2.6 vs 2.4

2.6 vs 2.4

[Srinivas Murthy]

[-- Attachment #1: Type: text/plain, Size: 275 bytes --]

Hi,
 Wolfgang Denk has a good listing of issues to consider betn linux 2.4 vs
2.6 for ppc82xx based platforms (http://www.denx.de/wiki/Know/Linux24vs26).

 Would this recommendation still hold?

 Are there any patches, developments in this area?

Thanks,
_Srinivas

[-- Attachment #2: Type: text/html, Size: 493 bytes --]

Re: 2.6 vs 2.4

[Benjamin Herrenschmidt]

On Thu, 2006-05-11 at 23:51 -0700, Srinivas Murthy wrote:
> Hi,
>  Wolfgang Denk has a good listing of issues to consider betn linux 2.4
> vs 2.6 for ppc82xx based platforms
> (http://www.denx.de/wiki/Know/Linux24vs26).
>  
>  Would this recommendation still hold?
>  
>  Are there any patches, developments in this area?

The main open question is: is somebody still maintaining the 8xx kernel
port ?

Ben.

Re: 2.6 vs 2.4

[Benjamin Herrenschmidt]

On Fri, 2006-05-12 at 23:57 +1000, Benjamin Herrenschmidt wrote:
> On Thu, 2006-05-11 at 23:51 -0700, Srinivas Murthy wrote:
> > Hi,
> >  Wolfgang Denk has a good listing of issues to consider betn linux 2.4
> > vs 2.6 for ppc82xx based platforms
> > (http://www.denx.de/wiki/Know/Linux24vs26).
> >  
> >  Would this recommendation still hold?
> >  
> >  Are there any patches, developments in this area?
> 
> The main open question is: is somebody still maintaining the 8xx kernel
> port ?

Oh , and to give a better reply after reading that web page... I'd say
there might be a reason why 2.6 is slower (or appears slower) than 2.4
on some of those embedded platforms. It's definitely faster in a number
of areas on 6xx and other desktop CPUs. A lot of embedded folks will
then bitch and complain about the slowdown on their 8xx processor, but
very few of them if not none popped up actually identifying the actual
issues and coming up with solution for them during the 2.6 dev. cycle.

Ben.

Re: 2.6 vs 2.4

[Felix Domke]

Benjamin Herrenschmidt wrote:
>> Wolfgang Denk has a good listing of issues to consider betn linux 2.4
>>vs 2.6 for ppc82xx based platforms
>>(http://www.denx.de/wiki/Know/Linux24vs26).
>> Would this recommendation still hold?
>> Are there any patches, developments in this area?
> The main open question is: is somebody still maintaining the 8xx kernel
> port ?
The 4xx port has the same problem.

When we switched from 2.4 to 2.6, IDE performance ("hdparm -t" to have a
single number) was reduced by about 25% (even after trying to finetune
the IDE driver - the time between the end of one transfer and the begin
of the next transfer was just too long to saturate the harddisk).

Because of my lack of knowledge of the block device layer's internals I
wasn't able to track that down. My initial plan was to build a trace
using the RiscTrace environment, to see any instruction executed between
requests, but my company moved away from PPC4xx hardware
(unfortunately), so I was never able to complete this. I still believe
that other platforms have the same problems.

On a 300MHz embedded mips machine, saturating a 100MBit network link via
ftp is not easy. Our old 252MHZ PPC machines never performed better than
around 5MB/s (but had a non-DMA NIC, which however could be satured in
theory with about 70% cpu load, based on the bus bandwidth. Why are the
remaining 30% not enough to do IDE DMA and the TCP overhead? Memory
performance? But why did we had better numbers with 2.4 then?).

On that mips machine, both IDE and network support DMA (agreed, it's a
RTL8139, so it requires another memcpy), and memcpy() performance is
>100MB/s. Where is the bottleneck? And, much more important: how do
measure it?

I'm sorry that I can't really much do anything else than complaining,
but all my attempts to track down the problems were futile. I'm not sure
if testing "ftp performance" (as a completely non-synthetic benchmark -
FTP speed is/was a real issue on our platform) is "the right test", but
it's one of the things where i personally want linux to become better.

Or is this just a misconfiguration? Are there mysterious IO scheduler
default parameters which are just suboptimal for our case?

regards,
Felix

Re: 2.6 vs 2.4

[Thiago Galesi]

> Hi,
>  Wolfgang Denk has a good listing of issues to consider betn linux 2.4 vs
> 2.6 for ppc82xx based platforms
> (http://www.denx.de/wiki/Know/Linux24vs26).
>
>  Would this recommendation still hold?

Not much. This article is almost one year old. Even though some things
do apply (like bigger footprint, etc), the kernel improved a lot on
the months that followed that article.

I've participated in a project that used a (somewhat patched) 2.6.10
kernel, with soccessful results.

Not the mention the advantages of the 2.6 kernel in general.

>The main open question is: is somebody still maintaining the 8xx kernel
>port ?

LINUX FOR POWERPC EMBEDDED PPC8XX
P:      Marcelo Tosatti
S:      Maintained

(this has been determined AFAIK in OLS 2005)

Re: 2.6 vs 2.4

[Felix Domke]

Thiago Galesi wrote:
> Not much. This article is almost one year old. Even though some things
> do apply (like bigger footprint, etc), the kernel improved a lot on
> the months that followed that article.
> I've participated in a project that used a (somewhat patched) 2.6.10
> kernel, with soccessful results.
Can you point me to a list of changes which where done? I really like to
check if they apply to my performance problems with 2.6 kernels, i.e. if
it's worth porting the board specific changes to a newer kernel.

Felix

2.6 vs 2.4

[Thiago Galesi]

> Can you point me to a list of changes which where done? I really like to
> check if they apply to my performance problems with 2.6 kernels, i.e. if
> it's worth porting the board specific changes to a newer kernel.
>

I'm sorry, I was thinking (specifically) of the 8XX port that went
from broken to maintained.

It would help if you could mention some ofthe problems you are having.

>When we switched from 2.4 to 2.6, IDE performance ("hdparm -t" to have a
>single number) was reduced by about 25% (even after trying to finetune
>the IDE driver

>On a 300MHz embedded mips machine, saturating a 100MBit network link via
>ftp is not easy.

Unfortunatelly my knowlege of IDE is almost 0. Maybe it has something
to do with schedulers being used?

About network, I had better results with a much slower processor than
your reported 5Mb/s (60MHz PPC in 2.6.10)


Thiago