Commit Graph

5315 Commits

Author SHA1 Message Date
Poul-Henning Kamp
e46eeb89b9 Do not employ timecounter hardware if our hz does not support their
correct rewinding.
2002-09-04 19:32:18 +00:00
Poul-Henning Kamp
e7fa55af89 Give up on calling tc_ticktock() from a timeout, we have timeout
functions which run for several milliseconds at a time and getting
in queue behind one or more of those makes us miss our rewind.

Instead call it from hardclock() like we used to do, but retain the
prescaler so we still cope with high HZ values.
2002-09-04 10:15:19 +00:00
Matthew Dillon
21c2d0479c Alright, fix the problems with the elf loader for the Alpha. It turns
out that there is no easy way to discern the difference between a text
segment and a data segment through the read-only OR execute attribute
in the elf segment header, so revert the algorithm to what it was before.

Neither can we account for multiple data load segments in the vmspace
structure (at least not without more work), due to assumptions obreak()
makes in regards to the data start and data size fields.

Retain RLIMIT_VMEM checking by using a local variable to track the
total bytes of data being loaded.

Reviewed by:	peter
X-MFC after:	ASAP
2002-09-04 04:42:12 +00:00
Peter Wemm
9782ecbab0 Make the text segment locating heuristics from rev 1.121 more reliable
so that it works on the Alpha.  This defines the segment that the entry
point exists in as 'text' and any others (usually one) as data.

Submitted by: tmm
Tested on: i386, alpha
2002-09-03 21:18:17 +00:00
John Baldwin
5fc3031366 - Change falloc() to acquire an fd from the process table last so that
it can do it w/o needing to hold the filelist_lock sx lock.
- fdalloc() doesn't need Giant to call free() anymore.  It also doesn't
  need to drop and reacquire the filedesc lock around free() now as a
  result.
- Try to make the code that copies fd tables when extending the fd table in
  fdalloc() a bit more readable by performing assignments in separate
  statements.  This is still a bit ugly though.
- Use max() instead of an if statement so to figure out the starting point
  in the search-for-a-free-fd loop in fdalloc() so it reads better next to
  the min() in the previous line.
- Don't grow nfiles in steps up to the size needed if we dup2() to some
  really large number.  Go ahead and double 'nfiles' in a loop prior
  to doing the malloc().
- malloc() doesn't need Giant now.
- Use malloc() and free() instead of MALLOC() and FREE() in fdalloc().
- Check to see if the size we are going to grow to is too big, not if the
  current size of the fd table is too big in the loop in fdalloc().  This
  means if we are out of space or if dup2() requests too high of a fd,
  then we will return an error before we go off and try to allocate some
  huge table and copy the existing table into it.
- Move all of the logic for dup'ing a file descriptor into do_dup() instead
  of putting some of it in do_dup() and duplicating other parts in four
  different places.  This makes dup(), dup2(), and fcntl(F_DUPFD) basically
  wrappers of do_dup now.  fcntl() still has an extra check since it uses
  a different error return value in one case then the other functions.
- Add a KASSERT() for an assertion that may not always be true where the
  fdcheckstd() function assumes that falloc() returns the fd requested and
  not some other fd.  I think that the assertion is always true because we
  are always single-threaded when we get to this point, but if one was
  using rfork() and another process sharing the fd table were playing with
  the fd table, there might could be a problem.
- To handle the problem of a file descriptor we are dup()'ing being closed
  out from under us in dup() in general, do_dup() now obtains a reference
  on the file in question before calling fdalloc().  If after the call to
  fdalloc() the file for the fd we are dup'ing is a different file, then
  we drop our reference on the original file and return EBADF.  This
  race was only handled in the dup2() case before and would just retry
  the operation.  The error return allows the user to know they are being
  stupid since they have a locking bug in their app instead of dup'ing
  some other descriptor and returning it to them.

Tested on:	i386, alpha, sparc64
2002-09-03 20:16:31 +00:00
John Baldwin
0d975d6341 Add some KASSERT()'s to ensure that we don't perform spin mutex ops on
sleep mutexes and vice versa.  WITNESS normally should catch this but
not everyone uses WITNESS so this is a fallback to catch nasty but easy
to do bugs.
2002-09-03 18:25:16 +00:00
David Xu
35c32a76f9 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
Ian Dowse
48b52b7a32 Split up __getcwd so that kernel callers of the internal version
can specify whether the buffer is in user or system space.
2002-09-02 22:40:30 +00:00
Ian Dowse
49c2ff159f Split fcntl() into a wrapper and a kernel-callable kern_fcntl()
implementation. The wrapper is responsible for copying additional
structure arguments (struct flock) to and from userland.
2002-09-02 22:24:14 +00:00
Matthew Dillon
05ef87980a Grammer cleanup 2002-09-02 17:27:30 +00:00
David Xu
67bdda9718 fix bogus CTR3 message.
Reviewed by: julian@freebsd.org (mentor)
2002-09-02 07:55:06 +00:00
Jake Burkholder
5fe3ed629a Moved elf brand identification into a function. Fully identify the
brand early in the process of loading an elf file, so that we can
identify the sysentvec, and so that we do not continue if we do not
have a brand (and thus a sysentvec).  Use the values in the sysentvec
for the page size and vm ranges unconditionally, since they are all
filled in now.
2002-09-02 04:50:57 +00:00
Alan Cox
8a59b15cd4 o Synchronize updates to struct vm_page::cow with the page queues lock. 2002-09-02 04:04:12 +00:00
Jake Burkholder
8cf034521b Fixed more indentation bugs. 2002-09-02 02:41:26 +00:00
Jake Burkholder
f36ba45234 Added fields for VM_MIN_ADDRESS, PS_STRINGS and stack protections to
sysentvec.  Initialized all fields of all sysentvecs, which will allow
them to be used instead of constants in more places.  Provided stack
fixup routines for emulations that previously used the default.
2002-09-01 21:41:24 +00:00
Ian Dowse
8f19eb88df Split out a number of mostly VFS and signal related syscalls into
a kernel-internal kern_*() version and a wrapper that is called via
the syscall vector table. For paths and structure pointers, the
internal version either takes a uio_seg parameter or requires the
caller to copyin() the data to kernel memory as appropiate. This
will permit emulation layers to use these syscalls without having
to copy out translated arguments to the stack gap.

Discussed on:		-arch
Review/suggestions:	bde, jhb, peter, marcel
2002-09-01 20:37:28 +00:00
Matthew Dillon
cac4515267 Implement data, text, and vmem limit checking in the elf loader and svr4
compat code.  Clean up accounting for multiple segments.  Part 1/2.

Submitted by:	Andrey Alekseyev <uitm@zenon.net> (with some modifications)
MFC after:	3 days
2002-08-30 18:09:46 +00:00
Peter Wemm
447b3772dc Change hw.physmem and hw.usermem to unsigned long like they used to be
in the original hardwired sysctl implementation.

The buf size calculator still overflows an integer on machines with large
KVA (eg: ia64) where the number of pages does not fit into an int.  Use
'long' there.

Change Maxmem and physmem and related variables to 'long', mostly for
completeness.  Machines are not likely to overflow 'int' pages in the
near term, but then again, 640K ought to be enough for anybody.  This
comes for free on 32 bit machines, so why not?
2002-08-30 04:04:37 +00:00
Julian Elischer
472be95807 Rejig the code to figure out estcpu and work out how long a KSEGRP has been
idle. What was there before was surprisingly ALMOST correct.

Peter and I fried our brains on this for a couple of hours figuring out
what this actually means in the context of multiple threads.

Reviewed by:	peter@freebsd.org
2002-08-30 00:25:49 +00:00
Peter Wemm
ee92a1ab51 Actually remove the a.out kld loader. While I am not 100% sure, I believe
it is broken.  It certainly has been suffering neglect.  It is not needed
because we never shipped a.out kld's and they never really worked right.
2002-08-29 23:04:05 +00:00
Julian Elischer
88151aa3f5 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
Mitsuru IWASAKI
3aea1e1405 Add sanity check seeing if adjusted start address exceeds end address
after boundary and alignment adjustment.
2002-08-29 12:39:21 +00:00
Jake Burkholder
bafbd49201 Renamed poorly named setregs to exec_setregs. Moved its prototype to
imgact.h with the other exec support functions.
2002-08-29 06:17:48 +00:00
Jake Burkholder
f3bec5d746 Don't require that sysentvec.sv_szsigcode be non-NULL. 2002-08-29 01:28:27 +00:00
Jake Burkholder
b17c50db93 Unrot SPARSE_MAPPING code (vm_map_pageable -> vm_map_wire). 2002-08-29 01:16:14 +00:00
Peter Wemm
d13947c3b0 updatepri() works on a ksegrp (where the scheduling parameters are), so
directly give it the ksegrp instead of the thread.  The only thing it used
to use in the thread was the ksegrp.

Reviewed by:	julian
2002-08-28 23:45:15 +00:00
Archie Cobbs
f2f03122c3 accept(2) on a socket that has been shutdown(2) normally returns
ECONNABORTED. Make this happen in the non-blocking case as well.
The previous behavior was to return EAGAIN, which (a) is not
consistent with the blocking case and (b) causes the application
to think the socket is still valid.

PR:		bin/42100
Reviewed by:	freebsd-net
MFC after:	3 days
2002-08-28 20:56:01 +00:00
Bruce Evans
8302d183f3 Include <sys/lockmgr.h> for the definitions of the locking interfaces that
are implemented here instead of depending on namespace pollution in
<sys/lock.h>.  Fixed nearby include messes (1 disordered include and 1
unused include).
2002-08-27 09:59:47 +00:00
Ian Dowse
02bd1bcd2a Add a new KTR type KTR_CONTENTION, and use it in the mutex code to
log the start and end of periods during which mtx_lock() is waiting
to acquire a sleep mutex. The log message includes the file and
line of both the waiter and the holder.

Reviewed by:	jhb, jake
2002-08-26 18:39:38 +00:00
Ian Dowse
9261400aa2 Add WITNESS_FILE() and WITNESS_LINE(), which allow users of witness
to print out the file and line from the lock object. These will be
used shortly by CTR() calls in the mutex code.

Reviewed by:	jhb, jake
2002-08-26 18:31:26 +00:00
Julian Elischer
b39f32841b move the assert to cover more cases 2002-08-26 05:02:56 +00:00
Jake Burkholder
81f223ca02 Fixed most indentation bugs. 2002-08-25 22:36:52 +00:00
Jake Burkholder
ca0387ef9f Fixed placement of operators. Wrapped long lines. 2002-08-25 20:48:45 +00:00
Philippe Charnier
93b0017f88 Replace various spelling with FALLTHROUGH which is lint()able 2002-08-25 13:23:09 +00:00
Jake Burkholder
fd559a8a39 Fixed white space around operators, casts and reserved words.
Reviewed by:	md5
2002-08-24 22:55:16 +00:00
Jake Burkholder
a7cddfed7f return x; -> return (x);
return(x); -> return (x);

Reviewed by:	md5
2002-08-24 22:01:40 +00:00
Marcel Moolenaar
5cf8741861 Work around a GCC optimization bug on ia64: In link_elf_symbol_values(),
a pointer to a symbol is given and we have to find the containing symbol
table. We do this by bounds checking. For some strange reason (ie I
haven't found the root cause) the first test succeeded for said symbol,
implying that the symbol came from the .dynsym table. In reality however
the symbol actually resided in the .symtab table. Needless to say that
all that was returned was junk.

The upper bounds check was: (symptr - baseptr) < symtab_size
This has been rewritten to: symptr < (baseptr + symtab_size)

As a side-effect, slightly more optimal (and still correct :-) code can
be generated on ia64.
2002-08-24 05:01:33 +00:00
Peter Wemm
2149c527f5 Move the TAILQ_INIT(&td->td_selq) before the retry: label. Otherwise in
some circumstances when we get a select collision, we can end up with
cases where we do not clear some sip->si_thread on the way out, leading to
page faults in selwakeup().  This should solve the problem where postfix
can crash the kernel during select collisions.

Reviewed by: alfred
2002-08-23 22:43:28 +00:00
Julian Elischer
d9d6e34fd0 Don't re-lock the sched lock if we didn't unlock it.
Original error by: David Xu <bsddiy@yahoo.com>
Fix by:	David Xu <bsddiy@yahoo.com>
Completely failed to spot it: Julian Elischer <julian@freebsd.org>
2002-08-23 07:23:44 +00:00
Jeff Roberson
ad32f726db - Fix a mistake in my last few commits. The PDROP flag stops msleep from
re-acquiring the mutex.

Pointy hat to:	me
Noticed by:	tegge
2002-08-23 00:32:03 +00:00
Peter Wemm
c6d6cf1772 s/sus/sys/ in the a.out kernel case.
Submitted by:	julian
2002-08-22 22:01:53 +00:00
Julian Elischer
49539972e9 slight cleanup of single-threading code for KSE processes 2002-08-22 21:45:58 +00:00
Archie Cobbs
4a6a94d8d8 Replace (ab)uses of "NULL" where "0" is really meant. 2002-08-22 21:24:01 +00:00
Peter Wemm
3e4517beb6 Instead of grabbing the userland a.out.h/link.h (or worse, from
/usr/include!), use sys/nlist_aout.h, machine/reloc.h, sys/imgact_aout.h
and sys/link_aout.h.
2002-08-22 20:43:07 +00:00
Peter Wemm
f99803876e Instead of nlist.h and link.h, use sys/nlist_aout.h and sys/link_elf.h
This avoids reaching out into userland sources (or worse: /usr/include!)
for building the kernel.
2002-08-22 20:39:30 +00:00
Robert Watson
1c39a77468 Spell proprly properly:
failed to set signal flags proprly for ast()
  failed to set signal flags proprly for ast()
  failed to set signal flags proprly for ast()
  failed to set signal flags proprly for ast()
2002-08-22 14:36:03 +00:00
Bruce Evans
5fd65482e0 Include <sys/systm.h> for the declarations of many things instead of
depending on namespace pollution in <sys/mumble.h>.
2002-08-22 12:47:22 +00:00
Alan Cox
0a179f8025 o Remove the AIOCBLIST_ASYNCFREE flag and related code. It's never set.
Submitted by:	Romer Gil <rgil@cs.rice.edu>
2002-08-22 08:50:15 +00:00
Jeff Roberson
4b6049cafa - Closer inspection revealed a possible deadlock situation in vn_lock() that
was introduced by my last commit but not caught by stress testing.  Fix
   that and slightly restructure the code so that it is more readable.
2002-08-22 07:57:43 +00:00
Jeff Roberson
9abf54f032 - Make vn_lock() vget() and VOP_LOCK() all behave the same way WRT
LK_INTERLOCK.  The interlock will never be held on return from these
   functions even when there is an error.  Errors typically only occur when
   the XLOCK is held which means this isn't the vnode we want anyway.  Almost
   all users of these interfaces expected this behavior even though it was
   not provided before.
2002-08-22 07:44:45 +00:00