[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: kernel 2.1.129 on Alpha Miata



>I have a similar system, but I've had no drive problems.  The only 
problem
>I have is getting the built-in tulip chip to work on 2.1.129. Maybe 
Loic's
>patches don't work on exp. kernels?

Hmmm... actually, I had some problems like that too... but not exactly. 
I got the tulip driver to work with the onboard Ethernet using version 
0.89H (I think). My problem was that when I added a second ethernet 
card(or third).. the onboard died. ifconfig reported all ethernets up 
and go, but the link light was gone, and I couldn't get pass the 
interface out to the network. At first I thought it was some problem 
with the tulip driver and having two DC21x4x based ethernet 
controllers.... I thought somehow it was getting confused... but that 
wasn't the problem, since I tried mixing the onboard with some 3com 
cards. I also tried a variety of combinations of compiling the drivers 
as modules and into the kernel. None of this helped. I also read some 
french dude's posting who was also having the same problems. (recall, 
the problem ONLY exists for me with an add-on ethernet in combo with the 
onboard) He basically gave up after several months... well, since May 
98. From what he could tell, there was something odd hardware wise with 
the onboard ethernet of the PWS a/au series. It's based on the DC21143 
chip... other than that, I'm clueless.

>As for the keyboard locking out on you, I had to 'correct' the sound
>configuration on 2.0.35.  Don't ask, I have no clue why....
>
>What compiler?  egcs-1.1 and up seem to have a bit of trouble 
generating
>totally stable code.  They won't compile a certain module in 2.0.36, 
and
>the X SVGA server trashes the console, the only fix is a soft boot.

I just installed RH5.2... and haven't played with it that much... just 
trying to get some essentials going... so it's egcs 1.0.3a-14. I'm using 
the 2.0.35 source that came with RH5.2. By the way, has anyone else 
noticed that some of the header files in the kernel source are different 
from the ones that get installed on the system? I keep getting 
'redefinition' warnings during compile because of this.... the kernel 
source has 'unsigned long' in place of 'u_int32_t' that are in the 
identical files in the system include dir. However, I noticed that in 
another header file, typedef is used to alias u_int32_t with 'unsigned 
long'. So it seems like it should okay... but egcs is complaining... I'm 
no professional programmer... just learning as I go.... but that seems a 
bit odd... anyone care to explain please?


>Some comments on what others wrote (some clipping):
>> I've had this problem since day one when I installed RH5.2... at 
first I 
>> thought it might be a problem with my SCSI... but now that I'm not 
the 
>> only one.... hmmm... anyways.. yes, it takes about 2 to 3 minutes 
just 
>> to mount my root partition.... on a similar Intel system, it takes 
less 
>> than 10 seconds.....
>
>It would help to know if you've tried RH5.1, and make sure you are sure
>the SCSI system is connected and terminated correctly.  Do you have a
>QLogic or Adaptec SCSI?  The Qlogic ones gave me trouble.  I heard they
>are better, but at that time, Linux's support for it & PWS wasn't good.

Hmmm... well it's an adaptec 3940UW.... with a aic7880 controller... two 
of them. Termination could be a problem... as the cable is a really old 
one... and it looks in bad shape. I will try to find a newer cable and 
see what happens... 

>
>> >	3) does something bad so that after I ctrl-alt-del and go back to 
milo
>> >and run a good kernel, the keyboard is disabled and the cd-rom 
flakes.
>> 
>> Yup, I've also had this problem since day one.... if I reboot from 
>> MILO... my system hangs during boot... if I cold boot it (reset 
>> switch)... I'm fine.... My system is total SCSI... 
>
>My situation isn't as severe that I remember, but I get some similar
>problems on a soft (warm) reset of an experimental kernel.
>
>> >	These behaviors seem odd since the kernel is supposed to be 
reasonably
>> >stable...
>
>Ditto.
>


______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index] []