From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.dev.rtsoft.ru (unknown [85.21.88.2]) by ozlabs.org (Postfix) with SMTP id 33C3EDDEE1 for ; Sat, 5 May 2007 01:36:20 +1000 (EST) Message-ID: <463B5422.6030105@ru.mvista.com> Date: Fri, 04 May 2007 19:41:22 +0400 From: Andrei Konovalov MIME-Version: 1.0 To: Mirek23 Subject: Re: zImage.elf loads but does not start References: <10079346.post@talk.nabble.com> <10142161.post@talk.nabble.com> <406A31B117F2734987636D6CCC93EE3C0159BAAE@ehost011-3.exch011.intermedia.net> <10165159.post@talk.nabble.com> <10323147.post@talk.nabble.com> In-Reply-To: <10323147.post@talk.nabble.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Cc: linuxppc-embedded@ozlabs.org List-Id: Linux on Embedded PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Mirek23 wrote: > I came to the point with my kernel 2.6 (patched by Grant) struggle on ml403 > that it reaches the > stage to mount nfs root file system from the computer which is on another > segment of the network: > > [ 2.565946] eth%d: XTemac: PHY detected at address 3. > [ 2.629348] eth0: Xilinx TEMAC #0 at 0x81200000 mapped to 0xC5020000,rq=1 > [ 2.713113] eth0: XTemac id 1.0f, block id 5, type 8 > [ 2.774335] mice: PS/2 mouse device common for all mice > [ 2.837152] TCP cubic registered > [ 2.876220] NET: Registered protocol family 1 > [ 2.928569] NET: Registered protocol family 17 > [ 3.485990] eth0: XTemac: Options: 0xb8f2 > [ 7.517827] eth0: XTemac: We renegotiated the speed to: 100 > [ 7.584504] eth0: XTemac: speed set to 100Mb/s > [ 8.652764] IP-Config: Unable to set interface netmask (-22). > [ 8.724733] Looking up port of RPC 100003/2 on 129.129.130.106 > > It cannot however go through the gateway because the macaddress is not > properly set (ie. it is set to 00:00:00:00). > My command line argument looks like: > > CONFIG_CMDLINE="console=ttyUL0,9600 root=/dev/nfs rw > nfsroot=129.129.130.106:/opt/eldk41/ppc_4xx,tcp macaddr=00:50:c2:46:d8:01 > ip=129.129.129.29:129.129.130.106:129.129.129.1:255.255.255.0:ML403:eth0:off > panic=1" > > Does somebody know why the macaddress is not taken into account. > > Mirek > Not sure how "macaddr=" in the cmdline is processed, but here is what I've used to play with TEMAC (this is essentially a copy from 2.4 kernels). The patch is few month old, and could not apply cleanly. Especially now when the Xilinx Virtex boards has been switched to the u-boot's bd_t structure (a very recent change by Grant Likely; probably not yet in the tree you are using). So this patch is rather an illustration. Thanks, Andrei --- linux-2.6.20.orig/arch/ppc/boot/simple/embed_config.c +++ linux-2.6.20/arch/ppc/boot/simple/embed_config.c @@ -778,6 +778,20 @@ embed_config(bd_t ** bdp) bd->bi_intfreq = XPAR_CORE_CLOCK_FREQ_HZ; bd->bi_busfreq = XPAR_PLB_CLOCK_FREQ_HZ; bd->bi_pci_busfreq = XPAR_PCI_0_CLOCK_FREQ_HZ; + + /* Set the address to Xilinx's preferred default. + * However, first to eliminate a compiler warning + * because we don't really use def_enet_addr, + * we'll reference it. The compiler + * optimizes it away so no harm done. */ + bd->bi_enetaddr[0] = def_enet_addr[0]; + bd->bi_enetaddr[0] = 0x00; + bd->bi_enetaddr[1] = 0x0A; + bd->bi_enetaddr[2] = 0x35; + bd->bi_enetaddr[3] = 0x00; + bd->bi_enetaddr[4] = 0x22; + bd->bi_enetaddr[5] = 0x00; + timebase_period_ns = 1000000000 / bd->bi_tbfreq; /* see bi_tbfreq definition in arch/ppc/platforms/4xx/xilinx_ml300.h */ }