2.3.30-pre3

2.3.30-pre3

[Geert Uytterhoeven]

  - Plain 2.3.30-pre3 screws your .config because of two missing `fi's in the
    config scripts.
  - Kill one more warning

This does not mean 2.3.30-pre3 is running on my box, I'm just compiling...
Paul's unstable rsync tree (linux-pmac-devel) doesn't compile, so I wanted to
give Linus' tree a try

--- chrp-2.3.30/arch/ppc/config.in.orig	Sun Nov 28 20:29:50 1999
+++ chrp-2.3.30/arch/ppc/config.in	Sun Nov 28 22:31:04 1999
@@ -85,12 +85,16 @@
 
 if [ "$CONFIG_APUS" = "y" ]; then
   define_bool CONFIG_PCI n
-else if [ "$CONFIG_OAK" = "y" ]; then
-  define_bool CONFIG_PCI n
-else if [ "$CONFIG_8xx" = "y" ]; then
-    bool 'QSpan PCI' CONFIG_PCI
 else
-    define_bool CONFIG_PCI y
+  if [ "$CONFIG_OAK" = "y" ]; then
+    define_bool CONFIG_PCI n
+  else
+    if [ "$CONFIG_8xx" = "y" ]; then
+      bool 'QSpan PCI' CONFIG_PCI
+    else
+      define_bool CONFIG_PCI y
+    fi
+  fi
 fi
 
 bool 'Networking support' CONFIG_NET


--- chrp-2.3.30/arch/ppc/chrpboot/piggyback.c.orig	Tue Jan 13 00:18:13 1998
+++ chrp-2.3.30/arch/ppc/chrpboot/piggyback.c	Sun Nov 28 22:27:31 1999
@@ -1,8 +1,9 @@
 #include <stdio.h>
+#include <unistd.h>
 
 extern long ce_exec_config[];
 
-main(int argc, char *argv[])
+int main(int argc, char *argv[])
 {
 	int i, cnt, pos, len;
 	unsigned int cksum, val;

Gr{oetje,eeting}s,
--
Geert Uytterhoeven -- Linux/{m68k~Amiga,PPC~CHRP} -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[Shaw Terwilliger]

Geert Uytterhoeven wrote:
> This does not mean 2.3.30-pre3 is running on my box, I'm just compiling...
> Paul's unstable rsync tree (linux-pmac-devel) doesn't compile, so I wanted to
> give Linus' tree a try

I'm running Paul's unstable tree now, but it took a bit of hacking to make it
compile.  fs/hfs is still using the 2.2 VFS calls, so I just stole fs/hfs
from Linus's tree, stuck it in, and it works.  The Mac keyboard driver
was also all busted (quite a few half-written / missing functions), so I
just wrote them like I thought they should go, and I'm typing through the
keyboard right now.  

I think I might have also had to tweak some other part of the code (I'm
thinking atyfb, but I can't remember).

I can make diffs if anyone wants them; I'm sure some other people would
rather like to build from Paul's unstable tree.  :)  I'd been using
his stable tree, but for the last 3 months, I can't boot one of his
stable kernels without getting a machine check somewhere in init.
The unstable ones have no problems with it, although my USB mouse seems
to have stopped working... (so now I'm trying to make Xpmac use some
keys to emulate mouse buttons, but they seem hard-coded, so it'll be
a little harder than I thought).

-- 
Shaw Terwilliger (sterwill@io.nu)

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[Eric Benard]

Hi all,

is there someone here who manage to run a 2.3.X kernel on a B&W G3 ?

I ve also tried to put the 2.3.x usb stack on a 2.2.13 and this compiles well,
it boots well, but it fails to initialize the hub and leave the usb part with
an error (-4) and boots well ... without the keyboardnor the mous e;-))

If someone has got a hack ...

Eric
 
______________________________________________________________________
Message envoye depuis iFrance >>  http://www.ifrance.com
Gratuit >> Hebergement (50 Mo)/Vos emails (POP&HTML,20 Mo)
Votre agenda online gratuit >> http://agenda.ifrance.com
NOUVEAU : Faxez gratuitement ! >> http://fax.ifrance.com


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[BenH]

On Sun, Nov 28, 1999, Shaw Terwilliger <sterwill@io.nu> wrote:

>I can make diffs if anyone wants them; I'm sure some other people would
>rather like to build from Paul's unstable tree.  :)  I'd been using
>his stable tree, but for the last 3 months, I can't boot one of his
>stable kernels without getting a machine check somewhere in init.
>The unstable ones have no problems with it, although my USB mouse seems
>to have stopped working... (so now I'm trying to make Xpmac use some
>keys to emulate mouse buttons, but they seem hard-coded, so it'll be
>a little harder than I thought).

Yes, please, send me your patches. Paul will be away for a few weeks and
I want to move all my patches (there are more coming ;) to 2.3.x asap.

BTW, dmasound is broken in Paul's current stable 2.2.14pre. I don't know
yet what's going on but it looks like the driver had major changes
(backported sound in from 2.3.x ?). I can't get any sound output on my
wallstreet anymore. I'll try to look into this next week.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[Shaw Terwilliger]

[-- Attachment #1: Type: text/plain, Size: 662 bytes --]

BenH wrote:
> Yes, please, send me your patches. Paul will be away for a few weeks and
> I want to move all my patches (there are more coming ;) to 2.3.x asap.

OK... here they are.  They're against paul's devel tree as of an hour
ago.  I'm running a fresh-built kernel from a tree with these applied;
everything's working well.

Besides being attached here, the individual patches (and the README) are
available at http://www.abisource.com/~sterwill/patches/.

If anyone else is interested in bulding the kernel, just grab the
devel tree and do "patch -p0 < all.patch" from /usr/src (assuming
the tree is /usr/src/linux).

-- 
Shaw Terwilliger (sterwill@io.nu)

[-- Attachment #2: all.patch --]
[-- Type: text/plain, Size: 2543 bytes --]

--- linux.paul/CREDITS	Sat Oct 16 02:56:04 1999
+++ linux/CREDITS	Sun Nov 28 21:20:04 1999
@@ -1978,6 +1978,15 @@
 S: Seattle, WA  98195-1800
 S: USA
 
+N: Shaw Terwilliger
+E: sterwill@io.nu
+D: macintosh adb keyboard fixes
+D: ppc chipsfb misc fixes 
+D: misc hfs maintenance
+S: 1116 Broadmoor Drive #3
+S: Champaign, IL  61821
+S: USA
+
 N: Corey Thomas
 E: corey@world.std.com
 W: http://world.std.com/~corey/index.html
--- linux.paul/drivers/video/chipsfb.c	Sat Oct 16 05:05:00 1999
+++ linux/drivers/video/chipsfb.c	Sun Nov 28 20:25:32 1999
@@ -417,7 +417,7 @@
 	disp->visual = fix->visual;
 	disp->var = *var;
 
-#if (defined(CONFIG_PMAC_PBOOK) || defined(CONFIG_FB_COMPAT_XPMAC))
+#if (defined(CONFIG_FB_COMPAT_XPMAC))
 	display_info.depth = bpp;
 	display_info.pitch = fix->line_length;
 #endif
--- linux.paul/fs/hfs/file.c	Thu Jul  1 00:01:54 1999
+++ linux/fs/hfs/file.c	Fri Nov  5 17:53:26 1999
@@ -498,7 +498,6 @@
 				written = -EFAULT;
 			break;
 		}
-		update_vm_cache(inode,pos,p,c);
 		pos += c;
 		written += c;
 		buf += c;
--- linux.paul/include/asm/keyboard.h	Sun Nov 28 19:31:57 1999
+++ linux/include/asm/keyboard.h	Sun Nov 28 19:45:46 1999
@@ -27,14 +27,32 @@
 #define DISABLE_KBD_DURING_INTERRUPTS	0
 #define INIT_KBD
 
+/* 
+ * I'm not a PPC keyboard driver expert, but one day kbd_init_hw()
+ * disappeared from this file (include/asm-ppc/keyboard.h), so I
+ * cloned it from a version found in a 2.2 kernel.  It works for me.
+ * - Shaw Terwilliger <sterwill@io.nu>
+ */
+
+static inline void kbd_init_hw(void)
+{
+#ifdef CONFIG_PREP
+	if (prep_kbd_present)
+#endif /* CONFIG_PREP */
+		ppc_md.kbd_init_hw();
+}
+
 static inline int kbd_setkeycode(unsigned int scancode, unsigned int keycode)
 {
+	/*
+	 * There used to be a stray "else" clause here that might
+	 * have been a left-over fragment from a copy-and-paste.
+	 * - Shaw Terwilliger <sterwill@io.nu>
+	 */
+	   
 	if ( ppc_md.kbd_setkeycode )
 		return ppc_md.kbd_setkeycode(scancode, keycode);
 	else
-		retureycode )
-		return ppc_md.kbd_setkeycode(scancode, keycode);
-	else
 		return 0;
 }
   
@@ -65,9 +83,16 @@
   
 static inline void kbd_leds(unsigned char leds)
 {
+	/*
+	 * The commented line below was, at one time, part of 
+	 * the code flow.  I've translated it to what I think its author
+	 * meant.
+	 * - Shaw Terwilliger <sterwill@io.nu>
+	 */
+
 	if ( ppc_md.kbd_leds )
-		ppc_md.kbd_ledt_hw )
-		ppc_md.kbd_init_hw();
+	/*	ppc_md.kbd_ledt_hw ) */
+		ppc_md.kbd_leds(leds);
 }
 
 #define kbd_sysrq_xlate	(ppc_md.ppc_kbd_sysrq_xlate)

[-- Attachment #3: README --]
[-- Type: text/plain, Size: 1643 bytes --]

Available here are patches to Paul Mackerras's latest (as of Sunday
November 28, 1999, CST) development (unstable) PPC Linux kernel mirror.
Get the code with "rsync -arvuz samba.anu.edu.au::linux-pmac-devel <tgtdir>"
where <tgtdir> is where you want your source tree to end up.

Then apply these patches:

all.patch:
	All of the below in one hunk.

chipsfb.c.patch:

	[linux/drivers/video/chipsfb.c]

	Changed chipsfb driver to store some display_info data only
	when CONFIG_FB_COMPAT_XPMAC is defined (and not as it was previously
	enabled--when _either_ CONFIG_PMAC_PBOOK or ...COMPAT_XPMAC).

	*** THIS PATCH IS UNTESTED *** ... I made these changes so I could
	compile chipsfb.c with my configuration, which enables 
	CONFIG_PMAC_PBOOK, but does not need Xpmac compatability.

hfs.file.c.patch:

	[linux/fs/hfs/file.c]

	Synced hfs code with code from Linus's 2.3.28 tree, which simply
	removes the call to update_vm_cache().

	*** HFS SEEMS BROKEN *** ... HFS works for _simple_ operations	
	with the kernels built from this code.  Mounts (ro and rw) work,
	file cp works, things like mv do NOT.  Don't try it unless you 
	really like xmon.

	I don't know if HFS has just been neglected through kernel 2.3,
	or if it's been this way for a while.

keyboard.h.patch:

	[linux/include/asm-ppc/keyboard.h]

	My version of this file came in just about half-baked.
	Added missing kbd_init_hw() function, cleaned up some
	stray chunks of code elsewhere in the file.

CREDITS.patch:

	Just tacked myself in there for the changes to keyboard and
	HFS, both of which I'd like to devote some of my time to 
	make work reliably on my PowerBook.  :)

Re: 2.3.30-pre3

[Martin Costabel]

Shaw Terwilliger wrote:
> 
> BenH wrote:
> > Yes, please, send me your patches. Paul will be away for a few weeks and
> > I want to move all my patches (there are more coming ;) to 2.3.x asap.
> 
> OK... here they are.  They're against paul's devel tree as of an hour
> ago.  I'm running a fresh-built kernel from a tree with these applied;
> everything's working well.
> 
> Besides being attached here, the individual patches (and the README) are
> available at http://www.abisource.com/~sterwill/patches/.
> 
> If anyone else is interested in bulding the kernel, just grab the
> devel tree and do "patch -p0 < all.patch" from /usr/src (assuming
> the tree is /usr/src/linux).

<???> 
Are there several different versions of Paul's devel tree? The one I am
getting is only at version 2.3.29 and has been nicely compiling and
running for me for a while now. I don't know about chipsfb.c which I am
not using, but the others of your patches are essentially those that A.
Joshi has published a long time ago. And they don't apply to my version
of Paul's devel tree, because they are already built in. 
</???>

--
Martin

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[Paul Mackerras]

On Mon, 29 Nov 1999, Martin Costabel wrote:

> Are there several different versions of Paul's devel tree? The one I am
> getting is only at version 2.3.29 and has been nicely compiling and
> running for me for a while now. I don't know about chipsfb.c which I am
> not using, but the others of your patches are essentially those that A.
> Joshi has published a long time ago. And they don't apply to my version
> of Paul's devel tree, because they are already built in. 

This is (partly at least) my fault.  I upgraded to 2.3.30-pre3 (a merge of
my tree and Cort's), found it didn't work, and reverted the devel tree to
2.3.29, which does work just fine for me.

I think that Shaw possibly got a corrupted include/asm-ppc/keyboard.h,
maybe I was updating it as he was fetching it or something.  I have
changed the ifdef in chipsfb.c as Shaw suggested.

Paul.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[Shaw Terwilliger]

Paul Mackerras wrote:
> This is (partly at least) my fault.  I upgraded to 2.3.30-pre3 (a merge of
> my tree and Cort's), found it didn't work, and reverted the devel tree to
> 2.3.29, which does work just fine for me.

Well, I was just using the old rsync repository at samba.anu.edu.au, not
at linuxcare.com.au.  However, the keyboard.h file had been that
way for a few weeks, probably unnoticed while everyone used the newer
repository.  I did a little patching and that kernel's running fine
on my laptop.

I would ask that perhaps samba.anu.edu.au::linux-pmac-[stable/devel]
be turned off (or print a nice MOTD with the new location) since there 
are countless web pages out there that still point to it.

> maybe I was updating it as he was fetching it or something.  I have
> changed the ifdef in chipsfb.c as Shaw suggested.

It looked right to me, but I didn't have a chips device to test it with.
Just seemed like an unnecessary test and kept me (and presumably others)
from compiling it.

-- 
Shaw Terwilliger (sterwill@io.nu)

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: 2.3.30-pre3

[Ani Joshi]

On Wed, 1 Dec 1999, Paul Mackerras wrote:

> I think that Shaw possibly got a corrupted include/asm-ppc/keyboard.h,
> maybe I was updating it as he was fetching it or something.  I have
> changed the ifdef in chipsfb.c as Shaw suggested.

	actually, Shaw's chips fb patch is wrong, in the sense that it
isn't a complete fix.  as Martin noted, i "fixed" chipsfb a while ago back
in 2.3.25 or so, there are some structs which are undefined which are used
in xpmac_compat code, please refer to the 2.3.25-ppc.patch in my directory
on devel, devel.linuxppc.org/users/ajoshi for a proper chipsfb patch.



ani


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/