[expert] HELP: Unbootable kernels after upgrade (9.1beta)
Sorry to bother you again, but is there nobody out there running Mandrake able to give me some pointers or answers so that new and old kernels can boot in the mdk9.1beta environment?? Eg. looking at the thread entitled: First 9.1beta *not* frustrating experience there seems to be problems with lilo booting. This is exactly the case for me. EXT2 does not boot and EXT3 is not even tried by the kernels/lilo... Please help... Which packages to change for a _downgrade_ to 9.0 or lower or what to do to boot the new 9.1beta: scrap lilo and go for grub or ... Thanks, Svante PS: Thank you to the people who already have replied, your help is appreciated. Svante Signell writes: Hello, I have been following the cooker development for some time now regularly upgrading. However, due to different reasons the upgrading was halted for the last half year. When starting the upgrade again in December from a situation around some time before 9.0 and halfway through I got stuck. No kernels boot any longer, the following message is displayed: EXT2-fs: ide0(3,65): couldn't mount beacuse of unsupported optional features (4). Kernel panic: VFS: unbootable mount root fs on 03:41. I have been able to mount the partitions (including /) using rescue mode kernels with ext3 fs, both for 8.1 and 9.0 CD's. All partitions have been checked and I can use chroot to simulate my system. Also the zip drive can be mounted so upgrading using packages from 8.1, 9.0 using CD and the cooker directory using the zip drive have been attempted. No success so far :-( The problem is that the kernel tries to mount the partitions in EXT2 mode and some features does not seem to be backwards compatible... Also mounting in EXT2 mode for the 8.1 CD using the 2.4.8 kernel gives the same error message as above. The /etc/modules file consist of the following entries: ext3 (no difference without) ide-floppy (for the zip) but the boot process does not come far enough for this file to be read... A couple of questions: 1. Which is the minimal list of packages needed for a succesful boot? 2. depmod -a for the wanted kernels result in a lot of unresolved symbols. Does this have any effect on the problems to boot the root partition. Following the boot process on other boxes it seems that depmod is run _after_ root is mounted, meaning that having correct dependencies are not so important _before_ the root partiion is mounted. 3. Which boot scripts are executed up to mounting of the partitions given in fstab, which packages contain these scripts and which configuration files are needed? 4. Is support for ext2 compiled in and ext3 is a module for kernels 2.4.18-16,-18, 2.4.20-2? How to get the ext3 module loaded correctly?? 5. How do you most easily switch between lilo and grub (not using any *Drake tools)? Packages installed (partial list): SysVinit-2.84-1mdk.i586.rpm lilo-22.3.4-1mdk.i586.rpm basesystem-9.0-3mdk.i586.rpm bash-2.05b-11mdk.i586.rpm e2fsprogs-1.32-2mdk.i586.rpm modutils-2.4.22-3mdk.i586.rpm glibc-2.3.1-6mdk.i586.rpmmount-2.11x-2mdk.i586.rpm glibc-profile-2.3.1-6mdk.i586.rpmnscd-2.3.1-6mdk.i586.rpm initscripts-6.91-18mdk.i586.rpm setup-2.2.0-33mdk.i586.rpm ldconfig-2.3.1-6mdk.i586.rpm timezone-2.3.1-6mdk.i586.rpm libext2fs2-1.32-2mdk.i586.rpmutil-linux-2.11x-2mdk.i586.rpm libext2fs2-devel-1.32-2mdk.i586.rpm zlib1-1.1.4-3mdk.i586.rpm Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] Unbootable kernels after upgrade
Hello, I have been following the cooker development for some time now regularly upgrading. However, due to different reasons the upgrading was halted for the last half year. When starting the upgrade again in December from a situation around some time before 9.0 and halfway through I got stuck. No kernels boot any longer, the following message is displayed: EXT2-fs: ide0(3,65): couldn't mount beacuse of unsupported optional features (4). Kernel panic: VFS: unbootable mount root fs on 03:41. I have been able to mount the partitions (including /) using rescue mode kernels with ext3 fs, both for 8.1 and 9.0 CD's. All partitions have been checked and I can use chroot to simulate my system. Also the zip drive can be mounted so upgrading using packages from 8.1, 9.0 using CD and the cooker directory using the zip drive have been attempted. No success so far :-( The problem is that the kernel tries to mount the partitions in EXT2 mode and some features does not seem to be backwards compatible... Also mounting in EXT2 mode for the 8.1 CD using the 2.4.8 kernel gives the same error message as above. The /etc/modules file consist of the following entries: ext3 (no difference without) ide-floppy (for the zip) but the boot process does not come far enough for this file to be read... A couple of questions: 1. Which is the minimal list of packages needed for a succesful boot? 2. depmod -a for the wanted kernels result in a lot of unresolved symbols. Does this have any effect on the problems to boot the root partition. Following the boot process on other boxes it seems that depmod is run _after_ root is mounted, meaning that having correct dependencies are not so important _before_ the root partiion is mounted. 3. Which boot scripts are executed up to mounting of the partitions given in fstab, which packages contain these scripts and which configuration files are needed? 4. Is support for ext2 compiled in and ext3 is a module for kernels 2.4.18-16,-18, 2.4.20-2? How to get the ext3 module loaded correctly?? 5. How do you most easily switch between lilo and grub (not using any *Drake tools)? Packages installed (partial list): SysVinit-2.84-1mdk.i586.rpm lilo-22.3.4-1mdk.i586.rpm basesystem-9.0-3mdk.i586.rpm bash-2.05b-11mdk.i586.rpm e2fsprogs-1.32-2mdk.i586.rpm modutils-2.4.22-3mdk.i586.rpm glibc-2.3.1-6mdk.i586.rpmmount-2.11x-2mdk.i586.rpm glibc-profile-2.3.1-6mdk.i586.rpmnscd-2.3.1-6mdk.i586.rpm initscripts-6.91-18mdk.i586.rpm setup-2.2.0-33mdk.i586.rpm ldconfig-2.3.1-6mdk.i586.rpm timezone-2.3.1-6mdk.i586.rpm libext2fs2-1.32-2mdk.i586.rpmutil-linux-2.11x-2mdk.i586.rpm libext2fs2-devel-1.32-2mdk.i586.rpm zlib1-1.1.4-3mdk.i586.rpm Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] X Access Problem
[EMAIL PROTECTED] writes: On Thu, 6 Jun 2002, Tibbetts, Ric wrote: Ok, this one is probably really obvious. But I'm in a really stupid state today, and can't see it I just loaded a fresh install of 8.2 on a box. I also loaded KDE 3. Everything seems to be going good, except: When I telnet to another box, and export the display back, it gets refused (yes, I ran xhost on the box first). I also cannot X -query to another host, because of the same problem. Ok, so it's a built in security issue. Looks like X is turned off for remote A few things that you might check: Is there a firewall running on either of the machines? If so, either allow the ports (IIRC, 6000) or if possible, disable the firewalling temporarily. Is name resolution working correctly between both machines? I.e., either have the entries in the hosts files or setup DNS. On a similar note, try disabling access control completely on the local machine with: xhost + then try running the app again. If it works, lock everything out with: xhost - then specifically allow the remote host. Are you exporting to the correct display? You might also try using an SSH session instead, since this may get around any blocked ports. Sorry for stepping in here, but I have been following this list on this subject for a while. My problem is probably related. Using ssh I can export the display as non-root to another machine, but not as root. Not even when starting an application, e.g. MandrakeUpdate, and entering the root password on the exported display, the application does not start. In fact _nothing_ tells you where the access violation is made. Recently, I also had problems starting X from the login console, but this one I have solved. My msec setting is 3 (normal). BTW: I'm following the cooker updates. Svante Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
RE: [expert] X Access Problem
Tibbetts, Ric writes: msec... Level 1 or above, and only localhost can access X -=SPLAT=- I just got hit with a BFO (Blind Flash of the Obvious). Thanks, I needed that! -Original Message- From: et [mailto:[EMAIL PROTECTED]] MSEC? level 3 or above? man msec says this: You must be root to run msec . Launch msec x to set you security level to x (x=[0-5]). It'll modify your system according to security level x features. For a fine description of each security level, consult the documentation under /usr/share/doc/msec-0.19/security.txt. If you want to make changes to the current level, use /etc/security/msec/perm.local to override the permis sions/owners/groups and /etc/security/msec/level.local to override the rules (see mseclib(3) for details). How to just change the X access from a remote box usung ssh (nothing else) and still keeping msec at level 3? Maybe its just RTFM (or security.txt). My Mandrake box is currently in power-off state. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] root partition and ext3
Hi, After successful transfer of all partitions to ext3, using tune2fs -j /dev/hda?, I thought everything was OK, also for the root partition. But its NOT! While having ext2 compiled into the kernel and ext3 as a module, as with kernel-2.4.18.16, the root partition is mounted as ext2, not ext3. This has been verified when rebooting after problems with shutdown (especially with upgrading glibc etc) when the root partition is uncleanly unmounted. How to resolve this, the fstab entry is useless since it's on the not yet mounted / partition...? I assume mounting of the root partition is taken care of by the kernel.. mount /dev/hda1 on / type ext3,ext2 (rw) ... cat /etc//fstab: /dev/hda1 / ext3,ext2 defaults 0 1 ... Thanks, Svante Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] Host SMBus controller not enabled!??
Hello, I'm trying to get lm-sensors to work on a Compaq 5670 Desktop running Mandrake Cooker (kernel-2.4.18) with partial success. Modules identified with sensors_detect and loading OK are: i2c-dev 4096 0 (unused) i2c-voodoo3 2824 0 (unused) eeprom 3104 0 (unused) adm1021 5408 0 (unused) i2c-proc6496 0 [eeprom adm1021] tvaudio10080 1 (autoclean) tuner 8612 1 (autoclean) msp340014288 1 (autoclean) bttv 59776 0 (unused) i2c-algo-bit7244 3 [i2c-voodoo3 bttv] i2c-core 13568 0 [i2c-dev eeprom adm1021 i2c-proc tvaudio bttv tuner msp3400 i2c-algo-bit] videodev4896 3 [bttv] The only module that fails to load is: i2c-piix4.o Some of the modules are used by a Hauppauge WinTV card and the ones detected by sensors-detect are: i2c-piix4, i2c-voodoo3, eeprom, adm1021. I found nothing about this problem in the FAQ, and trying different BIOS settings did not change anything (The Compaq BIOS doe not have many options available). Thanks in advance, Svante Signell root modprobe i2c-piix4 /lib/modules/2.4.18-7mdk/kernel/drivers/i2c/i2c-piix4.o.gz: init_module: No such device Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters modprobe: insmod /lib/modules/2.4.18-7mdk/kernel/drivers/i2c/i2c-piix4.o.gz failed modprobe: insmod i2c-piix4 failed rootdmesg i2c-piix4.o version 2.6.2 (2008) i2c-piix4.o: Found PIIX4 device SMBUS: Error: Host SMBus controller not enabled! i2c-piix4.o: Device not detected, module not inserted. rootlspci -v 00:00.0 Host bridge: Intel Corp. 440BX/ZX - 82443BX/ZX Host bridge (rev 02) Flags: bus master, medium devsel, latency 64 Memory at 4400 (32-bit, prefetchable) [size=64M] Capabilities: [a0] AGP version 1.0 00:14.0 ISA bridge: Intel Corp. 82371AB PIIX4 ISA (rev 02) Flags: bus master, medium devsel, latency 0 00:14.1 IDE interface: Intel Corp. 82371AB PIIX4 IDE (rev 01) (prog-if 80 [Master]) Flags: bus master, medium devsel, latency 64 I/O ports at 2420 [size=16] 00:14.3 Bridge: Intel Corp. 82371AB PIIX4 ACPI (rev 02) Flags: medium devsel, IRQ 9 Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] Sites with up to date apt *.cooker files?
Hello, After trying to use the Mandrake update tools urpmi, rpmdrake, etc and finding them more or less unusable (at least to follow cooker), I tried apt(-rpm) and these tools are. Or should be, as expected from the apt tools in Debian (used on my other boxes; no probs eg to follow unstable), _if_ the ftp sites had updated *.cooker files! But they do not, for example ftp.sunet.se shows: mandrake-devel/cooker/i586/Mandrake: RPMS.cooker 0 04/28/01 02:00:00 mandrake-devel/cooker/i586/Mandrake/base: pkglist.cooker 10719066 12/24/01 20:54:00 srclist.cooker 883823 07/27/01 02:00:00 The same situation applies for other Mandrake mirrors (and other sites). Whatever the vendors say, now when apt is ported to rpm based distributions (by connectiva), why can't we use best tools available for the task at hand, in this case to follow the unstable (or development branches) of Debian, Mandrake, Redhat etc. The situation is maybe different for people purchasing a distribution and doing bug/security/etc updates from there, using Mandrake rpmdrake, redhat up2date etc, thats another issue. Just a thought, sincerely Svante (Recently replaced a box running Redhat RawHide with Debian unstable/sid. Still keeping the box with Mandrake Cooker, but who knows?). (Also, maybe apt should be moved from RPMS2 to RPMS, at least for cooker??) Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Playing DVD's
There is also the VideoLan Client (and servers). rpms and debs are available. http://www.videolan.org/ Tobias Marx writes: Larry Sword wrote: Ken Thompson wrote: What is the group opinion of the best program to play DVD's?? I just got an older 5spd DVD to fool around with and now the Number 1 son wants to watch movies G. Any tips to make it work would also help a bunch -- Ken Thompson, North West Antique Autos Payette, Idaho Email: [EMAIL PROTECTED] http://www.nwaa.com ICQ 988809 Sales and brokering of antique autos and parts. xine-0.4.3.d4d-1 i've got better experiences with ogle (more stable). speed is about 95% of windvd on (you guess it) windows. just install the rpms from here (those for rh7.x work just fine on my mdk8.1 box): http://www.dtek.chalmers.se/groups/dvd/ you'll need: libdvdcss libdvdread liba52 Ogle Ogle_gui have fun! -- Free Dmitry Sklyarov, NOW! Fuck the DMCA! Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] ssh reverse mapping problem
Shall I take this seriously? Which programs to check? Where to look? Feb 1 20:20:10 computer sshd[864]: WARNING: /etc/ssh/primes does not exist, using old prime Feb 1 20:20:15 computer sshd[864]: reverse mapping checking getaddrinfo for aa.bb.cc failed - POSSIBLE BREAKIN ATTEMPT! Feb 1 20:20:15 compaq sshd[864]: Accepted password for user from xx.yy.zz port 1020 ssh2
[expert] 2.4.0 file lock support?
When installing packages with 2.4.0 kernels running, the following error occurs: Update-menus[2403]: Cannot lock /var/run/update-menus.pid: No locks available Aborting. (not present for 2.2.x kernels) uname -a Linux ... 2.4.0-5mdk #1 Mon Jan 15 16:05:02 CET 2001 i686 unknown rpm -qf /usr/bin/update-menus menu-2.1.5-45mdk rpm -q rpm rpm-4.0-15mdk
[expert] Re: [Dri-devel] DRI direct sloow, how to debug/trace (please)?
Nathan, thanks for your reply. See below. Nathan Hand writes: On Tue, Jan 16, 2001 at 12:58:16AM +0100, Svante Signell wrote: Q3A: OK, no problem, using libGL.so.1,2 from Mesa-3.4 HG2: OK Heretic 2: OK after full screen patch 1.06a sof-demo: OK descent3-demo: not OK In what way is descent 3 not OK? Back later on this one, problably same as for UT (SW fallbacks?). ut(436): not OK. The introductory video 'never' finishes. Sound hacks on and off and video runs at normal speed, the suddenly slows down, normal speed again, slows down, creeps...ESC. Same when in game mode. Since the Glide libraries in UT requires libglide.so.2 and this is not supported in X4, this option is no alternative!? You can install both glide libraries, but you'll have to close X in order to use a glide application. ... Please, anyone give me a hint how to proceed the tracking down the causes of problems. More info on request. UT is just slow. But you can check the following... $ cd /home/nathanh/.loki/ut/System $ emacs UnrealTournament.ini In the section marked [Engine.Engine] make sure you have GameRenderDevice=SDLGLDrv.SDLGLRenderDevice WindowedRenderDevice=SDLGLDrv.SDLGLRenderDevice RenderDevice=SDLGLDrv.SDLGLRenderDevice Yes, I have. In the section marked [SDLGLDrv.SDLGLRenderDevice] you need OpenGLLibName=/usr/lib/libGL.so.1 Yes, the same libGl.so.1.2 as for Q3A, see above, is used. Finally, make sure you're using 16bpp mode. Unreal Tournament is unbearably slow in 32bpp mode. Yes, I'm using 16 bpp. I don't beleive UT would be this slow. The driver seems to switch back and forth between SW and HW mode?? The previous Glide+3.3.x solution on the same computer was running at a reasonable speed, at least at 640x480x16bits. On another computer (2x450MHz) with a G400 DH no max, I don't experience this problem: getting around 15 fps avg @800x600x16bit for utbench. /Svante
[expert] DRI direct sloow, how to debug/trace (please)?
Hello list, I've been trying for a while now to get my 3dfx Voodoo Banshee to work in openGL mode together with X4/DRI and 2.2/2.4 kernels. Starting with MDK 7.2 (X-4.0.1, kernel-2.2.17, glibc-2.1) and upgrading to the cooker distribution (X-4.0.2, kernel 2.4.0/2.2.18, glibc-2.2) still some games (such as UT) don't run fully HW accelerated :( For X-3.3.x Glide was a usable alternative for those games supporting glide2x/3x. The situation is as follows: glxinfo reports direct rendering gears: OK around 600 fps Q3A: OK, no problem, using libGL.so.1,2 from Mesa-3.4 HG2: OK Heretic 2: OK after full screen patch 1.06a sof-demo: OK descent3-demo: not OK ut(436): not OK. The introductory video 'never' finishes. Sound hacks on and off and video runs at normal speed, the suddenly slows down, normal speed again, slows down, creeps...ESC. Same when in game mode. Since the Glide libraries in UT requires libglide.so.2 and this is not supported in X4, this option is no alternative!? The following list is what I'm suspecting to be problematic: 1. IRQ? Both the sound card and the Banshee (AGP) card share the same interrupt, 5. Earlier a 100MB Ethernet card was in the slot, _also_ sharing IRQ 5. Removing the NIC did not change anything, now the NIC is on IRQ 3 and the slot with IRQ 5 occupies a pcTV card. 2. Sound card? The sound card (on board) shares the same IRQ 5 with the Banshee card. Unloading all sound drivers before starting X did not change anything. 3. Memory/disk swapping? For the problematic games more than 100 % of total memory (128 MB) is used. If virtual memory is used swapping to disk slows down?? Also two processes are running, is this normal (single CPU, 450MHz)? 4. AGP? It it normal for the 4.0.2 Banshee driver not to use agpgart? # /sbin/lsmod ModuleSizeUsed by tdfx 37360 2 agpgart 18960 0 [tdfx] | not used?? Please, anyone give me a hint how to proceed the tracking down the causes of problems. More info on request. Thanks in advance, Svante
Re: [expert] hdd errors with kernel 2.4.0(att:: Svante Signell)(solved)
civileme, Thank you for your help. Trying with the noautotune option made the 2.4-kernel successfully boot without errors and hdparm -t /dev/hda gave 12.50 MB/sec. After that timing of hdb resulted in the 'hdb: dma_intr: ...' errors and gave 4.3 MB/sec. The log file showed: ... hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error } hdb: dma_intr: error=0x84 { DriveStatusError BadCRC } hda: DMA disabled Apparently the first (master) disk shuts off DMA mode and goes to pio!? After that also the first disk was back to 4.3 MB/sec. Encouraged by the speeds available I took a closer at the cable connecting the two disks. Due to cable length problems the cable sitting there is a maxitower cable with four disk contacts (only two are needed). Changing the contact for the second (slave) disk and positioning the cable solved the problem, hooray! Now both disks run with DMA and speeds between 12 and 12.5 MB/sec, and the noautotune option to ide0 is not needed any longer for booting kernel-2.4. /Svante civileme writes: On Sunday 07 January 2001 19:09, you wrote: civileme, Enclosed please find the information requested. The boot log is from kernel-2.4.0-0.15mdk obtained by logging in remote via ssh (it takes a loong time to get the bash prompt). Still cannot login from the console :( /Svante Hrm, I would say try installing with linux ide0=noautotune ide1=noautotune Obviously the kernel is overzealous in setting the disks to the highest rate they'll go, so let's tell it to stop that behavior Just stick that into the append line for lilo or grub append="ide0=noautotune" should do it. If it autotunes your other devices, that's OK. Civileme
[expert] How to switch between lilo and grub?
The last months or so DrakConf has been updating the grub /boot/grub/menu.lst and syncing with /etc/lilo.conf. Suddenly today I got the lilo boot menu instead the grub boot menu. How to change from lilo to grub and vice versa? Settings in DrakConf?
[expert] Cannot login to 2.4.0 kernels!
I'm currently upgrading from MDK7.2 (glibc-2.1) to the cooker development distribution (glibc-2.2). In this process user login with kernels 2.4.0-0.x is non-functional. Also a an older 2.3.48 kernel shows the same behavior. Everything seem OK, but after supplying username and password _nothing_ happens until I'm back at the login prompt after a _long_ time-out. Remote logging in works, but also here with a long delay before getting the bash prompt. No problem occurs with 2.2.x kernels! Which packages are involved in the login process? I've upgraded the following ones, which packages has still to be upgraded (and which are redundant, wrt the login phase)? MAKEDEV-23.0.6-4mdk bash-2.0.4-15mdk getty_ps-2.0.7j-13mdk libncurses5-5.2-12mdk mingetty-0.9.4-14mdk mgetty-1.1.22-2mdk modutils-2.3.23-3mdk ncurses-5.2-12mdk pam-0.72-14mdk perl-5.6000-24mdk pwdb-0.61-4mdk shadow-utils-2826-3mdk util-linux-2.10r-2mdk
Re: [expert] kernel-2.4 hdd errors
Below are the results of hdparm running 2.2.18-1mdk. (I cannot login to the 2.4.0-0.15mdk kernel, see separate posting. My HUB is also broken so remote login is not possible either) I have not found any boot-up message indicating ide0=autotune, nor any files with this text. Where to look for this, specific to kernel 2.4? Note that I'm running the disks with udma2 for the 2.2.18 kernel, see below. No problems for kerenels-2.2.x! /boot/grub/menu.lst does not set autotune for ide0: title mdk72++ kernel (hd1,0)/boot/vmlinuz-2.4.0-0.15mdk root=/dev/hdb1 #hdparm -i /dev/hdb /dev/hdb: Model=IBM-DTTA-351010, FwRev=T56OA73A, SerialNo=WF0WFJJ2557 Config={ HardSect NotMFM HdSw15uSec Fixed DTR10Mbs } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=34 BuffType=DualPortCache, BuffSize=466kB, MaxMultSect=16, MultSect=16 CurCHS=17475/15/63, CurSects=16513875, LBA=yes, LBAsects=19807200 IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 DMA modes: sdma0 sdma1 sdma2 mdma0 mdma1 mdma2 udma0 udma1 *udma2 #hdparm -t /dev/hdb /dev/hdb: Timing buffered disk reads: 64 MB in 14.83 seconds = 4.32 MB/sec #hdparm -X66 /dev/hdb /dev/hdb: setting xfermode to 66 (UltraDMA mode2) #fdparm -t /dev/hdb /dev/hdb: Timing buffered disk reads: 64 MB in 14.78 seconds = 4.33 MB/sec civileme writes: On Saturday 30 December 2000 11:54, you wrote: Svante Signell wrote: I'm getting errors like this when booting kernel 2.4.0-x, eg x=0.14mdk They does not happen with kernels 2.2.x. Do I have to compile a kernel on my own, not using the precompiled ones? hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error } hdb: dma_intr: error=0x84 { DriveStatusError BadCRC } more of the above... You need to remove any autotune instructions to the kernel in the boot-up procedures they will look like "ide0=autotune" You need to use hdparm -i on the drives to see what mode they are running Use hdparm -t to test the speed Use hdparm -Xxy where xy=64+udma mode to set the udma mode lower Keep testing speed as you change the settings. Your disks are set faster than they are capable of reliably running--and they are set as the manufacturer advertised them. Civileme
Re: [expert] IRQ steering?
Thank you Bug! for your reply, The three devices using IRQ 5 are: Sound card on main board: 00:06.0 Multimedia audio controller: ESS Technology ES1968 Maestro 2 Subsystem: Compaq Computer Corporation: Unknown device b0b8 Flags: bus master, medium devsel, latency 64, IRQ 5 I/O ports at 2000 Ethernet controller card on PCI bus: 00:03.0 Ethernet controller: Lite-On Communications Inc LNE100TX (rev 20) Subsystem: Lite-On Communications Inc LNE100TX Flags: bus master, medium devsel, latency 66, IRQ 5 I/O ports at 2800 Video card on AGP (=PCI?) bus: 01:00.0 VGA compatible controller: 3Dfx Interactive, Inc. Voodoo Banshee (rev 03) (prog-if 00 [VGA]) Subsystem: Guillemot Corporation Maxi Gamer Phoenix Flags: VGA palette snoop, 66Mhz, fast devsel, IRQ 5 Memory at 4000 (32-bit, non-prefetchable) Memory at 4200 (32-bit, prefetchable) I/O ports at 1000 Capabilities: [54] AGP version 1.0 I'm using X4.0.2 with DRI enabled. According to the log files and glxinfo the card is HW accelerated. However, even though Q3A and gears run OK, HG2, Heretic2, UT does _not_. Do the Banshee (tdfx) driver require an IRQ? If not everything should be OK, or can three devices share the same IRQ? W98 on the same computer have no problems with three devices at IRQ5. Further comments below. Thanks, Svante Bug Hunter writes: If you have PCI cards, IRQ sharing will work. Otherwise, you may have to force the card not to be PNP and then tell it what IRQ to be on. You usually have to run the DOS setup disk to get this to happen. I've not found any option to control the PNP behavior in the BIOS. Do the IRQ steering apply to ISA cards? This computer does not have any ISA slots :( On Fri, 29 Dec 2000, Svante Signell wrote: Anyone knows how IRQ 5 is chosen both for my eth0 (tulip, external card) driver and sound (maestro, on motherboard) on my Compaq Presario 5640/5670 when other interrupts are available: 4,6,7,9,10,11? ... How to steer away one of the units from IRQ 5 to avoid interrupt sharing/conflicts? In the documentation to both the tulip and maestro driver I have not found any option to control the IRQ. What unit takes care of allocating IRQ's if the card driver does not require one, the kernel or the BIOS? The BIOS for this computer has very limited number of options, and as mentioned above the sound hardware is on the motherboard. Also only two PCI slots are available (one with the ethernet card, the other with a PCTV card, IRQ 3), so shuffling cards is not much to try.
RE: [expert] IRQ steering?
Thank you JMS for your very informative reply. Further comments/questions below. Se also my reply to Bug Hunter (funny name :). Thanks, Svante [EMAIL PROTECTED] writes: Initially the BIOS handles setting up the IRQ's for your cards as the system comes up. The OS can modify this allocation as needed, within certain limits. The problem is that you are running up against one of the most often mis-understood limits. Here goes: Yes, as people here will indicate, you can CHANGE the IRQ utilized for a card, -BUT- in so doing you will not change the assignment of that one card. You will change the assignment of MULTIPLE devices or slots! Huh? PCI was initially designed to handle only three devices. As a result as manufacturers started adding PCI slots, they took to wiring two or more PCI slots together. Thus your AGP slot almost always shares an IRQ with the first PCI slot. Why? Well in their infinite wisdom, they deemed that if you were using an AGP video card, you would NOT also be using a video card in slot one and vice versa. Ugh, so you normally can use one or the other, but not both without resorting to some trickery called IRQ Steering. On newer motherboards two or more PCI slots may share an IRQ allocation with integrated devices. Thus, in your case the PCI slot you put the ethernet card into WILL ALWAYS share the same IRQ with your motherboard's integrated audio. NO MATTER WHAT YOU DO, this will be the case. You can change the allocation to say 11, but now your audio and eth0 will go to 11... ugh. This would be OK if the video card does not require an IRQ, or can more than two devices share the same IRQ? (W98 on the same machine have no problems with three devices at IRQ5) I'm using X4.0.2 with DRI enabled. According to the log files and glxinfo the card is HW accelerated. However, even though Q3A and gears run OK, HG2, Heretic2, UT does _not_. Do the Banshee (tdfx) driver require an IRQ? If not everything should be OK, or something else is causing the complete slowdown, maybe DMA? Your only choice is to move the ethernet card to another slot to avoid sharing IRQ's. ISA PNP cards don't suffer from this. In a way they are more flexible with resource allocation. What happens if I exchange the Ethernet card (IRQ 5) with the PCTV card (IRQ 3)? Then according to your info IRQ 5 will be allocated to the sound device, the TV card and the Banshee (AGP) card and IRQ 3 to the Ethernet card. If all devices requires an IRQ, no solution is obtained this way!? If you are using an ISA PNP card you might be able to change the IRQ it uses with software provided by the manufacturer... This software modifies the "hint" that the card reports to the bios during boot. The preferred IRQ is stored on the card itself, and the bios sets up the card to use the IRQ "hinted" at by the on board NVRAM (or in some cases other forms of non-volatile RAM). No ISA slots are avilable on this computer :( Since you've already stated that you have limited slots available, your only recourse is to either move things around, remove a device/card, get a new non-conflicting device or card, or a new computer. Welcome to the world of IBM based design architecture. Needless to say this all needs a MAJOR overhaul, but who is going to set the standards? -JMS -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Svante Signell Sent: Friday, December 29, 2000 11:15 AM To: [EMAIL PROTECTED] Subject: [expert] IRQ steering? Anyone knows how IRQ 5 is chosen both for my eth0 (tulip, external card) driver and sound (maestro, on motherboard) on my Compaq Presario 5640/5670 when other interrupts are available: 4,6,7,9,10,11? ...
[expert] IRQ steering?
No response so far, trying again. Anyone knows how IRQ 5 is chosen both for my eth0 (tulip, external card) driver and sound (maestro, on motherboard) on my Compaq Presario 5640/5670 when other interrupts are available: 4,6,7,9,10,11? Also the graphics card is allocated to IRQ5: 01:00.0 VGA compatible controller: 3Dfx Interactive, Inc. Voodoo Banshee (rev 03) (prog-if 00 [VGA]) Subsystem: Guillemot Corporation Maxi Gamer Phoenix Flags: VGA palette snoop, 66Mhz, fast devsel, IRQ 5 How to steer away one of the units from IRQ 5 to avoid interrupt sharing/conflicts? In the documentation to both the tulip and maestro driver I have not found any option to control the IRQ. What unit takes care of allocating IRQ's if the card driver does not require one, the kernel or the BIOS? The BIOS for this computer has very limited number of options, and as mentioned above the sound hardware is on the motherboard. Also only two PCI slots are available (one with the ethernet card, the other with a PCTV card, IRQ 3), so shuffling cards is not much to try.
[expert] kernel-2.4 hdd errors
I'm getting errors like this when booting kernel 2.4.0-x, eg x=0.14mdk They does not happen with kernels 2.2.x. Do I have to compile a kernel on my own, not using the precompiled ones? hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error } hdb: dma_intr: error=0x84 { DriveStatusError BadCRC } more of the above...
[expert] IRQ steering?
Anyone knows why is IRQ 5 chosen both for my eth0 (tulip, external card) driver and sound (maestro, on motherboard) on my Compaq Presario 5640/5670 when other interrupts are free: 4,6,7,9,10,11? How to steer away one of the units from IRQ 5 to avoid interrupt sharing? In the documentation to both the tulip and maestro driver I have not found any option to control the IRQ. What unit takes care of allocating IRQ's if the card driver does not require one, the kernel or the BIOS? The BIOS for this computer has very limited number of options, and as mentioned above the sound hardware is on the motherboard. Also only two PCI slots are available (one with the ethernet card, the other with a PCTV card, IRQ 3), so shuffling cards is not much to try.
[expert] Re: [Dri-devel] DRI direct sloow
Nathan Hand writes: Svante Signell writes: Thank you all for your replies. The gears problem is solved, but the slowness still remains, see below. setting FX_GLIDE_SWAPINTERVAL=0 make gears jump to 611 fps q3a demo001 is OK, both with libMesaVoodoGL (33.9 fps) from Q3A 1.11 and libGL from XFree86-4.0.1h (31.8 fps) But ut, heretic2, hg2, sof-demo all seem to render in SW (or at creeping speeds in HW). Especially starting the games is really slow. For sof-demo and hg2 no sound as user, but as root running X (not as root in X terminal). Can the problems be due to the sound card or the mouse? Possibly, but unlikely. More likely the ut, heretic2, hg2, sof binaries are finding the wrong libGL.so. At least heretic2 you can use +set gl_driver to tell it explicitly which libGL.so to use. Otherwise go into the games directory for each game and symlink libMesa's and libGL's to the /usr/lib/libGL.so (DRI version). I have already linked the real /usr/X11R6/lib/libGL.so.1.2 to libGl.so for heretic. Today I also tried the +set gl_driver by editing the config.cfg file. Same as before: The game start up in windowed mode (not full screen, as edited in the config file??). It refuses to go to full screen! Most noiceably, there is a delay between the key pressing and sound when moving in the setup menus. Last in this mail, please find the startup log of heretic 2 and the config file. There is nothing wrong with your setup now. The 611fps in gears proves this. Your games not being accelerated means they've got or found the wrong libGL.so. I don't think so. Something else is wrong. Why is Q3 working flawlessly (except for hacking sound)? heretic_startup.log heretic_config.cfg Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Re: [Dri-devel] DRI direct sloow
Mark B. Allan writes: Svante Signell wrote: ... But ut, heretic2, hg2, sof-demo all seem to render in SW (or at creeping speeds in HW). Especially starting the games is really slow. For sof-demo and hg2 no sound as user, but as root running X (not as root in X terminal). Can the problems be due to the sound card or the mouse? Last in this mail also find the glxinfo output. (This program was copied from RH, I did not find it in the Mesa-demos-3.4-4mdk package, hint MDK, Mesa is not installed, to avoid dual GL libs) Check /usr/X11R6/lib for libGLwrapper.* - it's a Mandrake invention which is supposed to make switching between the various GL implementations easier. My experience has been that it just screws things up more - you can kill it without any problems (although back it up just in case). The libGLwrapper.* library is in the Mesa-3.4-4mdk package which not installed, see above. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Re: [Dri-devel] DRI direct sloow
Thank you all for your replies. The gears problem is solved, but the slowness still remains, see below. setting FX_GLIDE_SWAPINTERVAL=0 make gears jump to 611 fps q3a demo001 is OK, both with libMesaVoodoGL (3398 fps) from Q3A 1.11 and libGL from XFree86-4.0.1h (31.8 fps) But ut, heretic2, hg2, sof-demo all seem to render in SW (or at creeping speeds in HW). Especially starting the games is really slow. For sof-demo and hg2 no sound as user, but as root running X (not as root in X terminal). Can the problems be due to the sound card or the mouse? Last in this mail also find the glxinfo output. (This program was copied from RH, I did not find it in the Mesa-demos-3.4-4mdk package, hint MDK, Mesa is not installed, to avoid dual GL libs) /Svante Nathan Hand writes: On Mon, Dec 11, 2000 at 09:03:37AM +0100, Svante Signell wrote: ... Hello, can somebody please enlighten me on why DRI is so terribly slow, eg gears run at 60-70 fps ( 500 is expected). glxinfo report direct rendering and all seems OK according to the XFree86.0.log file (attached). The swapbuffers procedure is tied to the vertical refresh on the glide cards (tdfx DRI driver included). This will limit your effective frame rate to 70fps. You can disable this with an environment variable. $ export FX_GLIDE_SWAPINTERVAL=0 $ gears You should get ~600fps in the Mesa gears demo, given your hardware. OS: MDK 7.2 + parts from mandrake-devel (eg XFree86-4.0.1h) Computer: Compaq 5640/5670, 450 MHZ, 128MB RAM. Video card: Voodoo Banshee glxinfo.out Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] DRI direct sloow
Content-Type: text/plain; charset=us-ascii Content-Description: message body text Content-Transfer-Encoding: 7bit Hello, can somebody please enlighten me on why DRI is so terribly slow, eg gears run at 60-70 fps ( 500 is expected). glxinfo report direct rendering and all seems OK according to the XFree86.0.log file (attached). OS: MDK 7.2 + parts from mandrake-devel (eg XFree86-4.0.1h) Computer: Compaq 5640/5670, 450 MHZ, 64MB RAM. Video card: Voodoo Banshee Note: the number of BIOS functions in this computer is _very_ limited. For example, there is no option to enable an IRQ for the video card :( Workarounds? attachement Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Grub vs lilo
Is it sufficient to edit /etc/lilo.conf and the changes are directly reflected in the boot menu? What about /boot/grub/menu.xxx It seems that lilo is non-functional. Why is it still installed? Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Networking problems(solved)
Lars Nordin writes: 1) Check the rc script for syslog. /etc/rc3.d/S**syslog linked to /etc/rc.d/init.d/syslog Thanks, fixed in ntsysv. The syslog daemon was not started at boot. 2) autologin? (Do you mean the graphical login? login as root start X-Windows and in the DrakeTools is an icon for modifying the X-Windows login screen) No, you are logged in as the user chosen in the upgrade menu (with X running and all). BTW, why is graphical login so terribly slow? I have time to log in as an odrinary user before X starts? - Original Message - From: "Svante Signell" [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Friday, December 01, 2000 6:07 PM Subject: Re: [expert] Networking problems(solved) Thank you Lars for your help, (I also appreciate help from EagleIce regarding boot levels) Now the install is OK and the upgrades are made (including kdebase). Networking is working too. Next step: 1. No log files (like messages, kernel.log etc) are created. Where to change that? 2. Where to disconnect/change user of autologin (except rpm -e) Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Broken upgrade 7.1 to 7.2
Boot logs show that char-major-4 and char-major-180 are missing. These corresponds to tty and usb (right?). The device files are there. Which module files correspond to these devices? Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Grub vs lilo
My question is more oriented how lilo and grub work together (or don't). Last time I tried lilo was in 7.1, where the lilo.conf file was edited and updated (I thought) by a lilo run. But what happened was not as expected, lilo did not upgrade according to the conf file. On the other hand, at the next boot grub came up again and the changes made in lilo.conf were reflected in the grub boot menu. Same happened after the upgrade. I edited the lilo.conf and the changes were reflected in grub boot menu (after running the Drake tool, don't remember the name). So, where do you change between lilo and grub, and which files are config files for the two tools in MDK7.2? (I've been using lilo for a long time now, but I am a newbie in grub.) Michael writes: Unless you told it to uninstall lilo why wouldn't it be installed? Linux is usually better about actually waiting for you to tell it what you want before it does it. Also a lot of times programs leave modified config files when uninstalled so if you need them later you have them. I still run LILO at home just because it's been functioning fine on this machine for several years and I see no reason to play with it but I use Grub on my new servers. It seems a lil more work to config but is overall slightly nicer. *^*^*^* Have the courage to take your own thoughts seriously, for they will shape you. -- Albert Einstein On Sun, 3 Dec 2000, Svante Signell wrote: Is it sufficient to edit /etc/lilo.conf and the changes are directly reflected in the boot menu? What about /boot/grub/menu.xxx It seems that lilo is non-functional. Why is it still installed? Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Log files?
The syslog daemon was not started at boot time. Solved by running eg. ntsysv. Mark Weaver writes: I would be seriously interested in finding out the answer to this myself, since I've recently downloaded the ISO's for Mandrake 7.2 and "had" planned on installing them as soon as I get the CD's burned. I'm not so sure if there are weird things like that going on. I've come to really rely on those logs as a source of system information as to what's going on on my system. -- Mark .. On Sun, 3 Dec 2000 Svante Signell had this to say! After upgrade from 7.1 to 7.2 no (interesting) log files are created any longer in /var/log. Is logging disabled or are the logs written somewhere else? If they are disabled, where to enable them? Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Getting 3D to work?
I need advice how to get XFree86-4.0.x to work in 3D with my Banshee graphics card. Several alternatives are possible!? 1. Stay with Glide_V3 Device3Dfx as with 3.3.x. Does this still work? 2. Download Glide_V3-DRI, tdfx_dri tdfx_drm from linux.3dfx.com and following instructions there. 3. Stay with Glide_V3-DRI-cvs XFree86-4.0.1 with glide-module. This combination does not work. I'm getting X server hangs. 4. Upgrade to the Glide_V3-DRI-cvs XFree86-4.0.1g from the development dirctories. (4.0.1g runs like charm on my debian-woody box with a G400.) 5. Other solutions/suggestions? Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] XFree86-4 questions
Finally took the step to leave 3.3.x for 4.0.x. Some questions: 1. Got HW accelaration almost working (4.1.0, Banshee). Quake3 starts accelerated, but exits due to errors with 4.0.1, i.e. with MDK 7.2 and libgL. Should I go for Glide, or GL? 2. Another alternative is to upgrade to cooker X4 (4.0.1e??) and use ligGL. Is this workable, or is Glide still the solution? 3. Does Quake need to be upgraded to 1.17? 4. Which version of UT works? 5. Anuthing else? Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Networking problems(solved)
Thank you Lars for your help, (I also appreciate help from EagleIce regarding boot levels) Now the install is OK and the upgrades are made (including kdebase). Networking is working too. Next step: 1. No log files (like messages, kernel.log etc) are created. Where to change that? 2. Where to disconnect/change user of autologin (except rpm -e) 3. XFree86 V4 Banshee, separate posting. 4. grub: is it sufficient to edit /etc/lilo.conf? What about /boot/grub/menu.?? lilo is non-functional? Why is it installed? Thank you for your help, maybe I won't switch to another distribution. I'm still not fluent enough in Mandrake;( Lars Nordin writes: Like I posted before check the links in /etc/rc.d/rc3.d or /etc/rc.5d/ for the script "network" in /etc/rc.d/init.d. I had the same problem upgrading. I was able to get the network running again by simply running "ln -s /etc/rc.d/network/etc/rc.d/rc5.d/SXXnetwork" where 'XX' is some two digit number. Run 'tksysv' if you like/need to use a graphical tool to fix this. - Original Message ----- From: "Svante Signell" [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Thursday, November 30, 2000 4:48 PM Subject: [expert] Networking problems After upgrade from 7.1 to 7.2 networking in non-functional. Any pointers? Everything is a mess, the upgrade did not complete either. Maybe it's time to switch to another distribution, I am using debian and RH on other boxes? This graphical/framebuffer thing is maybe good for newbies, but I feel like I've lost control of the configuration :( Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list. Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] How to disable graphical login?
After upgrade from 7.1 to 7.2 the graphical login is even slower to start :( What package is running the graphical shell? Got rid of gdm, but kdm is still there. The problem is that kdm is in kdebase and not easily removed. How to disable? I'd like to configure from scratch again. (Don't like linuxconf either) Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Networking problems
After upgrade from 7.1 to 7.2 networking in non-functional. Any pointers? Everything is a mess, the upgrade did not complete either. Maybe it's time to switch to another distribution, I am using debian and RH on other boxes? This graphical/framebuffer thing is maybe good for newbies, but I feel like I've lost control of the configuration :( Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] 7.2 upgrade problems
Upgrading from 7.1 to 7.2 (.iso files) goes well until the configuration of X. Probing is OK and everything is well, but finishing gives the following error message: /mnt/dev/null/.kderc not a directory and a hang here. Well we all know that dev files are not directories :( Something wrong with the Perl script for the installer? I had to kill the install process and manually reboot. 7.2 seems to run, but in first attempt the boot hangs on NFS (no network at boot time), solved by uninstalling nfs-utils-*. Also complaints on char-major-4 not found and default.png also not found for gnome to start. More to come... BTW: The upgrade of .rpms (500 MB) took 5.5 hours!! Even if my first hard disk is very slow, should it take that long? Going from 7.0 to 7.1 did not. /Svante Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Where are ALL the SRPMS?
How many SRPMS for MDK-7.1 do you find at: ftp://ftp.sunet.se/pub/Linux/distributions/mandrake/SRPMS/ You are referring to the mandrake-devel directory not 7.1. Alexander Skwar writes: On Sun, Aug 27, 2000 at 12:58:11AM +0100, Fergal Daly wrote: Hi, I've searched the archives of the mailing lists and seen this question arise several times but it doesn't seem to have ever been answered properly! Either the question has been ignored or the link provided isn't any use. The SRPMS directory on the mirror sites only has about 100 packages. Can anyone tell me where all the others are? Thanks, Dunno what you're talking about, the main mirror at ftp://ftp.sunet.se/pub/Linux/distributions/mandrake-devel/SRPMS contains right now 936 SRPMS. Don't know, but it seems like you're using a quite outdated mirror. Alexander Skwar -- Homepage:http://www.digitalprojects.com | http://www.dp.ath.cx Sichere Mail?Mail an [EMAIL PROTECTED] fuer GnuPG Keys ICQ: 7328191
[expert] HW accel with mdk7.1??
Need advice how to reenable HW acceleration on my Comapq with mdk71 and a Voodo Banshee card. What do you propose to install: mkd71 linux.3dfx.com = == Device3Dfx-2.3.4-5mdk.src.rpm Device3Dfx-2.3-5.src.rpm Glide_V3-2.3.4-5mdk Glide_V3-2.60-16 or should I use tdfx_dri, tdfx_drm, Glide_V3-3.10-6 from 3dfx?? Do I have to manually insmod the module 3dfx.o? Is Mesa in 7.1 (3.1-14) compiled for HW support. If not do I have to recompile it? Before 7.1 XFree86-3.3.x was distributed without problems for my 3dfx cards( banshee and voodoo2). Now with 7.1 XFree86-4.0 is mixed with 3.3.x servers. All this is confusing. Any FAQ's or links with help? (Have tried different alternatives for games, such as Q3, HG2, Heretic2, with warious success. Another issue is whether to use Glide option or go for GL and 4.0.x when installing the games!!) [EMAIL PROTECTED]
[expert] HW accel with mdk7.1??
Need advice how to reenable HW acceleration on my Comapq with mdk71 and a Voodo Banshee card. What do you propose to install: mkd71 linux.3dfx.com = == Device3Dfx-2.3.4-5mdk.src.rpm Device3Dfx-2.3-5.src.rpm Glide_V3-2.3.4-5mdk Glide_V3-2.60-16 or should I use tdfx_dri, tdfx_drm, Glide_V3-3.10-6 from 3dfx?? Do I have to manually insmod the module 3dfx.o? Is Mesa in 7.1 (3.1-14) compiled for HW support. If not do I have to recompile it? Before 7.1 XFree86-3.3.x was distributed without problems for my 3dfx cards( banshee and voodoo2). Now with 7.1 XFree86-4.0 is mixed with 3.3.x servers. All this is confusing. Any FAQ's or links with help? (Have tried different alternatives for games, such as Q3, HG2, Heretic2, with warious success. Another issue is whether to use Glide option or go for GL and 4.0.x when installing the games!!) [EMAIL PROTECTED]
[expert] mdk71+ut425+banshee problems
Unreal Tournamentwith Glide option does not start, complaints of not being able to allocate memory for colors. Other games such as q3, heretic2 works OK. Installed: XFree86-glide-module-4.0-6mdk,XFree86-server-4.0-6mdk, XFree86-server-common-3.3.6-14mdk, XFree86-SVGA-3.3.6-14mdk Glide_V3-2.60.15-5mdk, Device3Dfx-2.3-5 Installing with GL option instead of Glide results in a black screen and sound running. Which option to use: GL or Glide? Which modules should be unstalled?? Also, starting up gnome takes forever. It seems that something is wrong with X, taking up all memory and most cpu time. Hoping for help. Svante Signell
Re: [expert] Win TV/radio
Take a look at xawtv: http://www.in-berlin.de/User/kraxel/xawtv.html This one should work both in kde and gnome. Haven't got radio working yet though (Hauppauge WinTV Theater) Svante Traci Collins writes: Klar Brian D Contr MSG/SWS wrote: This may be out of this area's scope. However, I have a WinTV/Radio card (Hauppauge) and the TV works great in Linux MDK 6.1 and KDE. I was wandering if anyone knows how to get the radio to work in linux? Brian D. Klar - CVE OTS WPAFB (937)257-5773 937-973-3125 (Pager) Did you just do a standard install or did you have to install anything special to make this work? I am living in Gnome but I might move over to the KDE side if I could get my WinTV working reliably under Linux. -- Traci Collins, MA Professor of Computer Education Colorado Mountain College http://www.rof.net/wp/tcollins/traci.html
[expert] xawtv
I have recently installed an Hauppauge TV WinTV Theater card on a Compaq 5640/5670 running LM6.1. The card has Dolby surround stereo, radio tuner and remote control. TV seems to work OK, but overlay is not functional, see output below. Seems to be a conflict between DGA and v4l about the memory address to use for my Maxigamer Banshee card. Also, the radio does not work. What about the remote control? Do I have to compile a new hernel, either a patched 2.2.x or 2.3.x? The same situation remains after an upgrade to LM7.0 and xaw-3.0.7. Best regards, [EMAIL PROTECTED] /proc/pci: Bus 1, device 0, function 0: VGA compatible controller: 3Dfx Banshee (rev 3). Fast devsel. Fast back-to-back capable. IRQ 5. Non-prefetchable 32 bit memory at 0x4000 [0x4000]. Prefetchable 32 bit memory at 0x4200 [0x4208]. I/O at 0x1000 [0x1001]. /etc/conf.modules: alias char-major-81 videodev alias char-major-81-0 bttv pre-install bttvmodprobe -k msp3400; modprobe -k tuner options i2c scan=1 options bttvradio=1 vidmem=0x4200 pll=2 options tuner type=5 /sbin/lsmod: Module Size Used by bttv 39896 0 (unused) tuner 2092 1 (autoclean) msp340012512 1 (autoclean) i2c-old 3776 3 [bttv tuner msp3400] videodev2560 3 [bttv] /var/log/messages: Dec 29 17:51:33 compaq kernel: Linux video capture interface: v1.00 Dec 29 17:51:53 compaq kernel: i2c: initialized Dec 29 17:51:53 compaq kernel: bttv0: Brooktree Bt878 (rev 2) bus: 0, devfn: 32, irq: 3, memory: 0x4810. Dec 29 17:51:53 compaq kernel: PCI: Enabling bus mastering for device 00:20 Dec 29 17:51:53 compaq kernel: bttv: 1 Bt8xx card(s) found. Dec 29 17:51:53 compaq kernel: bttv0: Hauppauge eeprom: tuner=Philips FM1216 (5) Dec 29 17:51:53 compaq kernel: bttv0: audio chip: TDA9840 Dec 29 17:51:53 compaq kernel: bttv0: fader chip: TEA6300 Dec 29 17:51:53 compaq kernel: bttv0: model: BT878(Hauppauge new) Dec 29 17:51:53 compaq kernel: msp3400: daemon started Dec 29 17:51:53 compaq kernel: msp3400: init: chip=MSP3410D-B4, has NICAM support xawtv: This is xawtv-3.03, running on Linux/i686 (2.2.13-22mdk) visual: id=0x22 class=4 (TrueColor), depth=16 x11: 1280x1024, 16 bit/pixel, 2560 byte/scanline, DGA, VidMode waitpid: No child processes waitpid: No child processes v4l: 1280x1024, 16 bit/pixel, 2560 byte/scanline WARNING: v4l and dga disagree about the framebuffer base WARNING: Is v4l-conf installed correctly? ov_fbuf.base=(nil), base=0x4200 WARNING: overlay mode disabled wmhooks: kde
[expert] 7.0: How to download crypto SW outside install menus.
One of the menus of the upgrade from 6.1 to 7.0 concerns crypto code. Since for unknown reasons the network card did not work during the install (it works otherwise!!) I could not download crypto software. After the upgrade I tried to start drakxtools but did not find anything useful there. Which tool to use? Urls? Svante S.
[expert] Console mouse not OK in 7.0
After upgrade to 7.0 the console mouse (gpm -t imps2) does not work any longer. What is wrong? Svante S.
[expert] Re: Proposal: Source file package format (summary)
Hi, Here is a summary of the proposal for a common source file format: - Good idea! - Waste of time, Use configure; make; make install, Most packages are for Unix, not only Linux. - Source management problems, no-one is interested in BOTH .rpms and .debs! What about experimental versions? - For experimental packages, use dselect to put a hold on the new version. - For debian, directory structure is important, not .dsc files. - Good suggestion, decision is up to the package author!! - How to ensure the .spec files are valid and functional? - Reducing incompatibility between the variants of the GNU/Linux OS'es is a useful job. - Hard to build good rpm's and deb's. Install to standard directories? What about FHS?? - GNU people participating in LSB work? - Debian is not GNU!? - After rms comments about LSB/GNU/Linux/... this thread turned into a flame war!! No more comments of technical nature any more. Sorry I thought it was an idea worth a better faith!! Binary formats: - Alien can be used, at least from .rpm to .deb - rpm format to be used for binary packages in LSB. People who replied: Iain Wade [EMAIL PROTECTED] George Toft [EMAIL PROTECTED] Oliver Elphick [EMAIL PROTECTED] Adam C Powell IV [EMAIL PROTECTED] cyberclay [EMAIL PROTECTED] Richard Stallman [EMAIL PROTECTED] Chris Siebenmann [EMAIL PROTECTED] Daniel Quinlan [EMAIL PROTECTED] Flaming replies not included!! PS. This mail has been written using GNU Emacs and VM. Wonderful software. What about gemacs, a gnomified version!? DS. Happy New Millennium to everybody! Best regards, [EMAIL PROTECTED] Svante Signell writes: Greetings, What do you think of the following proposal: I order to simplify for package authors/maintainers and to reduce duplication, distribute the source file packages in .tar.gz (or .tar.bz2) format. This avoids the need to provide both .tar.gz, .src.rpm and debian source files. Included in these tarballs add .spec and .dsc files together with the original .tar.gz package and .diff.gz files. Then everybody interested can build source/binary files for their own preferred distribution using the same source files!! Also the GNU packages could contain .deb and .spec files, as is already the case (.spec-files) for gnome packages. Advantages == ++ Enables convergence towards Linux Standard Base (LSB) ++ Simplifies a lot for package maintainers, distribution specific files, .spec, .dsc etc could be supplied by the different vendors. ++ Faster feedback to package authors for patches incorporated into the main distribution. ++ Reduces the risk for patch divergence. (A lot of distribution-specific patches) + Useful for all .rpm-based systems, rpm -t? is already there. + Useful for all .deb-based systems, with minor changes in relevant tools. + Useful for .tar.gz-based systems, no changes necessary. + ... Drawbacks = - No easy way to see if .spec and .dsc files are included in the .tar.gz package (except using tar, but that requires a download) - Distribution profiling more difficult :-( - Package naming has to be agreed upon!! - ... Another issue is to merge the binary file formats .deb and .rpm :-( I'm currently running Rawhide, Redhat 6.1, Debian 2.2, Suse 6.3 and Mandrake 6.1 on different computers and disks, and would really appreciate a common format at least for source packages. Most often I recompile the sources myself. Please feel free to forward this mail to other interested parties not reached by the list here. I would like to start the discussion on this subject. No flame wars please! [EMAIL PROTECTED]
RE: [expert] Q3TEST/Linux/Banshee/GL problem
q3demoTEST-1.10-5 works OK on a rawhide/Matrox Millenuim/Voodoo 2 PC using libMesaVoodoGL.so.3.3, as well as q3test-1.0.8 with libMesaVoodooGL.so.3.1. Rickard Åberg writes: I have problems with q3demo too... I get flickering textures and stuff, tho' the sound works fine. The q3test versions before the q3demo worked just fine. If you use OpenSoundSystem the sound MAY become fucked up, at least from my own experience. I installed the kernel modules for my card instead of using OSS, then q3 ran just fine. But I have (had) a voodoo2... so maybe the problem is global, ie. they really fucked up the release? I can't test anymore since my voodoo2 card seems to have hit the ground for good. :( Don't know why, but one day it suddenly stopped working. Rickard Åberg [EMAIL PROTECTED] "Why do you necessarily have to be wrong just because a few million people think you are?" On 25-Nov-99 Svante Signell wrote: Hi, sending this to the list as well. Did not get any reply from ID. Anybody having the same experience? Svante Signell The combination of Mandrake6.1, Voodoo Banshee and libMesaVoodooGL.so.3.3 does not work with the latest demo: q3demoTEST-1.10-5.i386.rpm. The older demo, q3test-1.08-glibc-2.i386.rpm, worked OK. However that version used the library libMesaVoodooGL.so.3.1. After trying, it was found that the new demo works OK with the old library. Symptoms: The first screen with the flashing Quake III text seems to update the screen every two seconds or so. Also the sound hangs and the mouse is stuck in the upper left corner. Moving to play mode with the arrows results in normal mouse behaviour. However, after a few second of play, the sound hangs again and no movement is possible.
[expert] Suggestion for distribution of source file packages.
Hi all, I have a suggestion: I order to simplify for package authors/maintainers and to reduce duplication, distribute the source files in .tar.gz or .tar.bz2 format. This avoids the need to provide both .tar.gz, .src.rpm and debian source files. Included in these tarballs add .spec and/or .dsc files together with the original .tar.gz package and .diff.gz files. Then everybody interested can build source/binary files for their own preferred distribution using the same source files!! Another issue is to merge the binary file formats .deb and .rpm :( This mail is sent to the lists I'm subscribing to. I'm currently running Rawhide-19991117, Redhat 6.1, Debian 2.1, Suse 6.2 and Mandrake 6.1 on different computers and disks. Comments? Svante Signell
[expert] Q3TEST/Linux/Banshee/GL problem
Hi, sending this to the list as well. Did not get any reply from ID. Anybody having the same experience? Svante Signell The combination of Mandrake6.1, Voodoo Banshee and libMesaVoodooGL.so.3.3 does not work with the latest demo: q3demoTEST-1.10-5.i386.rpm. The older demo, q3test-1.08-glibc-2.i386.rpm, worked OK. However that version used the library libMesaVoodooGL.so.3.1. After trying, it was found that the new demo works OK with the old library. Symptoms: The first screen with the flashing Quake III text seems to update the screen every two seconds or so. Also the sound hangs and the mouse is stuck in the upper left corner. Moving to play mode with the arrows results in normal mouse behaviour. However, after a few second of play, the sound hangs again and no movement is possible.
[expert] Configuring a small local network and internet connection
Hello, Probably this is an FAQ. How to set up a local network with one machine connected to the internet via ppp? What about proxies/routers? The local network is made up from ethernet 100 Mbps cards and a 5-port 100 Mbps HUB. The internet connection is an ISDN modem with 64 kbps access.