Commit Graph

79 Commits

Author SHA1 Message Date
julian
e839899222 Don't use local variable 'p' in a debug statement.. we removed it. 2002-09-23 14:06:12 +00:00
julian
bcb38a31ff slightly clean up the thread_userret() and thread_consider_upcall() calls.
also some slight changes for TDF_BOUND testing and small style changes
Should ONLY affect KSE programs

Submitted by:	davidxu
2002-09-23 06:14:30 +00:00
peter
2586a4ce95 Argh. I've been reading makefiles for too long. Change comment to a
C-style comment.
2002-09-17 07:41:30 +00:00
peter
70885af57c Stub out the calls to get_mcontext and set_mcontext which only exist on
i386.  This stuff should not be prototyped in MD inludes if the interface
is expected to be MI.
2002-09-17 07:40:15 +00:00
mini
94ac5d965f Add kernel support needed for the KSE-aware libpthread:
- Use ucontext_t's to store KSE thread state.
	- Synthesize state for the UTS upon each upcall, rather than
	  saving and copying a trapframe.
	- Deliver signals to KSE-aware processes via upcall.
	- Rename kse mailbox structure fields to be more BSD-like.
	- Store the UTS's stack in struct proc in a stack_t.

Reviewed by:	bde, deischen, julian
Approved by:	-arch
2002-09-16 19:26:48 +00:00
julian
c7e9e7e892 Allocate KSEs and KSEGRPs separatly and remove them from the proc structure.
next step is to allow > 1 to be allocated per process. This would give
multi-processor threads. (when the rest of the infrastructure is
in place)

While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc
are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
julian
2689f8b40e Apparently something down in the guts of vm/uvm still needs giant
Obtained from:	 mini via P4 KSE tree.
2002-09-14 06:23:43 +00:00
julian
5702a380a5 Completely redo thread states.
Reviewed by:	davidxu@freebsd.org
2002-09-11 08:13:56 +00:00
julian
e26b912512 fix braino..
was clearing part of wrong thread structure..
2002-09-07 12:58:44 +00:00
julian
d191c82c92 fix misplaced schedlock
Submitted by:	davidxu@freebsd.org
2002-09-07 01:48:53 +00:00
julian
4446570abf Use UMA as a complex object allocator.
The process allocator now caches and hands out complete process structures
*including substructures* .

i.e. it get's the process structure with the first thread (and soon KSE)
already allocated and attached, all in one hit.

For the average non threaded program (non KSE that is) the allocated thread and its stack remain attached to the process, even when the process is
unused and in the process cache. This saves having to allocate and attach it
later, effectively bringing us (hopefully) close to the efficiency
of pre-KSE systems where these were a single structure.

Reviewed by:	davidxu@freebsd.org, peter@freebsd.org
2002-09-06 07:00:37 +00:00
davidxu
b1d94c37f7 s/SGNL/SIG/
s/SNGL/SINGLE/
s/SNGLE/SINGLE/

Fix abbreviation for P_STOPPED_* etc flags, in original code they were
inconsistent and difficult to distinguish between them.

Approved by: julian (mentor)
2002-09-05 07:30:18 +00:00
davidxu
de678b0952 In the kernel code, we have the tsleep() call with the PCATCH argument.
PCATCH means 'if we get a signal, interrupt me!" and tsleep returns
either EINTR or ERESTART depending on the circumstances.  ERESTART is
"special" because it causes the system call to fail, but right as it
returns back to userland it tells the trap handler to move %eip back a
bit so that userland will immediately re-run the syscall.
This is a syscall restart. It only works for things like read() etc where
nothing has changed yet. Note that *userland* is tricked into restarting
the syscall by the kernel. The kernel doesn't actually do the restart. It
is deadly for things like select, poll, nanosleep etc where it might cause
the elapsed time to be reset and start again from scratch.  So those
syscalls do this to prevent userland rerunning the syscall:
  if (error == ERESTART) error = EINTR;

Fake "signals" like SIGTSTP from ^Z etc do not normally invoke userland
signal handlers. But, in -current, the PCATCH *is* being triggered and
tsleep is returning ERESTART, and the syscall is aborted even though no
userland signal handler was run.
That is the fault here.  We're triggering the PCATCH in cases that we
shouldn't.  ie: it is being triggered on *any* signal processing, rather
than the case where the signal is posted to userland.
	--- Peter

The work of psignal() is a patchwork of special case required by the process
debugging and job-control facilities...
	--- Kirk McKusick
	"The design and impelementation of the 4.4BSD Operating system"
	Page 105

in STABLE source, when psignal is posting a STOP signal to sleeping
process and the signal action of the process is SIG_DFL, system will
directly change the process state from SSLEEP to SSTOP, and when
SIGCONT is posted to the stopped process, if it finds that the process
is still on sleep queue, the process state will be restored to SSLEEP,
and won't wakeup the process.

this commit mimics the behaviour in STABLE source tree.

Reviewed by: Jon Mini, Tim Robbins, Peter Wemm
Approved by: julian@freebsd.org (mentor)
2002-09-03 12:56:01 +00:00
julian
a9901abd65 Fix crack-smoking code that was panicing on the quad xeon:
- If either of proc or kse are NULL during thread_exit(), then
          the kernel is going to fault because parts of the function
          assume they aren't NULL.  Instead, just assert they aren't NULL
          (as well as the kse group) and assume they are in all of the
          code.  It doesn't make sense for them to be NULL here anyways.
        - Move the PROC_UNLOCK(p) up above clearing td_proc, etc. since
          otherwise we will panic if the proc's lock is contested.

Submitted by:	jhb@freebsd.org
2002-08-29 19:49:53 +00:00
julian
169932bd89 slight cleanup of single-threading code for KSE processes 2002-08-22 21:45:58 +00:00
julian
c99effb6f5 Revert some suspension/sleep/signal code from KSE-III
We need to rethink a bit of this and it doesn't matter if
we break the KSE test program for now as long
as non-KSE programs act as expected.

Submitted by:	David Xu <bsddiy@yahoo.com>
	(this guy's just asking to get hit with a commit bit..)
2002-08-21 20:03:55 +00:00
julian
a1b63b5aa6 Fix a comment. 2002-08-01 19:10:40 +00:00
julian
891c9fcb89 get suspension counting right.
fix an error message

Submitted by:	David Xu <bsddiy@yahoo.com>
2002-07-25 03:21:35 +00:00
julian
0374f9c1bb fix some style problems and remove a mis-merged assert. 2002-07-25 00:27:39 +00:00
julian
7e6f866aa8 Add some locking asserts and some comments 2002-07-24 23:21:05 +00:00
julian
1308e5a426 When single threading a multithreaded program, awaken the
'single threading thread' when the last other thread suspends.
I had this code in there before but it seems to have been
accidentally deleted somewhere along the way.  This would only affect
multithreaded processes.

Reviewed by:	David Xu <bsddiy@yahoo.com>
2002-07-24 19:50:08 +00:00
julian
83080a2d6a When suspending a thread, update the appropriate (sic) statistic. 2002-07-24 07:29:16 +00:00
peter
df4e1dcc74 ia64 does not have the same degree of stealth include file nesting,
so it needs an explicit #include <machine/frame.h> to get 'struct
trapframe'.  The fact that it needs this at this level is rather bogus
but it will not compile without it.
2002-07-17 23:43:55 +00:00
peter
d6b8f1f302 Pacify gcc on ia64 2002-07-17 23:32:13 +00:00
julian
d84464f213 Thinking about it I came to the conclusion that the KSE states were incorrectly
formulated.  The correct states should be:
IDLE:  On the idle KSE list for that KSEG
RUNQ:  Linked onto the system run queue.
THREAD: Attached to a thread and slaved to whatever state the thread is in.

This means that most places where we were adjusting kse state can go away
as it is just moving around because the thread is..
The only places we need to adjust the KSE state is in transition to and from
the idle and run queues.

Reviewed by:	jhb@freebsd.org
2002-07-14 03:43:33 +00:00
peter
b73c441dad Collect all the (now equivalent) pmap_new_proc/pmap_dispose_proc/
pmap_swapin_proc/pmap_swapout_proc functions from the MD pmap code
and use a single equivalent MI version.  There are other cleanups
needed still.

While here, use the UMA zone hooks to keep a cache of preinitialized
proc structures handy, just like the thread system does.  This eliminates
one dependency on 'struct proc' being persistent even after being freed.
There are some comments about things that can be factored out into
ctor/dtor functions if it is worth it.  For now they are mostly just
doing statistics to get a feel of how it is working.
2002-07-07 23:05:27 +00:00
peter
feeea8100f Make this compile on 64 bit platforms 2002-07-07 22:27:40 +00:00
arr
5bf62873eb - In thread_userret(), remove the Giant locking and unlocking around the
call to thread_alloc().

Approved by:	julian
Reviewed by:	jake, jeff
2002-07-01 03:15:16 +00:00
julian
a06b68b34f Add files that are new for KSE. 2002-06-29 07:04:59 +00:00