Questions about the CMD640 and RZ1000 bugfix support options

Questions about the CMD640 and RZ1000 bugfix support options

[Christoph Anton Mitterer]

[-- Attachment #1.1: Type: text/plain, Size: 914 bytes --]

Hi.

First of all: A happy new year in advance!

Now to my question:
In the kernel-configuration there are the two options:
CONFIG_BLK_DEV_CMD640        CMD640 chipset bugfix/support
and
CONFIG_BLK_DEV_RZ1000        RZ1000 chipset bugfix/support

At least the second of those two seems to cause some little slowdown 
("This may slow disk throughput by a few percent, but at least things 
will operate 100% reliably.").
1) Is the bigfixing code only activated when a buggy chipset is detected 
or do I even have a slowdown with proper a properchipset?
2) Can some help me to find out if my system has one of those two 
chipsets. This might be a dumb question but I'm not sure if 
CMD640/RZ1000 are only some little IDE controller chipsets or if they 
were some kind of Southbridge/Nortbridge.
If got a Epox 8KTA2 mainboard (having VIA VT82C686A/B chipset).

Thanks in advance!

Greetings,
Christoph Anton Mitterer.

[-- Attachment #1.2: cam.vcf --]
[-- Type: text/x-vcard, Size: 450 bytes --]

begin:vcard
fn:Mitterer, Christoph Anton
n:Mitterer;Christoph Anton
org:Munich University of Applied Sciences;Department of Mathematics and Computer Science
adr;quoted-printable;quoted-printable:;;Lothstra=C3=9Fe 34;M=C3=BCnchen;Freistaat Bayern;80335;Federal Republic of Germany
email;internet:cam@mathematica.scientia.net
tel;home:+49 89 24409390
tel;cell:+49 172 8617341
x-mozilla-html:TRUE
url:http://fhm.edu/
version:2.1
end:vcard


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Michelle Konzack]

[-- Attachment #1: Type: text/plain, Size: 1014 bytes --]

Am 2004-12-31 23:26:14, schrieb Christoph Anton Mitterer:
> Hi.

> CONFIG_BLK_DEV_CMD640        CMD640 chipset bugfix/support
> CONFIG_BLK_DEV_RZ1000        RZ1000 chipset bugfix/support

> 2) Can some help me to find out if my system has one of those two 
> chipsets. This might be a dumb question but I'm not sure if 
> CMD640/RZ1000 are only some little IDE controller chipsets or if they 
> were some kind of Southbridge/Nortbridge.
> If got a Epox 8KTA2 mainboard (having VIA VT82C686A/B chipset).

The CMD640 and RZ1000 are realy old...
I think it wat the time around P1 or something like this.

Forget it. - You do not need it.

> Thanks in advance!
> 
> Greetings,
> Christoph Anton Mitterer.

Greetings
Michelle

-- 
Linux-User #280138 with the Linux Counter, http://counter.li.org/ 
Michelle Konzack   Apt. 917                  ICQ #328449886
                   50, rue de Soultz         MSM LinuxMichi
0033/3/88452356    67100 Strasbourg/France   IRC #Debian (irc.icq.com)

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Alan Cox]

On Gwe, 2004-12-31 at 22:26, Christoph Anton Mitterer wrote:
> Hi.
> 
> First of all: A happy new year in advance!
> 
> Now to my question:
> In the kernel-configuration there are the two options:
> CONFIG_BLK_DEV_CMD640        CMD640 chipset bugfix/support
> and
> CONFIG_BLK_DEV_RZ1000        RZ1000 chipset bugfix/support
> 
> At least the second of those two seems to cause some little slowdown 
> ("This may slow disk throughput by a few percent, but at least things 

They only trigger for the affected chipsets

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Andrey Melnikoff]

Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> On Gwe, 2004-12-31 at 22:26, Christoph Anton Mitterer wrote:
> > Hi.
> > 
> > First of all: A happy new year in advance!
> > 
> > Now to my question:
> > In the kernel-configuration there are the two options:
> > CONFIG_BLK_DEV_CMD640        CMD640 chipset bugfix/support
> > and
> > CONFIG_BLK_DEV_RZ1000        RZ1000 chipset bugfix/support
> > 
> > At least the second of those two seems to cause some little slowdown 
> > ("This may slow disk throughput by a few percent, but at least things 
> They only trigger for the affected chipsets
But enabled by default. Maybe disable it by default ? Or make depend with 
CONFIG_M586 || CONFIG_M586TSC || CONFIG_M586MMX ?

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Alan Cox]

On Mer, 2005-01-05 at 18:58, Andrey Melnikoff wrote:
> > > At least the second of those two seems to cause some little slowdown 
> > > ("This may slow disk throughput by a few percent, but at least things 
> > They only trigger for the affected chipsets
> But enabled by default. Maybe disable it by default ? Or make depend with 
> CONFIG_M586 || CONFIG_M586TSC || CONFIG_M586MMX ?

They should be enabled by default. That makes it safer for default
compiles, and their code size is close to if not nil because it can all
be __devinit or __init

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Andrey Melnikoff]

Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> On Mer, 2005-01-05 at 18:58, Andrey Melnikoff wrote:
> > > > At least the second of those two seems to cause some little slowdown 
> > > > ("This may slow disk throughput by a few percent, but at least things 
> > > They only trigger for the affected chipsets
> > But enabled by default. Maybe disable it by default ? Or make depend with 
> > CONFIG_M586 || CONFIG_M586TSC || CONFIG_M586MMX ?

> They should be enabled by default. 
Why? This is really _OLD_ and _BUGGY_ chips. As I see in google - it used 
in Asustek Pentium MB PCI/I-P54SP4 and some Intel mb for Pentium with
Neptune chipsets. All of this MB - for classic Pentium 75/90/100MHz. 

> That makes it safer for default compiles, and their code size is 
> close to if not nil because it can all be __devinit or __init
At this time, no modern MB use this buggy chipsets.

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Christoph Anton Mitterer]

Andrey Melnikoff wrote:
 > Alan Cox wrote:
 > >They should be enabled by default.
 >
 > Why? This is really _OLD_ and _BUGGY_ chips. As I see in google - it used
 > in Asustek Pentium MB PCI/I-P54SP4 and some Intel mb for Pentium with
 > Neptune chipsets. All of this MB - for classic Pentium 75/90/100MHz.

What about the following idea?
Both stay enabled by default but the help text explains exactly (as
far as possible) which systems are affected.
This would help newbies like me to decide if those bugfixes are
necessary or not.

Greetings,
cam.

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Alan Cox]

On Iau, 2005-01-06 at 22:40, Andrey Melnikoff wrote:
> > They should be enabled by default. 
> Why? This is really _OLD_ and _BUGGY_ chips. As I see in google - it used 
> in Asustek Pentium MB PCI/I-P54SP4 and some Intel mb for Pentium with
> Neptune chipsets. All of this MB - for classic Pentium 75/90/100MHz. 

And the cost to you is .. zero.

> > That makes it safer for default compiles, and their code size is 
> > close to if not nil because it can all be __devinit or __init
> At this time, no modern MB use this buggy chipsets.

And lots of people run old motherboards. The base kernel configuration
by default is a mix between "things everyone wants, things defaulting on
for safety reasons and linus computer"

Alan

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Alan Cox]

On Iau, 2005-01-06 at 23:10, Christoph Anton Mitterer wrote:
> What about the following idea?
> Both stay enabled by default but the help text explains exactly (as
> far as possible) which systems are affected.
> This would help newbies like me to decide if those bugfixes are
> necessary or not.

Its the ideal solution. The diff for this is available on your computer
already - its kept in /dev/null 8)

          The PC-Technologies RZ1000 IDE chip is used on many common 486
and
          Pentium motherboards, usually along with the "Neptune"
chipset.
          Unfortunately, it has a rather nasty design flaw that can
cause
          severe data corruption under many conditions. Say Y here to
include
          code which automatically detects and corrects the problem
under
          Linux. This may slow disk throughput by a few percent, but at
least
          things will operate 100% reliably.

          The CMD-Technologies CMD640 IDE chip is used on many common
486 and
          Pentium motherboards, usually in combination with a "Neptune"
or
          "SiS" chipset. Unfortunately, it has a number of rather nasty
          design flaws that can cause severe data corruption under many
common
          conditions. Say Y here to include code which tries to
automatically
          detect and correct the problems under Linux. This option also
          enables access to the secondary IDE ports in some CMD640 based
          systems.
                                                                                
  

Re: Questions about the CMD640 and RZ1000 bugfix support options

[Christoph Anton Mitterer]

Alan Cox wrote:

 >On Iau, 2005-01-06 at 23:10, Christoph Anton Mitterer wrote:
 >
 >>What about the following idea?
 >>Both stay enabled by default but the help text explains exactly (as
 >>far as possible) which systems are affected.
 >>This would help newbies like me to decide if those bugfixes are
 >>necessary or not.
 >
 >
 >Its the ideal solution. The diff for this is available on your computer
 >already - its kept in /dev/null 8)

Ok,... so here's my patch which I request for inclusion:
--- begin of patch ---

--- end of patch ---
(Ahh,.. my first kernel patch,... *lol* )

;-)


Seriously,.. I thought about adding something like:
"Most users of modern computers won't need this, but it is safer to say 
Y here."

But according to you answer I think that you like the way it's at the 
moment. =)
Never mind! But if you think it's ok I could make a (real) patch which 
adds such a text.

Best wishes,
cam.