Add a new vnode op (cachedlookup) so that filesystems can plug into
a global vfs_cache check. The rest of this change will come when the
current zero size file problem is resolved.
Garrett Wollman [Mon, 25 Aug 1997 16:57:05 +0000 (16:57 +0000)]
Print out the two new fields in the ICMP stats. While we're at it, also
convert icmp_stats() to use sysctl(3) to retrieve the information
rather than kvm. This makes it easy to also print whether ICMP address
mask responses are enabled, so do so.
Garrett Wollman [Mon, 25 Aug 1997 16:29:27 +0000 (16:29 +0000)]
ICMP Timestamp Request messages could have harbored the same sort of
problem as Echo Requests when broad/multicast. When multicast echo responses
are disabled, also do the same for timestamp responses.
Warner Losh [Mon, 25 Aug 1997 05:02:11 +0000 (05:02 +0000)]
Change from using sys/param.h and MAXPATHLEN to using limits.h and PATH_MAX
to avoid the NOFILE problem. Back out NOFILE kludge. Fix unexploitable
buffer overflow in tipout.c.
Reviewed by: millert@openbsd.org (tipout.c)
Submitted by: millert@openbsd.org (rest of idea)
Garrett Wollman [Mon, 25 Aug 1997 01:25:31 +0000 (01:25 +0000)]
Configurably don't reply to broadcast or multicast echos. There are still
potential problems with other automatic-reply ICMPs, but some of them may
depend on broadcast/multicast to operate. (This code can simply be
moved to the `reflect' label to generalize it.)
Steve Passe [Sun, 24 Aug 1997 20:33:32 +0000 (20:33 +0000)]
A clean fix for the spl "deadlock before smp_active" problem.
Added a new variable, 'bsp_apic_ready', which is set as soon as the bootstrap
CPU has initialized its local APIC. Conditionalize the GENSPLR functions
to call ss_lock ONLY after bsp_apic_ready is TRUE; This should prevent
any problems with races between the time the 1st AP becomes ready and the
time smp_active is set.
KATO Takenori [Sun, 24 Aug 1997 11:23:17 +0000 (11:23 +0000)]
Code cleanup. Removed !FreeBSD code arround sysctl stuff. Renamed
doclusterread/doclusterwrite into ext2_doclusterread and
ext2_doclusterwrite, which are unique names. Moved #include of
<sys/sysctl.h> to the top of the file.
Stefan Eßer [Sun, 24 Aug 1997 06:24:51 +0000 (06:24 +0000)]
Go back to previous definition of FE_CACHE_SET, since the problem
caused by read-multiple on the 53c810a should have been fixed by
adjusting the alignment of the global header in rev 1.104 of ncr.c.
Steve Passe [Sun, 24 Aug 1997 00:05:37 +0000 (00:05 +0000)]
The last of the encapsolation of cpl/spl/ipending things into a critical
region protected by the simplelock 'cpl_lock'.
Notes:
- this code is currently controlled on a section by section basis with
defines in machine/param.h. All sections are currently enabled.
- this code is not as clean as I would like, but that can wait till later.
- the "giant lock" still surrounds most instances of this "cpl region".
I still have to do the code that arbitrates setting cpl between the
top and bottom halves of the kernel.
- the possibility of deadlock exists, I am committing the code at this
point so as to exercise it and detect any such cases B4 the "giant lock"
is removed.
Stefan Eßer [Sat, 23 Aug 1997 22:01:49 +0000 (22:01 +0000)]
Some cleanup and a fix for an assumed chip bug:
- Do not malloc SCRIPTS memory for those parts of the microcode that
are to be loaded into the on-chip SRAM of the 53c825a or 875 ...
- Modify ncr_chip_lookup to make adding new entries easier.
- Disable use of on-chip SRAM for the 53c825 rev 0x10 to 0x12, since
there seems to be a problem with rev 0x11, while 0x13 is known to
work. (Tested by Chuck Robey <chuckr@glue.umd.edu>).
This code will be merged into 2.2-stable after a few more days of
testing in -current.
Stefan Eßer [Sat, 23 Aug 1997 21:53:47 +0000 (21:53 +0000)]
Minor corrections to the code added in rev. 1.100 and 1.101:
- fix features of 53c860
- correctly adjust data structure to cache line boundary (32 bytes)
Peter Wemm [Fri, 22 Aug 1997 06:57:30 +0000 (06:57 +0000)]
Fix a braino in the Attic/ removal code for $CVSHeader$ that originated
from XFree86. I think this was harmless, the strncmp happened always
rather than being optimised out.
Peter Wemm [Fri, 22 Aug 1997 04:42:12 +0000 (04:42 +0000)]
Add -R (rescan or refresh) option to rescan and rebuild the hints
file based on the previous list of directories stored there which
should overcome a weakness of the '-m' switch which can only add
libs. This is an ideal way of updating the hints list after adding
or removing a shlib since it will remove entries that are gone and
doesn't need to have all the directories spelled out each time.
(eg: rm -f /usr/lib/libtcl75*; ldconfig -R) This only works for
version 2 hints files (which we've been generating for a year or
so) which store the path.