Re: IDE DMA hard lock at boot time (KT266A chpiset)

Re: IDE DMA hard lock at boot time (KT266A chpiset)

[Wolly]

Hi Mark,

Thank you very much for your quick responses to my problem. 

Due to your help I was now able to solve both problems. 

Now everything works great (including UDMA and the ethernet NIC 
and ACPI)
[however, the rtl8139 still refuses to work, but I can use the eepro100]. 

The three core things were: 
- switch off apic stuff
- you may enable ACPI but you must _disable_ ide power down 
  in the BIOS (seems to be a bug somewhere)
- This message is normal and can be ignored: 
  PCI: No IRQ known for interrupt pin A of device 00:11.1. Please try using 
    pci=biosirq.

Regards,
Wolly

[Sorry; everyone who did not CC me did not reach me because I seem 
to be `mysteriously' unsubscribed from lkml since Feb 01 when my 
mail box ran out of space...]

> maybe I wasn't clear on this: your problem seems to be irq-routing
> or irq lossage.  others have reported the same thing, and have fixed
> it by turning off the (spurious UP apic usage), turning off apics
> in bios, etc.

IDE DMA hard lock at boot time (KT266A chpiset)

[Wolly]

Hi all, 

I'm using an elitegroup mobo (K7VTA3 / KT266A chipset; Athlon XP) 
and I'm experiencing a hard lock at boot time when configuring the 
kernel using `via82cxx support'. (No ctrl-alt-del, no sysrq, only hw reset)

[Without that I can boot without problems but I don't have DMA and thus 
disk throughput is at less than 3 Mb/sec.]

The hard lock seems to happen during the first hd access and the hd light 
stays on when the system is crashing. 

Is that 
- a known issue (hopefully not), 
- a Linux kernel problem or
- do I have to blame the board (this is not completely unlikely as 
  I also have problems with two network cards which work without 
  problems in other boards and same kernel)

Is the KT266A chipset's DMA supported at all? (The vt82cxx driver 
source code lists several chipsets including KT266 but not KT266A.)

--------<Boot screen>--------
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
VP_IDE: IDE controller on PCI bus 00 dev 89
PCI: No IRQ known for interrupt pin A of device 0:11.1. Please try using 
pci=biosirq
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
VP_IDE: VIA vt8233 (rev 00) IDE UDMA100 controller on pci00:11.1
    ide0: BM-DMA at 0xd400-0xd407, BIOS settings: hda:DMA, hdb:pio
    ide1: BM-DMA at 0xd408-0xd40f, BIOS settings: hdc:DMA, hdb:pio
hda: IBM-DTLA-xyz, ATA DISK drive
hdc: CD-xyz, ATAPI CD/DVD-ROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
hda: xx sectors (30000 MB) [...], UDMA(100)
hdc: ATAPI 52X CD-ROM drive, 128kB Cache, DMA
Uniform CD-ROM driver Revision: 3.12
Partition check:
 hda:_
---------------------
(hardlock occurs; cursor at `_')

Regards, 
Wolly

Re: IDE DMA hard lock at boot time (KT266A chpiset)

[Jose Luis Domingo Lopez]

On Saturday, 23 February 2002, at 19:12:18 +0100,
Wolly wrote:

> Hi all, 
> 
> I'm using an elitegroup mobo (K7VTA3 / KT266A chipset; Athlon XP) 
> and I'm experiencing a hard lock at boot time when configuring the 
> kernel using `via82cxx support'. (No ctrl-alt-del, no sysrq, only hw reset)
> [...]
>
I have a similar hardware setup here (Soltek SL-75DRV2 with KT266A
chipset + Athlon XP 1700+), and although the box boots and runs fine
(even with Athlon optimizations and VIA82CXXX compiled in), from time to
time the machine locks solid (no kind of response from the box, HDD
light keeps on, power cicle needed).

K7 optimizations and VIA chipsets seem to be quite problematic pieces of
hardware, but in my case you have to add to this a GeForce2 MX video
card, with NO closed-binary nVidia drivers.

Machine freezes seems to follow the following pattern: working on X
environment, watching TV (PCI TV capture card), listening to digital
audio music (aka MP3 :) and some sort of disk access. A high number of
freezes seem to happen when opening a medium sized mailbox in Maildir
format (aproximately 800 messages/files inside). As the box has 256 MB
of RAM, and the mailbox has been read before, the whole directory is
maybe cached, except for newly arrived messages.

"mutt" starts reading the mailbox, and after some hundreds of messages,
the box locks solid. Digital audio stops, overlaid TV capture stops,
SysRq does nothing, network unresponsive. Looking at the logs shows
nothing, even after having tried SysRq+everything after the lock.

The cause of all this ?. Maybe Athlon optimizations, maybe a severe flaw
in the chipset and/or motherboard BIOS, maybe ALSA sound (VIA8233 sound
doesn't work with OSS modules) not fine, maybe crappy nVidia-based video
card....who knows. No overclocking of any kind by the way, and
temperature seems to be under control.

I'll try setting AGP speed to x2 instead of x4, adding Option "NvAgp" "0"
to my X 4.x configuration, disabling DRI and any kind of 3D
acceleration, etc... and whatever I devise to get again a box I can
trust. Damn, I should not have asked for a VIA-based board...

--
José Luis Domingo López
Linux Registered User #189436     Debian Linux Woody (P166 64 MB RAM)
 
jdomingo AT internautas DOT   org  => Spam at your own risk