Still on kexec for SH4 architecture, please help
matteo_pampolini at yahoo.it
Fri Feb 1 08:35:57 EST 2008
my name is Matteo and I'm writing from Italy. As you
can probably understand from the subject, I work with
Paolo Cortesi that recently asked for help on this
issue, unfortunately he has problems posting a
readable message, so here I am.
We are working on an SH4 architecture and we have to
need to replace our running kernel with a new one,
that's why we are using kexec, of course (BTW, a
stupid question: the new kernel will have all the RAM
available as the first one, am I right? No memory
pages are wasted during the transition, aren't they?).
We downloaded kexec-tools-testing-20071030 package, we
enabled kexec support in kernel, and launched:
./kexec --load --type=zImage-sh
and the result is always an oops:
Starting new kernel
segment: 0x0c001000 - 0x0c002000 (0x00001000)
segment: 0x0c210000 - 0x0c387000 (0x00177000)
start : 0x8c210000
illegal slot instruction: 01a0 [#1]
Pid : 92, Comm: kexec
PC is at 0x8c210008
PC : 8c210008 SP : 8a710000 SR : 400080f0 TEA :
2966e480 Not tainted
R0 : 00000004 R1 : fffffff0 R2 : a0000000 R3 :
R4 : aa7105ec R5 : ac387000 R6 : 8c210000 R7 :
R8 : 00001000 R9 : a0000000 R10 : 8a89c600 R11 :
R12 : 8441ab00 R13 : 0a710002 R14 : 00000000
MACH: 00000000 MACL: 00000000 GBR : 29573b18 PR :
Apparently is the old kernel that crashes during
relocation, but I'm not sure of it.
The empty_zero value is grabbed from System.map, but
please tell me, why start is 8c210000? I had a look at
kexec sources for SH, saw that some masks are applied
to addresses, but did not understand why.
Any help is really really appreciated, thank you all.
L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail: http://it.docs.yahoo.com/nowyoucan.html
More information about the kexec