Date: Sun, 13 Jun 2010 22:20:18 +0200 From: Andreas Tobler <andreast-list@fgznet.ch> To: Marius Strobl <marius@alchemy.franken.de> Cc: freebsd-sparc64@freebsd.org Subject: Re: U60 doesn't boot anymore with -CURRENT Message-ID: <4C153D82.9050302@fgznet.ch> In-Reply-To: <4C1530F1.6070207@fgznet.ch> References: <4C14ED7C.2030809@fgznet.ch> <20100613191518.GA95797@alchemy.franken.de> <4C1530F1.6070207@fgznet.ch>
next in thread | previous in thread | raw e-mail | index | archive | help
On 13.06.10 21:26, Andreas Tobler wrote: > On 13.06.10 21:15, Marius Strobl wrote: >> On Sun, Jun 13, 2010 at 04:38:52PM +0200, Andreas Tobler wrote: >>> Hi, >>> >>> my gecko testmachine refuses to boot with -CURRENT. >>> >>> Here the verbose part of the messages: >>> >>> --- >>> FreeBSD 9.0-CURRENT #0 r209123M: Sun Jun 13 14:07:28 CEST 2010 >>> >>> andreast@u60.andreas.nets:/usr/obj/export/devel/src/sys/GENERIC sparc64 >>> Preloaded elf kernel "/boot/kernel/kernel" at 0xc0bbc000. >>> real memory = 1610612736 (1536 MB) >>> avail memory = 1554571264 (1482 MB) >>> machine: SUNW,Ultra-60 >>> cpu0: Sun Microsystems UltraSparc-II Processor (449.99 MHz CPU) >>> mask=0xa0 maxtl=5 maxwin=7 >>> initalizing intr_countp >>> --- >>> >>> And here it hangs. >>> >>> Any idea on hwat is going wrong? >>> >> >> Nothing obvious as no sparc64 specific changes went in lately. >> If you use a kernel built with options ALT_BREAK_TO_DEBUGGER, >> can you break into the debugger via the alternate break sequence >> on the serial console and get backtrace when it hangs? What is >> the last known working revision? Can you bisect which revsion >> triggers the problem? > > I tracked it down to r207537, especially mp_locore.S. > If I use r207536 from mp_locore.S then my machine boots again. But this brings me to boot again: Index: mp_locore.S =================================================================== --- mp_locore.S (revision 207537) +++ mp_locore.S (working copy) @@ -207,7 +207,7 @@ bl %icc, 2f nop cmp %l1, CPU_IMPL_ULTRASPARCIII - bl %icc, 3f + bl %icc, 4f nop 2: mov CPU_STICKSYNC, %l2 membar #StoreLoad I think it is an obvious one, you changed the labels but you forgot to adjust this one. Am I right? Gruss, Andreas
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4C153D82.9050302>