In the last mail Timothy Baldwin said:
> In message <[EMAIL PROTECTED]>
> Russell King - ARM Linux Admin <[EMAIL PROTECTED]> wrote:
> You believe correctly. Further investigation shows that IscaFS was
> incorrectly reading files written by binutils (but not cp). It seems
> my confidence in IscaFS was misplaced, I will be submitting a bug
> report.
Aaargh. I've got (at least) 4 problems, 2 of which are now solved
1: IscaFS didn't understand holey files (binutils makes a zImage with holes
(parts of file with no disc storage allocated - read as all zeros))
Phil has fixed this (we think, but)
2: IscaFS still doesn't read the same data for zImage and the result of
cat zImage; (to remove the holes) - got a problem starting at block 400 ish
(206th double indirect block, IIRC)
3: egcs was generating LDRH instructions, because it was told I have a
StrongARM (True, but counterproductive on a RiscPC_)
4: something else is making buggy code, hence my kernel won't boot
> I recitified this by copying the kernel to RISC OS via a MS-DOS formmated
> floppy, however I now get a crc error when trying to boot a compressed
> kernel.
RISC OS seems to be pretty poor when it comes to reading MS-DOG disks.
cat zImage >someotherfile
should do the trick.
Or gzip or zip the zImage, and "uncompress" it on this side of IscaFS, so that
you get the CRC check to verify that it came across intact.
> BTW, the machine in question is 20M DRAM + 1M VRAM Risc PC with RISC OS 3.7
> and an ARM610 processor.
"Very interesting" (why a 610?)
Nick
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
