feedback and request for help with x86_64 preview
by Jonathan Wakely
First of all, thanks very much to everyone who's working hard to get FC1
working on x86_64 - your efforts are much appreciated.
I've been trying to install the fc1_x86_64 preview on my AMD Athlon 64
3200+ box, with MSI-6702 (a.k.a K8T Neo) m/b, which uses a VIA K8T800
chipset + VIA VT8237, and also has a Promise 20378 SATA controller.
(c.f.
http://www.msicomputer.com/product/detail_spec/product_detail.asp?model=K...
). For the record I have 512MB DDR RAM and the graphics card is ATI
9600. I have one SATA disk attached to the VT8237 controller which has a
single partition containing Windows XP Home. Leaving aside the fact that
this seems to be a bad combination of hardware in terms of linux
support, here is my story:
I downloaded the http://fedora.linux.duke.edu/fc1_x86_64 tree to another
machine on my home network and burnt the boot.iso to a CD.
I tried a "linux rescue" installation so I could get a shell and use
parted to resize the existing partition and add ext3/swap partitions for
linux.
I had no problem choosing HTTP installation and configuring network
settings (except for Bug #112683), but after retrieving stage2.img
anaconda says I have no hard disks and drops me to a shell, so it seems
my VIA SATA controller is not being found/recognised for some reason.
>From the shell prompt I couldn't get parted to find the SATA disk -
possibly because I was using the wrong device name, possibly because the
controller's not being found.
I downloaded the fedora/core/development/x86_64 (nee rawhide) tree
instead and tried installing that, but got lots of error messages about
slab corruption on the console and so abandoned the development version
and rebooted.
I installed an old 3.2GB IDE disk I had lying around, to try and install
fc1_x86_64 onto that (in order to compile a working via_sata.o and then
move the system to the SATA disk) which worked to a point, but when
anaconda was about to partition the IDE disk it managed to completely
destroy the disk so that even the BIOS can't see it (the disk was old
and I suspect it was about to fail anyway - this is unlikely to be
anything to do with FC1). This left me back where I started, with
anaconda not finding my SATA disk.
At this point I'm about to borrow another old IDE disk, am downloading
the http://fedora.linux.duke.edu/fc1_x86_64/jmforbes/ tree (which looks
like the latest one, is that right?) to try to install tonight, and will
also ensure my BIOS is up to date (possibly not).
So, my first question is has anyone successfully installed fc1_x86_64 on
an MSI mobo?
The list archives seem to indicate it's been tried on ASUS, Tyan and
others, but no explicit mention of MSI or K8T Neo.
Why didn't the VIA SATA controller get found?
I was under the impression (thanks to
http://www.redhat.com/archives/fedora-devel-list/2003-November/msg01216.html
and other posts) that the fc1_x86_64 preview included support for VIA
and promise SATA controllers.
(Note that the disk might be found if my BIOS is out-of-date and I flash
it tonight, the latest version contains fixes for both the VIA and
Promise controllers)
Should I be trying the /fc1_x86_64/jmforbes/ tree? What about the
/fc1_x86_64/preview/ tree?
What is the correct device name for the first SATA disk likely to be,
/dev/sda? It's possible I just failed to call parted with the right
name! Should anaconda have found the existing windows partition, even
though it's on a SATA disk? (I thought the answer to this was yes, and
if anaconda can't find it then neither will parted, is that right?)
Sorry this feedback is a little incomplete, but I can't remember exactly
what I've tried so far (e.g. can't remember what device names i did and
didn't try with parted) and wanted to send this before I try a third set
of installation images. I'd be very grateful for any help or advice, or
even flames if anything I've done stands out as particularly stupid
(except not flashing the BIOS earlier - I know that's stupid :-)
jon
http://www.mintel.com
London Office:
Mintel International Group Ltd (Mintel)
18-19 Long Lane
London
EC1A 9PL
UK
Tel: 020 7606 4533
Fax: 020 7606 5932
Chicago Office:
Mintel International Group Ltd (Mintel)
213 West Institute Place
Suite #208
Chicago
IL 60610
USA
Tel: 312 932 0400
Fax: 312 932 0469
Notice
********
This email may contain information that is privileged,
confidential or otherwise protected from disclosure. It
must not be used by, or its contents copied or disclosed
to, persons other than the addressee. If you have received
this email in error please notify the sender immediately
and delete the email. Any views or opinions expressed in
this message are solely those of the author, and do not
necessarily reflect those of Mintel.
No Mintel staff are authorised to make purchases using
email or over the internet, and any contracts so performed
are invalid.
Warning
**********
It is the responsibility of the recipient to ensure that
the onward transmission, opening or use of this message
and any attachments will not adversely affect their systems
or data. Please carry out such virus and other checks, as
you consider appropriate.
20 years, 3 months
rawhide report: 20031230 changes
by Build System
Updated Packages:
bind-9.2.3-4
------------
* Mon Dec 22 2003 Daniel Walsh <dwalsh(a)redhat.com> 9.2.3-4
- Break out library package
busybox-1.00.pre5-1
-------------------
* Mon Dec 29 2003 Dan Walsh <dwalsh(a)redhat.com> 1.00-pre5.1
-Latest update
dhcp-3.0pl2-6.18
----------------
* Mon Dec 29 2003 Dan Walsh <dwalsh(a)redhat.com> 1:3.0pl2-6.18
- Add BuildRequires groff
- Replace resolv.conf if renew and data changes
gcc-3.3.2-5
-----------
* Fri Dec 19 2003 Jakub Jelinek <jakub(a)redhat.com> 3.3.2-5
- update from gcc-3_3-branch
- PRs target/11992, target/13122, optimization/13037,
target/13256, target/12598, optimization/13318,
optimization/13060, optimization/12965, target/13354,
optimization/4490, target/13150, middle-end/10060,
driver/13211, target/13302, target/11322, target/12467,
12969, target/8407, 10239, 11640, c++/12253,
c++/13262, c++/13323, fortran/12633, libstdc++/6243,
libstdc++/11612, libstdc++/12496, libstdc++/13290,
libstdc++/9371, libstdc++/9546, libstdc++/10093,
libstdc++/10095
- fix __builtin_expect in C++ code
- fix unwinding through SA_ONSTACK signal frames on IA-64
* Fri Dec 12 2003 Jakub Jelinek <jakub(a)redhat.com> 3.3.2-4
- fix unwinding through signal frames on IA-64
* Wed Dec 03 2003 Jakub Jelinek <jakub(a)redhat.com> 3.3.2-3
- update from gcc-3_3-branch
- PRs optimization/11634, other/12505, optimization/13041,
target/12900, optimization/12926, optimization/12953,
target/12865, optimization/10467, optimization/11741,
c++/2094, libobjc/11433, c++/2094, libstdc++/12297,
libstdc++/12594
- BuildRequire texinfo (#111168)
- Require /sbin/install-info for libgcj (#110904)
- fix structure initialization with const fields and mostly zeros
in the initializer (#110966)
- fix gcj on PPC64 not emitting needed nop after branch and link
to non-local Java method (Richard Henderson)
- some more > 2GB handling fixes (Jan Hubicka)
glibc-2.3.3-2
-------------
* Mon Dec 29 2003 Jakub Jelinek <jakub(a)redhat.com> 2.3.3-2
- update from CVS
- faster getpid () in NPTL builds
- fix to make pthread_setcancelstate (PTHREAD_CANCEL_DISABLE, )
really disable cancellation (#112512)
- more regex fixes and speedups
- fix nextafter*/nexttoward*
- handle 6th syscall(3) argument on AMD64
- handle memalign/posix_memalign in mtrace
- fix linuxthreads memory leak (#112208)
- remove throw () from cancellation points in linuxthreads (#112602)
- fix NPTL unregister_atfork
- fix unwinding through alternate signal stacks
* Mon Dec 01 2003 Jakub Jelinek <jakub(a)redhat.com> 2.3.3-1
- update from CVS
- 2.3.3 release
- lots of regex fixes and speedups (#110401)
- fix atan2
- fix pshared condvars in NPTL
- fix pthread_attr_destroy for attributes created with
pthread_attr_init(a)GLIBC_2.0
- for the time being, include both nb_NO* and no_NO* as locales
so that the distribution can catch up with the no_NO->nb_NO
transition
- add BuildPrereq texinfo (#110252)
* Tue Nov 18 2003 Jakub Jelinek <jakub(a)redhat.com> 2.3.2-102
- update from CVS
- fix getifaddrs (CAN-2003-0859)
- fix ftw fd leak
- fix linuxthreads sigaction (#108634)
- fix glibc 2.0 stdio compatibility
- fix uselocale (LC_GLOBAL_LOCALE)
- speed up stdio locking in non-threaded programs on IA-32
- try to maintain correct order of cleanups between those
registered with __attribute__((cleanup))
and with LinuxThreads style pthread_cleanup_push/pop (#108631)
- fix segfault in regex (#109606)
- fix RE_ICASE multi-byte handling in regex
- fix pthread_exit in libpthread.a (#109790)
- FTW_ACTIONRETVAL support
- lots of regex fixes and speedups
- fix ceill/floorl on AMD64
rpmdb-fedora-1.90-0.20031230
----------------------------
20 years, 3 months
Re: Thumbdrive/vfat and noatime.
by Nils O. Selåsdal
On Tue, 2003-12-30 at 10:33, Peter Robinson wrote:
> > I just got one of those USB thumbdrives, and was pleased to see
> > that Fedora discovered it, made /mnt/flash/ and and entry in /etc/fstab.
> > However, reading some about the thumbdrive, it seems to have a limited
> > lifetime(quite a few thousands writes though). Which makes me want to
> > mount it with -o noatime -so it doesn't get unneeded writes-
> > (and actually sync also..).
> >
> > Shouldn't Fedora make such devices noatime by default ?
> > (If not, how do I add options that's automatically added to
> > /etc/fstab/ ?)
> >
> > It's also a vfat filesystem on the device. Why isn't noatime
> > the default for mounting fat/vfat.. ? It seems atime actually updates
> > the creation/modification time on fat filesystems anyway. Default
> > noatime makes sense for mount.vfat !?
>
> There was an article titled 'floppies for the new millenium' on Linux Journal
> recently, you can find it here... http://www.linuxjournal.com/article.php?sid=6867
Thanks, that deals with setting up the thumbdrive. However Fedora
already does that automatically. I'd like to know where it does that so
I can fiddle with the mount flags, and if not the noatime flag shold be
default for such devices.
--
Vennlig hilsen/Best Regards
Nils Olav Selåsdal
System Engineer
UtelSystems a/s
Tlf: 370 45 431
Mob: 943 01 380
w w w . u t e l s y s t e m s . c o m
20 years, 3 months
Thumbdrive/vfat and noatime.
by Nils O. Selåsdal
I just got one of those USB thumbdrives, and was pleased to see
that Fedora discovered it, made /mnt/flash/ and and entry in /etc/fstab.
However, reading some about the thumbdrive, it seems to have a limited
lifetime(quite a few thousands writes though). Which makes me want to
mount it with -o noatime -so it doesn't get unneeded writes-
(and actually sync also..).
Shouldn't Fedora make such devices noatime by default ?
(If not, how do I add options that's automatically added to
/etc/fstab/ ?)
It's also a vfat filesystem on the device. Why isn't noatime
the default for mounting fat/vfat.. ? It seems atime actually updates
the creation/modification time on fat filesystems anyway. Default
noatime makes sense for mount.vfat !?
.. comments ?
--
Vennlig hilsen/Best Regards
Nils Olav Selåsdal
System Engineer
UtelSystems a/s
w w w . u t e l s y s t e m s . c o m
20 years, 3 months
Linksys router problems with 2.6 kernels
by Mike Chambers
Since using the 2.6 kernels, it seems I have problems accessing my
linksys router via web/http whereas with 2.4 I can access it.
Anyone heard of this as of yet? Am going to talk with Linksys tech
support but wanted to try this first to see if anyone had any ideas or
suggestions.
--
Mike Chambers
Madisonville, KY
"It's funny until someone gets hurt...then it's hilarious!"
20 years, 3 months
initscripts: Can it restart a service automatically if one crashes?
by Tony Yat-Tung Cheung
Hi,
I am going to write an init script for a 3rd party application. I would
like the application to restart itself automatically if it crashes or
core dumps (up to a predefined limit). Does the current initscripts
provide similar functionality?
I assume it is important, as we would like to make sure certain services
continue to be available in the system.
Thank you. Any idea?
Best Regards,
Tony Cheung
20 years, 3 months
May I contribute?
by Bob Billing
I've been using Linux since I bought a copy of RedHat 4, about seven
years ago.
I've written some packages for my own use, which I'd like to
contribute to the Fedora effort.
The first of these is tsplit, a utility which simplifies splitting
large files over mutltiple volumes. It can deal with media changes
automatically.
I have a tarball, RPM and SRPM available. It's fairly minimal, just a
C source, Makefile, man page and spec file to build the RPM.
And yes, I am the same Robert Billing who wrote "Teach Yourself
Linux", which will be coming out as part of Hodder's UK "Teach
Yourself" series later this month.
--
Robert Billing, Tanglewood,
01344-772849 64, Pinehill Road,
rbilling(a)tnglwood.demon.co.uk Crowthorne,
http://www.tnglwood.demon.co.uk/ Berks, RG45 7JR
20 years, 3 months
rawhide report: 20031229 changes
by Build System
New package libxfce4util
Utility library for the XFce4 desktop environment
Updated Packages:
rpm-4.2.2-0.8
-------------
* Sun Dec 28 2003 Jeff Johnson <jbj(a)jbj.org> 4.2.2-0.8
- convert ja and ko man pages to utf8 (#106050).
- man page corrections (#106415).
- perl.req typo (#106672).
- fix: wrong package count for trigger scriptlet 1st arg (#100509).
- fix: don't break header SHA1 if non-existent user/group (#97727).
- remove fuids/fgids from rpmfi, easier to lookup fuser/fgroup instead.
- merge sensible parts of openpkg rpm.patch.bugfix (#104780).
- mark _javadocdir as documentation (#102898).
- flush pipe before exit 1 in check-files (#103867).
- perl.req: avoid regex misfire on '^use' in "= <<" assign (#109934).
- find-debuginfo.sh: permit stripping unwritable by non-root (#112429).
- missing build dependency (#111104).
rpmdb-fedora-1.90-0.20031229
----------------------------
20 years, 3 months