Re: Syncin with Git

On Thu, Sep 28, 2006 at 12:55:01PM +0900, Paul Mundt wrote:
[...]
> Given that, I would recommend that people start testing current git so
> we can get the rest of the mess cleaned up. I'll be killing the arch and
> CVS trees in short order.

I've just booted 2.6.18-git18 on the titan board:

 <5>Linux version 2.6.18-git18 (lenehan <at> nynaeve.twibble.org) (gcc version 3.4.4) #3 Tue Oct 3 11:12:39
EST 2006

and everything looks to be working fine. Nice.

Only two minor changes were needed. I deleted the two lines in
arch/sh/kernel/entry.S that were defining errno values. These must be
getting defined prior to their definition in entry.S via one of the
included files:

    sh4-linux-gcc -Wp,-MD,arch/sh/kernel/.entry.o.d  -nostdinc -isystem
/data/oe/build/titan-glibc-24/tmp/cross/lib/gcc/sh4-linux/3.4.4/include -D__KERNEL__
-Iinclude  -include include/linux/autoconf.h -D__ASSEMBLY__ -ml -Wa,-isa=sh4 -m4 -m4-nofpu    -c -o
arch/sh/kernel/entry.o arch/sh/kernel/entry.S
  arch/sh/kernel/entry.S: Assembler messages:
  arch/sh/kernel/entry.S:57: Error: junk at end of line, first unrecognized character is `3'
  arch/sh/kernel/entry.S:58: Error: junk at end of line, first unrecognized character is `2'
  make[1]: *** [arch/sh/kernel/entry.o] Error 1
  make: *** [arch/sh/kernel] Error 2

And I added a #include <linux/fs.h> to get the definition of
MS_RDONLY to arch/sh/kernel/setup.c:

Re: Syncin with Git

On Tue, Oct 03, 2006 at 04:38:21PM +1000, Jamie Lenehan wrote:
> I've just booted 2.6.18-git18 on the titan board:
> 
>  <5>Linux version 2.6.18-git18 (lenehan <at> nynaeve.twibble.org) (gcc version 3.4.4) #3 Tue Oct 3
11:12:39 EST 2006
> 
> and everything looks to be working fine. Nice.
> 
> Only two minor changes were needed. I deleted the two lines in
> arch/sh/kernel/entry.S that were defining errno values. These must be
> getting defined prior to their definition in entry.S via one of the
> included files:
> 
[snip]

> And I added a #include <linux/fs.h> to get the definition of
> MS_RDONLY to arch/sh/kernel/setup.c:
> 
Yes, I caught these this morning, and prepared a pull for Linus. If you
want to make sure that everything is still sensible for titan, you can
feel free to pull directly from:

	git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git

This should be sorted out by -rc1 at least.

> The config used was:
> 
>   http://twibble.org/dist/misc/titan-config-2.6.18-git18
> 

Re: Syncin with Git

On Tue, Oct 03, 2006 at 04:13:00PM +0900, Paul Mundt wrote:
> If you're feeling more adventurous, it would also be good to have some
> -mm testing. These updates will be auto-pulled in to 2.6.18-mm3, so that
> might be a good time to try it out.
> 
.. ok, make that -mm4 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Syncin with Git

On Tue, Oct 03, 2006 at 04:13:00PM +0900, Paul Mundt wrote:
[...]
> Yes, I caught these this morning, and prepared a pull for Linus. If you
> want to make sure that everything is still sensible for titan, you can
> feel free to pull directly from:
> 
> 	git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git

About time I started to use git I guess... I'll look into getting git
working and then test this out.

[...]
> > The config used was:
> > 
> >   http://twibble.org/dist/misc/titan-config-2.6.18-git18
> > 
> This config is slightly different from the one I have in configs/,
> mostly in relation to network settings. Did you want this config as the
> defacto standard, or is the one in the tree now ok?

Ummm, no - keep the current one. It looks like I generated this from
a slightly an older config. I just supplied it in case the errors
were configuration specific. I should run the current default through
make oldconfig to update it. How often do you think these defconfigs
should be updated? In most cases it'll just be answering all the new
questions.

> If you're feeling more adventurous, it would also be good to have some
> -mm testing. These updates will be auto-pulled in to 2.6.18-mm3, so that

Re: Syncin with Git

On Tue, Oct 03, 2006 at 05:39:49PM +1000, Jamie Lenehan wrote:
> On Tue, Oct 03, 2006 at 04:13:00PM +0900, Paul Mundt wrote:
> [...]
> > Yes, I caught these this morning, and prepared a pull for Linus. If you
> > want to make sure that everything is still sensible for titan, you can
> > feel free to pull directly from:
> > 
> > 	git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git
> 
> About time I started to use git I guess... I'll look into getting git
> working and then test this out.
> 
You can either clone this or Linus's tree directly and then pull from
mine, or setup a branch for it or whatever else. Though I'm not
intending that people use the sh-2.6.git tree in the general case, as
about the only time I'll be updating it will be when I have new changes
to merge in to current git via the pull tree.

> I should run the current default through make oldconfig to update it.
> How often do you think these defconfigs should be updated? In most
> cases it'll just be answering all the new questions.
> 
I've got some scripts now that run through all of them for updating and
building to make sure things are sensible, I'm planning on automating
these for nightly builds, though I don't forsee submitting defconfig
updates too frequently, oldconfig usually does the right thing anyways.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your

Needing an Up-to-date Howto for Dreamcast

It seems the only real how to for Linux and the Dreamcast is the Bill
Gatliff article (http://www.linuxdevices.com/articles/AT7466555948.html)
which is 5 years old and only covers the 2.4 series. I ask this because
I've had problems with booting 2.6 kernels on my Dreamcast. (It either
reboots the Dreamcast or does nothing that I can tell.)

The things it should cover IMO would be:
-How to use recent versions of the kernel
-A brief section on the cross compiler (if it's better to build one from
glibc or uclibc for instance. If uclibc, is it better to build it
yourself or use buildroot. And anything else that may be a source of
problems.)
-How to  boot kernels (is it still required to burn a new CD for every
kernel change, or can you use an uploader CD like dc-tools or dc-slave.)
-And any other pointers you may have. (Maybe common problems and fixes
to those problems.)
-Maybe even debugging techniques (be it remote or local)

Would anyone be willing to help out here?

Git changes & Development board (sh3)

Greetings,

Glad to see git merging is progressing well. I expect to try out the Jornada 
6xx support shortly.
I know I asked this a few months ago but anyone wanting to sell a sh3 
development board, drop me a mail.

Best wishes
Kristoffer

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Need for an Up-to-date Howto for Dreamcast

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
linuxsh-dev mailing list
linuxsh-dev <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxsh-dev

Re: Syncin with Git

On Tue, Oct 03, 2006 at 04:51:21PM +0900, Paul Mundt wrote:
[...]
> You can either clone this or Linus's tree directly and then pull from
> mine, or setup a branch for it or whatever else. Though I'm not
> intending that people use the sh-2.6.git tree in the general case, as
> about the only time I'll be updating it will be when I have new changes
> to merge in to current git via the pull tree.

Yep, I went with a clone of Linus's tree, which I think makes more
sense to follow in general. Anyway, we have a new problem today:

    sh4-linux-gcc -Wp,-MD,arch/sh/drivers/pci/.pci-sh7751.o.d  -nostdinc -isystem
/data/oe/build/titan-glibc-24/tmp/cross/lib/gcc/sh4-linux/3.4.4/include -D__KERNEL__
-Iinclude -Iinclude2 -I/usr/src/git/linux-2.6/include -include include/linux/autoconf.h
-I/usr/src/git/linux-2.6/arch/sh/drivers/pci -Iarch/sh/drivers/pci -Wall -Wundef
-Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing -fno-common -O2 -pipe -ml -Wa,-isa=sh4
-m4 -m4-nofpu -fomit-frame-pointer -Wdeclaration-after-statement  -D"KBUILD_STR(s)=#s"
-D"KBUILD_BASENAME=KBUILD_STR(pci_sh7751)" -D"KBUILD_MODNAME=KBUILD_STR(pci_sh7751)" -c -o
arch/sh/drivers/pci/pci-sh7751.o /usr/src/git/linux-2.6/arch/sh/drivers/pci/pci-sh7751.c
  /usr/src/git/linux-2.6/arch/sh/drivers/pci/pci-sh7751.c: In function `sh7751_pcic_init':
  /usr/src/git/linux-2.6/arch/sh/drivers/pci/pci-sh7751.c:158: error: `SH4_PCI_IO_BASE'
undeclared (first use in this function)
  /usr/src/git/linux-2.6/arch/sh/drivers/pci/pci-sh7751.c:158: error: (Each undeclared
identifier is reported only once
  /usr/src/git/linux-2.6/arch/sh/drivers/pci/pci-sh7751.c:158: error: for each function it appears in.)
  make[3]: *** [arch/sh/drivers/pci/pci-sh7751.o] Error 1
  make[2]: *** [arch/sh/drivers/pci] Error 2
  make[1]: *** [arch/sh/drivers] Error 2
  make: *** [_all] Error 2

SH4_PCI_IO_BASE isn't defined anywhere. Previously pr_debug was
defined as an emtpy macro but that was changed by this:

  commit 8b2a1fd1b394c60eaa2587716102dd5e9b4e5990
  Author: Zach Brown <zach.brown <at> oracle.com>
  Date:   Tue Oct 3 01:16:15 2006 -0700

      [PATCH] pr_debug: check pr_debug() arguments

      check pr_debug() arguments

      When DEBUG isn't defined pr_debug() is defined away as an empty macro.  By
      throwing away the arguments we allow completely incorrect code to build.

      Instead let's make it an empty inline which checks arguments and mark it so gcc
      can check the format specification.
  ...

I just changed this to SH7751_PCI_IO_BASE since this is in pci-sh7751.c:

diff --git a/arch/sh/drivers/pci/pci-sh7751.c b/arch/sh/drivers/pci/pci-sh7751.c
index dbe8378..85e1ee2 100644
--- a/arch/sh/drivers/pci/pci-sh7751.c
+++ b/arch/sh/drivers/pci/pci-sh7751.c
 <at>  <at>  -155,7 +155,7  <at>  <at>  int __init sh7751_pcic_init(struct sh4_p
 	 */
 	pr_debug("PCI: Mapping IO address 0x%x - 0x%x to base 0x%x\n",
 		 PCIBIOS_MIN_IO, (64 << 10),
-		 SH4_PCI_IO_BASE + PCIBIOS_MIN_IO);
+		 SH7751_PCI_IO_BASE + PCIBIOS_MIN_IO);

 	/*
 	 * XXX: For now, leave this board-specific. In the event we have other
 <at>  <at>  -163,7 +163,7  <at>  <at>  int __init sh7751_pcic_init(struct sh4_p
 	 */
 #ifdef CONFIG_SH_BIGSUR
 	bigsur_port_map(PCIBIOS_MIN_IO, (64 << 10),
-			SH4_PCI_IO_BASE + PCIBIOS_MIN_IO, 0);
+			SH7751_PCI_IO_BASE + PCIBIOS_MIN_IO, 0);
 #endif

 	/* Make sure the MSB's of IO window are set to access PCI space

With that it builds and boots fine.

--

-- 
 Jamie Lenehan <lenehan <at> twibble.org>

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

[PATCH] RTC new I/F support in landisk and etc.

Hi Paul-san

Please apply following patch.

(1)RTC RICHO RS5C313
I prepared the RTC patch of LANDISK that suited a new interface specification.

Patch and source
 http://eggplant.ddo.jp/www/download/kernel26/patch-2.6.18-git22/rtc_newif.patch
 http://eggplant.ddo.jp/www/download/kernel26/patch-2.6.18-git22/landisk_defconfig

(2)arch/sh/boards/landisk/landisk_pwc.c
 The name of landisk_pwc.c is changed to landisk_drv.c.
 The automatic allocation function of the driver number is added.
 The button detection routine was rewritten to "Select/poll" correspondence.

Patch and source
 http://eggplant.ddo.jp/www/download/kernel26/patch-2.6.18-git22/button_led_etc.patch

(3)The kexec function doesn't work
When the following patch is appropriated, it comes to operate correctly.

diff -urpN linux-2.6.18-git22/arch/sh/kernel/process.c linux-2.6.18-git22mod/arch/sh/kernel/process.c
--- linux-2.6.18-git22/arch/sh/kernel/process.c	2006-10-05 09:31:36.000000000 +0900
+++ linux-2.6.18-git22mod/arch/sh/kernel/process.c	2006-10-05 13:57:57.000000000 +0900
 <at>  <at>  -81,6 +81,16  <at>  <at>  void cpu_idle(void)

 void machine_restart(char * __unused)
 {
+
+#ifdef CONFIG_KEXEC
+	struct kimage *image;
+	image = xchg(&kexec_image, 0);
+	if (image) {
+		machine_shutdown();
+		machine_kexec(image);
+	}
+#endif
+
 	/* SR.BL=1 and invoke address error to let CPU reset (manual reset) */
 	asm volatile("ldc %0, sr\n\t"
 		     "mov.l  <at> %1, %0" : : "r" (0x10000000), "r" (0x80000001));

kogiidena

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV