On Fri, Jul 31, 2009 at 07:05:16AM +0200, Martin Michlmayr wrote:
* Moritz Muehlenhoff j...@inutil.org [2009-07-30 22:16]:
That's strange because I booted a kernel just fine when I added the
patches to 2.6.22-6. I'll test the .deb tomorrow.
What is the status of this bug, does it
On Sun, Nov 18, 2007 at 07:11:53PM +0100, Martin Michlmayr wrote:
* Florian Lohoff f...@rfc822.org [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because I booted a kernel just fine
* Moritz Muehlenhoff j...@inutil.org [2009-07-30 22:16]:
That's strange because I booted a kernel just fine when I added the
patches to 2.6.22-6. I'll test the .deb tomorrow.
What is the status of this bug, does it still apply to Lenny?
I think the bottom line is that we don't support SGI
On Mon, Nov 26, 2007 at 12:21:49PM +, Thiemo Seufer wrote:
With proper support in the installer, different images, and even further
increased buildd time: Quite a lot.
The point is that there are probably a lot ore IP22 old-r4k users than
there are matla 4kc or 5kc users ;) And those have a
Florian Lohoff wrote:
On Mon, Nov 26, 2007 at 12:21:49PM +, Thiemo Seufer wrote:
With proper support in the installer, different images, and even further
increased buildd time: Quite a lot.
The point is that there are probably a lot ore IP22 old-r4k users than
there are matla 4kc or
On Fri, Nov 23, 2007 at 05:32:56PM +, Thiemo Seufer wrote:
Gcc 4.2 has -mfix-r4000 and -mfix-r4400 options which work around some
of those errata. Maybe that's enough to make it work.
Nope - not enough - Neither -mfix-r4000 nor both let me create a working
64bit kernel ...
I guess very old
On Fri, Nov 23, 2007 at 07:04:17PM +, Thiemo Seufer wrote:
Unless we use the compiler options mentioned before, which is probably
not sensible given the performance impcat for other machines.
However, keeping the kernel for IP22 at 64 bit sounds valuable to me,
as most machines can make
Florian Lohoff wrote:
On Fri, Nov 23, 2007 at 07:04:17PM +, Thiemo Seufer wrote:
Unless we use the compiler options mentioned before, which is probably
not sensible given the performance impcat for other machines.
However, keeping the kernel for IP22 at 64 bit sounds valuable to me,
* Thomas Bogendoerfer [EMAIL PROTECTED] [2007-11-25 02:55]:
Okay - i comes past the typical trip points but later when going
to userspace throws errors a lot ...
hmm, I trusted Ralf, that the MAP_BASE change is needed for getting
modules to work, but it looks like it breaks vmalloc. And it
* Florian Lohoff [EMAIL PROTECTED] [2007-11-23 19:04]:
Can you test a kernel if I compile one with that change?
Definitly - tested and rebooted a couple hundret times the last days ;)
Can you try
http://merkel.debian.org/~tbm/tmp/linux-image-2.6.23-1-r4k-ip22_2.6.23-1~experimental.1_mips.deb
On Sat, Nov 24, 2007 at 01:01:44PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-23 19:04]:
Can you test a kernel if I compile one with that change?
Definitly - tested and rebooted a couple hundret times the last days ;)
Can you try
On Sat, Nov 24, 2007 at 09:34:03PM +0100, Florian Lohoff wrote:
Okay - i comes past the typical trip points but later when going
to userspace throws errors a lot ...
hmm, I trusted Ralf, that the MAP_BASE change is needed for getting
modules to work, but it looks like it breaks vmalloc. And it
On Fri, Nov 23, 2007 at 06:18:00PM +0100, Martin Michlmayr wrote:
Thomas Bogendoerfer said a select I8259 for IP22 should work around
this (he's working on a proper fix now). I'll do that for 2.6.23
since there won't be another upload of 2.6.22.
Yep
Can you test a kernel if I compile one
Florian Lohoff wrote:
On Fri, Nov 23, 2007 at 05:54:52PM +, Thiemo Seufer wrote:
That's a decision Thiemo would have to make. Thiemo, why is our IP22
kernel 64 bit again?
The idea is to provide 64-bit kernels where possible, so n32/n64
can work alongside o32.
The errata
On Fri, Nov 23, 2007 at 07:04:17PM +, Thiemo Seufer wrote:
Unless we use the compiler options mentioned before, which is probably
not sensible given the performance impcat for other machines.
However, keeping the kernel for IP22 at 64 bit sounds valuable to me,
as most machines can make
Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-23 17:59]:
SGI IP22 Indy r4k 100Mhz will not be able to boot unmodified 64bit
kernels. There are 64bit erratas for R4000SC Revision 3.0 which
prohibit running 64bit code without kernel and gcc modifications.
The only
Florian Lohoff wrote:
On Sun, Nov 18, 2007 at 07:11:53PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because
* Florian Lohoff [EMAIL PROTECTED] [2007-11-23 17:34]:
That's strange because I booted a kernel just fine when I added the
patches to 2.6.22-6. I'll test the .deb tomorrow.
Okay - So one of the bugs is the EISA irq initialization. Disabling
CONFIG_EISA works on my I2 in respect to booting.
On Sun, Nov 18, 2007 at 07:11:53PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because I booted a kernel just fine
On Sun, Nov 18, 2007 at 07:11:53PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because I booted a kernel just fine
* Florian Lohoff [EMAIL PROTECTED] [2007-11-23 17:59]:
SGI IP22 Indy r4k 100Mhz will not be able to boot unmodified 64bit
kernels. There are 64bit erratas for R4000SC Revision 3.0 which
prohibit running 64bit code without kernel and gcc modifications.
The only solution will be to run a 32bit
On Fri, Nov 23, 2007 at 05:54:52PM +, Thiemo Seufer wrote:
That's a decision Thiemo would have to make. Thiemo, why is our IP22
kernel 64 bit again?
The idea is to provide 64-bit kernels where possible, so n32/n64
can work alongside o32.
The errata Thomas ad-hoc found was for example
On Sun, Nov 18, 2007 at 07:11:53PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because I booted a kernel just fine
* Florian Lohoff [EMAIL PROTECTED] [2007-11-19 11:10]:
I have tested on 2 machines now and both behave the same way.
Are you probably using a different arcboot or tftp or something?
Strange. It works for me:
arcsboot: ARCS Linux ext2fs loader 0.3.8.8
Loading Linux from
Can you build a kernel from linux-mips git and see if that works?
--
Martin Michlmayr
http://www.cyrius.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
On Mon, Nov 19, 2007 at 12:14:09PM +0100, Martin Michlmayr wrote:
Can you build a kernel from linux-mips git and see if that works?
After trying to produce a new toolchain:
net/sched/em_meta.c: In function 'meta_int_loadavg_0':
net/sched/em_meta.c:127: error: PRINT_OPERAND, invalid operand
On Mon, Nov 19, 2007 at 12:13:29PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-19 11:10]:
I have tested on 2 machines now and both behave the same way.
Are you probably using a different arcboot or tftp or something?
Strange. It works for me:
arcsboot:
Florian Lohoff wrote:
On Sun, Nov 18, 2007 at 07:11:53PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because
On Mon, Nov 19, 2007 at 12:14:09PM +0100, Martin Michlmayr wrote:
Can you build a kernel from linux-mips git and see if that works?
Okay - it even gets more strange - here is a table:
IP22 Indy R5k 150Mhz
r4k-ip22-2.6.22-6 works
current git 2.6.24-rc2 breaks in Zilog serial driver (stops)
Florian Lohoff wrote:
On Mon, Nov 19, 2007 at 12:14:09PM +0100, Martin Michlmayr wrote:
Can you build a kernel from linux-mips git and see if that works?
After trying to produce a new toolchain:
net/sched/em_meta.c: In function 'meta_int_loadavg_0':
net/sched/em_meta.c:127: error:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-19 14:01]:
After trying to produce a new toolchain:
Set CONFIG_BUILD_ELF64=y in the kernel and this won't happen.
--
Martin Michlmayr
http://www.cyrius.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble?
On Mon, Nov 19, 2007 at 03:36:28PM +0100, Martin Michlmayr wrote:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-19 14:01]:
After trying to produce a new toolchain:
Set CONFIG_BUILD_ELF64=y in the kernel and this won't happen.
Its doesnt seem to exist anymore ...
[EMAIL
* Florian Lohoff [EMAIL PROTECTED] [2007-11-19 15:19]:
Okay - it even gets more strange - here is a table:
Can you send that information to [EMAIL PROTECTED]
--
Martin Michlmayr
http://www.cyrius.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble?
Package: linux-2.6
Version: 2.6.22-6
Severity: grave
Justification: renders package unusable
Hi,
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
Starting ELF64 kernel
And then its dead - Last lines:
boot
60928+176+320 entry:
* Florian Lohoff [EMAIL PROTECTED] [2007-11-18 19:02]:
Version: 2.6.22-6
the kernel linux-image-2.6.22-3-r4k-ip22_2.6.22-6_mips.deb dies
early on an SGI IP22 (mips) with
That's strange because I booted a kernel just fine when I added the
patches to 2.6.22-6. I'll test the .deb tomorrow.
--
35 matches
Mail list logo
