Slackware 14.2 doesn't install in athlon palomino

slack-atlhonxp-palomino · 07-03-2016, 10:36 PM

I have an Athlon xp 1600+ ( Palomino ) with 640MB Ram that has installed Slackware 14.1 and run it satisfactorily, except Firefox and Seamonkey that were a lot unstable ( less unstable Firefox 24.5 and Seamonkey 29 ) .
Before a 13.37 ran perfect.
Any of two 14.2 testing got installed but didn't run well because a lot of segmentation faults when libc was called by programs.
Now the 14.2 released version even no installing because the error appears at installation time ( mainly in dd , numberfmt ) until a kernel crash happens.
I know no SSSE-2 instructions in Palomino. Is it the only reason that a 586 compiled version doesn't work ?

Thanks for answers in advance

PS. I had a multiboot system with Debian. Debian stop working and after installing in testing v. 7.0. In that multiboot 13.37 continues running perfect although obsolete libraries and programs.

blancamolinos · 07-04-2016, 11:24 AM

do you see something as 'Illegal instruction' during the errors?

slack-atlhonxp-palomino · 07-04-2016, 02:39 PM

Thanks for your answer.

No illegal instruction messages.

Installing 14.2 all the errors are segmentation faults with registers detail ( x ip , y bp ... ), many of dd or numberfmt(?) , one of du, when they call a glibc routine. Around 50 I saw in dmesg. Finally a kernel error with a
long message.

In testing 14.2 lot of segmentation fault messages in several packages, too unstable so barely usable.

In 14.1 mozilla browsers are unstable with segmentation faults till they just break before showing 1st screen. So for stability I have to use old versions ( 24.5 Firefox ). When I have opportunity seeing error same graphic
library involved ( I don't remember now which one ). Midnight commander is quite unstable too.
Kernels installed are 3.14.67 ( disk 1 ) and 3.10.101 (
disk2 and disk3 multiboot ), using only 3.10.17huge when changing hardware, while compiling new one.

In 13.37 all was smooth when used years ago.

All versions using Fvwm.

If you need system-kernel logs -I have them- or mozilla message, I can look for it.

Best regards

slack-atlhonxp-palomino · 07-04-2016, 02:56 PM

PS. I got only illegal instruction messages in skype, google talk and such

blancamolinos · 07-04-2016, 04:35 PM

have you tested the memory with memtest (this utility is in the slackware-14.2 dvd or first cdrom; you have to write memtest after the boot: prompt) ?

slack-atlhonxp-palomino · 07-04-2016, 04:48 PM

Not since a long time ago. I'll do it and post result

slack-atlhonxp-palomino · 07-04-2016, 04:52 PM

Anyway in text mode or working without net 14.1 is stable for hours or days, except mc, and gcc

Best regards

slack-atlhonxp-palomino · 07-04-2016, 05:03 PM

PS I had a nvidia tnt2 , nv05, change from xorg level -nvidia driver- to kernel level -nouveau- was a headache and a bunch of problems and kernel crashes.

slack-atlhonxp-palomino · 07-04-2016, 11:27 PM

Memtest with Slack 14.2 crashes in boot : just blue screen
Same for Salck 14.2 testing, and systemreccd.

I compiled memtest myself with arch=athon-xp and same result.

I tried passmark version and boot mormal same result. Boot traced break at 2nd step : 1. Begin 2. MemS

Best regards

bassmadrigal · 07-05-2016, 02:54 AM

What about memtest from 13.37?

keefaz · 07-05-2016, 04:28 AM

How is hardware condition from inside? Is it clean from dust, no excessive temperature seen in bios?

55020 · 07-05-2016, 06:44 AM

Ok, I've tested a 14.2 install on an Athlon XP Barton (using the hugesmp kernel). It installs and runs without any problems.

The instruction set is (I think) the same as the Palomino -- it has sse, but not sse2. You can compare your /proc/cpuinfo with mine, to check that the flags are the same.

Code:

crapbox# cat /proc/cpuinfo 
processor	: 0
vendor_id	: AuthenticAMD
cpu family	: 6
model		: 10
model name	: AMD Athlon(tm) XP  3000+
stepping	: 0
cpu MHz		: 2162.802
cache size	: 512 KB
fdiv_bug	: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 1
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow
bogomips	: 4325.60
clflush size	: 32
cache_alignment	: 32
address sizes	: 34 bits physical, 32 bits virtual
power management: ts

It looks like your memory is the problem. The old Athlons always had difficult requirements for memory timing and power supplies, and if it didn't like your memory, it usually showed up as errors in the memtest block moves -- which is probably also the sort of thing that 'dd' would do.

slack-atlhonxp-palomino · 07-05-2016, 09:15 AM

To 55520 :
Differences in cpuinfo just
stepping : 0 palomino, 2 barton
flags : vmmcall palomino , apic barton

slack-atlhonxp-palomino · 07-05-2016, 09:27 AM

To Keefaz :
Temp1, cpu sensor, doesn't work long ago, so result is absurd:

"via686a-isa-6000
Adapter: ISA adapter
Vcore: +1.75 V (min = +0.06 V, max = +3.10 V)
in1: +0.27 V (min = +2.36 V, max = +2.61 V) ALARM
+3.3V: +3.33 V (min = +3.12 V, max = +3.45 V)
+5V: +4.93 V (min = +4.73 V, max = +5.20 V)
+12V: +12.30 V (min = +11.35 V, max = +12.48 V)
fan1: 3750 RPM (min = 0 RPM, div = 8)
fan2: 0 RPM (min = 0 RPM, div = 8)
temp1: +142.7°C (high = +146.2°C, hyst = -70.9°C)
temp2: +31.6°C (high = +146.2°C, hyst = -70.9°C)
temp3: +24.6°C (high = +146.2°C, hyst = +146.2°C)"

Case is open both sides, years ago, because of noise of fan2 and fan3, both off with case open. With the case open both sides dust is not absent.
In1 alarm is there from sensors reporting started 10 years ago

Best regards

slack-atlhonxp-palomino · 07-05-2016, 09:37 AM

More info in case be relevant : till years ago ( I think till version 5.3 Open BSD ) I used both Linuxes and OpenBSD in multiboot. I had to give up using OpenBSD as from a version a lot of processes from usual programs showed segmentation faults ( from wget to mc, even in text mode, out of wm ). I supposed at the moment that algorithm of randomizing memory blocks allocation in processes was the cause. Not problems at that moment with linuxes.