness is defined so that the
  panic message can contain the name of the offending system call.
- Add assertions that Giant and sched_lock are not held when returning from
  a system call, which were missing for alpha and ia64.
2000-12-12 01:14:32 +00:00
Poul-Henning Kamp
9b67714311 Remove DDB, it leaked in here with another commit.
Submitted by:	bde
2000-12-11 14:02:00 +00:00
Mike Smith
bb0d0a8efc Next phase in the PCI subsystem cleanup.
- Move PCI core code to dev/pci.
 - Split bridge code out into separate modules.
 - Remove the descriptive strings from the bridge drivers.  If you
   want to know what a device is, use pciconf.  Add support for
   broadly identifying devices based on class/subclass, and for
   parsing a preloaded device identification database so that if
   you want to waste the memory, you can identify *anything* we know
   about.
 - Remove machine-dependant code from the core PCI code.  APIC interrupt
   mapping is performed by shadowing the intline register in machine-
   dependant code.
 - Bring interrupt routing support to the Alpha
   (although many platforms don't yet support routing or mapping
   interrupts entirely correctly).  This resulted in spamming
   <sys/bus.h> into more places than it really should have gone.
 - Put sys/dev on the kernel/modules include path.  This avoids
   having to change *all* the pci*.h includes.
2000-12-08 22:11:23 +00:00
David Malone
7cc0979fd6 Convert more malloc+bzero to malloc+M_ZERO.
Submitted by:	josh@zipperup.org
Submitted by:	Robert Drehmel <robd@gmx.net>
2000-12-08 21:51:06 +00:00
Jake Burkholder
1b00f92039 Revert the previous change I made to cpu_switch. It doesn't help as
much as I thought it would and according to bde was a pessimization.
2000-12-08 21:31:52 +00:00
Poul-Henning Kamp
959b7375ed Staticize some malloc M_ instances. 2000-12-08 20:09:00 +00:00
Jake Burkholder
cc3f51b537 Fix a jump to the wrong label, <sigh>. Put a period at the end of a
sentence in a comment.

Submitted by:	bde
2000-12-08 19:53:37 +00:00
John Baldwin
31ae9b450a Argh, revert the clobber changes. Since %ecx and %edx aren't call safe,
calling the C functions mtx_enter_hard() and mtx_exit_hard() clobbers them.
Note that %eax is also not call safe, but it is already clobbered due to
cmpxchg.  However, now we are back to not compiling again, so these macros
are still left disabled for now.
2000-12-08 18:21:06 +00:00
Jake Burkholder
96d57f35b2 Change the calling conventions of the MTX_ENTER macro to match
that of MTX_EXIT.  Don't assume that the reg parameter to MTX_ENTER
holds curproc, load it explicitly.  Put semi-colons at the end of
the macros to be more consistent and so its harder to forget them
when these change.
2000-12-08 08:49:36 +00:00
John Baldwin
5e028585be Well, the previous commit wasn't entirely correct either. For now, just
disable the optimized mutex micro-operations for the non-I386_CPU case
and fall back to the C stubs that call the atomic_foo() inlines.
2000-12-08 05:03:34 +00:00
Poul-Henning Kamp
45b031c5e4 Move extern tsc_present outside function to quelch a warning. 2000-12-07 22:30:11 +00:00
Mitsuru IWASAKI
b52617dc0e Create a pmtimer device instance for GENERIC and NEWCARD kernels by default.
Submitted by:	Masayuki FUKUI <fukui@sonic.nm.fujitsu.co.jp>
2000-12-07 14:27:02 +00:00
John Baldwin
a423d0122b Fix broken register restraints that needlessly clobbered registers %ecx
and %edx resulting in gcc not having enough registers left to work with.
2000-12-07 02:23:16 +00:00
Peter Wemm
4366ac52ad This is kind of a nasty hack, but it appears to solve the Compaq DL360
SMP problem.  Compaq, in their infinite wisdom, forgot to put the IO apic
intpin #0 connection to the 8259 PIC into the mptable.  This hack is to
look and see if intpin #0 has *no* table entry and adds a fake ExtInt
entry for the remap routines to use.  isa/clock.c will still test the
interrupts.  This entry is only ever used on an already broken system.
2000-12-06 03:47:14 +00:00
Peter Wemm
a263238c86 Move io_apic_{read,write} from apic_ipl.s (where they do not belong) into
mpapic.c.  This gives us the benefit of C type checking.  These functions
are not called in any critical paths and are not used by the interrupt
routines.
2000-12-06 01:04:02 +00:00
Peter Wemm
9a18987b73 GC unused assembler function apic_eoi() 2000-12-06 00:38:04 +00:00
Jake Burkholder
1eb44f0270 Remove the last of the MD netisr code. It is now all MI. Remove
spending, which was unused now that all software interrupts have
their own thread.  Make the legacy schednetisr use an atomic op
for setting bits in the netisr mask.

Reviewed by:	jhb
2000-12-05 00:36:00 +00:00
Peter Wemm
5ee171d264 Cleanup some leftover lint from the old interrupt system.
Also, while here, run up to 32 interrupt sources on APIC systems.
Normalize INTREN/INTRDIS so they are the same on both UP and SMP systems
rather than sometimes a macro, and sometimes a function.

Reviewed by:  jhb, jakeb
2000-12-04 21:15:14 +00:00
Jake Burkholder
f315dbdbd4 (1) Allow a stray lock prefix to be compiled out with the
MPLOCKED macro
(2)	Use decimal 12 rather than hex 0xc in an addl
(3)	Implement MTX_ENTER for the I386_CPU case
(4)	Use semi-colons between instructions to allow MTX_ENTER
	and MTX_ENTER_WITH_RECURSION to be assembled
(5)	Use incl instead of incw to increment the recusion count
(6)	10 is not a valid label, use 7, 8 and 9 rather than 8, 9 and 10
(7)	Sort numeric labels

Submitted by:	bde (2, 4, and 5)
2000-12-04 12:38:03 +00:00
Jake Burkholder
1306962a43 Change cpu_switch to explicitly popl the callers program counter and
pushl that of the new process, rather than doing a movl (%esp) and
assuming that the stack has been setup right.  This make the initial
stack setup slightly more sane, and will make it easier to stick
an interrupted process onto the run queue without its knowing.
2000-12-03 01:09:59 +00:00
Mark Murray
4a3a2f0704 Namespace cleanup. Remove some #includes in favour of an explicit
declaration.

Asked for by:	bde
2000-12-02 17:59:41 +00:00
John Baldwin
1bcb20def1 Fix this slightly better by using NON_GPROF_RET instead of duplicating
hard-coded asm.

Suggested by:	bde
2000-12-01 05:29:21 +00:00
Jake Burkholder
21ad98bca8 Change doreti to take a trapframe instead of an intrframe.
Remove associated pushes of dummy units to convert frame.

Reviewed by:	jhb
2000-12-01 02:09:45 +00:00
John Baldwin
e60877b3ae Revert the previous change to this file. We have to hardcode in the opcode
for return because we do Evil Things(tm) with a 'ret' macro in
asmacros.h.

Noticed by:	markm
2000-11-30 20:07:49 +00:00
Marcel Moolenaar
d034d459da Don't use p->p_sigstk.ss_flags to keep state of whether the
process is on the alternate stack or not. For compatibility
with sigstack(2) state is being updated if such is needed.

We now determine whether the process is on the alternate
stack by looking at its stack pointer. This allows a process
to siglongjmp from a signal handler on the alternate stack
to the place of the sigsetjmp on the normal stack. When
maintaining state, this would have invalidated the state
information and causing a subsequent signal to be delivered
on the normal stack instead of the alternate stack.

PR: 22286
2000-11-30 05:23:49 +00:00
Peter Wemm
083e9ed543 Increase NKPT from 17 to 30. This fixes the 4GB ram boot panic on both
-current and RELENG_4 with GENERIC.

NKPT is the number of initial bootstrap page table pages we create for
the kernel during startup. Once VM is up, we resize it as needed, but
with 4G ram, the size of the vm_page_t structures was pushing it over
the limit.  The fact that trimmed down kernels boot on 4G ram machines
suggests that we were pretty close to the edge.

The "30" is arbitary, but smaller than the 'nkpt' variable on all
machines that I checked.
2000-11-30 01:53:02 +00:00
John Baldwin
8d9888d37a Don't wait forever for CPUs to stop or restart. Instead, give up after a
timeout.  If DIAGNOSTIC is turned on, then display a message to the console
with a map of which CPUs failed to stop or restart.  This gives an SMP box
at least a fighting chance of getting into DDB if one of the other CPUs has
interrupts disabled.
2000-11-28 23:52:36 +00:00
John Baldwin
bf8bfade99 Use atomic ops to close a race condition on the in_Debugger variable used
to only allow 1 CPU at a time to (non-recursively) enter the debugger.
2000-11-28 23:15:44 +00:00
Jonathan Lemon
e82ac18e52 Revert the last commit to the callout interface, and add a flag to
callout_init() indicating whether the callout is safe or not.  Update
the callers of callout_init() to reflect the new interface.

Okayed by: Jake
2000-11-25 06:22:16 +00:00
Jake Burkholder
553629ebc9 Protect the following with a lockmgr lock:
allproc
	zombproc
	pidhashtbl
	proc.p_list
	proc.p_hash
	nextpid

Reviewed by:	jhb
Obtained from:	BSD/OS and netbsd
2000-11-22 07:42:04 +00:00
Mark Murray
39413503a4 Assembler fixes.
Fix opcodes that were typed as ".byte 0xNN, 0xMM" when an older
assembler could not recognise the newer Pentium instructions.
Reviewed by:	jhb
2000-11-21 20:16:49 +00:00
Mark Murray
5855006767 Add a consistent API to a feature that most modern CPUs have; a fast
counter register in-CPU.

This is to be used as a fast "timer", where linearity is more important
than time, and multiple lines in the linearity caused by multiple CPUs
in an SMP machine is not a problem.

This adds no code whatsoever to the FreeBSD kernel until it is actually
used, and then as a single-instruction inline routine (except for the
80386 and 80486 where it is some more inline code around nanotime(9).

Reviewed by:	bde, kris, jhb
2000-11-21 19:55:21 +00:00
John Baldwin
0fb3a41200 Stop handcoding a couple of instructions since gas 2.10 can properly
assemble 16-bit code.

Noticed by:	markm
2000-11-21 18:36:10 +00:00
Jake Burkholder
fa2fbc3dac - Protect the callout wheel with a separate spin mutex, callout_lock.
- Use the mutex in hardclock to ensure no races between it and
  softclock.
- Make softclock be INTR_MPSAFE and provide a flag,
  CALLOUT_MPSAFE, which specifies that a callout handler does not
  need giant.  There is still no way to set this flag when
  regstering a callout.

Reviewed by:	-smp@, jlemon
2000-11-19 06:02:32 +00:00
Jake Burkholder
7da6f97772 - Split the run queue and sleep queue linkage, so that a process
may block on a mutex while on the sleep queue without corrupting
it.
- Move dropping of Giant to after the acquire of sched_lock.

Tested by:	John Hay <jhay@icomtek.csir.co.za>
		jhb
2000-11-17 18:09:18 +00:00
John Baldwin
835a748f30 - Change extra sanity checks in cpu_switch() to be conditional on INVARIANTS
instead of DIAGNOSTIC.
- Remove the p_wchan check as it no longer applies since a process may be
  switched out during CURSIG() within msleep() or mawait().
- Remove an extra sanity check only needed during the early SMPng work.
2000-11-17 17:37:43 +00:00
John Baldwin
20cdcc5b73 Don't release and acquire Giant in mi_switch(). Instead, release and
acquire Giant as needed in functions that call mi_switch().  The releases
need to be done outside of the sched_lock to avoid potential deadlocks
from trying to acquire Giant while interrupts are disabled.

Submitted by:	witness
2000-11-16 02:16:44 +00:00
John Baldwin
7c06c69188 Assert that Giant is not owned during the main loop of ithd_loop(). 2000-11-15 22:03:26 +00:00
Warner Losh
e655ea71de Add pmtimer device, necessary for proper time keeping when apm or
other power management devices are enabled.
2000-11-15 18:36:24 +00:00
John Baldwin
4ae465d099 Always enable interrupts during fork_trampoline() after releasing the
sched_lock.  This is needed for kernel threads that are created before
interrupts are enabled.  kthreads created by kld's that are created at
SI_SUB_KLD such as the random kthread.

Tested by:	phk
2000-11-14 23:01:24 +00:00
Jordan K. Hubbard
4dc665b8bf Proper capitalization of PCMCIA (and avoid matching pcm) 2000-11-14 01:13:57 +00:00
Jordan K. Hubbard
68ce54fbde In the year 2000, I think it's perfectly reasonable to include audio
support by default in GENERIC.
2000-11-14 01:11:13 +00:00