PATCH & call for help: Marking ISA only drivers

PATCH & call for help: Marking ISA only drivers

[Andi Kleen]

Hallo,

This patch tries to make most ISA only drivers dependent on CONFIG_ISA. 

The motivation is that it is a lot of work to get old drivers to compile
meaningfully (at least without warnings, not even testing them) on x86-64
and a lot of them are obviously not 64bit safe.  As it is very unlikely
that x86-64 boxes will ever have ISA slots[1] one simple way for that
is just removing the old ISA drivers from the configuration.

BTW I think CONFIG_ISA would be an useful configuration option for 
i386 too - at least most modern PCs do not come with ISA slots anymore.

Back to the patch: 

I left drivers/net alone because it looked like it already used CONFIG_ISA
properly.  I went over drivers/scsi and some other drivers directories
and just made those that looked like ISA only dependent on CONFIG_ISA.
Note that I usually don't know much about these devices, I just took
a quick look at the source of the driver. Of course I may have made mistakes
during that, so if you see any driver wrongly marked please tell me.

You can test that on i386 by applying the patch, change the
"define_bool CONFIG_ISA y" in arch/i386/config to n  and run your
favourite configuration tool and look for any ISA only drivers still appearing
as configuration choices.

I left out drivers/media/video and drivers/input/joystick - if there are any 
ISA only devices in there please tell me.

Patch for 2.5.14

Thanks,
-Andi

[1] If someone later will be ever crazy enough to buy an x86-64 machine
and put an ISA bridge into it they'll have to pay for that and fix 
all the ISA drivers they need ;)

diff -X ../../KDIFX -burp linux-vanilla/arch/x86_64/config.in linux/arch/x86_64/config.in
--- linux-vanilla/arch/x86_64/config.in	Mon May  6 13:11:56 2002
+++ linux/arch/x86_64/config.in	Thu May  9 19:41:56 2002
@@ -2,12 +2,15 @@
 # For a description of the syntax of this configuration file,
 # see Documentation/kbuild/config-language.txt.
 #
+# Note: ISA is disabled and will hopefully never be enabled.
+#
+
 mainmenu_name "Linux Kernel Configuration"
 
 define_bool CONFIG_X86_64 y
 
 define_bool CONFIG_X86 y
-define_bool CONFIG_ISA y
+define_bool CONFIG_ISA n
 define_bool CONFIG_SBUS n
 
 define_bool CONFIG_UID16 y
@@ -94,7 +99,7 @@ source drivers/mtd/Config.in
 
 source drivers/parport/Config.in
 
-source drivers/pnp/Config.in
+#source drivers/pnp/Config.in
 
 source drivers/block/Config.in
 
@@ -142,9 +147,10 @@ if [ "$CONFIG_NET" = "y" ]; then
    bool 'Network device support' CONFIG_NETDEVICES
    if [ "$CONFIG_NETDEVICES" = "y" ]; then
       source drivers/net/Config.in
-      if [ "$CONFIG_ATM" = "y" ]; then
-         source drivers/atm/Config.in
-      fi
+# ATM seems to be largely 64bit unsafe and also unmaintained.
+#      if [ "$CONFIG_ATM" = "y" ]; then
+#         source drivers/atm/Config.in
+#      fi
    fi
    endmenu
 fi
Only in linux/arch/x86_64: config.in-o
diff -X ../../KDIFX -burp linux-vanilla/drivers/char/Config.in linux/drivers/char/Config.in
--- linux-vanilla/drivers/char/Config.in	Fri May  3 13:22:24 2002
+++ linux/drivers/char/Config.in	Thu May  9 19:30:51 2002
@@ -39,7 +39,7 @@ if [ "$CONFIG_SERIAL_NONSTANDARD" = "y" 
    if [ "$CONFIG_DIGIEPCA" = "n" ]; then
       tristate '  Digiboard PC/Xx Support' CONFIG_DIGI
    fi
-   tristate '  Hayes ESP serial port support' CONFIG_ESPSERIAL
+   dep_tristate '  Hayes ESP serial port support' CONFIG_ESPSERIAL CONFIG_ISA
    tristate '  Moxa Intellio support' CONFIG_MOXA_INTELLIO
    tristate '  Moxa SmartIO support' CONFIG_MOXA_SMARTIO
    if [ "$CONFIG_EXPERIMENTAL" = "y" ]; then
Only in linux/drivers/char: Config.in-o
Only in linux/drivers/char: conmakehash
diff -X ../../KDIFX -burp linux-vanilla/drivers/media/radio/Config.in linux/drivers/media/radio/Config.in
--- linux-vanilla/drivers/media/radio/Config.in	Sun Apr 14 21:18:54 2002
+++ linux/drivers/media/radio/Config.in	Thu May  9 20:07:45 2002
@@ -4,6 +4,7 @@
 mainmenu_option next_comment
 comment 'Radio Adapters'
 
+if [ "$CONFIG_ISA" = "y" ]; then
 dep_tristate '  ADS Cadet AM/FM Tuner' CONFIG_RADIO_CADET $CONFIG_VIDEO_DEV
 dep_tristate '  AIMSlab RadioTrack (aka RadioReveal) support' CONFIG_RADIO_RTRACK $CONFIG_VIDEO_DEV
 if [ "$CONFIG_RADIO_RTRACK" = "y" ]; then
@@ -21,9 +22,11 @@ dep_tristate '  GemTek Radio Card suppor
 if [ "$CONFIG_RADIO_GEMTEK" = "y" ]; then
    hex '    GemTek i/o port (0x20c, 0x30c, 0x24c or 0x34c)' CONFIG_RADIO_GEMTEK_PORT 34c
 fi
+fi
 dep_tristate '  GemTek PCI Radio Card support' CONFIG_RADIO_GEMTEK_PCI $CONFIG_VIDEO_DEV $CONFIG_PCI
 dep_tristate '  Guillemot MAXI Radio FM 2000 radio' CONFIG_RADIO_MAXIRADIO $CONFIG_VIDEO_DEV
 dep_tristate '  Maestro on board radio' CONFIG_RADIO_MAESTRO $CONFIG_VIDEO_DEV
+if [ "$CONFIG_ISA" = "y" ]; then
 dep_tristate '  miroSOUND PCM20 radio' CONFIG_RADIO_MIROPCM20 $CONFIG_VIDEO_DEV $CONFIG_SOUND_ACI_MIXER
 dep_tristate '    miroSOUND PCM20 radio RDS user interface (EXPERIMENTAL)' CONFIG_RADIO_MIROPCM20_RDS $CONFIG_RADIO_MIROPCM20 $CONFIG_EXPERIMENTAL
 dep_tristate '  SF16FMI Radio' CONFIG_RADIO_SF16FMI $CONFIG_VIDEO_DEV
@@ -48,6 +51,7 @@ fi
 dep_tristate '  Zoltrix Radio' CONFIG_RADIO_ZOLTRIX $CONFIG_VIDEO_DEV
 if [ "$CONFIG_RADIO_ZOLTRIX" = "y" ]; then
    hex '    ZOLTRIX I/O port (0x20c or 0x30c)' CONFIG_RADIO_ZOLTRIX_PORT 20c
+fi
 fi
 
 endmenu
Only in linux/drivers/media/radio: Config.in-o
diff -X ../../KDIFX -burp linux-vanilla/drivers/net/hamradio/Config.in linux/drivers/net/hamradio/Config.in
--- linux-vanilla/drivers/net/hamradio/Config.in	Sun Apr 14 21:18:42 2002
+++ linux/drivers/net/hamradio/Config.in	Thu May  9 19:59:53 2002
@@ -4,8 +4,10 @@ dep_tristate 'Serial port KISS driver' C
 dep_tristate 'Serial port 6PACK driver' CONFIG_6PACK $CONFIG_AX25
 dep_tristate 'BPQ Ethernet driver' CONFIG_BPQETHER $CONFIG_AX25
     
-dep_tristate 'High-speed (DMA) SCC driver for AX.25' CONFIG_DMASCC $CONFIG_AX25
-dep_tristate 'Z8530 SCC driver' CONFIG_SCC $CONFIG_AX25
+if [ "$CONFIG_ISA" = "y" ]; then
+  dep_tristate 'High-speed (DMA) SCC driver for AX.25' CONFIG_DMASCC $CONFIG_AX25
+  dep_tristate 'Z8530 SCC driver' CONFIG_SCC $CONFIG_AX25
+fi
 if [ "$CONFIG_SCC" != "n" ]; then
    bool '  additional delay for PA0HZP OptoSCC compatible boards' CONFIG_SCC_DELAY
    bool '  support for TRX that feedback the tx signal to rx' CONFIG_SCC_TRXECHO
Only in linux/drivers/net/hamradio: Config.in-o
diff -X ../../KDIFX -burp linux-vanilla/drivers/net/wan/Config.in linux/drivers/net/wan/Config.in
--- linux-vanilla/drivers/net/wan/Config.in	Mon May  6 13:11:57 2002
+++ linux/drivers/net/wan/Config.in	Thu May  9 19:59:54 2002
@@ -9,11 +9,12 @@ bool 'Wan interfaces support' CONFIG_WAN
 if [ "$CONFIG_WAN" = "y" ]; then
 # There is no way to detect a comtrol sv11 - force it modular for now.
 
+   if [ "$CONFIG_ISA" = "y" ]; then
    dep_tristate '  Comtrol Hostess SV-11 support' CONFIG_HOSTESS_SV11 m
-
 # The COSA/SRP driver has not been tested as non-modular yet.
 
    dep_tristate '  COSA/SRP sync serial boards support' CONFIG_COSA m
+   fi
 
 #
 # COMX drivers
Only in linux/drivers/net/wan: Config.in-o
Only in linux/drivers/pci: gen-devlist
diff -X ../../KDIFX -burp linux-vanilla/drivers/scsi/Config.in linux/drivers/scsi/Config.in
--- linux-vanilla/drivers/scsi/Config.in	Fri May  3 13:22:27 2002
+++ linux/drivers/scsi/Config.in	Thu May  9 19:54:28 2002
@@ -42,11 +42,17 @@ fi
 if [ "$CONFIG_PCI" = "y" ]; then
    dep_tristate '3ware Hardware ATA-RAID support' CONFIG_BLK_DEV_3W_XXXX_RAID $CONFIG_SCSI
 fi
-dep_tristate '7000FASST SCSI support' CONFIG_SCSI_7000FASST $CONFIG_SCSI
+if [ "$CONFIG_ISA" = "y" ]; then
+   dep_tristate '7000FASST SCSI support' CONFIG_SCSI_7000FASST $CONFIG_SCSI
+fi
 dep_tristate 'ACARD SCSI support' CONFIG_SCSI_ACARD $CONFIG_SCSI
-dep_tristate 'Adaptec AHA152X/2825 support' CONFIG_SCSI_AHA152X $CONFIG_SCSI
-dep_tristate 'Adaptec AHA1542 support' CONFIG_SCSI_AHA1542 $CONFIG_SCSI
-dep_tristate 'Adaptec AHA1740 support' CONFIG_SCSI_AHA1740 $CONFIG_SCSI
+if [ "$CONFIG_ISA" = "y" ]; then
+   dep_tristate 'Adaptec AHA152X/2825 support' CONFIG_SCSI_AHA152X $CONFIG_SCSI
+   dep_tristate 'Adaptec AHA1542 support' CONFIG_SCSI_AHA1542 $CONFIG_SCSI
+fi
+if [ "$CONFIG_EISA" = "y" ]; then
+  dep_tristate 'Adaptec AHA1740 support' CONFIG_SCSI_AHA1740 $CONFIG_SCSI
+fi
 source drivers/scsi/aic7xxx/Config.in
 if [ "$CONFIG_SCSI_AIC7XXX" != "y" ]; then
    dep_tristate 'Old Adaptec AIC7xxx support' CONFIG_SCSI_AIC7XXX_OLD $CONFIG_SCSI
@@ -56,10 +62,16 @@ if [ "$CONFIG_SCSI_AIC7XXX" != "y" ]; th
       bool '  Collect statistics to report in /proc' CONFIG_AIC7XXX_OLD_PROC_STATS
    fi
 fi
-dep_tristate 'Adaptec I2O RAID support ' CONFIG_SCSI_DPT_I2O $CONFIG_SCSI
+# All the I2O code and drivers do not seem to be 64bit safe.
+if [ "$CONFIG_X86_64" != "y" ]; then
+  dep_tristate 'Adaptec I2O RAID support ' CONFIG_SCSI_DPT_I2O $CONFIG_SCSI
+fi
 dep_tristate 'AdvanSys SCSI support' CONFIG_SCSI_ADVANSYS $CONFIG_SCSI
 dep_tristate 'Always IN2000 SCSI support' CONFIG_SCSI_IN2000 $CONFIG_SCSI
-dep_tristate 'AM53/79C974 PCI SCSI support' CONFIG_SCSI_AM53C974 $CONFIG_SCSI $CONFIG_PCI
+# does not use pci dma and seems to be isa/onboard only for old machines
+if [ "$CONFIG_X86_64" != "y" ]; then
+  dep_tristate 'AM53/79C974 PCI SCSI support' CONFIG_SCSI_AM53C974 $CONFIG_SCSI $CONFIG_PCI
+fi
 dep_tristate 'AMI MegaRAID support' CONFIG_SCSI_MEGARAID $CONFIG_SCSI
 
 dep_tristate 'BusLogic SCSI support' CONFIG_SCSI_BUSLOGIC $CONFIG_SCSI
@@ -70,7 +82,9 @@ if [ "$CONFIG_PCI" = "y" ]; then
    dep_tristate 'Compaq Fibre Channel 64-bit/66Mhz HBA support' CONFIG_SCSI_CPQFCTS $CONFIG_SCSI
 fi
 dep_tristate 'DMX3191D SCSI support' CONFIG_SCSI_DMX3191D $CONFIG_SCSI $CONFIG_PCI
-dep_tristate 'DTC3180/3280 SCSI support' CONFIG_SCSI_DTC3280 $CONFIG_SCSI
+if [ "$CONFIG_ISA" != "y" ]; then
+  dep_tristate 'DTC3180/3280 SCSI support' CONFIG_SCSI_DTC3280 $CONFIG_SCSI
+fi
 dep_tristate 'EATA ISA/EISA/PCI (DPT and generic EATA/DMA-compliant boards) support' CONFIG_SCSI_EATA $CONFIG_SCSI
 if [ "$CONFIG_SCSI_EATA" != "n" ]; then
    bool '  enable tagged command queueing' CONFIG_SCSI_EATA_TAGGED_QUEUE
@@ -84,12 +98,14 @@ if [ "$CONFIG_MCA" = "y" ]; then
    dep_tristate 'Future Domain MCS-600/700 SCSI support' CONFIG_SCSI_FD_MCS $CONFIG_SCSI
 fi
 dep_tristate 'Intel/ICP (former GDT SCSI Disk Array) RAID Controller support' CONFIG_SCSI_GDTH $CONFIG_SCSI
-dep_tristate 'Generic NCR5380/53c400 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 $CONFIG_SCSI
-if [ "$CONFIG_SCSI_GENERIC_NCR5380" != "n" ]; then
+if [ "$CONFIG_ISA" = "y" ]; then
+  dep_tristate 'Generic NCR5380/53c400 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 $CONFIG_SCSI
+  if [ "$CONFIG_SCSI_GENERIC_NCR5380" != "n" ]; then
    bool '  Enable NCR53c400 extensions' CONFIG_SCSI_GENERIC_NCR53C400
    choice 'NCR5380/53c400 mapping method (use Port for T130B)' \
 		"Port CONFIG_SCSI_G_NCR5380_PORT \
 		 Memory CONFIG_SCSI_G_NCR5380_MEM" Port
+  fi
 fi
 if [ "$CONFIG_MCA" = "y" ]; then
    dep_tristate 'IBMMCA SCSI support' CONFIG_SCSI_IBMMCA $CONFIG_SCSI
@@ -111,7 +127,9 @@ if [ "$CONFIG_PARPORT" != "n" ]; then
       bool  '  ppa/imm option - Assume slow parport control register' CONFIG_SCSI_IZIP_SLOW_CTR
    fi
 fi
-dep_tristate 'NCR53c406a SCSI support' CONFIG_SCSI_NCR53C406A $CONFIG_SCSI
+if [ "$CONFIG_ISA" = "y" ]; then
+  dep_tristate 'NCR53c406a SCSI support' CONFIG_SCSI_NCR53C406A $CONFIG_SCSI
+fi
 if [ "$CONFIG_MCA" = "y" ]; then
    dep_tristate 'NCR Dual 700 MCA SCSI support' CONFIG_SCSI_NCR_D700 $CONFIG_SCSI
    if [ "$CONFIG_SCSI_NCR_D700" != "n" ]; then
@@ -164,11 +182,17 @@ fi
 if [ "$CONFIG_MCA" = "y" ]; then
    dep_tristate 'NCR MCA 53C9x SCSI support' CONFIG_SCSI_MCA_53C9X $CONFIG_SCSI
 fi
-dep_tristate 'PAS16 SCSI support' CONFIG_SCSI_PAS16 $CONFIG_SCSI
+if [ "$CONFIG_ISA" = "y" ]; then
+   dep_tristate 'PAS16 SCSI support' CONFIG_SCSI_PAS16 $CONFIG_SCSI
+fi
 dep_tristate 'PCI2000 support' CONFIG_SCSI_PCI2000 $CONFIG_SCSI
 dep_tristate 'PCI2220i support' CONFIG_SCSI_PCI2220I $CONFIG_SCSI
-dep_tristate 'PSI240i support' CONFIG_SCSI_PSI240I $CONFIG_SCSI
-dep_tristate 'Qlogic FAS SCSI support' CONFIG_SCSI_QLOGIC_FAS $CONFIG_SCSI
+if [ "$CONFIG_ISA" = "y" ]; then
+   dep_tristate 'PSI240i support' CONFIG_SCSI_PSI240I $CONFIG_SCSI
+fi
+if [ "$CONFIG_ISA" = "y" ]; then
+   dep_tristate 'Qlogic FAS SCSI support' CONFIG_SCSI_QLOGIC_FAS $CONFIG_SCSI
+fi
 if [ "$CONFIG_PCI" = "y" ]; then
    dep_tristate 'Qlogic ISP SCSI support' CONFIG_SCSI_QLOGIC_ISP $CONFIG_SCSI
    dep_tristate 'Qlogic ISP FC SCSI support' CONFIG_SCSI_QLOGIC_FC $CONFIG_SCSI
@@ -177,24 +201,31 @@ if [ "$CONFIG_PCI" = "y" ]; then
    fi
    dep_tristate 'Qlogic QLA 1280 SCSI support' CONFIG_SCSI_QLOGIC_1280 $CONFIG_SCSI
 fi
-if [ "$CONFIG_X86" = "y" ]; then
+if [ "$CONFIG_X86" = "y" -a "$CONFIG_ISA" = "y" ]; then
    dep_tristate 'Seagate ST-02 and Future Domain TMC-8xx SCSI support' CONFIG_SCSI_SEAGATE $CONFIG_SCSI
 fi
-dep_tristate 'Simple 53c710 SCSI support (Compaq, NCR machines)' CONFIG_SCSI_SIM710 $CONFIG_SCSI
-dep_tristate 'Symbios 53c416 SCSI support' CONFIG_SCSI_SYM53C416 $CONFIG_SCSI
+# definitely looks note 64bit safe:
+if [ "$CONFIG_ISA" = "y" -o "$CONFIG_MCA" = "y" -a "$CONFIG_X86_64" != "y" ]; then
+  dep_tristate 'Simple 53c710 SCSI support (Compaq, NCR machines)' CONFIG_SCSI_SIM710 $CONFIG_SCSI
+fi
+if [ "$CONFIG_ISA" = "y" ]; then
+  dep_tristate 'Symbios 53c416 SCSI support' CONFIG_SCSI_SYM53C416 $CONFIG_SCSI
+fi
 if [ "$CONFIG_PCI" = "y" ]; then
    dep_tristate 'Tekram DC390(T) and Am53/79C974 SCSI support' CONFIG_SCSI_DC390T $CONFIG_SCSI
    if [ "$CONFIG_SCSI_DC390T" != "n" ]; then
       bool '  _omit_ support for non-DC390 adapters' CONFIG_SCSI_DC390T_NOGENSUPP
    fi
 fi
-dep_tristate 'Trantor T128/T128F/T228 SCSI support' CONFIG_SCSI_T128 $CONFIG_SCSI
+if [ "$CONFIG_ISA" = "y" ]; then
+  dep_tristate 'Trantor T128/T128F/T228 SCSI support' CONFIG_SCSI_T128 $CONFIG_SCSI
+fi
 dep_tristate 'UltraStor 14F/34F support' CONFIG_SCSI_U14_34F $CONFIG_SCSI
-   if [ "$CONFIG_SCSI_U14_34F" != "n" ]; then
+if [ "$CONFIG_SCSI_U14_34F" != "n" ]; then
       bool '  enable elevator sorting' CONFIG_SCSI_U14_34F_LINKED_COMMANDS
       int  '  maximum number of queued commands' CONFIG_SCSI_U14_34F_MAX_TAGS 8
-   fi
-if [ "$CONFIG_X86" = "y" ]; then
+fi
+if [ "$CONFIG_X86" = "y" -a "$CONFIG_ISA" = "y" ]; then
    dep_tristate 'UltraStor SCSI support' CONFIG_SCSI_ULTRASTOR $CONFIG_SCSI
 fi
 #
Only in linux/drivers/scsi: Config.in-o

Re: PATCH & call for help: Marking ISA only drivers

[William Stearns]

Good evening, Andi,

On Thu, 9 May 2002, Andi Kleen wrote:

> This patch tries to make most ISA only drivers dependent on CONFIG_ISA. 

	I did quite a bit of this work for CML2 - bus dependencies can be 
found in the CML2 sources.
	Cheers,
	- Bill

---------------------------------------------------------------------------
        "Do you smell something burning or is it me?"
        -- Joan of Arc
--------------------------------------------------------------------------
William Stearns (wstearns@pobox.com).  Mason, Buildkernel, named2hosts, 
and ipfwadm2ipchains are at:                        http://www.stearns.org
--------------------------------------------------------------------------

Re: PATCH & call for help: Marking ISA only drivers

[Martin Dalecki]

Uz.ytkownik Andi Kleen napisa?:
> Hallo,
> 
> This patch tries to make most ISA only drivers dependent on CONFIG_ISA. 

If only for the fact that it allows you to don't look at archaic
hardware configuration options makes it a good idea I think.
Bus is bus if we have CONFIG_PCI, we should have CONFIG_ISA as well.

> The motivation is that it is a lot of work to get old drivers to compile
> meaningfully (at least without warnings, not even testing them) on x86-64
> and a lot of them are obviously not 64bit safe.  As it is very unlikely
> that x86-64 boxes will ever have ISA slots[1] one simple way for that
> is just removing the old ISA drivers from the configuration.
> 
> BTW I think CONFIG_ISA would be an useful configuration option for 
> i386 too - at least most modern PCs do not come with ISA slots anymore.

Plase add mcd and mcdx - CD-ROM drivers. Both of them required
an special "controller" card, which was indeed ISA based.

CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[Tomas Szepe]

> I did quite a bit of this work for CML2 - bus dependencies can be
> found in the CML2 sources.

Btw, what happened to CML2?
I haven't seen any updates since about February and haven't almost
certainly stumbled upon a post from ESR for quite long either.

T.

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[Dave Jones]

On Thu, May 09, 2002 at 10:08:19PM +0200, Tomas Szepe wrote:
 > > I did quite a bit of this work for CML2 - bus dependencies can be
 > > found in the CML2 sources.
 > Btw, what happened to CML2?

The majority seemed to be of the opinion it was too overfeatured whilst
lacking some of the more basic functionality desired from those who
would end up using it on a daily basis in favour of eye candy for
users who compile a kernel once per release.

 > I haven't seen any updates since about February and haven't almost
 > certainly stumbled upon a post from ESR for quite long either.

Yes, signal has been quite high on Linux Kernel last few months.

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[Alexander Viro]

On Thu, 9 May 2002, Tomas Szepe wrote:

> > I did quite a bit of this work for CML2 - bus dependencies can be
> > found in the CML2 sources.
> 
> Btw, what happened to CML2?

This is an ex-parrot

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[Tomas Szepe]

> [Alexander Viro <viro@math.psu.edu>, May-09 2002, Thu, 16:23 -0400]
> 
> > > I did quite a bit of this work for CML2 - bus dependencies can be
> > > found in the CML2 sources.
> > 
> > Btw, what happened to CML2?
> 
> This is an ex-parrot

If you're referring to what I think you are (uhhh), then maybe
"he's just resting"?

I mean, people certainly don't want to see a slug in its place.

:)


T.

Re: PATCH & call for help: Marking ISA only drivers

[Alan Cox]

> BTW I think CONFIG_ISA would be an useful configuration option for 
> i386 too - at least most modern PCs do not come with ISA slots anymore.

ISA slots != ISA devices

And ISA stuff hanging off things like chipset glue busses is everywhere

> +if [ "$CONFIG_ISA" = "y" ]; then
> +   dep_tristate 'Adaptec AHA152X/2825 support' CONFIG_SCSI_AHA152X $CONFIG_SCSI

2825 is not ISA bus

> +# does not use pci dma and seems to be isa/onboard only for old machines
> +if [ "$CONFIG_X86_64" != "y" ]; then
> +  dep_tristate 'AM53/79C974 PCI SCSI support' CONFIG_SCSI_AM53C974 $CONFIG_SCSI $CONFIG_PCI
> +fi

Thats PCI.

> +if [ "$CONFIG_ISA" = "y" ]; then
> +  dep_tristate 'Generic NCR5380/53c400 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 $CONFIG_SCSI
> +  if [ "$CONFIG_SCSI_GENERIC_NCR5380" != "n" ]; then

This is used in multiple non ISA situations.

> +if [ "$CONFIG_X86" = "y" -a "$CONFIG_ISA" = "y" ]; then
>     dep_tristate 'UltraStor SCSI support' CONFIG_SCSI_ULTRASTOR $CONFIG_SCSI
>  fi

It makes an ugly mess of Config.in - I guess 2.5 is the time to do what
Russell and Keith wanted and make undefined imply "N" in dep_*

Alan

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[toon]

On Thu, May 09, 2002 at 10:08:19PM +0200, Tomas Szepe wrote:
> > I did quite a bit of this work for CML2 - bus dependencies can be
> > found in the CML2 sources.
> 
> Btw, what happened to CML2?
> I haven't seen any updates since about February and haven't almost
> certainly stumbled upon a post from ESR for quite long either.

Eric was blasted away from the linux-kernel mailing list.
I understand that people tend to disagree about some subject every
once in a while. But he certainly didn't deserve the treatment he got.
Specially considering the enormous amount of work he put in it.

And I miss his signatures and the discussions about Aunt Tillie and stuff.  :-)

Toon.
-- 
 /"\                             |
 \ /     ASCII RIBBON CAMPAIGN   |
  X        AGAINST HTML MAIL     |
 / \

Re: PATCH & call for help: Marking ISA only drivers

[Andi Kleen]

On Thu, May 09, 2002 at 09:51:54PM +0200, Martin Dalecki wrote:
> Uz.ytkownik Andi Kleen napisa?:
> >Hallo,
> >
> >This patch tries to make most ISA only drivers dependent on CONFIG_ISA. 
> 
> If only for the fact that it allows you to don't look at archaic
> hardware configuration options makes it a good idea I think.

And often the old drivers do not work anymore anyways...

> >The motivation is that it is a lot of work to get old drivers to compile
> >meaningfully (at least without warnings, not even testing them) on x86-64
> >and a lot of them are obviously not 64bit safe.  As it is very unlikely
> >that x86-64 boxes will ever have ISA slots[1] one simple way for that
> >is just removing the old ISA drivers from the configuration.
> >
> >BTW I think CONFIG_ISA would be an useful configuration option for 
> >i386 too - at least most modern PCs do not come with ISA slots anymore.
> 
> Plase add mcd and mcdx - CD-ROM drivers. Both of them required
> an special "controller" card, which was indeed ISA based.

I removed the complete drivers/cdrom directory for x86-64 now, as it seems to 
contain no PCI code at all. For i386 it can be made dependent on CONFIG_ISA.

Also I did this change for the IDE configuration (assuming that there 
are no boxes with no ISA but a VLB slot). I hope this covers all 
ISA/VLB only adapters for IDE.

You may want to incorporate it.

-Andi

--- linux/drivers/ide/Config.in-o	Mon May  6 13:11:47 2002
+++ linux/drivers/ide/Config.in	Thu May  9 20:42:45 2002
@@ -131,7 +131,8 @@
 	      EXT_DIRECT	CONFIG_IDE_EXT_DIRECT"	8xx_PCCARD
    fi
 
-   bool '  Other IDE chipset support' CONFIG_IDE_CHIPSETS
+   # assume no ISA -> also no VLB
+   dep_bool '  Other ISA/VLB IDE chipset support' CONFIG_IDE_CHIPSETS CONFIG_ISA
    if [ "$CONFIG_IDE_CHIPSETS" = "y" ]; then
       comment 'Note: most of these also require special kernel boot parameters'
       bool '    ALI M14xx support' CONFIG_BLK_DEV_ALI14XX

Re: PATCH & call for help: Marking ISA only drivers

[Andi Kleen]

On Thu, May 09, 2002 at 10:33:21PM +0200, Alan Cox wrote:
> > BTW I think CONFIG_ISA would be an useful configuration option for 
> > i386 too - at least most modern PCs do not come with ISA slots anymore.
> 
> ISA slots != ISA devices
> 
> And ISA stuff hanging off things like chipset glue busses is everywhere
> 
> > +if [ "$CONFIG_ISA" = "y" ]; then
> > +   dep_tristate 'Adaptec AHA152X/2825 support' CONFIG_SCSI_AHA152X $CONFIG_SCSI
> 
> 2825 is not ISA bus

What then ?

> 
> > +# does not use pci dma and seems to be isa/onboard only for old machines
> > +if [ "$CONFIG_X86_64" != "y" ]; then
> > +  dep_tristate 'AM53/79C974 PCI SCSI support' CONFIG_SCSI_AM53C974 $CONFIG_SCSI $CONFIG_PCI
> > +fi
> 
> Thats PCI.

But seems to be not 64bit safe and miss pci dma support (note the CONFIG_X86_64). 

> 
> > +if [ "$CONFIG_ISA" = "y" ]; then
> > +  dep_tristate 'Generic NCR5380/53c400 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 $CONFIG_SCSI
> > +  if [ "$CONFIG_SCSI_GENERIC_NCR5380" != "n" ]; then
> 
> This is used in multiple non ISA situations.

Only on ancient motherboards (I remember having it on some really old EISA
machine) and non PC devices, no ? 
I'm hoping the non PC users will add their perspective architecture again.

-Andi

Re: PATCH & call for help: Marking ISA only drivers

[Andi Kleen]

On Fri, May 10, 2002 at 01:18:21AM +0200, Alan Cox wrote:
> > > > +if [ "$CONFIG_ISA" = "y" ]; then
> > > > +   dep_tristate 'Adaptec AHA152X/2825 support' CONFIG_SCSI_AHA152X $CONFIG_SCSI
> > > 
> > > 2825 is not ISA bus
> > 
> > What then ?
> 
> Vesa local bus

Ok, I'm assuming that there are no boxes with no ISA slots but VLB slots.
I guess that's safe. If someone really has a weird box were that is not true
I guess they'll have to live with defining CONFIG_ISA. 

In theory one could introduce an CONFIG_VLB, but I don't think it is 
worth it.

> > > > +if [ "$CONFIG_ISA" = "y" ]; then
> > > > +  dep_tristate 'Generic NCR5380/53c400 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 $CONFIG_SCSI
> > > > +  if [ "$CONFIG_SCSI_GENERIC_NCR5380" != "n" ]; then
> > > 
> > > This is used in multiple non ISA situations.
> > 
> > Only on ancient motherboards (I remember having it on some really old EISA
> > machine) and non PC devices, no ? 
> 
> On just about anything. If you have some old (or new) random weird box
> then so long as you know the address this works. NCR5380 macrocells are
> still being used I'm afraid to say, and attached to pretty much any bus
> people can find.

Ok thanks, will change it.

-Andi

Re: PATCH & call for help: Marking ISA only drivers

[Alan Cox]

> > > +if [ "$CONFIG_ISA" = "y" ]; then
> > > +   dep_tristate 'Adaptec AHA152X/2825 support' CONFIG_SCSI_AHA152X $CONFIG_SCSI
> > 
> > 2825 is not ISA bus
> 
> What then ?

Vesa local bus

> But seems to be not 64bit safe and miss pci dma support (note the CONFIG_X86_64). 

My fault

> > > +if [ "$CONFIG_ISA" = "y" ]; then
> > > +  dep_tristate 'Generic NCR5380/53c400 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 $CONFIG_SCSI
> > > +  if [ "$CONFIG_SCSI_GENERIC_NCR5380" != "n" ]; then
> > 
> > This is used in multiple non ISA situations.
> 
> Only on ancient motherboards (I remember having it on some really old EISA
> machine) and non PC devices, no ? 

On just about anything. If you have some old (or new) random weird box
then so long as you know the address this works. NCR5380 macrocells are
still being used I'm afraid to say, and attached to pretty much any bus
people can find.

Alan

Re: PATCH & call for help: Marking ISA only drivers

[Alan Cox]

> Ok, I'm assuming that there are no boxes with no ISA slots but VLB slots.
> I guess that's safe. If someone really has a weird box were that is not true
> I guess they'll have to live with defining CONFIG_ISA. 
> In theory one could introduce an CONFIG_VLB, but I don't think it is 
> worth it.

Agreed

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[jeff millar]

Some personal thoughts from a Linux user.  Seems like several possibilities
for Eric's attitude...

    1. waiting for people to see the need for CML2
    2. waiting for people who said they could do it better to put up
    3. totally disgusted by the attitude and behavior on the list
    4. all of the above

I never got CML2 to work correctly and found Eric to act helpful but to not
really help.  It surprised me that he knew little about how people (me)
actually used kernel config and build scripts.

Free software has to attract people...developers can't force it.  CML2
required powerful, knowledgeable kernel developers to change and they didn't
want to.

----- Original Message -----
From: "Tomas Szepe" <szepe@pinerecords.com>
To: "William Stearns" <wstearns@pobox.com>
Cc: "Andi Kleen" <ak@muc.de>; "ML-linux-kernel"
<linux-kernel@vger.kernel.org>
Sent: Thursday, May 09, 2002 4:08 PM
Subject: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]


> > I did quite a bit of this work for CML2 - bus dependencies can be
> > found in the CML2 sources.
>
> Btw, what happened to CML2?
> I haven't seen any updates since about February and haven't almost
> certainly stumbled upon a post from ESR for quite long either.
>
> T.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: PATCH & call for help: Marking ISA only drivers

[Martin Dalecki]

Uz.ytkownik Alan Cox napisa?:
>>Ok, I'm assuming that there are no boxes with no ISA slots but VLB slots.
>>I guess that's safe. If someone really has a weird box were that is not true
>>I guess they'll have to live with defining CONFIG_ISA. 
>>In theory one could introduce an CONFIG_VLB, but I don't think it is 
>>worth it.
> 
> 
> Agreed

In esp. in face of the fact that VLB was in reality merely the
486 CPU to memmory bus hanging out of the system... so it didn't
have it's own protocol different from inb outb - which logically corresponds
to ISA.

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[Sean Hunter]

On Thu, May 09, 2002 at 10:49:28PM +0200, toon@vdpas.hobby.nl wrote:
> On Thu, May 09, 2002 at 10:08:19PM +0200, Tomas Szepe wrote:
> > > I did quite a bit of this work for CML2 - bus dependencies can be
> > > found in the CML2 sources.
> > 
> > Btw, what happened to CML2?
> > I haven't seen any updates since about February and haven't almost
> > certainly stumbled upon a post from ESR for quite long either.
> 
> Eric was blasted away from the linux-kernel mailing list.
> I understand that people tend to disagree about some subject every
> once in a while. But he certainly didn't deserve the treatment he got.
> Specially considering the enormous amount of work he put in it.

Perhaps that will be a lesson to the self-styled "Hacker of social systems"
that there is a reason he has two ears and only one mouth.  

As a low-evel hacker he needed to elicit support from others to get his work
recognised and included.  He was unwilling to listen to advice and criticism
from those who know a lot better than him, and his work was not of such obvious
brilliance that people were prepared to accept it without change.

He was not prepared to make the changes folks wanted, so he took his ball and
went away in a strop.

> And I miss his signatures and the discussions about Aunt Tillie and stuff.  :-)

I certainly don't miss his signatures.  I considered a lot of them to be
offensively off-topic and needlessly provocative.

Sean

Re: CML2 [was Re: PATCH & call for help: Marking ISA only drivers]

[Andrew Rodland]

[-- Attachment #1: Type: text/plain, Size: 1213 bytes --]

On Thu, 9 May 2002 22:49:28 +0200
toon@vdpas.hobby.nl wrote:

> On Thu, May 09, 2002 at 10:08:19PM +0200, Tomas Szepe wrote:
> > > I did quite a bit of this work for CML2 - bus dependencies can be
> > > found in the CML2 sources.
> > 
> > Btw, what happened to CML2?
> > I haven't seen any updates since about February and haven't almost
> > certainly stumbled upon a post from ESR for quite long either.
> 
> Eric was blasted away from the linux-kernel mailing list.
> I understand that people tend to disagree about some subject every
> once in a while. But he certainly didn't deserve the treatment he got.
> Specially considering the enormous amount of work he put in it.

Agreed. I did a lot of testing and some grunt-level coding on CML2, and
I think it addressed a lot of problems that existed, and was getting
steadily better. Sure, it had some flaky spots, but I really don't see
all of the problems that people had with it. Even autoconfig had this
tendency to do the Right Thing. I think it was just about ready to start
getting some wide testing and use, when everyone abandoned it... And now
nobody wants to use it, because of course patches don't play nice with
it, and etc... Pretty sad.

Oh well.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]