1) Create intermediate directories with 755, not 777
2) Exit with error diagnostic if file exists but not a directory (-d)
3) Do chmod independently of chown (-d)
Bruce Evans [Tue, 28 Oct 1997 11:43:57 +0000 (11:43 +0000)]
Don't include <machine/cputypes.h> or declare cputype/class interfaces
in <machine/cpu.h>. Moved the declarations to <machine/cputypes.h>.
Fixed style bugs in the moved code. Fixed everything that depended on
the nested include. Don't include <machine/cpu.h> (in the changed files)
unless something in it is used directly.
Joerg Wunsch [Tue, 28 Oct 1997 07:27:48 +0000 (07:27 +0000)]
Removed another comment about not-so-mandatory devices i've missed in
the previous commit. It's perfectly legal to build a kernel without
any video device driver (and even without any console driver at all if
desired).
1) Ignore -C with -d, not fail (for $(INSTALL) macro from /etc/make.conf)
2) When uid/gid not specified, not try to set 0:0 owner, just do nothing.
It makes possible to use install -d by users without setting
their uid:gid each time.
3) Be more specific where install_dir fails
Bruce Evans [Mon, 27 Oct 1997 17:23:18 +0000 (17:23 +0000)]
Moved declaration of etext from <machine/md_var.h> to <machine/cpu.h>
and fixed everything that dependended on it being declared in the old
place. It is used in "machine-independent" code in subr_prof.c.
Moved declaration of btext from subr_prof.c to <machine/cpu.h>. It
is machine-dependent.
Move nobody to daemon class, otherwise it is impossible to start fingerd
while Apache is running, it effectively eats all default class limits for
nobody
Bruce Evans [Mon, 27 Oct 1997 15:26:23 +0000 (15:26 +0000)]
Use 127 instead of CHAR_MAX for the limit on the sequence count. The
limit doesn't have anything to do with characters. The count mainly
needs to fit in the VOP_READ() ioflag after being left shifted by 16.
Moved vn_lock() before vn_closefile(). vn_lock() was mismerged from
Lite2.
Bruce Evans [Mon, 27 Oct 1997 14:55:49 +0000 (14:55 +0000)]
Use bread() instead of cluster_read() for reading the last block
in a file. There was a (harmless, I think) off-by-1 error. This
was fixed in ufs long ago (rev.1.21 of ufs_readwrite.c) but not
in cd9660.
cd9660_read() has stagnated in many other ways. It is closer to
the Net/2 ufs_read() (which is was cloned from) than ufs_read()
itself is.
Joerg Wunsch [Mon, 27 Oct 1997 06:15:10 +0000 (06:15 +0000)]
Move the EP_ID_PORT out to 0x110, so it doesn't conflict with other port
usage at 0x100. Quoted Justin's quotation from the manual as well, to
explain the technical background.
PR: kern/4559
Submitted by: Stephen J. Roznowski <sjr@home.net>
Steve Price [Mon, 27 Oct 1997 04:22:13 +0000 (04:22 +0000)]
Offending statement removed from STANDARDS section. From what I
can gather from the Posix spec that I have (which is very old)
all one character options are extensions to Posix.
Nate Williams [Sun, 26 Oct 1997 21:08:42 +0000 (21:08 +0000)]
- Instead of relying on a functional call to register PCARD-capable drivers,
use a Linker Set. Note, if a driver is loaded as an LKM if will have
to use the function call, but since none of the existing drivers
are loadable, this made things cleaner and boot messages nicer.
Rename vn_default_error to vop_defaultop all over the place.
Move vn_bwrite from vfs_bio.c to vfs_default.c and call it vop_stdbwrite.
Use vop_null instead of nullop.
Move vop_nopoll from vfs_subr.c to vfs_default.c
Move vop_sharedlock from vfs_subr.c to vfs_default.c
Move vop_nolock from vfs_subr.c to vfs_default.c
Move vop_nounlock from vfs_subr.c to vfs_default.c
Move vop_noislocked from vfs_subr.c to vfs_default.c
Use vop_ebadf instead of *_ebadf.
Add vop_defaultop for getpages on master vnode in MFS.
Always initialize the syscall vectors for our "private" syscalls (not
just in the LKM case).
Plug nqnfs_vop_lease_check directly into the default_vnodeop_p table.
Bill Paul [Sun, 26 Oct 1997 18:47:31 +0000 (18:47 +0000)]
In clntudp_call(), it is possible that xdr_replymsg() might fail
partway through its attempt to decode the result structure sent by
the server. If this happens, it can leave the result partially
populated with dynamically allocated memory. In this event, the
xdr_replymsg() failure is detected and RPC_CANTDECODERES is returned,
but the memory in the partially populated result struct is not
free()d.
The end result is that memory is leaked when an RPC_CANTDECODERES
error occurs. (This condition can occur if a CLIENT * handle is created
using clntudp_bufcreate() with a receive buffer size that is too small
to handle the result sent by the server.)
Fixed by setting reply_xdrs.x_op to XDR_FREE and calling
xdr_replymsg() again to free the memory if an RPC_CANTDECODERES error
is detected.
I suspect that the clnt_tcp.c, clnt_unix.c and clnt_raw.c modules
may ha a similar problem, but I haven't duplicated the condition with
those yet.
John Dyson [Sun, 26 Oct 1997 12:33:39 +0000 (12:33 +0000)]
Change the initial amount of memory allocated for pv_entries to be proportional
to the amount of system memory. Also, clean-up some of the new pv_entry
mgmt code.
Kazutaka YOKOTA [Sun, 26 Oct 1997 07:35:18 +0000 (07:35 +0000)]
- The daemon might go off the screen and crashed the system if the
screen size was changed while the screen saver was inactive. Adjust
the positions of the daemon and the text and clip them accordingly
each time.
- Don't call set_border() too often. Some video chip may produce
flicker.
Pointed out by tony@dell.com
- Don't fill the entire screen with blank char every time the saver is
called. Blank only the part of the screen where the daemon and the
text was previously printed.
Nate Williams [Sun, 26 Oct 1997 06:06:54 +0000 (06:06 +0000)]
- 'Beep' support now happens in it's own separate file, so you can mess
around with different noises for the different events and not have it
affect other files.
Nate Williams [Sun, 26 Oct 1997 04:54:16 +0000 (04:54 +0000)]
- Functional changes to PCCARD support.
* Kill individual drivers 'suspend' routines, since there's no simple/safe
way to suspend/resume a card w/out going through the complete probe
at initialization time.
* Default to using the apm_pccard_resume sysctl code, which basically
pretends the card was removed, and then re-inserted. Suspend/resume
is now 'emulated' with a fake insert/removal. (Hence we no longer
need the driver-specific suspend routines.)
Nate Williams [Sun, 26 Oct 1997 04:36:24 +0000 (04:36 +0000)]
- Do a bunch of gratuitous changes intended to make the code easier to
follow.
* Rename/reorder all of the pccard structures, change many of the member
names to be descriptive, and follow more closely other 'bus' drivers
naming schemes.
* Rename a bunch of parameter and local variable names to be more
consistant in the code.
* Renamed the PCCARD 'crd' device to be the 'card' device
* KNF and make the code consistant where it was obvious.
* ifdef'd out some unused code
Brian Somers [Sun, 26 Oct 1997 01:04:02 +0000 (01:04 +0000)]
Cosmetic (no functional changes):
o Add missing $Id$s
o Move extern decls from .c -> .h files
o Staticize
o Remove #includes from .h files
o style(9)ify includes
o bcopy -> memcpy
bzero -> memset
bcmp -> memcmp
index -> strchr
rindex -> strrchr
o Move timeout.h -> timer.h (making it consistent w/ timer.c)
o Add -Wmissing-prototypes
John Dyson [Sat, 25 Oct 1997 02:41:56 +0000 (02:41 +0000)]
Support garbage collecting the pmap pv entries. The management doesn't
happen until the system would have nearly failed anyway, so no signficant
overhead is added. This helps large systems with lots of processes.
Mark Murray [Fri, 24 Oct 1997 16:35:34 +0000 (16:35 +0000)]
(Null commit - the previous commit blew up in my face).
Link this against -lcrypt. In the case where a user does not have
a ticket available on the server, this may cause chunks to be blown
when the local kinit does not find crypt(3).
Mark Murray [Fri, 24 Oct 1997 16:27:46 +0000 (16:27 +0000)]
Link this against -lcrypt. In the case where the user has no key available
on the keyserver, the local kinit will blow chunks when it tries a
crypt(3) and finds it unavailable.