Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Adam J. Richter]

[-- Attachment #1: Type: text/plain, Size: 1055 bytes --]

	linux-2.5.30/include/linux/serialP.h needs struct async_icount,
which is defined in <linux/serial.h>, causing
linux-2.5.30/drivers/serial/8250.c not to compile, among other problems.
In linux-2.5.30, you cannot compile a file that includes <linux/serialP.h>
without including <linux/serial.h>.  So, I think the solution is for
serialP.h to #include serial.h.  I have attached a patch that does this.

	From the comments in serialP.h, it looks like there was some
effort in linux-2.2 to allow inclusion of serialP.h without serial.h,
but I see no indication of what benefit that was supposed to provide.

	Ted (or whowever gathers drivers/serial patches for Linus), do
you want to shepherd this change to Linus, do you want me to submit it
directly, or do you want to do something else?

-- 
Adam J. Richter     __     ______________   575 Oroville Road
adam@yggdrasil.com     \ /                  Milpitas, California 95035
+1 408 309-6081         | g g d r a s i l   United States of America
                         "Free Software For The Rest Of Us."

[-- Attachment #2: serialP.diff --]
[-- Type: text/plain, Size: 451 bytes --]

--- linux-2.5.30/include/linux/serialP.h	2002-08-01 14:16:07.000000000 -0700
+++ linux/include/linux/serialP.h	2002-08-02 14:51:03.000000000 -0700
@@ -24,11 +24,7 @@
 #include <linux/tqueue.h>
 #include <linux/circ_buf.h>
 #include <linux/wait.h>
-#if (LINUX_VERSION_CODE < 0x020300)
-/* Unfortunate, but Linux 2.2 needs async_icount defined here and
- * it got moved in 2.3 */
 #include <linux/serial.h>
-#endif
 
 struct serial_state {
 	int	magic;

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Russell King]

On Fri, Aug 02, 2002 at 03:49:24PM -0700, Adam J. Richter wrote:
> 	linux-2.5.30/include/linux/serialP.h needs struct async_icount,
> which is defined in <linux/serial.h>, causing
> linux-2.5.30/drivers/serial/8250.c not to compile, among other problems.
> In linux-2.5.30, you cannot compile a file that includes <linux/serialP.h>
> without including <linux/serial.h>.  So, I think the solution is for
> serialP.h to #include serial.h.  I have attached a patch that does this.

Ack.  I've just found why I and many other people can build it, and
other people can't.  I can tell you that you're building 8250.c as a
module.

Why?

When I build 8250.c into the kernel, linux/module.h doesn't include
linux/version.h, so when we include linux/serialP.h, the compiler
assumes that LINUX_VERSION_CODE is zero.  So we end up including
linux/serial.h.

However, when building as a module, linux/module.h does include
linux/version.h, so when we don't include linux/serial.h.

Oh, the problems of trying to reduce the includes...  I think we should
re-include linux/serial.h and eliminate linux/serialP.h.

Hmm, I wonder how many other oddities like this are in the tree today.
It sounds like we want to create a rule similar to the one for using
CONFIG_* symbols.  Does this sound reasonable: if you use
LINUX_VERSION_CODE, you must include linux/version.h into that very
same file to guarantee that it is defined.

Well, I took checkconfig.pl and created checkversion.pl (attached).
Oh god, can I please put the worms back in the can?  Now?  I think
there's lots of work to do here; lots of stuff including linux/version.h
for the hell of it, and a comparitively small number not including it
when they use LINUX_VERSION_CODE.

(No, I'm just off to zzz, so this must be a nightmare, maybe it'll go
away by the time I wake up later today.) 8/

> 	Ted (or whowever gathers drivers/serial patches for Linus), do
> you want to shepherd this change to Linus, do you want me to submit it
> directly, or do you want to do something else?

It's more a thing for me than Ted.  One of the things I'd like to
do is to eventually kill off serialP.h

Ok, here's a fix for the 8250.c build problem (please don't send it
to Linus; I've other changes that'll be going via BK and patch to
lkml pending):

--- orig/drivers/serial/8250.c	Fri Aug  2 21:13:31 2002
+++ linux/drivers/serial/8250.c	Sat Aug  3 00:28:47 2002
@@ -31,7 +31,8 @@
 #include <linux/console.h>
 #include <linux/sysrq.h>
 #include <linux/serial_reg.h>
-#include <linux/serialP.h>
+#include <linux/circ_buf.h>
+#include <linux/serial.h>
 #include <linux/delay.h>
 
 #include <asm/io.h>
--- orig/drivers/serial/8250.h	Sat Jul 27 13:55:21 2002
+++ linux/drivers/serial/8250.h	Sat Aug  3 00:28:21 2002
@@ -59,3 +59,15 @@
 #else
 #define SERIAL8250_SHARE_IRQS 0
 #endif
+
+#if defined(__alpha__) && !defined(CONFIG_PCI)
+/*
+ * Digital did something really horribly wrong with the OUT1 and OUT2
+ * lines on at least some ALPHA's.  The failure mode is that if either
+ * is cleared, the machine locks up with endless interrupts.
+ */
+#define ALPHA_KLUDGE_MCR  (UART_MCR_OUT2 | UART_MCR_OUT1)
+#else
+#define ALPHA_KLUDGE_MCR 0
+#endif
+


And here's the (hacked up) checkversion.pl script - its used in the same
way as checkconfig.pl in the makefile, so use like this:

$ cd linux
$ find * -name SCCS -prune -o -name BitKeeper -prune -o -name '*.[chS]' \
 -type f -print | sort | xargs perl -w ../checkversion.pl

I'll clean it up _after_ sleep, unless someone doesn't get there first.

#! /usr/bin/perl
#
# checkconfig: find uses of CONFIG_* names without matching definitions.
# Copyright abandoned, 1998, Michael Elizabeth Chastain <mailto:mec@shout.net>.

use integer;

$| = 1;

foreach $file (@ARGV)
{
    # Open this file.
    open(FILE, $file) || die "Can't open $file: $!\n";

    # Initialize variables.
    my $fInComment    = 0;
    my $fInString     = 0;
    my $fUseKernCode  = 0;
    my $iLinuxVersion = 0;

    LINE: while ( <FILE> )
    {
	# Strip comments.
	$fInComment && (s+^.*?\*/+ +o ? ($fInComment = 0) : next);
	m+/\*+o && (s+/\*.*?\*/+ +go, (s+/\*.*$+ +o && ($fInComment = 1)));

	# Pick up definitions.
	if ( m/^\s*#/o )
	{
	    $iLinuxVersion     = $. if m/^\s*#\s*include\s*"linux\/version\.h"/o;
	}

	# Strip strings.
	$fInString && (s+^.*?"+ +o ? ($fInString = 0) : next);
	m+"+o && (s+".*?"+ +go, (s+".*$+ +o && ($fInString = 1)));

	# Pick up definitions.
	if ( m/^\s*#/o )
	{
	    $iLinuxVersion     = $. if m/^\s*#\s*include\s*<linux\/version\.h>/o;
	}

	# Look for usages.
	next unless m/LINUX_VERSION_CODE/o;
	$fUseKernCode = 1;
	last LINE if $iLinuxVersion;
	print "$file: $.: need linux/version.h\n";
	last LINE
    }

    # Report superfluous includes.
    if ( $iLinuxVersion && ! $fUseKernCode )
	{ print "$file: $iLinuxVersion: linux/version.h not needed.\n"; }

    close(FILE);
}


-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Russell King]

On Sat, Aug 03, 2002 at 12:56:26AM +0100, Russell King wrote:
> 	# Look for usages.
> 	next unless m/LINUX_VERSION_CODE/o;

It should probably look for the other two macros in linux/version.h as
well btw, so it could be generating false negatives.

-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Adam J. Richter]

On Sat, 3 Aug 2002 01:12:10 +0100, Russell King wrote:
>Ok, here's a fix for the 8250.c build problem (please don't send it
>to Linus; I've other changes that'll be going via BK and patch to
>lkml pending):
>
>--- orig/drivers/serial/8250.c  Fri Aug  2 21:13:31 2002
>+++ linux/drivers/serial/8250.c Sat Aug  3 00:28:47 2002
>@@ -31,7 +31,8 @@
> #include <linux/console.h>
> #include <linux/sysrq.h>
> #include <linux/serial_reg.h>
>-#include <linux/serialP.h>
>+#include <linux/circ_buf.h>
>+#include <linux/serial.h>
> #include <linux/delay.h>
> 
> #include <asm/io.h>

	Your patch still results in a compilation error for me.
It looks like 8250.c needs <linux/serialP.h> for ALPHA_KLUDGE_MCR:

  gcc -Wp,-MD,./.8250.o.d -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=i386 -nostdinc -iwithprefix include -DMODULE -include /usr/src/linux/include/linux/modversions.h   -DKBUILD_BASENAME=8250 -DEXPORT_SYMTAB  -c -o 8250.o 8250.c
8250.c: In function `serial8250_set_mctrl':
8250.c:1061: `ALPHA_KLUDGE_MCR' undeclared (first use in this function)
8250.c:1061: (Each undeclared identifier is reported only once
8250.c:1061: for each function it appears in.)

Adam J. Richter     __     ______________   575 Oroville Road
adam@yggdrasil.com     \ /                  Milpitas, California 95035
+1 408 309-6081         | g g d r a s i l   United States of America
                         "Free Software For The Rest Of Us."

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Kai Germaschewski]

On Sat, 3 Aug 2002, Russell King wrote:

> When I build 8250.c into the kernel, linux/module.h doesn't include
> linux/version.h, so when we include linux/serialP.h, the compiler
> assumes that LINUX_VERSION_CODE is zero.  So we end up including
> linux/serial.h.
> 
> However, when building as a module, linux/module.h does include
> linux/version.h, so when we don't include linux/serial.h.
> 
> Oh, the problems of trying to reduce the includes...  I think we should
> re-include linux/serial.h and eliminate linux/serialP.h.
> 
> Hmm, I wonder how many other oddities like this are in the tree today.
> It sounds like we want to create a rule similar to the one for using
> CONFIG_* symbols.  Does this sound reasonable: if you use
> LINUX_VERSION_CODE, you must include linux/version.h into that very
> same file to guarantee that it is defined.
> 
> Well, I took checkconfig.pl and created checkversion.pl (attached).
> Oh god, can I please put the worms back in the can?  Now?  I think
> there's lots of work to do here; lots of stuff including linux/version.h
> for the hell of it, and a comparitively small number not including it
> when they use LINUX_VERSION_CODE.
> 
> (No, I'm just off to zzz, so this must be a nightmare, maybe it'll go
> away by the time I wake up later today.) 8/

Hmmh, note that checkconfig.pl is actually not necessary anymore, it could 
be replaced by just including config.h unconditionally, or even include 
the macros from the commandline (-imacro include/linux/autoconf.h, I think 
you yourself did suggest that ;). The old build system relied on
#include <linux/config.h> to get "make dep" right, but that's not 
necessary anymore.

So I'm not sure introducing such a kind-of-weird rule for "version.h" is
the right way to go. An alternative would be to follow kaos' suggestion
and split version.h into uts_release.h (for the UTS_RELEASE macro) and
version.h for the numeric version. Then just include version.h from
kernel.h and be done with it.

--Kai

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Russell King]

On Fri, Aug 02, 2002 at 05:20:28PM -0700, Adam J. Richter wrote:
> On Sat, 3 Aug 2002 01:12:10 +0100, Russell King wrote:
> >Ok, here's a fix for the 8250.c build problem (please don't send it
> >to Linus; I've other changes that'll be going via BK and patch to
> >lkml pending):
> >
> >--- orig/drivers/serial/8250.c  Fri Aug  2 21:13:31 2002
> >+++ linux/drivers/serial/8250.c Sat Aug  3 00:28:47 2002
> >@@ -31,7 +31,8 @@
> > #include <linux/console.h>
> > #include <linux/sysrq.h>
> > #include <linux/serial_reg.h>
> >-#include <linux/serialP.h>
> >+#include <linux/circ_buf.h>
> >+#include <linux/serial.h>
> > #include <linux/delay.h>
> > 
> > #include <asm/io.h>
> 
> 	Your patch still results in a compilation error for me.
> It looks like 8250.c needs <linux/serialP.h> for ALPHA_KLUDGE_MCR:

Your quote above didn't include the patch for 8250.h which was in my
mail directly after 8250.c.  Did you specifically miss it for a reason?

-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

Re: Linux 2.5.30: [SERIAL] build fails at 8250.c

[Adam J. Richter]

On Sat, 3 Aug 2002 09:05:23 +0100, Russell King wrote:
>On Fri, Aug 02, 2002 at 05:20:28PM -0700, Adam J. Richter wrote:
>> On Sat, 3 Aug 2002 01:12:10 +0100, Russell King wrote:
>> >Ok, here's a fix for the 8250.c build problem (please don't send it
>> >to Linus; I've other changes that'll be going via BK and patch to
>> >lkml pending):
>> >
>> >--- orig/drivers/serial/8250.c  Fri Aug  2 21:13:31 2002
>> >+++ linux/drivers/serial/8250.c Sat Aug  3 00:28:47 2002
>> >@@ -31,7 +31,8 @@
>> > #include <linux/console.h>
>> > #include <linux/sysrq.h>
>> > #include <linux/serial_reg.h>
>> >-#include <linux/serialP.h>
>> >+#include <linux/circ_buf.h>
>> >+#include <linux/serial.h>
>> > #include <linux/delay.h>
>> > 
>> > #include <asm/io.h>
>> 
>> 	Your patch still results in a compilation error for me.
>> It looks like 8250.c needs <linux/serialP.h> for ALPHA_KLUDGE_MCR:

>Your quote above didn't include the patch for 8250.h which was in my
>mail directly after 8250.c.  Did you specifically miss it for a reason?

	Doh!  Sorry, my mistake.

Adam J. Richter     __     ______________   575 Oroville Road
adam@yggdrasil.com     \ /                  Milpitas, California 95035
+1 408 309-6081         | g g d r a s i l   United States of America
                         "Free Software For The Rest Of Us."