Re: [PATCH] hp100 -- fixes for new probing.

[Jeff Garzik <jgarzik@pobox.com>]

Stephen Hemminger wrote:
> Fixes to net-drivers-2.5-exp patches for hp100
> 	* EISA device id table needs a terminating string.
> 	* if one driver built for all variations (ISA, EISA, PCI)
> 	  then try to have sane error handling on probe.
> 
> diff -Nru a/drivers/net/hp100.c b/drivers/net/hp100.c
> --- a/drivers/net/hp100.c	Mon Nov 24 15:34:53 2003
> +++ b/drivers/net/hp100.c	Mon Nov 24 15:34:53 2003
> @@ -201,6 +201,7 @@
>  	{ "HWP1990" }, /* HP J2577 */
>  	{ "CPX0301" }, /* ReadyLink ENET100-VG4 */
>  	{ "CPX0401" }, /* FreedomLine 100/VG */
> +	{ "" }
>  };
>  MODULE_DEVICE_TABLE(eisa, hp100_eisa_tbl);
>  #endif
> @@ -3045,10 +3046,16 @@
>  	err = hp100_isa_init();
>  
>  #ifdef CONFIG_EISA
> -	err |= eisa_driver_register(&hp100_eisa_driver);
> +	if (err && err != -ENODEV) 
> +		return err;
> +
> +	err = eisa_driver_register(&hp100_eisa_driver);
>  #endif
>  #ifdef CONFIG_PCI
> -	err |= pci_module_init(&hp100_pci_driver);
> +	if (err && err != -ENODEV)
> +		return err;
> +
> +	err = pci_module_init(&hp100_pci_driver);
>  #endif
>  	return err;
>  }


Valid changes... but it looks like there should be some *_unregister_* 
calls in this last patch chunk, to clean up on error...

	Jeff