John Baldwin [Sun, 19 May 2002 00:14:50 +0000 (00:14 +0000)]
Change p_can{debug,see,sched,signal}()'s first argument to be a thread
pointer instead of a proc pointer and require the process pointed to
by the second argument to be locked. We now use the thread ucred reference
for the credential checks in p_can*() as a result. p_canfoo() should now
no longer need Giant.
Robert Watson [Sun, 19 May 2002 00:11:08 +0000 (00:11 +0000)]
Remove IFS from 5.0-CURRENT. This facilitates introducing UFS2 as
IFS had its fingers deep in the belly of the UFS/FFS split. IFS
will be reimplemented by the maintainer at a later date.
Robert Watson [Sun, 19 May 2002 00:09:30 +0000 (00:09 +0000)]
Remove IFS from 5.0-CURRENT. This facilitates introducing UFS2 as
IFS had its fingers deep in the belly of the UFS/FFS split. IFS
will be reimplemented by the maintainer at a later date once the UFS
implementation is in place.
Ian Dowse [Sat, 18 May 2002 22:18:17 +0000 (22:18 +0000)]
Add an ext2_uninit() routine that undoes the actions performed by
ext2_init(). This permits the ext2fs module to be unloaded without
causing panics and leaking memory.
Ian Dowse [Sat, 18 May 2002 21:33:07 +0000 (21:33 +0000)]
Fix two off-by-one errors when sanity-checking inode numbers. In
ext2fs, inode numbers start at 1, so the maximum valid inode number
is (s_inodes_per_group * s_groups_count), not one less. This is
just a minimal change to avoid unnecessary panics and errors; some
other related bugs that Bruce Evans mentioned to me are not addressed.
Ian Dowse [Sat, 18 May 2002 21:10:40 +0000 (21:10 +0000)]
Unconditionally update the maximum field width statistics when we
refetch the filesystem information in MNT_WAIT mode. This avoids
incorrect column alignment that sometimes occurs with NFS filesystems.
Ian Dowse [Sat, 18 May 2002 18:58:27 +0000 (18:58 +0000)]
Fix a typo where sizeof(daddr_t) was specified instead of sizeof(doff_t).
Now that daddr_t is 64-bit, this caused hash blocks to be allocated
twice as large as they need to be.
Bruce A. Mah [Sat, 18 May 2002 18:44:24 +0000 (18:44 +0000)]
New release notes: Perl deletion (along with a caution that some work
is still ongoing), catman(1) rewrite, makewhatis(1) rewrite,
spkrtest(8) rewrite.
Modified release note: killall(1) for markup.
Deleted release notes: Perl upgrade, style.perl(7) manpage, one note
about killall(1) was obsolete.
Improve chances that we correctly compile LIB1ASMSRC on all architectures.
sparc64 looked for the nonexistent sparc64/lb1spc.asm file instead
of the sparc/lb1spc.asm file.
arm probably looked for arm/arm/lib1funcs.asm instead of arm/lib1funcs.asm
ia64 probably looked for ia64/ia64/lib1funcs.asm instead of ia64/lib1funcs.asm
Marcel Moolenaar [Sat, 18 May 2002 07:53:56 +0000 (07:53 +0000)]
In msgrcv(), set msgtyp correctly. Hardwiring 0 as the message type
yields incorrect behaviour. The hardwiring was present in the very
first commit that implemented msgrcv() (revision 1.4) and hasn't been
changed since. The native implementation was complete at that time,
so there doesn't seem to be a reason for the hardwiring from a
technical point of view.
Alan Cox [Sat, 18 May 2002 07:48:06 +0000 (07:48 +0000)]
o Remove GIANT_REQUIRED from vm_map_madvise(). Instead, acquire and
release Giant around vm_map_madvise()'s call to pmap_object_init_pt().
o Replace GIANT_REQUIRED in vm_object_madvise() with the acquisition
and release of Giant.
o Remove the acquisition and release of Giant from madvise().
Gordon Tetlow [Sat, 18 May 2002 00:26:10 +0000 (00:26 +0000)]
Fix the order of shutdown scripts so it not only reverses the order of the
files in each ${local_startup} directory, it also reverses the order of the
directories.
Doug Ambrisko [Fri, 17 May 2002 20:18:48 +0000 (20:18 +0000)]
Add TAG_VENDOR_INDENTIFIER (option 60) to our DHCP request done by the
kernel BOOTP option. The format will be:
FreeBSD:<MACHINE>:<osrelease>
this way people can tune their DHCP server to server up root file systems
via the OS, machine type and version.
Brian Somers [Fri, 17 May 2002 13:38:36 +0000 (13:38 +0000)]
Change ``dmesg -a'' to ``dmesg''.
The change was introduced in src/etc/security 1.53 almost a year ago
in an attempt to see ipfw deny message logs.
However, ipfw deny/reject logs have been displayed since version 1.13
of the same file as a separate ``job'' and have since moved to
src/etc/periodic/security/500.ipfwdenied.
Juli Mallett [Fri, 17 May 2002 12:27:02 +0000 (12:27 +0000)]
Group the output chunks differently, to try to match BSD/OS output. Switch
to using arithmatic to determine buffer sizes to encode into.
Diane Bruce pointed out to me that BSD/OS did MIME too, so I want to match
their output, too, since my decision of 8 output groups was wholly arbitrary.
Fix the case:
cp file nonexistent/
which create nonextstent as file while trailing slash clearly indicates
that nonexistent must be a directory.
Also fix the case:
cp file1 file2/
which should produce error.
Juli Mallett [Fri, 17 May 2002 11:16:49 +0000 (11:16 +0000)]
Revert part of the previous revision. Changing exit() to return() in main()
is wrong, even though our C compiler doesn't understand that exit() eats
control of the program, and as such can sometimes complain that main() reaches
the bottom of its body without an explicit return().
Aside from that, the functional usefulness of changing it is effectively nil,
so back it out.
Peter Wemm [Fri, 17 May 2002 08:59:13 +0000 (08:59 +0000)]
Move LIB1ASMFUNCS from the SYMS variable and explicitly add it to OBJS
later. Otherwise make will try and build the supposedly assembler .o
files from libgcc2.c - which does not work too well (the .o's have no
content)
Juli Mallett [Fri, 17 May 2002 07:55:24 +0000 (07:55 +0000)]
Simplify prerun() in the case where there are no arguments. Can't do -I or -J
substitution if `utility' is not specified and we're using the buildin echo;
It has no arguments for us to abuse.
Peter Wemm [Fri, 17 May 2002 07:55:08 +0000 (07:55 +0000)]
Put on peril-sensitive sunglasses and risk touching gcc related build glue
again. Try and deal with platforms that provide their own crtbegin/end asm
files (ia64 for example). crtstuff.c does not actually work on ia64 since
libgcc.a doesn't have a few key support functions when built on ia64 so it
is compulsory to use crtbegin.asm and crtend.asm.
John Baldwin [Fri, 17 May 2002 05:45:39 +0000 (05:45 +0000)]
- Apparently, the Alpha ABI mandates that arguments be passed sign-extended
regardless of if they are signed or unsigned since it is easier to work
with sign-extended values. Thus, remove the disabled zapnot to
zero-extend the sign-extended value we read from *p in atomic_cmpset_32()
since the cmpval we are comparing against should already be
sign-extended.
- To ensure that the compiler knows to sign-extend the upper 32 bits of
cmpval rather than leaving garbage in there, cast the appropriately in
the constraints section.
Ruslan Ermilov [Fri, 17 May 2002 05:41:47 +0000 (05:41 +0000)]
Back out revision 1.30 change.
cc1plus can apparently be built if you happen to have
/usr/bin/gperf, or set CXX to point to a C++ compiler
that can build gperf(1) in the bootstrap-tools stage
of buildworld.
Juli Mallett [Fri, 17 May 2002 05:27:52 +0000 (05:27 +0000)]
Stop this program's abuse of malloc(3). Its return value doesn't need these
ugly explicit casts, and its argument doesn't need explicitly cast to u_int,
especially if sizeof() is being used.
Tim J. Robbins [Fri, 17 May 2002 05:20:30 +0000 (05:20 +0000)]
Un-deprecate od(1): rename the `deprecated' variable to `odmode', remove the
deprecation warning from the utility and manual page. Since this utility
is required by POSIX, it's not likely to be removed any time soon.
This is leading up to the addition of the P1003.1-2001 -s -A -j -N -t options.
Juli Mallett [Fri, 17 May 2002 05:11:07 +0000 (05:11 +0000)]
Clean up malloc(3)'s argument. Remove casts which do nothing when we're
using sizeof() anyway. Use slightly more consistent (per-file) error
reporting for malloc(3) returning NULL. If "malloc failed" was being printed,
don't use err(3). If a NULL format is being used, use err(3). In one case
errx(3) was being used with strerror(3), so just use err(3).
David E. O'Brien [Fri, 17 May 2002 03:43:47 +0000 (03:43 +0000)]
Update to Binutils 2.12.0 and stick to the native emulation for now.
This lets us get on the with the task at hand -- porting FreeBSD.
We can get fancy later on.