Re: qemu for rpi emulation

[Sergio López <slp%sinrega.org@localhost>]

2013/8/15 Jeff Rizzo <riz%tastylime.net@localhost>:
> I started a wiki page at http://wiki.netbsd.org/ports/evbarm/qemu_arm -
> just a skeleton, add information as you see fit.  Information on how
> developers can edit the wiki is here: http://wiki.netbsd.org/wiki/start/

Thanks. I'll start filling up the blanks next week.

>> On QEMU 1.4.0, that kernel hangs in the same place, while an older one
>> boots fine, so it doesn't look like a problem with the emulation. I'm
>> going to build -current to take a look into this.
>>

The problem seems to be a typo in "sys/arch/arm/arm32/arm32_machdep.c":

--- a/src/sys/arch/arm/arm32/arm32_machdep.c    2013-06-13
03:02:04.000000000 +0000
+++ b/src/sys/arch/arm/arm32/arm32_machdep.c    2013-08-15
21:28:25.000000000 +0000
@@ -138,7 +138,7 @@
         */
 #ifndef ARM_HAS_VBAR
        if (va == ARM_VECTORS_LOW
-           && (armreg_pfr1_read() && ARM_PFR1_SEC_MASK) != 0) {
+           && (armreg_pfr1_read() & ARM_PFR1_SEC_MASK) != 0) {
 #endif
                extern const uint32_t page0rel[];
                vector_page = (vaddr_t)page0rel;

On a side note, the option VERBOSE_INIT_ARM is quite useful for this
kind of situations, and I see some configs have it enabled by default.
Would it make sense to do the same for INTEGRATOR_CP (which is mainly
used with QEMU) or is preferable to use it only when it's really
needed?

Sergio.