Re: [PATCH] SCSI: Let users disable SCSI_WAIT_SCAN to be built

[Stefan Richter <stefanr@s5r6.in-berlin.de>]

I wrote:
> --- linux-2.6.22-rc1.orig/drivers/scsi/Kconfig
> +++ linux-2.6.22-rc1/drivers/scsi/Kconfig
> @@ -241,11 +241,19 @@ config SCSI_SCAN_ASYNC
>  	  You can override this choice by specifying "scsi_mod.scan=sync"
>  	  or async on the kernel's command line.
>  
> -config SCSI_WAIT_SCAN
> +config SCSI_WAIT_SCAN_NO_Y
>  	tristate
>  	default m
> -	depends on SCSI
> -	depends on MODULES

SCSI_WAIT_SCAN_NO_Y is unnecessary, as Randy Dunlap pointed out in "Re:
How to force Kconfig tristate into range n..m?".
http://lkml.org/lkml/2007/5/15/320

> +
> +config SCSI_WAIT_SCAN
> +	tristate "Pseudo driver which waits for SCSI scanning to finish"
> +	depends on MODULES && SCSI && SCSI_WAIT_SCAN_NO_Y

	depends on SCSI && m

> +	help
> +	  When loaded, this module will do nothing else than wait for
> +	  SCSI low-level drivers to finish asynchronous scanning.
> +	  The module will be called scsi_wait_scan.

There should be explained that it is the command "modprobe
scsi_wait_scan" which is doing the waiting, and that this is useful or
required in initrds.

> +	  Most people can say n here.

This sentence should probably be omitted, as it may be wrong in the
future and unsafe already now.

I will resend an updated patch.
-- 
Stefan Richter
-=====-=-=== -=-= =----
http://arcgraph.de/sr/