Re: [gentoo-user] What's the story with 2.6.35 kernel?
Ajai Khattri a...@bway.net writes: I upgraded several machines and some failed to boot 2.6.35, they just hang after grub starts loading the kernel. Some of these I managed to fix by comparing kernel configs with working machines, others dont work at all. The worst case is one where Ive upgraded udev to the latest which only works with kernels 2.6.25 or higher, and the last working kernel on that machine is 2.6.24 (grr!). Wondering if I can download a binary package of udev-149 from somewhere (or can I build it on another machine in a sandbox and package it with quickpkg?). Im kind of under pressure to fix this box but dont have udev-149 on any other machine available :-( That'll teach me to upgrade to the latest and greatest too fast... I think that you may be suffering from bug https://bugs.gentoo.org/show_bug.cgi?id=334269
Re: [gentoo-user] Re: undetected DVD r/w device
Selon walt w41...@gmail.com: On 09/10/2010 07:25 AM, alain.didierj...@free.fr wrote: Selon alain.didierj...@free.fr: Selon Strollerstrol...@stellar.eclipse.co.uk: On 6 Sep 2010, at 09:55, alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, ... symbolic links get created, no /dev/sr0. k3b still refuses to work. BUT when I # mount --bind / temporary I do get a /temporary/dev/sr0 !! What give ? I'm reluctant to add a NAME=sr0 rule as it should be the default; What's going on ? AFAICT, you didn't say if you read the earlier thread cited by Stroller: http://www.gossamer-threads.com/lists/gentoo/user/216290 Sure did. didn't help much.
[gentoo-user] IOAPIC kernel error and oops!
I just compiled and installed gentoo 2.6.35-r4 on an x86 machine and it failed miserably on boot up. As this kernel configuration of mine is not materially different to the 2.6.34 version, I am wondering if it is suffering from some congenital defect. Anyone else having such problems? -- Regards, Mick signature.asc Description: This is a digitally signed message part.
[gentoo-user] Re: IOAPIC kernel error and oops!
On Saturday 11 September 2010 07:57:32 you wrote: I just compiled and installed gentoo 2.6.35-r4 on an x86 machine and it failed miserably on boot up. As this kernel configuration of mine is not materially different to the 2.6.34 version, I am wondering if it is suffering from some congenital defect. Anyone else having such problems? I just saw other posts on this in the M/L, which I somehow missed on my first search. It seems that it may be a gcc bug. Fair enough. I am somehow confused though how I have ended up with this kernel installed on my machine. eix -l gentoo-sources shows that it is masked! Did the mask flip on off while I was between emerges or something? -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Properly handling missing files when downloading files from ftp:// via ISA proxy (emerge/wget)
On Friday 10 September 2010 17:13:44 Maciej Grela wrote: 2010/9/10 Paul Hartman paul.hartman+gen...@gmail.com: On Fri, Sep 10, 2010 at 9:43 AM, Maciej Grela maciej.gr...@gmail.com wrote: Hi, Is there any way to make emerge (wget) correctly behave when it tries to download a non-existing file from FTP in a network using ISA as ftp_proxy ? I have one of these at work and it's really annoying because of situations If you are allowed to bypass the proxy, try --use-proxy=off in your wget command line. Unfortunately proxy is the only way to access the Internet. It seems your FTP is proxied over HTTP and your client needs to support this. I'm not sure if wget supports HTTP proxies for FTP. Here is a document from Microsoft about configuring ISA and various clients: http://technet.microsoft.com/en-us/library/bb794745.aspx I'll read that, thanks. I think that you can try exporting your proxy's http address to ftp like so: # export ftp_proxy=http://my_proxy.com:1234; and then run emerge to see if you can get through. If your proxy requires a username/passwd you'll need to add these on the command line just as the handbook advises. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
[gentoo-user] Moving / around...
Hi, I plan to convert (==reinstall) my system to be 64bit. Since I have an already working and configure 32bit Gentoo- system I would like to do the migration as follows: Create another / partition somwhere on my harddisk Install/Create a new 64bit Gentoo root there. If everything works fine: Delete 32bit-/ and move (cp -a or something like that) the 64bit-/ onto the now empty 32bit-/. BUT: Are there any -- especiall system-related -- binaries or such, which get an hardcoded compiled in, so they would fail to work after / is moved to another place than where it was created? Best regards mcc
[gentoo-user] sudo in kernel config ?
few months ago, I read linux kernel in a nutschell, and the author wrote we shouldn't do kernel operations (config and build) as root. Is sudo (or kdesudo ?) a good replacement to that ? Kdesudo works good to have xconfig, which is more comfortable that menuconfig. But is it a good manner of making things ? -- Stéphane Guedon page web : http://www.22decembre.eu/ carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] How to correctly read CPU temperature ?
On Saturday 11 September 2010, meino.cra...@gmx.de wrote: Hi, with the command sensors (lm_sensor) I can read out the temperatures/voltages of my mobo/cpu: atk0110-acpi-0 Adapter: ACPI interface Vcore Voltage: +1.35 V (min = +0.80 V, max = +1.60 V) CPU/NB Voltage:+1.16 V (min = +0.80 V, max = +1.60 V) CPU VDDA Voltage: +2.50 V (min = +2.00 V, max = +3.00 V) DRAM Voltage: +1.51 V (min = +1.40 V, max = +1.90 V) HT Voltage:+1.20 V (min = +0.80 V, max = +1.50 V) NB Voltage:+1.10 V (min = +0.90 V, max = +1.35 V) SB Voltage:+1.11 V (min = +0.80 V, max = +1.50 V) +3.3V Voltage: +3.34 V (min = +2.97 V, max = +3.63 V) +5V Voltage: +4.97 V (min = +4.50 V, max = +5.50 V) +12V Voltage: +12.21 V (min = +10.20 V, max = +13.80 V) CPU Temperature: +35.0 C (high = +40.0 C, crit = +90.0 C) MB Temperature:+31.0 C (high = +35.0 C, crit = +95.0 C) NB Temperature:+46.0 C (high = +65.0 C, crit = +95.0 C) SB Temperature:+40.0 C (high = +35.0 C, crit = +75.0 C) OPT_TEMP1 Temperature: +0.0 C (high = +0.0 C, crit = +90.0 C) OPT_TEMP2 Temperature: +0.0 C (high = +0.0 C, crit = +90.0 C) OPT_TEMP3 Temperature: +0.0 C (high = +0.0 C, crit = +90.0 C) But there is something I worry about: The CPU temperature. On the Inet I found some, but not very clear infos, which say, that the temperature sensing diodes of the AMD Phenom II x6 T1090 were wrong. Second thing is, when idleing the CPU of my box has only 34 degree C -- which would be nice if true, but I dont believe that: The CPU is cooled with a Scythe Mulgen 2 Rev.B or with other words its only a fan and therefore only air cooling... that line is not from the diode. You need k8temp to read the cpu temperature. Also it is not wrong for all phenom II. Just some. this is with an X4: sensors k10temp-pci-00c3 Adapter: PCI adapter temp1: +55.2°C (high = +70.0°C) w83627ehf-isa-0290 Adapter: ISA adapter Vcore: +0.98 V (min = +0.00 V, max = +1.74 V) in1: +0.14 V (min = +2.04 V, max = +2.04 V) ALARM AVCC:+3.30 V (min = +2.98 V, max = +3.63 V) VCC: +3.30 V (min = +2.98 V, max = +3.63 V) in4: +1.68 V (min = +1.53 V, max = +2.04 V) in5: +1.69 V (min = +1.91 V, max = +2.04 V) ALARM in6: +1.86 V (min = +2.04 V, max = +2.01 V) ALARM 3VSB:+3.28 V (min = +2.98 V, max = +3.63 V) Vbat:+3.28 V (min = +2.70 V, max = +3.30 V) in9: +1.65 V (min = +2.04 V, max = +1.78 V) ALARM fan1: 0 RPM (min = 703 RPM, div = 128) ALARM fan2:547 RPM (min =0 RPM, div = 32) fan3: 0 RPM (min =0 RPM, div = 128) fan5: 0 RPM (min =0 RPM, div = 128) temp1: +35.0°C (high = -5.0°C, hyst = +125.0°C) sensor = thermistor temp2: +49.0°C (high = +80.0°C, hyst = +75.0°C) sensor = thermistor temp3: +47.5°C (high = +80.0°C, hyst = +75.0°C) sensor = thermistor cpu0_vid: +0.375 V cooled with a little Scythe Shuriken. On idle these CPUs don't need much current, which means that they are pretty cool. Bios says the same. So the temps are close to the truth. (temp 2 is cpu). When booted into the BIOS the hardware info shows a higher temperature for the CPU while idleing than the above shown command output. But: From where the BIOS takes its informations? the same sensors. BUT: bios does ZERO powersaving. So the CPU is at its highest clock and highest voltage setting, not even idling, so it sucks much more power.
[gentoo-user] static-libs
synce few days, I have a message of portage suggestiung me to use the static- libs USE flag for media-libs/jpeg-6b. What may be the consequence ? Please be gentle with explaining this sorte of things, as I have not the knowledges to understand the full compile process, otherwise I am a little bit familiar with it ! Thanks -- Stéphane Guedon page web : http://www.22decembre.eu/ carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] sudo in kernel config ?
On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. What the author is saying is that, to an extent, in theory no one should compile anything as root, or really do anything non-system-adminly as root. You should only do as root what is critically necessary (e.g. make install) as root. In a perfect, tidy world we'd all do that. This world, however does not exist. Even portage, by default does configure and make as root (albeit in a sandbox so it is safe(r). What the author means is theoretically the config/compile phase could unintentionally cause some kind of harm to your system. In practice I have never seen this or heard of it. The kernel devs are bright enough to ensure that the compilation does nothing outside the source tree itself. It's a good guideline but, like the government's dietary guidelines, not ones I intend to follow religiously. Is sudo (or kdesudo ?) a good replacement to that ? sudo runs things as root, so effectively you've done nothing but add a password prompt to the mix. Gentoo actually makes this a bit more difficult, because usually one uses portage to install the kernel sources, and they get installed as root-owned, and only root has write access to the kernel tree. Some people, such as myself, use kernel sources outside of portage (I follow a git repo) and do so as a non-root user. In this case the kernel tree is not owned by root and the config/compile is easily done as a non-root user. If you are super-paranoid. You can make a non-root copy of /usr/src/linux and compile it as a non-root user. But there really isn't any point in using sudo. It's effectively doing the same thing that you are trying to avoid.
Re: [gentoo-user] Moving / around...
meino.cra...@gmx.de writes: I plan to convert (==reinstall) my system to be 64bit. Since I have an already working and configure 32bit Gentoo- system I would like to do the migration as follows: Create another / partition somwhere on my harddisk Install/Create a new 64bit Gentoo root there. If everything works fine: Delete 32bit-/ and move (cp -a or something like that) the 64bit-/ onto the now empty 32bit-/. I would mount -o bind / /mnt/binroot and cp -a or something /mnt/bindroot to the 32bit-/, so you don't copy things as /proc and /dev as well. BUT: Are there any -- especiall system-related -- binaries or such, which get an hardcoded compiled in, so they would fail to work after / is moved to another place than where it was created? No. The path would still be the same, whatever the underlying device is. Wonko
Re: [gentoo-user] sudo in kernel config ?
Le Saturday 11 September 2010 11:46:59, Albert Hopkins a écrit : On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. What the author is saying is that, to an extent, in theory no one should compile anything as root, or really do anything non-system-adminly as root. You should only do as root what is critically necessary (e.g. make install) as root. In a perfect, tidy world we'd all do that. This world, however does not exist. Even portage, by default does configure and make as root (albeit in a sandbox so it is safe(r). What the author means is theoretically the config/compile phase could unintentionally cause some kind of harm to your system. In practice I have never seen this or heard of it. The kernel devs are bright enough to ensure that the compilation does nothing outside the source tree itself. It's a good guideline but, like the government's dietary guidelines, not ones I intend to follow religiously. Is sudo (or kdesudo ?) a good replacement to that ? sudo runs things as root, so effectively you've done nothing but add a password prompt to the mix. Gentoo actually makes this a bit more difficult, because usually one uses portage to install the kernel sources, and they get installed as root-owned, and only root has write access to the kernel tree. Some people, such as myself, use kernel sources outside of portage (I follow a git repo) and do so as a non-root user. In this case the kernel tree is not owned by root and the config/compile is easily done as a non-root user. If you are super-paranoid. You can make a non-root copy of /usr/src/linux and compile it as a non-root user. But there really isn't any point in using sudo. It's effectively doing the same thing that you are trying to avoid. I am not paranoid anymore, just asking to knowing persons... Ok ! thanks for your answer ! -- Stéphane Guedon page web : http://www.22decembre.eu/ carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] undetected DVD r/w device
Le Monday 06 September 2010 17:11:17, alain.didierj...@free.fr a écrit : Selon Stroller strol...@stellar.eclipse.co.uk: On 6 Sep 2010, at 09:55, alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, ... Current kernels usually call optical drives /dev/sr0 (/dev/sr1, c). There's no sr* device on my system !!! Only way to reach the drive: /dev/hda. Help So, you're not using the latest drivers (scsi emultation of ata hdd)... -- Stéphane Guedon page web : http://www.22decembre.eu/ carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Moving / around...
Alex Schuster wo...@wonkology.org [10-09-11 12:08]: meino.cra...@gmx.de writes: I plan to convert (==reinstall) my system to be 64bit. Since I have an already working and configure 32bit Gentoo- system I would like to do the migration as follows: Create another / partition somwhere on my harddisk Install/Create a new 64bit Gentoo root there. If everything works fine: Delete 32bit-/ and move (cp -a or something like that) the 64bit-/ onto the now empty 32bit-/. I would mount -o bind / /mnt/binroot and cp -a or something /mnt/bindroot to the 32bit-/, so you don't copy things as /proc and /dev as well. BUT: Are there any -- especiall system-related -- binaries or such, which get an hardcoded compiled in, so they would fail to work after / is moved to another place than where it was created? No. The path would still be the same, whatever the underlying device is. Wonko I think there is some misunderstanding: Before migration to 64bit: /dev/sda3 is mounted on / and contains the 32bit Gentoo /dev/sda10 is mounted on /home/mcc/migration and will contain the stuff of the 64bit Gentoo After migration I will *not* mount /dev/sda10 on / but will clear all stuff from /dev/sda3 and move the contents from /dev/sda10 to /dev/sda3. Is still valid what you said under this premissions, Wonko? Thanks a lot for your help in advance! Best regards mcc
Re: [gentoo-user] Moving / around...
On Sat, 11 Sep 2010 11:08:49 +0200, Alex Schuster wrote: I would mount -o bind / /mnt/binroot and cp -a or something /mnt/bindroot to the 32bit-/, so you don't copy things as /proc and /dev as well. Or use the -x option with cp, although I prefer to use rsync -ax for this type of thing. -- Neil Bothwick WindowError:01B Illegal error. Do NOT get this error. signature.asc Description: PGP signature
Re: [gentoo-user] Re: undetected DVD r/w device
On 11 Sep 2010, at 07:38, alain.didierj...@free.fr wrote: ... AFAICT, you didn't say if you read the earlier thread cited by Stroller: http://www.gossamer-threads.com/lists/gentoo/user/216290 Sure did. didn't help much. No, it wouldn't if you didn't bother to read it or actually follow its advice. Where is the the output of `dmesg`? You have not sent it. Stroller.
Re: [gentoo-user] sudo in kernel config ?
On Saturday 11 September 2010, Stéphane Guedon wrote: Le Saturday 11 September 2010 11:46:59, Albert Hopkins a écrit : On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. What the author is saying is that, to an extent, in theory no one should compile anything as root, or really do anything non-system-adminly as root. You should only do as root what is critically necessary (e.g. make install) as root. In a perfect, tidy world we'd all do that. This world, however does not exist. Even portage, by default does configure and make as root (albeit in a sandbox so it is safe(r). What the author means is theoretically the config/compile phase could unintentionally cause some kind of harm to your system. In practice I have never seen this or heard of it. The kernel devs are bright enough to ensure that the compilation does nothing outside the source tree itself. It's a good guideline but, like the government's dietary guidelines, not ones I intend to follow religiously. Is sudo (or kdesudo ?) a good replacement to that ? sudo runs things as root, so effectively you've done nothing but add a password prompt to the mix. Gentoo actually makes this a bit more difficult, because usually one uses portage to install the kernel sources, and they get installed as root-owned, and only root has write access to the kernel tree. Some people, such as myself, use kernel sources outside of portage (I follow a git repo) and do so as a non-root user. In this case the kernel tree is not owned by root and the config/compile is easily done as a non-root user. If you are super-paranoid. You can make a non-root copy of /usr/src/linux and compile it as a non-root user. But there really isn't any point in using sudo. It's effectively doing the same thing that you are trying to avoid. I am not paranoid anymore, just asking to knowing persons... Ok ! thanks for your answer ! well, some years ago someone made a mistake causing some people doing make as root loosing /dev/null or something like that. But not even everybody was hit. /me prefers loosing /dev/null over having /home/$USER overwritten.
Re: [gentoo-user] Moving / around...
On Sat, Sep 11, 2010 at 5:19 AM, meino.cra...@gmx.de wrote: SNIP I think there is some misunderstanding: Before migration to 64bit: /dev/sda3 is mounted on / and contains the 32bit Gentoo /dev/sda10 is mounted on /home/mcc/migration and will contain the stuff of the 64bit Gentoo After migration I will *not* mount /dev/sda10 on / but will clear all stuff from /dev/sda3 and move the contents from /dev/sda10 to /dev/sda3. Is still valid what you said under this premissions, Wonko? Thanks a lot for your help in advance! Best regards mcc Why not mount /dev/sda10 as root and be done with it.? No need to move anything. Do the 64-bit install as you are suggesting. Do NOT install grub. Place the 64-bit kernel in the current /boot pointing at /dev/sda10. Modify grub.conf to allow you to boot either /dev/sda3 (your 32-bit install) or /dev/sda10. (your 64-bit install) Boot both installs a few times and test that each is working. (They will be) Use the 64-bit install for a few days and make sure it's working. When it is don't boot 32-bit for a week or two, just leaving it there on the drive because almost certainly you will have forgotten to copy something over. (I always do...) Only when you are comfortable that 64-bit is working correctly delete the 32-bit on /dev/sda3 if you need the disk space. Remember, leaving /home out of the picture a Gentoo install takes maybe 10GB. It's not that large. Probably less if you shared the portage distfiles directory between the two. It doesn't hurt very much to have multiple installs on the same drive in different partitions. It's what I did playing with a stable and a testing install. I eventually deleted the testing install and just went with stable and a few testing application packages. (I still don't understand why any normal user wants a ~amd64 install but that's just me!) ;-) Hope this helps, Mark
Re: [gentoo-user] Moving / around...
Mark Knecht markkne...@gmail.com [10-09-11 17:08]: On Sat, Sep 11, 2010 at 5:19 AM, meino.cra...@gmx.de wrote: SNIP I think there is some misunderstanding: Before migration to 64bit: /dev/sda3 is mounted on / and contains the 32bit Gentoo /dev/sda10 is mounted on /home/mcc/migration and will contain the stuff of the 64bit Gentoo After migration I will *not* mount /dev/sda10 on / but will clear all stuff from /dev/sda3 and move the contents from /dev/sda10 to /dev/sda3. Is still valid what you said under this premissions, Wonko? Thanks a lot for your help in advance! Best regards mcc Why not mount /dev/sda10 as root and be done with it.? No need to move anything. ...because data access at the outer partitions are faster than those in the middle... Do the 64-bit install as you are suggesting. Do NOT install grub. Place the 64-bit kernel in the current /boot pointing at /dev/sda10. Modify grub.conf to allow you to boot either /dev/sda3 (your 32-bit install) or /dev/sda10. (your 64-bit install) Boot both installs a few times and test that each is working. (They will be) Use the 64-bit install for a few days and make sure it's working. When it is don't boot 32-bit for a week or two, just leaving it there on the drive because almost certainly you will have forgotten to copy something over. (I always do...) Only when you are comfortable that 64-bit is working correctly delete the 32-bit on /dev/sda3 if you need the disk space. In the docs on gentoo-wiki (or? somewhere else?) I read that some kind of data are not portable namely databases... Remember, leaving /home out of the picture a Gentoo install takes maybe 10GB. It's not that large. Probably less if you shared the portage distfiles directory between the two. It doesn't hurt very much to have multiple installs on the same drive in different partitions. It's what I did playing with a stable and a testing install. I eventually deleted the testing install and just went with stable and a few testing application packages. (I still don't understand why any normal user wants a ~amd64 install but that's just me!) ;-) The normal user like me want 64bit application to access more than 2GB per task. In my case: Rendering and simulation takes a LOT of memory especially when it comes to huge counts of vertice or particle interactions. Therefore I plan to install 8GByte RAM. Hope this helps, Mark Yes Mark, it helps! Thanks a lot! :) Best regards mcc
Re: [gentoo-user] sudo in kernel config ?
On Sat, 2010-09-11 at 05:46 -0400, Albert Hopkins wrote: In a perfect, tidy world we'd all do that. This world, however does not exist. Even portage, by default does configure and make as root (albeit in a sandbox so it is safe(r). I suppose one could compile the kernel sources as root but inside sandbox, though I've never tried that.
Re: [gentoo-user] undetected DVD r/w device
Selon Joerg Schilling joerg.schill...@fokus.fraunhofer.de: alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, but /dev/cdrom is not created at boot time and k3b returns No optical drive found. K3b did not find any optical device in your system. Solution : Make sure HAL daemon is running, it is used by K3b for finding devices. If you call cdrecord (release 3.00): cdrecord -scanbus or cdrecord -checkdrive and it finds a drive, then there is a bug in k3b. al...@isba ~ $ cdrecord --checkdrive Cdrecord-ProDVD-ProBD-Clone 3.00 (x86_64-unknown-linux-gnu) Copyright (C) 1995-2010 Jörg Schilling Linux sg driver version: 3.5.27 Using libscg version 'schily-0.9'. No target specified, trying to find one... Using dev=1000,0,0. Device type: Removable CD-ROM Version: 0 Response Format: 2 Capabilities : Vendor_info: 'TSSTcorp' Identifikation : 'CD/DVDW SH-W162C' Revision : 'TS10' Device seems to be: Generic mmc2 DVD-R/DVD-RW/DVD-RAM. Using generic SCSI-3/mmc CD-R/CD-RW driver (mmc_cdr). Driver flags : MMC-3 SWABAUDIO BURNFREE Supported modes: TAO PACKET SAO SAO/R96P SAO/R96R RAW/R16 RAW/R96P RAW/R96R cdrecord: Warning: Cannot read drive buffer. cdrecord: Warning: The DMA speed test has been skipped. al...@isba ~ $ dmesg | grep hda hda: TSSTcorpCD/DVDW SH-W162C, ATAPI CD/DVD-ROM drive hda: host max PIO5 wanted PIO255(auto-tune) selected PIO4 hda: UDMA/33 mode selected ide-cd: hda: ATAPI 48X DVD-ROM DVD-R CD-R/RW drive, 2048kB Cache -- ~adj~
Re: [gentoo-user] undetected DVD r/w device
Selon Stéphane Guedon steph...@22decembre.eu: Le Monday 06 September 2010 17:11:17, alain.didierj...@free.fr a écrit : Selon Stroller strol...@stellar.eclipse.co.uk: On 6 Sep 2010, at 09:55, alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, ... Current kernels usually call optical drives /dev/sr0 (/dev/sr1, c). There's no sr* device on my system !!! Only way to reach the drive: /dev/hda. Help So, you're not using the latest drivers (scsi emultation of ata hdd)... -- Stéphane Guedon page web : http://www.22decembre.eu/ carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc I use * ATA/ATAPI/MFM/RLL support (DEPRECATED) --- in Device Drivers as I've done in the past, and it used to work fine || ~adj~
Re: [gentoo-user] Re: What's the story with 2.6.35 kernel?
On Sat, 11 Sep 2010, Nikos Chantziaras wrote: Extremely bad idea. It's no wonder something did break. You can't just silentoldconfig between kernel versions and expect it to always work. At least oldconfig would catch new options (which might be options replacing old ones.) silentoldconfig shows all the new options too. I think the main problem was the changes in udev. -- A
Re: [gentoo-user] Moving / around...
On Sat, Sep 11, 2010 at 8:23 AM, meino.cra...@gmx.de wrote: SNIP Why not mount /dev/sda10 as root and be done with it.? No need to move anything. ...because data access at the outer partitions are faster than those in the middle... OK, assuming it's really measurable in real life, but I'll point out that you don't necessarily have to 'copy' data from partition to partition to achieve that. I've used gparted to first delete what you are terming /dev/sda3, then enlarge /dev/sda10 toward the side of the drive where you want it, then shrink sda10 when you get it there. Takes a lot of time but works for a dummy like me, and no need to mess with fstab, etc., because it just remains sda10. Granted, that simple example assumes there's nothing in the middle. If there is then I typically shrink and move it also. Not an ideal solution, but it works. But the point remains that you can probably exist with both installs on the drive for some _long_ period of time before you ever get around to these steps for the sake of performance. Certainly don't get rid of the working 32-bit install before you are _completely_ sure the 64-bit is working. - Mark
[gentoo-user] Re: S/MIME passphrase problem with Kleopatra
On Thursday 13 May 2010 11:08:48 you wrote: In the last two weeks I renewed an SSL certificate from Comodo for email usage. This time round Kleopatra is having problems with recognising the passphrase I use. I partially suspect a gnupg bug here probably relating to mime characters, but I am not sure how to troubleshoot it. This is a sequence of events that show how the problem occurs: I export the SSL cert from Firefox as a pkcs12 file. It asks for a passphrase to encrypt it with. It will accept my passphrase and saves the exported .p12 bundle as a file on my hard drive. Then I try to import this into Kleopatra. This is what I have come across here: If I have used a short passphrase when exporting from Firefox (say 8 characters long) there's no problem importing it into Kleopatra. If I use a long passphrase then it fails every time: Please enter a passphrase to unprotect the PKCS#12 object. p4ssPhr4se An error occurred while trying to import the certificate - Decryption failed. The log shows: == [2010-05-12T19:51:45] Log cleared 6 - 2010-05-12 19:52:12 gpg-agent[13563]: failed to unprotect the secret key: Bad passphrase 6 - 2010-05-12 19:52:12 gpg-agent[13563]: failed to read the secret key 6 - 2010-05-12 19:52:12 gpg-agent[13563]: command pksign failed: Bad passphrase 6 - 2010-05-12 19:52:12 gpg-agent[13563.6] DBG: - ERR 67108875 Bad passphrase GPG Agent 4 - 2010-05-12 19:52:12 gpgsm[16759]: error creating signature: Bad passphrase GPG Agent 4 - 2010-05-12 19:52:12 gpgsm[16759.0] DBG: - ERR 67108875 Bad passphrase GPG Agent 4 - 2010-05-12 19:52:12 gpgsm[16759.0] DBG: - BYE 4 - 2010-05-12 19:52:12 gpgsm[16759.0] DBG: - OK closing connection [client at fd 4 disconnected] 5 - 2010-05-12 19:52:12 dirmngr[16760.0] DBG: - [EOF] 6 - 2010-05-12 19:52:12 gpg-agent[13563.6] DBG: - [EOF] 6 - 2010-05-12 19:52:12 gpg-agent[13563]: handler 0xbf04c0 for fd 6 terminated [client at fd 5 disconnected] == Now, as I said above if I use a short passphrase to encrypt the certificate bundle when exporting it from Firefox, I manage to import it into Kleopatra and then I can re-encrypt it with either with the same short passphrase or with a longer passphrase. Kleopatra will accept any length at that stage and import it happily. However, even if I import it into Kleopatra I can't use it thereafter! Every time I try to use it in Kmail to sign/encrypt/decrypt a message it will fail when I enter the passphrase. :-( I have tried to convert the exported pkcs12 file into a pem bundle, but Kleopatra then fails to import it right from the start with a BER error - it doesn't even ask for a passphrase to decrypt it: == [2010-05-07T22:24:22] Log cleared [client at fd 4 connected] 4 - 2010-05-07 22:24:25 gpgsm[14692]: enabled debug flags: assuan 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - # Home: ~/.gnupg 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - # Config: /home/michael/.gnupg/gpgsm.conf 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - # AgentInfo: /tmp/gpg-yRFiu9/S.gpg-agent:13728:1 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - # DirmngrInfo: [not set] 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OK GNU Privacy Guard's S/M server 2.0.14 ready 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OPTION display=:0.0 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OK 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OPTION enable-audit-log=1 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OK 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - INPUT FD=21 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OK 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - IMPORT 4 - 2010-05-07 22:24:25 gpgsm[14692]: invalid radix64 character 2d skipped 4 - 2010-05-07 22:24:25 gpgsm[14692]: invalid radix64 character 3a skipped 4 - 2010-05-07 22:24:25 gpgsm[14692]: invalid radix64 character 2c skipped 4 - 2010-05-07 22:24:25 gpgsm[14692]: invalid radix64 character 2d skipped 4 - 2010-05-07 22:24:25 gpgsm[14692]: invalid radix64 character 3a skipped 4 - 2010-05-07 22:24:25 gpgsm[14692]: invalid radix64 character 2d skipped 4 - 2010-05-07 22:24:25 gpgsm[14692]: total number processed: 0 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - S IMPORT_RES 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - ERR 150995078 BER error KSBA 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - BYE 4 - 2010-05-07 22:24:25 gpgsm[14692.0] DBG: - OK closing connection [client at fd 4 disconnected] == Any idea why Kleopatra fails with this new Comodo certificate? It had/has no problem using the expired certificate by the same CA (of course it is shown as expired now). How could I troubleshoot this thing? Some things I have tried so far: I have imported and used this SSL cert on a
Re: [gentoo-user] undetected DVD r/w device
On Saturday 11 September 2010 16:57:18 alain.didierj...@free.fr wrote: Selon Joerg Schilling joerg.schill...@fokus.fraunhofer.de: alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, but /dev/cdrom is not created at boot time and k3b returns No optical drive found. K3b did not find any optical device in your system. Solution : Make sure HAL daemon is running, it is used by K3b for finding devices. If you call cdrecord (release 3.00): cdrecord -scanbus or cdrecord -checkdrive and it finds a drive, then there is a bug in k3b. al...@isba ~ $ cdrecord --checkdrive Cdrecord-ProDVD-ProBD-Clone 3.00 (x86_64-unknown-linux-gnu) Copyright (C) 1995-2010 Jörg Schilling Linux sg driver version: 3.5.27 Using libscg version 'schily-0.9'. No target specified, trying to find one... Using dev=1000,0,0. This looks fishy ... On mine it is: Using dev=1,0,0. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] undetected DVD r/w device
On Saturday 11 September 2010 17:01:29 alain.didierj...@free.fr wrote: Selon Stéphane Guedon steph...@22decembre.eu: Le Monday 06 September 2010 17:11:17, alain.didierj...@free.fr a écrit : Selon Stroller strol...@stellar.eclipse.co.uk: On 6 Sep 2010, at 09:55, alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, ... Current kernels usually call optical drives /dev/sr0 (/dev/sr1, c). There's no sr* device on my system !!! Only way to reach the drive: /dev/hda. Help So, you're not using the latest drivers (scsi emultation of ata hdd)... -- Stéphane Guedon page web : http://www.22decembre.eu/ carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc I use * ATA/ATAPI/MFM/RLL support (DEPRECATED) --- in Device Drivers as I've done in the past, and it used to work fine Deselect that which is deprecated and will be removed soon and select the appropriate SCSI drivers for your drives. There's been a few messages on this list explaining how to go about it. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
[gentoo-user] Re: sudo in kernel config ?
On 09/11/2010 11:24 AM, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell, and the author wrote we shouldn't do kernel operations (config and build) as root. Is sudo (or kdesudo ?) a good replacement to that ? Kdesudo works good to have xconfig, which is more comfortable that menuconfig. But is it a good manner of making things ? Why sudo? Simply chown -R the whole kernel tree. Only the modules_install and install targets will need root. I've done it like this for as long as I can remember.
Re: [gentoo-user] undetected DVD r/w device
Mick michaelkintz...@gmail.com wrote: If you call cdrecord (release 3.00): cdrecord -scanbus or cdrecord -checkdrive and it finds a drive, then there is a bug in k3b. al...@isba ~ $ cdrecord --checkdrive Cdrecord-ProDVD-ProBD-Clone 3.00 (x86_64-unknown-linux-gnu) Copyright (C) 1995-2010 Jörg Schilling Linux sg driver version: 3.5.27 Using libscg version 'schily-0.9'. No target specified, trying to find one... Using dev=1000,0,0. This looks fishy ... On mine it is: Using dev=1,0,0. This is the result of having a non-orthogonal driver model for SCSI devices in Linux. Some drivers are fully integrated into the SCSI sub-system and receive low SCSI bus numbers from libscg and other drivers are separated and are mapped to high SCSI bus numbers. Some drivers are even accessible from two or more competing drivers. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [gentoo-user] undetected DVD r/w device
Selon Mick michaelkintz...@gmail.com: On Saturday 11 September 2010 17:01:29 alain.didierj...@free.fr wrote: Selon Stéphane Guedon steph...@22decembre.eu: Le Monday 06 September 2010 17:11:17, alain.didierj...@free.fr a écrit : Selon Stroller strol...@stellar.eclipse.co.uk: On 6 Sep 2010, at 09:55, alain.didierj...@free.fr wrote: For some unknown reason, my DVD r/w device is not detected as such by udev: I can mount /dev/hda and read a data CD, ... Current kernels usually call optical drives /dev/sr0 (/dev/sr1, c). There's no sr* device on my system !!! Only way to reach the drive: /dev/hda. Help So, you're not using the latest drivers (scsi emultation of ata hdd)... I use * ATA/ATAPI/MFM/RLL support (DEPRECATED) --- in Device Drivers as I've done in the past, and it used to work fine Deselect that which is deprecated and will be removed soon and select the appropriate SCSI drivers for your drives. There's been a few messages on this list explaining how to go about it. Done. Works both for cdrom unit and old IDE disk, whch now are know as sr0 (ex hda) and sdc (wax hdc). cdrecord --checkdrive says Cdrecord-ProDVD-ProBD-Clone 3.00 (x86_64-unknown-linux-gnu) Copyright (C) 1995-2010 Jörg Schilling Linux sg driver version: 3.5.34 Using libscg version 'schily-0.9'. No target specified, trying to find one... Using dev=4,0,0. Device type: Removable CD-ROM Version: 5 Response Format: 2 Capabilities : Vendor_info: 'TSSTcorp' Identifikation : 'CD/DVDW SH-W162C' Revision : 'TS10' Device seems to be: Generic mmc2 DVD-R/DVD-RW/DVD-RAM. Using generic SCSI-3/mmc CD-R/CD-RW driver (mmc_cdr). Driver flags : MMC-3 SWABAUDIO BURNFREE Supported modes: TAO PACKET SAO SAO/R96P SAO/R96R RAW/R16 RAW/R96P RAW/R96R cdrecord: Warning: Cannot read drive buffer. cdrecord: Warning: The DMA speed test has been skipped. But that silly k3b returns No optical drive found. K3b did not find any optical device in your system. Solution : Make sure HAL daemon is running, it is used by K3b for finding devices. and on the terminal QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied QStringList Solid::Backends::Hal::HalManager::findDeviceByDeviceInterface(const Solid::DeviceInterface::Type) error: org.freedesktop.DBus.Error.AccessDenied k3b(5361)/kdecore (services) KMimeTypeFactory::parseMagic: Now parsing /usr/share/mime/magic I'm lost ! -- ~adj~
Re: [gentoo-user] undetected DVD r/w device
On Saturday 11 September 2010 18:50:04 Mick wrote: Deselect that which is deprecated and will be removed soon I hope it won't. I have one box whose drives the new drivers cannot detect - at least, nothing I've tried has worked so far. and select the appropriate SCSI drivers for your drives. There's been a few messages on this list explaining how to go about it. -- Rgds Peter. Linux Counter 5290, 1994-04-23.
Re: [gentoo-user] Moving / around...
Mark Knecht markkne...@gmail.com [10-09-11 20:40]: On Sat, Sep 11, 2010 at 8:23 AM, meino.cra...@gmx.de wrote: SNIP Why not mount /dev/sda10 as root and be done with it.? No need to move anything. ...because data access at the outer partitions are faster than those in the middle... OK, assuming it's really measurable in real life, but I'll point out that you don't necessarily have to 'copy' data from partition to partition to achieve that. I've used gparted to first delete what you are terming /dev/sda3, then enlarge /dev/sda10 toward the side of the drive where you want it, then shrink sda10 when you get it there. Takes a lot of time but works for a dummy like me, and no need to mess with fstab, etc., because it just remains sda10. Granted, that simple example assumes there's nothing in the middle. If there is then I typically shrink and move it also. Not an ideal solution, but it works. But the point remains that you can probably exist with both installs on the drive for some _long_ period of time before you ever get around to these steps for the sake of performance. Certainly don't get rid of the working 32-bit install before you are _completely_ sure the 64-bit is working. - Mark Hi Mark, sorry, but with gparted Co, I made some experiences which let me leave those tools alone. Maybe the problem sits right in front of my monitor, but... In my case, there is something in the middle, thats why it is /dev/sda3 and /dev/sd10 and not /dev/sda3 and /dev/sd4... So things getting even more complex and especially complexer than cp and friends... Is there any automagical check whic does some basic checking, to find the biggest bugs in a fresh 64bit-installation? (Beside those, which are identical on 32bit and 64bit -- like emerge and such...) ? Best regards, mcc
Re: [gentoo-user] undetected DVD r/w device
Apparently, though unproven, at 20:24 on Saturday 11 September 2010, Peter Humphrey did opine thusly: On Saturday 11 September 2010 18:50:04 Mick wrote: Deselect that which is deprecated and will be removed soon I hope it won't. I have one box whose drives the new drivers cannot detect - at least, nothing I've tried has worked so far. That's the beauty of Linux. When Linus decides he's had enough keeping old stuff hanging around and rips it out, you just put it right back in your own copies. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] sudo in kernel config ?
Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? -- alan dot mckinnon at gmail dot com
[gentoo-user] Re: sudo in kernel config ?
On 09/11/2010 11:18 PM, Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? chown -R
Re: [gentoo-user] sudo in kernel config ?
Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. Glad only root can get to the sources. ;-) Dale :-) :-)
[gentoo-user] Re: sudo in kernel config ?
On 09/11/2010 11:35 PM, Dale wrote: Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. Glad only root can get to the sources. ;-) No, any user can't edit them; only the user you assign the files to. If you assign them to root, only root can edit them. If you assign them to kerneluser, only kerneluser can edit them. This is Unix 101 :)
Re: [gentoo-user] sudo in kernel config ?
On Sat, 11 Sep 2010 15:35:58 -0500 Dale rdalek1...@gmail.com wrote: If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. This was actually a potential risk once upon a time: http://attrition.org/security/advisory/gobbles/GOBBLES-16.txt
Re: [gentoo-user] sudo in kernel config ?
Apparently, though unproven, at 22:28 on Saturday 11 September 2010, Etaoin Shrdlu did opine thusly: On Sat, 11 Sep 2010 15:35:58 -0500 Dale rdalek1...@gmail.com wrote: If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. This was actually a potential risk once upon a time: http://attrition.org/security/advisory/gobbles/GOBBLES-16.txt More like an actual risk all the time. Which is why: # ls -al /usr/src/ total 2 drwxr-xr-x 3 root root 136 2010-09-01 11:41 . drwxr-xr-x 17 root root 480 2010-08-23 01:44 .. -rw-r--r-- 1 root root0 2008-06-17 19:37 .keep lrwxrwxrwx 1 root root 18 2010-09-01 11:30 linux - linux-2.6.35-ck-r2 drwxr-xr-x 24 root root 1584 2010-09-01 02:12 linux-2.6.35-ck-r2 -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Re: sudo in kernel config ?
Nikos Chantziaras wrote: On 09/11/2010 11:35 PM, Dale wrote: Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. Glad only root can get to the sources. ;-) No, any user can't edit them; only the user you assign the files to. If you assign them to root, only root can edit them. If you assign them to kerneluser, only kerneluser can edit them. This is Unix 101 :) My point was, if the sources are say in the user group, then any user can edit them? Right now, they are in the root group and owned my root which for security reasons is a good idea. That way a regular user can't edit or modify the kernel sources. Dale :-) :-)
Re: [gentoo-user] Re: sudo in kernel config ?
Apparently, though unproven, at 22:34 on Saturday 11 September 2010, Nikos Chantziaras did opine thusly: On 09/11/2010 11:18 PM, Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? chown -R I utterly fail to see the point of this prohibition against building as root. Sure, it makes sense when I'm installing perl stuff for my users and last command is sudo make install. But for everything else? -- alan dot mckinnon at gmail dot com
[gentoo-user] Re: static-libs
On 09/11/2010 02:13 AM, Stéphane Guedon wrote: synce few days, I have a message of portage suggestiung me to use the static- libs USE flag for media-libs/jpeg-6b. What may be the consequence ? Please be gentle with explaining this sorte of things, as I have not the knowledges to understand the full compile process, otherwise I am a little bit familiar with it ! I have jpeg-8b, so I can't be sure about 6b. I just turned on the static-libs USE flag and re-installed jpeg. The only difference is that the 'static' lib /usr/lib/libjpeg.a wasn't there before and it is now, that's all. Is your system trying to upgrade jpeg to a newer version? I notice that 6b doesn't use any USE flags, and the newer versions do use the static-libs flag. (Just re-installing 6b shouldn't complain about USE flags because the package doesn't look for them.) Any program that uses the dynamic libjpeg.so would need to be re-compiled if the version of jpeg changes. If the static library is used instead, the program no longer needs libjpeg.so because the static library is linked into the binary executable at compile-time. The price you pay is a larger binary executable, but you never need to worry about future jpeg version changes. I don't know how portage chooses between static and dynamic libs while building a package. Anyone else know?
[gentoo-user] Re: sudo in kernel config ?
On 09/11/2010 11:49 PM, Dale wrote: Nikos Chantziaras wrote: On 09/11/2010 11:35 PM, Dale wrote: Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. Glad only root can get to the sources. ;-) No, any user can't edit them; only the user you assign the files to. If you assign them to root, only root can edit them. If you assign them to kerneluser, only kerneluser can edit them. This is Unix 101 :) My point was, if the sources are say in the user group, then any user can edit them? Right now, they are in the root group and owned my root which for security reasons is a good idea. That way a regular user can't edit or modify the kernel sources. The group can only write if the files have the group write permission set. Still in Unix 101 domain, hehe :)
[gentoo-user] Re: sudo in kernel config ?
On 09/11/2010 11:51 PM, Alan McKinnon wrote: Apparently, though unproven, at 22:34 on Saturday 11 September 2010, Nikos Chantziaras did opine thusly: On 09/11/2010 11:18 PM, Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? chown -R I utterly fail to see the point of this prohibition against building as root. Sure, it makes sense when I'm installing perl stuff for my users and last command is sudo make install. But for everything else? Well, running GCC and Make as root raises the same concerns as running any other program as root. In the case of Gentoo, this isn't too important though, since in Gentoo, you don't build your software in your home dir and then sudo make install, but portage will run GCC as root anyway. With other distros, you never run GCC/Make/etc as root. In Gentoo you do, so there's no point in reading too much into this.
Re: [gentoo-user] Moving / around...
meino.cra...@gmx.de writes: I think there is some misunderstanding: Before migration to 64bit: /dev/sda3 is mounted on / and contains the 32bit Gentoo /dev/sda10 is mounted on /home/mcc/migration and will contain the stuff of the 64bit Gentoo After migration I will *not* mount /dev/sda10 on / but will clear all stuff from /dev/sda3 and move the contents from /dev/sda10 to /dev/sda3. Is still valid what you said under this premissions, Wonko? That's how I understood it, although I assumed the temproary 64bit install would be on a 2nd drive, thus you would copy it back once it seems to work. No, I see no problem with this. About performance: I'm not sure it will be even noticeable. Yes, most drives (but not all) are organized so the first partitions go to the outside, which is faster. With LVM, I used to create two volume groups on my drive, a group for swap and the system, and another one for data. But then I thought it's not worth the effort, and I lose some of the LVM benefits. Well, with everything encrypted I don't get full performance anyway, so my case might be a little different. But the performance increase is only true when reading lots of data. I'm not sure how big the role of this is in real life. Access time is not influenced, it will on average take half a turn of the drive till the heads can access the data, and to me it looks like typical stuff a linux system does is reading many not so large files, cluttered around in the file system. But that's my guess only. And I understand that you like to optimize stuff - I like to do this too. But sometimes I think that the potential benefit might not be so large, compared to the time I spend moving data around to the ideal place, or the time I would need to spend thinking about how to tune things. Or the time you need to fix a problem that you know was working in the old system, but this is gone now and you cannot have a quick look at it, or just boot into it. You lose the opportunity to start your old system in order to compare the times of your big renderings. And maybe at one point you need to create some true 32bit applications? Happened to me. So I just chroot into my old system and build there. Oh, and you mentioned databases. Yes, mysql stores itsa data in machine- depenent form. You will need to dump the data and re-import it in the new system. You will be happy to still have the 32bit system in such a case :) Wonko
Re: [gentoo-user] Re: sudo in kernel config ?
Apparently, though unproven, at 23:01 on Saturday 11 September 2010, Nikos Chantziaras did opine thusly: On 09/11/2010 11:49 PM, Dale wrote: Nikos Chantziaras wrote: On 09/11/2010 11:35 PM, Dale wrote: Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. Glad only root can get to the sources. ;-) No, any user can't edit them; only the user you assign the files to. If you assign them to root, only root can edit them. If you assign them to kerneluser, only kerneluser can edit them. This is Unix 101 :) My point was, if the sources are say in the user group, then any user can edit them? Right now, they are in the root group and owned my root which for security reasons is a good idea. That way a regular user can't edit or modify the kernel sources. The group can only write if the files have the group write permission set. Still in Unix 101 domain, hehe :) And you need write permission on the containing directory to create new files or delete existing ones. Nothing to do with the permissions on the file itself. With this, I have moved us on to Unix 101a :-) -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Re: sudo in kernel config ?
Nikos Chantziaras wrote: On 09/11/2010 11:49 PM, Dale wrote: Nikos Chantziaras wrote: On 09/11/2010 11:35 PM, Dale wrote: Alan McKinnon wrote: Apparently, though unproven, at 11:46 on Saturday 11 September 2010, Albert Hopkins did opine thusly: On Sat, 2010-09-11 at 10:24 +0200, Stéphane Guedon wrote: few months ago, I read linux kernel in a nutschell(sic), and the author wrote we shouldn't do kernel operations (config and build) as root. I call bullsh*t. I've been compiling kernels for 17 years and for the most part have done it as root without any problems. Same here. The root user (sometimes portage) creates /usr/src/linux-* Someone tell me again exactly how user alan is supposed to build those sources? If they are accessible by a user, couldn't a user then edit or add something that would then cause a security problem? If they can edit them and no one know it, then root comes along and builds a shiney new kernel with a really nice security hole. Glad only root can get to the sources. ;-) No, any user can't edit them; only the user you assign the files to. If you assign them to root, only root can edit them. If you assign them to kerneluser, only kerneluser can edit them. This is Unix 101 :) My point was, if the sources are say in the user group, then any user can edit them? Right now, they are in the root group and owned my root which for security reasons is a good idea. That way a regular user can't edit or modify the kernel sources. The group can only write if the files have the group write permission set. Still in Unix 101 domain, hehe :) I know that. Why would a person want anyone BUT root to be able to access and change the kernel sources? Lets see if asking it this way makes more sense. lol Dale :-) :-)
Re: [gentoo-user] undetected DVD r/w device
On Saturday 11 September 2010 21:15:12 Alan McKinnon wrote: Apparently, though unproven, at 20:24 on Saturday 11 September 2010, Peter Humphrey did opine thus: On Saturday 11 September 2010 18:50:04 Mick wrote: Deselect that which is deprecated and will be removed soon I hope it won't. I have one box whose drives the new drivers cannot detect - at least, nothing I've tried has worked so far. That's the beauty of Linux. When Linus decides he's had enough keeping old stuff hanging around and rips it out, you just put it right back in your own copies. Hmm. Not sure how I'd do that. I'd just assumed that I'd stick with a kernel version that worked. You know: mask anything later than version 99.94. And yes, I do remember that version, but in the 1.n.n series. -- Rgds Peter. Linux Counter 5290, 1994-04-23.
Re: [gentoo-user] sudo in kernel config ?
On Saturday 11 September 2010 21:28:13 Etaoin Shrdlu wrote: This was actually a potential risk once upon a time: Sorry to drift from the topic, but would somebody please explain to me what a potential risk is? How does it differ from a risk? (Not getting at you, Etaoin; the world is just full of woolly thinking that threatens to submerge us all. Or not thinking, in most cases.) -- Rgds Peter. Linux Counter 5290, 1994-04-23.
Re: [gentoo-user] sudo in kernel config ?
On Sat, 11 Sep 2010 23:05:22 +0100 Peter Humphrey pe...@humphrey.ukfsn.org wrote: On Saturday 11 September 2010 21:28:13 Etaoin Shrdlu wrote: This was actually a potential risk once upon a time: Sorry to drift from the topic, but would somebody please explain to me what a potential risk is? How does it differ from a risk? (Not getting at you, Etaoin; the world is just full of woolly thinking that threatens to submerge us all. Or not thinking, in most cases.) I suppose that a risk is potential because it's possible that it's, um risky only under certain circumstances. If those circumstances are not true for you, there is no risk; if they are true, there is a risk. Once you know that there is a risk (thus it's no longer potential, but it's actual), it still take somebody or something to exploit it to actually have a problem. Makes sense?
Re: [gentoo-user] Properly handling missing files when downloading files from ftp:// via ISA proxy (emerge/wget)
On Sat, Sep 11, 2010 at 5:44 PM, Mick michaelkintz...@gmail.com wrote: On Friday 10 September 2010 17:13:44 Maciej Grela wrote: 2010/9/10 Paul Hartman paul.hartman+gen...@gmail.compaul.hartman%2bgen...@gmail.com : On Fri, Sep 10, 2010 at 9:43 AM, Maciej Grela maciej.gr...@gmail.com wrote: Hi, Is there any way to make emerge (wget) correctly behave when it tries to download a non-existing file from FTP in a network using ISA as ftp_proxy ? I have one of these at work and it's really annoying because of situations If you are allowed to bypass the proxy, try --use-proxy=off in your wget command line. Unfortunately proxy is the only way to access the Internet. It seems your FTP is proxied over HTTP and your client needs to support this. I'm not sure if wget supports HTTP proxies for FTP. Here is a document from Microsoft about configuring ISA and various clients: http://technet.microsoft.com/en-us/library/bb794745.aspx I'll read that, thanks. I think that you can try exporting your proxy's http address to ftp like so: # export ftp_proxy=http://my_proxy.com:1234; and then run emerge to see if you can get through. If your proxy requires a username/passwd you'll need to add these on the command line just as the handbook advises. -- Regards, Mick Hi You might find the problem is NTLM authentication. I have found net-proxy/*ntlmaps **to work around this issue.* * * *Kind regards* * * *Brett Freer* *www.rhapsody.com.au* * *
Re: [gentoo-user] undetected DVD r/w device
On Saturday 11 September 2010 22:57:12 Peter Humphrey wrote: On Saturday 11 September 2010 21:15:12 Alan McKinnon wrote: Apparently, though unproven, at 20:24 on Saturday 11 September 2010, Peter Humphrey did opine thus: On Saturday 11 September 2010 18:50:04 Mick wrote: Deselect that which is deprecated and will be removed soon I hope it won't. I have one box whose drives the new drivers cannot detect - at least, nothing I've tried has worked so far. That's the beauty of Linux. When Linus decides he's had enough keeping old stuff hanging around and rips it out, you just put it right back in your own copies. Hmm. Not sure how I'd do that. I'd just assumed that I'd stick with a kernel version that worked. You know: mask anything later than version 99.94. And yes, I do remember that version, but in the 1.n.n series. Is it possible to create one's own patch for drivers from old kernel(s) and use that to patch later versions?! O_O -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] sudo in kernel config ?
On Saturday 11 September 2010 23:03:14 Etaoin Shrdlu wrote: Makes sense? Not convinced. Sorry. -- Rgds Peter. Linux Counter 5290, 1994-04-23.
Re: [gentoo-user] Re: sudo in kernel config ?
Apparently, though unproven, at 23:47 on Saturday 11 September 2010, Dale did opine thusly: My point was, if the sources are say in the user group, then any user can edit them? Right now, they are in the root group and owned my root which for security reasons is a good idea. That way a regular user can't edit or modify the kernel sources. The group can only write if the files have the group write permission set. Still in Unix 101 domain, hehe :) I know that. Why would a person want anyone BUT root to be able to access and change the kernel sources? Lets see if asking it this way makes more sense. lol Gentoo does things different. If you read Documentation/* in the kernel sources, you will not find there what Gentoo has. /usr/src/linux was intended by the kernel devs[1] to be where the system headers are stored - what glibc uses to build. Like everything else in /usr/ this is obviously writeable for root only (usually). The intent is that you download kernel sources to ~, build there and sudo make install. Gentoo needs a kernel tree (not just headers) for all manner of stuff to build against. These days many distros also do it this way to accommodate the needs of getting nvidia-drivers and vm products to build their drivers etc. This must obviously also be writeable only for root. So, the ancient advice about not building as root is bullshit. It might have been good advice once but like all advice it's time is past. To answer your question: You wouldn't. Anything else is just daft. [1] this itself might be ancient cruft and hopelessly out of date -- alan dot mckinnon at gmail dot com
[gentoo-user] Re: [OT] sudo in kernel config ?
On Sun, 12 Sep 2010 00:06:04 +0100 Peter Humphrey pe...@humphrey.ukfsn.org wrote: On Saturday 11 September 2010 23:03:14 Etaoin Shrdlu wrote: Makes sense? Not convinced. Sorry. The Merriam-Webster gives this definition of potential: existing in possibility : capable of development into actuality which is exactly what I meant in my original post. I linked a document (if you read that), and the exploit described there could happen only if one installed the kernel sources downloaded from kernel.org. Hence the potential in the above meaning: if one did not use those sources, there was no risk for that specific exploit. But since you're not convinced, now it would be nice, for my own education, and perhaps someone else's, that you elaborated a bit more. What exactly do you find non convincing in that usage of the adjective? How would you express the concept better?
[gentoo-user] Re: static-libs
On 09/11/2010 01:52 PM, walt wrote: I don't know how portage chooses between static and dynamic libs while building a package. Aha! Grepping through /usr/portage/eclass/* for 'static' taught me something: xorg-2.eclass: myopts+= $(use_enable static-libs static) So there is a static useflag in addition to static-libs, which I'd never noticed before. I haven't actually tried the experiment yet, but I'm speculating that the difference between 'static' and 'static-libs' is something like this: Many packages install libraries so that other packages can use them. The 'static-libs' useflag tells a package to build and install the static version of its own libraries *in addition to* the dynamic ones. The 'static' useflag tells portage to build a package and link it against the static versions of libraries that were installed by *other* packages. Quiz for sober people: (That excludes moi ;) What happens if portage builds a 'static' package that links against a library that was installed without the 'static-libs' useflag being set?
[gentoo-user] Re: undetected DVD r/w device
On 09/11/2010 11:24 AM, Peter Humphrey wrote: On Saturday 11 September 2010 18:50:04 Mick wrote: Deselect that which is deprecated and will be removed soon I hope it won't. I have one box whose drives the new drivers cannot detect - at least, nothing I've tried has worked so far. That seems very odd. Is there something exotic about those drives? The new driver sees the controller but the controller doesn't see the drives? Or, if the kernel doesn't see the controller then maybe consider adding the SATA_AHCI driver as a test. I'm still a bit confused when trying to figure out which controller uses which driver, but I've been tripped up with that particular driver before.
Re: [gentoo-user] Re: static-libs
What happens if portage builds a 'static' package that links against a library that was installed without the 'static-libs' useflag being set? http://tldp.org/HOWTO/Program-Library-HOWTO/introduction.html writes: DL libraries aren't really a different kind of library format (both static and shared libraries can be used as DL libraries); If I understand this at all, the term static library is a litle misleading. * Libraries would be shared (in memory) and unshared. * Executables would link static or dynamic (against both types). The term static library would refer to unshared libraries, because are typically linked statically (but must not). Please correct me if I am wrong. I am new in the field. Al
Re: [gentoo-user] Moving / around...
Alex Schuster wo...@wonkology.org [10-09-12 04:13]: meino.cra...@gmx.de writes: I think there is some misunderstanding: Before migration to 64bit: /dev/sda3 is mounted on / and contains the 32bit Gentoo /dev/sda10 is mounted on /home/mcc/migration and will contain the stuff of the 64bit Gentoo After migration I will *not* mount /dev/sda10 on / but will clear all stuff from /dev/sda3 and move the contents from /dev/sda10 to /dev/sda3. Is still valid what you said under this premissions, Wonko? That's how I understood it, although I assumed the temproary 64bit install would be on a 2nd drive, thus you would copy it back once it seems to work. No, I see no problem with this. About performance: I'm not sure it will be even noticeable. Yes, most drives (but not all) are organized so the first partitions go to the outside, which is faster. With LVM, I used to create two volume groups on my drive, a group for swap and the system, and another one for data. But then I thought it's not worth the effort, and I lose some of the LVM benefits. Well, with everything encrypted I don't get full performance anyway, so my case might be a little different. But the performance increase is only true when reading lots of data. I'm not sure how big the role of this is in real life. Access time is not influenced, it will on average take half a turn of the drive till the heads can access the data, and to me it looks like typical stuff a linux system does is reading many not so large files, cluttered around in the file system. But that's my guess only. And I understand that you like to optimize stuff - I like to do this too. But sometimes I think that the potential benefit might not be so large, compared to the time I spend moving data around to the ideal place, or the time I would need to spend thinking about how to tune things. Or the time you need to fix a problem that you know was working in the old system, but this is gone now and you cannot have a quick look at it, or just boot into it. You lose the opportunity to start your old system in order to compare the times of your big renderings. And maybe at one point you need to create some true 32bit applications? Happened to me. So I just chroot into my old system and build there. Oh, and you mentioned databases. Yes, mysql stores itsa data in machine- depenent form. You will need to dump the data and re-import it in the new system. You will be happy to still have the 32bit system in such a case :) Wonko I also cannot evaluate the real impact the position of the /-partition on the harddisk has on system performance. I read about it years ago and since than I always put the partitions always in the sequence of boot,swap,root,home onto the harddisks. May be its only a tradition nowadays... ;) Do you know of any other kind of data beside databses, which may be machinedependant or cause trouble while migrating to 64bit? Best regards, mcc