Ultra 5 Boot Hang Update.

[Yoshihiko Iwama]

  This mail does not describe the solution but I believe it contains useful 
content.

  I saw the boot -v log and thought that it contained two important 
sections.
These are:

from the line "pcib1: <APB PCI-PCI bridge> at device 1.1 on pci0" to last 
line

and

from the line "found-> vendor=0x1002, dev=0x4754, revid=0x9a" to
last line(included above).

  The memory range decoded by APB is displied "0xe0000000-0xffffffff" but
assigned range is form "0x0"(to "0xfff") because displied by next line:

"  map[18]: type Memory, range 32, base 0, size 12, enabled"

  Obviously, 0x0 is not within the range of 0xe0000000-0xffffffff, so this 
is
probably an incorrect value.
  I also look the comment in /sys/dev/fb/machfb.c:

===
/*
* Depending on the firmware version the VGA I/O and/or memory
* resources of the Mach64 chips come up disabled.  These will be
* enabled by pci(4) when activating the resource in question but
* this doesn't necessarily mean that the resource is valid.
* Invalid resources seem to have in common that they start at
* address 0.  We don't allocate the VGA memory in this case in
* order to avoid warnings in apb(4) and crashes when using this
* invalid resources.  X.Org is aware of this and doesn't use the
* VGA memory resource in this case (but demands it if it's valid).
*/
===

  Here, compared with the boot -v log of PR121539, assigned rande is
from "0xe2000000"(to "0xe2000fff") by below line:

"  map[18]: type Memory, range 32, base rxe2000000, size 12, enabled"

  I believe there may be problems with allocating memory on the bus.


  The story changes, about the second section, this is about graphics chips.
The vender ID=0x1002 and dev ID=0x4754 means AMD(formerly ATI) and
Rage 3D II Graphics Accelerator by http://pcidatabase.com/ or source code.
  Compared with the boot -v log of PR121539, the dev ID is 0x4750 in
PR121539. This means ATI 3D Rage Pro 215GP.
  Perhaps there is a problem with FreeBSD and may not be able to cope well
with ATI 3D Rage Pro 215 GP.


  So, what to do the most recently, I suggest you connect serial console to
Ultra 5. Of course, even if you simply connect to the serial console, it 
will hang
if ATI 3D Rage Pro 215GP is detected on FreeBSD. Therefore, disable ATI 3D
Rage Pro 215GP. This method is described below:

https://docs.oracle.com/cd/E19455-01/806-4629-05/6jdmq4im5/index.html

  Unfortunately, since I do not own Ultra 5, I have not tried the above 
method
in practice. Therefore, the above method merely invalidates the graphic chip
on the OBP, and it may still be detected by FreeBSD. In this way, Ultra 5 
will
hang as well.
  I advise that you should make sure you can make serial console connections
before disabling graphics chips.

---
Yoshihiko Iwama

-----Original Message----- 
From: Zaphod Beeblebrox
Sent: Tuesday, May 9, 2017 6:11 AM
To: freebsd-sparc64 at freebsd.org
Subject: Ultra 5 Boot Hang Update.

so the whole boot -v is attached via the link below.  Please help.  Hangs
forever after the last pcib1 line (and STOP-A doesn't drop to a prompt)
... just to jog everyone's memory, to get this going I (translate / to
enter):

set-defaults/1 0 mkp/80 1 mkp/8 2 mkp/0 3 mkp/20 4 mkp/c0 5 mkp/ff 6 mkp/ee
7 mkp/0 8 mkp/0 9 mkp/0 a mkp/0 b mkp/c0 c mkp/ff d mkp/ee e mkp/0 f 0 do i
idprom@ xor loop f mkp

... does that initialization pose some problem for FreeBSD?

https://owncloud.towernet.ca/index.php/s/14awIqSzdOBexok
_______________________________________________
freebsd-sparc64 at freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-sparc64
To unsubscribe, send any mail to "freebsd-sparc64-unsubscribe at freebsd.org"