[PATCH] 2.4.25: Get rid of obsolete LMC driver

[PATCH] 2.4.25: Get rid of obsolete LMC driver

[Daniel Egger]

[-- Attachment #1.1: Type: text/plain, Size: 752 bytes --]

Hi Marcello,

This patch plus an additional
rm -r drivers/net/wan/lmc
gets rid of the obsolete LMC WAN driver and all references to it in the
Configure.in and Makefiles and MAINTAINERS.

When LMC was taken over by SBE their kernel version of the driver went
effectively unmaintained after Alan Cox turned down their LMC->SBE
rename patch. Today SBE recommends their own version of the driver only
which unfortunately needs different tools and also clashes with the LMC
driver when not compiled as module. This general recommendation from
their support effectively makes the in-kernel code obsete.

Their drivers can be retrieved from:
ftp://ftp.sbei.com/pub/OpenSource/Linux/sbe_driver/sbe_linux-4.0a.tgz

-- 
Servus,
       Daniel

[-- Attachment #1.2: lmc.diff --]
[-- Type: text/x-patch, Size: 3138 bytes --]

--- linux-2.4.25-testbed.old/Documentation/Configure.help	2004-02-28 12:17:06.000000000 +0100
+++ linux-2.4.25-testbed/Documentation/Configure.help	2004-02-28 12:26:12.000000000 +0100
@@ -11043,32 +11043,6 @@
 
   Say Y if yout card supports this feature.
 
-LanMedia Corp. serial boards (SSI/V.35, T1/E1, HSSI, T3)
-CONFIG_LANMEDIA
-  This is a driver for the following Lan Media family of serial
-  boards.
-
-  LMC 1000 board allows you to connect synchronous serial devices (for
-  example base-band modems, or any other device with the X.21, V.24,
-  V.35 or V.36 interface) to your Linux box.
-
-  LMC 1200 with on board DSU board allows you to connect your Linux
-  box directly to a T1 or E1 circuit.
-
-  LMC 5200 board provides a HSSI interface capable of running up to
-  52 mbits per second.
-
-  LMC 5245 board connects directly to a T3 circuit saving the
-  additional external hardware.
-
-  To change setting such as syncPPP vs cisco HDLC or clock source you
-  will need lmcctl.  It is available at <ftp://ftp.lanmedia.com/>.
-
-  This code is also available as a module called lmc.o ( = code
-  which can be inserted in and removed from the running kernel
-  whenever you want). If you want to compile it as a module, say M
-  here and read <file:Documentation/modules.txt>.
-
 Fibre Channel driver support
 CONFIG_NET_FC
   Fibre Channel is a high speed serial protocol mainly used to connect
--- linux-2.4.25-testbed.old/MAINTAINERS	2004-02-28 12:22:28.000000000 +0100
+++ linux-2.4.25-testbed/MAINTAINERS	2004-02-28 12:21:08.000000000 +0100
@@ -1084,12 +1084,6 @@
 W:	http://www.cse.unsw.edu.au/~neilb/oss/knfsd/
 S:	Maintained
 
-LANMEDIA WAN CARD DRIVER
-P:      Andrew Stanley-Jones
-M:      asj@lanmedia.com
-W:      http://www.lanmedia.com/
-S:      Supported
- 
 LAPB module
 P:	Henner Eisen
 M:	eis@baty.hanse.de
--- linux-2.4.25-testbed.old/drivers/net/setup.c	2004-01-15 20:34:16.000000000 +0100
+++ linux-2.4.25-testbed/drivers/net/setup.c	2004-02-28 12:28:05.000000000 +0100
@@ -76,10 +76,6 @@
 	{comx_init, 0},
 #endif
 	 
-#if defined(CONFIG_LANMEDIA)
-	{lmc_setup, 0},
-#endif
-	 
 /*
 *
 *	Wireless non-HAM
--- linux-2.4.25-testbed.old/drivers/net/wan/Config.in	2004-02-28 12:17:40.000000000 +0100
+++ linux-2.4.25-testbed/drivers/net/wan/Config.in	2004-02-28 12:28:42.000000000 +0100
@@ -46,12 +46,6 @@
    fi
 
 #
-# Lan Media's board. Currently 1000, 1200, 5200, 5245
-#
-
-   tristate '  LanMedia Corp. SSI/V.35, T1/E1, HSSI, T3 boards' CONFIG_LANMEDIA
-
-#
 # ATI's sync/async boards. Currently 2520, 4020, 4520, 8520 -- 0 in second digit means async only
 #
 
--- linux-2.4.25-testbed.old/drivers/net/wan/Makefile	2004-02-28 12:17:40.000000000 +0100
+++ linux-2.4.25-testbed/drivers/net/wan/Makefile	2004-02-28 12:25:29.000000000 +0100
@@ -50,12 +50,6 @@
 obj-$(CONFIG_SYNCLINK_SYNCPPP)	+=		syncppp.o
 obj-$(CONFIG_X25_ASY)		+= x25_asy.o
 
-subdir-$(CONFIG_LANMEDIA) += lmc
-
-ifeq ($(CONFIG_LANMEDIA),y)
-  obj-y += lmc/lmc.o
-endif
-
 subdir-$(CONFIG_ATI_XX20) += 8253x
 
 ifeq ($(CONFIG_ATI_XX20),y)

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 481 bytes --]

Re: [PATCH] 2.4.25: Get rid of obsolete LMC driver

[Francois Romieu]

Daniel Egger <de@axiros.com> :
[removal of LMC wan driver]

Shall one understand that:
- in-kernel LMC driver has been proved dysfunctionnal ?
- SBE code will be turned into something adequate for kernel integration ?

Just curious.

--
Ueimor

Re: [PATCH] 2.4.25: Get rid of obsolete LMC driver

[Daniel Egger]

[-- Attachment #1: Type: text/plain, Size: 1441 bytes --]

Am Sam, den 28.02.2004 schrieb Francois Romieu um 14:46:

> Shall one understand that:
> - in-kernel LMC driver has been proved dysfunctionnal ?

Hard to tell. Standard Raw HDLC mode works fine but I couldn't find the
correct tools to setup PPP instead. And SBE support is not really
helpful unless you're using their drivers, which had been recently
updated to a version that actually works with late 2.4 kernels.

> - SBE code will be turned into something adequate for kernel integration ?

Their package consists of tools, a patch and a standalone module source.
Depending on the kernel version the patch needs to be applied, but it
should be in the latest kernels already. The source can be compiled
against any kernel tree and works reasonably well. It probably could be
integrated into the kernel, however given the attitude of SBE and the
senselessness of the rename and some other changes I don't see a good
reason why the community should help them by doing their homework.

Unless there's a reliable driver with working tools I don't see a good
reason to have this driver around. It already bit me hard in the rear
and I'd rather see someone bitch against SBE because their driver is not
working than hunt hours for mistakes and eventually end up with SBE
anyway.

Other than that I've learnt my lesson and selfish as I am, I will not
take it further than I already did.

-- 
Servus,
       Daniel

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 481 bytes --]

Re: [PATCH] 2.4.25: Get rid of obsolete LMC driver

[Jeff Garzik]

Daniel Egger wrote:
> Hi Marcello,
> 
> This patch plus an additional
> rm -r drivers/net/wan/lmc
> gets rid of the obsolete LMC WAN driver and all references to it in the
> Configure.in and Makefiles and MAINTAINERS.
> 
> When LMC was taken over by SBE their kernel version of the driver went
> effectively unmaintained after Alan Cox turned down their LMC->SBE
> rename patch. Today SBE recommends their own version of the driver only
> which unfortunately needs different tools and also clashes with the LMC
> driver when not compiled as module. This general recommendation from
> their support effectively makes the in-kernel code obsete.
> 
> Their drivers can be retrieved from:
> ftp://ftp.sbei.com/pub/OpenSource/Linux/sbe_driver/sbe_linux-4.0a.tgz

Alan Cox vetoed a rename patch, so you want to rip out the driver 
instead???  For an unreviewed out-of-tree driver?

Without a suitable replacement, I don't give a shit about what SBE 
recommends.

Veto.

	Jeff

Re: [PATCH] 2.4.25: Get rid of obsolete LMC driver

[Daniel Egger]

[-- Attachment #1: Type: text/plain, Size: 1096 bytes --]

Am Sam, den 28.02.2004 schrieb Jeff Garzik um 17:54:

> > Their drivers can be retrieved from:
> > ftp://ftp.sbei.com/pub/OpenSource/Linux/sbe_driver/sbe_linux-4.0a.tgz

> Alan Cox vetoed a rename patch, so you want to rip out the driver 
> instead???  For an unreviewed out-of-tree driver?

Indeed. However the unreviewed out-of-tree driver is based on the kernel
one. But since:
- it's close to impossible to compile tools to change the operation mode
  of the card
- SBE card owners will not notice that the LMC driver might work for
  them
- the LMC and the SBE driver will clash if both compiled

the best we can do for our users is to recommend using the SBE drivers
and complain to SBE about problems.

> Without a suitable replacement, I don't give a shit about what SBE 
> recommends.

There is a suitable replacement: the SBE driver. But since there's no
volunteer to keep it updated in the kernel it'd be better to drop it.
I've spent several hours untangling the mess and I'd really like to help
other not getting hit by this.

-- 
Servus,
       Daniel

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 481 bytes --]