Bug 8372 - [PPC64] kernel won't boot with GXT2000 if FB_OF and FRAMEBUFFER_CONSOLE are enabled
Summary: [PPC64] kernel won't boot with GXT2000 if FB_OF and FRAMEBUFFER_CONSOLE are e...
Status: RESOLVED OBSOLETE
Alias: None
Product: Drivers
Classification: Unclassified
Component: Console/Framebuffers (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: James Simmons
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-04-26 07:16 UTC by o587914
Modified: 2013-12-23 18:40 UTC (History)
5 users (show)

See Also:
Kernel Version: 2.6.27
Tree: Mainline
Regression: No


Attachments
/proc/device-tree with GXT2000 (tar.bz2) (11.25 KB, application/x-bzip)
2007-04-26 22:25 UTC, o587914
Details

Description o587914 2007-04-26 07:16:43 UTC
Most recent kernel where this bug did *NOT* occur: 2.6.17.14, 
Distribution: Gentoo Linux (PowerPC64, 64bit kernel) + kernel.org kernel
Hardware Environment: IBM RS/6000 44P Model 170 Power3
Software Environment: gcc-4.1.1
Problem Description:

If I enable both FB_OF and FRAMEBUFFER_CONSOLE my machine won't show anything
more than this on the serial line (console=ttyS0,9600 is set) and the machine
won't come up at all (you know, maybe it would show something on the not
connected screen), so I think the kernel just stops at this point:

###############################################################################
Please wait, loading kernel...
   Elf64 kernel loaded...
Loading System.map ...
System.map loaded at 01a00000, size: 1024 Kbytes
OF stdout device is: /pci@fef00000/isa@b/serial@i3f8
command line: root=/dev/sda3 ro console=ttyS0,9600 root=/dev/sda3 
memory layout at init:
  alloc_bottom : 00000000018f6000
  alloc_top    : 0000000020000000
  alloc_top_hi : 0000000020000000
  rmo_top      : 0000000020000000
  ram_top      : 0000000020000000
Looking for displays
found display   : /pci@fee00000/display@b, opening ... done
opening PHB /pci@fef00000... done
opening PHB /pci@fee00000... done
instantiating rtas at 0x000000001efb5000 ... done
copying OF device tree ...
Building dt strings...
Building dt structure...
Device tree strings 0x0000000001bf7000 -> 0x0000000001bf7fc4
Device tree struct  0x0000000001bf8000 -> 0x0000000001bfd000
Calling quiesce ...
returning from prom_init
Using pSeries machine description
Starting Linux PPC64 #1 Thu Apr 26 13:08:40 UTC 2007
-----------------------------------------------------
ppc64_pft_size                = 0x0                  
physicalMemorySize            = 0x20000000
ppc64_caches.dcache_line_size = 0x80      
ppc64_caches.icache_line_size = 0x80
htab_address                  = 0xc00000001e000000
htab_hash_mask                = 0xffff            
-----------------------------------------------------
Linux version 2.6.18.8 (root@power3) (gcc version 4.1.1 (Gentoo 4.1.1-r3)) #1
Thu Apr 26 13:08:40 UTC 27
[boot]0012 Setup Arch                                                          
                       
No ramdisk, default root is /dev/sda2
Python workaround: reg0: 18e3388     
Python workaround: reg0: 18eb788
PPC64 nvram contains 110592 bytes
[boot]0015 Setup Done            
Built 1 zonelists.  Total pages: 131072
Kernel command line: root=/dev/sda3 ro console=ttyS0,9600 root=/dev/sda3 
mpic: Setting up MPIC " MPIC     " version 1.2 at ffc00000, max 8 CPUs   
mpic: ISU size: 16, shift: 4, mask: f                                 
mpic: Initializing for 32 sources    
mpic: Detected reversed IPI registers
i8259 legacy interrupt controller initialized
PID hash table entries: 4096 (order: 12, 32768 bytes)
###############################################################################

The machine displays this on the front panel: "4B27D4C4 \n P1-C1"

If I use my GXT4500 then the machine boots just fine with the exact same kernel.
Comment 1 o587914 2007-04-26 08:31:26 UTC
same for 2.6.21
Comment 2 Antonino Daplas 2007-04-26 14:01:03 UTC
Most recent kernel where this bug did *NOT* occur: 2.6.17.14, 
Distribution: Gentoo Linux (PowerPC64, 64bit kernel) + kernel.org kernel
Hardware Environment: IBM RS/6000 44P Model 170 Power3
Software Environment: gcc-4.1.1
Problem Description:

If I enable both FB_OF and FRAMEBUFFER_CONSOLE my machine won't show anything
more than this on the serial line (console=ttyS0,9600 is set) and the machine
won't come up at all (you know, maybe it would show something on the not
connected screen), so I think the kernel just stops at this point:

###############################################################################
Please wait, loading kernel...
   Elf64 kernel loaded...
Loading System.map ...
System.map loaded at 01a00000, size: 1024 Kbytes
OF stdout device is: /pci@fef00000/isa@b/serial@i3f8
command line: root=/dev/sda3 ro console=ttyS0,9600 root=/dev/sda3 
memory layout at init:
  alloc_bottom : 00000000018f6000
  alloc_top    : 0000000020000000
  alloc_top_hi : 0000000020000000
  rmo_top      : 0000000020000000
  ram_top      : 0000000020000000
Looking for displays
found display   : /pci@fee00000/display@b, opening ... done
opening PHB /pci@fef00000... done
opening PHB /pci@fee00000... done
instantiating rtas at 0x000000001efb5000 ... done
copying OF device tree ...
Building dt strings...
Building dt structure...
Device tree strings 0x0000000001bf7000 -> 0x0000000001bf7fc4
Device tree struct  0x0000000001bf8000 -> 0x0000000001bfd000
Calling quiesce ...
returning from prom_init
Using pSeries machine description
Starting Linux PPC64 #1 Thu Apr 26 13:08:40 UTC 2007
-----------------------------------------------------
ppc64_pft_size                = 0x0                  
physicalMemorySize            = 0x20000000
ppc64_caches.dcache_line_size = 0x80      
ppc64_caches.icache_line_size = 0x80
htab_address                  = 0xc00000001e000000
htab_hash_mask                = 0xffff            
-----------------------------------------------------
Linux version 2.6.18.8 (root@power3) (gcc version 4.1.1 (Gentoo 4.1.1-r3)) #1
Thu Apr 26 13:08:40 UTC 27
[boot]0012 Setup Arch                                                          
                       
No ramdisk, default root is /dev/sda2
Python workaround: reg0: 18e3388     
Python workaround: reg0: 18eb788
PPC64 nvram contains 110592 bytes
[boot]0015 Setup Done            
Built 1 zonelists.  Total pages: 131072
Kernel command line: root=/dev/sda3 ro console=ttyS0,9600 root=/dev/sda3 
mpic: Setting up MPIC " MPIC     " version 1.2 at ffc00000, max 8 CPUs   
mpic: ISU size: 16, shift: 4, mask: f                                 
mpic: Initializing for 32 sources    
mpic: Detected reversed IPI registers
i8259 legacy interrupt controller initialized
PID hash table entries: 4096 (order: 12, 32768 bytes)
###############################################################################

The machine displays this on the front panel: "4B27D4C4 \n P1-C1"

If I use my GXT4500 then the machine boots just fine with the exact same kernel.
Comment 3 Benjamin Herrenschmidt 2007-04-26 14:30:40 UTC
Please send me a tarball of /proc/device-tree with the GXT2000 in
Comment 4 o587914 2007-04-26 22:25:06 UTC
Created attachment 11281 [details]
/proc/device-tree with GXT2000 (tar.bz2)
Comment 5 o587914 2007-04-26 22:29:40 UTC
mhh.. bugzilla doesn't seem to like tar.bz2 that much.. (always downlods as
"attachment.cgi") I'll send you a mail with the tarball, too.
Comment 6 Natalie Protasevich 2007-07-04 13:19:48 UTC
Any update on this so far, have you tried most recent kernels?
Thanks.
Comment 7 o587914 2007-08-13 12:26:35 UTC
2.6.22.2 also shows this behaviour.
Comment 8 Greg Trigg 2007-09-23 05:13:21 UTC
I'm using a GXT2000P and a GXT3000P.  Both show the same problems as those mentioned by Markus Rothe.  I have four 44P-170's and they all fail to boot in the same way as comment #2.  I have two of each cards I mentioned.  As mentioned by Markus, kernel 2.6.17 does not have this problem.(In reply to comment #7)
> 2.6.22.2 also shows this behaviour.
> 
Comment 9 Greg Trigg 2007-11-03 07:21:38 UTC
Correction to my previous comment:
The kernel will boot with the GXT3000P, it just won't show any video.  I was able to ssh into it.  The GXT2000, however, will not boot.

With kernel 2.6.22.6 and the GXT2000 adapter, my 44P-170 front panel is showing the same error.  P1-C1 is the CPU.  Having the Service manual and the same error for the 44P-170, I looked up the 8-digit code.

4B27D4C4 is a system bus time-out error (checkstop) for the 400MHz processor.  It signifies that an adapter has failed.

The service manual test procedures say to remove adapters one by one and see if the same error code comes up.  I removed the video card.  The machine booted and I was able to ssh into it.
Comment 10 Alan 2008-09-26 05:50:14 UTC
Closing out stale bugs, please re-open and update version if still present
Comment 11 Greg Trigg 2008-10-13 04:15:38 UTC
I have re-tested this with the latest kernels (2.6.24,2.6.25,2.6.26,2.6.27) and the problem still exists.  Other cards will still allow the RS6000 to boot even if the video is not shown.  The GXT2000P will not allow boot up at all.

I tried to change the version information but was not allowed.
Comment 12 Alan 2013-12-23 18:40:55 UTC

Closing out stale bugs, please re-open and update version if still present

Note You need to log in before you can comment on or make changes to this bug.