July 1999 - openSUSE Users - openSUSE Mailing Lists

RE: [suse-linux-e] Re: [SuSE Linux] 1024-Cylinder Limit Question
by Scott.Worley＠COMPAQ.com 27 Jul '99

27 Jul '99

Michael, The original PC-AT BIOS int13 call uses a 10bit field for cylinder number which caps the max cylinder to 1023. LILO uses the original call to maintain compatibility across platforms, i.e. old 386 systems and up. SCSI BIOS replaces the system BIOS int13 interface but must provide same calls so it's also limited to 1023 cylinders. The SCSI controller handles the actual logical to physical driver translation but on PC's it must live with the old int13 interface. Some SCSI cards, like the Adaptec 2940, have a setting to turn off drive translation. I've ran linux in the past with drive tranlation off but LILO is still limited to <1023 cylinders. Translation maps driver geometry so that the number of cylinders is <= 1023 usually by increasing the number of logical heads. Of course different BIOS implement the translation differently which can result in in a hard drive's partitioning not being recognized on another system. For example, many BIOS will cap the number of heads to 240 whereas my Adaptec 2940 SCSI controller caps at 255 heads. One problem with tranlation is the 8GB barrier. The max translated geometry for the old int13 interface, 1023x255x63, results in ~7.8GB. The int13 interface was extended a few years ago with new calls which allow the drive to be accessed with larger CHS values or using sector addressing which is somewhat independent of geometry. I guess someone could modify LILO to use the extended interface but you only have about 470 bytes of code space in the MBR. There's also another open source boot manager on the SuSE CD's, chos. I've looked at the code for chos and it also seems to use the old int13 interface. The other thing to mention is that IDE drives have multiple fields containing drive geometries. The Compaq BIOS limits the geometry in the old int13 interface to 1023x240x63(7.8GB). The new int13 functions are available to access the whole drive. Also, during boot up the BIOS programs the logical cylinders, logical heads, and logical sectors in words 54-56 in the IDE drive's ID structure. Words 57-58 contain the total number of sectors on the drive. So for your Armada or any other ibm type PC any software, LILO, which uses the original int13 interface will be limited to 1023 cylinders and a size of 7.8GB. Any software written to take advantage of the extended int13 interface and/or new IDE drive data fields can access the whole drive. I tested linux with kernel 2.0.36 on an Armada with 10 & 14GB drives and found that linux fdisk would not let me create a partition > 7.8GB unless I used the advanced menu and manually changed the disk geometry. Once the partitions were created, boot partition < 1023, linux had no problems accessing all the drive. Kernel 2.2.X and the newer disk utilities fixed this. I wrote all this out because this drive translation stuff still confuses me sometimes. Hope it explained the issue. -----Original Message----- From: Michael Malone [mailto:Michael.Malone@globalone.net] Sent: Tuesday, July 27, 1999 3:42 AM To: suse-linux-e(a)suse.de Subject: [suse-linux-e] Re: [SuSE Linux] 1024-Cylinder Limit Question I have been pondering this same question about my new Compaq Armada 7800. Is the Bios new enough to overcome the 1024 cylinder limit? It's the one with the 14gig, not the 8gig drive. Anyone know? Thanks, Michael Dennis Soper wrote: > Hi all, > > Chapter 4 in the > SuSE 6.1 (pg 108) manual makes a kind of offhand reference that > certain BIOSes and SCSI drives are not subject to the 1024- > Cylinder limit for the boot machinery. --------------------------------------------------------------------- To unsubscribe, e-mail: suse-linux-e-unsubscribe(a)suse.com For additional commands, e-mail: suse-linux-e-help(a)suse.com

1 0

Loss of TCP/IP
by lists＠tcob1.tele2.co.uk 27 Jul '99

27 Jul '99

Hi, I have noticed that recently my system 'loses' tcp/ip and the only way to regain is to reboot the system. I checked my network and it appears okay. My connection is via an ethernet card, and using iptraf I noticed that I was still getting ARPs etc but no TCP stuff. Would anyone have a suggestion as to what might be causing this. Sean -- Sysop and Webmaster of TCOB1 | Fidonet: 2:252/300 2:252/0 http://212.19.67.58 | Free Fidonet Feeds available telnet://212.19.67.58 2300 | Linux User: #124682 ICQ: 679813 ftp://212.19.67.58 | irc: thecivvie on irc.sysopnet.org Powered by SuSe Linux 6.0 and a 128k Tele2 connection Soon to be the father of TWINS PHORUMS online: http://212.19.67.58/phorum/

1 0

Re: [SuSE Linux] Dumb Kernel Question
by tomas＠primenet.com 27 Jul '99

27 Jul '99

"Michael J. McGillick" wrote: > > Evening: > > Am I correct in assuming that I can download the linux-2.2.10.tar.gz > kernel and use it to replace the SuSE-linux kernel ther now? If I'm not > mistaken, two or three symlinks need to be changed to point the new kernel > source, not the old, but other than that, am I ok doing this, or do I need > to wait for a special kernel released by SuSE? > Yep, piece of cake to do also. 1. download the source 2. create a directory md /usr/src/linux-2.2.10 3. remove the old symlink rm -rf /usr/src/linux 4. create a new symlink ln -s /usr/src/linux-2.2.10 /usr/src/linux 5. copy .config to new linux directory or run make mrproper to start from scratch. 6. configure and make kernel/modules Another option to save bandwidth is the following 1. cp -dpr /usr/src/linux-2.2.7 /usr/src/linux-old (in case of oops) 2. download patches 2.2.8 - 2.2.10 3. patch old kernel 4. rename /usr/src/linux-2.2.7 /usr/src/linux-2.2.10 5. do steps 3 - 6 from above procedure This second is assuming you used the the standard kernel sources not the suse modified sources. Otherwise I would recommend the first procedure. -- S.Toms - tomas(a)primenet.com - http://www.primenet.com/~tomas S.u.S.E. Linux v6.1+ - Kernel 2.2.10 All Finagle Laws may be bypassed by learning the simple art of doing without thinking. -- To get out of this list, please send email to majordomo(a)suse.com with this text in its body: unsubscribe suse-linux-e Check out the SuSE-FAQ at http://www.suse.com/Support/Doku/FAQ/ and the archive at http://www.suse.com/Mailinglists/suse-linux-e/index.html

4 4

Is anybody out there?
by jonski＠cachalot74.freeserve.co.uk 27 Jul '99

27 Jul '99

Hi SuSE A BIT OFF TOPIC! Is there anyone out there? Is it just me or has anyone else noticed how quiet this list has gone. Or has eneryone gone off on holiday? BTW looking forward to August 9th. John Skinner

2 1

forget passwd
by snieto＠sc.sas.cica.es 27 Jul '99

27 Jul '99

Hello, How can I enter in a linux machine if I have forgotten the root password ? Please help me.

3 2

pitfalls of upgrading 5.3 to 6.2?
by tlambley＠cspops.co.uk 27 Jul '99

27 Jul '99

Hi again all, If I was to upgrade my SuSE 5.3 to 6.2 what will actually change? I've never had to upgrade a dist' before so I've no idea what to expect. Will I have to recreate the user accounts and reinstall the software, even the packages that aren't supplied on a dist'? Will my all of my mod's to the various config files have to be redone? I recall having a lot of problems with ppp and sendmail - I'd rather not go through all that again! I run a mailing list for my local footy club and would prefer not to be offline for a couple of weeks while I reconfig everything. Any input gratefully received, T.

3 2

/var/log/messages
by alexm＠hr-60.aus.tx.i3s.net 27 Jul '99

27 Jul '99

Hi, # more /var/log/messages --> Jul 27 08:36:28 hr-60 -- MARK -- Jul 27 08:56:28 hr-60 -- MARK -- Jul 27 09:16:28 hr-60 -- MARK -- Jul 27 09:36:28 hr-60 -- MARK -- Jul 27 09:56:28 hr-60 -- MARK -- Jul 27 10:16:28 hr-60 -- MARK -- Jul 27 10:36:28 hr-60 -- MARK -- Jul 27 10:56:28 hr-60 -- MARK -- What is it refering to? thanks, -alexm

1 0

Re-entry as suse-linux-e
by jonski＠cachalot74.freeserve.co.uk 27 Jul '99

27 Jul '99

Hi To All Glad to be back in suse-linux-e land. I missed you all for a couple of days. Re-sync'ed into new list this morning. TVM to JL, JLK and SC. I was a bit worried that we might have been MicroShafted. John Skinner

1 0

CD Writing in Linux
by michael.norman1＠which.net 27 Jul '99

27 Jul '99

Hi Has anybody succeeded in getting an HP 7200i (IDE/ATAPI) cd writer recognised and working under Linux. I've tried re-compiling the kernel with SCSI Emulation etc but I have not been able to get cdrecord to recognise the drive. The drive is cofigured as slave (/dev/hdd). TIA Mike

2 1

Re: [suse-linux-e] Pt. II ATAPI zip drive mounting; non-root shutdown et. al.
by markgray＠iago.nac.net 27 Jul '99

27 Jul '99

"Rod Johnson" <rjohns(a)otenet.gr> writes: > > Hi, > > > Thanks for all the replies (especially to our resident GNU-bee...he must be > awfully busy). Even with all the help I still have a few problems. > > 1) My ATAPI ZIP drive (hdd) has an entry in fstab- > > /dev/hdd /zip auto noauto,user 0 0 Should be: /dev/hdd4 /zip auto noauto,user 0 0 (For some reason iomega partitions the zipdisk and puts the data on partition 4) [snip] > Finally, can anyone point me to a site that has up-to-date xwindows > drivers for n-Vidia cards. I contacted the STB tech support and > they sent me to a site that wanted to sell the drivers and an Xfree > version for $100. I'm sure it is a viable alternative, but if I am > going to have to pay for something, I would rather spend my money on > hardware (I bought a winmodem-stupid stupid stupid me). nvidia's website: ftp://ftp1.detonator.nvidia.com/pub/drivers/english/riva-tnt-tnt2-vanta/lin… (Their updates might be in the latest XFree86 release from ftp.xfree86.org also -- but I am happy with the nvidia one.)

1 0