[PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[Luis de Bethencourt]

These drivers only have runtime but no build time dependencies, so they can
be built for testing purposes if the Kconfig COMPILE_TEST option is enabled.

This is useful to have more build coverage and make sure that drivers are
not affected by changes that could cause build regressions.

Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
---
 drivers/scsi/Kconfig | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
index d2f480b..0d8ee16 100644
--- a/drivers/scsi/Kconfig
+++ b/drivers/scsi/Kconfig
@@ -416,7 +416,7 @@ config SCSI_ACARD
 
 config SCSI_AHA152X
 	tristate "Adaptec AHA152X/2825 support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	select SCSI_SPI_ATTRS
 	select CHECK_SIGNATURE
 	---help---
@@ -509,7 +509,7 @@ config SCSI_ADVANSYS
 
 config SCSI_IN2000
 	tristate "Always IN2000 SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	help
 	  This is support for an ISA bus SCSI host adapter.  You'll find more
 	  information in <file:Documentation/scsi/in2000.txt>. If it doesn't work
@@ -657,7 +657,7 @@ config SCSI_DMX3191D
 
 config SCSI_DTC3280
 	tristate "DTC3180/3280 SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	select SCSI_SPI_ATTRS
 	select CHECK_SIGNATURE
 	help
@@ -776,7 +776,7 @@ config SCSI_ISCI
 
 config SCSI_GENERIC_NCR5380
 	tristate "Generic NCR5380/53c400 SCSI PIO support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	select SCSI_SPI_ATTRS
 	---help---
 	  This is a driver for the old NCR 53c80 series of SCSI controllers
@@ -796,7 +796,7 @@ config SCSI_GENERIC_NCR5380
 
 config SCSI_GENERIC_NCR5380_MMIO
 	tristate "Generic NCR5380/53c400 SCSI MMIO support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	select SCSI_SPI_ATTRS
 	---help---
 	  This is a driver for the old NCR 53c80 series of SCSI controllers
@@ -970,7 +970,7 @@ config SCSI_IZIP_SLOW_CTR
 
 config SCSI_NCR53C406A
 	tristate "NCR53c406a SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	help
 	  This is support for the NCR53c406a SCSI host adapter.  For user
 	  configurable parameters, check out <file:drivers/scsi/NCR53c406a.c>
@@ -1238,7 +1238,7 @@ config SCSI_NCR53C8XX_NO_DISCONNECT
 
 config SCSI_PAS16
 	tristate "PAS16 SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE) && SCSI
 	select SCSI_SPI_ATTRS
 	---help---
 	  This is support for a SCSI host adapter.  It is explained in section
@@ -1252,7 +1252,7 @@ config SCSI_PAS16
 
 config SCSI_QLOGIC_FAS
 	tristate "Qlogic FAS SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	---help---
 	  This is a driver for the ISA, VLB, and PCMCIA versions of the Qlogic
 	  FastSCSI! cards as well as any other card based on the FASXX chip
@@ -1321,7 +1321,7 @@ config SCSI_SIM710
 
 config SCSI_SYM53C416
 	tristate "Symbios 53c416 SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	---help---
 	  This is support for the sym53c416 SCSI host adapter, the SCSI
 	  adapter that comes with some HP scanners. This driver requires that
@@ -1372,7 +1372,7 @@ config SCSI_AM53C974
 
 config SCSI_T128
 	tristate "Trantor T128/T128F/T228 SCSI support"
-	depends on ISA && SCSI
+	depends on (ISA || COMPILE_TEST) && SCSI
 	select SCSI_SPI_ATTRS
 	select CHECK_SIGNATURE
 	---help---
-- 
2.5.1

Re: [PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[Randy Dunlap]

On 10/14/15 14:13, Luis de Bethencourt wrote:
> These drivers only have runtime but no build time dependencies, so they can
> be built for testing purposes if the Kconfig COMPILE_TEST option is enabled.
> 
> This is useful to have more build coverage and make sure that drivers are
> not affected by changes that could cause build regressions.
> 
> Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
> ---
>  drivers/scsi/Kconfig | 20 ++++++++++----------
>  1 file changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
> index d2f480b..0d8ee16 100644
> --- a/drivers/scsi/Kconfig
> +++ b/drivers/scsi/Kconfig

> @@ -1238,7 +1238,7 @@ config SCSI_NCR53C8XX_NO_DISCONNECT
>  
>  config SCSI_PAS16
>  	tristate "PAS16 SCSI support"
> -	depends on ISA && SCSI
> +	depends on (ISA || COMPILE) && SCSI

COMPILE_TEST

>  	select SCSI_SPI_ATTRS
>  	---help---
>  	  This is support for a SCSI host adapter.  It is explained in section


-- 
~Randy

Re: [PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[James Bottomley]

On Wed, 2015-10-14 at 22:13 +0100, Luis de Bethencourt wrote:
> These drivers only have runtime but no build time dependencies, so they can
> be built for testing purposes if the Kconfig COMPILE_TEST option is enabled.
> 
> This is useful to have more build coverage and make sure that drivers are
> not affected by changes that could cause build regressions.

I don't think I understand the purpose.  If the object is to build a
load of old ISA drivers, why not just turn on CONFIG_ISA in the test
build? That's how I build test SCSI on my systems.

James

Re: [PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[Luis de Bethencourt]

On 14/10/15 23:42, James Bottomley wrote:
> On Wed, 2015-10-14 at 22:13 +0100, Luis de Bethencourt wrote:
>> These drivers only have runtime but no build time dependencies, so they can
>> be built for testing purposes if the Kconfig COMPILE_TEST option is enabled.
>>
>> This is useful to have more build coverage and make sure that drivers are
>> not affected by changes that could cause build regressions.
> 
> I don't think I understand the purpose.  If the object is to build a
> load of old ISA drivers, why not just turn on CONFIG_ISA in the test
> build? That's how I build test SCSI on my systems.
> 
> James
> 
> 

Hi James,

The idea is that the more drivers are built with make allyesconfig, the better.

I added COMPILE_TEST to all the drivers that can be built without having ISA,
since they depend on runtime and not in build time. I understand your point,
which raises the question of why isn't CONFIG_ISA set to Y in make allyesconfig.

Would this be possible?

Thanks for the review,
Luis

Re: [PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[James Bottomley]

On Thu, 2015-10-15 at 00:30 +0100, Luis de Bethencourt wrote:
> On 14/10/15 23:42, James Bottomley wrote:
> > On Wed, 2015-10-14 at 22:13 +0100, Luis de Bethencourt wrote:
> >> These drivers only have runtime but no build time dependencies, so they can
> >> be built for testing purposes if the Kconfig COMPILE_TEST option is enabled.
> >>
> >> This is useful to have more build coverage and make sure that drivers are
> >> not affected by changes that could cause build regressions.
> > 
> > I don't think I understand the purpose.  If the object is to build a
> > load of old ISA drivers, why not just turn on CONFIG_ISA in the test
> > build? That's how I build test SCSI on my systems.
> > 
> > James
> > 
> > 
> 
> Hi James,
> 
> The idea is that the more drivers are built with make allyesconfig, the better.
> 
> I added COMPILE_TEST to all the drivers that can be built without having ISA,
> since they depend on runtime and not in build time. I understand your point,
> which raises the question of why isn't CONFIG_ISA set to Y in make allyesconfig.
> 
> Would this be possible?

It happens today, but it depends on architecture.  An allyesconfig on
ia32 should have it set and possibly a few others.  There's no reason an
x86-64 couldn't have it set, it's just that x86-64 was the architecture
break where ISA was discontinued, so there's no real point building that
configuration.

James

Re: [PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[kbuild test robot]

Hi Luis,

[auto build test WARNING on scsi/for-next -- if it's inappropriate base, please suggest rules for selecting the more suitable base]

url:    https://github.com/0day-ci/linux/commits/Luis-de-Bethencourt/scsi-kconfig-When-possible-compile-drivers-with-COMPILE_TEST/20151015-051558
reproduce:
        # apt-get install sparse
        make ARCH=x86_64 allmodconfig
        make C=1 CF=-D__CHECK_ENDIAN__


sparse warnings: (new ones prefixed by >>)

>> drivers/scsi/NCR5380.c:1751:9: sparse: context imbalance in 'NCR5380_transfer_dma' - unexpected unlock

vim +/NCR5380_transfer_dma +1751 drivers/scsi/NCR5380.c

^1da177e Linus Torvalds 2005-04-16  1735  	NCR5380_setup(instance);
^1da177e Linus Torvalds 2005-04-16  1736  
^1da177e Linus Torvalds 2005-04-16  1737  	if ((tmp = (NCR5380_read(STATUS_REG) & PHASE_MASK)) != p) {
^1da177e Linus Torvalds 2005-04-16  1738  		*phase = tmp;
^1da177e Linus Torvalds 2005-04-16  1739  		return -1;
^1da177e Linus Torvalds 2005-04-16  1740  	}
^1da177e Linus Torvalds 2005-04-16  1741  #if defined(REAL_DMA) || defined(REAL_DMA_POLL)
^1da177e Linus Torvalds 2005-04-16  1742  #ifdef READ_OVERRUNS
^1da177e Linus Torvalds 2005-04-16  1743  	if (p & SR_IO) {
^1da177e Linus Torvalds 2005-04-16  1744  		c -= 2;
^1da177e Linus Torvalds 2005-04-16  1745  	}
^1da177e Linus Torvalds 2005-04-16  1746  #endif
52a6a1cb Finn Thain     2014-03-18  1747  	dprintk(NDEBUG_DMA, "scsi%d : initializing DMA channel %d for %s, %d bytes %s %0x\n", instance->host_no, instance->dma_channel, (p & SR_IO) ? "reading" : "writing", c, (p & SR_IO) ? "to" : "from", (unsigned) d);
^1da177e Linus Torvalds 2005-04-16  1748  	hostdata->dma_len = (p & SR_IO) ? NCR5380_dma_read_setup(instance, d, c) : NCR5380_dma_write_setup(instance, d, c);
^1da177e Linus Torvalds 2005-04-16  1749  #endif
^1da177e Linus Torvalds 2005-04-16  1750  
^1da177e Linus Torvalds 2005-04-16 @1751  	NCR5380_write(TARGET_COMMAND_REG, PHASE_SR_TO_TCR(p));
^1da177e Linus Torvalds 2005-04-16  1752  
^1da177e Linus Torvalds 2005-04-16  1753  #ifdef REAL_DMA
^1da177e Linus Torvalds 2005-04-16  1754  	NCR5380_write(MODE_REG, MR_BASE | MR_DMA_MODE | MR_ENABLE_EOP_INTR | MR_MONITOR_BSY);
^1da177e Linus Torvalds 2005-04-16  1755  #elif defined(REAL_DMA_POLL)
^1da177e Linus Torvalds 2005-04-16  1756  	NCR5380_write(MODE_REG, MR_BASE | MR_DMA_MODE);
^1da177e Linus Torvalds 2005-04-16  1757  #else
^1da177e Linus Torvalds 2005-04-16  1758  	/*
^1da177e Linus Torvalds 2005-04-16  1759  	 * Note : on my sample board, watch-dog timeouts occurred when interrupts

:::::: The code at line 1751 was first introduced by commit
:::::: 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 Linux-2.6.12-rc2

:::::: TO: Linus Torvalds <torvalds@ppc970.osdl.org>
:::::: CC: Linus Torvalds <torvalds@ppc970.osdl.org>

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation

Re: [PATCH] scsi: kconfig: When possible, compile drivers with COMPILE_TEST

[Luis de Bethencourt]

On 15/10/15 00:39, James Bottomley wrote:
> On Thu, 2015-10-15 at 00:30 +0100, Luis de Bethencourt wrote:
>> On 14/10/15 23:42, James Bottomley wrote:
>>> On Wed, 2015-10-14 at 22:13 +0100, Luis de Bethencourt wrote:
>>>> These drivers only have runtime but no build time dependencies, so they can
>>>> be built for testing purposes if the Kconfig COMPILE_TEST option is enabled.
>>>>
>>>> This is useful to have more build coverage and make sure that drivers are
>>>> not affected by changes that could cause build regressions.
>>>
>>> I don't think I understand the purpose.  If the object is to build a
>>> load of old ISA drivers, why not just turn on CONFIG_ISA in the test
>>> build? That's how I build test SCSI on my systems.
>>>
>>> James
>>>
>>>
>>
>> Hi James,
>>
>> The idea is that the more drivers are built with make allyesconfig, the better.
>>
>> I added COMPILE_TEST to all the drivers that can be built without having ISA,
>> since they depend on runtime and not in build time. I understand your point,
>> which raises the question of why isn't CONFIG_ISA set to Y in make allyesconfig.
>>
>> Would this be possible?
> 
> It happens today, but it depends on architecture.  An allyesconfig on
> ia32 should have it set and possibly a few others.  There's no reason an
> x86-64 couldn't have it set, it's just that x86-64 was the architecture
> break where ISA was discontinued, so there's no real point building that
> configuration.
> 
> James
> 
> 

Hi James,

That makes sense.

Thanks for the information, I understand this a lot better now.
Luis