2.4 broken on 486SX

2.4 broken on 486SX

[Ross Vandegrift]

Hello all,

	I've tried many versions of 2.4 kernels on a 486SX that I have, and none of them will boot.  When trying to boot, I see 'Loading...............' with the dots continuing to appear.  Finally, the expected carriage return, and bam the system is dead.  The box doesn't even make it to 'Uncompressing Linux...'.  What I'm wondering, is how on earth am I supposed to figure out what is going on and fix it?  I've tried booting both with loadlin and directly off a floppy, both produce the same result.  The machine has 16M of RAM.  I do not know offhand the make of the CPU.

	Ross Vandegrift
	ross@willow.seitz.com

Re: 2.4 broken on 486SX

[Richard B. Johnson]

On Fri, 24 Aug 2001, Ross Vandegrift wrote:

> Hello all,
> 
> 	I've tried many versions of 2.4 kernels on a 486SX that I have,
> and none of them will boot.
[SNIPPED stuff with no new-lines]

You need to compile as a '386. Modify ../linux/.config to enable
CONFIG_M386=y under "Processor type and features". Delete all other
lines until "General Setup". Then execute:

	make oldconfig

Linux 2.4.1 boots fine on a '386. If that doesn't work, you may
have to back-rev your 'C' compiler.

I'm using egcs-2.91.66 which works okay. Some newer versions may
generate code that can't run on a `386.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.1 on an i686 machine (799.53 BogoMips).

    I was going to compile a list of innovations that could be
    attributed to Microsoft. Once I realized that Ctrl-Alt-Del
    was handled in the BIOS, I found that there aren't any.

Re: 2.4 broken on 486SX

[Ross Vandegrift]

Ack, that was a horrific problem report.  Here are a few relevant data points I left out:

Kernels have always been compiled only with 386 optimizations.
	(tried 486 a few times, no difference)
Always built including soft float support (not that it matters so early)
Always built with gcc 2.95.2
Same kernel binaries work perfectly on a number of other boxen:
	A 386DX, a number of Pentium 100's, another 486SX, and a K6-2


Ross Vandegrift
ross@willow.seitz.com

Re: 2.4 broken on 486SX

[Brian Gerst]

Ross Vandegrift wrote:
> 
> Ack, that was a horrific problem report.  Here are a few relevant data points I left out:
> 
> Kernels have always been compiled only with 386 optimizations.
>         (tried 486 a few times, no difference)
> Always built including soft float support (not that it matters so early)
> Always built with gcc 2.95.2
> Same kernel binaries work perfectly on a number of other boxen:
>         A 386DX, a number of Pentium 100's, another 486SX, and a K6-2

What was the last known working kernel for this machine?  Have you tried
any -ac kernels?

--

				Brian Gerst

Re: 2.4 broken on 486SX

[Steve Kieu]

By the way, I got this strange thing too but with
2.2.19. I got the kernel 2.2.19 source from slackware
site, compile it using gcc 2.95.3 and select CPU as
486, but the compiled image can not run in my 486 CPU
(it does run fine in the machine I compile it i686).
There is no way. Ironically 2.4.x works for me, if I
choose 486, I can run it on 486. 

I noticed the differrence in the cflags in 2.2.19 it
is like  -m486 -DCPU=i486  and in 2.4.x it is
-march=i486 instead of -DCPU=i486. According to gcc
documentation it is the same. Why it is like that ? 



=====
S.KIEU

_____________________________________________________________________________
http://shopping.yahoo.com.au - Father's Day Shopping
- Find the perfect gift for your Dad for Father's Day

Re: 2.4 broken on 486SX

[Cliff Albert]

On Sat, Aug 25, 2001 at 12:46:27PM +1000, Steve Kieu wrote:

> By the way, I got this strange thing too but with
> 2.2.19. I got the kernel 2.2.19 source from slackware
> site, compile it using gcc 2.95.3 and select CPU as
> 486, but the compiled image can not run in my 486 CPU
> (it does run fine in the machine I compile it i686).
> There is no way. Ironically 2.4.x works for me, if I
> choose 486, I can run it on 486. 
> 
> I noticed the differrence in the cflags in 2.2.19 it
> is like  -m486 -DCPU=i486  and in 2.4.x it is
> -march=i486 instead of -DCPU=i486. According to gcc
> documentation it is the same. Why it is like that ? 

Linux 2.4.8-ac3 runs just fine on my Compaq Prosignia VS (which has a 486DX66)

-- 
Cliff Albert		| RIPE:	     CA3348-RIPE | www.oisec.net
cliff@oisec.net		| 6BONE:     CA2-6BONE	 | icq 18461740