oops in i2c_keywest_init

Olaf Hering olh at suse.de
Sun May 22 19:24:28 EST 2005


 On Sat, May 21, Benjamin Herrenschmidt wrote:

> Ok, it looks like the device-tree is bogus on these and lacks interrupt
> informations. That plus the i2c-keywest driver doesn't properly test for
> the presence of an interrupt in the device node and crashes if there is
> none. Can you test this patch ? I haven't even tried building so it may
> need some fixup. It tries to "fixup" the device-tree at boot, and also
> adds some guard to i2c-keywest. Let me know if it fixes booting and if
> i2c works propertly.

It allows booting with 'nosmp' at least. lsprop and dmesg sent in
private mail.



More information about the Linuxppc64-dev mailing list