amd(4) module

[Scott Long]

Ruslan Ermilov wrote:
> The amd(4) module is pretty unusable at the moment.  The attached
> patch fixed it for me.  Should I commit it?

The patch looks fine.

> 
> BTW, the "camcontrol rescan all" only works if amd.ko is loaded
> from the loader(8); loading it by kldload(8) causes the camcontrol
> to stuck.  Is this a known issue?

How does it get stuck?  A backtrace would help.  However, I'm working on
porting the NetBSD esp(4) driver, which will replace amd(4) and also add
support for many other variants of the hardware.

Scott

> 
> 
> Cheers,
> 
> 
> ------------------------------------------------------------------------
> 
> Index: amd.c
> ===================================================================
> RCS file: /home/ncvs/src/sys/dev/amd/amd.c,v
> retrieving revision 1.21
> diff -u -r1.21 amd.c
> --- amd.c	1 Jul 2003 15:51:53 -0000	1.21
> +++ amd.c	31 Jul 2003 15:45:51 -0000
> @@ -2497,3 +2497,4 @@
>  
>  static devclass_t amd_devclass;
>  DRIVER_MODULE(amd, pci, amd_driver, amd_devclass, 0, 0);
> +MODULE_DEPEND(amd, cam, 1, 1, 1);