Guy Helmer [Fri, 28 May 1999 01:54:38 +0000 (01:54 +0000)]
Where the fstab page discusses filesystem options, add a reference
to the '-o' flag in the mount(8) and filesystem specific
mount pages where information can be obtained about the available
options.
Bruce Evans [Thu, 27 May 1999 12:58:41 +0000 (12:58 +0000)]
Fixed usage message to almost match the man page. The old -d option and
the new -i option were missing.
Fixed style bugs in previous commit:
(1) initialisation of a local variable in its declaration.
(2) inconsistency of (1) with style of nearby code.
(3) disorder of declaration for (1).
(4) a line longer than 80 characters.
(5) bitrot in the printf() -> err() cleanups to help bloat the line in (4).
David Greenman [Thu, 27 May 1999 12:24:21 +0000 (12:24 +0000)]
Added net.inet.tcp.path_mtu_discovery variable which when set to 0
(default 1) disables PMTUD globally. Although PMTUD can be disabled in
the standard case by locking the MTU on a static route (including the
default route), this method doesn't work in the face of dynamic routing
protocols like gated.
Andrew Gallatin [Wed, 26 May 1999 23:22:14 +0000 (23:22 +0000)]
Add support for the DEC_ST6600 family of alphas. This includes the
Compaq XP1000, AlphaServer DS20, AlphaServer DS10, and DP264
This has been tested *only* on XP1000's. I'll be interested to hear from
owners of other types of DEC_ST6600 alphas.
I'd like to thank Don Rice of Compaq for providing the documentation required
to support this platform on FreeBSD. I'd also like to thank Doug Rabson for newbus,
and for helping me get a multiple hoses working with newbus.
Andrew Gallatin [Wed, 26 May 1999 23:05:23 +0000 (23:05 +0000)]
Forgotten in previous commit:
Allow chipset drivers to specify the direct-mapped DMA window's mask in
preparation for tsunami support. Previous chipsets' direct-mapped DMA
mask was always 1024*1024*1024. The Tsunami chipset needs it to be
2*1024*1024*1024
Andrew Gallatin [Wed, 26 May 1999 23:01:57 +0000 (23:01 +0000)]
Allow chipset drivers to specify the direct-mapped DMA window's mask in
preparation for tsunami support. Previous chipsets' direct-mapped DMA
mask was always 1024*1024*1024. The Tsunami chipset needs it to be
2*1024*1024*1024
Bill Paul [Wed, 26 May 1999 22:56:22 +0000 (22:56 +0000)]
Two small tweaks:
- Clear the IFF_OACTIVE flag when al_txeof() runs down the last TX mbuf chain.
- Mark the workaround for the transmitter stalling bug with
#ifdef AL_TX_STALL_WAR/#endif.
Justin T. Gibbs [Wed, 26 May 1999 16:59:17 +0000 (16:59 +0000)]
Ignore subvendor and subdevice ids for all ai7880 parts. These chips
only support 'mirroring' the vendor and device ids, so we don't
lose any information. Certain revisions of the aic7880 will not
perform the mirroring so to match all possiblities would double
the number of table entries. This change also allows us to match
things like the 2944B which I missed in the original table.
Roger Hardiman [Wed, 26 May 1999 13:24:35 +0000 (13:24 +0000)]
Updated options for the Bt848/Bt878 driver
This includes the BKTR_430_FX_MODE and BKTR_SIS_VIA_MODE options which make
Bt878/879 cards work better on 430FX and old SiS/VIA/OPTi boards
Justin T. Gibbs [Tue, 25 May 1999 20:17:29 +0000 (20:17 +0000)]
Run xpt_action at splsoftcam() so that it cannot be re-entered by our SWI.
The XPT doesn't have a problem with this itself, but some controllers
drivers may have been caught off guard by the old behavior.
XPT_CONT_TARGET_IO is also a valid ccb type for cam_periph_unmapmem.
Justin T. Gibbs [Tue, 25 May 1999 20:15:20 +0000 (20:15 +0000)]
The 1542 cards do not allow adapter commands to be queued while mailbox
commands are outstanding. You'd think they'd just clear the IDLE bit,
but alas, no. Delay until all pending mailbox commands have completed
in aha_cmd to work around this.
Report sync rates correctly on Fast Adaptec cards. Clones may still be
reported incorrectly since there is no documenation on how they report
extended sync values.
Dmitrij Tejblum [Tue, 25 May 1999 19:32:53 +0000 (19:32 +0000)]
From our i386 counterpart:
|revision 1.13
|date: 1995/09/15 23:49:23; author: davidg; state: Exp; lines: +15 -2
|Check for page being resident when doing I/O with /dev/kmem and return
|EFAULT if it is not resident. This prevents the system from manufacturing
|a zero-fill page for unused but allocated areas of the kernel's VM.
Kenneth D. Merry [Tue, 25 May 1999 17:10:04 +0000 (17:10 +0000)]
Fix the read/write permission checks in cam_periph_mapmem(). If we're
writing, we want to be able to read the buffer. If we're reading, we want
to be able to write to the buffer.
PR: kern/11870
Submitted by: Andrew Mobbs <amobbs@allstor-sw.co.uk>
Peter Wemm [Tue, 25 May 1999 11:14:33 +0000 (11:14 +0000)]
Add a -i localname switch (restricted to root) to specify a different
local "login" name for rcmd(3). This is particularly useful for things
like portslave and other packages with terminal server functionality
where a login can either run ppp locally or get shunted off to another
box via rlogin depending on radius authentication etc. Quite often the
local box doesn't even have accounts, so a flag such as this is needed.
Obviously this is restricted to callers with uid == 0.
Fix i4bteld entry, the section of the sh case for "i4teld*)" should be
BEFORE the case for "i4tel*)". (match the longest prefix first!)
Recognized by: Harry Starr
Brian Somers [Mon, 24 May 1999 16:39:17 +0000 (16:39 +0000)]
Correct the ``ignoring sync/async'' warnings so that they show
up with the correct device type.
Reassign the correct tcpdevice or execdevice after transfering
a link in MP server mode.
Bill Paul [Mon, 24 May 1999 14:56:55 +0000 (14:56 +0000)]
Fix bug that can cause transmit corruption. There are actually two 'rings'
in the transmit code: the TX descriptor ring, and a 'shadow' ring of mbuf
pointers, one for each TX descriptor. When transmitting a packet that
consists of several fragments in an mbuf chain, we link each fragment
to a descriptor in the TX ring, but we only save a pointer to the mbuf
chain. This pointer is saved in the shadow ring entry which corresponds
to the first fragment in the packet. Later, ti_txeof() can release the
whole chain with a single m_freem() call. (We need the second ring to
keep track of the virtual addresses of the mbuf chains.)
The problem with this is that the Tigon isn't actually through with the
mbuf chain until it reaches the last fragment (which has the TI_BDFLAG_END
bit set), however the current scheme releases the mbuf chain as soon as
the first fragment is consumed. This is wrong, since the mbufs can then
be yanked out from under the Tigon and modified before the other fragments
can be transmitted.
The fix is to make a one line change to ti_encap() so that it saves the
mbuf chain pointer in the shadow ring entry that corresponds to the last
fragment in TX ring instead of the first. This prevents the mbufs from
being released until the last fragment is transmitted.
Painstakingly diagnosed and fixed by: Robert Picco <picco@mail.wevinc.com>
Brought to my attention by: dg
Luigi Rizzo [Mon, 24 May 1999 10:01:22 +0000 (10:01 +0000)]
close pr 10889:
+ add a missing call to dn_rule_delete() when flushing firewall
rules, thus preventing possible panics due to dangling pointers
(this was already done for single rule deletes).
+ improve "usage" output in ipfw(8)
+ add a few checks to ipfw pipe parameters and make it a bit more
tolerant of common mistakes (such as specifying kbit instead of Kbit)
Peter Wemm [Mon, 24 May 1999 03:08:46 +0000 (03:08 +0000)]
Fix a [start,end] vs [start,count] botch that corrupted the resource
manager and prevented IOPort allocation beyond the first EISA slot from
working. subr_rman.c should have trapped this on the way into the system
rather than tripping over the wreckage.
Head banged into wall repeatedly by: "Matthew N. Dodd" <winter@jurai.net>
John Birrell [Mon, 24 May 1999 00:37:26 +0000 (00:37 +0000)]
Back out my previous change (phk didn't like it) in favour of setting
rootdev in the mfs initialisation code iff MFS_ROOT (which Bruce doesn't
like). Damned if I do - damned if I don't.
John Birrell [Mon, 24 May 1999 00:34:10 +0000 (00:34 +0000)]
Remove the test for bdevsw(dev) == NULL from bdevvp() because it fails
if there is no character device associated with the block device. In this
case that doesn't matter because bdevvp() doesn't use the character
device structure.
John Birrell [Mon, 24 May 1999 00:30:49 +0000 (00:30 +0000)]
- Make setroot() conditional on FFS etc, to avoid a compiler warning
on systems with no FFS.
- Remove all references to mfs from cpu_rootconf(). mfs_init is
called prior to cpu_rootconf(), so it can set mountrootfsname to mfs
and (more imporantly) set rootdev using the (bogus in Bruce's opinion)
special major number of 255.
John Birrell [Mon, 24 May 1999 00:27:12 +0000 (00:27 +0000)]
- Back out Luoqi's cdevsw stuff. It panics on my system and is not required.
- Fix an error message.
- Do the MFS_ROOT setting of mountrootfsname in mfs_init() instead of
cpu_rootconf().
- Set rootdev in mfs_init instead of later in mfs_mount() iff MFS_ROOT.
Roger Hardiman [Sun, 23 May 1999 21:40:51 +0000 (21:40 +0000)]
Added rgb_vbi_prog() to capture VBI data and video at the
same time. To capture VBI data, /dev/vbi must be opened
before starting video capture.
A partly on code from Hiroki Mori <mori@infocity.co.jp>
Justin T. Gibbs [Sun, 23 May 1999 18:57:29 +0000 (18:57 +0000)]
Rely on the statistics in XPT_GDEV_STATS instead of the versions still
retained in XPT_GDEV_TYPE for binary compatibility. Mark the legacy
structure values for removal when we bump the major CAM revision.
Justin T. Gibbs [Sun, 23 May 1999 18:54:34 +0000 (18:54 +0000)]
Poll for interrupts in bt_cmd in case they are masked. A completing
mailbox command will block the completion of an immediate command,
so we must service them for our command to succeed.
xargs was spliting the input in a very different way from everyone else,
and was also a bit inconsistent: leading blanks, or any double blanks
generated empty arguments, but a trailing blank did not.
PR: bin/2630, bin/10914
Submitted by: Arne Henrik Juul <arnej@imf.unit.no>
Construct the profile file name from the name of the executable. A program
compiled with -pg and run will generate a file <executable-filename>.gmon,
not gmon.out.
Justin T. Gibbs [Sat, 22 May 1999 22:04:11 +0000 (22:04 +0000)]
aic7xxx.c:
Honor the 'bus reset at startup' option now that the XPT properly
handles transfer negotiation in this scenario.
Honor the sync rate settings on Ultra2 controllers. We would
always negotiate at the fastest speed. Oops.
aic7xxx.h:
Whitespace.
aic7xxx.seq:
Fix a minor nit that would cause the controller to miss the update
of the negotiation required bitmask causing the negotiation to
be delayed by a command.