DECstation R3000 boot error

DECstation R3000 boot error

[Martin Michlmayr]

We're getting the following boot error on a DECstation with R3K CPU.
It simply hangs after the "high precision timer" message.  Maciej, do
you have some time to look into this issue, or does anyone else have
any idea what's going on there?

This is with current git.


Linux version 2.6.15 (tbm@deprecation) (gcc version 4.0.3 20051201 (prerelease) (Debian 4.0.2-5)) #2 Mon Jan 23 12:41:27 GMT 2006
This is a DECstation 5000/2x0
CPU revision is: 00000230
FPU revision is: 00000340
Determined physical RAM map:
 memory: 0c000000 @ 00000000 (usable)
Built 1 zonelists
Kernel command line: rd_start=2154876928 rd_size=
Primary instruction cache 64kB, linesize 4 bytes.
Primary data cache 64kB, linesize 4 bytes.
Synthesized TLB refill handler (17 instructions).
Synthesized TLB load handler fastpath (31 instructions).
Synthesized TLB store handler fastpath (31 instructions).
Synthesized TLB modify handler fastpath (25 instructions).
PID hash table entries: 1024 (order: 10, 16384 bytes)
Using 24.999 MHz high precision timer.�

-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Kaj-Michael Lang]

On Mon, 23 Jan 2006, Martin Michlmayr wrote:

> We're getting the following boot error on a DECstation with R3K CPU.
> It simply hangs after the "high precision timer" message.  Maciej, do

It's trying to use a timer source that does not exist on the DECstation 
you're trying to use, most likely. What model is it you have ? I've 
tracked it down when I got my first (working) DECstation (5000/133) a
month or so ago. Forgot all about it under the holidays. I think
I have a patch somewhere to fix it, I'll do some digging.

-- 
Kaj-Michael Lang

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 613 bytes --]

On Mon, 23 Jan 2006, Martin Michlmayr wrote:

> We're getting the following boot error on a DECstation with R3K CPU.
> It simply hangs after the "high precision timer" message.  Maciej, do
> you have some time to look into this issue, or does anyone else have
> any idea what's going on there?
[...]
> Using 24.999 MHz high precision timer.ďż˝

 It looks like a problem with initialization of the serial driver -- this 
is where the console is being set up.  See also the rubbish at the end -- 
that's a hint.  I am assuming you've got your console configuration right.

 I'll see if I can reproduce it.

  Maciej

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Tue, 24 Jan 2006, Kaj-Michael Lang wrote:

> > We're getting the following boot error on a DECstation with R3K CPU.
> > It simply hangs after the "high precision timer" message.  Maciej, do
> 
> It's trying to use a timer source that does not exist on the DECstation you're
> trying to use, most likely. What model is it you have ? I've tracked it down

 Well, I don't think it's possible to calculate frequency of something 
that's inexistent, yet it does...  That's a /240.

> when I got my first (working) DECstation (5000/133) a
> month or so ago. Forgot all about it under the holidays. I think
> I have a patch somewhere to fix it, I'll do some digging.

 The /133 (as all 3MINs) may have an older revision of the I/O ASIC that 
may not have the free-running counter indeed.  It's handled correctly 
regardless, except from missing timestamp precision, obviously.

 What kind of a patch do you need anyway and why isn't it yet in my mail?

  Maciej

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Maciej W. Rozycki <macro@linux-mips.org> [2006-01-24 11:09]:
>  It looks like a problem with initialization of the serial driver -- this 
> is where the console is being set up.  See also the rubbish at the end -- 
> that's a hint.  I am assuming you've got your console configuration right.

We had the following settings: console=s and osconsole=3
The terminal emulation program should've been configured properly
too.

-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Tue, 24 Jan 2006, Martin Michlmayr wrote:

> We had the following settings: console=s and osconsole=3
> The terminal emulation program should've been configured properly
> too.

 Well, the console device was not specified at the Linux's command line 
and the kernel currently does not query REX variables to infer which 
device to use.  Please retry with "console=ttyS2"; you may specify line 
parameters as required, too.  That may be compiled in if you run out of 
the BBU RAM space (which only allows up to 37 characters for kernel 
arguments).

 The output you got was produced with the early console which uses REX 
primitives.

  Maciej

Re: DECstation R3000 boot error

[Peter 'p2' De Schrijver]

[-- Attachment #1: Type: text/plain, Size: 3824 bytes --]

Hi,

> 
>  Well, the console device was not specified at the Linux's command line 
> and the kernel currently does not query REX variables to infer which 
> device to use.  Please retry with "console=ttyS2"; you may specify line 
> parameters as required, too.  That may be compiled in if you run out of 
> the BBU RAM space (which only allows up to 37 characters for kernel 
> arguments).
> 
>  The output you got was produced with the early console which uses REX 
> primitives.

Ok. This results in the following panic :

>>boot 3/tftp/vmlinux-pmax console=ttyS2
 
 -tftp boot(3), bootp 10.42.0.190:/tftpboot/vmlinux-pmax
 -tftp load 3080326+0+114586
 Linux version 2.6.15 (tbm@deprecation) (gcc version 4.0.3 20051201
 (prerelease) (Debian 4.0.2-5)) #2 Tue Jan 24 23:08:47 GMT 2006
 This is a DECstation 5000/2x0
 CPU revision is: 00000230
 FPU revision is: 00000340
 Determined physical RAM map:
memory: 0c000000 @ 00000000 (usable)
Built 1 zonelists
Kernel command line: console=ttyS2
Primary instruction cache 64kB, linesize 4 bytes.
Primary data cache 64kB, linesize 4 bytes.
Synthesized TLB refill handler (17 instructions).
Synthesized TLB load handler fastpath (31 instructions).
Synthesized TLB store handler fastpath (31 instructions).
Synthesized TLB modify handler fastpath (25 instructions).
PID hash table entries: 1024 (order: 10, 16384 bytes)
Using 24.999 MHz high precision timer.}SZ·]Œ0 @ 00000000 (usable)
Built 1 zonelists
Kernel command line: console=ttyS2
Primary instruction cache 64kB, linesize 4 bytes.
Primary  cache 64kB, linesize 4 bytes.
Synthesized TLB refill handler (17 instructions).
Synthesized TLB load handler fastpath (31 instructions).
Synthesized TLB store handler fastpath (31 instructions).
Synthesized TLB modify handler fastpath (25 instructions).
PID hash table entries: 1024 (order: 10, 16384 bytes)
Using 24.999 MHz high precision timer.
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Reserved instruction in kernel code[#1]:
Cpu 0
$ 0   : 00000000 1008ec00 80000000 00000010
$ 4   : 8034e000 00000001 8034f000 00000000
$ 8   : 0000bae0 802e1740 00000044 ffffc000
$12   : 00000078 81178018 802e17f4 00000005
$16   : 810069c0 00000001 810069c0 80340000
$20   : 00000000 1008ec01 00000001 00000001
$24   : 802a69d8 00000000
$28   : 802dc000 802ddea0 00029cf1 8009a88c
Hi    : 00000000
Lo    : 00000000
epc   : 80332384 0x80332384     Not tainted
ra    : 8009a88c get_page_from_freelist+0x4b0/0x548
Status: 1008ec04    IEp
Cause : 30000028
PrId  : 00000230
Modules linked in:
Process swapper (pid: 0, threadinfo=802dc000, task=802de000)
Stack : 00000048 8117a018 802e17c4 00000005 00000044 8117a800 00000000 00000000
        00000000 802e1be0 802e1be0 802de000 000080d0 a000fcec 00000000 802e1be0
        00000000 fdbcbd7e 00000000 8009a9b8 000280d0 00000000 00000000 00000044
        8031cd84 8031cd4c 00000010 00000000 000280d0 00000000 00000080 802ddf18
        00000000 80340000 80340000 a000fcec 00000000 00000003 00000000 fdbcbd7e
        ...
Call Trace:
 [<8009a9b8>] __alloc_pages+0x94/0x3cc
 [<8031cd84>] __free_pages_bootmem+0x128/0x144
 [<8031cd4c>] __free_pages_bootmem+0xf0/0x144
 [<8009b440>] __get_free_pages+0x3c/0x8c
 [<8004e414>] setup_zero_pages+0x34/0x1ac
 [<80313b74>] mem_init+0x5c/0x1ac
 [<8031ed8c>] inode_init_early+0xa8/0xc0
 [<8030c440>] start_kernel+0x140/0x21c
 [<8030c000>] _sinittext+0x0/0x7c
 [<8030c448>] start_kernel+0x148/0x21c
 [<8030c530>] unknown_bootoption+0x0/0x374


Code: 00000000  00000000  24861000 <fc800000> fc800008  fc800010 fc800018  24840040  fc80ffe0
Kernel panic - not syncing: Attempted to kill the idle task!

L & L

p2.

-- 
goa is a state of mind

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

Re: DECstation R3000 boot error

[Peter 'p2' De Schrijver]

[-- Attachment #1: Type: text/plain, Size: 1860 bytes --]

Hi,

> Hi,
> 
> Cpu 0
> $ 0   : 00000000 1008ec00 80000000 00000010
> $ 4   : 8034e000 00000001 8034f000 00000000
> $ 8   : 0000bae0 802e1740 00000044 ffffc000
> $12   : 00000078 81178018 802e17f4 00000005
> $16   : 810069c0 00000001 810069c0 80340000
> $20   : 00000000 1008ec01 00000001 00000001
> $24   : 802a69d8 00000000
> $28   : 802dc000 802ddea0 00029cf1 8009a88c
> Hi    : 00000000
> Lo    : 00000000
> epc   : 80332384 0x80332384     Not tainted
> ra    : 8009a88c get_page_from_freelist+0x4b0/0x548
> Status: 1008ec04    IEp
> Cause : 30000028
> PrId  : 00000230
> Modules linked in:
> Process swapper (pid: 0, threadinfo=802dc000, task=802de000)
> Stack : 00000048 8117a018 802e17c4 00000005 00000044 8117a800 00000000 00000000
>         00000000 802e1be0 802e1be0 802de000 000080d0 a000fcec 00000000 802e1be0
>         00000000 fdbcbd7e 00000000 8009a9b8 000280d0 00000000 00000000 00000044
>         8031cd84 8031cd4c 00000010 00000000 000280d0 00000000 00000080 802ddf18
>         00000000 80340000 80340000 a000fcec 00000000 00000003 00000000 fdbcbd7e
>         ...
> Call Trace:
>  [<8009a9b8>] __alloc_pages+0x94/0x3cc
>  [<8031cd84>] __free_pages_bootmem+0x128/0x144
>  [<8031cd4c>] __free_pages_bootmem+0xf0/0x144
>  [<8009b440>] __get_free_pages+0x3c/0x8c
>  [<8004e414>] setup_zero_pages+0x34/0x1ac
>  [<80313b74>] mem_init+0x5c/0x1ac
>  [<8031ed8c>] inode_init_early+0xa8/0xc0
>  [<8030c440>] start_kernel+0x140/0x21c
>  [<8030c000>] _sinittext+0x0/0x7c
>  [<8030c448>] start_kernel+0x148/0x21c
>  [<8030c530>] unknown_bootoption+0x0/0x374
> 
> 
> Code: 00000000  00000000  24861000 <fc800000> fc800008  fc800010 fc800018  24840040  fc80ffe0
> Kernel panic - not syncing: Attempted to kill the idle task!
> 

This is on a Decstation 5000/240.

L & L

p2.

-- 
goa is a state of mind

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

Re: DECstation R3000 boot error

[Kaj-Michael Lang]

On Tue, 24 Jan 2006, Maciej W. Rozycki wrote:

> On Tue, 24 Jan 2006, Kaj-Michael Lang wrote:
>
> The /133 (as all 3MINs) may have an older revision of the I/O ASIC that
> may not have the free-running counter indeed.  It's handled correctly
> regardless, except from missing timestamp precision, obviously.

Sorry, my bad. This problem is under 2.4, not 2.6. 2.6 has other problems.
And yes, the 3MIN does not have a free-running counter in the ASIC.
(http://mail-index.netbsd.org/port-pmax/1995/01/28/0006.html)

> What kind of a patch do you need anyway and why isn't it yet in my mail?

Do still you take patches for 2.4 ? I'll try to find it.

-- 
Kaj-Michael Lang

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Kaj-Michael Lang <milang@tal.org> [2006-01-25 12:42]:
> Sorry, my bad. This problem is under 2.4, not 2.6. 2.6 has other problems.

Which problems does 2.6 have?
-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Wed, 25 Jan 2006, Peter 'p2' De Schrijver wrote:

> Code: 00000000  00000000  24861000 <fc800000> fc800008  fc800010 fc800018  24840040  fc80ffe0
> Kernel panic - not syncing: Attempted to kill the idle task!

 Hmm, the code is:

00000000	nop
00000000	nop
24861000	addiu	a2,a0,4096
fc800000	sd	zero,0(a0)
fc800008	sd	zero,8(a0)
fc800010	sd	zero,16(a0)
fc800018	sd	zero,24(a0)
24840040	addiu	a0,a0,64
fc80ffe0	sd	zero,-32(a0)

and it obviously cannot work on a MIPS I processor.  That's probably from 
broken assembly code somewhere -- I should have my sources updated within 
a few days and I'll see if I can reproduce the problem.

  Maciej

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Wed, 25 Jan 2006, Kaj-Michael Lang wrote:

> And yes, the 3MIN does not have a free-running counter in the ASIC.
> (http://mail-index.netbsd.org/port-pmax/1995/01/28/0006.html)

 Well, that mail doesn't actually say anything about the 3MIN, but you can 
take my word anyway.

> Do still you take patches for 2.4 ? I'll try to find it.

 Possibly -- it depends on what the problem is.  If it stops a system from 
working, then certainly.

  Maciej

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Maciej W. Rozycki <macro@linux-mips.org> [2006-01-25 11:10]:
> and it obviously cannot work on a MIPS I processor.  That's probably from 
> broken assembly code somewhere -- I should have my sources updated within 
> a few days and I'll see if I can reproduce the problem.

Did you have a chance to look at this?
-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Fri, 3 Feb 2006, Martin Michlmayr wrote:

> > and it obviously cannot work on a MIPS I processor.  That's probably from 
> > broken assembly code somewhere -- I should have my sources updated within 
> > a few days and I'll see if I can reproduce the problem.
> 
> Did you have a chance to look at this?

 Not yet, sorry.  I have updated my tree and most of my local patches, but 
I have a few to go yet.  My time is limited these days, but I'll try hard 
to get at the problems you reported by the coming weekend at the very 
latest.

  Maciej

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Mon, 6 Feb 2006, Maciej W. Rozycki wrote:

>  Not yet, sorry.  I have updated my tree and most of my local patches, but 
> I have a few to go yet.  My time is limited these days, but I'll try hard 
> to get at the problems you reported by the coming weekend at the very 
> latest.

 The following change fixes the problem for me -- it looks like an 
omission from some recent changes elsewhere.  Ralf, please apply.

  Maciej

Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>

patch-mips-2.6.15-20060124-mips-cpu-isa-0
diff -up --recursive --new-file linux-mips-2.6.15-20060124.macro/include/asm-mips/cpu.h linux-mips-2.6.15-20060124/include/asm-mips/cpu.h
--- linux-mips-2.6.15-20060124.macro/include/asm-mips/cpu.h	2005-12-10 05:58:16.000000000 +0000
+++ linux-mips-2.6.15-20060124/include/asm-mips/cpu.h	2006-02-12 02:34:24.000000000 +0000
@@ -204,9 +204,9 @@
  */
 #define MIPS_CPU_ISA_I		0x00000001
 #define MIPS_CPU_ISA_II		0x00000002
-#define MIPS_CPU_ISA_III	0x00000003
-#define MIPS_CPU_ISA_IV		0x00000004
-#define MIPS_CPU_ISA_V		0x00000005
+#define MIPS_CPU_ISA_III	0x00000004
+#define MIPS_CPU_ISA_IV		0x00000008
+#define MIPS_CPU_ISA_V		0x00000010
 #define MIPS_CPU_ISA_M32R1	0x00000020
 #define MIPS_CPU_ISA_M32R2	0x00000040
 #define MIPS_CPU_ISA_M64R1	0x00000080

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Maciej W. Rozycki <macro@linux-mips.org> [2006-02-13 09:15]:
>  The following change fixes the problem for me -- it looks like an 
> omission from some recent changes elsewhere.  Ralf, please apply.

Current git and the standard decstation config fails to build with:

  LD      .tmp_vmlinux1
arch/mips/kernel/built-in.o: In function `einval':arch/mips/kernel/scall32-o32.S:(.text+0xb6c0): undefined reference to `sys_newfstatat'

This is with binutils 2.16.1cvs20060117-1 and gcc 4.0.3 20051201.
-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Martin Michlmayr <tbm@cyrius.com> [2006-02-13 22:39]:
>   LD      .tmp_vmlinux1
> arch/mips/kernel/built-in.o: In function `einval':arch/mips/kernel/scall32-o32.S:(.text+0xb6c0): undefined reference to `sys_newfstatat'
> 
> This is with binutils 2.16.1cvs20060117-1 and gcc 4.0.3 20051201.

I see a fix for this just went into Linus' git tree.  Can we have this
in the linux-mips tree too please.

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=326a625748535c4cdb1c632b1dcb07030989a393

-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Martin Michlmayr]

BTW, turning CONFIG_VT on leads to these compile errors:

  CC      drivers/tc/lk201.o
drivers/tc/lk201.c:19:26: error: linux/kbd_ll.h: No such file or directory
drivers/tc/lk201.c:23:26: error: asm/keyboard.h: No such file or directory
drivers/tc/lk201.c: In function ‘parse_kbd_rate’:
drivers/tc/lk201.c:189: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:190: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:190: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:196: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:197: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:198: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:199: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:200: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c:201: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c: In function ‘write_kbd_rate’:
drivers/tc/lk201.c:210: error: ‘struct kbd_repeat’ has no member named ‘rate’
drivers/tc/lk201.c: In function ‘lk201_rx_char’:
drivers/tc/lk201.c:363: warning: implicit declaration of function ‘handle_scancode’
drivers/tc/lk201.c: In function ‘lk201_init’:
drivers/tc/lk201.c:407: error: invalid lvalue in assignment
drivers/tc/lk201.c:408: error: invalid lvalue in assignment
make[2]: *** [drivers/tc/lk201.o] Error 1


--- old-config	2006-02-13 22:55:32.000000000 +0000
+++ .config	2006-02-13 22:55:36.000000000 +0000
@@ -530,7 +530,9 @@
 #
 # Character devices
 #
-# CONFIG_VT is not set
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
 # CONFIG_SERIAL_NONSTANDARD is not set
 CONFIG_SERIAL_DEC=y
 CONFIG_SERIAL_DEC_CONSOLE=y
@@ -631,6 +633,13 @@
 # CONFIG_FB_VIRTUAL is not set
 
 #
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+CONFIG_DUMMY_CONSOLE=y
+# CONFIG_FRAMEBUFFER_CONSOLE is not set
+
+#
 # Logo configuration
 #
 CONFIG_LOGO=y

-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Martin Michlmayr <tbm@cyrius.com> [2006-02-13 22:59]:
> BTW, turning CONFIG_VT on leads to these compile errors:

OK, I got it down to:

  CC      drivers/tc/lk201.o
drivers/tc/lk201.c: In function ‘lk201_rx_char’:
drivers/tc/lk201.c:361: warning: implicit declaration of function ‘handle_scancode’
drivers/tc/lk201.c: In function ‘lk201_init’:
drivers/tc/lk201.c:405: error: invalid lvalue in assignment
drivers/tc/lk201.c:406: error: invalid lvalue in assignment
make[2]: *** [drivers/tc/lk201.o] Error 1

But this driver really needs to be ported to the new input interface.


--- a/drivers/tc/lk201.c~	2006-02-13 22:59:57.000000000 +0000
+++ b/drivers/tc/lk201.c	2006-02-13 23:21:54.000000000 +0000
@@ -16,11 +16,9 @@
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/delay.h>
-#include <linux/kbd_ll.h>
 #include <linux/kbd_kern.h>
 #include <linux/vt_kern.h>
 
-#include <asm/keyboard.h>
 #include <asm/dec/tc.h>
 #include <asm/dec/machtype.h>
 #include <asm/dec/serial.h>
@@ -186,28 +184,28 @@
 {
 	if (r->delay <= 0)
 		r->delay = kbdrate.delay;
-	if (r->rate <= 0)
-		r->rate = kbdrate.rate;
+	if (r->period <= 0)
+		r->period = kbdrate.period;
 
 	if (r->delay < 5)
 		r->delay = 5;
 	if (r->delay > 630)
 		r->delay = 630;
-	if (r->rate < 12)
-		r->rate = 12;
-	if (r->rate > 127)
-		r->rate = 127;
-	if (r->rate == 125)
-		r->rate = 124;
+	if (r->period < 12)
+		r->period = 12;
+	if (r->period > 127)
+		r->period = 127;
+	if (r->period == 125)
+		r->period = 124;
 }
 
 static int write_kbd_rate(struct kbd_repeat *rep)
 {
-	int delay, rate;
+	int delay, period;
 	int i;
 
 	delay = rep->delay / 5;
-	rate = rep->rate;
+	period = rep->period;
 	for (i = 0; i < 4; i++) {
 		if (lk201_hook.poll_tx_char(lk201_handle,
 					    LK_CMD_RPT_RATE(i)))
@@ -216,7 +214,7 @@
 					    LK_PARAM_DELAY(delay)))
 			return 1;
 		if (lk201_hook.poll_tx_char(lk201_handle,
-					    LK_PARAM_RATE(rate)))
+					    LK_PARAM_RATE(period)))
 			return 1;
 	}
 	return 0;


-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Maciej W. Rozycki <macro@linux-mips.org> [2006-02-13 09:15]:
> The following change fixes the problem for me -- it looks like an
> omission from some recent changes elsewhere.  Ralf, please apply.

Apparently it boots with current git and your fix.  I think p2 will
confirm later.

However, it doesn't boot with 2.6.15 from a few weeks ago plus some
random fixes plus your patch.  Any idea what else might have changed?
-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Martin Michlmayr <tbm@cyrius.com> [2006-02-13 23:27]:
> Apparently it boots with current git and your fix.  I think p2 will
> confirm later.

Here's the boot log from p2:


>>boot 3/tftp/vmlinux-pmax console=ttyS2

-tftp boot(3), bootp 10.42.0.190:/tftpboot/vmlinux-pmax
-tftp load 2986117+0+118683
Linux version 2.6.16-rc3 (tbm@deprecation) (gcc version 4.0.3 20051201 (prerelease) (Debian 4.0.2-5)) #3 Mon Feb 13 22:40:31 GMT 2006
This is a DECstation 5000/2x0
CPU revision is: 00000230
FPU revision is: 00000340
Determined physical RAM map:
 memory: 0c000000 @ 00000000 (usable)
Built 1 zonelists
Kernel command line: console=ttyS2
Primary instruction cache 64kB, linesize 4 bytes.
Primary data cache 64kB, linesize 4 bytes.
Synthesized TLB refill handler (17 instructions).
Synthesized TLB load handler fastpath (31 instructions).
Synthesized TLB store handler fastpath (31 instructions).
Synthesized TLB modify handler fastpath (25 instructions).
PID hash table entries: 1024 (order: 10, 16384 bytes)
Using 25.000 MHz high precision timer.
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 191488k/196608k available (2350k kernel code, 5044k reserved, 417k data, 148k init, 0k highmem)
Mount-cache hash table entries: 512
Checking for 'wait' instruction...  unavailable.
NET: Registered protocol family 16
SCSI subsystem initialized
TURBOchannel rev. 1 at 25.0 MHz (without parity)
    slot 0: DEC      PMAZ-AA  V5.3d
    slot 1: DEC      PMAGB-BA V1.0
    slot 2: DEC      PMAG-BA  V5.3a
Initializing Cryptographic API
io scheduler noop registered
io scheduler anticipatory registered (default)
io scheduler deadline registered
io scheduler cfq registered
fb0: PMAG-BA frame buffer device in slot 2
fb1: PMAGB-BA frame buffer device in slot 1
fb1: Osc0: 130.808MHz, Osc1: disabled, Osc0 selected
DECstation Z8530 serial driver version 0.09
ttyS00 at 0xbf900001 (irq = 14) is a Z85C30 SCC
ttyS01 at 0xbf900009 (irq = 14) is a Z85C30 SCC
ttyS02 at 0xbf980001 (irq = 15) is a Z85C30 SCC
ttyS03 at 0xbf980009 (irq = 15) is a Z85C30 SCC
rtc: Digital DECstation epoch (2000) detected
Real Time Clock Driver v1.12a
declance.c: v0.009 by Linux MIPS DECstation task force
declance0: IOASIC onboard LANCE, addr = 08:00:2b:38:72:b4, irq = 16
declance0: registered as eth0.
SCSI ID 7 Clk 25MHz CCF=5 TOut 167 NCR53C9x(esp236)
SCSI ID 7 Clk 25MHz CCF=5 TOut 167 NCR53C9x(esp236)
ESP: Total of 2 ESP hosts found, 2 actually in use.
scsi0 : ESP236 (NCR53C9x)
scsi1 : ESP236 (NCR53C9x)
mice: PS/2 mouse device common for all mice
NET: Registered protocol family 2
IP route cache hash table entries: 2048 (order: 1, 8192 bytes)
TCP established hash table entries: 8192 (order: 3, 32768 bytes)
TCP bind hash table entries: 8192 (order: 3, 32768 bytes)
TCP: Hash tables configured (established 8192 bind 8192)
TCP reno registered
TCP bic registered
NET: Registered protocol family 1
NET: Registered protocol family 17
Root-NFS: No NFS server available, giving up.
VFS: Unable to mount root fs via NFS, trying floppy.
VFS: Cannot open root device "<NULL>" or unknown-block(2,0)
Please append a correct "root=" boot option
Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(2,0)

(He didn't specify root= so that's fine)
-- 
Martin Michlmayr
http://www.cyrius.com/

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

On Mon, 13 Feb 2006, Martin Michlmayr wrote:

> @@ -631,6 +633,13 @@
>  # CONFIG_FB_VIRTUAL is not set
>  
>  #
> +# Console display driver support
> +#
> +CONFIG_VGA_CONSOLE=y
> +CONFIG_DUMMY_CONSOLE=y
> +# CONFIG_FRAMEBUFFER_CONSOLE is not set
> +
> +#
>  # Logo configuration
>  #
>  CONFIG_LOGO=y

 Hmm, VGA_CONSOLE doesn't make much sense for the DECstation, does it?  
You might want to select FRAMEBUFFER_CONSOLE instead.

  Maciej

Re: DECstation R3000 boot error

[Maciej W. Rozycki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1344 bytes --]

On Mon, 13 Feb 2006, Martin Michlmayr wrote:

> OK, I got it down to:
> 
>   CC      drivers/tc/lk201.o
> drivers/tc/lk201.c: In function ‘lk201_rx_char’:
> drivers/tc/lk201.c:361: warning: implicit declaration of function ‘handle_scancode’
> drivers/tc/lk201.c: In function ‘lk201_init’:
> drivers/tc/lk201.c:405: error: invalid lvalue in assignment
> drivers/tc/lk201.c:406: error: invalid lvalue in assignment
> make[2]: *** [drivers/tc/lk201.o] Error 1
> 
> But this driver really needs to be ported to the new input interface.

 The driver has been ported by JBG (thanks!) -- it's the zs.c driver that 
needs to be ported to the new serial infrastructure.  But that's tough if 
to be done properly (DMA and synchronous modes are not handled well by the 
serial core), so not at the moment, sorry.  I'll think about minimal 
functionality to keep it going though and perhaps lk201.c could be changed 
to work with current zs.c as is (dz.c has already been ported -- I have no 
way of testing it, so I somewhat lack incentive to go through it and 
verify if it's at least as good as the old driver in 2.4.)...

 I'm not sure what your patch is meant to do, but if changing rate to 
period you need to invert the values as well -- period = 1 / rate -- and 
the hardware expects rates in cps IIRC; mind the units, though.

  Maciej

Re: DECstation R3000 boot error

[Ralf Baechle]

On Mon, Feb 13, 2006 at 10:39:51PM +0000, Martin Michlmayr wrote:

>   LD      .tmp_vmlinux1
> arch/mips/kernel/built-in.o: In function `einval':arch/mips/kernel/scall32-o32.S:(.text+0xb6c0): undefined reference to `sys_newfstatat'

Fixed yesterday, didn't get around to push it out until today.

  Ralf

Re: DECstation R3000 boot error

[Ralf Baechle]

On Mon, Feb 13, 2006 at 10:47:33PM +0000, Martin Michlmayr wrote:

> >   LD      .tmp_vmlinux1
> > arch/mips/kernel/built-in.o: In function `einval':arch/mips/kernel/scall32-o32.S:(.text+0xb6c0): undefined reference to `sys_newfstatat'
> > 
> > This is with binutils 2.16.1cvs20060117-1 and gcc 4.0.3 20051201.
> 
> I see a fix for this just went into Linus' git tree.  Can we have this
> in the linux-mips tree too please.
> 
> http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=326a625748535c4cdb1c632b1dcb07030989a393

Ah, Yoichi at it again.  He really loves sending patches upstream without
letting me know.

  Ralf

Re: DECstation R3000 boot error

[Yoichi Yuasa]

On Tue, 14 Feb 2006 16:27:22 +0000
Ralf Baechle <ralf@linux-mips.org> wrote:

> On Mon, Feb 13, 2006 at 10:47:33PM +0000, Martin Michlmayr wrote:
> 
> > >   LD      .tmp_vmlinux1
> > > arch/mips/kernel/built-in.o: In function `einval':arch/mips/kernel/scall32-o32.S:(.text+0xb6c0): undefined reference to `sys_newfstatat'
> > > 
> > > This is with binutils 2.16.1cvs20060117-1 and gcc 4.0.3 20051201.
> > 
> > I see a fix for this just went into Linus' git tree.  Can we have this
> > in the linux-mips tree too please.
> > 
> > http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=326a625748535c4cdb1c632b1dcb07030989a393
> 
> Ah, Yoichi at it again.  He really loves sending patches upstream without
> letting me know.

Sorry, I forgot add Cc: .

Yoichi

Re: DECstation R3000 boot error

[Martin Michlmayr]

* Maciej W. Rozycki <macro@linux-mips.org> [2006-02-14 10:01]:
> The driver has been ported by JBG (thanks!) -- it's the zs.c driver
> that needs to be ported to the new serial infrastructure.  But
> that's tough if to be done properly (DMA and synchronous modes are
> not handled well by the serial core), so not at the moment, sorry.
> I'll think about minimal functionality to keep it going though and
> perhaps lk201.c could be changed to work with current zs.c as is
> (dz.c has already been ported -- I have no way of testing it, so I
> somewhat lack incentive to go through it and verify if it's at least
> as good as the old driver in 2.4.)...

Do you think you'll have time to port zs.c soon?  I'd like to move
away from 2.4 in Debian and DECstation is currently the only platform
that isn't supported in 2.6 (or at least not fully).
-- 
Martin Michlmayr
http://www.cyrius.com/