Re: revival of NetBSD/sandpoint

In article <000301c83c70$7deb8590$1718a8c0@tpad3>,
Toru Nishimura <[email protected]> wrote:
>NetBSD/sandpoint has revived.  It's up and running on AmPro
>EnCorePP1 board in mult-user mode.  It was found that the source
>of whole issue was i8259 interrupt handling.  Tim Rightnour made
>a fix around there and brought refreshing improvement for this port.
>Without the fix the initial inquery of /bin/sh path can not take any
>console input.
>
>http://mail-index.netbsd.org/source-changes/2007/12/11/0070.html
>
>This port needs more work to iron out some stability issue during
>heavy load, however, overall efforts should be done easier since
>standard working NetBSD environment is now available for deep
>kernel debugging.
>
>Due to the very low profile of HW popularity the significance of
>this port is limited.  During this development efforts, we made
>LIBSA ELF bootloader by NFS.  It's known useful combined with
>Moto DINK32 and PPCboot (popular U-Boot predecessor) to
>make possible a solid and comfortable development cycle.  The
>byproduct should be a good vihecle to bring up any NetBSD
>porting.  It can become a self-standing complete bootloader
>when IDE block device driver is added.
>
>This revival was made possible by innovative work by Micrel Lorenz
>and Tim Rightnour.  The "extensible PIC framework," which is
>found in powerpc/pic/ directory, is so versatile and has very large
>significance.  Using this framework powerpc can be easily
>adapted for new interrupt controller schemes.  Examples out of my
>head are;
>- Freescale MPC8349/8315 family
>the interrupt controller, now called IPIC, has a different design from
>original OpenPIC standard.
>- Tundra Semiconductor Tsi109/110 
>descendents of Moto MPC107/108 system controller and have
>OpenPIC-like extended interrupt design.
>
>powerpc/pic totally eliminated agonizing extintr.c adaption.  All to do
>is to make and add new pic_xxx.c in powerpc/pic/ directory.
>

Very nicely done!

christos