[parisc-linux] jfs-1.0.23 as well xfs-cvs break fb console on k-2.4.19-pa22

[jsoe0708@tiscali.be]

Oooh,...

>
>Joel Soete wrote:
>> The very first patch 2.4.19 I still have was pa5 and the same problem
occur.
>> It seems definitely that JFS and STI FB in 2.4.19 are incompatible.
>> Where do I have to start my investigation?
>...
>> Really no idea?
>
>Sorry. I don't. And until 2.4.19 boots on the A500-6X,
>I'm not interested in chasing the JFS/STI issue.
>
>My advice is to start adding printk's in the main vfs_ops entry
>points to JFS until you can narrow down which entry point is
>having problems. Then start looking STI to see if anything
>in the JFS code uses the same services or something.
>

it sounds like a gcc bug:

simply adding "printk("VC: %s: %d\n", __FUNCTION__, __LINE__);" some where
in the middle of take_over_console (to check the way) and the problem is
solved???

I met a similar problem in a evms tools where there was two solutions:
a. changing -O2 to -O0 or -O1
b. using gcc-3.2

But here gcc-3.2 produce well a bootable kernel but afaik there is still
a network connection problem (I try to hack this but the only think I reach
to find is that the last called function was show_stack and that was riht
:-))) )

Hmm I currious to try -O0 on the only console part (if possible)?

Regards,
    Joel