Re: New problem with an hp CD5750
Le lundi 02 avril 2007 à 09:59 -0400, Lennart Sorensen a écrit : > On Sat, Mar 31, 2007 at 12:35:01PM +0200, [EMAIL PROTECTED] wrote: > > I have always some problems with the HP dc5750. After installing > > "wildly" the system (by copying a Knoppix 32 bits CD on the HD). > > The system runs, slowly..., even X and KDE are OK, but I encounter a new > > problem. > > Ths disk is deadly slow : > > hdparm -t gives 1.43 Mo/s > > on an 23 years old Amiga 2000 running Linux I get 3.4 Mo/s > > so 1.43 Mo/s is quite a pain. > > I want to compile a new kernel with only the needed modules , eg a > > better disk driver, and no unused modules. The problem is that I can't > > get a working kernel on the machine, the kernel does not boot with a > > panic "does not find the disk ... append a correct root=... ". > > I try with a 2.6.20 kernel, take the 2.6.19 config found in the Knoppix > > install and a make oldconfig. The kernel compile is OK but panics while > > booting not finding the boot disk. > > > > There is something which does not work properly with the disk > > driver ... > > Any help will be appreciated. > > Usually less than 2MB/s means you don't have DMA enabled, which usually > means you are using the wrong ide driver or have a bad configuration > somewhere. You can check the settings in /proc/ide/hda/settings or > whatever your HD is. > > As for installing knoppix that way, I have no idea. knoppix is a > tolerable live cd but I would hate to have it as my normal use system. Hello, Somebody on the list give me an advice to switch (in the BIOS) IDE emulation from "Legacy IDE" to "Native" I noticed that when I select "Legacy IDE" the access turns to UDMA 100 with no other choice. When I try it I get a kernel panic at boot ... not syncing ... please append a correct "root=". So it is a misfit between the drivers and the system or with the way the SATA drivers works. Are the SATA hd devices always called HDx with the SATA driver working or are they called SDx like the Scsi hd ? On one of my home systems I lost an SCSI disk and replace it by a standard IDE one and the disk is seen as SDE after the SCSI disks (kernel 2.6.20). Knoppix is Debian based, so the system seems to be OK, and I update it with apt-get set to etch. I recompile the last 2.6.19 with SATA set to "Off" and the machine boots (Legacy IDE) with the same bad disk performance. A simple "vi" on a file is long as on the venerable Amiga which I used as reference. (Yes an Amiga 2000 works with Linux, this one has known its firts kernel before the 2.0 and runs with a Debian since 2.0.x, I think x was 10 to 12 many years ago, it runs now 2.6.18 with full success). On my "big" (2x AMD 2Ghz) system the same "hdparm -t /dev/sde1" gives 66.46 Mo/s on the IDE disk. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: xen on amd64 on dual core amd box
On Mon, Apr 02, 2007 at 10:01:49AM -0400, Lennart Sorensen wrote: > On Mon, Apr 02, 2007 at 08:17:34AM +1000, Alex Samad wrote: > > nope 2.6.18 > > debxen:~# COLUMNS=150 dpkg -l '*xen*' | grep ^ii > > ii linux-image-2.6-xen-amd64 2.6.18+6Linux > > kernel 2.6 image on AMD64 > > ii linux-image-2.6.18-4-xen-amd64 2.6.18.dfsg.1-12Linux > > 2.6.18 image on AMD64 > > ii linux-modules-2.6.18-4-xen-amd6 2.6.18.dfsg.1-12Linux > > 2.6.18 modules on AMD64 > > ii xen-hypervisor-3.0.3-1-amd643.0.3-0-2 The Xen > > Hypervisor on AMD64 > > ii xen-linux-system-2.6.18-4-xen-a 2.6.18.dfsg.1-12XEN > > system > > with Linux 2.6.18 image on AMD64 > > ii xen-tools 3.1-1 Tools to > > manage debian XEN virtual servers > > ii xen-utils-3.0.3-1 3.0.3-0-2 XEN > > administrative tools > > ii xen-utils-common3.0.3-0-2 XEN > > administrative tools - common files > > > > > > > > Apr 2 06:48:15 debxen kernel: Timer ISR/3: Time went backwards: > > delta=-63867412 delta_cpu=1824132588 shadow=536987137206 off=781032024 > > processed=53783200 cpu_processed=53594400 > > Apr 2 06:48:15 debxen kernel: 0: 53783200 > > Apr 2 06:48:15 debxen kernel: 1: 53588400 > > Apr 2 06:48:15 debxen kernel: 2: 53775200 > > Apr 2 06:48:15 debxen kernel: 3: 53594400 > > Apr 2 06:48:21 debxen kernel: printk: 7 messages suppressed. > > I wonder if that kernel is before they dealt with the TSC being unsyched > on dual core athlon 64's. I remember quite a bit of discussions and > patches happening some months ago about TSC support on dual core > systems. strange same box running the stock debian kernel doesn't exhibit any problems! > > -- > Len Sorensen > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > signature.asc Description: Digital signature
Re: Problems booting from SATA in a HP DC5750 (ATI SB600 ?)
On Mon, Apr 02, 2007 at 05:19:05PM -0300, Joao Carlos de Lima Roscoe wrote: > Dear Srs, > > I've just suceeded performing the base install from an etch snapshot CD1, > in a HP DC5750, which uses an ATI SB600 to provide SATA. > I had to choose "Native IDE" in the bios settings do install properly > (other options are "Legacy IDE" and "SATA" - not the case, I have only > one HD). > > The install went well - no glitch, and grub was put on MBR, as usual. > > However, during boot, the machine reports a lot of errors: > > * > ata2: SATA Link up 1.5 Gbps (SStatus 113 SControl 300) > ata2.00: Configured for UDMA/100 > ata2: EH Complete > ata2.00 exception Emask 0x40 SAct 0x0 SErr 0x800 action 0x2 > ata2.00 (irq_stat 0x4001) > ata2.00 tag 0 cmd 0xa0 Emask 0x40 stat 0x51 err 0x20 (internal error) > softresetting port > * > > The pattern above is repeated 20 or so times, then boot proceeds and > succeeds!!! > Once up, the box runs ok, as far as I could evaluate. > > I've been googling around but could not find anything useful so far. > > Has anyone seen something like this? Any hints? > Thank you in advance, Well it's an ATI chipset, so well, not surprising. There was some recent patches on the kernel mailing list dealing with SATA issues on the SB600. I don't remember if they are in 2.6.20 or will go in 2.6.21. Usually having to set the bios to anything other than native SATA/enhanced/whatever they call the setting that does no emulation, is not a good sign of having the right working drivers. I think the debian kernel team has experimental 2.6.20 kernel packages that you could try to see if it is any better. -- Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Problems booting from SATA in a HP DC5750 (ATI SB600 ?)
Dear Srs, I've just suceeded performing the base install from an etch snapshot CD1, in a HP DC5750, which uses an ATI SB600 to provide SATA. I had to choose "Native IDE" in the bios settings do install properly (other options are "Legacy IDE" and "SATA" - not the case, I have only one HD). The install went well - no glitch, and grub was put on MBR, as usual. However, during boot, the machine reports a lot of errors: * ata2: SATA Link up 1.5 Gbps (SStatus 113 SControl 300) ata2.00: Configured for UDMA/100 ata2: EH Complete ata2.00 exception Emask 0x40 SAct 0x0 SErr 0x800 action 0x2 ata2.00 (irq_stat 0x4001) ata2.00 tag 0 cmd 0xa0 Emask 0x40 stat 0x51 err 0x20 (internal error) softresetting port * The pattern above is repeated 20 or so times, then boot proceeds and succeeds!!! Once up, the box runs ok, as far as I could evaluate. I've been googling around but could not find anything useful so far. Has anyone seen something like this? Any hints? Thank you in advance, Joao -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Weired things on Dual-Core AMD64
On Mon, Apr 02, 2007 at 06:33:21PM +0200, Hans-J. Ullrich wrote: > At the moment I am running the stock kernel. It is 2.6.18-4-amd64. A > selfbuilt > kernel crashed sometimes at bootup. I changed nothing but kernel-timer from > 250Hz to 1000Hz. This might be the problem (there was another post relating > to this behaviour by someone else.) The stock kernel is running fine. Well I have a realtek card that I am playing around with. I have managed to connect and such but last I played with it the speed dropped to something awful not long after connecting. There have been driver updates since though so I should try again. > Most of the users use ndiswrapper, but I am not a friend of it. Formerly I > had > broadcom, which ran fine with native drivers. The support by the vendors of > wireless cards is sucking ! Meanwhile I checked a little bit more and > discovered the problem either at madwifi-tools or the drivers itself. It > might be an initialisation problem. It is fancy: Doing /etc/init.d/networking > twice ore three times does the trick. The madwifi.driver does not see the AP > each time, so it does not get the dhcpleases. Yeah I think I will stay away from ndiswrapper. > I never had problems with it so far. Maybe it is on Intel-cpus , I always use > AMD. CPU doesn't matter. Video chip and X driver does. I buy whichever CPU is currently best (which has been Athlon and Athlon 64 for a long time, but is now the Core 2 Duo). > BTW: I managed to get a new Brother printer (printer+scanner) running in > debian-amd64. It is DCP-115c aka MFC-210C (including some others , too). To > do this, I had to rebuilt a redhat-package for 64-bit and add an entry > in /etc/udev/libsane. It was a little tricky. I could write a little how-to > for debian-amd64. There was none in the web, only for I386. The printerpart I > could not get running on debian-amd64, but I could not find out, why. Maybe > it is a simple problem of rights or the driver is incompatible (it is a > binary) with 64-bit. > > So, please leave me a note, if this howto is interesting and someone likes to > make it popular for the community. Not sure. I don't liek all in pnes personally, and I despise brother products (past experience with their laser printers was too painful, not to mention their idiotic fax machines). I am tempted to buy a xerox phaser 6120N since it is a nice colour laser, with good print quality, native postscript level 3, and ethernet connection for under $400. -- Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Weired things on Dual-Core AMD64
Am Montag, 2. April 2007 15:54 schrieb Lennart Sorensen: > On Sat, Mar 31, 2007 at 03:49:56PM +0200, Hans-J. Ullrich wrote: > > I just got a brandnew notebook with AMD64 X2 cpu. Formerly I had a single > > core AMD64 cpu. Running the new one, I dicsovered some new things. Maybe > > you can solve this or even explain it. > > > > 1. The easy thing: On websides, animated gifs or pictures are running > > twice as fast as they should. Maybe this is a konqueror related problem. > > Does someone discover this, too ? > > There have been problems with some chipsets (mainly ATI, but also a few > nvidia ones) where the system time would run at double speed. I believe > it was caused by an irq routing issue (acpi/bios bugs mainly it seems) > which caused the timer irq to be received twice, which made time move at > double speed. Newer kernels seem to have fixed that. Which kernel did > you install? At the moment I am running the stock kernel. It is 2.6.18-4-amd64. A selfbuilt kernel crashed sometimes at bootup. I changed nothing but kernel-timer from 250Hz to 1000Hz. This might be the problem (there was another post relating to this behaviour by someone else.) The stock kernel is running fine. > > > 2. The one, I am looking for a week now: It is not possible, to get my > > Atheros AR5000G card running at startup. The card does never see the > > accesspoint at startup. Strange: When the system is up, makling a > > "/etc/init.d/network restart" let it work. Where is the difference to the > > startup sequence ? I could not find. Any help is preceated. > > No idea. I still haven't had much luck getting wireless to work with > linux. Most of the users use ndiswrapper, but I am not a friend of it. Formerly I had broadcom, which ran fine with native drivers. The support by the vendors of wireless cards is sucking ! Meanwhile I checked a little bit more and discovered the problem either at madwifi-tools or the drivers itself. It might be an initialisation problem. It is fancy: Doing /etc/init.d/networking twice ore three times does the trick. The madwifi.driver does not see the AP each time, so it does not get the dhcpleases. > > > 3. Last there is a wish: Is it possible, to activate the logo at > > kernel-start in the stock kernels in the future ? I like those TWO > > penguins at boot ! Yes, natuarally I could build my own kernel, but it > > would be nice, if they would be in the stock kernel, too. I do not see > > any technical problems with it, or are there any ? > > I find framebuffer drivers screw up so many X drivers, that they aren't > worth the hassle. And I hardly ever reboot (I guess laptops might be > rebooted more often than a desktop). > I never had problems with it so far. Maybe it is on Intel-cpus , I always use AMD. > > 4. Just a general question, not amd64-related: module-assistant makes a > > link to the kernel-headers, not to the kernel-sources itself. I do not > > quite understand the differnce between the kernel-headers and the > > kernel-sources. Can someone explain the difference between the two ones > > for me, and what policy Debian is running ? > > The kernel headers package contains the configuration and header files > matching your current kernel. The kernel source contains unconfigured > sources (which include headers). The configuration part is rather > important when building modules to match a kernel. And when building > other things you only ever need the headers, since the source files are > only used for building the kernel. Hence the need for kernel-headers > package and no need for kernel sources. > Ah, that does it explain. Kernel-headers are need to build kernel-modules, kernel-sources are need to build the kernel itself. > After all you don't need the sources to libc to compile a c program, > just the standard c library header files. > > -- > Len Sorensen BTW: I managed to get a new Brother printer (printer+scanner) running in debian-amd64. It is DCP-115c aka MFC-210C (including some others , too). To do this, I had to rebuilt a redhat-package for 64-bit and add an entry in /etc/udev/libsane. It was a little tricky. I could write a little how-to for debian-amd64. There was none in the web, only for I386. The printerpart I could not get running on debian-amd64, but I could not find out, why. Maybe it is a simple problem of rights or the driver is incompatible (it is a binary) with 64-bit. So, please leave me a note, if this howto is interesting and someone likes to make it popular for the community. Best regards Hans -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New problem with an hp CD5750
On Mon, Apr 02, 2007 at 10:08:16AM -0400, [EMAIL PROTECTED] wrote: > On Sat, Mar 31, 2007 at 04:11:00PM +0200, Hans-J. Ullrich wrote: > > Am Samstag, 31. M?rz 2007 12:35 schrieb [EMAIL PROTECTED]: > > > hdparm -t gives 1.43 Mo/s > > > on an 23 years old Amiga 2000 running Linux I get 3.4 Mo/s > > > so 1.43 Mo/s is quite a pain. > > I didn't know an Amiga 2000 *could* run Linux. Doesn't Linux need > memory-mapping hardware? Or did Commodore sneak that in without telling > me? Many people had 68030 or 68040 accelerator cards, many of which had real CPUs rather tahn EC versions. -- Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New problem with an hp CD5750
On Sat, Mar 31, 2007 at 04:11:00PM +0200, Hans-J. Ullrich wrote: > Am Samstag, 31. März 2007 12:35 schrieb [EMAIL PROTECTED]: > > hdparm -t gives 1.43 Mo/s > > on an 23 years old Amiga 2000 running Linux I get 3.4 Mo/s > > so 1.43 Mo/s is quite a pain. I didn't know an Amiga 2000 *could* run Linux. Doesn't Linux need memory-mapping hardware? Or did Commodore sneak that in without telling me? -- hendrik -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: xen on amd64 on dual core amd box
On Mon, Apr 02, 2007 at 08:17:34AM +1000, Alex Samad wrote: > nope 2.6.18 > debxen:~# COLUMNS=150 dpkg -l '*xen*' | grep ^ii > ii linux-image-2.6-xen-amd64 2.6.18+6Linux > kernel 2.6 image on AMD64 > ii linux-image-2.6.18-4-xen-amd64 2.6.18.dfsg.1-12Linux > 2.6.18 image on AMD64 > ii linux-modules-2.6.18-4-xen-amd6 2.6.18.dfsg.1-12Linux > 2.6.18 modules on AMD64 > ii xen-hypervisor-3.0.3-1-amd643.0.3-0-2 The Xen > Hypervisor on AMD64 > ii xen-linux-system-2.6.18-4-xen-a 2.6.18.dfsg.1-12XEN system > with Linux 2.6.18 image on AMD64 > ii xen-tools 3.1-1 Tools to > manage debian XEN virtual servers > ii xen-utils-3.0.3-1 3.0.3-0-2 XEN > administrative tools > ii xen-utils-common3.0.3-0-2 XEN > administrative tools - common files > > > > Apr 2 06:48:15 debxen kernel: Timer ISR/3: Time went backwards: > delta=-63867412 delta_cpu=1824132588 shadow=536987137206 off=781032024 > processed=53783200 cpu_processed=53594400 > Apr 2 06:48:15 debxen kernel: 0: 53783200 > Apr 2 06:48:15 debxen kernel: 1: 53588400 > Apr 2 06:48:15 debxen kernel: 2: 53775200 > Apr 2 06:48:15 debxen kernel: 3: 53594400 > Apr 2 06:48:21 debxen kernel: printk: 7 messages suppressed. I wonder if that kernel is before they dealt with the TSC being unsyched on dual core athlon 64's. I remember quite a bit of discussions and patches happening some months ago about TSC support on dual core systems. -- Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New problem with an hp CD5750
On Sat, Mar 31, 2007 at 12:35:01PM +0200, [EMAIL PROTECTED] wrote: > I have always some problems with the HP dc5750. After installing > "wildly" the system (by copying a Knoppix 32 bits CD on the HD). > The system runs, slowly..., even X and KDE are OK, but I encounter a new > problem. > Ths disk is deadly slow : > hdparm -t gives 1.43 Mo/s > on an 23 years old Amiga 2000 running Linux I get 3.4 Mo/s > so 1.43 Mo/s is quite a pain. > I want to compile a new kernel with only the needed modules , eg a > better disk driver, and no unused modules. The problem is that I can't > get a working kernel on the machine, the kernel does not boot with a > panic "does not find the disk ... append a correct root=... ". > I try with a 2.6.20 kernel, take the 2.6.19 config found in the Knoppix > install and a make oldconfig. The kernel compile is OK but panics while > booting not finding the boot disk. > > There is something which does not work properly with the disk > driver ... > Any help will be appreciated. Usually less than 2MB/s means you don't have DMA enabled, which usually means you are using the wrong ide driver or have a bad configuration somewhere. You can check the settings in /proc/ide/hda/settings or whatever your HD is. As for installing knoppix that way, I have no idea. knoppix is a tolerable live cd but I would hate to have it as my normal use system. -- Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Weired things on Dual-Core AMD64
On Sat, Mar 31, 2007 at 03:49:56PM +0200, Hans-J. Ullrich wrote: > I just got a brandnew notebook with AMD64 X2 cpu. Formerly I had a single > core > AMD64 cpu. Running the new one, I dicsovered some new things. Maybe you can > solve this or even explain it. > > 1. The easy thing: On websides, animated gifs or pictures are running twice > as > fast as they should. Maybe this is a konqueror related problem. Does someone > discover this, too ? There have been problems with some chipsets (mainly ATI, but also a few nvidia ones) where the system time would run at double speed. I believe it was caused by an irq routing issue (acpi/bios bugs mainly it seems) which caused the timer irq to be received twice, which made time move at double speed. Newer kernels seem to have fixed that. Which kernel did you install? > 2. The one, I am looking for a week now: It is not possible, to get my > Atheros > AR5000G card running at startup. The card does never see the accesspoint at > startup. Strange: When the system is up, makling a "/etc/init.d/network > restart" let it work. Where is the difference to the startup sequence ? I > could not find. Any help is preceated. No idea. I still haven't had much luck getting wireless to work with linux. > 3. Last there is a wish: Is it possible, to activate the logo at kernel-start > in the stock kernels in the future ? I like those TWO penguins at boot ! > Yes, natuarally I could build my own kernel, but it would be nice, if they > would be in the stock kernel, too. I do not see any technical problems with > it, or are there any ? I find framebuffer drivers screw up so many X drivers, that they aren't worth the hassle. And I hardly ever reboot (I guess laptops might be rebooted more often than a desktop). > 4. Just a general question, not amd64-related: module-assistant makes a link > to the kernel-headers, not to the kernel-sources itself. I do not quite > understand the differnce between the kernel-headers and the kernel-sources. > Can someone explain the difference between the two ones for me, and what > policy Debian is running ? The kernel headers package contains the configuration and header files matching your current kernel. The kernel source contains unconfigured sources (which include headers). The configuration part is rather important when building modules to match a kernel. And when building other things you only ever need the headers, since the source files are only used for building the kernel. Hence the need for kernel-headers package and no need for kernel sources. After all you don't need the sources to libc to compile a c program, just the standard c library header files. -- Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: gcc-4.1
Hi Rob And thanks for the answer. > >Has anyone noticed that gcc is older in experimental for amd64 than > > for other architectures. It seems to me like dependencies that go in > > circles. If so, does anyone have an explanation? I wanted to compile the > > newest version of openoffice but I need this version of gcc for that. > > The packages in experimental don't seem to build for all architectures - > another example is gnome-applets, which has been built for i386 only. > > Sometimes it is because of failure to build on that architecture. Sometimes > it just doesn't build. > > Look at the per-package build logs at buildd.debian.org for more > information. I can in fact wait a bit for the gcc but the page you pointed out is interesting and I should have known. But I could not find the experimental build logs in there. I play with compiling openoffice myself because I sometimes have to wait months for the official amd64 debian compiler to do it (I'm not complaining, thanks for the job you do if you read it :) and I want to have the newest version to check its compatibility with MS-Word myself. Regards Gudjon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]