On Tue, 18 Jan 2000, Juergen Bierlein wrote:
>Rich Payne wrote:
>> Which SCSI card are you using?
>
>Ok here is my configuration:
>PC164 with 500 Mhz 21164a CPU
>256 MB RAM (4x64 MB)
>Adaptec 2940UW (PCI)
>Teles ISDN 16.3 (ISA)
>Soundblaster 16ASP (ISA)
>Matrox Millenium I 8MB
>internal SCSI-CDROM NEC 3xi
>2 x 4 GB SCSI Disks (IBM)
>1 x 18 GB SCSI Disks (IBM)
>
>OS is SuSE 6.1 AXP (previous version was Redhat 5.1).
>Current Kernel is 2.2.14 with the named problems.
Are you using modules? If so please use 2.2.14aa1 (patch against 2.2.14).
I fixed an alpha bug that may explain weirs lockups using modules (related
to self-modifying code and icache-coherency and icache virtual-caches).
ftp://ftp.*.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.2/2.2.14aa1.gz
The 2.2.14aa1 will also extend the max limit of memory available on alpha
from 2G to 2T and lots of other VM/SMP quite necessary things. Just have a
look at the process I run over it :)
andrea@D104:~ > cat /proc/7366/status
Name: leak
State: S (sleeping)
Pid: 7366
PPid: 6687
Uid: 557 557 557 557
Gid: 50 50 50 50
Groups: 50
VmSize: 9439640 kB
VmLck: 0 kB
VmRSS: 9278640 kB
VmData: 9437216 kB
VmStk: 16 kB
VmExe: 8 kB
VmLib: 1304 kB
SigPnd: 0000000000000000
SigBlk: 0000000000000000
SigIgn: 0000000000000000
SigCgt: 0000000000000000
CapInh: 00000000fffffeff
CapPrm: 0000000000000000
CapEff: 0000000000000000
BTW, the strict patch for the module-load bug is here:
ftp://ftp.*.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.2/2.2.14aa1/alpha-imb-1.gz
Andrea
