For those who don't follow linuxppc-embedded, any ideas. I realized this list might be in a better position to answer this thread.
Thanks,
-- Brian -----Original Message----- From: brian.auld@adic.com [mailto:brian.auld@adic.com] Sent: Friday, January 31, 2003 6:01 PM To: linuxppc-embedded@lists.linuxppc.org Subject: RE: Trying to program u-boot into flash using BDI2000
Yep, I have these (peripheral bus setups) set as per the BDI config file. I'll go ahead and attach the whole thing instead of just [HOST] and [FLASH].
The default boot setup on Ebony is to use a 0.5MB chunk of flash at the very top of the memory map (0xfff80000 --> 0xffffffff). I also just realized that the u-boot.bin that I'm trying to load is 524788bytes. Given that 0.5MB is not exactly 500000bytes and is right around the 524788 number, is it that:
(a) 0.5MB is not enough to hold the u-boot binary?? (b) u-boot is designed to fit like a glove into 0.5MB of flash. It sounds like too much of a coincidence to be right around the 524K number.
Others that have installed u-boot/ppcboot into ebony/walnut flash, have the used the default top 0.5MB chunk or RAM, or did they have to change dip switches to achieve a configuration that provided a larger boot flash area??
-- Brian
== ATTACHMENT OF FULL CONFIG FILE (WITH COMMENTS REMOVE FOR EASY READING ==
[INIT] ; Setup TLB WTLB 0xF0000095 0x1F00003F WTLB 0x00000098 0x0000003F
; Setup Peripheral Bus WDCR 0x12 0x00000010 WDCR 0x13 0x9B015480 WDCR 0x12 0x00000000 WDCR 0x13 0xFFF18000
; Setup SDRAM Controller (DDR SDRAM) ; single-sided,non-buffered, 12x10(4), 128MB DIMM WDCR 0x10 0x00000082 WDCR 0x11 0x40000000 WDCR 0x10 0x00000080 WDCR 0x11 0x410A4012 WDCR 0x10 0x00000081 WDCR 0x11 0x8080082B WDCR 0x10 0x00000040 WDCR 0x11 0x000A4001 WDCR 0x10 0x00000030 WDCR 0x11 0x08200000 WDCR 0x10 0x00000020 WDCR 0x11 0x06000000 WDCR 0x11 0x86000000
[TARGET] JTAGCLOCK 1 CPUTYPE 440 BDIMODE AGENT BREAKMODE SOFT STEPMODE JTAG
[HOST] IP 172.16.77.151 FILE u-boot.bin FORMAT BIN START 0xfff80000 LOAD MANUAL DEBUGPORT 2001
[FLASH] CHIPTYPE AM29F CHIPSIZE 0x80000 BUSWIDTH 8 WORKSPACE 0x000040000 FILE u-boot.bin FORMAT BIN 0x00000000 ERASE 0xFFF80000 ERASE 0xFFF90000 ERASE 0xFFFA0000 ERASE 0xFFFB0000 ERASE 0xFFFC0000 ERASE 0xFFFD0000 ERASE 0xFFFE0000 ERASE 0xFFFF0000
[REGS] IDCR1 0x010 0x011 IDCR2 0x012 0x013 IDCR3 0x014 0x015 IDCR4 0x016 0x017 FILE reg440gp.def
============================= END OF ATTACHMENT ===========================
-----Original Message----- From: Chris Zimman [mailto:chris@cryptoapps.com] Sent: Friday, January 31, 2003 5:28 PM To: brian.auld@adic.com Subject: Re: Trying to program u-boot into flash using BDI2000
On Fri, Jan 31, 2003 at 01:18:32PM -0800, brian.auld@adic.com wrote:
J39 is confirmed open (i.e. write enabled).
Are the peripheral bus setups correct?
These are what I was using
; Setup Peripheral Bus WDCR 0x12 0x00000010 ;Select EBC0_B0AP WDCR 0x13 0x9B015480 ;B0AP: Flash and SRAM WDCR 0x12 0x00000000 ;Select EBC0_B0CR WDCR 0x13 0xFFF18000 ;B0CR: 1MB at 0xFFF00000, r/w, 8bit
--Chris
-----Original Message----- From: Chris Zimman [mailto:chris@cryptoapps.com] Sent: Friday, January 31, 2003 3:05 PM To: brian.auld@adic.com Subject: Re: Trying to program u-boot into flash using BDI2000
On Fri, Jan 31, 2003 at 11:40:36AM -0800, brian.auld@adic.com wrote:
Just as a first -- did you check the flash write protect jumper (J39)? Should be open.
--Chris
The embedded linux on ebony(440GP) saga continues:
- ELDK in place (check)
- kernel compiled (check)
- u-boot compiled (check)
- ebony in place (check ... dip switch settings confirmed)
- BDI2000 in hand (arrived this morning)
OK, I'm trying to program flash with u-boot (I'm using u-boot.bin). I am using the 440GP config file provided by abatron, plus the [FLASH]
parameters
from the example in the bdiGDB users manual for programming flash on the 405. It looks like the memory map as it pertains to flash as well as the part (AMD29F040) was consistent from the 405 --> 440, so it seemed
logical
to use these FLASH values. The 440GP config file provided does not have
any
FLASH parameters set, which is why I made this change.
Well, I successfully erased the flash from 0xfff8000 - 0xffffffff and checked the locations after the erase ... they were all 'F's. I then
tried
to program the flash with my u-boot.bin and got the following:
===== start of bdi clip ===== BDI>prog 0xfff80000 u-boot.bin bin= Programming u-boot.bin , please wait .... # Programming flash memory failed at 0xfff80000 BDI>prog 0xfff80000 u-boot.bin bin Programming u-boot.bin , please wait .... # Programming flash memory failed at 0xfff80000 BDI> ===== end of bdi clip =====
Any ideas why this failed. Attached is my config file. I'll just attach
the
[HOST] and [FLASH] sections to make it a little easier on the eyes. For anyone familiar with BDI sample 440GP confile file, you will see that it
is
this file with the [FLASH] additions I mentioned above. Actually, after looking at the file, I remembered that I also trimmed information in the [HOST] section ... specifically stuff that was presuming path to Windows information. My host is Linux.
===== start of config file ([HOST] and [FLASH] sections) ===== ;bdiGDB configuration file for IBM 440GP Reference Board
-- cut --
[HOST] IP 172.16.77.151 ;Linux host ;FILE u-boot.bin ;FORMAT BIN ;START 0x200000 LOAD MANUAL ;load code MANUAL or AUTO after reset DEBUGPORT 2001 DUMP dump.bin ;Linux: dump.bin must already exist and public writable
[FLASH] CHIPTYPE AM29F ; flash part = AMD29F040B CHIPSIZE 0x80000 ;The size of one flash chip in bytes
(0.5MB)
BUSWIDTH 8 ;The width of the flash memory bus in bits
(8
|
16 | 32) WORKSPACE 0x000040000 ; RAM buffer for fast flash programming FILE u-boot.bin ;The file to program FORMAT BIN 0x00000000 ERASE 0xFFF80000 ERASE 0xFFF90000 ERASE 0xFFFA0000 ERASE 0xFFFB0000 ERASE 0xFFFC0000 ERASE 0xFFFD0000 ERASE 0xFFFE0000 ERASE 0xFFFF0000
-- cut --
===== end of config file ( with cuts) =====
-- Brian
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/