[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: panic: pmap_map_chunk: no L2 table for VA 0xc0400000
Chris Gilbert wrote:
Rene: I routinely build using Fedora 8, so that is most likely not the
Chris Gilbert wrote:
"Can you enable VERBOSE_INIT_ARM by adding an options VERBOSE_INIT_ARM to
your kernel config. It'll give more tracing during bootup"
To be honest I don't know how to do That
Chris: can you give me a short explanation on how to do that if building
is possible on "Fedora 8" of course.
uncomment the line:
from the kernel config file (NSLU2), and rebuild (you must do a clean
build, otherwise the option won't be picked up) the kernel.
In theory building on fedora core should be fine, we support cross
I suspect that the kernel has sneaked over the 4MB size boundary, and
that's what's trigger the problem (I saw this on acorn32 a while back)
However I can't see how that's kicking in as the code looks to (but I
could be mis-reading it) setup 16MB of space for the kernel.
Having just built a fresh NSLU2 kernel, it's below the 4MB limit, so I
am puzzled now:
text data bss dec hex filename
2283971 479328 217924 2981223 2d7d67
I think VERBOSE_INIT_ARM is probably worth a try.
You probably should keep port-arm in the loop, as there are lurkers out
there who can help :)
I rebuilt from scratch and had the same problem as Rene. The last
version I built was 4.99.60, which worked ok (This is version 4.99.62).
Then rebuilt with the VERBOSE_INIT_ARM option turned on. Here is the
output I get with that option:
NetBSD/evbarm (Linksys NSLU2) booting ...
physmemory: 8192 pages at 0x10000000 -> 0x11ffffff
Allocating page tables
physical_start = 0x10000000, physical_end = 0x12000000
IRQ stack: p0x1000f000 v0x8000f000
ABT stack: p0x10010000 v0x80010000
UND stack: p0x10011000 v0x80011000
SVC stack: p0x10012000 v0x80012000
Creating L1 page table at 0x10000000
pmap_map_chunk: pa=0x10200000 va=0x80200000 size=0x4022f000
Let me know if there is more you want to try.
Main Index |
Thread Index |