Re: [Openzaurus-devel] Zaurus Kernel 2.6.5

[John L. <je...@st...>]

On 2004.04.21 15:57, Dirk Opfer wrote:
> Dear John,
> 
> please find atached my patch to get the Kernel running on the collie.
> What is your current status?
> 
>

ANNOUNCE: patch-2.6.5-jl2.diff.gz is available from
http://www.cs.wisc.edu/~lenz/zaurus

Ok.  After enabling the floating point stuff, I was able to mount a  
root filesystem using the CFI driver and init booted and I was able to  
log in and run some commands in busybox.  Except when I tried to write  
a file, and then it spawned some error messages.  So it seems that CFI  
driver is close to working (just has some problems with erase blocks).

So I then went to the pre-CFI sharp driver that you mailed to me, and  
that seems to be working fine.  Note: I merged the changes into the  
drivers/mtd/chips/sharp.c file.  There are a few bugs that I noticed in  
collie_sharp.c (most notably, using jiffies < timeo instead of  
time_after and not freeing allocated memory correctly in the case of an  
error).

I am going to start a discussion on the mtd mailing list about this,  
since the KConfig file says that eventually CFI should support  
everything.  If you notice, the 0x8080 come back only for the bytes  
past the end of the last eraseblock (the ones added since we pass a -- 
pad option to mkfs.jffs2).

In any case, I was able to boot into an initrd.bin built by oe and log  
in using the keyboard and framebuffer.

So I have updated my patch to 2.6.5-jl2, which just includes these few  
MTD changes. (Brad, I haven't looked at the shepherd stuff you emailed  
me yet, I will look at that soon tho)

I haven't tried 2.6.6-rc2 yet either :)

John