[parisc-linux] Some 2.5-testing

[Rüdiger Scholz]

John David Anglin schrieb:

>Until issues related to the canonicalization of function pointers
>and the alignment of stack locals are resolved, it is best to stick
>with debian gcc-3.0.4 for kernel building.
>
So I installed gcc-3.04 from Debian, did "make distclean", "make 
oldconfig" (selecting "PA7100LC") and booted...
And got the same loop again ;)

---------- SNIP---------------

     md: linear personality registered as nr 1
    Badness in local_bh_enable at at kernel/softirq.c:109
    Kernel addresses on the stack:

     [tirq.c:109
    <4>Kernel addresses on the stack:
    <4> [<10124edcc8>]  [<1010ea4c>]
    <4> [<1012503c>]  [<1028ac08>]  [<10124cc>]
    <4> [<1029a2a4>]  [<103ad4d4>]  [<10100284>]  [<10109c5ab8>] 
    [<101289a8>]  [<1028ac20>]
     [<1028b428>]  [<101070f28ac08>]  [<10124f10>]  [<1028adcc>]
     [<1029a2a4>]  [<103adc>]
    <4> [<1029a2a4>]  [<103ad4d4>]  [<10100284>]  [<10109c>] 
    [<103ad4d4>]  [<10100284>]  [<10109c5c>]


---------- SNAP---------------

Complete Bootlog and System.map is here: rscholz.bluehash.de/parisc .
Can I do something else to help finding the bug?

>  Loops of this kind can
>happen when lock structures are misaligned on PA 1.1 machines.
>
>Dave
>  
>

Rüdiger