svn commit: r326377 - in head/sys: amd64/conf arm/conf i386/conf mips/conf modules/aic7xxx/ahc modules/aic7xxx/ahd powerpc/conf sparc64/conf

[Scott Long]

> On Nov 29, 2017, at 6:19 PM, Bruce Evans <brde at optusnet.com.au> wrote:
> 
> On Wed, 29 Nov 2017, Scott Long wrote:
> 
>> Log:
>> It's time to retire AHC_REG_PRETTY_PRINT and AHD_REG_PRETTY_PRINT from
>> the standard kernels.  They are still available as custom compile
>> options.
> 
> MI NOTES still gives MI sizes for these options.  It seems unlikely
> that these sizes are still correct even for 1 arch where they were measured
> on.
> 

Today, I measured a savings of ~130KiB when both options were removed.  I
also did the original size measurements back in 2000-whatever.  I’m guessing
that the difference is due to improvements in compiler technology.

>> Modified: head/sys/amd64/conf/GENERIC
>> ==============================================================================
>> --- head/sys/amd64/conf/GENERIC	Wed Nov 29 23:28:40 2017	(r326376)
>> +++ head/sys/amd64/conf/GENERIC	Wed Nov 29 23:41:49 2017	(r326377)
>> @@ -123,11 +123,7 @@ device		siis			# SiliconImage SiI3124/SiI3132/SiI3531
>> 
>> # SCSI Controllers
>> device		ahc			# AHA2940 and onboard AIC7xxx devices
>> -options 	AHC_REG_PRETTY_PRINT	# Print register bitfields in debug
>> -					# output.  Adds ~128k to driver.
>> device		ahd			# AHA39320/29320 and onboard AIC79xx devices
>> -options 	AHD_REG_PRETTY_PRINT	# Print register bitfields in debug
>> -					# output.  Adds ~215k to driver.
> 
> k is a bogus unit for memory sizes.  NOTES has the same bug.
> 
>> Modified: head/sys/i386/conf/GENERIC
>> ==============================================================================
>> --- head/sys/i386/conf/GENERIC	Wed Nov 29 23:28:40 2017	(r326376)
>> +++ head/sys/i386/conf/GENERIC	Wed Nov 29 23:41:49 2017	(r326377)
>> @@ -120,11 +120,6 @@ device		siis			# SiliconImage SiI3124/SiI3132/SiI3531
>> 
>> # SCSI Controllers
>> device		ahc			# AHA2940 and onboard AIC7xxx devices
>> -options 	AHC_REG_PRETTY_PRINT	# Print register bitfields in debug
>> -					# output.  Adds ~128k to driver.
>> -device		ahd			# AHA39320/29320 and onboard AIC79xx devices
>> -options 	AHD_REG_PRETTY_PRINT	# Print register bitfields in debug
>> -					# output.  Adds ~215k to driver.
>> device		esp			# AMD Am53C974 (Tekram DC-390(T))
>> device		hptiop			# Highpoint RocketRaid 3xxx series
>> device		isp			# Qlogic family
>> 
> 
> This also removes ahd.  This bug is only new on i386/GENERIC.

Oops!

>  Other arches
> except amd64 never had ahd.  arm/conf/CRB was most broken -- it had 2 PRETTY
> options but neither ahc or ahd.  Perhaps these old drivers don't belong in
> any GENERIC.

There’s a good argument to be made for doing a house-cleaning of GENERIC
between now and when FreeBSD 12.0 is released.

Scott