Hi Richard,
Thanks for pointing these out.
Is the linker error message that Jeff first
pointed out bogus w.r.t. calc_bus_time()?
Do these have nothing to do with each other?
Or are they closely related but the text offset
is just confusing?
> > drivers/usb/usbdrv.o: In function `calc_bus_time':
> > usb.c(.text.init+0x4a4): relocation truncated to fit:
> BRADDR text.exit
> > make: *** [vmlinux] Error 1
Georg, I'll leave it up to you whether you think this
needs to be fixed. It's wrong, but does it matter?
I don't know why an Alpha couldn't be built with a UHCI
controller (like one from VIA), but I guess it's not
likely.
~Randy
> alloc_uhci (in __init) calls uhci_cleanup_dev (in __exit).
>
> This is a no-no.
>
> The theory of __exit is that it is only used for module unloading,
> and as such the kernel proper is allowed to discard that data at
> link time. Trying to make a call to a function that has been
> discarded
> is what produces the admittedly somewhat less than informative error.
>
> You don't see the linker error on x86 because their call instruction
> is capable of spanning the entire 32-bit address space, and so a call
> to NULL is representable. ;-)
>
> r~
>
> PS. Jeff, for future reference, UHCI is for Intel core logics only.
> OHCI is what would be used by the chips on alphas.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
