[Linux-ia64] running X with 2.4.18 kernel

[Linux-ia64] running X with 2.4.18 kernel

[Chris McDermott]

Anybody running X with a 2.4.18 (either 0226 or 0410 ia64 patch) kernel? 
I discovered a couple of weeks ago that it didn't work, at least with 
the kernels I'm building. Now, I'm finally getting around to trying this 
again. The systems that fail are running RH 7.2 (with XFree86-4.1.0-8 
or XFree86-4.1.0-15). I have also built and installed XFree86 4.2.0, 
same results. I'm just now starting to debug this, but thought I'd take 
the lazy way out first and see if anybody else has seen/solved this. Here
is what I see in the log file:

(II) LoadModule: "scanpci"
(II) Loading /usr/X11R6/lib/modules/libscanpci.a
(II) Module scanpci: vendor="The XFree86 Project"
	compiled for 4.2.0, module version = 0.1.0
	ABI class: XFree86 Video Driver, version 0.5

Fatal server error:
Caught signal 11.  Server aborting


Thanks,

Chris

Re: [Linux-ia64] running X with 2.4.18 kernel

[David Mosberger]

>>>>> On Tue, 23 Apr 2002 13:53:48 -0700, Chris McDermott <lcm@us.ibm.com> said:

  Chris> Anybody running X with a 2.4.18 (either 0226 or 0410 ia64
  Chris> patch) kernel?

Yes, I am.

  Chris> I discovered a couple of weeks ago that it
  Chris> didn't work, at least with the kernels I'm building. Now, I'm
  Chris> finally getting around to trying this again. The systems that
  Chris> fail are running RH 7.2 (with XFree86-4.1.0-8 or
  Chris> XFree86-4.1.0-15). I have also built and installed XFree86
  Chris> 4.2.0, same results. I'm just now starting to debug this, but
  Chris> thought I'd take the lazy way out first and see if anybody
  Chris> else has seen/solved this.

Search the earlier kernel patch announcements for chatr.  Then do:

	chatr -E /usr/X11R6/bin/XFree86

BTW: I did try to make the announcement as conspicuous as possible.  I
guess it still didn't work...

	--david

Re: [Linux-ia64] running X with 2.4.18 kernel

[Bjorn Helgaas]

Have you chatr'd your X executable?  Stack pages are not
executable by default starting in 2.4.18.  See David's patch
announcement for the fix.
-- 
Bjorn Helgaas - bjorn_helgaas at hp.com
Linux Systems Operation R&D
Hewlett-Packard

RE: [Linux-ia64] running X with 2.4.18 kernel

[Matt_Domsch]

> Anybody running X with a 2.4.18 (either 0226 or 0410 ia64 
> patch) kernel? 
> I discovered a couple of weeks ago that it didn't work, at least with 
> the kernels I'm building. 

See the release notes for the first 2.4.18 patch that David made.  He
included a program to allow executable stack for certain apps (like X),
while keeping it non-executable by default.  X needs to have an executable
stack for scanpci to work.  After running David's program, it works fine for
me.

Thanks,
Matt

--
Matt Domsch
Sr. Software Engineer
Dell Linux Solutions www.dell.com/linux
Linux on Dell mailing lists @ http://lists.us.dell.com
#1 US Linux Server provider for 2001!

RE: [Linux-ia64] running X with 2.4.18 kernel

[David Mosberger]

>>>>> On Tue, 23 Apr 2002 16:07:53 -0500, Matt_Domsch@Dell.com said:

  Matt> X needs to have an executable stack for scanpci to
  Matt> work.  After running David's program, it works fine for me.

BTW: a fix for the X server which will obviate the need for chatr has
been submitted by the Debian maintainer of the X11 package.  I'm not
tracking the status (too many other things to worry about), but I
expect that the X maintainers will eventually get around to fixing
this.

	--david

Re: [Linux-ia64] running X with 2.4.18 kernel

[Chris McDermott]

On Tue, Apr 23, 2002 at 14:03 -0700, David Mosberger wrote: 
> >>>>> On Tue, 23 Apr 2002 13:53:48 -0700, Chris McDermott <lcm@us.ibm.com> said:
> 
>   Chris> Anybody running X with a 2.4.18 (either 0226 or 0410 ia64
>   Chris> patch) kernel?
> 
> Yes, I am.
> 
>   Chris> I discovered a couple of weeks ago that it
>   Chris> didn't work, at least with the kernels I'm building. Now, I'm
>   Chris> finally getting around to trying this again. The systems that
>   Chris> fail are running RH 7.2 (with XFree86-4.1.0-8 or
>   Chris> XFree86-4.1.0-15). I have also built and installed XFree86
>   Chris> 4.2.0, same results. I'm just now starting to debug this, but
>   Chris> thought I'd take the lazy way out first and see if anybody
>   Chris> else has seen/solved this.
> 
> Search the earlier kernel patch announcements for chatr.  Then do:
> 
> 	chatr -E /usr/X11R6/bin/XFree86
> 
> BTW: I did try to make the announcement as conspicuous as possible.  I
> guess it still didn't work...
> 
> 	--david
Well, it was very conspicuous if I would have RTFMed in the first place.
Sorry to waste everyone's time.

Chris

[Linux-ia64] running X with 2.4.18 kernel

[Zac Morris]

I'n trying to run X with 2.4.18 kernel and get the
following error in the log file:

(II) LoadModule: "scanpci"
(II) Loading /usr/X11R6/lib/modules/libscanpci.a
(II) Module scanpci: vendor="The XFree86 Project"
	compiled for 4.2.0, module version = 0.1.0
	ABI class: XFree86 Video Driver, version 0.5

Fatal server error:
Caught signal 11.  Server aborting

I have read all the postings on this archive regarding
this problem and tried running the chatr script (from
ftp://ftp.hpl.hp.com/pub/linux-ia64/chatr-v0.0.tar.gz)
that David had posted, I'm getting a whole lot of
syntax errors. What could be the problem?

Where can I get the release notes for the first 2.4.18
patch that David made?

Thanks

Zac




__________________________________________________
Do you Yahoo!?
Yahoo! Web Hosting - establish your business online
http://webhosting.yahoo.com

Re: [Linux-ia64] running X with 2.4.18 kernel

[Bjorn Helgaas]

On Tuesday 11 March 2003 7:45 am, Zac Morris wrote:
> I'n trying to run X with 2.4.18 kernel ...

That's a pretty old kernel!  The ia64 patch was released over a
year ago.  All else being equal, it will be much easier to help
you if you're running a more current kernel.  The 2.4.20 ia64
patch was released on 12/10/2002, and has no major known issues.

> I have read all the postings on this archive regarding
> this problem and tried running the chatr script (from
> ftp://ftp.hpl.hp.com/pub/linux-ia64/chatr-v0.0.tar.gz)
> that David had posted, I'm getting a whole lot of
> syntax errors. What could be the problem?

Did you compile chatr?  It compiles fine for me:

    09:15:43 lp:~$ socksify wget -q ftp://ftp.hpl.hp.com/pub/linux-ia64/chatr-v0.0.tar.gz
    09:15:59 lp:~$ tar -zxvf chatr-v0.0.tar.gz
    chatr.c
    09:16:02 lp:~$ cc chatr.c
    09:16:08 lp:~$ 

If you're having trouble compiling it, please include the actual
error messages you're seeing.

> Where can I get the release notes for the first 2.4.18
> patch that David made?

They're included in the diff file (linux-2.4.18-ia64-020226.gz).
Here they are:

        - better ia64 IRQ affinity support (Erich Focht)
        - include Firewire configuration for ia64 (Bdale Garbee/Grant Grundler)
        - fix ia32 GDT initialization (Don Dugger)
        - bump gcc3.x inlining threshold to 2000 to get things to
          compile with gcc3.1
        - don't make ia64 stack and data executable by default and
          drop lazy-execute bit support; restore original interface
          for page-fault handler
        - fix clone2() stack initialization bug found by Asit Mallick
        - implement ia32 emulation for SIOCGIFCONF (Don Dugger)
        - fix siginfo kernel data leaking
        - make page fault handler pass reason for a SIGSEGV/SIGILL in
          siginfo.si_isr (see include/asm-ia64/siginfo.h for details)
        - /proc/efivars update (Matt Domsch)
        - fix I/O SAPIC irq routing (makes ACPI power button events work)
          (J.I. Lee)
        - fix alternate TLB miss handlers to not lose Exception Deferral bit
        - MCA update (Jenna Hall)
        - update perfmon to v1.0 (Stephane Eranian)
        - add PTRACE_GETREGS/PTRACE_SETREGS (Yoav Zach)
        - fix handling of .label_state/.copy_state in kernel unwinder
        - fix exception handler to not assume that archdata_start is
          non-NULL for all modules (Bjorn Helgaas)
        - move initial kernel stack from region 4 down to region 3
        - don't define _HAVE_ARCH_IPV6_CSUM as we don't really have it
          (Bdale Garbee/Grant Grunderl)
        - drop .bias from ld4.bias instruction in spinlock code to avoid
          CPU errata (Asit Mallick)