From malekith at pld.org.pl Fri Mar 8 12:15:53 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 12:15:53 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) Message-ID: <20020308111551.GA4599@ep09.kernel.pl> Hi, Yesterday qboosh suggested me some additions to bsp to allow pivot_root style boot using it. I thought it would be funny to have system booted up this way, so I did it :) There were some more or less severe problems during it. 1) old root (initrd) has to be placed somewhere on real root, I suggest /var/lib/bsp/oldroot [RFC] 2) we cannot unmount old root while bsp is running, I first thought we can unmount and freeramdisk old root in rc-scripts, however there is no standalone freeramdisk utility and even beside it, unmounting in rc-scripts seems messy, so I decided to use two stage bsp bootup. Bsp from initrd, at end of its life does exec /sbin/bsp (from real root filesystem, in the meantime we have to close /dev/console but this ain't problem), that unmounts oldroot, frees ramdisk and does exec /sbin/init. This works :) anyway RFC. 3) there is problem with devfs. We can mount it in stage 1, unmount before trying to unmount oldroot, and mount it again in real root, so we don't have to have /dev/ram0 and /dev/console on root filesystem, but what then? leave it mounted? what about people not using it? and what if there is not devfs in kernel? RFC 4) after boot: [malekith at roke ~]$ ls -l /dev/root lr-xr-xr-x 1 root root 4 Jan 1 1970 /dev/root -> rd/0 [malekith at roke ~]$ I guess this is problem with devfsd. 5) we have to use root=/dev/ram0 in lilo, and put bsp as /sbin/init on initrd. (otherwise kernel would never free up memory used for bootup). 6) directory to put old root in, could be created automagicly by bsp, we would have to mount root read-write, create the directory, and in stage 2 remove it and mount root read-only. But I guess this isn't very bright idea (to mess with writing root before fsck has a chance to run). Tests were done on 2.4.17-2, almost stock (rpm -bb kernel.spec), but compiled with gcc-3.0.4. Any comments are welcome. More detailed information can be found in cvs://bsp/README -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From qboosh at pld.org.pl Fri Mar 8 12:33:48 2002 From: qboosh at pld.org.pl (Jakub Bogusz) Date: Fri, 8 Mar 2002 12:33:48 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308111551.GA4599@ep09.kernel.pl> References: <20020308111551.GA4599@ep09.kernel.pl> Message-ID: <20020308113348.GB31524@gruby.cs.net.pl> On Fri, Mar 08, 2002 at 12:15:53PM +0100, Michal Moskal wrote: > 6) directory to put old root in, could be created automagicly by bsp, > we would have to mount root read-write, create the directory, > and in stage 2 remove it and mount root read-only. But I guess this > isn't very bright idea (to mess with writing root before fsck > has a chance to run). Somebody wrote at news:pl.comp.os.linux that in some distribution (RedHat or Mandrake) there is persistent empty /initrd directory, and when he removed it, system boot failed. Guess why :) This may be an answer - but I think it's at least a little nasty... -- Jakub Bogusz From malekith at pld.org.pl Fri Mar 8 12:43:37 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 12:43:37 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308113348.GB31524@gruby.cs.net.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> Message-ID: <20020308114337.GB5098@ep09.kernel.pl> On Fri, Mar 08, 2002 at 12:33:48PM +0100, Jakub Bogusz wrote: > On Fri, Mar 08, 2002 at 12:15:53PM +0100, Michal Moskal wrote: > > 6) directory to put old root in, could be created automagicly by bsp, > > we would have to mount root read-write, create the directory, > > and in stage 2 remove it and mount root read-only. But I guess this > > isn't very bright idea (to mess with writing root before fsck > > has a chance to run). > > Somebody wrote at news:pl.comp.os.linux that in some distribution > (RedHat or Mandrake) there is persistent empty /initrd directory, > and when he removed it, system boot failed. Guess why :) > > This may be an answer - but I think it's at least a little nasty... Having "unused" directory in / is like asking for truble. But if it's in /var/lib/bsp/initrd, then nobody will remove it. And if he will... well he's out of luck :) Removing /sbin/bsp will also prevent system from booting in this setting, by the way. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From qboosh at pld.org.pl Fri Mar 8 12:52:56 2002 From: qboosh at pld.org.pl (Jakub Bogusz) Date: Fri, 8 Mar 2002 12:52:56 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308114337.GB5098@ep09.kernel.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> Message-ID: <20020308115256.GC31524@gruby.cs.net.pl> On Fri, Mar 08, 2002 at 12:43:37PM +0100, Michal Moskal wrote: > On Fri, Mar 08, 2002 at 12:33:48PM +0100, Jakub Bogusz wrote: > > On Fri, Mar 08, 2002 at 12:15:53PM +0100, Michal Moskal wrote: > > > 6) directory to put old root in, could be created automagicly by bsp, > > > we would have to mount root read-write, create the directory, > > > and in stage 2 remove it and mount root read-only. But I guess this > > > isn't very bright idea (to mess with writing root before fsck > > > has a chance to run). > > > > Somebody wrote at news:pl.comp.os.linux that in some distribution > > (RedHat or Mandrake) there is persistent empty /initrd directory, > > and when he removed it, system boot failed. Guess why :) > > > > This may be an answer - but I think it's at least a little nasty... > > Having "unused" directory in / is like asking for truble. > But if it's in /var/lib/bsp/initrd, then nobody will remove it. But it will be inaccessible if /var isn't on root filesystem (and it isn't rare case) so... it must be something on root... in /etc? -- Jakub Bogusz From malekith at pld.org.pl Fri Mar 8 13:10:55 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 13:10:55 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308115256.GC31524@gruby.cs.net.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> <20020308115256.GC31524@gruby.cs.net.pl> Message-ID: <20020308121055.GA2381@ep09.kernel.pl> On Fri, Mar 08, 2002 at 12:52:56PM +0100, Jakub Bogusz wrote: > On Fri, Mar 08, 2002 at 12:43:37PM +0100, Michal Moskal wrote: > > On Fri, Mar 08, 2002 at 12:33:48PM +0100, Jakub Bogusz wrote: > > > On Fri, Mar 08, 2002 at 12:15:53PM +0100, Michal Moskal wrote: > > > > 6) directory to put old root in, could be created automagicly by bsp, > > > > we would have to mount root read-write, create the directory, > > > > and in stage 2 remove it and mount root read-only. But I guess this > > > > isn't very bright idea (to mess with writing root before fsck > > > > has a chance to run). > > > > > > Somebody wrote at news:pl.comp.os.linux that in some distribution > > > (RedHat or Mandrake) there is persistent empty /initrd directory, > > > and when he removed it, system boot failed. Guess why :) > > > > > > This may be an answer - but I think it's at least a little nasty... > > > > Having "unused" directory in / is like asking for truble. > > But if it's in /var/lib/bsp/initrd, then nobody will remove it. > > But it will be inaccessible if /var isn't on root filesystem (and it > isn't rare case) so... it must be something on root... in /etc? Right. Maybe /lib or /boot ? -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From qboosh at pld.org.pl Fri Mar 8 13:15:44 2002 From: qboosh at pld.org.pl (Jakub Bogusz) Date: Fri, 8 Mar 2002 13:15:44 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308121055.GA2381@ep09.kernel.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> <20020308115256.GC31524@gruby.cs.net.pl> <20020308121055.GA2381@ep09.kernel.pl> Message-ID: <20020308121544.GD31524@gruby.cs.net.pl> On Fri, Mar 08, 2002 at 01:10:55PM +0100, Michal Moskal wrote: > On Fri, Mar 08, 2002 at 12:52:56PM +0100, Jakub Bogusz wrote: > > On Fri, Mar 08, 2002 at 12:43:37PM +0100, Michal Moskal wrote: > > > On Fri, Mar 08, 2002 at 12:33:48PM +0100, Jakub Bogusz wrote: > > > > On Fri, Mar 08, 2002 at 12:15:53PM +0100, Michal Moskal wrote: > > > > > 6) directory to put old root in, could be created automagicly by bsp, > > > > > we would have to mount root read-write, create the directory, > > > > > and in stage 2 remove it and mount root read-only. But I guess this > > > > > isn't very bright idea (to mess with writing root before fsck > > > > > has a chance to run). > > > > > > > > Somebody wrote at news:pl.comp.os.linux that in some distribution > > > > (RedHat or Mandrake) there is persistent empty /initrd directory, > > > > and when he removed it, system boot failed. Guess why :) > > > > > > > > This may be an answer - but I think it's at least a little nasty... > > > > > > Having "unused" directory in / is like asking for truble. > > > But if it's in /var/lib/bsp/initrd, then nobody will remove it. > > > > But it will be inaccessible if /var isn't on root filesystem (and it > > isn't rare case) so... it must be something on root... in /etc? > > Right. > > Maybe /lib or /boot ? Not /boot - I thought about it before I wrote "/etc", but /boot can be on separate fs too (and during boot only bootloader can access it, using BIOS to disk read operations). Maybe if it would be mounted in initrd - but it would require more devices, possibly more drivers and... more problems :) -- Jakub Bogusz From orzech at pld.org.pl Fri Mar 8 13:19:04 2002 From: orzech at pld.org.pl (Tomek Orzechowski) Date: Fri, 8 Mar 2002 13:19:04 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308121055.GA2381@ep09.kernel.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> <20020308115256.GC31524@gruby.cs.net.pl> <20020308121055.GA2381@ep09.kernel.pl> Message-ID: <20020308121904.GA27951@alpha.hq.cs.net.pl> On 03/08/02 Michal Moskal wrote: >> > > > 6) directory to put old root in, could be created automagicly by bsp, >> > > > we would have to mount root read-write, create the directory, >> > > > and in stage 2 remove it and mount root read-only. But I guess this >> > > > isn't very bright idea (to mess with writing root before fsck >> > > > has a chance to run). >> > > >> > > Somebody wrote at news:pl.comp.os.linux that in some distribution >> > > (RedHat or Mandrake) there is persistent empty /initrd directory, >> > > and when he removed it, system boot failed. Guess why :) >> > > >> > > This may be an answer - but I think it's at least a little nasty... >> > >> > Having "unused" directory in / is like asking for truble. >> > But if it's in /var/lib/bsp/initrd, then nobody will remove it. >> >> But it will be inaccessible if /var isn't on root filesystem (and it >> isn't rare case) so... it must be something on root... in /etc? > >Right. > >Maybe /lib or /boot ? /boot - no; some old machines requires small /boot partition on big disks. /lib may be good. I have never thought about moving it form rootfs to another partition. -- Tomek Orzechowski From wiget at pld.org.pl Fri Mar 8 13:25:50 2002 From: wiget at pld.org.pl (Artur Frysiak) Date: Fri, 8 Mar 2002 13:25:50 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308114337.GB5098@ep09.kernel.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> Message-ID: <20020308122550.GA12748@dziadek.home> On Fri, Mar 08, 2002 at 12:43:37PM +0100, Michal Moskal wrote: > On Fri, Mar 08, 2002 at 12:33:48PM +0100, Jakub Bogusz wrote: > > On Fri, Mar 08, 2002 at 12:15:53PM +0100, Michal Moskal wrote: > > > 6) directory to put old root in, could be created automagicly by bsp, > > > we would have to mount root read-write, create the directory, > > > and in stage 2 remove it and mount root read-only. But I guess this > > > isn't very bright idea (to mess with writing root before fsck > > > has a chance to run). > > > > Somebody wrote at news:pl.comp.os.linux that in some distribution > > (RedHat or Mandrake) there is persistent empty /initrd directory, > > and when he removed it, system boot failed. Guess why :) > > > > This may be an answer - but I think it's at least a little nasty... > > Having "unused" directory in / is like asking for truble. > But if it's in /var/lib/bsp/initrd, then nobody will remove it. > And if he will... well he's out of luck :) echo "Don't remove this dir. Its used by bootup proces" >/initrd/README /initrd is sugested by Werner Almesberger and Hans Lermen in linux/Documentation/initrd.txt -- Artur Frysiak http://www.pld.org.pl/ From alchemyx at uznam.net.pl Fri Mar 8 13:33:30 2002 From: alchemyx at uznam.net.pl (Michal Margula) Date: Fri, 8 Mar 2002 13:33:30 +0100 (CET) Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308122550.GA12748@dziadek.home> Message-ID: On Fri, 8 Mar 2002, Artur Frysiak wrote: > > echo "Don't remove this dir. Its used by bootup proces" >/initrd/README > /initrd is sugested by Werner Almesberger > and Hans Lermen in linux/Documentation/initrd.txt And to be sure chattr +i /initrd ;-) -- Michal Margula, alchemyx at uznam.net.pl, ICQ UIN 12267440, +) http://alchemyx.uznam.net.pl/, Polish section of Linux Counter maintainer From blues at ds6.pg.gda.pl Fri Mar 8 13:36:11 2002 From: blues at ds6.pg.gda.pl (Blues) Date: Fri, 8 Mar 2002 13:36:11 +0100 (CET) Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: Message-ID: On Fri, 8 Mar 2002, Michal Margula wrote: > > echo "Don't remove this dir. Its used by bootup proces" >/initrd/README > > /initrd is sugested by Werner Almesberger > > and Hans Lermen in linux/Documentation/initrd.txt > And to be sure chattr +i /initrd ;-) BTW... maybe we should to do this on some system files/directories...? -- --------------------------------- pozdr. Pawe? Go?aszewski --------------------------------- CPU not found - software emulation... From baggins at sith.mimuw.edu.pl Fri Mar 8 13:40:16 2002 From: baggins at sith.mimuw.edu.pl (Jan Rekorajski) Date: Fri, 8 Mar 2002 13:40:16 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308111551.GA4599@ep09.kernel.pl> Message-ID: <20020308134016.B16565@sith.mimuw.edu.pl> [pi?tek, 08 marzec 2002], Michal Moskal napisa?(a): > Hi, > > Yesterday qboosh suggested me some additions to bsp to allow pivot_root > style boot using it. I thought it would be funny to have system booted up > this way, so I did it :) > > There were some more or less severe problems during it. > > 1) old root (initrd) has to be placed somewhere on real root, I suggest > /var/lib/bsp/oldroot [RFC] Why not /boot/root ? > 2) we cannot unmount old root while bsp is running, I first thought we > can unmount and freeramdisk old root in rc-scripts, however there is > no standalone freeramdisk utility and even beside it, unmounting in > rc-scripts seems messy, so I decided to use two stage bsp bootup. Works for me and my Rescue CD. freeramdisk = /sbin/blockdev --flushbufs /dev/rd/* > Bsp from initrd, at end of its life does exec /sbin/bsp (from real > root filesystem, in the meantime we have to close /dev/console but > this ain't problem), that unmounts oldroot, frees ramdisk and does > exec /sbin/init. This works :) anyway RFC. I can send you linuxrc.c from Rescue CD. > 3) there is problem with devfs. We can mount it in stage 1, unmount > before trying to unmount oldroot, and mount it again in real root, > so we don't have to have /dev/ram0 and /dev/console on root filesystem, > but what then? leave it mounted? what about people not using it? > and what if there is not devfs in kernel? RFC Just use normal /dev at first stage. Janek -- Jan R?korajski | ALL SUSPECTS ARE GUILTY. PERIOD! bagginsmimuw.edu.pl | OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY? BOFH, MANIAC | -- TROOPS by Kevin Rubio From malekith at pld.org.pl Fri Mar 8 13:46:54 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 13:46:54 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308134016.B16565@sith.mimuw.edu.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308134016.B16565@sith.mimuw.edu.pl> Message-ID: <20020308124654.GA25888@ep09.kernel.pl> On Fri, Mar 08, 2002 at 01:40:16PM +0100, Jan Rekorajski wrote: > [pi?tek, 08 marzec 2002], Michal Moskal napisa?(a): > > > Hi, > > > > Yesterday qboosh suggested me some additions to bsp to allow pivot_root > > style boot using it. I thought it would be funny to have system booted up > > this way, so I did it :) > > > > There were some more or less severe problems during it. > > > > 1) old root (initrd) has to be placed somewhere on real root, I suggest > > /var/lib/bsp/oldroot [RFC] > > Why not /boot/root ? Bacause it can be on separate partition. /lib/oldroot seems more reasonable. > > 2) we cannot unmount old root while bsp is running, I first thought we > > can unmount and freeramdisk old root in rc-scripts, however there is > > no standalone freeramdisk utility and even beside it, unmounting in > > rc-scripts seems messy, so I decided to use two stage bsp bootup. > > Works for me and my Rescue CD. > freeramdisk = /sbin/blockdev --flushbufs /dev/rd/* Oh, I didn't know that. > > Bsp from initrd, at end of its life does exec /sbin/bsp (from real > > root filesystem, in the meantime we have to close /dev/console but > > this ain't problem), that unmounts oldroot, frees ramdisk and does > > exec /sbin/init. This works :) anyway RFC. > > I can send you linuxrc.c from Rescue CD. I would appriciate that. Anyway, without any files open I couldn't unmount /oldroot while 1st stage bsp was running. > > 3) there is problem with devfs. We can mount it in stage 1, unmount > > before trying to unmount oldroot, and mount it again in real root, > > so we don't have to have /dev/ram0 and /dev/console on root filesystem, > > but what then? leave it mounted? what about people not using it? > > and what if there is not devfs in kernel? RFC > > Just use normal /dev at first stage. Yes, it can be done. But someday normal /dev will be removed from kernel. And in fact we don't need any normal device right now, bsp can start even without /dev/console, and mount devfs. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From malekith at pld.org.pl Fri Mar 8 13:48:11 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 13:48:11 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308122550.GA12748@dziadek.home> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> <20020308122550.GA12748@dziadek.home> Message-ID: <20020308124811.GB25888@ep09.kernel.pl> On Fri, Mar 08, 2002 at 01:25:50PM +0100, Artur Frysiak wrote: > echo "Don't remove this dir. Its used by bootup proces" >/initrd/README > /initrd is sugested by Werner Almesberger and > Hans Lermen in linux/Documentation/initrd.txt But is is so ugly... :) -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From baggins at sith.mimuw.edu.pl Fri Mar 8 13:54:50 2002 From: baggins at sith.mimuw.edu.pl (Jan Rekorajski) Date: Fri, 8 Mar 2002 13:54:50 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308124654.GA25888@ep09.kernel.pl> Message-ID: <20020308135450.A16772@sith.mimuw.edu.pl> [pi?tek, 08 marzec 2002], Michal Moskal napisa?(a): > On Fri, Mar 08, 2002 at 01:40:16PM +0100, Jan Rekorajski wrote: > > [pi?tek, 08 marzec 2002], Michal Moskal napisa?(a): > > > > > Hi, > > > > > > Yesterday qboosh suggested me some additions to bsp to allow pivot_root > > > style boot using it. I thought it would be funny to have system booted up > > > this way, so I did it :) > > > > > > There were some more or less severe problems during it. > > > > > > 1) old root (initrd) has to be placed somewhere on real root, I suggest > > > /var/lib/bsp/oldroot [RFC] > > > > Why not /boot/root ? > > Bacause it can be on separate partition. /lib/oldroot seems more > reasonable. There is /mnt directory, maybe /mnt/oldroot? > > > Bsp from initrd, at end of its life does exec /sbin/bsp (from real > > > root filesystem, in the meantime we have to close /dev/console but > > > this ain't problem), that unmounts oldroot, frees ramdisk and does > > > exec /sbin/init. This works :) anyway RFC. > > > > I can send you linuxrc.c from Rescue CD. > > I would appriciate that. Anyway, without any files open I couldn't > unmount /oldroot while 1st stage bsp was running. I did it from modified rc-scripts, at the start of rc umount /initrd ; flushbufs I send linuxrc separately. > > > 3) there is problem with devfs. We can mount it in stage 1, unmount > > > before trying to unmount oldroot, and mount it again in real root, > > > so we don't have to have /dev/ram0 and /dev/console on root filesystem, > > > but what then? leave it mounted? what about people not using it? > > > and what if there is not devfs in kernel? RFC > > > > Just use normal /dev at first stage. > > Yes, it can be done. But someday normal /dev will be removed from > kernel. And in fact we don't need any normal device right now, bsp > can start even without /dev/console, and mount devfs. Have you tried with /dev mounted automagically by kernel? Just curious. Janek -- Jan R?korajski | ALL SUSPECTS ARE GUILTY. PERIOD! bagginsmimuw.edu.pl | OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY? BOFH, MANIAC | -- TROOPS by Kevin Rubio From malekith at pld.org.pl Fri Mar 8 14:20:34 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 14:20:34 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308135450.A16772@sith.mimuw.edu.pl> References: <20020308124654.GA25888@ep09.kernel.pl> <20020308135450.A16772@sith.mimuw.edu.pl> Message-ID: <20020308132034.GA18070@ep09.kernel.pl> On Fri, Mar 08, 2002 at 01:54:50PM +0100, Jan Rekorajski wrote: > > Bacause it can be on separate partition. /lib/oldroot seems more > > reasonable. > > There is /mnt directory, maybe /mnt/oldroot? It is probably better suited, but more prone to removal by stupid user. But it could be solution (with echo "don't delete tralalal..." > /mnt/oldroot/README). > > > > Bsp from initrd, at end of its life does exec /sbin/bsp (from real > > > > root filesystem, in the meantime we have to close /dev/console but > > > > this ain't problem), that unmounts oldroot, frees ramdisk and does > > > > exec /sbin/init. This works :) anyway RFC. > > > > > > I can send you linuxrc.c from Rescue CD. > > > > I would appriciate that. Anyway, without any files open I couldn't > > unmount /oldroot while 1st stage bsp was running. > > I did it from modified rc-scripts, at the start of rc > umount /initrd ; flushbufs > I send linuxrc separately. This can be also done at the beginning of PLD rc-scripts (second stage bsp wouldn't be needed then), but I guess doing everything in bsp is less error prone (doesn't require recent rc-scripts and so on). > Have you tried with /dev mounted automagically by kernel? Just curious. No, but this should be ok. We would unmount it anyway (and mount new one) after changing root, but if kernel mounted it, we wouldn't have to have any devices on initrd. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From aredridel at nbtsc.org Fri Mar 8 17:16:58 2002 From: aredridel at nbtsc.org (Aredridel) Date: Fri, 8 Mar 2002 09:16:58 -0700 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308132034.GA18070@ep09.kernel.pl> References: <20020308124654.GA25888@ep09.kernel.pl> <20020308135450.A16772@sith.mimuw.edu.pl> <20020308132034.GA18070@ep09.kernel.pl> Message-ID: <20020308161658.GA5102@nbtsc.org> On Fri, Mar 08, 2002 at 02:20:34PM +0100, Michal Moskal wrote: > On Fri, Mar 08, 2002 at 01:54:50PM +0100, Jan Rekorajski wrote: > > > Bacause it can be on separate partition. /lib/oldroot seems more > > > reasonable. > > > > There is /mnt directory, maybe /mnt/oldroot? > > It is probably better suited, but more prone to removal by stupid > user. But it could be solution (with echo "don't delete tralalal..." > > /mnt/oldroot/README). What about using /mnt itself, since it won't stay mounted after system boot anyway? -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 232 bytes Desc: not available URL: From malekith at pld.org.pl Fri Mar 8 17:39:09 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 8 Mar 2002 17:39:09 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308161658.GA5102@nbtsc.org> References: <20020308124654.GA25888@ep09.kernel.pl> <20020308135450.A16772@sith.mimuw.edu.pl> <20020308132034.GA18070@ep09.kernel.pl> <20020308161658.GA5102@nbtsc.org> Message-ID: <20020308163909.GA24369@ep09.kernel.pl> On Fri, Mar 08, 2002 at 09:16:58AM -0700, Aredridel wrote: > On Fri, Mar 08, 2002 at 02:20:34PM +0100, Michal Moskal wrote: > > On Fri, Mar 08, 2002 at 01:54:50PM +0100, Jan Rekorajski wrote: > > > > Bacause it can be on separate partition. /lib/oldroot seems more > > > > reasonable. > > > > > > There is /mnt directory, maybe /mnt/oldroot? > > > > It is probably better suited, but more prone to removal by stupid > > user. But it could be solution (with echo "don't delete tralalal..." > > > /mnt/oldroot/README). > > What about using /mnt itself, since it won't stay mounted after system > boot anyway? Looks fine (iff it is always properly unmounted, very early). As unmounting won't be done by bsp (after some discussion with wiget, I guess rc-scripts is proper place to do that), this has to be doubly ensured, that rc-scripts won't mount -a before unmounting initrd. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From michal at michal.waw.pl Fri Mar 8 18:15:19 2002 From: michal at michal.waw.pl (Michal Kochanowicz) Date: Fri, 8 Mar 2002 18:15:19 +0100 Subject: 2.4 kernel style boot using pivot_root (RFC) In-Reply-To: <20020308124811.GB25888@ep09.kernel.pl> References: <20020308111551.GA4599@ep09.kernel.pl> <20020308113348.GB31524@gruby.cs.net.pl> <20020308114337.GB5098@ep09.kernel.pl> <20020308122550.GA12748@dziadek.home> <20020308124811.GB25888@ep09.kernel.pl> Message-ID: <20020308171519.GB2910@wieszak.mmm.ozarow-12.waw.pl> On Fri, Mar 08, 2002 at 01:48:11PM +0100, Michal Moskal wrote: > On Fri, Mar 08, 2002 at 01:25:50PM +0100, Artur Frysiak wrote: > > echo "Don't remove this dir. Its used by bootup proces" >/initrd/README > > /initrd is sugested by Werner Almesberger and > > Hans Lermen in linux/Documentation/initrd.txt > But is is so ugly... :) Yeah, I don't like it too. /lib is IMO better place. -- --= Michal Kochanowicz==--==--==BOFH==--==--==michal at michal.waw.pl =-- --= finger me for PGP public key or visit http://michal.waw.pl/PGP =-- --==--==--==--==--==-- Vodka. Connecting people.--==--==--==--==--==-- A chodzenie po g?rach SSIE!!! From malekith at pld.org.pl Fri Mar 15 18:38:37 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Fri, 15 Mar 2002 18:38:37 +0100 Subject: mdctl-initrd Message-ID: <20020315173837.GA4277@ep09.kernel.pl> I made initrd subpackge in mdctl, linked against dietlibc. It is meant to be build by default and put in distribution resources. I'm going to hack lvm and raidtools, and commit shortly. mdctl-init package contains just /sbin/initrd-mdctl. I'm talking about it, since other packages linked with diet/uc are simply done using changing %{__cc}. However, as these packages are going to be of big usage (they will be required in near future by geninitrd to generate initrd for raid/lvm systems properly), they cannot be built using home-grown methods. Therefore we need dietlibc{-devel,-static,} on builders. A note about raid support in bsp -- it will be removed shortly (it doesn't work anyway). BTW: when hacking packages to built with uClibc/dietlibc, please remove glibc-devel from chroot you're doing it in. It will ensure glibc is not used (this was the case with slang). -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From wrobell at ite.pl Fri Mar 15 19:25:14 2002 From: wrobell at ite.pl (wrobell) Date: Fri, 15 Mar 2002 19:25:14 +0100 Subject: new rpm groups; wrobell: SPECS python-bonobo.spec,NONE,1.1 In-Reply-To: <20020315181127.80964E1048@wun.zie.pg.gda.pl> References: <20020315181127.80964E1048@wun.zie.pg.gda.pl> Message-ID: <20020315182514.GF2327@pred.posexperts.com.pl> On Fri, Mar 15, 2002 at 06:11:25PM +0000, Artur Wr?blewski wrote: [...] > Group: Development/Languages/Python What about new group Libraries/Python or Python/Modules? Many of modules are just for applications. By analogy... there is group Libraries (Libraries/Modules) and Development/Libraries (Development/Languages/Python). The same idea applies to Perl and PHP in my opinion. Any objections? If none, then current state is a bug and we should fix it before 1.0 release. [...] wrobell From kloczek at rudy.mif.pg.gda.pl Sat Mar 16 19:33:34 2002 From: kloczek at rudy.mif.pg.gda.pl (=?ISO-8859-2?Q?Tomasz_K=B3oczko?=) Date: Sat, 16 Mar 2002 19:33:34 +0100 (CET) Subject: mdctl-initrd In-Reply-To: <20020315173837.GA4277@ep09.kernel.pl> Message-ID: On Fri, 15 Mar 2002, Michal Moskal wrote: > I made initrd subpackge in mdctl, linked against dietlibc. It is meant > to be build by default and put in distribution resources. I'm going to > hack lvm and raidtools, and commit shortly. > > mdctl-init package contains just /sbin/initrd-mdctl. > > I'm talking about it, since other packages linked with diet/uc are > simply done using changing %{__cc}. However, as these packages are going > to be of big usage (they will be required in near future by geninitrd > to generate initrd for raid/lvm systems properly), they cannot be built > using home-grown methods. Therefore we need dietlibc{-devel,-static,} on > builders. > > A note about raid support in bsp -- it will be removed shortly (it > doesn't work anyway). > > BTW: when hacking packages to built with uClibc/dietlibc, please remove > glibc-devel from chroot you're doing it in. It will ensure glibc is not > used (this was the case with slang). bsp is very specialized tool. The same initrd-mdctl. Both are used in the same time .. on system startup. IMHO logical is prepare one tool with additional initrd-mdctl functionality. When ths functionality will be integrated in one tol it will be much easier prepare them without hardcoding uClibc/dietlibc build dependencies (some neccessatry libc funcionality can be even coded inside bsp source code). kloczek -- ----------------------------------------------------------- *Ludzie nie maj? problem?w, tylko sobie sami je stwarzaj?* ----------------------------------------------------------- Tomasz K?oczko, sys adm @zie.pg.gda.pl|*e-mail: kloczek at rudy.mif.pg.gda.pl* From malekith at pld.org.pl Mon Mar 18 10:20:13 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Mon, 18 Mar 2002 10:20:13 +0100 Subject: mdctl-initrd In-Reply-To: References: <20020315173837.GA4277@ep09.kernel.pl> Message-ID: <20020318092013.GA15168@ep09.kernel.pl> On Sat, Mar 16, 2002 at 07:33:34PM +0100, Tomasz K?oczko wrote: > On Fri, 15 Mar 2002, Michal Moskal wrote: > > > I made initrd subpackge in mdctl, linked against dietlibc. It is meant > > to be build by default and put in distribution resources. I'm going to > > hack lvm and raidtools, and commit shortly. > > > > mdctl-init package contains just /sbin/initrd-mdctl. > > > > I'm talking about it, since other packages linked with diet/uc are > > simply done using changing %{__cc}. However, as these packages are going > > to be of big usage (they will be required in near future by geninitrd > > to generate initrd for raid/lvm systems properly), they cannot be built > > using home-grown methods. Therefore we need dietlibc{-devel,-static,} on > > builders. > > > > A note about raid support in bsp -- it will be removed shortly (it > > doesn't work anyway). > > > > BTW: when hacking packages to built with uClibc/dietlibc, please remove > > glibc-devel from chroot you're doing it in. It will ensure glibc is not > > used (this was the case with slang). > > bsp is very specialized tool. The same initrd-mdctl. Both are used in the > same time .. on system startup. IMHO logical is prepare one tool with > additional initrd-mdctl functionality. When ths functionality will be > integrated in one tol it will be much easier prepare them without > hardcoding uClibc/dietlibc build dependencies (some neccessatry libc > funcionality can be even coded inside bsp source code). I don't want to reinvent the wheel. However patches are welcome. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From malekith at pld.org.pl Mon Mar 18 13:11:56 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Mon, 18 Mar 2002 13:11:56 +0100 Subject: bootdisk issues Message-ID: <20020318121156.GA21729@ep09.kernel.pl> I've changes build system of bootdisk. The reason was, that the old one was based on BOOT packages, that are absent from CVS right now, and unmaitainable. Therefore the change had to take place before 1.0 (we're doing rc1 now, not final?). All tools are rebuilt with current CVS version. In addition parted and fdisk is now linked against uClibc. All this is likely to cause problems. So: PLEASE TEST IT!!! BTW: where are pdksh aliases? And where is kernel from ra/i386? -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From dobrek at itp.uni-hannover.de Tue Mar 19 13:48:13 2002 From: dobrek at itp.uni-hannover.de (Lukas Dobrek) Date: Tue, 19 Mar 2002 13:48:13 +0100 Subject: [kreutzm@itp.uni-hannover.de: printing, euro ...] Message-ID: <20020319134813.A13520@zibal.itp.uni-hannover.de> Some people don't understand history. But anyway it should be done. Lukasz ----- Forwarded message from Helge Kreutzmann ----- > Date: Mon, 18 Mar 2002 10:02:56 +0100 > From: Helge Kreutzmann > Subject: printing, euro ... > To: Pukasz Dobrek > > Hello Lukasz ! [cut] > Secondly I think it would be a good idea if PLD includes the euro sign > for latex, e.g. > ftp://ftp.dante.de/tex-archive/help/Catalogue/entries/eurosym.html > > So that the polish get used to the Euro as well even though it'll > probably at least four years until the official currency in Poland > switches ... > > Best greetings > > Helge > > P.S. No interest any more in CeBIT ? > > > -- > Helge Kreutzmann, Dipl.-Phys. Helge.Kreutzmann at itp.uni-hannover.de > gpg signed mail preferred gpg-key: finger kreutzm at rigel.itp.uni-hannover.de > 64bit GNU powered http://www.itp.uni-hannover.de/~kreutzm > Help keep free software "libre": http://www.freepatents.org/ ----- End forwarded message ----- -- ?ukasz Dobrek An optimist believes that we live in the best of all possible worlds. A pessimist is sure that this must be so. From wiget at pld.org.pl Tue Mar 19 14:42:56 2002 From: wiget at pld.org.pl (Artur Frysiak) Date: Tue, 19 Mar 2002 14:42:56 +0100 Subject: [kreutzm@itp.uni-hannover.de: printing, euro ...] In-Reply-To: <20020319134813.A13520@zibal.itp.uni-hannover.de> References: <20020319134813.A13520@zibal.itp.uni-hannover.de> Message-ID: <20020319134256.GA12324@dziadek.home> On Tue, Mar 19, 2002 at 01:48:13PM +0100, Lukas Dobrek wrote: > ----- Forwarded message from Helge Kreutzmann > ----- > > > Date: Mon, 18 Mar 2002 10:02:56 +0100 > > From: Helge Kreutzmann > > Subject: printing, euro ... > > To: Pukasz Dobrek > > > > Hello Lukasz ! > [cut] > > Secondly I think it would be a good idea if PLD includes the euro sign > > for latex, e.g. > > ftp://ftp.dante.de/tex-archive/help/Catalogue/entries/eurosym.html I plan adding eurosym package to LaTeX in near future. But not to upcomming PLD 1.0. Best regards -- Artur Frysiak http://www.pld.org.pl/ From malekith at pld.org.pl Wed Mar 20 18:10:08 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Wed, 20 Mar 2002 18:10:08 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> Message-ID: <20020320171008.GA13222@ep09.kernel.pl> On Wed, Mar 20, 2002 at 04:34:10PM +0000, Filip Kalinski wrote: > Module name: SPECS > Changes by: filon > > Changed RCS file: /cvsroot/SPECS/SDL.spec,v > ---------------------------- > revision 1.77 > date: 2002/03/20 16:34:07; author: filon; state: Exp; lines: +20 -12 > - removed not needed requirement for gtk+-devel > - enable nasm oprimizations on x86 ^^^^ This caused problems on k5 (segv on cpuid). I don't know if it's still true though. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From malekith at pld.org.pl Wed Mar 20 18:11:33 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Wed, 20 Mar 2002 18:11:33 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320171008.GA13222@ep09.kernel.pl> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> Message-ID: <20020320171133.GB13222@ep09.kernel.pl> On Wed, Mar 20, 2002 at 06:10:08PM +0100, Michal Moskal wrote: > On Wed, Mar 20, 2002 at 04:34:10PM +0000, Filip Kalinski wrote: > > Module name: SPECS > > Changes by: filon > > > > Changed RCS file: /cvsroot/SPECS/SDL.spec,v > > ---------------------------- > > revision 1.77 > > date: 2002/03/20 16:34:07; author: filon; state: Exp; lines: +20 -12 > > - removed not needed requirement for gtk+-devel > > - enable nasm oprimizations on x86 > ^^^^ > > This caused problems on k5 (segv on cpuid). I don't know if it's still > true though. Forgot: Test case: try running mpeg_play. -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From fk181140 at zodiac.mimuw.edu.pl Wed Mar 20 18:19:11 2002 From: fk181140 at zodiac.mimuw.edu.pl (Filip Kalinski) Date: Wed, 20 Mar 2002 18:19:11 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320171133.GB13222@ep09.kernel.pl> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> <20020320171133.GB13222@ep09.kernel.pl> Message-ID: <20020320171911.GA738@fk.localdomain> On day Wed, Mar 20, 2002 at 06:11:33PM +0100, Michal Moskal wrote what follows: > On Wed, Mar 20, 2002 at 06:10:08PM +0100, Michal Moskal wrote: > > On Wed, Mar 20, 2002 at 04:34:10PM +0000, Filip Kalinski wrote: > > > Module name: SPECS > > > Changes by: filon > > > > > > Changed RCS file: /cvsroot/SPECS/SDL.spec,v > > > ---------------------------- > > > revision 1.77 > > > date: 2002/03/20 16:34:07; author: filon; state: Exp; lines: +20 -12 > > > - removed not needed requirement for gtk+-devel > > > - enable nasm oprimizations on x86 > > ^^^^ > > > > This caused problems on k5 (segv on cpuid). I don't know if it's still > > true though. > > Forgot: Test case: try running mpeg_play. > I don't have any mpeg's to test if it works :-) Did you mean that it does or doesn't t work? If doesn't it will be strange, becouse it is by default on, and I think SDL developers know what they are doing... -- Filip Kali?ski From qboosh at pld.org.pl Wed Mar 20 18:21:15 2002 From: qboosh at pld.org.pl (Jakub Bogusz) Date: Wed, 20 Mar 2002 18:21:15 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320171008.GA13222@ep09.kernel.pl> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> Message-ID: <20020320172115.GA2875@satan.blackhosts> On Wed, Mar 20, 2002 at 06:10:08PM +0100, Michal Moskal wrote: > On Wed, Mar 20, 2002 at 04:34:10PM +0000, Filip Kalinski wrote: > > Module name: SPECS > > Changes by: filon > > > > Changed RCS file: /cvsroot/SPECS/SDL.spec,v > > ---------------------------- > > revision 1.77 > > date: 2002/03/20 16:34:07; author: filon; state: Exp; lines: +20 -12 > > - removed not needed requirement for gtk+-devel > > - enable nasm oprimizations on x86 > ^^^^ > > This caused problems on k5 (segv on cpuid). I don't know if it's still > true though. If so, nasm can be enabled on i586/i686/athlon only (k5 belongs to i486 arch). -- Jakub Bogusz http://prioris.mini.pw.edu.pl/~qboosh/ From malekith at pld.org.pl Wed Mar 20 18:23:39 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Wed, 20 Mar 2002 18:23:39 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320171911.GA738@fk.localdomain> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> <20020320171133.GB13222@ep09.kernel.pl> <20020320171911.GA738@fk.localdomain> Message-ID: <20020320172339.GA10459@ep09.kernel.pl> On Wed, Mar 20, 2002 at 06:19:11PM +0100, Filip Kalinski wrote: > I don't have any mpeg's to test if it works :-) > > Did you mean that it does or doesn't t work? If doesn't it will be > strange, becouse it is by default on, and I think SDL developers know > what they are doing... It caused SEGV when playing mpegs with mpeg_play on k5 box. Maybe I had buggy box :) -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From malekith at pld.org.pl Wed Mar 20 18:24:07 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Wed, 20 Mar 2002 18:24:07 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320172115.GA2875@satan.blackhosts> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> <20020320172115.GA2875@satan.blackhosts> Message-ID: <20020320172407.GB10459@ep09.kernel.pl> On Wed, Mar 20, 2002 at 06:21:15PM +0100, Jakub Bogusz wrote: > If so, nasm can be enabled on i586/i686/athlon only (k5 belongs to i486 > arch). I had *i586* packages on it, and everything worked fine. (beside this). -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From qboosh at pld.org.pl Wed Mar 20 19:56:06 2002 From: qboosh at pld.org.pl (Jakub Bogusz) Date: Wed, 20 Mar 2002 19:56:06 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320172407.GB10459@ep09.kernel.pl> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> <20020320172115.GA2875@satan.blackhosts> <20020320172407.GB10459@ep09.kernel.pl> Message-ID: <20020320185606.GA16171@satan.blackhosts> On Wed, Mar 20, 2002 at 06:24:07PM +0100, Michal Moskal wrote: > On Wed, Mar 20, 2002 at 06:21:15PM +0100, Jakub Bogusz wrote: > > If so, nasm can be enabled on i586/i686/athlon only (k5 belongs to i486 > > arch). > > I had *i586* packages on it, and everything worked fine. (beside this). Ouch, I thought about AMD 5x86. Of course AMD K5 is i586 (even with TSC according to docs I have). SEGV on cpuid sounds strange, because: 1. SDL even checks eflags for CPUID support, 2. SDL isn't the only library/program that uses cpuid... -- Jakub Bogusz http://prioris.mini.pw.edu.pl/~qboosh/ From malekith at pld.org.pl Thu Mar 21 16:16:29 2002 From: malekith at pld.org.pl (Michal Moskal) Date: Thu, 21 Mar 2002 16:16:29 +0100 Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320185606.GA16171@satan.blackhosts> References: <20020320163411.78E22E1048@wun.zie.pg.gda.pl> <20020320171008.GA13222@ep09.kernel.pl> <20020320172115.GA2875@satan.blackhosts> <20020320172407.GB10459@ep09.kernel.pl> <20020320185606.GA16171@satan.blackhosts> Message-ID: <20020321151629.GA13039@ep09.kernel.pl> On Wed, Mar 20, 2002 at 07:56:06PM +0100, Jakub Bogusz wrote: > On Wed, Mar 20, 2002 at 06:24:07PM +0100, Michal Moskal wrote: > > On Wed, Mar 20, 2002 at 06:21:15PM +0100, Jakub Bogusz wrote: > > > If so, nasm can be enabled on i586/i686/athlon only (k5 belongs to i486 > > > arch). > > > > I had *i586* packages on it, and everything worked fine. (beside this). > > Ouch, I thought about AMD 5x86. Of course AMD K5 is i586 (even with TSC > according to docs I have). Ok :) > SEGV on cpuid sounds strange, because: 1. SDL even checks eflags for > CPUID support, 2. SDL isn't the only library/program that uses cpuid... /proc/cpuinfo claimed that CPU has cpuid. But issing it caused SEGV. Unfortunetly I cannot check it right now... anybody with k5? -- : Michal Moskal :::::::: malekith/at/pld.org.pl : GCS {C,UL}++++$ a? !tv : PLD Linux ::::::: Wroclaw University, CS Dept : {E-,w}-- {b++,e}>+++ h From blues at ds6.pg.gda.pl Thu Mar 21 18:43:58 2002 From: blues at ds6.pg.gda.pl (Blues) Date: Thu, 21 Mar 2002 18:43:58 +0100 (CET) Subject: filon: SPECS SDL.spec,1.76,1.77 In-Reply-To: <20020320172115.GA2875@satan.blackhosts> Message-ID: On Wed, 20 Mar 2002, Jakub Bogusz wrote: > > This caused problems on k5 (segv on cpuid). I don't know if it's still > > true though. > If so, nasm can be enabled on i586/i686/athlon only (k5 belongs to i486 > arch). wrong. It's i586. -- --------------------------------- pozdr. Pawe? Go?aszewski --------------------------------- CPU not found - software emulation... From ser at metalab.unc.edu Sun Mar 24 02:52:18 2002 From: ser at metalab.unc.edu (--s+) Date: Sun, 24 Mar 2002 02:52:18 +0100 Subject: [apache-mod_msff.spec] Message-ID: <20020324015218.GA28690@szafa.hyperreal.pl> i've prepared apache-mod_msff.spec - essential apache module for real bofhs! -- http://ibiblio.org/ser/ | http://it-zone.org/ ________________________________________________________ linuxdoc.org|gnu.org|hyperreal.pl|pld.org.pl|it-zone.org "The moon is made of green cheese." -- John Heywood From michal at michal.waw.pl Sun Mar 24 10:22:01 2002 From: michal at michal.waw.pl (Michal Kochanowicz) Date: Sun, 24 Mar 2002 10:22:01 +0100 Subject: [apache-mod_msff.spec] In-Reply-To: <20020324015218.GA28690@szafa.hyperreal.pl> References: <20020324015218.GA28690@szafa.hyperreal.pl> Message-ID: <20020324092201.GC2998@wieszak.mmm.ozarow-12.waw.pl> On Sun, Mar 24, 2002 at 02:52:18AM +0100, --s+ wrote: > i've prepared apache-mod_msff.spec - essential apache > module for real bofhs! It looks like a very cool module (I'm going to install it on some machines, where nobody will shoot be for doing this ;( ) but I have a question. Where did you get this: #v+ tej firmy w pe?nej realizacji Wolnych Pi?tk?w [czyli dni lu?niejszych, w kt?rych _nawet_ mo?na przyj?? do pracy w d?insach :-)))) #v- from? -- --= Michal Kochanowicz==--==--==BOFH==--==--==michal at michal.waw.pl =-- --= finger me for PGP public key or visit http://michal.waw.pl/PGP =-- --==--==--==--==--==-- Vodka. Connecting people.--==--==--==--==--==-- A chodzenie po g?rach SSIE!!! From ser at metalab.unc.edu Sun Mar 24 12:01:33 2002 From: ser at metalab.unc.edu (--s+) Date: Sun, 24 Mar 2002 12:01:33 +0100 Subject: [apache-mod_msff.spec] In-Reply-To: <20020324092201.GC2998@wieszak.mmm.ozarow-12.waw.pl> References: <20020324015218.GA28690@szafa.hyperreal.pl> <20020324092201.GC2998@wieszak.mmm.ozarow-12.waw.pl> Message-ID: <20020324110133.GC4093@szafa.hyperreal.pl> --->[Quoting Michal Kochanowicz :] > Where did you get this: > #v+ > tej firmy w pe?nej realizacji Wolnych Pi?tk?w [czyli dni > lu?niejszych, w kt?rych _nawet_ mo?na przyj?? do pracy w > d?insach :-)))) > #v- > from? >From my mind. -- http://ibiblio.org/ser/ | http://it-zone.org/ ________________________________________________________ linuxdoc.org|gnu.org|hyperreal.pl|pld.org.pl|it-zone.org "The moon is made of green cheese." -- John Heywood From blues at ds6.pg.gda.pl Mon Mar 25 11:13:51 2002 From: blues at ds6.pg.gda.pl (Blues) Date: Mon, 25 Mar 2002 11:13:51 +0100 (CET) Subject: [MBT] new entry in pkg openssh "Xforwarding doesn't work" In-Reply-To: <20020325095944.11149.qmail@ep09.kernel.pl> Message-ID: On 25 Mar 2002 bugs at pld.org.pl wrote: > Date: 2002-03-25 10:59:36+01 Author: Sebastian Zagrodzki (zagrodzki) > Title: Xforwarding doesn't work > Ticket ID: #125 > Package: openssh-3.1p1-1 > Distribution: PLD-Ra.main > Category: doesn't work as it should > Current state: closed -- resolved > Text: > > Configuration file for sshd should contain: > X11UseLocalhost yes > if this line is commented out, sshd uses "::1" instead of "127.0.0.1" when creating socket, and X applications try to connect to "localhost" via IPv4. > *** State changed to 'closed -- resolved' Czy to naprawd? rozwi?zuje problem? Nie mam w tym momencie gdzie tego sprawdzi?. I je?eli to jest rozwi?zanie to w default configa powinno trafi?... -- --------------------------------- pozdr. Pawe? Go?aszewski --------------------------------- CPU not found - software emulation... From wrobell at ite.pl Thu Mar 28 14:42:16 2002 From: wrobell at ite.pl (wrobell) Date: Thu, 28 Mar 2002 14:42:16 +0100 Subject: [builder] vim 6.1 Message-ID: <20020328134215.GB14325@pred.posexperts.com.pl> Please send vim 6.1 to builders. There are only bug fixes between vim 6.0.208 (stable) and vim 6.1 (cvs). I am using vim 6.1 two days and it is stable for me. It can be put in PLD 1.0 in my opinion. wrobell