more on Re: Please review: bugfix for vinvalbuf()
:Hi,
:
:I've just tripped over an obviously long-standing (since about
:Jan. 1998) bug in vinvalbuf while looking into PR kern/26224. The
:problematic code looks like (on -CURRENT):
:
: /*
:* Destroy the copy in the VM cache, too.
:*/
: mtx_lock(&vp->v_interlock);
: if (VOP_GETVOBJECT(vp, &object) == 0) {
: vm_object_page_remove(object, 0, 0,
: (flags & V_SAVE) ? TRUE : FALSE);
: }
: mtx_unlock(&vp->v_interlock);
:
:The locks seem to be needed for file systems that don't perform real
:locking (on -STABLE, they are simplelocks).
:This, however, is incorrect because vm_object_page_remove may sleep.
:I've attached a patch that passes the interlock to
:vm_object_page_remove, which in turn passes it to a modified version
:of vm_page_sleep, which unlocks it around the sleep.
:I think that this is correct, because the object should be in a valid
:state when we sleep (and all checks are reexecuted in that case).
:
:Since I'm not very experienced with vfs and vm stuff, I'd be glad if
:this patch could get some review. In particular, is the lock/unlock
:pair really still needed, and are there notable differeces in -STABLE
:(because the fix would need the be MFC'ed)?
:
:Thanks,
: - thomas
Ok, I've looked at this more. What is supposed to happen is that
the 'while (vp->v_numoutput) ...' code just above the section you
cited is supposed to prevent the deadlock. The while code looks like
this:
while (vp->v_numoutput > 0) {
vp->v_flag |= VBWAIT;
tsleep(&vp->v_numoutput, PVM, "vnvlbv", 0);
}
However, as Ian points out in his followup, it doesn't work:
:...
:I've seen a related deadlock here in 4.x with NFS; vm_fault locks
:a page and then calls vput which aquires the v_interlock. This code
:in vinvalbuf locks the v_interlock first, and then vm_object_page_remove()
:locks the page. That's a simple lock-order reversal deadlock which I
:guess would still exist even with this patch.
It doesn't work for the simple reason that vm_fault isn't busying the
page for writing, it's busying it for reading, so v_numoutput will be
0.
I think the best solution is to change vinvalbuf's wait loop to
wait on vm_object->paging_in_progress instead of vp->v_numoutput,
or perhaps to wait for both conditions to be satisfied.
vm_object->paging_in_progress applies to reads and writes, while
vp->v_numoutput only applies to writes.
I know this isn't the most correct solution... there is still the
lock reversal if vm_object_remove_pages() were ever to sleep. The
idea is that it won't sleep if there is no paging in progress because
nobody will have any of the object's pages locked. I think it is
the best we can do for the moment. There are several places in
vm/vm_object.c where the same assumption is made (testing against
vm_object->paging_in_progress, though, which works, not vp->v_numoutput).
-
Thomas, if you are interested this could be a little project for you.
See if you can code-up another while() loop to also wait for
the object's paging_in_progress count to become 0 in the vinvalbuf()
code. Look in vm/vm_object.c for examples of how to construct such
a loop. I will be happy to review and test whatever you come up with
and I'm sure Ian will too!
-Matt
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
> > In the "default" case, it should attempt to obtain a DHCP lease, > > and, failing that, ask the user to give it settings, or let > > them do IPv4 stateless autoconfiguration. Ad Hoc networking > > should always "just work". > > If anyone is taking a vote, I disagree. I do not want any system > ever assuming anything about my network. Even Win checks with the > user before enabling DHCP. Er, you *are* kidding, right? All modern Windows versions will try to get a lease long before the user has any say in the matter. So does OS X. -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] V I C T O R Y N O T V E N G E A N C E To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: TCP Window Size
You made it way to big... On 10 Jul 2001, Joseph Lekostaj wrote: > > I've been trying to up my TCP window size from the default 16K and it's caused >nothing but problems. From the info I've found so far, these are the sysctl i've >changed: > > kern.ipc.maxsockbuffer=2097152 > net.inet.tcp.rfc1323=1 > net.inet.tcp.sendspace=524288 > net.inet.tcp.recvspace=524288 > > But if I do that, on boot I get all sorts of error messages about buffer space. >i.e.: > > Jul 9 11:53:20 ccn64 portmap[180]: cannot create tcp socket: No buffer space >available > Jul 9 11:53:21 ccn64 inetd[199]: shell/tcp: socket: No buffer space available > Jul 9 11:53:21 ccn64 inetd[199]: login/tcp: socket: No buffer space available > Jul 9 11:58:55 ccn64 RPC::PlClient[243]: Cannot connect: No buffer space available > Jul 9 11:58:55 ccn64 RPC::PlClient[246]: Cannot connect: No buffer space available > > Is there anything I'm missing? > > -- > Joe LeKostaj > - > Just don't create a file called -rf. > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-hackers" in the body of the message > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: vfs.vmiodirenable undocumented
If memory serves me right, "David Xu" wrote: > but why hasn't a complete sysctl manual? > I see OpenBSD has a better sysctl manual, our sysctl(8) is too bad, > except the command usage info is useful, all left is garbage > information and waste disk space. I'm sorry, I missed the part of your message containing the patches to fix this problem. Bruce. PGP signature
Re: vfs.vmiodirenable undocumented
* David Xu <[EMAIL PROTECTED]> [010710 20:10] wrote: > but why hasn't a complete sysctl manual? > I see OpenBSD has a better sysctl manual, our sysctl(8) is too bad, > except the command usage info is useful, all left is garbage > information and waste disk space. David, I know you're busy writing the nextgen version of the FreeBSD logging filesystem, but I'm hoping that you'll be able to fit in an improved sysctl(8) into your work queue as well. What do you say? -- -Alfred Perlstein [[EMAIL PROTECTED]] Ok, who wrote this damn function called '??'? And why do my programs keep crashing in it? To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: vfs.vmiodirenable undocumented
but why hasn't a complete sysctl manual? I see OpenBSD has a better sysctl manual, our sysctl(8) is too bad, except the command usage info is useful, all left is garbage information and waste disk space. Regards, David Xu - Original Message - From: Sheldon Hearn <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, July 10, 2001 8:21 PM Subject: Re: vfs.vmiodirenable undocumented > > > On Tue, 10 Jul 2001 14:13:23 +0200, Sheldon Hearn wrote: > > > Someone recently suggested that I tune vfs.vmiodirenable on a system > > with lots of memory. The CVS commit logs and the source tell me > > absolutely nothing about what this tunable does. > > > > Is anyone in a position to document it? > > Someone mailed me privately and pointed out that the sysctl is > documented in tuning(7). > > Thanks, > Sheldon. > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-hackers" in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
B/W quickcam..(anyone got the program?)
I had a copy of Xfqcam that used to run this B/W qcam I have sitting here but it seems that I must have got it from somewher estrange because I cannot find it anywhere these days... anyone still have a copy of it? (I think it was left on my on=ld machine at whistle when I left :-( ) julian To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
[ANNOUNCE] SPY-1.1 - syscall monitoring kernel module
Hi, I just uploaded an updated version of the SPY, which is a kernel module that allows to selectively monitor and/or block execution of any syscalls. This version works on relatively current -CURRENT (after the struct proc changes). You can get it from: http://people.freebsd.org/~abial See also the detailed description there. I should be able also to provide a version for 4-STABLE soon, depending on my time and availability of the machine... Enjoy! -- Andrzej // // Andrzej Bialecki <[EMAIL PROTECTED]>, Chief System Architect // WebGiro AB, Sweden (http://www.webgiro.com) // // <[EMAIL PROTECTED]> FreeBSD developer (http://www.freebsd.org) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: (a bit offtopic) KDE2.1.1 install
[ redirected to -questions ] On Mon, Jul 09, 2001 at 02:27:35PM -0500, Lakey, Jeremy # IHTUL ([EMAIL PROTECTED]) wrote: > I'm running a p233 with 128mgs of ram, and KDE 2.1.1 install has been > compiling ALL MORNING! > > Is this unusual? No. It takes 6 hours on my dual Pentium III 600MHz with 640MB RAM to install KDE 2.1.2[*] (from ports/x11/kde2 meta-port). I would imagine on your system it will take 12-18 hours. Use the packages if you don't like this. OR.. use something lighter on the system like Window Maker... 8) -- wca [*] 2.1.2 is just a security release for kdelibs only. Nevertheless, it's what's currently in ports. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Please review: bugfix for vinvalbuf()
Hi,
I've just tripped over an obviously long-standing (since about
Jan. 1998) bug in vinvalbuf while looking into PR kern/26224. The
problematic code looks like (on -CURRENT):
/*
* Destroy the copy in the VM cache, too.
*/
mtx_lock(&vp->v_interlock);
if (VOP_GETVOBJECT(vp, &object) == 0) {
vm_object_page_remove(object, 0, 0,
(flags & V_SAVE) ? TRUE : FALSE);
}
mtx_unlock(&vp->v_interlock);
The locks seem to be needed for file systems that don't perform real
locking (on -STABLE, they are simplelocks).
This, however, is incorrect because vm_object_page_remove may sleep.
I've attached a patch that passes the interlock to
vm_object_page_remove, which in turn passes it to a modified version
of vm_page_sleep, which unlocks it around the sleep.
I think that this is correct, because the object should be in a valid
state when we sleep (and all checks are reexecuted in that case).
Since I'm not very experienced with vfs and vm stuff, I'd be glad if
this patch could get some review. In particular, is the lock/unlock
pair really still needed, and are there notable differeces in -STABLE
(because the fix would need the be MFC'ed)?
Thanks,
- thomas
Index: kern/vfs_subr.c
===
RCS file: /home/ncvs/src/sys/kern/vfs_subr.c,v
retrieving revision 1.315
diff -u -r1.315 vfs_subr.c
--- kern/vfs_subr.c 2001/07/04 16:20:13 1.315
+++ kern/vfs_subr.c 2001/07/10 19:45:41
@@ -800,7 +800,7 @@
mtx_lock(&vp->v_interlock);
if (VOP_GETVOBJECT(vp, &object) == 0) {
vm_object_page_remove(object, 0, 0,
- (flags & V_SAVE) ? TRUE : FALSE);
+ (flags & V_SAVE) ? TRUE : FALSE, &vp->v_interlock);
}
mtx_unlock(&vp->v_interlock);
Index: vm/vm_map.c
===
RCS file: /home/ncvs/src/sys/vm/vm_map.c,v
retrieving revision 1.206
diff -u -r1.206 vm_map.c
--- vm/vm_map.c 2001/07/04 20:15:16 1.206
+++ vm/vm_map.c 2001/07/10 20:02:41
@@ -1875,7 +1875,7 @@
vm_object_page_remove(object,
OFF_TO_IDX(offset),
OFF_TO_IDX(offset + size + PAGE_MASK),
- FALSE);
+ FALSE, NULL);
}
VOP_UNLOCK(object->handle, 0, curproc);
vm_object_deallocate(object);
@@ -1991,7 +1991,7 @@
offidxend = offidxstart + count;
if ((object == kernel_object) || (object == kmem_object)) {
- vm_object_page_remove(object, offidxstart, offidxend, FALSE);
+ vm_object_page_remove(object, offidxstart, offidxend, FALSE,
+NULL);
} else {
pmap_remove(map->pmap, s, e);
if (object != NULL &&
@@ -1999,7 +1999,7 @@
(object->flags & (OBJ_NOSPLIT|OBJ_ONEMAPPING)) ==
OBJ_ONEMAPPING &&
(object->type == OBJT_DEFAULT || object->type ==
OBJT_SWAP)) {
vm_object_collapse(object);
- vm_object_page_remove(object, offidxstart, offidxend,
FALSE);
+ vm_object_page_remove(object, offidxstart, offidxend,
+FALSE, NULL);
if (object->type == OBJT_SWAP) {
swap_pager_freespace(object, offidxstart,
count);
}
@@ -2994,7 +2994,7 @@
/*
* Remove unneeded old pages
*/
- vm_object_page_remove(first_object, 0, 0, 0);
+ vm_object_page_remove(first_object, 0, 0, 0, NULL);
/*
* Invalidate swap space
Index: vm/vm_object.c
===
RCS file: /home/ncvs/src/sys/vm/vm_object.c,v
retrieving revision 1.196
diff -u -r1.196 vm_object.c
--- vm/vm_object.c 2001/07/04 20:15:16 1.196
+++ vm/vm_object.c 2001/07/10 20:03:23
@@ -1438,7 +1438,7 @@
* The object must be locked.
*/
void
-vm_object_page_remove(vm_object_t object, vm_pindex_t start, vm_pindex_t end,
boolean_t clean_only)
+vm_object_page_remove(vm_object_t object, vm_pindex_t start, vm_pindex_t end,
+boolean_t clean_only, struct mtx *interlk)
{
vm_page_t p, next;
unsigned int size;
@@ -1478,7 +1478,7 @@
* interrupt -- minimize the spl transitions
*/
- if (vm_page_sleep_busy(p, TRUE, "vmopar"))
+
Re: TCP Window Size
In message <[EMAIL PROTECTED]>, [EMAIL PROTECTED] writes: >I'd personally suggest just tuning net.inet.tcp.*space to 64k at most. 1. Depends on your application. To avoid TCP window size related slowdowns, it needs to be at least as large as the bandwidth delay product (multiply bandwidth by RTT). 2. net.inet.tcp.*space affects the _defaults_. Changing them means that you'll use huge amounts of memory even in places where it doesn't matter (like ftp control connections). Instead, use setsockopt with SO_SNDBUF and SO_RCVBUF where appropriate. -- http://www.poohsticks.org/drew/";>Home Page For those who do, no explanation is necessary. For those who don't, no explanation is possible. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Suggestions for sysinstall / disklabel
From: Garance A Drosihn <[EMAIL PROTECTED]> Subject: Re: Suggestions for sysinstall / disklabel Date: Tue, 10 Jul 2001 17:28:26 -0400 > However, there is some kind of column-counting mistake there, > such that if you are creating more than one columns-worth of > partitions, then turning on the softupdates flag for a partition > in the second column writes past column 80 on the screen (which > is to say, it writes on the first character of the following row). This was fixed post-4.3. - jordan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
I'm also interested in helping out with the installer. I have also just read Jordan's message (at http://docs.freebsd.org/cgi/getmsg.cgi?fetch=77877+0+archive/2000/freebsd-ha ckers/2917.freebsd-hackers as referenced in another email) and here is how I see it. #2 cannot happen without #1 happening first. Jordan outlines a number of underlying problems which simply writing a new installer, in terms of just a new frontend to what sysinstall currently does, cannot solve. Many of the underlying structures and mechanisms for distribution and packaging must be modified to resolve those problems. These problems are generally the same as the ones other people are bringing up, or are needed changes before other types of changes can be made (though some, such as adding "partition magic" style partition sizing could be added without these underlying changes). I think, from a "will be accepted by the community" perspective, what has to happen is the old installer must evolve, little by little, into something which is more modular and can be easily replaced or interfaced into by other alternate installers. As it becomes more modular, it will be easier to change some of the underlying mechanisms. As other people have said, the installer is the first exposure of new people to FreeBSD. That means it is really important, and therefore a lot of us will care about what happens to it. Major overhauls are not going to go over easily (think bikeshed, only instead of what color to paint the bikeshed, we are discussing what color to paint the large centerpiece at the entrance to the world headquarters). Gilbert - Original Message - From: "Bill Moran" <[EMAIL PROTECTED]> To: "Garance A Drosihn" <[EMAIL PROTECTED]> Cc: "Wes Peters" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Tuesday, July 10, 2001 1:30 PM Subject: Re: FreeBSD Mall now BSDCentral > *sigh* this has gotten way off track somehow. Looking back, I'm > probably primarily to blame. > > The fact is: *I* *AM* interested in replacing or helping out with > an effort to replace/improve sysinstall. However, there are two > critical things I must understand if I am to do anything truely > productive along this line: > > 1. Should sysinstall be fixed or replaced. > 2. What needs done to fix or improve sysinstall. > > Now, I've gotten several excellent suggestions with regard to #2. > If any of my replies to these seem like arguments, then that's > just a part of my inability to communicate, as what I'm trying to > do is *understand* what exactly needs done. If I only change what > *I* want, it might not be terribly useful overall (as can be seen > by my divergence from other's opinions) You make an excellent point > of this in your second paragraph below. > > What I was really interested in getting answered was #1, as it > was suggested by Terry Lambert that "sysinstall must go" (don't > remember his exact words) I have no desire to start repairing > something if a complete replacement is in order. > > The fact of the matter is that I don't understand the problem > enough to fix it. It may seem strange to people that I'm more > interested in understanding this problem than in fixing problems > that I already understand. Well, I always have been a little > strange ... > > Personally, I would appreciate it if folks would *not* back > off from this conversation, since I'm acutally beginning to > understand, thanks to the tireless efforts of many who already > understand. > > -Bill > > Garance A Drosihn wrote: > > The main problem I have seen with this discussion is that the > > current partitioning program works well for you (Bill). Every > > time anyone suggests something, you reply that it seems frivolous > > to you, or that there is no other installer that you personally > > have used which does a better job. We're not trying to suggest > > things to make partitioning easier for YOU, we're suggesting > > things which would make it nicer/easier/friendlier for US, or > > people that we deal with. > > > > Pretty much every time I have used the disk-partitioning software > > for freebsd I have hated it. The *reason* I hate it changes from > > run-to-run, but I've never been particularly happy with it. It > > is clear that you've had better luck with it than I have. I think > > I'll just leave it at that, instead of trying to interest you in > > making some changes to it. I do not mean this as a sarcastic jab, > > I am just making the observation that the current setup seems to > > work well for you, and thus you're not going to be as interested > > in changing it as we might have hoped. > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-hackers" in the body of the message > > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: TCP Window Size
On Tue, Jul 10, 2001 at 05:00:50PM -0500, Bill Fumerola wrote: > Good christ thats a lot of memory per socket. kern.ipc.maxsockets is > also fairly important to know here too. Nothing is going to prevent > the above values from getting out of hand, short of a huge amount of > memory dedicated to sockets.. see below... ignore this, I was mistaken at what time the memory gets allocated -- Bill Fumerola - security yahoo / Yahoo! inc. - [EMAIL PROTECTED] / [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: TCP Window Size
On Tue, Jul 10, 2001 at 02:54:41PM -0500, Joseph Lekostaj wrote: > > I've been trying to up my TCP window size from the default 16K and it's caused >nothing but problems. From the info I've found so far, these are the sysctl i've >changed: > > kern.ipc.maxsockbuffer=2097152 > net.inet.tcp.rfc1323=1 > net.inet.tcp.sendspace=524288 > net.inet.tcp.recvspace=524288 Good christ thats a lot of memory per socket. kern.ipc.maxsockets is also fairly important to know here too. Nothing is going to prevent the above values from getting out of hand, short of a huge amount of memory dedicated to sockets.. see below... > But if I do that, on boot I get all sorts of error messages about buffer space. >i.e.: > > Jul 9 11:53:20 ccn64 portmap[180]: cannot create tcp socket: No buffer space >available > Jul 9 11:53:21 ccn64 inetd[199]: shell/tcp: socket: No buffer space available > Jul 9 11:53:21 ccn64 inetd[199]: login/tcp: socket: No buffer space available > Jul 9 11:58:55 ccn64 RPC::PlClient[243]: Cannot connect: No buffer space available > Jul 9 11:58:55 ccn64 RPC::PlClient[246]: Cannot connect: No buffer space available > > Is there anything I'm missing? The memory to support blowing 524k of memory per socket? As I recall this memory is eaten as the sockets are created. If you really think you need that much memory (you don't), you're going to need to increase your socket memory to a lot more then 2097152 I'd personally suggest just tuning net.inet.tcp.*space to 64k at most. -- Bill Fumerola - security yahoo / Yahoo! inc. - [EMAIL PROTECTED] / [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Suggestions for sysinstall / disklabel
At 2:29 PM -0700 7/10/01, David O'Brien wrote: >On Tue, Jul 10, 2001, Bill Moran wrote: > > Now, I've never used partition magic, but I (personally) find > > the FreeBSD partition program in sysinstall to be the easiest > > one I've ever used. > > What should be changed to make it easier? > >Maybe not "easier" but better: > >+ Allow one to specify the partition letter, than assumeing `e'. >+ Allow one to specify the ordering of partitions that will be written. > Alpha users keep getting bit in the ass because sysinstall orders the > swap partition at the begining of the disk vs. after /. One cannot > tell which order the partitions will be written to disk. Hmm. Is it a different program on Alpha than i386? On i386, the order on the disk seems to always be the order they were created in disklabel (which might or might not match the order of the partition letters...). On i386, I know know the program seems to want to futz with the partition letters based on the name you give to the partition, which is sometimes annoying. Ie, it "wants" the root partition to be partition 'a', and swap to be partition 'b', but there are times when that's not what *I* want (for one reason or another). I sometimes create partitions with the wrong name (such as '/') so I can get the partition letter I want, and then rename the partition after it has assigned the letter. It may just be the weird way I operate, in that I create multiple 'fdisk partitions' which will hold freebsd "slices" (so I can boot between freebsd-stable and freebsd-current). Or I'll use sysinstall to repartition one disk while up-and-running on a different disk. -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Can someone verify this?
On Sat, 2001/07/07 at 18:57:15 -0400, Paul Halliday wrote:
> FreeBSD dissent.p450.box 4.3-RC FreeBSD 4.3-RC #3: Sun Jun 10 22:27:47
> EDT 2001 [EMAIL PROTECTED]:/usr/src/sys/compile/workstation
> i386
>
> FreeBSD useless.dell.box 4.3-STABLE FreeBSD 4.3-STABLE #6: Fri Jul 6
> 18:57:08 EDT 2001
> [EMAIL PROTECTED]:/usr/src/sys/compile/useless i386
>
> mount /dev/acd0c /cdrom > should obviously fail, yet causes...
>
> panic: vm -fault on nofault entry, addr: c3e1e000
>
> reboot.
> any ideas?
If it was an audio CD you were trying to mount: this is a known
problem. The attached patch fixes it for me by disallowing reading of
partial blocks; this could also be fixed by setting the buffer size
different from the transfer size in such a case.
- thomas
Index: dev/ata/atapi-cd.c
===
RCS file: /home/ncvs/src/sys/dev/ata/atapi-cd.c,v
retrieving revision 1.48.2.10
diff -u -r1.48.2.10 atapi-cd.c
--- dev/ata/atapi-cd.c 2001/02/25 21:35:20 1.48.2.10
+++ dev/ata/atapi-cd.c 2001/07/09 21:48:58
@@ -1126,9 +1126,7 @@
/* reject all queued entries if media changed */
if (cdp->atp->flags & ATAPI_F_MEDIA_CHANGED) {
bp->b_error = EIO;
- bp->b_flags |= B_ERROR;
- biodone(bp);
- return;
+ goto failure;
}
bzero(ccb, sizeof(ccb));
@@ -1149,7 +1147,11 @@
lastlba = cdp->info.volsize;
}
-count = (bp->b_bcount + (blocksize - 1)) / blocksize;
+if (bp->b_bcount % blocksize != 0) {
+ bp->b_error = EINVAL;
+ goto failure;
+}
+count = bp->b_bcount / blocksize;
if (bp->b_flags & B_READ) {
/* if transfer goes beyond range adjust it to be within limits */
@@ -1191,6 +1193,11 @@
atapi_queue_cmd(cdp->atp, ccb, bp->b_data, count * blocksize,
bp->b_flags & B_READ ? ATPR_F_READ : 0, 30, acd_done,bp);
+return;
+
+failure:
+bp->b_flags |= B_ERROR;
+biodone(bp);
}
static int
Re: FreeBSD Mall now BSDCentral
Laurence Berland wrote: > > On Tue, 10 Jul 2001, Wes Peters wrote: > > > Laurence Berland wrote: > > > > > > On Tue, 10 Jul 2001, Rasputin wrote: > > > > > > > I may be low on caffeine, but I don't see how breaking up the base system > > > > into packages makes it any easier to upgrade than using cvsup? > > > > > > I think the discussion is Re: binary upgrades, like putting in the CD and > > > hitting that upgrade option, which right now doesn't quite get you there > > > afaik. > > > > I don't think the goal was to make the system easier to upgrade, but rather > > easier to subset. Do we really NEED to have sendmail on every DNS server > > we put together? > > Also very true. But in the context of upgrades, which is what Rasputin > was coming from, it's a binary upgrade issue, not a source one, which is > why cvsup isn't a fair comparison. I definitely think it'd be nice to > have things like sendmail somewhat more separate. Right, but updates were NOT the driving REASON for removing chunks of FreeBSD to ports. Of course this change will have ramifications for updates. The primary change, from my viewpoint, is that I don't have to wait to update sendmail on my DNS servers, because sendmail won't be part of the default system anymore. Yippee! -- "Where am I, and what am I doing in this handbasket?" Wes Peters Softweyr LLC [EMAIL PROTECTED] http://softweyr.com/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
At 2:06 PM -0700 7/10/01, Jordan Hubbard wrote: >From: Garance A Drosihn <[EMAIL PROTECTED]> >Subject: Re: FreeBSD Mall now BSDCentral >Date: Tue, 10 Jul 2001 16:06:54 -0400 > > > I do wish I could find someone who was interested in making > > changes to it, though. > >Why can't we just nominate you and be done with it? :) I'm still slogging my way thru lpr/lpd code, and I dare say there's enough crappy code in there to keep me occupied for quite some time... -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, Jul 10, 2001 at 09:49:18AM -0400, Bill Moran wrote: > Now, I've never used partition magic, but I (personally) find the > FreeBSD > partition program in sysinstall to be the easiest one I've ever used. > What should be changed to make it easier? Maybe not "easier" but better: + Allow one to specify the partition letter, than assumeing `e'. + Allow one to specify the ordering of partitions that will be written. Alpha users keep getting bit in the ass because sysinstall orders the swap partition at the begining of the disk vs. after /. One cannot tell which order the partitions will be written to disk. -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Suggestions for sysinstall / disklabel
coming from thread "Re: FreeBSD Mall now BSDCentral"... At 4:30 PM -0400 7/10/01, Bill Moran wrote: >*sigh* this has gotten way off track somehow. Looking back, I'm >probably primarily to blame. > >The fact is: *I* *AM* interested in replacing or helping out with >an effort to replace/improve sysinstall. However, there are two >critical things I must understand if I am to do anything truely >productive along this line: > >1. Should sysinstall be fixed or replaced. >2. What needs done to fix or improve sysinstall. > >Now, I've gotten several excellent suggestions with regard to #2. >If any of my replies to these seem like arguments, then that's >just a part of my inability to communicate, as what I'm trying to >do is *understand* what exactly needs done. If I only change what >*I* want, it might not be terribly useful overall (as can be seen >by my divergence from other's opinions) You make an excellent point >of this in your second paragraph below. > >What I was really interested in getting answered was #1, as it >was suggested by Terry Lambert that "sysinstall must go" (don't >remember his exact words) I have no desire to start repairing >something if a complete replacement is in order. > >The fact of the matter is that I don't understand the problem >enough to fix it. It may seem strange to people that I'm more >interested in understanding this problem than in fixing problems >that I already understand. Well, I always have been a little >strange ... > >Personally, I would appreciate it if folks would *not* back >off from this conversation, since I'm actually beginning to >understand, thanks to the tireless efforts of many who already >understand. I am inclined to say "replace sysinstall", just because it has so many quirks in it's user-interface which have annoyed me. On the other hand, I do want something a lot like sysinstall, and maybe if enough of the little things got fixed up then I might not be so eager to have it completely replaced. Part of the install process is the disk-partitioning step, which is more of an issue with fdisk and disklabel than sysinstall. Every time I go to use those I end up hitting something I don't particularly like, although the specifics of what I dislike are different each time. Eventually I get it to do exactly what I want it to do, at which point I try to put the whole experience out of my mind... :-) The one (silly, trivial) thing that I *like* about Openbsd's disk-partitioning setup is that I say "I want about 50 meg for this partition", and it automatically rounds up to the nearest "even boundary". It figures out the next-largest number of sectors such that no disk space is wasted. Yes, I realize it's a 20-gig disk now, but it still annoys me when I see "800 sectors unusable" for each partition I create. I like the recent change so the partition-size can be displayed in sectors vs Kilobytes vs Megabytes (we may soon have to add Gigabytes to that!). It would also be nice to have percentage as display option there. I create several different partitions, and inevitably I get to the end and realize some partition is too large or too small. What I'd like to do at that point is say "just take partition , and increase it by 20 meg", and have it move all the partitions after it to match that. Instead, I end up having to delete all the partitions down to the one I want to change, and then recreate them all. I very much like the recent change so you can turn on the softupdates flag for a partition when you are creating it. However, there is some kind of column-counting mistake there, such that if you are creating more than one columns-worth of partitions, then turning on the softupdates flag for a partition in the second column writes past column 80 on the screen (which is to say, it writes on the first character of the following row). I know there are other things which have annoyed me when I've gone to use the disk editor, but I'm afraid I have not written them down anywhere, and those are the only things that I remember off the top of my head. -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Bill Moran wrote in list.freebsd-hackers: # [...] # 1. Should sysinstall be fixed or replaced. Read the fine manual. ;) sysinstall(8): >BUGS > This utility is a prototype which lasted several years past its expira > tion date and is greatly in need of death. # 2. What needs done to fix or improve sysinstall. See 1. And if you didn't already know Jordans "Installation and package tools document, version 1.0" you should read it also: http://docs.freebsd.org/cgi/getmsg.cgi?fetch=77877+0+archive/2000/freebsd-hackers/2917.freebsd-hackers HTH, Uwe To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
mail failed, returning to sender
|- Message log follows: ---|
| no valid recipients were found for this message |
|--|
| [EMAIL PROTECTED] - unknown user
|--|
German translation:
Sie haben Ihre eMail an die Adresse "[EMAIL PROTECTED]" gerichtet.
Der hintere Teil der Adresse ("Domain"), "@mail.hh.provi.de", ist gueltig,
der Benutzernamen jedoch nicht.
Ueberpruefen Sie bitte insbesondere diesen Teil der eMail-Adresse!
Sie erhalten im Anhang die ersten zehn Zeilen Ihrer Original-eMail zurueck.
Mit freundlichen Gruessen,
Point of Presence GmbH, Hamburg
>From [EMAIL PROTECTED] Tue Jul 10 23:22:35 2001
Return-path: <[EMAIL PROTECTED]>
Envelope-to: [EMAIL PROTECTED]
Delivery-date: Tue, 10 Jul 2001 23:22:35 +0200
Received: from mx2.freebsd.org ([216.136.204.119])
by mail.provi.de with smtp (Exim 3.20 #2)
id 15K4xm-mE-00
for [EMAIL PROTECTED]; Tue, 10 Jul 2001 23:22:34 +0200
Received: from hub.freebsd.org (hub.freebsd.org [216.136.204.18])
by mx2.freebsd.org (Postfix) with ESMTP
id 8FF5255DFF; Tue, 10 Jul 2001 14:22:22 -0700 (PDT)
(envelope-from [EMAIL PROTECTED])
Received: by hub.freebsd.org (Postfix, from userid 538)
id 0F14E37B408; Tue, 10 Jul 2001 14:22:22 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1])
by hub.freebsd.org (Postfix) with SMTP
id E28E92E81CA; Tue, 10 Jul 2001 14:22:21 -0700 (PDT)
(envelope-from owner-freebsd-hackers-digest)
Received: by hub.freebsd.org (bulk_mailer v1.12); Tue, 10 Jul 2001 14:22:21 -0700
From: [EMAIL PROTECTED] (freebsd-hackers-digest)
To: [EMAIL PROTECTED]
Subject: freebsd-hackers-digest V5 #175
Reply-To: [EMAIL PROTECTED]
Sender: [EMAIL PROTECTED]
Precedence: bulk
Message-ID: <[EMAIL PROTECTED]>
Date: Tue, 10 Jul 2001 14:22:22 -0700 (PDT)
freebsd-hackers-digest Tuesday, July 10 2001 Volume 05 : Number 175
In this issue:
Re: FreeBSD Mall now BSDCentral
wx0 jumbo frame support explosions..
Re: wx0 jumbo frame support explosions..
Re: best way to migrate to a new disk
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
Eclipse manufacturs 32-way xeon boxes too. http://www.eclipse.com/, recently bought out by Data General, (http://www.dg.com/, used to be the manufacturers of the Aviion/Eclipse mainframe machines... and their proprietory O/S 'unicos'. They still sell/manufacture the numa based 32-way xeon boxes, though they've stopped supporting the legacy mainframes. Nathan Vidican [EMAIL PROTECTED] http://Nathan.Vidican.com/ - Original Message - From: "Wilko Bulte" <[EMAIL PROTECTED]> To: "Josh M Osborne" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Tuesday, July 10, 2001 3:34 PM Subject: Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question) > On Tue, Jul 10, 2001 at 02:25:13PM -0400, Josh M Osborne wrote: > > On Tue, Jul 10, 2001 at 12:42:12PM -0400, [EMAIL PROTECTED] wrote: > > [...] > > > > Instead, AMD implemented the Intel APIC specification; > > > > I'm not sure if they did it by licensing the patent > > > > (Intel had a patent on the APIC design), or if it's > > > > just been long enough for it to come off patent > > > > > > The Athlon uses the Alpha's ev6 spec, not the intel spec. > > > That may explain why you can buy Alpha systems with 40+ CPUs, and > > 32 max, on Wildfire. > > > Intel XENON boxes with no more then eight (or is it four?). It is > > Eight, eg on a Compaq Proliant 8000. Or 32 for the Unisys (IIRC) CMP > machines. > > -- > | / o / / _ Arnhem, The Netherlandsemail: [EMAIL PROTECTED] > |/|/ / / /( (_) Bulte "Youth is not a time in life, it is a state of mind" > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Inconsistency with wchar_t / wint_t in current
On Tue, Jul 10, 2001 at 05:40:27PM +0200, Carlo Dapor wrote: > I stumbled over an inconsistency with the data types wchar_t and wint_t. > My machine is a FreeBSD-5.0 current as of July 8th, 2001. I thought these were fixed. I'll look into it. -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
From: Garance A Drosihn <[EMAIL PROTECTED]> Subject: Re: FreeBSD Mall now BSDCentral Date: Tue, 10 Jul 2001 16:06:54 -0400 > I do wish I could find someone who was interested in making changes > to it, though. Why can't we just nominate you and be done with it? :) - jordan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
*sigh* this has gotten way off track somehow. Looking back, I'm probably primarily to blame. The fact is: *I* *AM* interested in replacing or helping out with an effort to replace/improve sysinstall. However, there are two critical things I must understand if I am to do anything truely productive along this line: 1. Should sysinstall be fixed or replaced. 2. What needs done to fix or improve sysinstall. Now, I've gotten several excellent suggestions with regard to #2. If any of my replies to these seem like arguments, then that's just a part of my inability to communicate, as what I'm trying to do is *understand* what exactly needs done. If I only change what *I* want, it might not be terribly useful overall (as can be seen by my divergence from other's opinions) You make an excellent point of this in your second paragraph below. What I was really interested in getting answered was #1, as it was suggested by Terry Lambert that "sysinstall must go" (don't remember his exact words) I have no desire to start repairing something if a complete replacement is in order. The fact of the matter is that I don't understand the problem enough to fix it. It may seem strange to people that I'm more interested in understanding this problem than in fixing problems that I already understand. Well, I always have been a little strange ... Personally, I would appreciate it if folks would *not* back off from this conversation, since I'm acutally beginning to understand, thanks to the tireless efforts of many who already understand. -Bill Garance A Drosihn wrote: > The main problem I have seen with this discussion is that the > current partitioning program works well for you (Bill). Every > time anyone suggests something, you reply that it seems frivolous > to you, or that there is no other installer that you personally > have used which does a better job. We're not trying to suggest > things to make partitioning easier for YOU, we're suggesting > things which would make it nicer/easier/friendlier for US, or > people that we deal with. > > Pretty much every time I have used the disk-partitioning software > for freebsd I have hated it. The *reason* I hate it changes from > run-to-run, but I've never been particularly happy with it. It > is clear that you've had better luck with it than I have. I think > I'll just leave it at that, instead of trying to interest you in > making some changes to it. I do not mean this as a sarcastic jab, > I am just making the observation that the current setup seems to > work well for you, and thus you're not going to be as interested > in changing it as we might have hoped. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: The Foundation [was Re: FreeBSD Mall now BSDCentral]
At 10:30 AM -0700 7/7/01, Richard Hodges wrote: >Could I just ask exactly what the FreeBSD Foundation _is_? > >I read the announcement and bylaws, and it looks like it is supposed >to be the offical organization representing FreeBSD. On the other >hand, many people have suggested that there is no connection between >Core and the Foundation. It is a foundation which will fund some projects intended to benefit FreeBSD, and "the public good". People can send contributions to the foundation and get a tax write-off. For the most part, there could be a hundred different foundations, each contributing to freebsd in their own way, and any given FreeBSD user could contribute to the foundations that they "trust" the most, or which is funding the kinds of projects which that user is particularly interested in. The United Way collects money, and gives some of that to a local hospital. That does not mean the hospital IS the united way, or that the united way IS the hospital. They are separate entities. >Although I infer tacit consent, I have not heard any official >blessing of the Foundation, either. But there is more than a >hint of transfering the FreeBSD trademark to the Foundation. I'll admit that this trademark issue is the one very odd thing about the foundation. I like the idea of the foundation, and for the record I have already sent a small contribution to them. If good things come of that, I'll send more. Still, I am a bit uneasy about the trademark being turned over to them. Not that I worry about the individuals, but I wonder about the foundation itself. Let's say the trademark is moved, and then the foundation folds. What happens to the trademark then? What happens if the foundation gets the trademark, and then they DO have to legally defend it from someone who's just out to make trouble? Could such an action cause the demise of the foundation? On the other hand, I have no better course of action to suggest wrt the trademark. It's just that this is the one "odd thing" about the FreeBSD Foundation, at least in my mind. -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, 10 Jul 2001, Wes Peters wrote: > Laurence Berland wrote: > > > > On Tue, 10 Jul 2001, Rasputin wrote: > > > > > I may be low on caffeine, but I don't see how breaking up the base system > > > into packages makes it any easier to upgrade than using cvsup? > > > > I think the discussion is Re: binary upgrades, like putting in the CD and > > hitting that upgrade option, which right now doesn't quite get you there > > afaik. > > I don't think the goal was to make the system easier to upgrade, but rather > easier to subset. Do we really NEED to have sendmail on every DNS server > we put together? Also very true. But in the context of upgrades, which is what Rasputin was coming from, it's a binary upgrade issue, not a source one, which is why cvsup isn't a fair comparison. I definitely think it'd be nice to have things like sendmail somewhat more separate. > > -- > "Where am I, and what am I doing in this handbasket?" > > Wes Peters Softweyr LLC > [EMAIL PROTECTED] http://softweyr.com/ > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
At 3:42 PM -0400 7/10/01, [EMAIL PROTECTED] wrote: >In a message dated 07/10/2001 12:52:30 PM Eastern Daylight Time, >[EMAIL PROTECTED] writes: > >> BSDi had no effect on ftp.freebsd.org's services and kept things >> completely unchanged there > >which pretty much confirms my "no impact" statement in that area. I must admit, I am also pretty close to filtering out everything from your userid, as your comments are just a waste of time. That's pretty amazing, considering that most of this thread is a waste of time, and yet you're the only one who really strikes me as being worth a filter. -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
At 9:49 AM -0400 7/10/01, Bill Moran wrote: >Wes Peters wrote: > > > Oh, come now. FreeBSD's disk partitioning has always sucked. > > It does suck somewhat less than many Linux linstallers, and a > > lot less than the OpenBSD installer, actually, there are a few things about the openbsd partitioning setup which I kinda like... :-) > > but it still shoves way too many details and options at the > > average user. Something akin to PartitionMagic would be an > > ideal way to go, given unlimited resources to throw against > > this particular problem. > >Now, I've never used partition magic, but I (personally) find the >FreeBSD partition program in sysinstall to be the easiest one I've >ever used. What should be changed to make it easier? The main problem I have seen with this discussion is that the current partitioning program works well for you (Bill). Every time anyone suggests something, you reply that it seems frivolous to you, or that there is no other installer that you personally have used which does a better job. We're not trying to suggest things to make partitioning easier for YOU, we're suggesting things which would make it nicer/easier/friendlier for US, or people that we deal with. Pretty much every time I have used the disk-partitioning software for freebsd I have hated it. The *reason* I hate it changes from run-to-run, but I've never been particularly happy with it. It is clear that you've had better luck with it than I have. I think I'll just leave it at that, instead of trying to interest you in making some changes to it. I do not mean this as a sarcastic jab, I am just making the observation that the current setup seems to work well for you, and thus you're not going to be as interested in changing it as we might have hoped. I do wish I could find someone who was interested in making changes to it, though. -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
TCP Window Size
I've been trying to up my TCP window size from the default 16K and it's caused nothing but problems. From the info I've found so far, these are the sysctl i've changed: kern.ipc.maxsockbuffer=2097152 net.inet.tcp.rfc1323=1 net.inet.tcp.sendspace=524288 net.inet.tcp.recvspace=524288 But if I do that, on boot I get all sorts of error messages about buffer space. i.e.: Jul 9 11:53:20 ccn64 portmap[180]: cannot create tcp socket: No buffer space available Jul 9 11:53:21 ccn64 inetd[199]: shell/tcp: socket: No buffer space available Jul 9 11:53:21 ccn64 inetd[199]: login/tcp: socket: No buffer space available Jul 9 11:58:55 ccn64 RPC::PlClient[243]: Cannot connect: No buffer space available Jul 9 11:58:55 ccn64 RPC::PlClient[246]: Cannot connect: No buffer space available Is there anything I'm missing? -- Joe LeKostaj - Just don't create a file called -rf. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
On 2001-07-10, Josh M Osborne scribbled: # The current Intel's have a shared bus, and all memory traffic goes # over it, and some cache coherency traffic as well. The official names of Intel's bus include: GTL, GTL+, AGTL and AGTL+. The new iTanic (aka Itanium) processor uses the AGTL+ protocol whereas the Pentium II/III use the GTL protocol. The Pentium 4 uses the GTL+ which allows for the quad-pumped 100Mhz FSB. I could have mixed up which processor uses which... but you get the idea :) # The AMD's/EV6's have a memory bus PER CPU plus a coherency bus. # I think the coherency bus may even be point-to-point between the # CPU and coherency controller, not a all the CPUs with the coherency # controller being responsible for routing messages as needed. If I read the specs correctly on the EV6 protocol... each CPU has a separate connection to the 'northbridge' chip. It's up to the northbridge to provide connectivity to the memory. # It is clearly a more expensive, more complex system. It also allows # much higher memory bandwidth (if two CPUs are looking at different # chunks of the address space they get their own path to memory). If # the coherency "bus" really is point-to-point the coherency controller # has to have a big chunk of SRAM, but you should be able to get # dramatically more CPUs to access memory quickly. The biggest problem is the number of traces required... which is more than double of that found in a single-processor configuration. Also, there is a memory bandwidth bottleneck if you have both processors hitting memory... there isn't a lot of bandwidth left open for other devices ;-) # That may explain why you can buy Alpha systems with 40+ CPUs, and # Intel XENON boxes with no more then eight (or is it four?). It is # also part of why the big Alphas are costly, but only part of it... 32-way machines are built differently than your 2-way or 4-way servers. Some use cellular multi-processing, some use NUMA, and many other techologies and concepts to allow massive number of processors within a single server. You can build a 32-way Xeon machine (Unisys has... NUMA-Q... which used to be Sequent, I believe has a 32-way configuration available) but they are very, very expensive... mostly when each 'pod' or 'cell' requires 2+ Meg of coherency cache... plus the numerous amounts of memory channels. -- Linh Pham [[EMAIL PROTECTED]] // 404b - Brain not found To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
In a message dated 07/10/2001 12:52:30 PM Eastern Daylight Time, [EMAIL PROTECTED] writes: > BSDi had no effect on ftp.freebsd.org's services and kept things > completely unchanged there which pretty much confirms my "no impact" statement in that area. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
On Tue, Jul 10, 2001 at 02:25:13PM -0400, Josh M Osborne wrote: > On Tue, Jul 10, 2001 at 12:42:12PM -0400, [EMAIL PROTECTED] wrote: > [...] > > > Instead, AMD implemented the Intel APIC specification; > > > I'm not sure if they did it by licensing the patent > > > (Intel had a patent on the APIC design), or if it's > > > just been long enough for it to come off patent > > > > The Athlon uses the Alpha's ev6 spec, not the intel spec. > That may explain why you can buy Alpha systems with 40+ CPUs, and 32 max, on Wildfire. > Intel XENON boxes with no more then eight (or is it four?). It is Eight, eg on a Compaq Proliant 8000. Or 32 for the Unisys (IIRC) CMP machines. -- | / o / / _ Arnhem, The Netherlands email: [EMAIL PROTECTED] |/|/ / / /( (_) Bulte "Youth is not a time in life, it is a state of mind" To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
ie ethernet device driver
I have two machines which still use the ie ethernet driver. It seems to have been disabled, but I'm not sure why. Can someone explain why it has been removed, or perhaps if it got broken, if it is feasible to put it back into -STABLE? The driver is still listed in a number of places including man, LINT, the source, etc. However, it is unusable because it has been removed from isa_compat.h so it cannot be recognized by the OS. In addition, there is an intermittent bug (PR 16214) with the driver. There is a fix for both of these problems here: http://www.jfitz.com/tips/freebsd_etherexpress16.html It modifies the following files, which are the culprits for the current lack of support for the ie driver: /usr/src/sys/dev/ie/if_ie.c and /usr/src/sys/i386/isa/isa_compat.h The fix works great and I've had no issues with the machines running it, even though I put a heavy load on both of them for network services. They shuffle mail, answer DNS queries, and provide ntp and nfs services for months without a problem. These machines and cards are old, granted. But there is a certain pride that I can take in keeping a 486DX/100 and a 386SX/16 as two of the most stable machines in a production environment. Regards, Steven To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
On Tue, Jul 10, 2001 at 12:42:12PM -0400, [EMAIL PROTECTED] wrote: [...] > > Instead, AMD implemented the Intel APIC specification; > > I'm not sure if they did it by licensing the patent > > (Intel had a patent on the APIC design), or if it's > > just been long enough for it to come off patent > > The Athlon uses the Alpha's ev6 spec, not the intel spec. The APIC covers things like how I/O interrupts are routed. The thing AMD licensed from DEC (or Compaq) is the ev6 "bus" protocol for keeping the cache contents coherent between CPUs. That is basically invisible to even OS software (as long as it works), other then altering how long memory references take. The current Intel's have a shared bus, and all memory traffic goes over it, and some cache coherency traffic as well. The AMD's/EV6's have a memory bus PER CPU plus a coherency bus. I think the coherency bus may even be point-to-point between the CPU and coherency controller, not a all the CPUs with the coherency controller being responsible for routing messages as needed. It is clearly a more expensive, more complex system. It also allows much higher memory bandwidth (if two CPUs are looking at different chunks of the address space they get their own path to memory). If the coherency "bus" really is point-to-point the coherency controller has to have a big chunk of SRAM, but you should be able to get dramatically more CPUs to access memory quickly. That may explain why you can buy Alpha systems with 40+ CPUs, and Intel XENON boxes with no more then eight (or is it four?). It is also part of why the big Alphas are costly, but only part of it... -- Not speaking for much of anyone, maybe not even myself To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On 10 Jul 2001, Rajappa Iyer wrote: :Jamie Bowden <[EMAIL PROTECTED]> writes: : :> On 10 Jul 2001, Rajappa Iyer wrote: : :> :One of the nice things I like about FreeBSD (and I daresay I'm not :> :alone in this) is that when I install it, I know that I'll get a :> :kernel with a corresponding full and functional userland. I see the :> :packaging of this `base system' as a bunch of (meta)packages as the :> :thin edge of the wedge---pretty soon FreeBSD will resemble the :> :hodge-podge collection of different (often conflicting) packages that :> :Linux is. :> :> Where as I see the ability to incrementally upgrade only the parts of the :> OS that have changed from release to release as I can do right now in :> Irix. : :Yes, I understand the argument, but fear that it's all too easy to get :into the kind of package mess that all Linux dists have. It would :require considerable amount of release engineering and testing to get :everything right. Do you really think it's worth expending that :amount of effort for an arguably minor improvement? I wouldn't call having that kind of improvement minor. For all inst/swmgr's warts, I can use it on a graphics or serial console with no problem. I can use it to do network installs of new machines or network upgrades on existing machines. Part of what makes it as usefull as it is is tied to the SGI Prom having network support built in, which most PC's lack (I'd say all, but someone, somewhere, would present a motherboard with onboard networking and a bios that had network support). This can be overcome with a minimal kernel (and even SGI goes this route with the miniroot). Is Irix the best Unix in the world? No, but there are some damn nice aspects to it, and it wouldn't hurt to emulate it where it shines. It will of course install from local CDROM or tape drive as well. One of Irix's shining features is its installation and package management systems, which are one and the same. You can load scripts from within it to automate with subsystems are installed (guaranteeing consistency of the machines you have to manage). Part of my default installation handles instified third party software, because I can. I could automate the process to the level of Solaris' Jumpstart if I chose, but don't have enough machines here to warrant that level of automation. And at anytime, I can remove a subsystem (like printing for example if I decide I don't need it) without manually digging through the OS directory tree and deleting bits (that something else may require that I wasn't aware of). Jamie Bowden -- "It was half way to Rivendell when the drugs began to take hold" Hunter S Tolkien "Fear and Loathing in Barad Dur" Iain Bowen <[EMAIL PROTECTED]> To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Terry Lambert wrote: > > If anyone is taking a vote, I disagree. I do not want any system > > ever assuming anything about my network. Even Win checks with the > > user before enabling DHCP. > > FYI: The networking bootstrap process I described above > is derived from the process used by Windows 98 and above, > as it comes configured by default on systems with integral > network cards. Personally, I don't consider win98 a reference point by which to model OS design. When you say win98 and above do you include the NT line (win2k)? With the _current_ IPv4 network, I don't see any good reason for servers to use DHCP, and FreeBSD is primarily a server OS, so why should it default to DHCP? > The "link.local" draft RFC for doing the > IPv4 stateless autoconfiguration was coauthored by a > Microsoft employee. I'm not familiar with the standards you reference above, what's the RFC#? > See the IETF "ZEROCONF" working group for more details: > this stuff is going to be part of the standards soon. Possibly. But then again, IPv6 will change a number of the rules as we know them. Which will be adopted and come into widespread use first is a matter for fortune tellers. -Bill To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Jamie Bowden <[EMAIL PROTECTED]> writes: > On 10 Jul 2001, Rajappa Iyer wrote: > :One of the nice things I like about FreeBSD (and I daresay I'm not > :alone in this) is that when I install it, I know that I'll get a > :kernel with a corresponding full and functional userland. I see the > :packaging of this `base system' as a bunch of (meta)packages as the > :thin edge of the wedge---pretty soon FreeBSD will resemble the > :hodge-podge collection of different (often conflicting) packages that > :Linux is. > > Where as I see the ability to incrementally upgrade only the parts of the > OS that have changed from release to release as I can do right now in > Irix. Yes, I understand the argument, but fear that it's all too easy to get into the kind of package mess that all Linux dists have. It would require considerable amount of release engineering and testing to get everything right. Do you really think it's worth expending that amount of effort for an arguably minor improvement? Regards, Rajappa -- <[EMAIL PROTECTED]> a.k.a. Rajappa Iyer. They also surf who stand in the waves. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Terry Lambert wrote: > Bill Moran wrote: > > Now, I've never used partition magic, but I (personally) > > find the FreeBSD partition program in sysinstall to be the > > easiest one I've ever used. What should be changed to make > > it easier? > > 1) Buy a new laptop > 2) Make the Windows partition smaller > 3) Install FreeBSD Touche ... considering the last time I installed FreeBSD on a laptop, the procedure was: 1) Boot from FreeBSD CD 2) Delete existing partition 3) Install FreeBSD OTOH: I don't see this as causing sysinstall's partition editor to be bad/worthless. How many other installers allow partition resizing (I don't know) Just add this feature (I'm not saying it would be easy, I'm saying that it doesn't require scrapping the existing system to add it, and lack of it does not invalidiate the quality/usefulness of what currently exists.) > > I disagree. I use sysinstall constantly. There's no easier > > way to install packages. > > ??? > > You suggested that people keep up to date using "cvsup"; Don't remember saying that, but I probably did ;) > but doing that won't result in new categories showing up > in "sysinstall", nor in your local packages archive being > updated to match your "cvsup" sources. You missed my point. I'm not defending sysintall in the previous paragraph. I'm defending an overall system maintenance utility that can be used for general stuff like changing network config, adding users, adding/removing software, etc. sysinstall does this now (whether badly or not). My point is only this: Do NOT assume that sysinstall is ONLY used during initial installation. It currently has the ability to help out long after the system is installed. Any utility that replaces it should be able to do the same. -Bill To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
a DDB Question.
I'm trying to track down a memory leak using ddb. DDB is a really
neat piece of work, but its not very well documented. I notice that
the man page is dated 1996, and folks have added useful things like
'call', that are undocumented.
Specifically, what it the correct syntax to examine a static variable
in a kernel routine? I built a little bit of code to help me out,
like this:
static int my_test_routine() {
static int counter;
return(counter++);
}
So exactly what would be the syntax to examine 'counter'?
Thanks.
--
Robert Sexton - [EMAIL PROTECTED], Cincinnati OH, USA
There's safety in numbers... Large prime numbers. - John Gilmore
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Jordan Hubbard wrote: > From: [EMAIL PROTECTED] > > These mechanisms existed before without BSDi, so there was no "impact". > > Actually, ftp downloads got a LOT slower after BSDi took over, so i consider > > it a negative impact in that area. > BSDi had no effect on ftp.freebsd.org's services and kept things > completely unchanged there, it was merely other factors which changed. > The Internet started to suck more and changing economic realities in > the ISP space forced the archive to move to the east coast, where > things only got progressively worse. This would have occurred even > sooner had Walnut Creek CDROM been involved and, in fact, we probably > would have pulled the plug a lot sooner since WC had far less money to > spend on things like that. I'm going to go out on a limb here ... There are how many comitters now? and how many developers? And how many people like me who are VERY reliant on FreeBSD and are not (yet) serious developers? Add all those numbers up and you get a whole LOT of people. The upshot being that keeping that many people informed accurately as to (for example) whether or not the BSDi merger caused the ftp site to slow down is going to be tough. Very tough. Damn near impossible. Yet, just about every one of those folks is going to know that BSDi bought Walnut Creek, and just about every one of those people is going to notice when the ftp site slows down. And a frightening percentage of them are going to put two and two together and come up with an answer that sounds right, even if it isn't. In the lack of an authoritative announcement such as "Because of Internet infrastructure problems we are moving ftp.freebsd.org and we don't know whether this will improve the situation or hurt it." Along with the subsequent announcements that the change didn't particularly help. Look at it, how many sources of (mis)information are there? All the mailing lists? The "announce" section of the website? Daemonnews? trash^H^H^H^H^Hslashdot? (who sometimes seem to hate BSD as much as they hate Microsoft) With all these, the inner workings of how one merger or another affects the big picture are simply not going to be broadly available. Making them broadly avialable would be a wonderful thing for the community. Perhaps an announcements page on the web site that is updated with more frequent announcements? Maybe a special mailing list that can be used to send out frequent announments as to what's going on in general. How would you keep speculation from getting out of hand? Personally, I don't know. But I think (overall) at least part of the solution to the arguments that have been occurring is to improve the communication channels. So that's my "going out on a limb" for today. Hope this dissertation is actually helpful/useful. -Bill To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Jordan Hubbard wrote: > > David: please install 4.3 --R E L E A S E--, the last > > official --R E L E A S E--. > > I T I S T H E R E. Man, Terry, do you really get off on being so > publically misinformed or is there something more pathological at work > here? :-) Yeah, there something more pathological: what the code says is not what the code appears to do, in practice, during an upgrade. As I told David, now I just have to find out why... -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
David O'Brien wrote: > There you go Terry, async mounts for installs. Thanks. This does not appear to work on upgrades for things like /usr/ports. Now I will have to go find out why; how annoying: yet more work. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Wes Peters wrote: > > I think the discussion is Re: binary upgrades, like putting > > in the CD and hitting that upgrade option, which right now > > doesn't quite get you there afaik. > > I don't think the goal was to make the system easier to upgrade, > but rather easier to subset. Do we really NEED to have sendmail > on every DNS server we put together? A properly designed system would support both. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Bill Moran wrote: > > If you pick "default installation" or "full installation", it > > _should_ try to be smart; if you pick "custom installation", > > you chould have to babysit it like you do today. > > > > In the "default" case, it should attempt to obtain a DHCP lease, > > and, failing that, ask the user to give it settings, or let > > them do IPv4 stateless autoconfiguration. Ad Hoc networking > > should always "just work". > > If anyone is taking a vote, I disagree. I do not want any system > ever assuming anything about my network. Even Win checks with the > user before enabling DHCP. FYI: The networking bootstrap process I described above is derived from the process used by Windows 98 and above, as it comes configured by default on systems with integral network cards. The "link.local" draft RFC for doing the IPv4 stateless autoconfiguration was coauthored by a Microsoft employee. See the IETF "ZEROCONF" working group for more details: this stuff is going to be part of the standards soon. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Bill Moran wrote: > Now, I've never used partition magic, but I (personally) > find the FreeBSD partition program in sysinstall to be the > easiest one I've ever used. What should be changed to make > it easier? 1) Buy a new laptop 2) Make the Windows partition smaller 3) Install FreeBSD #1 & #3 are doable, even though #3 is counter to all the training we put people through to get them Windows savvy. #2 can not be done in the FreeBSD partition program, without destroying your Windows partition, but Partition Magic can do it, no problem. > I disagree. I use sysinstall constantly. There's no easier > way to install packages. ??? You suggested that people keep up to date using "cvsup"; but doing that won't result in new categories showing up in "sysinstall", nor in your local packages archive being updated to match your "cvsup" sources. > > OTOH, if you really want to take a stab at a FreeBSD > > problem that will make you famous should you succeed, > > we'd all like to see a lovely, simple installer that > > will run on both direct attached VGA and a serial console, > > is lovely and well thought out, and intuitive to use. > > Extra points if you can run it over X across the network. > > Obviously, these improvements would be good. Part of this is concurrent version management. Right now, I can't make concurrent distribution images for things like PicoBSD. If I can install components on a one-off basis, I can build a "PicoBSD" cafeteria-style. For the serial console to work, the 2.88M boot floppy image needs to include a "/boot.config" with a "-P" in it, so that it comes up on video and keyboard if a keyboard is present, but comes up on serial, if one is not (the "-P" approach isn't perfect; as noted in the handbook, the probe code can fail to probe the keyboard unless it's a 101/102 key or better: it should use the BIOS instead, as the BIOS always gets it right). -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Re. The Foundation [was Re: FreeBSD Mall now BSDCentral]
The Foundation has yet to approach Wind River about the Trademark, so I cannot speculate on their disposition. >>> >>>What are your plans to change this situation? >> >>The Foundation isn't planning to do anything about the trademark or >>in regards to any of its other proposed activities until sufficient >>donations have arrived. I believe that the financial statement released >>with our announcement makes it clear why this must be the case. > > How much do you think it will cost to transfer the trademark? We're seeking legal counsel on this issue now. We should have a better estimate shortly. -- Justin To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Can someone verify this?
Paul Halliday schrieb: > > FreeBSD dissent.p450.box 4.3-RC FreeBSD 4.3-RC #3: Sun Jun 10 22:27:47 > EDT 2001 [EMAIL PROTECTED]:/usr/src/sys/compile/workstation > i386 > > FreeBSD useless.dell.box 4.3-STABLE FreeBSD 4.3-STABLE #6: Fri Jul 6 > 18:57:08 EDT 2001 > [EMAIL PROTECTED]:/usr/src/sys/compile/useless i386 > > mount /dev/acd0c /cdrom > should obviously fail, yet causes... > > panic: vm -fault on nofault entry, addr: c3e1e000 # FreeBSD msgate.ms-agentur.de 4.3-Stable FreeBSD 4.3-Stable#1: Mon Jul 9 20:26:31 CEST 2001 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/MSGATE i386 # dmesg|grep acd acd0: CDROM ata ata1-master using PIO4 # mount /dev/acd0c /cdrom /** with an ISO CD-ROM **/ mount: /dev/acd0c on /cdrom: incorrect super block # mount /dev/acd0c /cdrom /** without any disc **/ mount: /dev/acd0c: device busy No panic, no problem. Sorry, did not try with audio CDs -- none at work available HTH -Christoph Sold To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Re. The Foundation [was Re: FreeBSD Mall now BSDCentral]
>>> The Foundation has yet to approach Wind River about the Trademark, >>> so I cannot speculate on their disposition. >> >>What are your plans to change this situation? > >The Foundation isn't planning to do anything about the trademark or >in regards to any of its other proposed activities until sufficient >donations have arrived. I believe that the financial statement released >with our announcement makes it clear why this must be the case. How much do you think it will cost to transfer the trademark? -DG David Greenman Co-founder, The FreeBSD Project - http://www.freebsd.org President, TeraSolutions, Inc. - http://www.terasolutions.com Pave the road of life with opportunities. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
SMP works on athlons... while internally AMD uses the Alpha EV6 bus to give each CPU a full 200MHz point-to-point bus between it and it's RAM, to the OS, it just looks like any other intel based SMP machine. (It just runs faster) Ken On Mon, 9 Jul 2001, Nathan Vidican wrote: > I seem to recall a few discussions about the Dual Athlon buzz some > while back which had stated that the Athlon would essentially require a > completely different SMP spec than that currently utilized by the Intel > procesors. Assuming that this was true, one would assume that the O/S > too would require a different kind of SMP support in order to function > with these CPUs. > Unfortuneately, a recent thread has me at a bit of a loss here; in > that people seem to be speaking about the processor/smp chipset as > though they function just like Intel's do. Assuming that this > conflicting information is indeed correct, then would it not be > feasible to assume that the code currently implemented for using SMP > implementations under FreeBSD would be portable to the new Athlon MP > processor line? > The threads I'm speaking of, were to freebsd-questions most > recently wherein someone had been asking if the new Tyan ThunderK7 > motherboard would work with FreeBSD. The general concencus was 'why > not', from the responses I had read... but no one who answered really > seemed to know for sure. > Just for the record, is it or is it not possible to run SMP with > the new Athlon MP Processors; or has no-one even tried yet? Currently > the only O/S I know of which is promoting the usage of such systems is > Novell Netware, and I am just curious if FreeBSD will (if it is not > currently) be capable of running on such a system? > > > -- > Nathan Vidican > [EMAIL PROTECTED] > http://Nathan.Vidican.com/ > > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-questions" in the body of the message > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Jamie Bowden wrote: > You're expecting the whole world to keep the source tree on disk and > recompile the OS. Once I've done this, I cannot regress. This is > unrealistic in production environments. I can update Irix without > shutting down, and a single reboot at the end to load the new kernel. > > Everything is tracked via inst/swmgr, any part can be upgraded or > downgraded as necessary, including dependancies. In a production environment, the "downgraded" part is oftem the most critical of the two, since "upgrades" often aren't, and being able to "undo" one that is more trouble than it's worth is often worth the entire price of the system, when you are up against a wall. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Wes Peters wrote: > > Bill Moran wrote: > > > > Wes Peters wrote: > > > > > Oh, come now. FreeBSD's disk partitioning has always sucked. It does suck > > > somewhat less than many Linux linstallers, and a lot less than the OpenBSD > > > installer, but it still shoves way too many details and options at the > > > average user. Something akin to PartitionMagic would be an ideal way to go, > > > given unlimited resources to throw against this particular problem. > > > > Now, I've never used partition magic, but I (personally) find the > > FreeBSD > > partition program in sysinstall to be the easiest one I've ever used. > > What should be changed to make it easier? > > "How much of this disk do you want FreeBSD to use? ___%" This would be a nice little addition (and probably easy to add) but I don't see a tremendious gain from this. > Better yet, a nice graphical > view of the disk and the 4 possible entries in the partition table. I would classify this as "fluff and glitz", but that may just be me. > Allow > the user to grab the ends of partitions we can manipulate and move them > around, either through keyboard navigation or with a mouse. Hmmm ... we're really aiming at novice users here now, aren't we? I suppose that's not a bad thing, but I just never thought of going so far with it. > Focus on the > task we're attempting to accomplish: slicing the disk into 1 to 4 differnt > logical parts, rather than on the crufty underlying details. Hmmm ... well, I was never upset with the "crufty details" ... I rather like to know what's going on under the hood all the time. Then again, that's me. If you're targeting newbies and other less-educated (or less "I sure would like to figure this out" inclined) people, then what you're suggesting would probably be a good idea. OTOH: if you made it so the "crufty details" were no longer visible at all, I would be upset to NOT be able to see what was going on. > "Where am I, and what am I doing in this handbasket?" You're on Earth. No further explanation needed. -Bill To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Rasputin wrote: > > Where as I see the ability to incrementally upgrade only > > the parts of the OS that have changed from release to > > release as I can do right now in Irix. > > I may be low on caffeine, but I don't see how breaking up > the base system into packages makes it any easier to upgrade > than using cvsup? > > Id have thought it would require more work to upgrade > under some system similar to the ports tree (at least > that's my experience) We're talking packages, not ports. We're also talking about being able to maintain basic configuration control, without having to screw around without compiling the sources yourself. Consider binary upgrades for things like security alerts, which could happen automatically, based on whatever criteria you specify (including "root exploit" or "Never Do Anything Without My Permission"). In the worst case, you could need to compile a newer version of "sendmail" or "bind" than that which came with the system, to resolve an exploit. You would still want to end up with a "RELEASE plus known patch sets" when you were done, so that you could feel both comfortable about your ability to reproduce your production system, should you need to replace it or to scale to more customers. Running "some snapshot of STABLE" is not really the way to do this. And there is the commercial support issue: what constitutes a "supported configuration"? Certainly not a "checkout your source tree from your local repository copy using this date tag: XXX". Also, you should be aware that in commercial deployment, having a compiler on board the system is often considered a bad thing, as it permits entre to exploiters bringing their own programs onto the system. One of the things that TrustedBSD played around with is binary signatures, where it is not possible to run a binary that does not have a corresponding approved signature. In such a system, it's really imperitive that configuration management occur through a centralized binary blessed to install only blessed binaries. That _really_ precludes rebuilding from sources. > But like I said, I've probably misread this post. > > I thought the OP was referring to X in particular, and > since that's upgraded via ports anyway, it does seem a > good candidate to be installed by pkg_add (it's quite > confusing for newbies to "pkg_info | grep XFree " and > have it return nothing, especially when you're sat in > Enlightenment...) This really demonstrates the problem with having base system components (I include X11, bind, sendmail, etc. in this) that are not easily upgraded. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
From: Terry Lambert <[EMAIL PROTECTED]> Subject: Re: FreeBSD Mall now BSDCentral Date: Tue, 10 Jul 2001 09:02:06 -0700 > David: please install 4.3 --R E L E A S E--, the last > official --R E L E A S E--. I T I S T H E R E. Man, Terry, do you really get off on being so publically misinformed or is there something more pathological at work here? :-) - Jordan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
terry tumbled, > but the motherboards are "jumbo sized", and take a strange > power connector, so you can only get the power supply from > one vendor (so far). two, actually. > Instead, AMD implemented the Intel APIC specification; > I'm not sure if they did it by licensing the patent > (Intel had a patent on the APIC design), or if it's > just been long enough for it to come off patent The Athlon uses the Alpha's ev6 spec, not the intel spec. hawk -- Prof. Richard E. Hawkins, Esq. /"\ ASCII ribbon campaign [EMAIL PROTECTED] Smeal 178 (814) 375-4700 \ / against HTML mail These opinions will not be those of Xand postings Penn State until it pays my retainer. / \ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
From: [EMAIL PROTECTED] Subject: Re: FreeBSD Mall now BSDCentral Date: Tue, 10 Jul 2001 11:54:59 EDT > These mechanisms existed before without BSDi, so there was no "impact". > Actually, ftp downloads got a LOT slower after BSDi took over, so i consider > it a negative impact in that area. This is typical "Elvis hasn't been seen lately. Aliens are hard to spot too. Conclusion: Elvis is an Alien" thinking. BSDi had no effect on ftp.freebsd.org's services and kept things completely unchanged there, it was merely other factors which changed. The Internet started to suck more and changing economic realities in the ISP space forced the archive to move to the east coast, where things only got progressively worse. This would have occurred even sooner had Walnut Creek CDROM been involved and, in fact, we probably would have pulled the plug a lot sooner since WC had far less money to spend on things like that. Bill's points are valid too - looking exclusively on the dark side of things is an engineer's predilection that's not always fair or right, and though an honest discussion on what went wrong and right is fine, let's try to keep the crack-smoking conspiracy theories to a minimum. Thank you. - Jordan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Jamie Bowden wrote: > Where as I see the ability to incrementally upgrade only > the parts of the OS that have changed from release to release > as I can do right now in Irix. > > You know, it's funny that you told me Irix is antiquated not long ago > Terry, it has most of the feature set you seem to be looking for. I don't think I said that, I think I said SGI was a dead husk, after their suicidal leap... certainly, Irix has better cluster scalability than most OS's of any vintage, with XFS and their volume manager. The only OS that's even close is AIX. The one really antiquated thing I can point to in Irix is the ability to stage a quota denial of service attack on someone by "giving" them core dump files using the SVR3 chown semantics, but I think that's optional now that people have complained about it so long. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Rajappa Iyer wrote: > > Well, I'd sorta like to *see* them before writing the coding > > equivalent of a blank check, but given reasonably functional > > implementations, sure, I'd be happy to commit your "sysinstall > > mountpoint auto-discovery" and "release package metadata" > > enhancements. > > One of the nice things I like about FreeBSD (and I daresay I'm not > alone in this) is that when I install it, I know that I'll get a > kernel with a corresponding full and functional userland. I see the > packaging of this `base system' as a bunch of (meta)packages as the > thin edge of the wedge---pretty soon FreeBSD will resemble the > hodge-podge collection of different (often conflicting) packages that > Linux is. That would be an option, but it would certianly not be the default. I can install a lot of stuff today that I would claim would put my system in the "hodge-podge" category, most of them from ports, some of them from the net. PS: Note that you only get a kernel if you create your release from the GENERIC config; using a different config with the "make release" process results in a CDROM that boots and installs your "/kernel.MYCONFIG", but doesn't do the final step of copying that file to "/kernel", which renderes the resulting system unbootable. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Giorgos Keramidas wrote: > > The base system is not registered into the packages > > system, because of sysinstall. > > It's not installed from /usr/ports but from /usr/src. > I don't know if it's a good idea to have a huge > freebsd_base-5.0-current-20010624 in the packages list, or a zillion > freebsd_base-bin, freebsd_base-etc, etc. installed. BTW: I would settle for being able to select between base system components that are default (e.g. "bind" and "sendmail" and "perl", etc.), and those components from ports (e.g. "djbdns", and "postfix" and "newer than base version of perl") on an equivalency basis, but I don't see how you would manage this without the registration of base system components into the same configuration management system as the replacements you are permitted to select from for any given service or role. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Nik Clayton wrote: > It's reasonable to want to control what get's called FreeBSD. I never said it wasn't; I would just like to draw the line on the far side of sysinstall being what shows up first thing when you boot from a CDROM. > The intent here is not to prevent third party installers -- > they can be open source, closed source, or whatever mix you > want. If you want to produce a commercial distribution of > FreeBSD that does not use sysinstall as the default > installation mechanism then go right ahead, make it the > default, have it come up automatically when your customers > boot from CD, and so on. > > However, if you want to call it FreeBSD, then, somewhere, > sysinstall (and whatever replaces it) must be available. > Put it on "boot-legacy.flp" if you want, and strongly urge > your customers not to use it. But make it available to > those that want it. This is amazingly more reasonable than previous posts, which have all suggested that it must be possible to boot the CDROM to sysinstall. Effectively, doing so would require that sysinstall take the front seat, and that you put another text menu entry on it to pick your installer. > Then I can make sure that the Handbook chapter on installation says, > right at the beginning: I also understand the documentation issue. Look, I've been programming professionally for ~22 years now, and I didn't just step off the turnup truck: you'll find my code in BSD all the way back to when I wrote the FAQ and patchkit for 386BSD 0.1. I know what professional software developement, and consistency in presentation to the user means for a product. ...Please look at it from the perspective of someone willing to work on improving the initial impression that FreeBSD leaves in a user's mind: FreeBSD is behind in the game from the start, since PCs come installed with Windows, and without a seperate partition that can be easily and immediately usable by a third party OS. It drops further behind because of the difficulty of transferring experience over to using the FreeBSD tools from people who have been trained up in the Windows style guide. I think that any attempt to make the initial experience less painful will require a lot of work, and the ability to license "Partittion Magic" or a similar tool, right out of the box, and have it be the first thing people see (or, preferrably, have it be one of the things that people see, as seamlessly integrated into the overall look and feel of the installation process as possible). The "Partition Magic for FreeBSD" isn't going to happen without $$$ being involved, I think. Walnut Creek sold a FreeBSD package that included one that ran under Windows. But the barrier to entry is still too high to be able to capture a reasonable mindshare. I _personally_ do not want to build such a distribution; I think it would open up whoever did that to extreme friction with the FreeBSD project: Hell, even the mere act of contemplating such a thing has practically set off a firestorm. No Thanks! I'll work on the periphery problems, hopefully enabling someone else to do the deed. Frankly, I don't see much of this work happening, unless there is at worst nose-thumbing ande grudging cooperation from the project. As Jordan says: Walnut Creek CDROM is dead; someone has to take up the mantle -- but not me... not today. There are too many people looking for a back to stick arrows into, and I'm happy to let the indians focus their scalping on Wind River Systems for now. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, Jul 10, 2001 at 09:02:06AM -0700, Terry Lambert wrote:
> David O'Brien wrote:
> >
> > On Mon, Jul 09, 2001 at 01:29:23PM -0700, Terry Lambert wrote:
> > > Most of the work should be done using async mounts, or
> >
> > IT DOES. Terry please read the code...
>
>
> David: please install 4.3 --R E L E A S E--, the last
> official --R E L E A S E--.
/*
* Miscellaneous support routines..
*
* $FreeBSD: src/release/sysinstall/misc.c,v 1.40 1999/11/27 14:33:07 phk Exp $
..snip...
if (mount("ufs", mountpoint, RunningAsInit ? MNT_ASYNC | MNT_NOATIME : 0,
(caddr_t)&ufsargs) == -1) {
There you go Terry, async mounts for installs.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Dear BSD"guru", Can you please refrain from trolling here? There is no point in bemoaning the past. Come up with a decent software package, put it in the ports tree, then maybe somebody will listen to you. The way you are acting just put your address into my mailers trash filter. Sorry about the wasted bandwith -Christoph Sold [EMAIL PROTECTED] schrieb: > > [lots of pointless lines deleted] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
David O'Brien wrote: > > On Mon, Jul 09, 2001 at 01:29:23PM -0700, Terry Lambert wrote: > > Most of the work should be done using async mounts, or > > IT DOES. Terry please read the code... David: please install 4.3 --R E L E A S E--, the last official --R E L E A S E--. Don't install STABLE, and Don't Install CURRENT; I am well aware that the problems have been addressed since, but they have not been addressed by a --R E L E A S E--, or by any CDROM you can get without having to burn your own. Thanks, -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
In a message dated 07/10/2001 11:14:58 AM Eastern Daylight Time, [EMAIL PROTECTED] writes: > "How much of this disk do you want FreeBSD to use? ___%" > > Was that really so difficult to imagine? Better yet, a nice graphical > view of the disk and the 4 possible entries in the partition table. Allow > the user to grab the ends of partitions we can manipulate and move them > around, either through keyboard navigation or with a mouse. Focus on the > task we're attempting to accomplish: slicing the disk into 1 to 4 differnt > logical parts, rather than on the crufty underlying details. > You are assuming the "average user" understands how much to allocate for root, usr, var and swap, which they dont. But you can certainly make suggestions based on the available space. An important question is: are you designing this for unix gurus, or for the guy in the bookstore that wants to use FreeBSD? (or a windows guy who doesnt understand unix partition concepts?). Is your goal to make it easier for the unix guy to install freebsd, or to expand the market by making FreeBSD usable by the less-than-clueful computer professional? Bryan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: more on latency
Jim Bryant wrote: > Terry Lambert wrote: > > Kenneth Wayne Culver wrote: > > > I think I found the reason that my FreeBSD box is performing > > > so poorly as a NATing router. When I do an ipnat -l to see > > > what "active connections" are there on the router, a list > > > about 3 pages long (using ipnat -l | more) appears. I think > > > maybe it's having trouble because for every packet coming in > > > and out of the router, it's got to look at that list of > > > active connections for the right one to send to and from. Is > > > there any way to make connections that aren't being used go > > > away from the NAT faster? Thanks a lot. > > > > Don't run unnecessary daemons. > > > > The pcb lookups are a linear traversal, as well, and for > > a large number of connections, the calllout wheel for > > timers sucks. > > Is there a way to get similar stats from natd? I don't know; you could look at the netstat output from the tun interface it uses, and that would give you some of the flow information. In general, FreeBSD doesn't completely track SNMP RFC mandated statistics; I've helped a local person hack code out of netstat to do things like reporting of the number of active connections using UCD SNMP, but it's not common to find FreeBSD keeping stats that match up 100% with the MIB entries people normally like to see from the generic MIBs. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
In a message dated 07/09/2001 8:02:03 PM Eastern Daylight Time, [EMAIL PROTECTED] writes: > BSDi really did make an impact. Just because you fail to see one doesn't mean > it didn't happen. You might want to read the cvs-all mailing list. If you'd > like, I can supply you with some procmail filters that will put all of the > commits of people who are/were supported by WC/BSDi/WRS into a folder called > 'direct-positive-result-of-bsdi-funding'. > > I'm also unclear on how you have installed FreeBSD in the past few years, > please clarify: > > [ ] purchased CD from BSDi > [ ] downloaded from ftp.freebsd.org (which BSDi was paying for) > [ ] downloaded from snapshot servers (partnership of USWest and WC/BSDi) > [ ] downloaded from FTP mirror (who got their data from ftp.freebsd.org..) > [ ] other (doesn't matter, the source that was used to generate the > binaries was stored on WC/BSDi equipment) > These mechanisms existed before without BSDi, so there was no "impact". Actually, ftp downloads got a LOT slower after BSDi took over, so i consider it a negative impact in that area. They made little difference as far as public perception is concerned. Obviously you got some dollars from them, which is good for you. BSDi wanted to put freebsd on the map, the same map as Linux. They failed. Come to terms with it. B To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Package based configuration control
Bill Moran wrote: > > > While "going on forever" would actually be required in order > to fix things, I'm more interested in why you think that > sysintall needs "to go" - or did I misunderstand your previous > post? Do you think that there are simply so many problems > with sysinstall that it's not worth fixing? I guess I was > looking at the basic menu-driven layout and wondering > what you saw wrong with that. Well, for one thing, the packages menu puts up the wrong text field from the packages as the detailed description; for another, there are the navigation issues that Eric Melville is currently working on (space vs. tab vs. return). In general, I also think that the menus are too deep, and are not organized orthogonally (there should be as little difference as possible between the appearance of ports vs. packages vs. base system components, when it comes to installing them. It's also annoying that packages have to have their category headers compiled into the program as static strings, or they show up as two-headed stepchildren. But the main problem I have is that there are some things which are tracked, and some things which aren't; it should be the case that everything is tracked. A surprising number of people have expressed an interest in working on a package-based install in private email, so it's probably an idea whose time has come. We should probably resurrect the -config mailing list for this purpose. I've Cc:'ed that list on this response. If anyone wants to join the list, send a message to [EMAIL PROTECTED] With the body: subscribe [EMAIL PROTECTED] -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
Nathan Vidican wrote: > > I seem to recall a few discussions about the Dual Athlon > buzz some while back which had stated that the Athlon would > essentially require a completely different SMP spec than that > currently utilized by the Intel procesors. Assuming that this > was true, one would assume that the O/S too would require a > different kind of SMP support in order to function with these > CPUs. > Unfortuneately, a recent thread has me at a bit of a loss here; in > that people seem to be speaking about the processor/smp chipset as > though they function just like Intel's do. Assuming that this > conflicting information is indeed correct, then would it not be > feasible to assume that the code currently implemented for using SMP > implementations under FreeBSD would be portable to the new Athlon MP > processor line? You should check the SMP list archives. The systems work fine with FreeBSD, as it currently sits, but the motherboards are "jumbo sized", and take a strange power connector, so you can only get the power supply from one vendor (so far). There was an attempt to create a non-Intel standard called OpenAPIC, but no one implemented motherboards that supported it, so that attempt died. Instead, AMD implemented the Intel APIC specification; I'm not sure if they did it by licensing the patent (Intel had a patent on the APIC design), or if it's just been long enough for it to come off patent (I seem to remember the external 386 APIC chips were out in 1984 or a little after that, which would put them after the 1998 date for 14 years from date of issue for the patents on them). -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Inconsistency with wchar_t / wint_t in current
Dear fellow hackers
I stumbled over an inconsistency with the data types wchar_t and wint_t.
My machine is a FreeBSD-5.0 current as of July 8th, 2001.
This following simple source code breaks unless You modify ,
see below.
/* stddef.h is included by curses.h, but here I want to make sure the
* modified version of stddef.h is picked up !
*/
#include
#include
int
main (char *argv[], int argc) {
printf ("size of wchar_t is %d\n", sizeof (wchar_t));
printf ("size of wint_t is %d\n", sizeof (wint_t));
exit (0);
return 0;
}
The error is:
gcc -I. -O -Os -pipe -s -o mess mess.c
mess.c: In function `main':
mess.c:8: `wint_t' undeclared (first use in this function)
mess.c:8: (Each undeclared identifier is reported only once
mess.c:8: for each function it appears in.)
*** Error code 1
Stop in /usr/tmp/wchar-mess.
There are two different definitions of the both data types wchar_t and wint_t.
Please refer to the output of the commands at the end of this mail.
We have (n)curses.h, defining them both unsigned long;
We have runetype,stddef,stdlib,wchar,wctype, defining them as
_BSD_WCHAR_T_ / _BSD_WINT_T_ resp.
_BSD_W{CHAR,INT}_T_ are defined in .
Also, (n)curses is FSF originated, as oulined in the disclaimer.
Now, I expected some symmetry here, note that wint_t is not defined in either
{runetype,stddef,stdlib,wchar}.h.
Also, wchar_t is not defined in wctype.h.
The definitions in curses.h are dangerous, they should use _BSD_WCHAR_T_ /
_BSD_WINT_T_, and both data types must be defined in the header files mentioned
above.
Is somebody working on reconciling the header files ?
Eventually I fixed the build by inserting the following lines in a local
stddef.h:
*** /usr/include/stddef.h Fri May 25 02:29:30 2001
--- stddef.hMon Jul 9 21:02:25 2001
***
*** 61
--- 62,66
+ #ifdef_BSD_WINT_T_
+ typedef _BSD_WINT_T_wint_t;
+ #undef_BSD_WINT_T_
+ #endif
+
Ciao, derweil,
--
Carlo
PS: Here are the revealing commands !
"fgrep wchar_t /usr/include/*.h /usr/include/sys/*.h | fgrep typedef" shows
/usr/include/curses.h:typedef unsigned long wchar_t;
/usr/include/ncurses.h:typedef unsigned long wchar_t;
/usr/include/runetype.h:typedef _BSD_WCHAR_T_ wchar_t;
/usr/include/stddef.h:typedef _BSD_WCHAR_T_ wchar_t;
/usr/include/stdlib.h:typedef _BSD_WCHAR_T_ wchar_t;
/usr/include/wchar.h:typedef_BSD_WCHAR_T_ wchar_t;
Whereas "fgrep wint_t /usr/include/*.h /usr/include/sys/*.h | fgrep typedef"
results in
/usr/include/curses.h:typedef long int wint_t;
/usr/include/ncurses.h:typedef long int wint_t;
/usr/include/wchar.h:typedef_BSD_WINT_T_wint_t;
/usr/include/wctype.h:typedef _BSD_WINT_T_wint_t;
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
Updated Aironet Sniffing patches for -stable
I have new patches for the Aironet sniffing and some major code clean-up of duplicated structures and defines it requires the latest -stable. http://www.ambrisko.com/doug/an/an.patch.cisco.rfmon2+ifconfig5 Applies to /usr/src if you don't have the linux ioctl patch from: http://www.ambrisko.com/doug/an/ installed then undef LINUX_COMPAT. Please test and then I will send-pr the patches in. Note this patch touches tcpdump & libpcap to add 802.11 support so a make world & kernel is recommended. This won't be needed when our tcpdump & libpcap is updated. Doug A. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Laurence Berland wrote: > > On Tue, 10 Jul 2001, Rasputin wrote: > > > I may be low on caffeine, but I don't see how breaking up the base system > > into packages makes it any easier to upgrade than using cvsup? > > I think the discussion is Re: binary upgrades, like putting in the CD and > hitting that upgrade option, which right now doesn't quite get you there > afaik. I don't think the goal was to make the system easier to upgrade, but rather easier to subset. Do we really NEED to have sendmail on every DNS server we put together? -- "Where am I, and what am I doing in this handbasket?" Wes Peters Softweyr LLC [EMAIL PROTECTED] http://softweyr.com/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Bill Moran wrote: > > Wes Peters wrote: > > > Oh, come now. FreeBSD's disk partitioning has always sucked. It does suck > > somewhat less than many Linux linstallers, and a lot less than the OpenBSD > > installer, but it still shoves way too many details and options at the > > average user. Something akin to PartitionMagic would be an ideal way to go, > > given unlimited resources to throw against this particular problem. > > Now, I've never used partition magic, but I (personally) find the > FreeBSD > partition program in sysinstall to be the easiest one I've ever used. > What should be changed to make it easier? "How much of this disk do you want FreeBSD to use? ___%" Was that really so difficult to imagine? Better yet, a nice graphical view of the disk and the 4 possible entries in the partition table. Allow the user to grab the ends of partitions we can manipulate and move them around, either through keyboard navigation or with a mouse. Focus on the task we're attempting to accomplish: slicing the disk into 1 to 4 differnt logical parts, rather than on the crufty underlying details. -- "Where am I, and what am I doing in this handbasket?" Wes Peters Softweyr LLC [EMAIL PROTECTED] http://softweyr.com/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, Jul 10, 2001 at 02:26:43PM +0200, Olaf Hoyer wrote: > We agreed to do something in Europe: > > - providing informational structure for BSD > - providing a channel where people that want to do booths at exhibitions etc > may contact and get some help, pre-financing (perhaps) and merchandise articles > to help financing the whole thing. Can I point you at the BSD EU Group mailing list, [EMAIL PROTECTED], where much the same discussions are happening at the moment. N -- FreeBSD: The Power to Serve http://www.freebsd.org/ FreeBSD Documentation Project http://www.freebsd.org/docproj/ --- 15B8 3FFC DDB4 34B0 AA5F 94B7 93A8 0764 2C37 E375 --- PGP signature
Re: wx0 jumbo frame support explosions..
Yay. :^) Timeframe? On Tue, 10 Jul 2001, Matthew Jacob wrote: > Not debugged yet. > > > On Mon, 9 Jul 2001, Geoff Mohler wrote: > > > When I enable jumbo frames in /usr/src/sys/pci.if_wx.c, and then set it > > via 'ifconfig wx0 mtu 9000' once the new kernel is booted..my system > > immediately goes zonkers...not even healthy enough to log. Just kernel > > panic and reboot. > > > > Idears? > > > > --- > > ** > > *New & Improved: http://www.speedtoys.com* > > ** > > > > > > To Unsubscribe: send mail to [EMAIL PROTECTED] > > with "unsubscribe freebsd-hackers" in the body of the message > > > > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-hackers" in the body of the message > --- ** *New & Improved: http://www.speedtoys.com* ** To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Terry Lambert wrote: > If you pick "default installation" or "full installation", it > _should_ try to be smart; if you pick "custom installation", > you chould have to babysit it like you do today. > > In the "default" case, it should attempt to obtain a DHCP lease, > and, failing that, ask the user to give it settings, or let > them do IPv4 stateless autoconfiguration. Ad Hoc networking > should always "just work". If anyone is taking a vote, I disagree. I do not want any system ever assuming anything about my network. Even Win checks with the user before enabling DHCP. -Bill To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Wes Peters wrote: > Oh, come now. FreeBSD's disk partitioning has always sucked. It does suck > somewhat less than many Linux linstallers, and a lot less than the OpenBSD > installer, but it still shoves way too many details and options at the > average user. Something akin to PartitionMagic would be an ideal way to go, > given unlimited resources to throw against this particular problem. Now, I've never used partition magic, but I (personally) find the FreeBSD partition program in sysinstall to be the easiest one I've ever used. What should be changed to make it easier? > The main problem with sysinstall all along has been that it is a one-off > program. There is no glitz in writing installers, and they're only used > once, then rarely if ever again. I disagree. I use sysinstall constantly. There's no easier way to install packages. I also direct newbies to use sysinstall to tweak parameters and install software. Especially on my home desktop system where I'm constantly installing/uninstalling software. As previously stated, I find the partitioning a perfect combination of easy to use/easy to control that I almost always use it when adding disks. While I won't disagree that there are things in sysinstall that can be improved (for example, you seem to get lost in menus at times) I don't think it should ever be pawned off as something that a user only sees once. Lurk -questions for a while and see how often a newbie is directed back to sysinstall to correct something or install a package. > OTOH, if you really want to take a stab at a FreeBSD problem that will > make you famous should you succeed, we'd all like to see a lovely, simple > installer that will run on both direct attached VGA and a serial console, > is lovely and well thought out, and intuitive to use. Extra points if you > can run it over X across the network. Obviously, these improvements would be good. -Bill To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, 10 Jul 2001, Rasputin wrote: > I may be low on caffeine, but I don't see how breaking up the base system > into packages makes it any easier to upgrade than using cvsup? I think the discussion is Re: binary upgrades, like putting in the CD and hitting that upgrade option, which right now doesn't quite get you there afaik. > > Id have thought it would require more work to upgrade under some system > similar to the ports tree (at least that's my experience) > > But like I said, I've probably misread this post. > > I thought the OP was referring to X in particular, and since that's > upgraded via ports anyway, it does seem a good candidate to be > installed by pkg_add (it's quite confusing for newbies to > "pkg_info | grep XFree " and have it return nothing, especially when > you're sat in Enlightenment...) L: http://www.isp.northwestern.edu/~laurence To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
> > > > ... or maybe we should remind ourselves that the only thing > > official about FreeBSD is the code. Let the CD vendors figure > > out ways to attract customers from each other, lets worry more > > about ways to attract 'customers' from other operating systems. > >Bill and Jordan are right on this one, guys. We (as in FreeBSD) can put >up the ISOs and every Tom, Dick and Harry with a CD burner can distribute >FreeBSD, and differentiate themselves on packaging, sales channel, and >customer service. Who knows, we might even get a few local shops to pre- >install FreeBSD on a machine or two, with their own FreeBSD discs thrown >in. It could happen. I'll talk to SuperDale and see if "Totally Awesome >Computers" will do this, they run their web site on FreeBSD. Hi folks! Well, I'm just returning from german Linuxtag, one of the biggest events concernung free software. (http://www.linuxtag.org) We ran also a booth there, providing information ( I also held some speech there, covering the history of BSD) and some contacts. We agreed to do something in Europe: - providing informational structure for BSD - providing a channel where people that want to do booths at exhibitions etc may contact and get some help, pre-financing (perhaps) and merchandise articles to help financing the whole thing. Basically, those people who showed up at our booth asked: Well, ok, it seems that this is a good OS, but where/whom can I contact in case of trouble? Who provides _commercial_ support? I'd love to buy a T-Shirt or a pin, can you sell me one? I already use BSD, but are there any books about it? ((No, not the handbook. People love to pay for additional literature helping make the bookshelf look cool.) So there certainly is money involved, and if we could organize it in a way, that money flows back to the project respectively into activities that help promoting BSD, that should be fine. This also includes the possibility (which needs to be checked for legal/trade commision issued yet) to provide any user with the possibility to order some CDs, T-Shirts whatsoever. BTW: BSD stands herein for: Free/Net/OpenBSD, BSD/OS, MacOS X, as far the BSD portion is concerned. Any input? Maybe we should take this over to -chat. Olaf To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: vfs.vmiodirenable undocumented
On Tue, 10 Jul 2001 14:13:23 +0200, Sheldon Hearn wrote: > Someone recently suggested that I tune vfs.vmiodirenable on a system > with lots of memory. The CVS commit logs and the source tell me > absolutely nothing about what this tunable does. > > Is anyone in a position to document it? Someone mailed me privately and pointed out that the sysctl is documented in tuning(7). Thanks, Sheldon. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: more on latency
On Monday, 9th July 2001, Terry Lambert wrote: >Kenneth Wayne Culver wrote: >> >> I think I found the reason that my FreeBSD box is performing >> so poorly as a NATing router. When I do an ipnat -l to see >> what "active connections" are there on the router, a list >> about 3 pages long (using ipnat -l | more) appears. I think >> maybe it's having trouble because for every packet coming in >> and out of the router, it's got to look at that list of >> active connections for the right one to send to and from. Is >> there any way to make connections that aren't being used go >> away from the NAT faster? Thanks a lot. > >Don't run unnecessary daemons. > >The pcb lookups are a linear traversal, as well, and for >a large number of connections, the calllout wheel for >timers sucks. I can't imagine even the most inefficiently coded linear traversal causing this problem given the beefy machine being used. I set up a cable sharing system for friends of mine and it is a Pentium 100 with 2 ISA NICs! That system adds no more than 2 or so ms to the latency with 3 simultaneous counterstrike players. I used ipfw and natd in a trivial configuration on 4.3-R. I wonder if the problem is a lack of mbufs or some similar misconfiguration tragedy. "netstat -m" and "top" output might be helpful. Stephen. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
vfs.vmiodirenable undocumented
Someone recently suggested that I tune vfs.vmiodirenable on a system with lots of memory. The CVS commit logs and the source tell me absolutely nothing about what this tunable does. Is anyone in a position to document it? Ciao, Sheldon. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, 10 Jul 2001, Rasputin wrote: :* Jamie Bowden <[EMAIL PROTECTED]> [010710 12:42]: :> On 10 Jul 2001, Rajappa Iyer wrote: : :> :One of the nice things I like about FreeBSD (and I daresay I'm not :> :alone in this) is that when I install it, I know that I'll get a :> :kernel with a corresponding full and functional userland. I see the :> :packaging of this `base system' as a bunch of (meta)packages as the :> :thin edge of the wedge---pretty soon FreeBSD will resemble the :> :hodge-podge collection of different (often conflicting) packages that :> :Linux is. : :> Where as I see the ability to incrementally upgrade only the parts of the :> OS that have changed from release to release as I can do right now in :> Irix. : :I may be low on caffeine, but I don't see how breaking up the base system :into packages makes it any easier to upgrade than using cvsup? :Id have thought it would require more work to upgrade under some system :similar to the ports tree (at least that's my experience) :But like I said, I've probably misread this post. You're expecting the whole world to keep the source tree on disk and recompile the OS. Once I've done this, I cannot regress. This is unrealistic in production environments. I can update Irix without shutting down, and a single reboot at the end to load the new kernel. Everything is tracked via inst/swmgr, any part can be upgraded or downgraded as necessary, including dependancies. Jamie Bowden -- "It was half way to Rivendell when the drugs began to take hold" Hunter S Tolkien "Fear and Loathing in Barad Dur" Iain Bowen <[EMAIL PROTECTED]> To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: g++ and substring extraction
On Tue, Jul 10, 2001 at 12:41:55PM +0200, Volker Sturm wrote: > Hi, > I am using g++ 2.95.3 on 4.3-STABLE. I want to do some string operations > in one > of my functions. The manual for g++ says that there are member functions > like > somestring.before(i); or > somestring.at(0, i); Did you try the substr method ? For example string s2 = s1.substr(pos,length); --gt PGP signature
Re: FreeBSD Mall now BSDCentral
* Jamie Bowden <[EMAIL PROTECTED]> [010710 12:42]: > On 10 Jul 2001, Rajappa Iyer wrote: > :One of the nice things I like about FreeBSD (and I daresay I'm not > :alone in this) is that when I install it, I know that I'll get a > :kernel with a corresponding full and functional userland. I see the > :packaging of this `base system' as a bunch of (meta)packages as the > :thin edge of the wedge---pretty soon FreeBSD will resemble the > :hodge-podge collection of different (often conflicting) packages that > :Linux is. > Where as I see the ability to incrementally upgrade only the parts of the > OS that have changed from release to release as I can do right now in > Irix. I may be low on caffeine, but I don't see how breaking up the base system into packages makes it any easier to upgrade than using cvsup? Id have thought it would require more work to upgrade under some system similar to the ports tree (at least that's my experience) But like I said, I've probably misread this post. I thought the OP was referring to X in particular, and since that's upgraded via ports anyway, it does seem a good candidate to be installed by pkg_add (it's quite confusing for newbies to "pkg_info | grep XFree " and have it return nothing, especially when you're sat in Enlightenment...) -- Oh, I don't blame Congress. If I had $600 billion at my disposal, I'd be irresponsible, too. -- Lichty & Wagner Rasputin :: Jack of All Trades - Master of Nuns :: To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: Athlon MP / AMD 760MP Chipset (Athlon SMP question)
"Nathan Vidican" wrote: > I seem to recall a few discussions about the Dual Athlon buzz some > while back which had stated that the Athlon would essentially require a > completely different SMP spec than that currently utilized by the Intel > procesors. Assuming that this was true, one would assume that the O/S > too would require a different kind of SMP support in order to function > with these CPUs. It works fine. AMD implemented Intel MPSPEC 1.4 for SMP and it is closer to compliance than most Intel / serverworks systems. There is no magic required. I have a thunder K7 for my desktop with dual 1.2GHz AthlonMP's. All 4.x+ releases will boot on it. The only gotcha is that the older releases dont recognize the 766 IDE controller and run in biosdma mode instead of UDMA66/100. http://people.freebsd.org/~peter/thunderk7.txt I asked the Tyan people about the special power connector.. That's there solely for the AGPPro support. Other motherboards that have AGPPro have a second power connector. The base system uses nowhere near the power that the 460W power supplies are capable of, unless you start using the 200Watt+ double height AGPPro slot with the extra fingers for power feeds. Cheers, -Peter -- Peter Wemm - [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] "All of this is for nothing if we don't go to the stars" - JMS/B5 To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On 10 Jul 2001, Rajappa Iyer wrote: :Jordan Hubbard <[EMAIL PROTECTED]> writes: : :> Well, I'd sorta like to *see* them before writing the coding :> equivalent of a blank check, but given reasonably functional :> implementations, sure, I'd be happy to commit your "sysinstall :> mountpoint auto-discovery" and "release package metadata" :> enhancements. : :One of the nice things I like about FreeBSD (and I daresay I'm not :alone in this) is that when I install it, I know that I'll get a :kernel with a corresponding full and functional userland. I see the :packaging of this `base system' as a bunch of (meta)packages as the :thin edge of the wedge---pretty soon FreeBSD will resemble the :hodge-podge collection of different (often conflicting) packages that :Linux is. Where as I see the ability to incrementally upgrade only the parts of the OS that have changed from release to release as I can do right now in Irix. You know, it's funny that you told me Irix is antiquated not long ago Terry, it has most of the feature set you seem to be looking for. Jamie Bowden -- "It was half way to Rivendell when the drugs began to take hold" Hunter S Tolkien "Fear and Loathing in Barad Dur" Iain Bowen <[EMAIL PROTECTED]> To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
On Tue, Jul 10, 2001 at 03:51:15AM +, Terry Lambert wrote: > ] I never said a "CDROM must boot to sysinstall" and I challenge you to > ] find a quote to that effect. What both Nik and I said was that it > ] must be an OPTION to do so, somehow, or you haven't provided a stock > ] FreeBSD experience and people are potentially going to be confused as > ] to what "FreeBSD" means. This holds especially true if you haven't > ] provided source code to your wizzy installer and they have no way of > ] figuring out how or why it's even misbehaving, which you can bet it > ] will since nobody ever writes perfect software. > > You know that the CDROM boot process is based on a 2.88M floppy > image. If it "must be an OPTION to do so", then you are saying > the same thing: that it must boot to sysinstall. Not at all. If you're thinking of shipping a CD, make sure that somewhere on that CD is a copy of the existing two floppy images, a copy of fdimage (or similar), and a paragraph of documentation that says: If you would prefer to use the standard FreeBSD installer (sysinstall) then take two blank floppies. Insert the first floppy, and run "fdimage mfsroot.flp a:", then insert the second floppy and run "fdimage boot.flp a:". Leave that floppy in the drive, and reboot (making sure to set your BIOS to "boot from floppy"). The FreeBSD Project installer will then start. Using this installer is not documented here (we strongly urge you to use the installer we provide and document) but it is covered in chapter 2 of the FreeBSD Handbook, at http://www.FreeBSD.org/. That would completely satisfy the requirements I've outlined. N -- FreeBSD: The Power to Serve http://www.freebsd.org/ FreeBSD Documentation Project http://www.freebsd.org/docproj/ --- 15B8 3FFC DDB4 34B0 AA5F 94B7 93A8 0764 2C37 E375 --- PGP signature
g++ and substring extraction
Hi, I am using g++ 2.95.3 on 4.3-STABLE. I want to do some string operations in one of my functions. The manual for g++ says that there are member functions like somestring.before(i); or somestring.at(0, i); Problem is that the compiler complains that these member functions don't exist. It only recognizes somestring.at(i); So I did a not very satisfying rebuild of the above functionality by a for(;;)-construct and concatenated the desired string character-wise. Is there a compiler option to be set so that I can access this advanced functionality? Does the FreeBSD-port miss some features? Volker Sturm To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: best way to migrate to a new disk
Steven Ames schrieb: > > > > Don't use tar. It loses devices, can't handle holey files well and a > > > number of other minor clitches. Use dump instead. > > Hrm... what about 'rsync'? Does it suffer from the same problems as 'tar'? > I use rsync a lot because its incremental. This is off topic from migrating > to > a new disk, but I am curious... rsync won't handle holey files at all. Think about the space you're wasting. Stay with dump, or peruse dd. HTH -Christoph Sold To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: wx0 jumbo frame support explosions..
Not debugged yet. On Mon, 9 Jul 2001, Geoff Mohler wrote: > When I enable jumbo frames in /usr/src/sys/pci.if_wx.c, and then set it > via 'ifconfig wx0 mtu 9000' once the new kernel is booted..my system > immediately goes zonkers...not even healthy enough to log. Just kernel > panic and reboot. > > Idears? > > --- > ** > *New & Improved: http://www.speedtoys.com* > ** > > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-hackers" in the body of the message > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
wx0 jumbo frame support explosions..
When I enable jumbo frames in /usr/src/sys/pci.if_wx.c, and then set it via 'ifconfig wx0 mtu 9000' once the new kernel is booted..my system immediately goes zonkers...not even healthy enough to log. Just kernel panic and reboot. Idears? --- ** *New & Improved: http://www.speedtoys.com* ** To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
Re: FreeBSD Mall now BSDCentral
Jordan Hubbard <[EMAIL PROTECTED]> writes: > Well, I'd sorta like to *see* them before writing the coding > equivalent of a blank check, but given reasonably functional > implementations, sure, I'd be happy to commit your "sysinstall > mountpoint auto-discovery" and "release package metadata" > enhancements. One of the nice things I like about FreeBSD (and I daresay I'm not alone in this) is that when I install it, I know that I'll get a kernel with a corresponding full and functional userland. I see the packaging of this `base system' as a bunch of (meta)packages as the thin edge of the wedge---pretty soon FreeBSD will resemble the hodge-podge collection of different (often conflicting) packages that Linux is. Rajappa -- <[EMAIL PROTECTED]> a.k.a. Rajappa Iyer. They also surf who stand in the waves. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
DDB 'kill' command
Hi folks,
The attached patch implements a 'kill' command in DDB. Previously, it
was possible to do `call psignal(xxx,yyy)` and have it DTRT. (This
was very useful when you accidently got your system so deep in the
hole that spawning kill(1) takes forever and even then possibly
doesn't succeed.) However, psignal() doesn't respect locking by
itself, so trying that now leads to all kinds of badness. This patch
basically wraps the psignal() call in a 'kill' command that respects
all necessary locks.
Actually, it isn't very clear exactly which locks it should respect.
The debugger is a special case in this way. This patch uses the
PROC_TRYLOCK macro; if it fails, the command bails out. Thus, it
can't use pfind()--the latter automatically does a PROC_LOCK--so it
has to walk the allproc list manually. It does *not* attempt to get a
shared lock on the allproc list. There is similar code in db_trace.c,
and it doesn't call sx_slock(), either; I asked jhb about this a while
ago, and all he said was that it is intentional.
Comments? Suggestions?
Thanks,
Dima Dorfman
[EMAIL PROTECTED]
Index: db_command.c
===
RCS file: /stl/src/FreeBSD/src/sys/ddb/db_command.c,v
retrieving revision 1.36
diff -u -r1.36 db_command.c
--- db_command.c2001/07/08 04:56:05 1.36
+++ db_command.c2001/07/10 07:19:45
@@ -36,7 +36,12 @@
*/
#include
#include
+#include
+#include
+#include
+#include
#include
+#include
#include
#include
@@ -62,6 +67,7 @@
static db_cmdfcn_t db_fncall;
static db_cmdfcn_t db_gdb;
+static db_cmdfcn_t db_kill;
/* XXX this is actually forward-static. */
extern struct command db_show_cmds[];
@@ -413,6 +419,7 @@
{ "show", 0, 0, db_show_cmds },
{ "ps", db_ps, 0, 0 },
{ "gdb",db_gdb, 0, 0 },
+ { "kill", db_kill,CS_OWN, 0 },
{ (char *)0, }
};
@@ -566,4 +573,44 @@
db_printf("Next trap will enter %s\n",
boothowto & RB_GDB ? "GDB remote protocol mode"
: "DDB debugger");
+}
+
+static void
+db_kill(dummy1, dummy2, dummy3, dummy4)
+ db_expr_t dummy1;
+ boolean_t dummy2;
+ db_expr_t dummy3;
+ char * dummy4;
+{
+ struct proc *p;
+ db_expr_t pid, sig;
+
+#define DB_ERROR(f)do { db_printf f; db_flush_lex(); return; } while (0)
+
+ /* Retrieve arguments. */
+ if (!db_expression(&sig))
+ DB_ERROR(("Missing signal number\n"));
+ if (!db_expression(&pid))
+ DB_ERROR(("Missing process ID\n"));
+ db_skip_to_eol();
+ if (sig < 0 || sig > _SIG_MAXSIG)
+ DB_ERROR(("Signal number out of range\n"));
+
+ /* Find the process in queston. */
+ /* sx_slock(&allproc_lock); */
+ LIST_FOREACH(p, &allproc, p_list)
+ if (p->p_pid == pid)
+ break;
+ /* sx_sunlock(&allproc_lock); */
+ if (p == NULL)
+ DB_ERROR(("Can't find process with pid %d\n", pid));
+
+ /* If it's already locked, bail; otherwise, do the deed. */
+ if (PROC_TRYLOCK(p) == 0)
+ DB_ERROR(("Can't lock process with pid %d\n", pid));
+ else {
+ psignal(p, sig);
+ PROC_UNLOCK(p);
+ }
+#undef DB_ERROR
}
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
