Fw: scsi_wait_scan Kconfig option

Fw: scsi_wait_scan Kconfig option

[Nick Warne]

[-- Attachment #1: Type: text/plain, Size: 1509 bytes --]


Hi all,

Please see below.

Nick

Begin forwarded message:

Date: Fri, 7 Dec 2007 19:12:42 +0000
From: Nick Warne <nick@ukfsn.org>
To: linux-kernel@vger.kernel.org
Subject: scsi_wait_scan Kconfig option



Hi all,

I am bringing this up again - primarily as I forgot about it after
patching my build tree ages ago:

http://lkml.org/lkml/2007/10/27/68

Today I built and installed git for the first time and cloned Linus'
tree (very trick!).

I try not to build a modular kernel, but only have modules ON due to
nVidia (sigh).  So I was semi-surprised when I saw the scsi_wait_scan
module being built again, yet NO WHERE in menuconfig is it present to
turn OFF.  Even if I hand edit .config, make puts it back again...


.config
# SCSI device support
#
# CONFIG_RAID_ATTRS is not set
CONFIG_SCSI=y
CONFIG_SCSI_DMA=y
# CONFIG_SCSI_TGT is not set
# CONFIG_SCSI_NETLINK is not set
# CONFIG_SCSI_PROC_FS is not set

#
# SCSI support type (disk, tape, CD-ROM)
#
CONFIG_BLK_DEV_SD=y
# CONFIG_CHR_DEV_ST is not set
# CONFIG_CHR_DEV_OSST is not set
# CONFIG_BLK_DEV_SR is not set
# CONFIG_CHR_DEV_SG is not set
# CONFIG_CHR_DEV_SCH is not set

#
# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
#
# CONFIG_SCSI_MULTI_LUN is not set
# CONFIG_SCSI_CONSTANTS is not set
# CONFIG_SCSI_LOGGING is not set
# CONFIG_SCSI_SCAN_ASYNC is not set
CONFIG_SCSI_WAIT_SCAN=m



I have attached my patch again :-)

Nick
-- 
Free Software Foundation Associate Member 5508

-- 
Free Software Foundation Associate Member 5508

[-- Attachment #2: scsi_wait.patch --]
[-- Type: text/x-patch, Size: 682 bytes --]

--- linux-current/drivers/scsi/Kconfig_old	2007-10-20 12:44:50.000000000 +0100
+++ linux-current/drivers/scsi/Kconfig	2007-10-20 12:57:13.000000000 +0100
@@ -248,10 +248,17 @@
 	  or async on the kernel's command line.
 
 config SCSI_WAIT_SCAN
-	tristate
+	tristate "Wait for SCSI scan completion"
 	default m
 	depends on SCSI
 	depends on MODULES
+	help
+	  The SCSI subsystem can probe for devices while the rest of the
+	  system continues booting, and even probe devices on different
+	  busses in parallel, leading to a significant speed-up.
+
+	  You can load the scsi_wait_scan module to ensure that all scans
+	  have completed.
 
 menu "SCSI Transports"
 	depends on SCSI

Re: Fw: scsi_wait_scan Kconfig option

[Matthew Wilcox]

On Fri, Dec 07, 2007 at 07:39:53PM +0000, Nick Warne wrote:
> I try not to build a modular kernel, but only have modules ON due to
> nVidia (sigh).  So I was semi-surprised when I saw the scsi_wait_scan
> module being built again, yet NO WHERE in menuconfig is it present to
> turn OFF.  Even if I hand edit .config, make puts it back again...

You have modules on ... which means you might decide to load a scsi
driver as a module.  Maybe one that isn't part of the source tree.
The scsi_wait_scan module is only 1500 bytes.  Apart from a sense of
ideological purity (odd in someone who chooses to use nVidia rather
than, say, nv or nouveau), this really isn't a problem.

-- 
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."

Re: scsi_wait_scan Kconfig option

[Nick Warne]

Hi Matthew,

On Fri, 7 Dec 2007 12:47:56 -0700
Matthew Wilcox <matthew@wil.cx> wrote:

> On Fri, Dec 07, 2007 at 07:39:53PM +0000, Nick Warne wrote:
> > I try not to build a modular kernel, but only have modules ON due to
> > nVidia (sigh).  So I was semi-surprised when I saw the
> > scsi_wait_scan module being built again, yet NO WHERE in menuconfig
> > is it present to turn OFF.  Even if I hand edit .config, make puts
> > it back again...
> 
> You have modules on ... which means you might decide to load a scsi
> driver as a module.  Maybe one that isn't part of the source tree.
> The scsi_wait_scan module is only 1500 bytes.  Apart from a sense of
> ideological purity (odd in someone who chooses to use nVidia rather
> than, say, nv or nouveau), this really isn't a problem.

Alas the card I have isn't supported, but that debate isn't for here.

OK, I only have SCSI set for USB storage etc.  I find it strange that
this one and only option in the whole of Kconfig builds a module, and
yet, even though I know I will never need it, I can't stop it.

Surely if people build their own kernels, they should know (and) be
allowed to configure such options?

Nick 
-- 
Free Software Foundation Associate Member 5508

Re: scsi_wait_scan Kconfig option

[Nick Warne]

On Sat, 08 Dec 2007 14:11:44 +0100
Clemens Koller <clemens.koller@anagramm.de> wrote:

> Nick Warne schrieb:
> > I am bringing this up again - primarily as I forgot about it after
> > patching my build tree ages ago:
> > 
> > http://lkml.org/lkml/2007/10/27/68

Subject: Re: Fw: scsi_wait_scan Kconfig option
Date: Fri, 7 Dec 2007 12:47:56 -0700

On Fri, Dec 07, 2007 at 07:39:53PM +0000, Nick Warne wrote:
> I try not to build a modular kernel, but only have modules ON due to
> nVidia (sigh).  So I was semi-surprised when I saw the scsi_wait_scan
> module being built again, yet NO WHERE in menuconfig is it present to
> turn OFF.  Even if I hand edit .config, make puts it back again...  

On Fri, 7 Dec 2007 12:47:56 -0700
Matthew Wilcox <matthew@wil.cx> wrote:

> You have modules on ... which means you might decide to load a scsi
> driver as a module.  Maybe one that isn't part of the source tree.
> The scsi_wait_scan module is only 1500 bytes.  Apart from a sense of
> ideological purity (odd in someone who chooses to use nVidia rather
> than, say, nv or nouveau), this really isn't a problem.
> 
> Please see the patch I sent some days ago, which does the very
> same thing: http://marc.info/?l=linux-kernel&m=119677244318528&w=2
> 
> True...

Well, that's two of us would that like to be able to stop it
building :-)

Nick
-- 
Free Software Foundation Associate Member 5508

Re: scsi_wait_scan Kconfig option

[Stefan Richter]

> Clemens Koller <clemens.koller@anagramm.de> wrote:
>> Nick Warne schrieb:
>>> I am bringing this up again - primarily as I forgot about it after
>>> patching my build tree ages ago:
>>>
>>> http://lkml.org/lkml/2007/10/27/68

>> Please see the patch I sent some days ago, which does the very
>> same thing: http://marc.info/?l=linux-kernel&m=119677244318528&w=2

Besides, in order for scsi_wait_scan to work, drivers need to support it
explicitly, don't they?  If so, simply kill the "default m" from config
SCSI_WAIT_SCAN and let any driver which is integrated with it select it.

(Not that I'm a friend of select, but here is a case where it won't hurt
too much.)
-- 
Stefan Richter
-=====-=-=== ==-- -=---
http://arcgraph.de/sr/

Re: scsi_wait_scan Kconfig option

[Matthew Wilcox]

On Sat, Dec 08, 2007 at 10:21:11PM +0100, Stefan Richter wrote:
> Besides, in order for scsi_wait_scan to work, drivers need to support it
> explicitly, don't they?  If so, simply kill the "default m" from config
> SCSI_WAIT_SCAN and let any driver which is integrated with it select it.

No.  All drivers which call scsi_scan_host() support it.

-- 
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."