Bruce Evans [Tue, 19 Mar 2002 08:20:44 +0000 (08:20 +0000)]
Backed out the part of the previous commit related to xargs. It just
broke things, since "name=value ... cmd ..." only works for simple
commands (not for pipelines).
Bruce Evans [Tue, 19 Mar 2002 06:30:24 +0000 (06:30 +0000)]
Removed env(1) commands. make(1) uses a real shell, and
"env name=value ... cmd ..." is just a pessimized way of doing
"name=value ... cmd ..." in real shells. Set the environment
(without using env(1)) before starting xargs so that env(1)
is not needed in "xargs env name=value ... cmd ..."
Bruce Evans [Tue, 19 Mar 2002 06:11:43 +0000 (06:11 +0000)]
Removed bogus env(1) commands. make(1) uses a real shell, so
"env name=value ... cmd ..." was just a pessimized way of doing
"name=value ... cmd ...". Note that make(1) can't optimize
either of these to an exec of env(1) or "cmd" even if the second
"..." is simple, since it can't tell that the shell metacharacter
in "name=value" is actually handled by env(1).
Alan Cox [Tue, 19 Mar 2002 04:54:30 +0000 (04:54 +0000)]
Eliminate unnecessary calls to grow_stack() and useracc() from linux_sendsig()
and linux_rt_sendsig(). (See i386/i386/machdep.c revisions 1.503 and 1.504.)
Bruce Evans [Tue, 19 Mar 2002 04:09:21 +0000 (04:09 +0000)]
Fixed some printf format errors (hopefully all of the remaining daddr64_t
ones for GENERIC, and all others on the same line as those). Reformat
the printfs if necessary to avoid new long lones or old format printf
errors.
Juli Mallett [Tue, 19 Mar 2002 02:05:27 +0000 (02:05 +0000)]
Fix handling of filenames with spaces in them in the input file to
uudecode(1), as this behaviour was explicitly added in revision 1.12 as
the result of PR 2882. Remove space (' ') from the delimiter characters
handed to strtok(3).
Reviewed by: mike
Approved by: mike
Pointy hat to: me
Crist J. Clark [Tue, 19 Mar 2002 01:56:04 +0000 (01:56 +0000)]
IPFilter may need to be re-sync'ed even if we are not filtering, but
only doing ipnat(8). Go back to using $ipfilter_active, but turn off
$ipfilter_active when loading ipl.ko has failed.
Submitted by: devet@devet.org (Arjan de Vet)
MFC after: 3 days
Juli Mallett [Tue, 19 Mar 2002 00:44:07 +0000 (00:44 +0000)]
Make uudecode(1) use setmode(3) and getmode(3) instead of just blindly
chmodding using an octal mode, as uudecode needs to handle symbolic modes
as chmod and such do.
Suggested by: Tim J. Robbins <tim@robbins.id.au>
Also move meta-character (~ in this case) globbing to only if we are reading
the filename from the encoded file, as otherwise it is the shell's job.
Add an ia64 category; we'll need it sooner or later. Also add a commented-
out entry for powerpc; I'll enable it when the powerpc port gets its own
mailing list.
Bruce Evans [Mon, 18 Mar 2002 23:38:16 +0000 (23:38 +0000)]
Fixed printf format errors in previous commit. %llu is no more suitable
than %u for printing signed 64-bit types. It fails on different machines,
and has the wrong signdness.
Fixed old printf format error on the same line. %u is not suitable for
printing 32-bit types on all machines.
"Fixed" format printf error in previous commit. This file is not
formatted in KNF. Partially restore bug for bug compatibility: indent
the printf args too much, but don't format them for 160-column terminals.
Bruce Evans [Mon, 18 Mar 2002 23:24:00 +0000 (23:24 +0000)]
Fixed printf format errors in previous commit. %llu is no more suitable
than %u for printing signed 64-bit types. It fails on different machines,
and has the wrong signdness.
Fixed old printf format error on the same line. %u is not suitable for
printing 32-bit types on all machines.
Brian Feldman [Mon, 18 Mar 2002 15:08:09 +0000 (15:08 +0000)]
Back out the modification of vm_map locks from lockmgr to sx locks. The
best path forward now is likely to change the lockmgr locks to simple
sleep mutexes, then see if any extra contention it generates is greater
than removed overhead of managing local locking state information,
cost of extra calls into lockmgr, etc.
Additionally, making the vm_map lock a mutex and respecting it properly
will put us much closer to not needing Giant magic in vm.
Crist J. Clark [Mon, 18 Mar 2002 12:55:28 +0000 (12:55 +0000)]
Fix world breakage introduced by my recent modifications to
chpass(8). The relations between libc, libpam, chpass, passwd, and
vipw are a mess and probably should be cleaned up.
Alan Cox [Mon, 18 Mar 2002 07:59:57 +0000 (07:59 +0000)]
Eliminate grow_stack() from (o)sendsig(). If the stack needs to grow,
copyout() will page fault and perform grow_stack() from trap_pfault().
These calls to grow_stack() accomplish nothing.
Andrew R. Reiter [Mon, 18 Mar 2002 07:45:30 +0000 (07:45 +0000)]
- Lock down the ``module'' structure by adding an SX lock that is used by
all the global bits of ``module'' data. This commit adds a few generic
macros, MOD_SLOCK, MOD_XLOCK, etc., that are meant to be used as ways
of accessing the SX lock. It is also the first step in helping to lock
down the kernel linker and module systems.
Warner Losh [Mon, 18 Mar 2002 07:39:27 +0000 (07:39 +0000)]
__MAKE_CONF?=/etc/make.conf
Use this where we are now using /etc/make.conf.
This allows people to override the current default of always including
/etc/make.conf. Setting __MAKE_CONF to /dev/null disables it
completely, while setting it to something else allows one to override
what is on the system. This can be desirable in situations where a
machine has many users and some of them want different defaults, or
defaults appropriate to cross building to be different than those for
normal building.
Warner Losh [Mon, 18 Mar 2002 07:23:41 +0000 (07:23 +0000)]
o remove __P
o Use ansi function definitions
o MAXPATHLEN already has the NUL at the end, so no need to add 1 (note that
MAXNAMLEN doesn't, so the + 1 there is correct).
o remove register.
Warner Losh [Mon, 18 Mar 2002 07:14:59 +0000 (07:14 +0000)]
o MAXPATHLEN is the correct constant to use for path names, it includes
the NULL.
o use snprintf in preference to unchecked strcat in a couple of places that
likely can't overflow. Makes it easier to grep for strcpy :-)
Kirk McKusick [Mon, 18 Mar 2002 05:39:04 +0000 (05:39 +0000)]
Cannot release vnode underlying the nullfs vnode in null_inactive
as it leaves the nullfs vnode allocated, but with no identity. The
effect is that a null mount can slowly accumulate all the vnodes
in the system, reclaiming them only when it is unmounted. Thus
the null_inactive state instead accelerates the release of the
null vnode by calling vrecycle which will in turn call the
null_reclaim operator. The null_reclaim routine then does the
freeing actions previosuly (incorrectly) done in null_inactive.