Re: Tried to run 2.6.0-test1

Re: Tried to run 2.6.0-test1

[Norman Diamond]

Again I cannot keep up with the mailing list.  If anyone wishes to advise or
has questions, please kindly contact me directly.

I wrote:

> I've downloaded modutils-2.4.21-18.src.rpm from Rusty Russell's directory
[...]
> Odds are I can probably gunzip and tar and compile the new modutils, but
> will the resulting executables get installed into the right places without
> using the rpm command and its spec file, I'm not sure I want to chance it.

I read the two README files and started to try building the first one
starting with "./configure --prefix=/" as directed.  It said:
  /home/ndiamond/modutils-2.4.21-18/SOURCES/module-init-tools-0.9.11a/missing: Unknown `--run' option
  Try `/home/ndiamond/modutils-2.4.21-18/SOURCES/module-init-tools-0.9.11a/missing --help' for more information

I tried that command for help.  Indeed it says there's no such thing as a
"--run" option.  Looks like I lied.  I don't know how to compile the new
module utilities.

One thing is interesting though.  This configure script figured out that my
cpu is an i586.

By the way, the last line in that README file says that if this is all too
complicated then install the source RPM.  Gee thanks.  I already tried
"rpm --rebuild" but it assumes an i686.

Re: Tried to run 2.6.0-test1

[Brett]

Norman Diamond said:
> Again I cannot keep up with the mailing list.  If anyone wishes to advise
> or
> has questions, please kindly contact me directly.
>
> By the way, the last line in that README file says that if this is all too
> complicated then install the source RPM.  Gee thanks.  I already tried
> "rpm --rebuild" but it assumes an i686.
>

you want rpmbuild

[brett@synapse brett]$ rpmbuild --help
Usage: rpmbuild [OPTION...]
<snip>
  --target=CPU-VENDOR-OS        override target platform

rpmbuild --target=i586 --rebuild modutils-*.src.rpm

thanks,

     / Brett

Re: Tried to run 2.6.0-test1

[Norman Diamond]

"Brett" <generica@email.com> replied to one of my desperate messages today.
I'm sending this personally and to the list.  Again I can't keep up with the
list, so if Brett or anyone else has further advice or questions, please
contact me directly.

> > By the way, the last line in that README file says that if this is all
> > too complicated then install the source RPM.  Gee thanks.  I already
> >  tried "rpm --rebuild" but it assumes an i686.
>
> you want rpmbuild

SuSE 8.1 doesn't seem to have it.

> [brett@synapse brett]$ rpmbuild --help
> Usage: rpmbuild [OPTION...]
> <snip>
>   --target=CPU-VENDOR-OS        override target platform

Yeah, that's better than ARCH-VENDOR-OS.  The rpm --rebuild command's
assumed arch of i486 was fine with me, the problem is the assumed cpu of
i686.

Any chance that the rpm --rebuild command's arch-vendor-os might really mean
cpu-vendor-os?  Nah, I guess you would have said so if it were.

> rpmbuild --target=i586 --rebuild modutils-*.src.rpm

Oh, it doesn't need --target=i586-intel-linux  :-?  (By the way, the reason
I come up with snarky questions like this one is that I RTFM.  Sigh.)

Re: Tried to run 2.6.0-test1

[Brett]

Norman Diamond said:
> "Brett" <generica@email.com> replied to one of my desperate messages
> today.
> I'm sending this personally and to the list.  Again I can't keep up with
> the
> list, so if Brett or anyone else has further advice or questions, please
> contact me directly.
>
>> > By the way, the last line in that README file says that if this is all
>> > too complicated then install the source RPM.  Gee thanks.  I already
>> >  tried "rpm --rebuild" but it assumes an i686.
>>
>> you want rpmbuild
>
> SuSE 8.1 doesn't seem to have it.
>

ahh, jumped in halfway thru thread and didn't know you were on suse

>> rpmbuild --target=i586 --rebuild modutils-*.src.rpm
>
> Oh, it doesn't need --target=i586-intel-linux  :-?  (By the way, the
> reason
> I come up with snarky questions like this one is that I RTFM.  Sigh.)
>

[root@synapse router_firmware]# rpmbuild --target=i586 --rebuild
/tbla/modutils-2.4.21-18.src.rpm

<snip>

Wrote: /usr/src/redhat/RPMS/i586/modutils-2.4.21-18.i586.rpm

works for me on rh9 ... sorry i can't give suse specific help

good luck,

     / Brett

Re: Tried to run 2.6.0-test1

[Gene Heskett]

On Monday 21 July 2003 02:23, Norman Diamond wrote:

>By the way, the last line in that README file says that if this is
> all too complicated then install the source RPM.  Gee thanks.  I
> already tried "rpm --rebuild" but it assumes an i686.

Is that not in the .spec file, and therefore editable?

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.26% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: Tried to run 2.6.0-test1

[Norman Diamond]

Again I cannot keep up with the mailing list.  If anyone wishes to advise or
has questions, please kindly contact me directly.

In addition to my previous message quoting myself, I lied yet again:

> I've downloaded modutils-2.4.21-18.src.rpm from Rusty Russell's directory
[...]
> Odds are I can probably gunzip and tar and compile the new modutils, but
> will the resulting executables get installed into the right places without
> using the rpm command and its spec file, I'm not sure I want to chance it.

I read the two README files.  BOTH OF THE GZIPPED TARBALLS
PRODUCE THE SAME EXECUTABLE FILENAMES, such as insmod.
If I can get both of them to build, how can I guess which one to install?

Besides, I tried building module-init-tools-0.9.11a even after its configure
script failed.  There are a lot more differences than just the type of cpu,
from the compilation options when trying "rpm --rebuild".  For example no -D
options for combining various stuff, no URL to explain tainting, etc.  Also
this one's "make install" doesn't install the lsmod command.

I really really lied.  I'll never figure out how to build these things
without help.  As I understand it, all I need at this point is the new
modutils for execution on an i586.

Re: Tried to run 2.6.0-test1

[Norman Diamond]

Again I cannot keep up with the mailing list.  If anyone wishes to advise or
has questions, please kindly contact me directly.

1.  Trying to upgrade modutils.

I've downloaded modutils-2.4.21-18.src.rpm from Rusty Russell's directory on
ftp.kernel.org.  Not being an rpm expert, I tried "rpm --rebuild" on it.
The compilation steps assume -mcpu=i686.  I need to specify that the cpu is
an i586.

"man rpm" and "info rpm" both say that arch-vendor-os can be designated,
but -march=i486 is already fine with me, the problem is the cpu not the
arch.  I'm not sure what it would mean to try specifying a vendor (Intel) or
OS (Linux).  There seems to be no way to specify the cpu in "rpm --rebuild".

I also tried rpm -ivh and finally found where three files got installed to.
Two gzipped tarballs and one spec file.  The spec file doesn't specify the
cpu, as far as I can tell.

Odds are I can probably gunzip and tar and compile the new modutils, but
will the resulting executables get installed into the right places without
using the rpm command and its spec file, I'm not sure I want to chance it.

My .config file for the kernel sources correctly designates the cpu as an
i586mmx, but I guess modutils shouldn't be expected to guess which kernel
sources to find and extract the cpu type from it.

2.  Blank screen during/after boot.

Two people recommended alternative locations for the "post-halloween"
document and I got one.

> Known gotchas.
> Certain known bugs are being reported over and over. Here are the
>     workarounds.
> - Blank screen after decompressing kernel?
>   Make sure your .config has
>   CONFIG_INPUT=y, CONFIG_VT=y, CONFIG_VGA_CONSOLE=y and
>     CONFIG_VT_CONSOLE=y

It has.  The screen is still blank after decompressing the kernel.  In SuSE
8.1 the screen goes completely blank and never recovers.  In Red Hat  the
screen continues displaying the message about decompressing the kernel but
freezes there until X comes up, and then returns to that same frozen text
screen during shutdown, so I have to guess when it's OK to turn off the
power.

Re: Tried to run 2.6.0-test1

[Gabor MICSKO]

2003-07-20, v keltezéssel Norman Diamond ezt írta:

> > And read http://www.codemonkey.org.uk/post-halloween-2.5.txt for
> > all the OTHER stuff you might need...
> 
> DNS error.

http://portal.fsn.hu/old/post-halloween/post-halloween-2.5.txt

-- 
Windows not found
(C)heers, (P)arty or (D)ance?
-----------------------------------
Micskó Gábor
Compaq Accredited Platform Specialist, System Engineer (APS, ASE)
Szintézis Computer Rendszerház Kft.      
H-9021 Gyor, Tihanyi Árpád út. 2.
Tel: +36-96-502-216
Fax: +36-96-318-658
E-mail: gmicsko@szintezis.hu
Web: http://www.hup.hu

Re: Tried to run 2.6.0-test1

[Diego Calleja García]

El Mon, 21 Jul 2003 00:48:54 +0900 "Norman Diamond" <ndiamond@wta.att.ne.jp> escribió:

> > And read http://www.codemonkey.org.uk/post-halloween-2.5.txt for
> > all the OTHER stuff you might need...
> 
> DNS error.

http://lwn.net/Articles/39901/

Re: Tried to run 2.6.0-test1

[Norman Diamond]

Valdis.Kletnieks@vt.edu replied to me.  I'm sending this personally and to
the list.  Again I can't keep up with the list, so if Mr./Ms. Kletnieks or
anyone else has further advice or questions, please contact me directly.

>> What does the thing need in order to use modules?
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/people/rusty/modules/

550 /pub/linux/kernel/people/people/rusty/modules/: No such file or
directory

Oh I see, two many people.  After fixing that, I see a list of
module-init-tools. Hmm, modutils-2.4.21-18.src.rpm is indeed newer
than SuSE 8.1.  Downloading it now and will try it after sleeping.

I guess it's just luck that Red Hat 9 had a sufficiently new modutils.

> And read http://www.codemonkey.org.uk/post-halloween-2.5.txt for
> all the OTHER stuff you might need...

DNS error.

>> By the way there are also no tones from PCMCIA during the
>> boot process so I'm pretty sure it's also not recognizing my LAN card,
>
> The 'no tones' confused me too.  In 'make menuconfig', go to 'Input Device
> support', select 'Misc' - this will add an entry 'PC Speaker Support'.
> Check that, it will add  CONFIG_INPUT_PCSPKR which should make the beeps
> return.

I have it configured as a module.  I guess it will be fixed automatically
when modules start working.

> If you hit other snags, *read the post-halloween file*.  Most of the
> gotchas are in there already.

All except the DNS error, which came up again while I was checking out the
rest of this.  I don't think post-halloween will get here tonight.

Re: Tried to run 2.6.0-test1

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 977 bytes --]

On Mon, 21 Jul 2003 00:03:57 +0900, Norman Diamond <ndiamond@wta.att.ne.jp>  said:

> What does the thing need in order to use modules?

 ftp://ftp.kernel.org/pub/linux/kernel/people/people/rusty/modules/

And read http://www.codemonkey.org.uk/post-halloween-2.5.txt for
all the OTHER stuff you might need...

> come up?  By the way there are also no tones from PCMCIA during the
> boot process so I'm pretty sure it's also not recognizing my LAN card,

The 'no tones' confused me too.  In 'make menuconfig', go to 'Input Device support',
select 'Misc' - this will add an entry 'PC Speaker Support'.  Check that, it will add
CONFIG_INPUT_PCSPKR which should make the beeps return.

If you hit other snags, *read the post-halloween file*.  Most of the gotchas are
in there already.  If you do what that says, and it STILL doesn't work, feel free to
post again - the whole point of the -testX series is so we can find all the remaining
gotchas and either fix or workaround them.

[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Tried to run 2.6.0-test1

[Norman Diamond]

I'm sorry, I cannot cope with the volume in this mailing list.
If anyone volunteers to help and/or if anyone has questions,
please e-mail me directly.

On Red Hat 9 without Qt, make xconfig wouldn't run.
Make menuconfig ran.  Compilation ran.
During make modules_install, there were warnings about
undefined symbols for a bunch of wireless LAN card modules.
But at least the thing ran.
It boots, it keeps the screen nearly blank during booting,
then X comes up and the thing runs.
During shutdown, the screen returns to the same nearly blank display
that it had during booting.  I have to guess when it's safe to shut
off the power.  But in between startup and shutdown, it runs.
On Red Hat 9.

Then I moved to SuSE 8.1 and tried configuring a few more options.

On SuSE 8.1, make xconfig ran.  Just before the xconfig windows
appeared, these warnings appeared in the terminal window:
  ./scripts/kconfig/qconf arch/i386/Kconfig
  boolean symbol BINFMT_ZFLAT tested for 'm'? test forced to 'n'
  #
  # using defaults found in arch/i386/defconfig
  #
  arch/i386/defconfig:68: trying to assign nonexistent symbol X86_SSE2
  arch/i386/defconfig:544: trying to assign nonexistent symbol NET_PCMCIA_RADIO
  arch/i386/defconfig:663: trying to assign nonexistent symbol INTEL_RNG
Maybe one of these provides a clue about wireless LAN card modules.

Compilation starts, but this is weird:
    AS      arch/i386/boot/setup.o
  /tmp/ccZht1mp.s: Assembler messages:
  /tmp/ccZht1mp.s:1903: Warning: value 0x37ffffff truncated to 0x37ffffff
That's some truncation, eh?

These look troublesome though:
    CC [M]  drivers/mtd/nftlcore.o
  drivers/mtd/nftlcore.c: In function `NFTL_foldchain':
  drivers/mtd/nftlcore.c:354: warning: passing arg 7 of pointer to function
    makes pointer from integer without a cast
  drivers/mtd/nftlcore.c:358: warning: passing arg 7 of pointer to function
    makes pointer from integer without a cast
  drivers/mtd/nftlcore.c:363: warning: passing arg 7 of pointer to function
    makes pointer from integer without a cast
  drivers/mtd/nftlcore.c: In function `nftl_writeblock':
  drivers/mtd/nftlcore.c:632: warning: passing arg 7 of pointer to function
    makes pointer from integer without a cast
  drivers/mtd/nftlcore.c: In function `nftl_readblock':
  drivers/mtd/nftlcore.c:696: warning: passing arg 7 of pointer to function
    makes pointer from integer without a cast
    CC [M]  drivers/mtd/nftlmount.o
  drivers/mtd/nftlmount.c: In function `find_boot_record':
  drivers/mtd/nftlmount.c:220: warning: passing arg 7 of pointer to function
    makes pointer from integer without a cast

Oh, and then compilation of modules aborts:
    CC [M]  drivers/mtd/devices/doc2000.o
  drivers/mtd/devices/doc2000.c: In function `DoC2k_init':
  drivers/mtd/devices/doc2000.c:567: warning: assignment from incompatible
    pointer type
  drivers/mtd/devices/doc2000.c:568: warning: assignment from incompatible
    pointer type
    CC [M]  drivers/mtd/devices/docprobe.o
    CC [M]  drivers/mtd/devices/docecc.o
    CC [M]  drivers/mtd/devices/slram.o
    CC [M]  drivers/mtd/devices/blkmtd.o
  drivers/mtd/devices/blkmtd.c:52:25: linux/iobuf.h: No such file or directory
  drivers/mtd/devices/blkmtd.c: In function `blkmtd_readpage':
  drivers/mtd/devices/blkmtd.c:219: warning: implicit declaration of function
    `alloc_kiovec'
  drivers/mtd/devices/blkmtd.c:236: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:239: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:240: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:241: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:242: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:243: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:265: warning: implicit declaration of function
    `brw_kiovec'
  drivers/mtd/devices/blkmtd.c:267: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:268: warning: implicit declaration of function
    `free_kiovec'
  drivers/mtd/devices/blkmtd.c:169: warning: unused variable `b'
  drivers/mtd/devices/blkmtd.c: In function `write_queue_task':
  drivers/mtd/devices/blkmtd.c:323: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:351: `KIO_MAX_SECTORS' undeclared (first use
    in this function)
  drivers/mtd/devices/blkmtd.c:351: (Each undeclared identifier is reported
    only once
  drivers/mtd/devices/blkmtd.c:351: for each function it appears in.)
  drivers/mtd/devices/blkmtd.c:369: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:370: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:382: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:384: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:392: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:393: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c:407: dereferencing pointer to incomplete type
  drivers/mtd/devices/blkmtd.c: In function `blkmtd_erase':
  drivers/mtd/devices/blkmtd.c:527: warning: unused variable `b'
  drivers/mtd/devices/blkmtd.c: In function `blkmtd_read':
  drivers/mtd/devices/blkmtd.c:640: warning: unused variable `b'
  drivers/mtd/devices/blkmtd.c: In function `blkmtd_write':
  drivers/mtd/devices/blkmtd.c:713: parse error before "e3"
  drivers/mtd/devices/blkmtd.c:712: warning: unused variable `b'
  drivers/mtd/devices/blkmtd.c: At top level:
  drivers/mtd/devices/blkmtd.c:932: parse error before "cleanup_blkmtd"
  drivers/mtd/devices/blkmtd.c:933: warning: return type defaults to `int'
  drivers/mtd/devices/blkmtd.c:1021: parse error before "calc_erase_regions"
  drivers/mtd/devices/blkmtd.c:1022: warning: return type defaults to `int'
  drivers/mtd/devices/blkmtd.c:1047: parse error before "__init"
  drivers/mtd/devices/blkmtd.c:1047: warning: type defaults to `int' in
    declaration of `__init'
  drivers/mtd/devices/blkmtd.c:1047: warning: data definition has no type
    or storage class
  drivers/mtd/devices/blkmtd.c:1050: parse error before "init_blkmtd"
  drivers/mtd/devices/blkmtd.c:1051: warning: return type defaults to `int'
  drivers/mtd/devices/blkmtd.c: In function `init_blkmtd':
  drivers/mtd/devices/blkmtd.c:1199: structure has no member named `module'
  drivers/mtd/devices/blkmtd.c:1060: warning: unused variable `b'
  drivers/mtd/devices/blkmtd.c: At top level:
  drivers/mtd/devices/blkmtd.c:1269: warning: type defaults to `int' in
    declaration of `module_init'
  drivers/mtd/devices/blkmtd.c:1269: warning: parameter names (without
    types) in function declaration
  drivers/mtd/devices/blkmtd.c:1269: warning: data definition has no type
    or storage class
  drivers/mtd/devices/blkmtd.c:1270: warning: type defaults to `int' in
    declaration of `module_exit'
  drivers/mtd/devices/blkmtd.c:1270: warning: parameter names (without
    types) in function declaration
  drivers/mtd/devices/blkmtd.c:1270: warning: data definition has no type
    or storage class
  make[3]: *** [drivers/mtd/devices/blkmtd.o] Error 1
  make[2]: *** [drivers/mtd/devices] Error 2
  make[1]: *** [drivers/mtd] Error 2
  make: *** [drivers] Error 2
  
OK, I made menuconfig again and removed some MTD stuff.  It made but
it wouldn't boot.  The screen was blank for the entire boot process
(after grub selected the kernel to boot).  After a while I pressed
Ctrl-Alt-Del and rebooted my old kernel.
/var/log/messages includes this:
  Jul 20 22:10:20 diamondpana kernel: Loaded 28312 symbols from
    /boot/System.map-2.6.0-test1.
  Jul 20 22:10:20 diamondpana kernel: Symbols match kernel version 2.6.0.
  Jul 20 22:10:20 diamondpana kernel: No module symbols loaded -
    kernel modules not enabled.
Huh???  What about these lines in .config:
  CONFIG_MODULES=y
  CONFIG_MODULE_UNLOAD=y
  CONFIG_MODULE_FORCE_UNLOAD=y
  CONFIG_OBSOLETE_MODPARM=y
  # CONFIG_MODVERSIONS is not set
  CONFIG_KMOD=y
And this output to ls /lib/modules/2.6.0-test1:
  build        modules.generic_string  modules.parportmap  modules.usbmap
  kernel       modules.ieee1394map     modules.pcimap
  modules.dep  modules.isapnpmap       modules.pnpbiosmap
What does the thing need in order to use modules?
Is this the reason why the screen is completely blank during the boot
process instead of only mostly blank?  Is it the reason why X doesn't
come up?  By the way there are also no tones from PCMCIA during the
boot process so I'm pretty sure it's also not recognizing my LAN card,
but when I reboot my old kernel then PCMCIA recognizes the LAN card
just fine during booting.

Oh... the README says:
   - Make sure you have gcc 2.95.3 available.
Unfortunately I only have gcc 3.2 available.  Sigh.
But I think that was the same on Red Hat 9, probably.

So what does the thing need in order to recognize its modules?