Robert Watson [Tue, 22 Feb 2000 03:56:58 +0000 (03:56 +0000)]
After much consulting with bde, concluded that this fix was the best fix
to the current jail/chflags interactions. This fix conditionalizes ``root
behavior'' in the chflags() case on not being in jail, so attempts to
perform a chflags in a jail are limited to what a normal user could do.
For example, this does allow setting of user flags as appropriate, but
prohibits changing of system flags.
Alfred Perlstein [Tue, 22 Feb 2000 01:40:43 +0000 (01:40 +0000)]
Don't scare the users.
Kirk argees that the comment about corruption caused by switching the flags
on an already mounted manpage are bogus, it doesn't happen.
Warner Losh [Mon, 21 Feb 2000 06:56:30 +0000 (06:56 +0000)]
[[ Forced commit due to /tmp/msg botch ]]
Minor improvements in pcic.
o Change IODF_ACTIVE to MDF_ACTIVE when dealing with memory. This is
a text only change.
o Better spl safty, ala comments from Bruce and the nomads.
o Default to resetting the pcic on resume. This is needed for newer
machines and may fix jordan's vaio problem.
o i/o mapping verbosity under bootverbose.
o Added comment about irqs that should have been in prior commit.
o Added sanity check on memory rid.
Gary Jennejohn [Mon, 21 Feb 2000 02:10:10 +0000 (02:10 +0000)]
Newbus-ify the USR Sportster TA Intern driver.
Enable the driver in sys/conf/files.i386.
In isa/isavar.h increase ISA_NPORT from 32 to 50. This is required
because this brain-damaged card maps 49 (!) port ranges. This does
not have a negative impact because this value only specifies the maximum
number of entries in a linked list and not the size of an array which
is allocated in all drivers.
The register/fifo access routines were not newbus-ified because
1) I knew that the old code worked and is simpler and more efficient
2) the if_ed driver does something similar and
3) the newbus macros collapse to inb/outb anyway.
Peter Wemm [Sun, 20 Feb 2000 15:01:37 +0000 (15:01 +0000)]
Followup to previous commit; change fd_mask from signed to unsigned as
it should have been to start with. The implications of that are not
pleasant when combined with >> as it breaks on multiples of NFDBITS.
Right shifting of a signed value fills with a copy of the sign bit,
not a zero.
Peter Wemm [Sun, 20 Feb 2000 13:36:26 +0000 (13:36 +0000)]
Fix select(2) for the Alpha. (!!) It was never returning true for
fd's in the range of 32-63, 96-127 etc. The first problem was the
FD_*() macros were shifting a 32 bit integer "1" left by more than
32 bits. The same problem happened in selscan(). ffs() also takes
an int argument and causes failure. For cases where int == long
(ie: the usual case for x86, but not always as gcc can have long
being a 64 bit quantity) ffs() could be used.
Kenneth D. Merry [Sun, 20 Feb 2000 04:42:44 +0000 (04:42 +0000)]
Fix 'camcontrol inquiry'. The inquiry data structure changes (increased to
256 bytes) caused it to break on many devices.
The SCSI spec says that for commands with 8-bit length fields, a value of 0
means 256 bytes. As it turns out, many devices don't deal with that
properly. Some interpret the 0 as 0, and return no data. Others return
more than 256 bytes of data, and cause an overrun.
The fix is to tell the device we've only allocated SHORT_INQUIRY_LENGTH (36
bytes) of inquiry data, instead of sizeof(struct scsi_inquiry_data).
camcontrol.c: Change inq_len in the call to scsi_inquiry() to
SHORT_INQUIRY_LENGTH, and add a long comment
explaining the reason for the change.
scsi_all.h: Add a comment above the definitinon of
SHORT_INQUIRY_LENGTH alerting people that it is
both the initial probe inquiry length, and the
minimum amount of data needed for scsi_print_inquiry()
to function.
scsi_all.c: Add a comment about SHORT_INQUIRY_LENGTH being the
minimum amount of data needed for
scsi_print_inquiry() to function.
Robert Watson [Sun, 20 Feb 2000 02:51:11 +0000 (02:51 +0000)]
- As jail(8) has been almost completely rewritten, prepend another copyright/
BSD-style license, as an add-on to phk's beerware license. Please fedex
some beer to phk.
- Add a ``make depend'' line to the jail-building, which fixes openssl,
among other things. Suggested by: kris
- Add ``newaliases'' to the list of things to do when setting up a new
jail, so that the jailed sendmail doesn't complain.
- Correct references to ``kern.jail.set_hostname_allowed'' which now read
``jail.set_hostname_allowed''.
- Add a reference to sysctl.conf where the sysctl can easily be set in
a persistent way.
Robert Watson [Sun, 20 Feb 2000 01:10:36 +0000 (01:10 +0000)]
Disable chflags() from within jail() so that root within jail can't make
a mess in securelevel environments. Results in one warning during
/etc/rc as it attempts to remove file flags, but this is harmless.
Yoshinobu Inoue [Sat, 19 Feb 2000 16:10:16 +0000 (16:10 +0000)]
Change IPv6 scoped addr format again based on recent standard discussion.
Sorry for the flapping, but no change will be done for 4.0 anymore.
Official standard will be published around April or later.
If different format would be adopted at that time, then support for
the new format will be added to the succeeding FreeBSD 4.x.
Doug Rabson [Sat, 19 Feb 2000 09:44:06 +0000 (09:44 +0000)]
Remove the vga-pci driver. It serves no purpose and it hides the hardware
from useful drivers such as the 3D DRI drivers I will be porting for
hardware accelerated OpenGL. The hardware will still be reported during
boot using the nomatch system.
Søren Schmidt [Fri, 18 Feb 2000 20:57:33 +0000 (20:57 +0000)]
Update the ata driver to take more advantage of newbus, this
was needed to make attach/detach of devices work, which is
needed for the PCCARD support.
(PCCARD support is still not working though, more to come on that)
Support the CMD646 chip which is used on many alphas, sadly only
in WDMA2 mode, as the silicon is broken beyond belief for UDMA modes.
Lots of cosmetic fixes here and there.
Sorry for the size of this megapatchfromhell but it was not
possible otherwise...
newbus patches based on work from: dfr (Doug Rabson)
Robert Watson [Fri, 18 Feb 2000 19:02:22 +0000 (19:02 +0000)]
Fix up a few documentation nits in jail(8), as well as improve the
instructions so as to reduce warnings during jail startup, etc.
Add a somewhat bolder warning recommending the use of
kern.jail.set_hostname to limit jail renamining.
Invert the meaning of two questions in the Standard installation so just
whapping "Yes" by default does not turn you into an anonymous FTP-supporting
gateway machine. Those aren't the right "defaults."
Robert Watson [Fri, 18 Feb 2000 05:49:08 +0000 (05:49 +0000)]
Modify MAKEDEV to create four bpf devices instead of one when doing a
MAKEDEV all, making DHCP on multiple interfaces happier, and allowing use
of tcpdump to to debug DHCP, without creating more devices.
(we need devfs)
Modify MAKEDEV to create four tun devices in MAKEDEV all as well, since
we're being gratuitous with network pseudo-devices.
(we need devfs)
Bill Swingle [Fri, 18 Feb 2000 03:36:52 +0000 (03:36 +0000)]
ack. Ignore me. Log should read:
Added references to related man pages that should be under the
"See Also" section. The related PR mentions some missing man pages
that Jason has added to his to do list.
Mike Smith [Thu, 17 Feb 2000 23:33:57 +0000 (23:33 +0000)]
Increase the time we spend waiting for the controller to become ready to
accept a new command; in high load cases it may be too busy for the old
value.
This loop needs something to tie it to real time, rather than just the CPU's
ability to fetch from the L1 data cache, but this hack works for now.
Mike Smith [Thu, 17 Feb 2000 23:32:08 +0000 (23:32 +0000)]
Change the mountroot prompt to something that doesn't look at all like a
firmware prompt. Several sleepy folk mistook the '>>>' for the SRM
prompt, which was never the desired idea.
Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
Approved by: jkh
Bruce Evans [Thu, 17 Feb 2000 18:37:45 +0000 (18:37 +0000)]
Don't include <machine/ipl.h> in <sys/systm.h> in the i386 case. This
fixes some namespace pollution in general and breakage of modules that
aren't in the sys tree in particular (<machine/ipl.h> includes further
headers that aren't installed under /usr/include).
Reimplemented SPLASSERT() so that it is more machine independent and
less bloated and doesn't require the <machine/ipl.h> include spam.
In particular, don't assume that `cpl' can be printed using %08x
format. The alpha arch doesn't even have `cpl'. SPLASSERT() was
harmless on alphas because it isn't actually used.
Daniel C. Sobral [Thu, 17 Feb 2000 17:59:41 +0000 (17:59 +0000)]
If no distribution file is found, abort the installation. Previous
behavior silently ignored the problem. This affects installations
from a filesystem, and possible other types of media.
Daniel C. Sobral [Thu, 17 Feb 2000 17:57:40 +0000 (17:57 +0000)]
If the installation media disappears in the process of installing
a distribution, recognize it and treat as fatal media error. This
happens in the case of a timeout on FTP installations where the
user chooses not to select another FTP site, and resulted in
segmentation fault.
The default rule in this file actually sent mail to root as its default
action when denying access to a service. Unfortunately, this also makes
a dandy denial-of-service attack possible. Change to just log the event
and shoot a "go away" response back down the socket.
Bill Swingle [Thu, 17 Feb 2000 03:40:21 +0000 (03:40 +0000)]
From the PR:
The description of -X option in csh(1) manpage uses a wording
that references the descriptions of -x, -v and -V. This might
be a little confusing. Changed this to a complete description
that does not reference other paragraphs.
Bill Swingle [Thu, 17 Feb 2000 03:23:39 +0000 (03:23 +0000)]
From the PR:
The first paragraph of "Argument list processing" says that an
argument of - will make csh be a login shell. However, running
csh with only a - as an argument fails with the error message.
Bill Swingle [Thu, 17 Feb 2000 03:17:03 +0000 (03:17 +0000)]
From the PR:
Three minor changes to the manpage of chmod(1).
1. At the description of -H option, I added that symlinks are
not followed _by default_ to show that links can be followed,
but the default chmod behavior is not to do so.
2. Moved a misplaced .Va file command up to the place it belongs.
3. Simplified the grammar that describes symbolic modes.
Robert Watson [Wed, 16 Feb 2000 23:50:43 +0000 (23:50 +0000)]
Modified jail.8 to correct a typo (inetd_flas vs. inetd_flags), and add
a comment to the effect that I'm responsible for the additional
documentation, et al, so that phk gets fewer messages about my errors.