2.4 vs 2.6

John dingo at coco2.arach.net.au
Mon Sep 13 16:52:44 CDT 2004


Cef wrote:

>On Tue, 14 Sep 2004 06:36, John wrote:
>
>  
>
>>I don't think testing with non-standard boot params is appropriate
>>because this is not how most people are going to run. Stuart, if you're
>>going to set everyone up with 'elevator=deadline,' then test that way.
>>    
>>
>
>Erm, I'm not going to remotely think of suggesting that we set 
>elevator=deadline, until the one problem case I know of (yours) has been 
>tested manually with it at least. This is why I asked.
>  
>

Not one case. Every system I've done the test on, covering one SiS and 
two Intel chipsets.

I have running five systems with CD drives, three of them exhibit this 
behaviour,one has other dramas and I'm not going to acept results from 
it, and I think the fifth does but I don't recall for sure but I will 
retest it when convenient.  If it does, it introduces a third chipset 
manufacturer.

I'm using Knoppix because it boots on everything, 3.6 because it's the 
latest I have _and_ it has both 2.5 and 2.6 kernels.

>If it doesn't fix the problem, then it's not going to be suggested. If it does 
>fix the problem, then it's a kernel issue and this needs to be addressed, 
>wether by making that a default kernel boot param for the moment (not 
>advisable), or fixing the problem in the kernel. At the moment, I'm not even 
>sure wether it's a kernel issue, or a drive setup issue that can be taken 
>care of in userspace (eg: via hdparm).
>
>  
>
>>A basic test is fine by me, it shows a probable problem that needs
>>checking.
>>    
>>
>
>Any chance you can try with the elevator=deadline boot param and report back 
>if it solves your issues? An answer on this line gives us a clue as to where 
>the problem is.
>  
>

I suggest you test it on your boxes too; it looks from here like a 
general problem, and that is why I suggested others try it.

I _might_ have an Evectra (I810) to test later today.

>Also, the output of 'hdparm -v /dev/hda' on the system would be good too, 
>using Ubuntu's 2.6 kernel, and any other kernels you can boot in the box (eg: 
>2.4 & 2.6 from another distribution, etc).
>
>  
>
Typically they have DMA on, unmasking off. Recently, other settings seem 
to me to have little impact.







More information about the sounder mailing list