Peter Wemm [Mon, 28 Jun 1999 09:38:09 +0000 (09:38 +0000)]
Add some extra alpha_pal_imb()'s in overkill mode. This makes this PC164SX
run quite reliably now. I've explicitly tagged them as /* XXX overkill? */
although one does actually check for VM_PROT_EXECUTE.
Based on a suggestion by: Dmitrij Tejblum <tejblum@arc.hq.cti.ru>
Peter Wemm [Mon, 28 Jun 1999 09:18:44 +0000 (09:18 +0000)]
Revert back to not using -DKERNEL - it exposes inlines etc.
Zap symbols.raw and glue to make symbols.* - it's not used on the ELF-only
alpha kernel. Symbol sorting is dead-end anyway once libkvm uses the
in-kernel linker symbol lookup.
David E. O'Brien [Mon, 28 Jun 1999 09:05:56 +0000 (09:05 +0000)]
Define "OBSTACK_CHUNK_SIZE" as getpagesize().
Our malloc can allocte pagesized blocks efficiently and the EGCS default size
of 4072 bytes is not optimal.
Kirk McKusick [Mon, 28 Jun 1999 07:54:58 +0000 (07:54 +0000)]
When requesting an exclusive lock with LK_NOWAIT, do not panic
if LK_RECURSIVE is not set, as we will simply return that the
lock is busy and not actually deadlock. This allows processes
to use polling locks against buffers that they may already
hold exclusively locked.
David E. O'Brien [Mon, 28 Jun 1999 04:09:14 +0000 (04:09 +0000)]
Amd now mounts to "/.amd_mnt" rather than "/net". "/net/<host>" now works
in addition to "/host/<host>". This make us consistant with Sun (as we
already were with SGI).
Peter Wemm [Sun, 27 Jun 1999 22:08:38 +0000 (22:08 +0000)]
Kirk missed a required BUF_KERNPROC(). Even though this is a non-async
transfer, the b_iodone hook causes biodone() to release it from interrupt
context.
Robert Nordier [Sun, 27 Jun 1999 19:29:15 +0000 (19:29 +0000)]
Drop the embedded boot code in favour of a -B option which reads
the boot code from /boot/mbr, or elsewhere as defined by the revised
"-b bootcode" option; use getopt(3); clarify usage(); partially
revise man page; etc.
Sheldon Hearn [Sun, 27 Jun 1999 18:05:34 +0000 (18:05 +0000)]
Add command-line option (-w), specified once to enable wrapping and
twice to enable wrapping for internal wrapping as well. If the option is
not specified wrapping is turned off so that inetd will behave exactly
as it used to before TCP Wrappers was imported.
Change etc/defaults/rc.conf so as to encourage wrapping on new systems.
Clarify the use of TCP Wrappers in the IMPLEMENTATION NOTES of the
manual page.
Doug Rabson [Sun, 27 Jun 1999 12:44:44 +0000 (12:44 +0000)]
* Tidy up a few things and fix some more warnings.
* Change the devsw declaration to standard form and add missing fields
* Change the filename from cdev_mod.ko to cdev.ko
Peter Wemm [Sun, 27 Jun 1999 11:40:03 +0000 (11:40 +0000)]
The BUF_*() routines must be internally splbio() protected otherwise they
can cause a biodone() from a disk interrupt to spin when the interrupt
code tries to grab the simplelock. Masking BIO here means buftimelock
and/or lk->lk_interlock shouldn't be held when an interrupt tries to grab
them.
Mike Pritchard [Sun, 27 Jun 1999 00:28:55 +0000 (00:28 +0000)]
Document that mlock() and munlock() can return EPERM if the
caller is not the super-user. Also document that we do not
currently support the per-process RLIMIT_MEMLOCK limit.
Mike Pritchard [Sat, 26 Jun 1999 22:53:22 +0000 (22:53 +0000)]
Actually describe all of the command line options in fsinfo(8).
Most of the text was take from the fsinfo section in info(1).
Also included a pointer to that information, since there is
so much detail about the format of the configuration files there.
Peter Wemm [Sat, 26 Jun 1999 14:44:24 +0000 (14:44 +0000)]
I'm tired of having a 'hanging root device'.. This isn't a "fix", just
a workaround for a specific case where cam interrupts right in the middle
of this printf.
Brian Somers [Sat, 26 Jun 1999 02:54:24 +0000 (02:54 +0000)]
If we run out of chat script immediately after a successfully executed
expect-send-expect sequence, finish gracefully, don't core dump.
This bug has been there for over a year - I could never reproduce it !
Straw provided by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
Kirk McKusick [Sat, 26 Jun 1999 02:47:16 +0000 (02:47 +0000)]
Convert buffer locking from using the B_BUSY and B_WANTED flags to using
lockmgr locks. This commit should be functionally equivalent to the old
semantics. That is, all buffer locking is done with LK_EXCLUSIVE
requests. Changes to take advantage of LK_SHARED and LK_RECURSIVE will
be done in future commits.
KATO Takenori [Fri, 25 Jun 1999 15:17:20 +0000 (15:17 +0000)]
From submitter:
The attached diff attempts to eliminate as much of the difference
between the i386 and the pc98 version of the file as possible. It
should not make any semantic difference (it consists of whitespace
changes, order changes, comment changes, changes of case for hex
constants, and merging in a couple of constants that hadn't made it
from the i386 version.)
Pierre Beyssac [Fri, 25 Jun 1999 13:43:30 +0000 (13:43 +0000)]
Never return the root node itself from rn_match(); return NULL instead.
This caused a panic in rtfreee() called with a root node from the
routing socket code (when processing a RTM_GET message looking for
the default route while there is none).
Since no existing code seems to have any use getting the root node
from rn_match(), it seems cleaner never to return it rather than
check for this condition at the caller's.
Søren Schmidt [Fri, 25 Jun 1999 09:03:07 +0000 (09:03 +0000)]
Nine'th update to the new ATA/ATAPI driver:
The atapi subsystem has gotten better error handeling and timeouts,
it also tries a REQUEST SENSE command when devices returns errors,
to give a little more info as to what went wrong. It might be a
little verbose for now, but I'm interested in as much feedback on
errors as possible, especially timeouts, as I'm a bit in doubt if
I've chosen resonable default values everywhere.
The disk driver has been changed a bit to prepare for tagged queing,
which is next on my list.
The disk driver has grown a dump routine, I got one implementation
from Darrell Anderson <anderson@cs.duke.edu> which also did
partial dumps (usefull on big memory machines) I left out the
partial stuff for now, and changed the rest alot to fit into the new
ad_request framework.
Some minor cleanups and rearrangements as well.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, again you have been warned :)
Notebook owners should be carefull that their machines dont suspend
as this might cause trouble...
John Polstra [Fri, 25 Jun 1999 02:53:59 +0000 (02:53 +0000)]
Fix a serious performance bug for large programs on the Alpha,
discovered by Hidetoshi Shimokawa. Large programs need multiple
GOTs. The lazy binding stub in the PLT can be reached from any of
these GOTs, but the dynamic linker only has enough information to
fix up the first GOT entry. Thus calls through the other GOTs went
through the time-consuming lazy binding process on every call.
This fix rewrites the PLT entries themselves to bypass the lazy
binding.
John-Mark Gurney [Thu, 24 Jun 1999 22:50:24 +0000 (22:50 +0000)]
add support to buildworld as a normal user:
-DNOFSCHG disables installation of libs with flag schg
GAMEGRP change the group with which games are installed
also organize the binary section into alphebetical order some what..