[linux-sparc-users] M3000 failing to boot with "ERROR: CPU RED_State Reset has occurred."

[Kyle K]

Hello,

I have been trying to get any SPARC based Linux distribution working 
without success on M3000 server with SPARC64 VII CPU.

Today I came across this Oracle's reference Linux distribution and was 
excited to try it out hoping it would work, but it is failing on same 
error that other distros did. I tried Debian 7.11.0 (which is the last 
SPARC based Debian ISO before support got dropped in 8), tried Gentoo 
latest minimal CD dated 20141201 and also tried older ISO from 2008.  
They all fail with "ERROR: CPU RED_State Reset has occurred."

---

XSCF> console -d 0 -y

Console contents may be logged.
Connect to DomainID 0?[y|n] :y

{0} ok boot cdrom
Boot device: /pci at 0,600000/pci at 0/pci at 0/scsi at 0/disk at 4,0:f  File and args:
SILO Version 1.4.14 - Rel: 4.0.18.el6
\
Use <TAB> to see valid boot labels or <RETURN> for default action.
Alternatively, type 'halt' to drop back to the OpenBoot prompt.

boot:
Allocated 64 Megs of memory at 0x40000000 for kernel
Loaded kernel version 4.1.8
Loading initial ramdisk (42887664 bytes at 0x4400000 phys, 0x40C00000 
virt)...
/
PROMLIB: Sun IEEE Boot Prom 'OBP 4.33.0.a 2011/04/26 11:32'
PROMLIB: Root node compatible:
Initializing cgroup subsys cpuset
Initializing cgroup subsys cpu
Initializing cgroup subsys cpuacct
Linux version 4.1.8-15.1.el6uek.sparc64 
(mockbuild at sparc-ol6-builder-01.us.oracle.com) (gcc version 4.4.7 
20120313 (Red Hat 4.4.7-4.0.13) (GCC) ) #1 SMP Mon Oct 5 17:48:42 PDT 2015
bootconsole [earlyprom0] enabled
ARCH: SUN4U
Ethernet address: 00:0b:5d:e2:5c:e7
MM: PAGE_OFFSET is 0xfffff80000000000 (max_phys_bits == 40)
MM: VMALLOC [0x0000000100000000 --> 0x0000060000000000]
MM: VMEMMAP [0x0000060000000000 --> 0x00000c0000000000]
Kernel: Using 4 locked TLB entries for main kernel image.
Remapping the kernel... done.
ERROR: CPU RED_State Reset has occurred.
Resetting...
POST Sequence 01 CPU Check
POST Sequence 02 Banner
LSB#00 (XSB#00-0): POST 2.15.0 (2010/10/06 14:25)
POST Sequence 03 Fatal Check
POST Sequence 04 CPU Register
POST Sequence 05 STICK
POST Sequence 06 MMU
POST Sequence 07 Memory Initialize
POST Sequence 08 Memory
POST Sequence 09 Raw UE In Cache
POST Sequence 0A Floating Point Unit
POST Sequence 0B SC
POST Sequence 0C Cacheable Instruction
POST Sequence 0D Softint
POST Sequence 0E CPU Cross Call
POST Sequence 0F CMU-CH
POST Sequence 10 PCI-CH
POST Sequence 11 Master Device
POST Sequence 12 DSCP
POST Sequence 13 SC Check Before STICK Diag
POST Sequence 14 STICK Stop
POST Sequence 15 STICK Start
POST Sequence 16 Error CPU Check
POST Sequence 17 System Configuration
POST Sequence 18 System Status Check
POST Sequence 19 System Status Check After Sync
POST Sequence 1A OpenBoot Start...
POST Sequence Complete.

SPARC Enterprise M3000 Server, using Domain console
Copyright (c) 1998, 2011, Oracle and/or its affiliates. All rights reserved.
Copyright (c) 2011, Oracle and/or its affiliates and Fujitsu Limited. 
All rights reserved.
OpenBoot 4.33.0.a, 32768 MB memory installed, Serial #10050791.
Ethernet address 0:b:5d:e2:5c:e7, Host ID: 80995ce7.



Rebooting with command: boot
Boot device: cdrom  File and args:
SILO Version 1.4.14 - Rel: 4.0.18.el6
\
Use <TAB> to see valid boot labels or <RETURN> for default action.
Alternatively, type 'halt' to drop back to the OpenBoot prompt.

boot:
Allocated 64 Megs of memory at 0x40000000 for kernel
Loaded kernel version 4.1.8
Loading initial ramdisk (42887664 bytes at 0x4400000 phys, 0x40C00000 
virt)...
/
PROMLIB: Sun IEEE Boot Prom 'OBP 4.33.0.a 2011/04/26 11:32'
PROMLIB: Root node compatible:
Initializing cgroup subsys cpuset
Initializing cgroup subsys cpu
Initializing cgroup subsys cpuacct
Linux version 4.1.8-15.1.el6uek.sparc64 
(mockbuild at sparc-ol6-builder-01.us.oracle.com) (gcc version 4.4.7 
20120313 (Red Hat 4.4.7-4.0.13) (GCC) ) #1 SMP Mon Oct 5 17:48:42 PDT 2015
bootconsole [earlyprom0] enabled
ARCH: SUN4U
Ethernet address: 00:0b:5d:e2:5c:e7
MM: PAGE_OFFSET is 0xfffff80000000000 (max_phys_bits == 40)
MM: VMALLOC [0x0000000100000000 --> 0x0000060000000000]
MM: VMEMMAP [0x0000060000000000 --> 0x00000c0000000000]
Kernel: Using 4 locked TLB entries for main kernel image.
Remapping the kernel... done.
ERROR: CPU RED_State Reset has occurred.
Resetting...

...

---



Here's a log from Debian 7.11.0

---

{0} ok boot cdrom
...
boot:
Allocated 64 Megs of memory at 0x40000000 for kernel
EXT2 superblock magic is wrong
Loaded kernel version 3.2.78
EXT2 superblock magic is wrong
Loading initial ramdisk (5041962 bytes at 0x7FE000000 phys, 0x40C00000 
virt)...
\
[    0.000000] PROMLIB: Sun IEEE Boot Prom 'OBP 4.33.0.a 2011/04/26 11:32'
[    0.000000] PROMLIB: Root node compatible:
[    0.000000] Initializing cgroup subsys cpuset
[    0.000000] Initializing cgroup subsys cpu
[    0.000000] Linux version 3.2.0-4-sparc64 
(debian-kernel at lists.debian.org) (gcc version 4.6.3 (Debian 4.6.3-14) ) 
#1 Debian 3.2.78-1
[    0.000000] bootconsole [earlyprom0] enabled
[    0.000000] ARCH: SUN4U
[    0.000000] Ethernet address: 00:0b:5d:e2:5c:e7
[    0.000000] Kernel: Using 2 locked TLB entries for main kernel image.
[    0.000000] Remapping the kernel... done.
ERROR: CPU RED_State Reset has occurred.
Resetting...

---



Here's a log from Gentoo 20141201

---

{0} ok boot cdrom
...
boot: gentoo
Allocated 64 Megs of memory at 0x40000000 for kernel
Loaded kernel version 3.14.14
Loading initial ramdisk (2221856 bytes at 0x7FE008000 phys, 0x40C00000 
virt)...
/
PROMLIB: Sun IEEE Boot Prom 'OBP 4.33.0.a 2011/04/26 11:32'
PROMLIB: Root node compatible:
Linux version 3.14.14-gentoo (root at bender) (gcc version 4.7.3 (Gentoo 
4.7.3-r1 p1.5, pie-0.5.5) ) #1 SMP Wed Dec 3 13:49:00 UTC 2014
bootconsole [earlyprom0] enabled
ARCH: SUN4U
Ethernet address: 00:0b:5d:e2:5c:e7
PAGE_OFFSET is 0xffffff0000000000 (max_phys_bits == 40)
Kernel: Using 4 locked TLB entries for main kernel image.
Remapping the kernel... done.
ERROR: CPU RED_State Reset has occurred.
Resetting...

---


The only Operating System beside Solaris that I found to bootup and work 
is OpenBSD. FreeBSD doesn't work as currently it does not support my 
SPARC64 VII from what I read due to lack of manpower. NetBSD probably 
works but I haven't tried since I'm more interested in GNU/Linux.


Please let me know if there's something that can be done to solve this 
issue. Maybe the error was corrected in newer kernel version? I searched 
git logs of Linus' tree for 'CPU RED_State Reset' but didn't find any 
matches.


Thank you.