On Jan 17, 2007, at 9:55 PM, [EMAIL PROTECTED] wrote: > [EMAIL PROTECTED] wrote on > 01/17/2007 > 07:42:09 AM: > >> >> On Jan 17, 2007, at 12:27 AM, [EMAIL PROTECTED] wrote: >> >>> anything in the source. I don't want to clutter up the mailing >>> list with >>> newbie questions so I'd like to get in touch with the person who >> >> Ask the questions on the list, thus they are around for others to >> learn from if need by. >> >> - kumar >> > Well, okay. I hate taking up bandwidth for something that's no doubt > trivial to most of you, but here goes. I'm trying to bring up the > 2.6.19.1 kernel using u-boot 1.1.6 on the MPC8349ITX board and I'm > having > problems with the open firmware part of it (which I am TOTALLY new > to). > Short version, I don't think the *.dts file for this board is > correct and > I'm wondering if anyone might have a newer version or good > documentation > as to the formatting of the *.dts files?
There should be an update dts in the kernel source under arch/powerpc/ boot/dts > Long version: While booting the kernel it seems to hang in the > do_div() > call on line 933 of .../arch/powerpc/kernel/time.c (yes I know it's > not > really a call but inline assembly) because ppc_tb_freq is zero. The > get_freq() calls in generic_calibrate_decr() find the 'timebase- > frequency' > and 'clock-frequency' entries in the device tree, but the value is > zero > and thus ppc_tb_freq is zero. Sounds like your u-boot isn't properly filling out parts of the devtree while booting. > Digging through the u-boot code, I found ft_setup() which is > supposed to > add some stuff to the device tree passed in on the bootm command > line. One > of the things it's supposed to do is fill in the 'timebase-frequency' > portion of the device tree. When it calls ft_get_prop() to find the > 'timebase-frequency' node it passes in the path > '/cpus/PowerPC,[EMAIL PROTECTED]/timebase-frequency' as what it's searching > for, but > what ft_get_prop() finds is > '///cpus//cpus/PowerPC,[EMAIL PROTECTED]/timebase-frequency', which of course > > doesn't > match, so it returns NULL to ft_setup() and the 'timebase- > frequency' never > gets updated before being passed to the kernel. Hmm, that's odd. I'm pretty sure the '/' are added while we parse the tree seems like some prefix is getting added on you '///cpus/' if I was to guess. I'd double check your .dts against the one in the kernel tree. > So I'm guessing the my problem is either the *.dts file or dtc, but > I'm > not sure which since I'm so new to this and I'm not sure if the > *.dts file > I've got is properly formatted. I suppose it's possible that > ft_get_prop() in u-boot is wrong, but for now I'm betting it's > right and > my *.dts file is formatted wrong. (FYI, I'm using > .../arch/powerpc/boot/dts/mpc8349emitx.dts from the 2.6.19.1 kernel > tree.) Well that's really odd than. I'd try grabbing u-boot 1.2.0 and see if you still have the same issue. Also, how are you invoking dtc to get the dtb? > Thanks for listening and for your feedback. - k _______________________________________________ Linuxppc-embedded mailing list [email protected] https://ozlabs.org/mailman/listinfo/linuxppc-embedded
