Re: ncr53c875: CACHE TEST FAILED or System stops with "cat" or "ls"

Gregory P. Smith (greg@nas.nasa.gov)
Tue, 27 Apr 1999 10:43:05 -0700


I've been having these problems as well. They started in the 2.2.x
(and the late 2.1.1xx but I can't verify that anymore) and are much
worse in the 2.2.6 SYM driver. I have an Asus 53c875 based card. In
addition to the times the card doesn't want to boot with "CACHE TEST
FAILED" I have encountered several times where the SCSI bus resets
while I'm using it. It usually goes downhill from there so I reboot
anyways. Turning the system off for a minute or so tends to prevent
the cache test failed message from occurring. I'm running on a Cyrix
M-II at [3x83] 250Mhz.

On my noname Alpha (with the 53c810 built in to the MB) I have not
been seeing problems. It is running 2.2.0-pre9/final.

I am at the point where I can't go beyond kernel 2.2.5 due to
filesystem damage potential and a need for a stable system to develop
on that won't lock up hard on a whim (same story; the scsi light stays
on).

I'll be removing one or two things from my SCSI bus soon to see if
that helps, but I never had problems under 2.0.

Greg

> I'm in trouble with some 2.2.x-kernels (2.2.6, too):
> The kernel had nearly nothing compiled in (even no networking), I've
> compiled with
> gcc 2.7.2.3
> and switched the computer on "fresh" (this is not necessary). Sometimes
> everything is okay. But
> too often it hangs with:
>
> ncr53c8xx: at PCI bus 0, device 12, function
> ncr53c8xx: 53c875 detected with Tekram NVRAM
> ncr53c875-0: rev rev=0x03, base=0xdf800000, io_port=0xb000, irq=11
> ncr53c875-0: Tekram format NVRAM, ID 7, Fast-20, Parity Checking
> ncr53c875-0: on-chip RAM at 0xdf000000
> CACHE TEST FAILED: script execution failed.
> start=00007230, pc=00007238, end=0000725c
> CACHE INCORRECTLY CONFIGURED.
> ncr53c875-0: detaching...
>
> And then the kernel fails mounting root.
>
> But if the kernel is doing "normal" its easily possible to crash the
> system (it just stops, the host-adapter-led is on and an led of one or
> more disks, too):
>
> while : ; do ls -laR / ; done ("/" is /dev/sda)
> on another terminal:
> cat /dev/sdc > /dev/null
>
> It takes about 20 seconds. I dont need the "cat" when I start
> "/sbin/update" before.
>
> An other way is:
> cat /dev/sda > /dev/null &
> cat /dev/sdc > /dev/null &
> cat /dev/sdd > /dev/null &
>
> But now I have to switch from one vt to another and back and forth and
> back... I really hold down Ctrl and Alt pressing: F1, F6, F10, F1, F6...
> I twiddle then about 10 sec.
> Maybe I dont need to change the virtual terminals, but after 10 min.
> "catting" I get bored and played again with Ctrl+Alt+Fx,Fy,Fz... - This
> took again only some seconds...
>
> Kernel 2.2.2 is less problematic during reboot (first problem), but here I
> tried only with kernel+raid-patches. - I'm now using this kernel (I depend
> on raid): uname -a
> Linux panorama 2.2.2 #6 Fri Mar 5 01:47:41 CET 1999 i586 unknown
>
> I had no problems (during reboot) with kernel 2.0.3x. And the second
> problem didn't appear. I updated all things mentioned in
> "Documentation/Changes" for 2.2.x
>
> Ah! I forgot to mention, that the SYM53C8XX-Driver was a little bit
> different: already one "ls -laR /" stopped the system...
>
> Who can help me? What should I further report? And most important:
> Which searchengine is _really_ capable looking for:
> ncr53c875 and "CACHE TEST FAILED"
> in the linux-mailing-lists? - I failed here, too.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/