The pkg_version -c output will now call for a "make clean" before and
after the port build/install. The former cleans up "dirty" port work
directories that happen to be lying around, and the latter cleans up
after we're done so that they won't trip up someone else.
Backed out previous commit. It cause massive filesystem corruption,
not to mention a compile-time warning about the critical function
becoming unused, by replacing spec_bmap() with vop_stdbmap().
ntfs seems to have the same bug.
The factor for converting specfs block numbers to physical block
numbers is 1, but vop_stdbmap() uses the bogus factor
btodb(ap->a_vp->v_mount->mnt_stat.f_iosize), which is 16 for ffs with
the default block size of 8K. This factor is bogus even for vop_stdbmap()
-- the correct factor is related to the filesystem blocksize which is not
necessarily the same to the optimal i/o size. vop_stdbmap() was apparently
cloned from nfs where these sizes happen to be the same.
There may also be a problem with a_vp->v_mount being null. spec_bmap()
still checks for this, but I think the checks in specfs are dead code
which used to support block devices.
Make rtprio work again.
- add a missing break which caused RTP_SET to always return EINVAL
- break instead of returning if p_can fails so proc_lock is always
dropped correctly
- only copyin data that is actually needed
- use break instead of goto
- make rtp_to_pri return EINVAL instead of -1 if the values are out
or range so we don't have to translate
Build system:
o Define a RELN_ROOT variable which points to the root of the relnotes
tree (i.e., src/release/doc).
o By default, define DOC_PREFIX in terms of RELN_ROOT; this gives a
bigger chance of finding the doc/ tree without help in the form of
setting DOC_PREFIX on the command line.
o Respect DOCDIR; `make install` works now.
o As part of the move to not maintaining copies of the vnode owning uid
and gid in the ACL, vaccess_acl_posix1e() was changed to accept
explicit file_uid and file_gid as arguments. However, in making the
change, I explicitly checked file_gid against cr->cr_groups[0], rather
than using groupmember, resulting in ACL_GROUP_OBJ entries being
compared to the caller's effective gid only, not the remainder of
its groups. This was recently corrected for the version of the
group call without privilege, but the second test (when privilege is
added) was missed. This change replaces an additiona cr->cr_groups[0]
check with groupmember().
Add a vop_stdbmap(), and make it part of the default vop vector.
Make 7 filesystems which don't really know about VOP_BMAP rely
on the default vector, rather than more or less complete local
vop_nopbmap() implementations.
peter [Sun, 29 Apr 2001 03:19:58 +0000 (03:19 +0000)]
Disable llib. The previous commit does not go anywhere near far enough
to enable this.
1: it was running xlint out of the object directory, which is not
safe (ie: run a 5.x binary on a 4.x world - no libc.so.5, or run an
alpha binary on x86).
2: lint has /usr/libexec/lint1 and /usr/libexec/lint2 hard coded in.
This is the same as problem 1.
3: lint has got /usr/bin/cc hard coded in as well. Also, see problem 1.
There are probably more problems, but these are enough of a showstopper.
This version has a step debugger, which now completely replaces the
old trace feature. Also, we moved all of the FreeBSD-specific MI
code to loader.c, reducing the diff between this and the official
FICL distribution.
Add category definitions (which were missing) for locale categories LC_MESSAGES,
LC_NUMERIC, LC_MONETARY. Information source is POSIX locale defintions
collection at ftp://dkuug.dk/i18n/WG15-collection/
Clean up the error handling code a bit. There is no need to "roll our
own" error() routine when errx() is available. This resolves a conflict
when linking statically.
alfred [Fri, 27 Apr 2001 22:24:45 +0000 (22:24 +0000)]
Address a number of problems with sysctl_vm_zone().
The zone allocator's locks should be leaflocks, meaning that they
should never be held when entering into another subsystem, however
the sysctl grabs the zone global mutex and individual zone mutexes
while holding the lock it calls SYSCTL_OUT which recurses into the
VM subsystem in order to wire user memory to do a safe copy. This
can block and cause lock order reversals.
To fix this:
lock zone global.
get a count of the number of zones.
unlock global.
allocate temporary storage.
format and SYSCTL_OUT the banner.
lock global.
traverse list.
make sure we haven't looped more than the initial count taken
to avoid overflowing the allocated buffer.
lock each nodes.
read values and format into buffer.
unlock individual node.
unlock global.
format and SYSCTL_OUT the rest of the data.
free storage.
return.
Other problems included not checking for errors when doing sysctl out
of the column header. Fixed.
Inconsistant termination of the copied string. Fixed.
Objected to by: des (for not using sbuf)
Since the output is not variable length and I'm actually over
allocating signifigantly and I'd like to get this fixed now, I'll
work on the sbuf convertion at a later date. I would not object
to someone else taking it upon themselves to convert it to sbuf.
I hold no MAINTIANER rights to this code (for now).
Better handling of ioctl(SIOCSIFFLAGS) failing in ifpromisc():
- Don't print the "promiscuous mode (enabled|disabled)" on failure
- Restore the reference count on failure
sos [Fri, 27 Apr 2001 19:47:33 +0000 (19:47 +0000)]
Move the disklabel construction to the attach function instead of
in open. This prevent panics when a the disklabel is accessed on
an ATA-RAID subdisk.
Overhaul of the SMP code. Several portions of the SMP kernel support have
been made machine independent and various other adjustments have been made
to support Alpha SMP.
- It splits the per-process portions of hardclock() and statclock() off
into hardclock_process() and statclock_process() respectively. hardclock()
and statclock() call the *_process() functions for the current process so
that UP systems will run as before. For SMP systems, it is simply necessary
to ensure that all other processors execute the *_process() functions when the
main clock functions are triggered on one CPU by an interrupt. For the alpha
4100, clock interrupts are delievered in a staggered broadcast fashion, so
we simply call hardclock/statclock on the boot CPU and call the *_process()
functions on the secondaries. For x86, we call statclock and hardclock as
usual and then call forward_hardclock/statclock in the MD code to send an IPI
to cause the AP's to execute forwared_hardclock/statclock which then call the
*_process() functions.
- forward_signal() and forward_roundrobin() have been reworked to be MI and to
involve less hackery. Now the cpu doing the forward sets any flags, etc. and
sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically
return so that they can execute ast() and don't bother with setting the
astpending or needresched flags themselves. This also removes the loop in
forward_signal() as sched_lock closes the race condition that the loop worked
around.
- need_resched(), resched_wanted() and clear_resched() have been changed to take
a process to act on rather than assuming curproc so that they can be used to
implement forward_roundrobin() as described above.
- Various other SMP variables have been moved to a MI subr_smp.c and a new
header sys/smp.h declares MI SMP variables and API's. The IPI API's from
machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h.
- The globaldata_register() and globaldata_find() functions as well as the
SLIST of globaldata structures has become MI and moved into subr_smp.c.
Also, the globaldata list is only available if SMP support is compiled in.
Make stathz match reality on platforms like the 4100 where hz!=1024
It might be more correct to make stathz as close as possible to 128,
but that would involve adding complexity to the clock intr path, which
I don't want to do.
o Remove the disabled p_cansched() test cases that permitted users to
modify the scheduling properties of processes with a different real
uid but the same effective uid (i.e., daemons, et al). (note: these
cases were previously commented out, so this does not change the
compiled code at al)
Initiate the deorbit burn sequence for sys/ipl.h by moving the software
interrupt priorities over to sys/interrupt.h next to the software interrupt
API.
The function __rpc_uaddr2taddr_af() converts an RPC "universal
address" string to a netbuf/sockaddr "transport address". In the
case of an AF_LOCAL address, it was missing the code to actually
point the netbuf at the newly allocated sockaddr_un, so the caller
ended up with a netbuf containing junk.