From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 18182 invoked by alias); 12 Aug 2004 16:27:54 -0000 Mailing-List: contact ecos-discuss-help@ecos.sourceware.org; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: ecos-discuss-owner@ecos.sourceware.org Received: (qmail 18167 invoked from network); 12 Aug 2004 16:27:52 -0000 Received: from unknown (HELO SIMON) (69.73.59.253) by sourceware.org with SMTP; 12 Aug 2004 16:27:52 -0000 Received: FROM [192.168.1.1] BY SIMON with hMailServer ; 12-Aug-2004 11:27:40 -0500 Message-ID: <9E9C10C1-7A11-4731-8A57-7E1514E18A2A@SIMON> Reply-To: From: To: Cc: , Date: Thu, 12 Aug 2004 16:27:00 -0000 Organization: BCS MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Subject: [ECOS] PowerPC debugging with Macraigor Raven X-SW-Source: 2004-08/txt/msg00215.txt.bz2 Hello, I am having some trouble with setting and responding to breakpoints while attempting to port eCos over to a MPC860 based platform using a Macraigor Raven BDM pod. I can load insight and get the board initialized so that I have full access to the SDRAM, but when I set a breakpoint after downloading the code, I cannot break at that location. The code is properly downloaded and I can do some tricks to get breakpoints to work, but it is not at all usable. The main problem is that I do not have an Ethernet port available for GDB/Redboot and serial debugging would be far too time consuming to be practical. I have followed all the directions from Macraigor (both their GCC/Insight images as well as ones I built myself). I have also seen most the emails related to this and it appears some people are using the Raven, but I must be missing some special settings in my ECC file or in the gdb init script. Using an Xscale based JTAG debugger from Macraigor (with libremote) on a different project seem to work with respect to debugging eCos based applications, so I feel certain that there may be an easy way to get this done. If anyone should happen to know this solution, I would appreciate it. It looks like my only two solutions is to can eCos/gnu or to covert my ELF file to DWARF and try to debug under Green Hill's MULTI environment. - Chip Boling Attached below is a edited version of my gdb init file in case it is of interest or help with this problem. If anyone is using the Raven pod successfully, a copy of your init script as well as your ECC file would be helpful. I may just be missing some checkmark or option. ################### begin GDB script ################# set endian big target remote localhost:8888 # We need reset to do; system reset, freeze timers, # and single instruction execution. # monitor reset monitor endian big # set the IMMR (spr638) monitor spr638 = 0xff000000 # turn off the watchdog timer --- disable always, enable # bus monitor timeout with maximum time setting # # to disable watchdog -> 0xffffff88 # for 2.6 second timeout-> 0xffffff8d # for 2 second timeout -> 0xbebcff8d # for 1.5 second timeout-> 0x8f0dff8d # for 1 second timeout -> 0x5f5eff8d # for 500 mS timeout -> 0x2fafff8d # for 250 mS timeout -> 0x17d7ff8d # for 100 mS timeout -> 0x0989ff8d # monitor long 0xff000004 0xffffff88 # disable the cache (ic_cst = 560) (dc_cst = spr 568) monitor spr560 = 0x04000000 monitor spr568 = 0x04000000 # Place the processor in its fundamental operating mode (deals with internal pipeline, etc) # ISCT_SER = 0: serialized with instruction fetch show cycles # ISCT_SER = 3: serialized with no instruction fetch show cycles # ISCT_SER = 7: non-serialized with no instruction fetch show cycles (normal mode) # (ictrl = spr158) monitor spr158 = 0x00000007 # Program the SIUMCR: Data Show Cycles disabled monitor long 0xff000000 0x00200400 # Program clock-related registers # The Multiplication Factor must be changed separately. It # will not work when changed within this script. It is only # here for completness. The write to the SCCR does work monitor long 0xff000280 0x01800000 monitor long 0xff000284 0x00400000 # TBD comment out the following code in order to make the setup code do it # later on. # Program miscellaneous SIU registers monitor long 0xff000320 0x55ccaa33 monitor long 0xff000324 0x55ccaa33 monitor long 0xff000328 0x55ccaa33 monitor long 0xff00032c 0x55ccaa33 monitor long 0xff000340 0x55ccaa33 monitor long 0xff000344 0x55ccaa33 monitor short 0xff000240 0x0080 monitor short 0xff000220 0x00c0 monitor short 0xff000200 0x00c0 # Setup UPM tables in order to initialize SDRAM # *** Note MDR, MCR, and MAR code deleted here to make this posting # a little shorter. Also commented out the BR/OR register # settings for the same reason. ######################################################################## # load our file load c:/cygwin/tmp/ecosTest/hello.elf symbol c:/cygwin/tmp/ecosTest/hello.elf ######################################################################## # set up some breakpoints b main b hal_hardware_init -- Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss