Via chipset KT133 and system instability

[Tim Wunder]

On Monday 23 December 2002 05:08 pm, someone claiming to be Net Llama! wrote:
> On Mon, 23 Dec 2002, Tim Wunder wrote:
> > On 12/23/2002 4:24 PM, someone claiming to be Net Llama! wrote:
> > > On Mon, 23 Dec 2002, Tim Wunder wrote:
> > >> Anyone running an AMD-based board with a Via KT133 chipset? I've
> > >> been experiencing a significant amount of system instabilty (X
> > >> crashing, SRPM rebuilds failing with internal compiler errors,
> > >> running kmix locking the system) and was leaning toward bad RAM
> > >> being the culprit. But as I've been googling, I've found several
> > >> references to the KT133 chipset, and the VT82C686 Southbrisge
> > >> wreaking havoc.
> > >>
> > >> I'll be updating my motherboard's BIOS tonight (an ABIT KT7A) and
> > >> replacing the RAM to see if that helps, but any insight y'all could
> > >> provide would be helpfull, especially if your running on a mobo
> > >> with a VIA chipset. I'm currently running RH8's stock Athlon
> > >> kernel, but would be willing to compile a new kernel if that would
> > >> help fix things.
> > >
> > > Another datapoint to consider is that RH8 isn't exactly famous for
> > > its stability.  Can you reliably reproduce the instability, or is it
> > > seemingly random?
> >
> > Seemingly random, although going to
> > http://www.backbayguide.com/newwine.htm in Mozilla seems to crash X
> > fairly regularly.
>
> It crashed X??  That is creepy.  Anyway, if you've got a Knoppix CD, you
> could always see if the behavior is reproducable there.  THat should help
> to rule out/in hardware as a factor.

Now *that* is a great idea!
BTW, I ran a memory test, updated the BIOS and now, when I start X-Chat, X 
crashes. Here's the tail end of an strace:
open("/usr/X11R6/lib/X11/locale/locale.alias", O_RDONLY) = 5
fstat64(5, {st_mode=S_IFREG|0444, st_size=45725, ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 
0x40125000
read(5, "#\t$Xorg: locale.alias,v 1.3 2000"..., 4096) = 4096
read(5, "CA.ISO_8859-1\t\t\t\ten_CA.ISO8859-1"..., 4096) = 4096
read(5, "8591\t\t\t\t\tfr_FR.ISO8859-1\nfr_FR.I"..., 4096) = 4096
read(5, "_GB.ISO8859-15\nlo\t\t\t\t\t\tlo_LA.MUL"..., 4096) = 4096
read(5, "\tsh_YU.ISO8859-2\nsh_HR.iso88592\t"..., 4096) = 4096
read(5, "CN.eucCN\nchinese-t\t\t\t\t\tzh_TW.euc"..., 4096) = 4096
read(5, "9-2\ncs_CZ.ISO_8859-2:\t\t\t\tcs_CZ.I"..., 4096) = 4096
read(5, "-1\nes_PE.iso88591:\t\t\t\t\tes_PE.ISO"..., 4096) = 4096
read(5, "HU.ISO-8859-2:\t\t\t\thu_HU.ISO8859-"..., 4096) = 4096
read(5, "n_NO.ISO8859-1\nny:\t\t\t\t\t\tny_NO.IS"..., 4096) = 4096
read(5, "scii:\t\t\t\t\tta_IN.TSCII-0\nta_IN.ts"..., 4096) = 4096
read(5, "n:\t\t\t\t\tsh_YU.ISO8859-2\nslovak:\t\t"..., 4096) = 669
read(5, "", 4096)                       = 0
close(5)                                = 0
munmap(0x40125000, 4096)                = 0
open("/usr/X11R6/lib/X11/locale/locale.dir", O_RDONLY) = 5
fstat64(5, {st_mode=S_IFREG|0444, st_size=27925, ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 
0x40125000
read(5, "#\t$Xorg: locale.dir,v 1.3 2000/0"..., 4096) = 4096
read(5, "LOCALE\t\t\tfr_BE.ISO8859-15\niso885"..., 4096) = 4096
read(5, ".TCVN\nvi_VN.viscii/XLC_LOCALE\t\tv"..., 4096) = 4096
read(5, "\t\t\tlv_LV.UTF-8\nen_US.UTF-8/XLC_L"..., 4096) = 4096
close(5)                                = 0
munmap(0x40125000, 4096)                = 0
writev(3, [{"1\0\25\0\1\0I\0", 8}, {"-zz_abiword-century schoolbook-r"..., 
73}, {"\0\0\0", 3}], 3) = 84
read(3, "\1\231a\0\23\0\0\0\1\0\377\277]\t\16\10\4\0\0\0000\0\0"..., 32) = 32
readv(3, [{"K-zz_abiword-century schoolbook-"..., 76}, {"", 0}], 2) = 76
writev(3, [{"1\0\25\0\1\0I\0", 8}, {"-zz_abiword-century schoolbook-r"..., 
73}, {"\0\0\0", 3}], 3) = 84
read(3, "\1\231b\0\23\0\0\0\1\0\377\277]\t\16\10\4\0\0\0000\0\0"..., 32) = 32
readv(3, [{"K-zz_abiword-century schoolbook-"..., 76}, {"", 0}], 2) = 76
write(3, "-\0\26\0\27\0\0\2K\0\0\0-zz_abiword-century "..., 100) = 100
read(3, 0xbffff640, 32)                 = -1 EAGAIN (Resource temporarily 
unavailable)
select(4, [3], NULL, NULL, NULL)        = 1 (in [3])

See anything interesting?
Looks like it's trying to read from a block of memory, read(3, 0xbffff640, 
32), but it's unavailable, whatever that means... I'm guessing at this point. 
I'll see if I can find that Knoppix CD I used to have...

Tim


-- 
RedHat Psyche 8.0, stock kernel, Gnome 2.x, Xfree86 4.2.0
  4:00pm  up  8:17,  3 users,  load average: 0.24, 0.07, 0.02
It's what you learn AFTER you know it all that counts