Add a new ddb command 'show pcpu' which lists some of the per-cpu data.
Specifically, the cpuid, curproc, curpcb, npxproc, and idleproc members.
Also, if witness is compiled into the kernel, then a list of all the spin
locks held by this CPU is displayed. By default the information for the
current CPU is displayed, but a decimal cpu id may be specified as a
parameter to obtain information on a specific CPU.
- Split out the functionality of displaying the contents of a single lock
list into a public witness_list_locks() function. Call this function
twice in witness_list() instead of using an evil goto.
- Adjust the 'show locks' command to take an optional parameter which
specifies the pid of a process to list the locks of. By default the
locks held by the current process are displayed.
Clean up telnet's argument processing a bit. autologin and encryption is
now the default, so ignore the arguments that turn it on. Add a new -y
argument to turn off encryption in case someone wants to do that. Sync
these changes with the man page (including removing the now obsolete
statement about availability only in the US and Canada).
Rev 1.10 actually only tested booting from floppies. Thus that is all
that was fixed.
There are other problems why the current sources for the Alpha `cdboot'
do not produce a working loader. Because we use the 4.0-RELEASE `cdboot'
binary, it will not get these fixes at the current time. Thus CD booting
on certain AlphaServer boxes is still broken.
Add IPv6 support to showmount(8). This replaces IPv4-specific code
with calls to the new protocol-independent clnt_*_create functions
provided by ti-rpc. Martin submitted a more complex patch to achieve
this, but it turns out that clnt_create() does everything we need.
Soften the dire warnings about this code. Things are kinda working
now and it does compile :-). There are still some issues, but it is a
good time to soften the warning.
joe [Thu, 5 Apr 2001 13:59:51 +0000 (13:59 +0000)]
Fix a few bugs in BSDPAN:
* Fix a bug which prevented the second invocation of overloaded
subs governed by SelfLoader from functioning.
* Fix a bug with XS modules. MakeMaker determines where the xsubpp
preprocessor is located by adding "ExtUtils" string to the Perl's
system path. At the same time, BSDPAN has to fool MakeMaker into
thinking that the Perl's system path is elsewhere. Now we
`reverse-adjust' the notion of the Perl's system path for a
moment, so xsubpp utility is found.
This should fix the breakage with some p5- ports.
Reported by: vanilla
Submitted by: Anton Berezin <tobez@tobez.org>
Don't call prom_open() multiple times. This confuses some versions of SRM
and makes it impossible to boot from floppy and CD on some AlphaServer
platforms.
Detective work by: Michael Richards <michael@fastmail.ca>
We don't use the "pseudo-device" configuration option any longer,
so update the example to use the correct definition.
Add an example for documenting kernel compile options, along with
a small example of how to reference them in the main text of the
man page (I.e. the .Dv macro).
Inspired-by: a brief exchange I saw in in the commit messages mail
o Rather than arbitrarily construct a credential in the nfs_statfs()
VFS operation, make use of the calling process's credential. This
solution may not be ideal (there are a number of other possible
proposals, including making use of the proc0 credential, adding a
credential argument to the VFSOP, and switching from a hard-coded
ucred to a hard-coded nfscred), it is simple and appears to
work. The arguments against using simply crget() are fairly
strong: it is the only place in the code (other than a nearly
identical invocation in ncp) where crget() is invoked, other than
in the process credential creation code; as ucred becomes extensible,
this use of crget() without appropriate context results in less and
less meaningful credential data. The implementation here will
probably be tweaked as a result of experimentation and further
exploration of the requirements. In the mean-time, it allows
progress to be made in ucred expansion for new security models without
causing a crash every time df is used on an NFS mounted file system.
This code has been interop tested against FreeBSD and Solaris NFS
servers. While using the process credentials should not introduce
interop problems, please let me know if any turn out to exist.
Change a couple of M_WAITOKs used in M_PREPEND() to M_TRYWAITs, which
is what they should be. As the returned mbuf is already checked for
failure of M_PREPEND even in the wait case, nothing more to be done
here.
Various style fixes.
Also place the macros under #ifdef _KERNEL. Equally hide the internal
structures such as the freelist structs which include condition variables.
- Add descriptions and cross-references for the ACL editing library
functions.
- Place the acl_dup() description in alphabetical order.
- Move the POSIX.1e descriptions under the ENVIRONMENT section to the
STANDARDS section.
ken [Wed, 4 Apr 2001 22:50:57 +0000 (22:50 +0000)]
Don't allow immediate values of 0 for operations that can take either an
immediate value or the accumulator. 0 is the chip's internal
representation for the accumulator, and so 0 is an invalid immediate value
when the accumulator can also be specified as an argument.
Redo a lot of the target mode infrastructure to be cognizant of Dual Bus
cards like the 1280 && the 12160. Cleanup isp_target_putback_atio.
Make sure bus and correct tag ids and firmware handles get propagated
as needed.
Complete some Ansification. Check to make sure, in tdma_mk, that we won't
overflow the request queue. The reason we want to do this is that we
now push out completed CTIOs as we complete them- this gets the QLogic
working on them quicker. So we need to know whether we can put the entire
burrito out before we start.
We now support conjoint status with data for the last CTIO for both Fibre
Channel and SCSI. Leave the old code in place in case we need to go back
(minor 3 line ifdef).
Ultra-ultra important- *don't* set rq->req_seg_count for non-data
target mode requests in isp_pci_dmasetup. D'oh- this is actually
the tag value area for a CTIO. What *was* I thinking? Boy howdy
does both aic7xxx and sym get awfully unhappy when on reconnect
you give them a constant '1' for a tag value.
Perform some more Ansification. Remove and then replace the isp_putback_atio
function- we did it a bit cleaner. We only use this if a CTIO completes with
!CT_OK state. We now have managed to get away without having to poke around
and trying to find the original ATIO- the csio we're using has the tag_id
and lun values with it which is mostly what we need when we do the putback.
Make sure we correctly propagate AT_TQAE->CT_TQAE for tags. Make sure
we call ISP_DMAFREE only if we had DATA to move.
Amazing. The bits to enable tagged queing in target mode, grok that a
tag is active for an ATIO, and say that you want to reconnect with
a tag value in a CTIO have *never* been exercised until now. This lossage
derived from Solaris code where this stuff originally came from that is
about 7 years old. Amazing.
We now bundle the incoming tag (legal values are 0..256) as the low
16 bits of the ccb_accept_tio's at_tagid while we put the firmware
handle for this ATIO in the top 16 bits- define some macros to make
this cleaner.
After loading f/w, for FC cards print out Firmware Attributes.
Redo establishment of default SCSI parameters whether or not
we've been compiled for target mode. Unfortunately, the Qlogic
f/w is confused so that if we set all targets to be 'safe' (i.e.,
narrow/async), it will also then report narrow, async if we're
contacted in target mode from that target (acting in initiator
role). D'oh!
Fix ISPCTL_TOGGLE_TMODE to correctly enable the right channel for
dual channel cards. Add some more opcodes. Fix a stupid NULL
pointer bug.
If we have and error and are booting verbosely, don't be complaining
if this was a non-retryable selection timeout- wading through 256 targets
worth of Fibre Channel 'selection timeouts' is tedious at best.
Activate build of posix1e extensions in libc and libc_r that have been
moved in from libposix1e, and deactivate build of the soon-to-be-removed
libposix1e.
Prepare for the inclusion of libposix1e into libc: retire the old
Makefile, add Makefile.inc needed for libc build; add
#include "namespace.h"/#include "un-namespace.h" pairs around the
includes of sys/acl.h and sys/capability.h, and an additional underscore
in front of the functions that will be overridden in libc_r.
Fixed a null pointer bug in rev.1.10. Rev.1.10 was supposed to to
move the "for safety" zeroing of unused members of timebuf to a better
place. It actually moved the zeroing to a worse place and didn't add
necessary braces.
Fixed a nearby older bug. timebuf.tm_gmtoff was sometimes used even
when timebuf was invalid. Even when it is zeroed, a failing mktime()
might set it to nonzero.
Simplify initialization and remove offending DMA channel resets there.
The resets trash whatever is pointed to DMA registers, but at cmi_attach()
time the DMA registers have not been initialized with valid addresses.
Reviewed by: Cameron Grant <gandalf@vilnya.demon.co.uk>