Re: Installation Problems on Dell PowerEdge 6100/200

[James C. Beyer]

We have a machine here in the lab that is the precurser of your machine
that we also had boot problems with.  

What I found was that the adaptec BIOS Disks > 1GB support was breaking
the linux and FreeBSD boot routine.  When I disabled Disks > 1GB my
problems were solved.  It seems the BIOS setting is for machines running
OSes that need help looking at large disks.

I hope this helps.  If you already have your machine working great; I hope
this my help someone else then.

james

On Tue, 8 Aug 2000, Daniel Lang wrote:

> Hi,
> 
> we've got a Dell PowerEdge 6100/200 here, which is an older
> SMP server, featuring 4 PPro 200 CPU's two internal AIC7880
> channels and an Adaptec 2940UW PCI controller. One of the
> AIC's is connected to a SCA Backplane that holds 4 disks.
> 
> Now before even having a chance to see if FreeBSD-SMP
> works with this box, I didn't manage to install FreeBSD
> (4.1) correctly. Well, installation seems to go well in
> any attempt:
> 
> - Floppy boots, Kernel finds all ahc's and disks
> - Installation seems to succeed (Partitioning, installation, etc)
> 
> But then, the machine won't boot. It seems to happen, that the
> MBR of the target device (da0, which is the first disk on the
> second controller, channel A of internal AIC's) is found, but
> the next stage of the bootstrapping process cannot be found.
> I played also around using boot0 and installing on other disks,
> this is like what I got:
> 
> Standard MBR and System on da0: -> Missing operating system
> boot0 MBR and System on da0:-> F1 -> *beep* (nothing else)
> boot0 MBR (and old System) on da0, standard MBR and System on da1:
> booting from da0: F1 -> *beep*, F5 (disk2) -> Missing operating system
> booting from da1: -> Missing operating system
> 
> So it seems no boot-block after the MBR can be found.
> 
> I tried: - installing and booting from different disks on internal ahc
>  - disabling some of the controllers 
>  - installing and booting from a disk on the 2940,
>as well while disabling the others
> 
> It all had no effect.
> 
> All adaptec's BIOS has Disks > 1GB and INT13 enabled (of course the
> BIOS itself is enabled, too)
> 
> Previously Solaris 7/x86 was running on this machine, there
> were no such problems, but we don't really want to run Solaris... :-}
> 
> Any clue ? 
> 
> Many thanks,
>  Daniel
> -- 
> IRCnet: Mr-Spock  - My name is Pentium of Borg, division is futile, you
> will be approximated. - 
> *Daniel Lang * [EMAIL PROTECTED] * +49 89 289 25735 * http://www.leo.org/~dl/*
> 
> -- 
> IRCnet: Mr-Spock  - May His Shadow fall upon thee - 
> *Daniel Lang * [EMAIL PROTECTED] * +49 89 289 25735 * http://www.leo.org/~dl/*
> 
> 
> 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

3.4 cdroms and xdm

[James C. Beyer]

A few questions:

Am I insane or is my FreeBSD 3.4 install disk from Walnut Creek bad?  I
can not get it to boot and boot disks created with it do not work.
Will boot disks created from the web source work?

I seem to vaguely remember something about this but can not remember where
or what was said.

Also,  I managed to get xdm to go active at boot time on my 3.3 machine
but it would not let anyone in.  I used the line in the ttys file and just
changed the off to on.  Is this problem covered somewhere?  I have not
found anything and would really like to avoid command line startx for my
users.


Any help will be greatly appreciated.

james



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: PR kern/14034: gettimeofday() returns negative value?

[James C. Beyer]

Actually, a line like  temp =t1->tv_usec + 100;
or using some other temp to save the t1 data before it is permuted will
make the program work perfectly.  t1->tv_usec += 100;  Kills the
correct value when t2 is updated.  The another option would be to stop
using pointers in the function call. ie:

#include 
#include 
#include 
#include 

#include 
#include 
#include 
#include 
#include 

int
time_elapsed(struct timeval t1, struct timeval t2)
{
int s, ms;
 
s = t1.tv_sec - t2.tv_sec;
assert((s >= 0));
if (s != 0) { 
if (t1.tv_usec < t2.tv_usec) {
s--;
t1.tv_usec += 100;
}
} 
ms = s * 100 + (t1.tv_usec - t2.tv_usec);
return (ms);
}

int 
main()
{
struct timeval tv1, tv2;
int run = 0;
int diff;

if (gettimeofday(&tv1, NULL) < 0) {
perror("gettimeofday");
}
tv2.tv_usec = tv1.tv_usec;
tv2.tv_sec = tv1.tv_sec;


for (;;) {
usleep(9000);
if (gettimeofday(&tv1, NULL) < 0) {
perror("gettimeofday");
}

/*
 * diff in usec.
 */
diff = time_elapsed(tv1, tv2);
if (diff < 0) {

printf("-ve tvdiff %d\n", diff);
printf("%ld %ld %ld %ld\n",
tv1.tv_sec, tv1.tv_usec,
tv2.tv_sec, tv2.tv_usec);
printf("run = %d\n", run);
assert(0);
}
tv2.tv_usec = tv1.tv_usec;
tv2.tv_sec = tv1.tv_sec;
run++;
}
return (0);
}
works fine been running for 5 minutes without a hitch.

Don't update a value if you don't really want it changed outside a
function.

jcb




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message