On Mon, Jun 6, 2011 at 2:19 PM, Nicolas Pitre <nicolas.pi...@linaro.org> wrote: > On Mon, 6 Jun 2011, Andy Green wrote: > >> On 06/04/2011 02:17 PM, Somebody in the thread at some point said: >> >> Hi - >> >> > When booted on a Beagleboard-xM, the resulting kernel appears to hang >> > after "Starting kernel...". I haven't investigated any further. >> >> Panda chokes the same way with previously OK omap4_defconfig.
I haven't found a precise cause yet, but this definitely looks like it could be a Thumb-2 related issue. An ARM kernel for Panda boots (without device tree, anyway), but a Thumb kernel hangs as you describe. This is similar to the behaviour I observed on mx51evk. >> > [...] >> >> > The .config I'm using is that described on the wiki at >> > https://wiki.linaro.org/Resources/HowTo/KernelDeploy#From_Linaro_sources >> >> with omap2plus_defconfig, after adding EXT4, the userland blows SIGILL -- and >> just SIGILL, not segfaults -- in an unpredictable but apparently >> deterministic >> way. For example ls at bash prompt is OK but ls -l is SIGILL. >> >> The rootfs is linaro Ubuntu autobuilt stuff from a few weeks ago that has >> been >> reliable until this kernel. >> >> I googled around and saw some other guy met this in April with no obvious >> result. It's possibly toolchain related since I am using gcc 4.5.1 Fedora >> ARM >> cross toolchain, but that's a bit subtle since omap4_defconfig is OK and >> expecting that to work is legit. Linus HEAD works fine with it too. >> >> http://xbmc-installer.pastebin.ca/2045257 >> http://pandaboard.org/pbirclogs/index.php?date=2011-04-11#T23:14:08 >> >> I attempted to adapt a few things in omap2plus_defconfig like CONFIG_EMBEDDED >> to omap4_defconfig style but it didn't help; in any event omap4_defconfig is >> broken differently. >> >> I have some dss patches on top but I rewound them for now and just pushed >> three small patches on top to my master (moving the old master to >> "tilt-linux-linaro-2.6.38"). > > Can you test with plain v2.6.39 to see if those problems are present, > and if no then it would be interesting to see the result of a git bisect > run. v2.6.39 does work for me in Thumb-2 on OMAP, so I will have a go at bisecting tomorrow... Cheers ---Dave _______________________________________________ linaro-dev mailing list linaro-dev@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-dev
