]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
28 years agoThis commit was generated by cvs2svn to compensate for changes in r14551,
Peter Wemm [Mon, 11 Mar 1996 19:17:01 +0000 (19:17 +0000)]
This commit was generated by cvs2svn to compensate for changes in r14551,
which included commits to RCS files with non-trunk default branches.

28 years agoImport 4.4BSD-Lite2 onto the vendor branch, note that in the kernel, all
Peter Wemm [Mon, 11 Mar 1996 19:17:01 +0000 (19:17 +0000)]
Import 4.4BSD-Lite2 onto the vendor branch, note that in the kernel, all
files are off the vendor branch, so this should not change anything.

A "U" marker generally means that the file was not changed in between
the 4.4Lite and Lite-2 releases, and does not need a merge.  "C" generally
means that there was a change.

28 years agoAdd FAILSAFE option for selecting extra conservativeness when such
Jordan K. Hubbard [Mon, 11 Mar 1996 18:47:39 +0000 (18:47 +0000)]
Add FAILSAFE option for selecting extra conservativeness when such
is more practical (like during installation).  Correspondingly, set the
option by default in GENERIC now.

28 years agoCleaned up uninitialized 'rt' warning properly
Bill Fenner [Mon, 11 Mar 1996 17:11:23 +0000 (17:11 +0000)]
Cleaned up uninitialized 'rt' warning properly
Make a copy of the header of a packet that gets queued due to
 lack of forwarding cache entry, so that nobody else can step
 on it.  Thanks to Mike Karels <karels@bsdi.com> for pointing
 this one out.

28 years agoBig clean-up job. Remove ancient and never-to-be used stuff.
Mark Murray [Mon, 11 Mar 1996 16:17:58 +0000 (16:17 +0000)]
Big clean-up job. Remove ancient and never-to-be used stuff.
The look much more like BSD Makefiles now.

28 years agoChanged socket code to use 4.4BSD queue macros. This includes removing
David Greenman [Mon, 11 Mar 1996 15:37:44 +0000 (15:37 +0000)]
Changed socket code to use 4.4BSD queue macros. This includes removing
the obsolete soqinsque and soqremque functions as well as collapsing
so_q0len and so_qlen into a single queue length of unaccepted connections.
Now the queue of unaccepted & complete connections is checked directly
for queued sockets. The new code should be functionally equivilent to
the old while being substantially faster - especially in cases where
large numbers of connections are often queued for accept (e.g. http).

28 years agoMove or add #include <queue.h> in preparation for upcoming struct socket
David Greenman [Mon, 11 Mar 1996 15:13:58 +0000 (15:13 +0000)]
Move or add #include <queue.h> in preparation for upcoming struct socket
changes.

28 years agoAdd a hint that a make world is done so people don't think that the last
Jordan K. Hubbard [Mon, 11 Mar 1996 14:59:10 +0000 (14:59 +0000)]
Add a hint that a make world is done so people don't think that the last
command (makewhatis) is running forver. :)

28 years agoInclude queue.h in preparation for struct socket changes.
David Greenman [Mon, 11 Mar 1996 13:02:44 +0000 (13:02 +0000)]
Include queue.h in preparation for struct socket changes.

28 years agoMove #include of queue.h before #include of socketvar.h in preparation for
David Greenman [Mon, 11 Mar 1996 13:01:12 +0000 (13:01 +0000)]
Move #include of queue.h before #include of socketvar.h in preparation for
struct socket changes.

28 years agoInclude queue.h in preparation of struct socket changes.
David Greenman [Mon, 11 Mar 1996 13:00:12 +0000 (13:00 +0000)]
Include queue.h in preparation of struct socket changes.

28 years agoInclude queue.h in preparation for upcoming struct socket changes.
David Greenman [Mon, 11 Mar 1996 12:56:07 +0000 (12:56 +0000)]
Include queue.h in preparation for upcoming struct socket changes.

28 years agoerr() --> errx() for non-errno related failures.
Joerg Wunsch [Mon, 11 Mar 1996 11:01:03 +0000 (11:01 +0000)]
err() --> errx() for non-errno related failures.

28 years agoAdd Andrew Gordon for his rpc.statd and stub rpc.lockd
Peter Wemm [Mon, 11 Mar 1996 10:59:00 +0000 (10:59 +0000)]
Add Andrew Gordon for his rpc.statd and stub rpc.lockd

28 years agoBring in most of Bruce's prot_to_real() changes into the APM BIOS
Nate Williams [Mon, 11 Mar 1996 06:48:57 +0000 (06:48 +0000)]
Bring in most of Bruce's prot_to_real() changes into the APM BIOS
startup code.

Reviewed by:  bde

28 years agoFix typo #ifdef -> .if defined().
Mark Murray [Mon, 11 Mar 1996 06:22:50 +0000 (06:22 +0000)]
Fix typo #ifdef -> .if defined().
Tidy uo this file a bit.

28 years agoFrom Lite2: proc LIST changes
Jeffrey Hsu [Mon, 11 Mar 1996 06:14:38 +0000 (06:14 +0000)]
From Lite2: proc LIST changes
call kern_proc.c:procinit() to initialize LISTs
call to usrinfoinit() subsumed by procinit()
Reviewed by: davidg & bde

28 years agoFor Lite2: proc LIST changes.
Jeffrey Hsu [Mon, 11 Mar 1996 06:13:09 +0000 (06:13 +0000)]
For Lite2: proc LIST changes.
Reviewed by: davidg & bde

28 years agoFor Lite2: proc LIST changes.
Jeffrey Hsu [Mon, 11 Mar 1996 06:11:43 +0000 (06:11 +0000)]
For Lite2: proc LIST changes.
Reviewed by:  davidg & bde

28 years agoMerge in Lite2: LIST changes
Jeffrey Hsu [Mon, 11 Mar 1996 06:10:11 +0000 (06:10 +0000)]
Merge in Lite2: LIST changes
stylistic changes to function prototypes
Note: sizeof(struct proc) is now exactly 256 bytes, no room to spare.

28 years agoFrom Lite2: proc LIST changes.
Jeffrey Hsu [Mon, 11 Mar 1996 06:05:03 +0000 (06:05 +0000)]
From Lite2: proc LIST changes.
Reviewed by: david & bde

28 years agoFrom Lite2: proc LIST changes
Jeffrey Hsu [Mon, 11 Mar 1996 06:04:20 +0000 (06:04 +0000)]
From Lite2: proc LIST changes
stylistic changes to function prototypes
Reviewed by: david & bde

28 years agoFor Lite2: proc LIST changes.
Jeffrey Hsu [Mon, 11 Mar 1996 05:55:56 +0000 (05:55 +0000)]
For Lite2: proc LIST changes.
Reviewed by: david & bde

28 years agoMerge in Lite2: proc LIST changes.
Jeffrey Hsu [Mon, 11 Mar 1996 05:52:50 +0000 (05:52 +0000)]
Merge in Lite2: proc LIST changes.
Reviewed by: david & bde

28 years agoMerge in Lite2: proc LIST changes
Jeffrey Hsu [Mon, 11 Mar 1996 05:48:57 +0000 (05:48 +0000)]
Merge in Lite2: proc LIST changes
64-bit fix for alpha
add debugging code for locking
Reviewed by: david & bde

28 years agoFrom Lite2: file LIST changes.
Jeffrey Hsu [Mon, 11 Mar 1996 05:41:43 +0000 (05:41 +0000)]
From Lite2: file LIST changes.

28 years agoFrom Lite2: proc and file LIST changes
Jeffrey Hsu [Mon, 11 Mar 1996 05:34:46 +0000 (05:34 +0000)]
From Lite2: proc and file LIST changes

28 years agoFix some grammer problems.
Mike Pritchard [Mon, 11 Mar 1996 04:50:13 +0000 (04:50 +0000)]
Fix some grammer problems.

28 years agoAdd the -x option to the usage string.
Mike Pritchard [Mon, 11 Mar 1996 03:36:10 +0000 (03:36 +0000)]
Add the -x option to the usage string.

Submitted by: Philippe Charnier <charnier@lirmm.fr>

28 years agoFix a couple of typos.
Mike Pritchard [Mon, 11 Mar 1996 03:31:51 +0000 (03:31 +0000)]
Fix a couple of typos.

Submitted by: Philippe Charnier <charnier@limrr.fr>

28 years agoAdd a reference to intro(9).
Mike Pritchard [Mon, 11 Mar 1996 03:28:43 +0000 (03:28 +0000)]
Add a reference to intro(9).

Submitted by: Philippe Charnier <charnier@lirmm.fr>

28 years agoChange an incorrect reference to the "-n" flag to "-#".
Mike Pritchard [Mon, 11 Mar 1996 03:09:00 +0000 (03:09 +0000)]
Change an incorrect reference to the "-n" flag to "-#".
Submitted by: Masanobu Saitoh <msaitoh@spa.is.uec.ac.jp>
Obtained from: NetBSD-bugs mailing list

28 years agoFrom Lite2: rename fs to vfs.
Jeffrey Hsu [Mon, 11 Mar 1996 03:08:51 +0000 (03:08 +0000)]
From Lite2: rename fs to vfs.

28 years agoFrom Lite2: rename fs to vfs.
Jeffrey Hsu [Mon, 11 Mar 1996 03:06:45 +0000 (03:06 +0000)]
From Lite2: rename fs to vfs.

28 years agoBe more careful in how we set SCSICONF. It was possible for a controller
Justin T. Gibbs [Mon, 11 Mar 1996 02:49:48 +0000 (02:49 +0000)]
Be more careful in how we set SCSICONF.  It was possible for a controller
set to a device id above 7 to screw up the data fifo setting.

28 years agoDisable the selection hardware before acknowledging a SELTO interrupt.
Justin T. Gibbs [Mon, 11 Mar 1996 02:48:41 +0000 (02:48 +0000)]
Disable the selection hardware before acknowledging a SELTO interrupt.
It seems the 7880 gets confused if you don't do this.

28 years agoFrom Lite2: moved acct_process() prototype over to acct.h.
Jeffrey Hsu [Mon, 11 Mar 1996 02:27:53 +0000 (02:27 +0000)]
From Lite2: moved acct_process() prototype over to acct.h.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: add whiteout flags to namei parameter descriptors
Jeffrey Hsu [Mon, 11 Mar 1996 02:25:04 +0000 (02:25 +0000)]
Merge in Lite2: add whiteout flags to namei parameter descriptors
add function prototype for relookup()
Spruce up number formatting.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: buf fix with comparison of wrong field against EXTPROC.
Jeffrey Hsu [Mon, 11 Mar 1996 02:24:39 +0000 (02:24 +0000)]
Merge in Lite2: buf fix with comparison of wrong field against EXTPROC.
Did not accept change of second argument to ioctl from int to u_long.
Reviewed by: davidg & bde

28 years agoFrom NetBSD: add #include <sys/acct.h> for acct_process() prototype.
Jeffrey Hsu [Mon, 11 Mar 1996 02:24:21 +0000 (02:24 +0000)]
From NetBSD: add #include <sys/acct.h> for acct_process() prototype.
Reviewed by: davidg & bde

28 years agoFrom Lite2: add typedef for register_t.
Jeffrey Hsu [Mon, 11 Mar 1996 02:23:56 +0000 (02:23 +0000)]
From Lite2: add typedef for register_t.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: add function prototype.
Jeffrey Hsu [Mon, 11 Mar 1996 02:23:33 +0000 (02:23 +0000)]
Merge in Lite2: add function prototype.
Note: Lite2 struct sysent changes need to go in FreeBSD sysent.h instead.
Leave out prototype for nosys() until convert makesyscalls.sh
  over to Lite2 format, if we ever do so.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use machine-independent types
Jeffrey Hsu [Mon, 11 Mar 1996 02:22:58 +0000 (02:22 +0000)]
Merge in Lite2: use machine-independent types
add kernel function prototype
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use machine-independent types.
Jeffrey Hsu [Mon, 11 Mar 1996 02:22:40 +0000 (02:22 +0000)]
Merge in Lite2: use machine-independent types.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: clean up function prototypes.
Jeffrey Hsu [Mon, 11 Mar 1996 02:22:23 +0000 (02:22 +0000)]
Merge in Lite2: clean up function prototypes.
Did not accept change of second argument to ioctl from int to u_long.
Reviewed by: davidg & bde

28 years agoFrom Lite2: change code parameter to u_long and initialize ps_sig.
Jeffrey Hsu [Mon, 11 Mar 1996 02:22:02 +0000 (02:22 +0000)]
From Lite2: change code parameter to u_long and initialize ps_sig.
Reviewed by: davidg & bde

28 years agoChange type of code argument to sendsig from unsigned to u_long to make it
Jeffrey Hsu [Mon, 11 Mar 1996 02:21:42 +0000 (02:21 +0000)]
Change type of code argument to sendsig from unsigned to u_long to make it
  consistent w/ signalvar.h and kern_sig.c.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: add function prototypes
Jeffrey Hsu [Mon, 11 Mar 1996 02:21:04 +0000 (02:21 +0000)]
Merge in Lite2: add function prototypes
remove bogus function prototype for issig()---no such function
fix comment
Lite2 changed type of ps_code from int to long.  We change it to u_long
  to make it consistent w/ its usage in kern_sig.c.
Did not accept type change to ps_addr field.  Delete it instead as it is
  not used anywhere.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: cosmetic indentation change.
Jeffrey Hsu [Mon, 11 Mar 1996 02:20:13 +0000 (02:20 +0000)]
Merge in Lite2: cosmetic indentation change.
We already have the other Lite2 changes, which consists of additional
  function prototypes.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types.
Jeffrey Hsu [Mon, 11 Mar 1996 02:19:02 +0000 (02:19 +0000)]
Merge in Lite2: use fixed-width types.
Did not accept changes to types of off_t and pid_t.  See
  <machine/ansi.h> for explanation.
Did not accept changes to major and minor macros upon bde's review.
  These changes are unnecessary.
Did not accept type change to fd_mask upon bde's review.
  This change is unnecessary as long as sizeof(fd_mask) divides FD_SETSIZE.
Reviewed by: davidg & bde

28 years agoFrom Lite2: rename fs to vfs.
Jeffrey Hsu [Mon, 11 Mar 1996 02:18:23 +0000 (02:18 +0000)]
From Lite2: rename fs to vfs.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: rename fs to vfs.
Jeffrey Hsu [Mon, 11 Mar 1996 02:17:53 +0000 (02:17 +0000)]
Merge in Lite2: rename fs to vfs.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: LIST replacement for f_filef, f_fileb, and filehead.
Jeffrey Hsu [Mon, 11 Mar 1996 02:17:30 +0000 (02:17 +0000)]
Merge in Lite2: LIST replacement for f_filef, f_fileb, and filehead.
Did not accept change of second argument to ioctl from int to u_long.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: LIST replacement for f_filef, f_fileb, and filehead.
Jeffrey Hsu [Mon, 11 Mar 1996 02:17:11 +0000 (02:17 +0000)]
Merge in Lite2: LIST replacement for f_filef, f_fileb, and filehead.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: no changes required. Just update CSRG id.
Jeffrey Hsu [Mon, 11 Mar 1996 02:16:19 +0000 (02:16 +0000)]
Merge in Lite2: no changes required. Just update CSRG id.
Our changes since Lite1 include all the added functionality of Lite2.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types.
Jeffrey Hsu [Mon, 11 Mar 1996 02:15:32 +0000 (02:15 +0000)]
Merge in Lite2: use fixed-width types.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types
Jeffrey Hsu [Mon, 11 Mar 1996 02:15:04 +0000 (02:15 +0000)]
Merge in Lite2: use fixed-width types
add POSIX compatibility
add union fs defines
Reviewed by: davidg & bde

28 years agoMerge in Lite2: fix comment
Jeffrey Hsu [Mon, 11 Mar 1996 02:14:38 +0000 (02:14 +0000)]
Merge in Lite2: fix comment
optimize TAILQ_INSERT_HEAD
Reviewed by: davidg & bde

28 years agoMerge in Lite2: clean up casts for architectures where the size of addresses
Jeffrey Hsu [Mon, 11 Mar 1996 02:14:16 +0000 (02:14 +0000)]
Merge in Lite2: clean up casts for architectures where the size of addresses
are bigger than the size of ints
add function prototypes
reorder some fields
Did not add prototype for m_reclaim().  It's only used in kern/uipc_mbuf.c and
  is declared static there.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: stylistic changes to function prototypes
Jeffrey Hsu [Mon, 11 Mar 1996 02:13:32 +0000 (02:13 +0000)]
Merge in Lite2: stylistic changes to function prototypes
add comments
Did not accept change of second argument of ioctl prototype from int to u_long.
Did not merge in changes to fields in bdevsw and cdevsw.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: add function prototypes.
Jeffrey Hsu [Mon, 11 Mar 1996 02:12:57 +0000 (02:12 +0000)]
Merge in Lite2: add function prototypes.
Did not accept prototypes for unp_attach(), unp_bind(), unp_connect(),
  unp_detach(), unp_discard(), unp_disconnect(), unp_drop(),
  np_gc(), unp_mark(), unp_scan(), and unp_shutdown().
They are used only in uipc_usrreq.c and declared static there.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: change manifest constant values for NOCRED and FSCRED
Jeffrey Hsu [Mon, 11 Mar 1996 02:11:48 +0000 (02:11 +0000)]
Merge in Lite2: change manifest constant values for NOCRED and FSCRED
stylistic changes to function prototypes
Reviewed by: davidg & bde

28 years agoMerge in Lite2: make timespec struct POSIX.4 compliant
Jeffrey Hsu [Mon, 11 Mar 1996 02:11:25 +0000 (02:11 +0000)]
Merge in Lite2: make timespec struct POSIX.4 compliant
stylistic changes to function prototypes
Reviewed by: davidg & bde

28 years agoMerge in Lite2: no changes necessary. Just update CSRG id.
Jeffrey Hsu [Mon, 11 Mar 1996 02:10:41 +0000 (02:10 +0000)]
Merge in Lite2: no changes necessary.  Just update CSRG id.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types.
Jeffrey Hsu [Mon, 11 Mar 1996 02:10:14 +0000 (02:10 +0000)]
Merge in Lite2: use fixed-width types.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: remove u_long cast from B_MAGICMASK and B_DEVMAGIC
Jeffrey Hsu [Mon, 11 Mar 1996 02:09:55 +0000 (02:09 +0000)]
Merge in Lite2: remove u_long cast from B_MAGICMASK and B_DEVMAGIC
cosmetic changes to comments and number formatting
Reviewed by: davidg & bde

28 years agoMerge in Lite2: bump BSD date from 199306 to 199506.
Jeffrey Hsu [Mon, 11 Mar 1996 02:09:24 +0000 (02:09 +0000)]
Merge in Lite2: bump BSD date from 199306 to 199506.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: sync up comments.
Jeffrey Hsu [Mon, 11 Mar 1996 02:09:09 +0000 (02:09 +0000)]
Merge in Lite2: sync up comments.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: resolve manifest constant numbering in favor of least
Jeffrey Hsu [Mon, 11 Mar 1996 02:08:50 +0000 (02:08 +0000)]
Merge in Lite2:  resolve manifest constant numbering in favor of least
  required changes and also add parentheses around BUCKETINDX macro body.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: only update CSRG id---the one change made in Lite2
Jeffrey Hsu [Mon, 11 Mar 1996 02:08:16 +0000 (02:08 +0000)]
Merge in Lite2: only update CSRG id---the one change made in Lite2
  to move the key_t typedef to types.h has already been made in a
    previous commit.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types.
Jeffrey Hsu [Mon, 11 Mar 1996 02:07:37 +0000 (02:07 +0000)]
Merge in Lite2: use fixed-width types.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types and reformat comments.
Jeffrey Hsu [Mon, 11 Mar 1996 02:06:30 +0000 (02:06 +0000)]
Merge in Lite2: use fixed-width types and reformat comments.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: use fixed-width types and add whiteout file type.
Jeffrey Hsu [Mon, 11 Mar 1996 02:05:52 +0000 (02:05 +0000)]
Merge in Lite2: use fixed-width types and add whiteout file type.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: rename B_APPENDWRITE to B_NEEDCOMMIT.
Jeffrey Hsu [Mon, 11 Mar 1996 02:04:27 +0000 (02:04 +0000)]
Merge in Lite2: rename B_APPENDWRITE to B_NEEDCOMMIT.
The other change to b_pfcent is no longer pertinent, since we've
  deleted that field.
Reviewed by: davidg & bde

28 years agoMerge in Lite2: fix typo in comment.
Jeffrey Hsu [Mon, 11 Mar 1996 02:03:11 +0000 (02:03 +0000)]
Merge in Lite2: fix typo in comment.
Reviewed by: davidg & bde

28 years agoOnly restrict the user from doing something like "mount /mnt /mnt"
Mike Pritchard [Mon, 11 Mar 1996 00:22:28 +0000 (00:22 +0000)]
Only restrict the user from doing something like "mount /mnt /mnt"
for file system types that actually cause a panic (ufs, msdos, cd9660).
This makes /proc mountable again.

28 years agoTweak the data/bss segment page count. The last version worked
Peter Wemm [Sun, 10 Mar 1996 23:44:40 +0000 (23:44 +0000)]
Tweak the data/bss segment page count. The last version worked
with all the test cases I tried, I'm sure this is more correct.

Tweak some prototypes.

28 years agoPrint 8bit chars from the net if current locale allows them
Andrey A. Chernov [Sun, 10 Mar 1996 23:33:48 +0000 (23:33 +0000)]
Print 8bit chars from the net if current locale allows them

28 years agoFix the vm_map_remove and vm_map_protect calls.. Somewhere along the
Peter Wemm [Sun, 10 Mar 1996 23:25:17 +0000 (23:25 +0000)]
Fix the vm_map_remove and vm_map_protect calls..  Somewhere along the
line, these had got (start, length) arguments instead of (start, end)
args.  This could be the cause of Robert Sanders lockups with ZMAGIC
binaries.

28 years agoImproved efficiency in pmap_remove, and also remove some of the pmap_update
John Dyson [Sun, 10 Mar 1996 23:06:00 +0000 (23:06 +0000)]
Improved efficiency in pmap_remove, and also remove some of the pmap_update
optimizations that were probably incorrect.

28 years agolinux_interp now moved into sys/i386/linux/linux_sysvec.c, but still
Peter Wemm [Sun, 10 Mar 1996 22:43:37 +0000 (22:43 +0000)]
linux_interp now moved into sys/i386/linux/linux_sysvec.c, but still
connected at load and unload time from here.

28 years agoMove the elf linux_interp structure from the lkm stub to here so that it
Peter Wemm [Sun, 10 Mar 1996 22:42:16 +0000 (22:42 +0000)]
Move the elf linux_interp structure from the lkm stub to here so that it
can be referenced when the linux emulator is statically compiled into the
kernel.

28 years agoFix some rounding problems.. In some (fairly rare) situtaions it mapped
Peter Wemm [Sun, 10 Mar 1996 22:37:34 +0000 (22:37 +0000)]
Fix some rounding problems..  In some (fairly rare) situtaions it mapped
one page too many, which caused obreak() to fail in vm_map_find() with
ENOMEM because of the conflicting page.

28 years agoImplement rudumentry support for the linux TIOC[SG]ETSERIAL ioctl's.
Peter Wemm [Sun, 10 Mar 1996 22:30:53 +0000 (22:30 +0000)]
Implement rudumentry support for the linux TIOC[SG]ETSERIAL ioctl's.
To complete this, some extra state has to be kept somewhere so that the
B38400 flag in Linux can be correctly translated to/from either 38400,
57600 or 115200.

Submitted by: Robert Sanders <rsanders@mindspring.com>

28 years agoFix the getdents() emulation, the Linux ELF libraries use this, and
Peter Wemm [Sun, 10 Mar 1996 22:27:51 +0000 (22:27 +0000)]
Fix the getdents() emulation, the Linux ELF libraries use this, and
this code was not quite right (linux has a readdir and getdents syscall,
with the same args. readdir only returns one entry and uses a mutant
dirent structure.  This code was also returning the mutant form for
getdents as well.  My fault for missing this before.)

28 years agoFix a (mostly harmless) bogon when calculating the start of the stack gap.
Peter Wemm [Sun, 10 Mar 1996 22:24:54 +0000 (22:24 +0000)]
Fix a (mostly harmless) bogon when calculating the start of the stack gap.

28 years agoFix a (mostly harmless) bogon when allocating space above the stack
Peter Wemm [Sun, 10 Mar 1996 22:23:51 +0000 (22:23 +0000)]
Fix a (mostly harmless) bogon when allocating space above the stack
in the stack gap..

28 years agoFix typo - -des -> -ldes
Mark Murray [Sun, 10 Mar 1996 21:15:00 +0000 (21:15 +0000)]
Fix typo - -des -> -ldes

28 years agoremove ``whatis''
Wolfram Schneider [Sun, 10 Mar 1996 18:59:11 +0000 (18:59 +0000)]
remove ``whatis''

28 years agoshare the code between apropos and whatis
Wolfram Schneider [Sun, 10 Mar 1996 18:52:33 +0000 (18:52 +0000)]
share the code between apropos and whatis

28 years agoMore missing int32 -> int32_t...
David Greenman [Sun, 10 Mar 1996 18:33:43 +0000 (18:33 +0000)]
More missing int32 -> int32_t...

28 years agoFixed some missing int32 -> int32_t
David Greenman [Sun, 10 Mar 1996 18:17:54 +0000 (18:17 +0000)]
Fixed some missing int32 -> int32_t

28 years agou_int8 -> u_int8_t
Jordan K. Hubbard [Sun, 10 Mar 1996 12:52:47 +0000 (12:52 +0000)]
u_int8 -> u_int8_t

28 years agoFirst attempt at FreeBSD & Linux ELF support.
Søren Schmidt [Sun, 10 Mar 1996 08:42:54 +0000 (08:42 +0000)]
First attempt at FreeBSD & Linux ELF support.

Compile and link a new kernel, that will give native ELF support, and
provide the hooks for other ELF interpreters as well.

To make native ELF binaries use John Polstras elf-kit-1.0.1..
For the time being also use his ld-elf.so.1 and put it in
/usr/libexec.

The Linux emulator has been enhanced to also run ELF binaries, it
is however in its very first incarnation.
Just get some Linux ELF libs (Slackware-3.0) and put them in the
prober place (/compat/linux/...).
I've ben able to run all the Slackware-3.0 binaries I've tried
so far.
(No it won't run quake yet :)

28 years agoReplace PATCH_PRFX with a more generic DIST_SUBDIR. If it is set to
Satoshi Asami [Sun, 10 Mar 1996 08:26:21 +0000 (08:26 +0000)]
Replace PATCH_PRFX with a more generic DIST_SUBDIR.  If it is set to
"foo", what this does is:

(1) Put all distfiles and patchfiles in /usr/ports/distfiles/foo

(2) Go to ftp://ftp.freebsd.org/pub/FreeBSD/distfiles/foo when the
    master site is down

When your port has a lot of dist/patchfiles, or has a file that does
not have a very port-specific name (e.g., "Makefile"), set this
variable instead of redefining DISTDIR.  (If you redefine DISTDIR, (1)
will work but (2) will not.)

Agreed that it's a good idea by: adam

28 years agoCleanse the SCSI subsystem of its internally defined types
Justin T. Gibbs [Sun, 10 Mar 1996 07:13:15 +0000 (07:13 +0000)]
Cleanse the SCSI subsystem of its internally defined types
u_int32, u_int16, u_int8, int32, int16, int8.
Use the system defined *_t types instead.

28 years agoaic7870.c, ncr.c:
Justin T. Gibbs [Sun, 10 Mar 1996 07:12:49 +0000 (07:12 +0000)]
aic7870.c, ncr.c:
Cleanse the SCSI subsystem of its internally defined types
u_int32, u_int16, u_int8, int32, int16, int8.
Use the system defined *_t types instead.

aic7870.c:
Handle Seeprom data a little better.

28 years agobt.c, aic7xxx.c:
Justin T. Gibbs [Sun, 10 Mar 1996 07:11:45 +0000 (07:11 +0000)]
bt.c, aic7xxx.c:
Cleanse the SCSI subsystem of its internally defined types
u_int32, u_int16, u_int8, int32, int16, int8.
Use the system defined *_t types instead.

aic7xxx.c:
Fix the reset code.
Instead of queing up all of the SCBs that timeout during timeout
processing, we take the first and have it champion the effort.
Any other scbs that timeout during timeout handling are given
another lifetime to complete in the hopes that once timeout
handing is finished, they will complete normally.  If one of
these SCBs times out a second time, we panic and Justin tries
again.

The other major change is to queue flag aborted SCBs during timeout
handling, and "ahc_done" them all at once as soon as we have the
controller back into a sane state.  Calling ahc_done any earlier
will cause the SCSI subsystem to toss the command right back at
us and the attempt to queue the command will conflict with what
the timeout routine is trying to accomplish.

The aic7xxx driver will now respond to bus resets initiated by
other devices.

28 years agoCleanse the SCSI subsystem of its internally defined types
Justin T. Gibbs [Sun, 10 Mar 1996 07:04:48 +0000 (07:04 +0000)]
Cleanse the SCSI subsystem of its internally defined types
u_int32, u_int16, u_int8, int32, int16, int8.
Use the system defined *_t types instead.

28 years agoaha1742.c:
Justin T. Gibbs [Sun, 10 Mar 1996 07:04:27 +0000 (07:04 +0000)]
aha1742.c:
Cleanse the SCSI subsystem of its internally defined types
u_int32, u_int16, u_int8, int32, int16, int8.
Use the system defined *_t types instead.

eisaconf.c:
Cosmetic formatting chagnes.