Hi,
short version: I have two aic7xxx SCSI-controllers in my machine and I
can reproduce hard lockups when accessing both simultaneously.
This phenomenon was not present on 2.4.18 vanilla and appears
since 2.4.19 up to (last one tested 2.4.21-pre7). 2.4.20-ac2 also
is affected.
I have two aic7xxx controllers (one 2940U, one UW) on a VIA KT266-based
motherboard with the following SCSI devices connected to them.
2940UW ----(internal 68pin)---- IBM DNES-318350W (Disk) on ID 0
|
+- (active terminator)
2940U ----(internal 50pin)---- PLEXTOR PX-20TS (CD) on ID 5
| |
+--- YAMAHA CRW4416S +- ARCHIVE Python 27871-XXX (Streamer)
(ext) (CD-Writer) |
+- (passive terminator)
This setup seemed to be stable for a long time.
Now I discovered (and verified) that I have problems
with concurrent access between the 2940UW and the 2940U.
A test looks like this:
1.) boot acpi=off aic7xxx=verbose console=ttyS1 init=/bin/sh
2.) init-2.05a# /sbin/devfsd /dev
3.) init-2.05a# cat /dev/scsi/host1/bus0/target6/lun0/cd >/dev/null &
4.) init-2.05a# ls /bin
(hearing harddisk clicking a little...)
---> tadaa: instant hang. No response whatsoever, cursor of the
matrox framebuffer stops blinking
It's also possible to try burning a CD, this also locks the machine
(and that's how I discovered it, who needs CDs anyway these days :-) ).
Permutations I have tested:
- acpi=off/on
- internal/external cable of the 2940U connected/disconnected
- reading Plextor/Yamaha
Affected kernels:
- kernel 2.4.18 works without any problems
- kernel 2.4.19, 20, 21-pre7, 20-ac2 hangs
I don't think that cabling of the SCSI-bus is an issue, because
the problem appears on both the internal and the external bus
independently and it always appears as soon as I access the *other*
SCSI adapter.... And using 2.4.18 I burned many CDs flawlessly...
/proc/interrupts:
0: 6647 XT-PIC timer
1: 135 XT-PIC keyboard
2: 0 XT-PIC cascade
4: 1 XT-PIC serial
8: 1 XT-PIC rtc
11: 3311 XT-PIC aic7xxx, VIA8233
12: 1084 XT-PIC PS/2 Mouse
14: 17 XT-PIC ide0
15: 1114 XT-PIC aic7xxx, eth0, Ensoniq AudioPCI, usb-uhci
NMI: 0
LOC: 6613
ERR: 10
MIS: 0
Attachments (too long to have inline conveniently):
- 2.4.21-p7-no_acpi.txt -> sample boot. 2.4.21-p7 with acpi
patch but acpi disabled
- lspci -> output of lspci -v -v -v
- config-2.4.21... -> .config from 2.4.21-p7 kernel used
If I can help with providing some more datapoints, feel free
to email me.
Chris
-- Gates' Law: Every 18 months, the speed of software halfes.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Apr 23 2003 - 22:00:24 EST