On Feb 8, 2008 7:53 AM, ron minnich rminnich@gmail.com wrote:
On Feb 8, 2008 3:51 AM, Carl-Daniel Hailfinger
c-d.hailfinger.devel.2006@gmx.net wrote:
On 08.02.2008 06:54, ron minnich wrote:
This patch (which is NOT signed off) adds (or tries to) PIRQ support for the alix1c.
it crashes and burns badly, however, I don' t know why.
You found the problem with LAR parsing of ELF. Any reason why this can't affect stage2 as well?
I am pretty sure it is. I am thinking on stage 2, however, to add a manual step in main memset(bss, 0, end-bss);
ok, no need: the .bss thing is no longer an issue. Here is LAR: normal/stage2/segment0 (191796 bytes, lzma compressed to 110 bytes @0x9500);loadaddress 0x0xb8a0 entry 0x0x2000 normal/stage2/segment1 (32284 bytes, lzma compressed to 17073 bytes @0x95c0);loadaddress 0x0x2000 entry 0x0x2000 normal/stage2/segment2 (6300 bytes, lzma compressed to 431 bytes @0xd8d0);loadaddress 0x0xa000 entry 0x0x2000
Note that stage2 segment 0 is bss! The LAR fix is buying us quite a lot. (that said, decompressing zeros may be a very expensive way to zero-fill a region ...)
This leaves me more puzzled, will try more tonight.
ron