Re: APIC priorities, can they be changed?
linux-os (Dick Johnson) wrote: On Thu, 15 Feb 2007, Mike Panetta wrote: > I am not on the list (corperate email sucks) so please CC any replies to > me. Thanks. [snip] > I have seen the preempt patches, but they touch a lot of files, and we > have gone through testing with the 2.6.16.19 kernel and do not wish to > change, and we cannot find a version of the patch for this kernel. Is > there something smaller/simpler I can do? My understanding is the > priority in the APIC is set in software via the interrupt vector number > (higher numbers have lower priority) is this true? If so, how hard > would it be for me to just change the vector numbers around? > [snip] > Thanks, > Mike Can't you muck with the BIOS settings? That's where the primary hardware gets 'connected'. I did try that. The BIOS only allows me to either allocate an IRQ to be a PCI interrupt, or reserve it (for what I have no idea). The IRQ's listed in the BIOS are also different from the ones Linux sees. I think the BIOS is seeing the XT-PIC IRQ numbers and Linux is seeing the APIC numbers. For example the little bios blurb that prints before the system boots says the USB controller I am interested in is assigned/using IRQ 10, Linux sees it using IRQ 18. I have found that I can keep Linux from using the APIC by disabling it with a kernel command line switch, but that does not help, it just makes Linux use the XT-PIC instead of the IO-APIC to do IRQ routing. So I guess I'm back to my original question of 'Would changing the vector numbers do what I want?' and if the answer is 'yes', how would I do it? Thanks, Mike Cheers, Dick Johnson Penguin : Linux version 2.6.16.24 on an i686 machine (5592.61 BogoMips). New book: http://www.AbominableFirebug.com/ _ The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> - and destroy all copies of this information, including any attachments, without reading or disclosing them. Thank you. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
APIC priorities, can they be changed?
I am not on the list (corperate email sucks) so please CC any replies to me. Thanks. I am working on a project that has run in to what seems to be an interrupt priority problem. We switched mainboards in our product and went from a system where the EHCI controller IRQ was of a fairly high priority to one where its the lowest in the system. Now we cannot get the data off our USB device in time and we drop alot of data. I was hoping someone here may have a simple(ish) solution to the problem. I have seen the preempt patches, but they touch a lot of files, and we have gone through testing with the 2.6.16.19 kernel and do not wish to change, and we cannot find a version of the patch for this kernel. Is there something smaller/simpler I can do? My understanding is the priority in the APIC is set in software via the interrupt vector number (higher numbers have lower priority) is this true? If so, how hard would it be for me to just change the vector numbers around? Ideally I would love to switch to the 2.6.20 kernel and use the preempt patch, but I doubt we have the time for that. Thanks, Mike - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
APIC priorities, can they be changed?
I am not on the list (corperate email sucks) so please CC any replies to me. Thanks. I am working on a project that has run in to what seems to be an interrupt priority problem. We switched mainboards in our product and went from a system where the EHCI controller IRQ was of a fairly high priority to one where its the lowest in the system. Now we cannot get the data off our USB device in time and we drop alot of data. I was hoping someone here may have a simple(ish) solution to the problem. I have seen the preempt patches, but they touch a lot of files, and we have gone through testing with the 2.6.16.19 kernel and do not wish to change, and we cannot find a version of the patch for this kernel. Is there something smaller/simpler I can do? My understanding is the priority in the APIC is set in software via the interrupt vector number (higher numbers have lower priority) is this true? If so, how hard would it be for me to just change the vector numbers around? Ideally I would love to switch to the 2.6.20 kernel and use the preempt patch, but I doubt we have the time for that. Thanks, Mike - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: APIC priorities, can they be changed?
linux-os (Dick Johnson) wrote: On Thu, 15 Feb 2007, Mike Panetta wrote: I am not on the list (corperate email sucks) so please CC any replies to me. Thanks. [snip] I have seen the preempt patches, but they touch a lot of files, and we have gone through testing with the 2.6.16.19 kernel and do not wish to change, and we cannot find a version of the patch for this kernel. Is there something smaller/simpler I can do? My understanding is the priority in the APIC is set in software via the interrupt vector number (higher numbers have lower priority) is this true? If so, how hard would it be for me to just change the vector numbers around? [snip] Thanks, Mike Can't you muck with the BIOS settings? That's where the primary hardware gets 'connected'. I did try that. The BIOS only allows me to either allocate an IRQ to be a PCI interrupt, or reserve it (for what I have no idea). The IRQ's listed in the BIOS are also different from the ones Linux sees. I think the BIOS is seeing the XT-PIC IRQ numbers and Linux is seeing the APIC numbers. For example the little bios blurb that prints before the system boots says the USB controller I am interested in is assigned/using IRQ 10, Linux sees it using IRQ 18. I have found that I can keep Linux from using the APIC by disabling it with a kernel command line switch, but that does not help, it just makes Linux use the XT-PIC instead of the IO-APIC to do IRQ routing. So I guess I'm back to my original question of 'Would changing the vector numbers do what I want?' and if the answer is 'yes', how would I do it? Thanks, Mike Cheers, Dick Johnson Penguin : Linux version 2.6.16.24 on an i686 machine (5592.61 BogoMips). New book: http://www.AbominableFirebug.com/ _ The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] - and destroy all copies of this information, including any attachments, without reading or disclosing them. Thank you. - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
2.4.4 + LFS what am I doing wrong?
I cannot for the life of me get LFS working. I have recompiled the glibc rpm (the 2.2.1 one) for i686 and installed it. I have recompiled fileutils against that glibc. It still tells me that the file is too large when I do a dd if=/dev/zero of=bigfile bs=8192 it stops at 2.0GB (ls -lh). What am I missing? I thought I had all the pieces in place. It works on another machine I have. I just don't understand why it does not work on this one... I have this gut feeling its something really stupid but for the life of me I cannot find out what. Thanks, Mike - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
2.4.4 + LFS what am I doing wrong?
I cannot for the life of me get LFS working. I have recompiled the glibc rpm (the 2.2.1 one) for i686 and installed it. I have recompiled fileutils against that glibc. It still tells me that the file is too large when I do a dd if=/dev/zero of=bigfile bs=8192 it stops at 2.0GB (ls -lh). What am I missing? I thought I had all the pieces in place. It works on another machine I have. I just don't understand why it does not work on this one... I have this gut feeling its something really stupid but for the life of me I cannot find out what. Thanks, Mike - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
ACPI broken in 2.4.4-ac6
ACPI seems to be broken on 2.4.4-ac6 or atleast poweroff is broken. During bootup all ACPI prints is that it was enabled, it used to (in plain jane 2.4.4) print the sleep levels supported by the bios but does not in ac6. What could be the cause? Thanks, Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
ACPI broken in 2.4.4-ac6
ACPI seems to be broken on 2.4.4-ac6 or atleast poweroff is broken. During bootup all ACPI prints is that it was enabled, it used to (in plain jane 2.4.4) print the sleep levels supported by the bios but does not in ac6. What could be the cause? Thanks, Mike -- - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
HPT366 IDE DMA error question.
What could cause this error? hdi: timeout waiting for DMA ide_dmaproc: chipset supported ide_dma_timeout func only: 14 hdi: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest } hdi: DMA disabled ide4: reset: success I get this message on all my off board HPT366 based controller cards. I am using these cards with seagate Barracuda ATA III Model ST320414A 20GB drives. Are there any known issues with these drives and the HPT366 based controllers? Are there any known issues with using 2-3 HPT366 cards in one system? There is only 1 drive per channel (2 per card). I am using this setup with Software RAID and needless to say no DMA=slow as hell. Just so you know the onboard IDE controller works fine. The drives report no errors on the onboard controllers and they have UDMA enabled (it was not disabled by the kernel). If there is any other info that is needed I will be glad to provide it. Thanks, Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
HPT366 IDE DMA error question.
What could cause this error? hdi: timeout waiting for DMA ide_dmaproc: chipset supported ide_dma_timeout func only: 14 hdi: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest } hdi: DMA disabled ide4: reset: success I get this message on all my off board HPT366 based controller cards. I am using these cards with seagate Barracuda ATA III Model ST320414A 20GB drives. Are there any known issues with these drives and the HPT366 based controllers? Are there any known issues with using 2-3 HPT366 cards in one system? There is only 1 drive per channel (2 per card). I am using this setup with Software RAID and needless to say no DMA=slow as hell. Just so you know the onboard IDE controller works fine. The drives report no errors on the onboard controllers and they have UDMA enabled (it was not disabled by the kernel). If there is any other info that is needed I will be glad to provide it. Thanks, Mike -- - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
kernel panic when using loop device on kernel 2.4.3
I have been getting kernel panics on kernel 2.4.3 when using the loop device on a rather regular basis. I get a kernel panic but no oops message. The kernel panic message says Kernel panic: invalid blocksize passed to set_blocksize. I saw that someone else on the list has had these problems as well but I have seen no response. Has a fix for this bug been posted that I have not seen? Is the fix in any current stable or AC kernel? As far as I know I am not using any strange options. The only thing I can think of would be that multiple people are using the loopback device at the same time. This is because the box that is having the problems is a build server for a custom software distro. We use the loopback filesystem to create the boot image. The box is a 4way xeon with 1GB of ram. Are there any known issues with loopback and SMP? Or even loopback and multiple mounts/usage? ver_linux output attached. Thanks, Mike -- If some fields are empty or look unusual you may have an old version. Compare to the current minimal requirements in Documentation/Changes. Linux optimus.applianceware.com 2.4.3 #2 SMP Wed Apr 18 06:00:29 PDT 2001 i686 unknown Gnu C 2.96 Gnu make 3.79.1 binutils 2.10.91.0.2 util-linux 2.10r modutils 2.4.2 e2fsprogs 1.19 reiserfsprogs 3.x.0b Linux C Library2.2.2 Dynamic linker (ldd) 2.2.2 Procps 2.0.7 Net-tools 1.57 Console-tools 0.3.3 Sh-utils 2.0 Modules Loaded
kernel panic when using loop device on kernel 2.4.3
I have been getting kernel panics on kernel 2.4.3 when using the loop device on a rather regular basis. I get a kernel panic but no oops message. The kernel panic message says Kernel panic: invalid blocksize passed to set_blocksize. I saw that someone else on the list has had these problems as well but I have seen no response. Has a fix for this bug been posted that I have not seen? Is the fix in any current stable or AC kernel? As far as I know I am not using any strange options. The only thing I can think of would be that multiple people are using the loopback device at the same time. This is because the box that is having the problems is a build server for a custom software distro. We use the loopback filesystem to create the boot image. The box is a 4way xeon with 1GB of ram. Are there any known issues with loopback and SMP? Or even loopback and multiple mounts/usage? ver_linux output attached. Thanks, Mike -- If some fields are empty or look unusual you may have an old version. Compare to the current minimal requirements in Documentation/Changes. Linux optimus.applianceware.com 2.4.3 #2 SMP Wed Apr 18 06:00:29 PDT 2001 i686 unknown Gnu C 2.96 Gnu make 3.79.1 binutils 2.10.91.0.2 util-linux 2.10r modutils 2.4.2 e2fsprogs 1.19 reiserfsprogs 3.x.0b Linux C Library2.2.2 Dynamic linker (ldd) 2.2.2 Procps 2.0.7 Net-tools 1.57 Console-tools 0.3.3 Sh-utils 2.0 Modules Loaded
kernel 2.4.2-ac14 in vmware - hangs
I am using VMware to test a linux install and since I have upgraded to 2.4.2-ac14 the VM locks up right after: calibrating APIC timer ... . CPU clock speed is 1152.4771 MHz. . host bus clock speed is 0. MHz. cpu: 0, clocks: 0, slice: 0 The last line is where it locks up... Its kind of suspicious that the clock speeds are wrong, The CPU speed should be 733 MHz (unless VMWare is using both my CPU's?) and I am almost 100% sure that the bus speed should not be 0. MHz! Can anyhone help me with this? Is it a bug on my part or what? Thanks, Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
kernel 2.4.2-ac14 in vmware - hangs
I am using VMware to test a linux install and since I have upgraded to 2.4.2-ac14 the VM locks up right after: calibrating APIC timer ... . CPU clock speed is 1152.4771 MHz. . host bus clock speed is 0. MHz. cpu: 0, clocks: 0, slice: 0 The last line is where it locks up... Its kind of suspicious that the clock speeds are wrong, The CPU speed should be 733 MHz (unless VMWare is using both my CPU's?) and I am almost 100% sure that the bus speed should not be 0. MHz! Can anyhone help me with this? Is it a bug on my part or what? Thanks, Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: EXPORT_SYMTAB and kernel 2.2.x
Nevermind... I figured it out... I added the modules to the MX_OBJS list and it worked fine. Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/
EXPORT_SYMTAB and kernel 2.2.x
I am trying to add some drivers to kernel 2.2.16 and I think I have everything done right except that when it tries to compile the driver I get an error message saying that EXPORT_SYMTAB is not defined. How do I fix this? I have added a target to the Config.in and I have added what I think to be appropriate lines to the Makefile (to check the config vars defined in Config.in) but I still get the error. Any help is greatly appreciated! Thanks, Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/
EXPORT_SYMTAB and kernel 2.2.x
I am trying to add some drivers to kernel 2.2.16 and I think I have everything done right except that when it tries to compile the driver I get an error message saying that EXPORT_SYMTAB is not defined. How do I fix this? I have added a target to the Config.in and I have added what I think to be appropriate lines to the Makefile (to check the config vars defined in Config.in) but I still get the error. Any help is greatly appreciated! Thanks, Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/
Re: EXPORT_SYMTAB and kernel 2.2.x
Nevermind... I figured it out... I added the modules to the MX_OBJS list and it worked fine. Mike -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/
Re: FWD: Re: Linux pipe question
Thanks for the reply! On Wed, Sep 20, 2000 at 07:29:06PM +0200, Jakob Østergaard wrote: > On Wed, Sep 20, 2000 at 12:31:25PM -0400, Mike Panetta wrote: > > Can anyone answer this? > > I am not sure if unnamed pipes in linux > > are pageable or not. If an unnamed pipe > > could be paged out what could be done > > to prevent that from happening? > > The pipe itself is not pageable, but the user programs > will use buffers to actually use the pipe, and user programs > are of course pageable. > > You might want to look into encrypted swap-space > or at least using mlock() to lock the user programs in > core. It depends on how secure you want it. Could someone > actually access the swap space (eg. steal the disk), or > could someone install compromised versions of the programs > unnoticed ? > > Most programs just fill their buffers with random data or > zeroes, after they're done with the confidential data. > > -- > ... > : [EMAIL PROTECTED] : And I see the elder races, : > :.: putrid forms of man: > : Jakob Østergaard : See him rise and claim the earth, : > :OZ9ABN : his downfall is at hand. : > :.:{Konkhra}...: > - > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to [EMAIL PROTECTED] > Please read the FAQ at http://www.tux.org/lkml/ -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/
FWD: Re: Linux pipe question
Can anyone answer this? I am not sure if unnamed pipes in linux are pageable or not. If an unnamed pipe could be paged out what could be done to prevent that from happening? TIA, Mike - Forwarded message from AW <[EMAIL PROTECTED]> - Date: Wed, 20 Sep 2000 12:27:05 -0400 From: AW <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Subject: Re: Linux pipe question > I am not sure... But would this be a named pipe or > not? This would be a UNnamed pipe, i.e., gen_confidential_data | gpg -e -r [EMAIL PROTECTED] ... The question is: is any of the clear text confidential data handled by the unnamed pipe at risk for being written to disk? Comments in the kernel code suggest that it's buffered in a single physical page but I suspect that it's actually a virtual page that could be paged out. Does the answer depend on if gen_confidential_data limits its write to not exceed PIPE_BUF (4096)? Clearly, gen_confidential_data is subject to being paged out unless it locks itself into memory. THANKS! Bob - End forwarded message - -- -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/
Re: FWD: Re: Linux pipe question
Thanks for the reply! On Wed, Sep 20, 2000 at 07:29:06PM +0200, Jakob Østergaard wrote: On Wed, Sep 20, 2000 at 12:31:25PM -0400, Mike Panetta wrote: Can anyone answer this? I am not sure if unnamed pipes in linux are pageable or not. If an unnamed pipe could be paged out what could be done to prevent that from happening? The pipe itself is not pageable, but the user programs will use buffers to actually use the pipe, and user programs are of course pageable. You might want to look into encrypted swap-space or at least using mlock() to lock the user programs in core. It depends on how secure you want it. Could someone actually access the swap space (eg. steal the disk), or could someone install compromised versions of the programs unnoticed ? Most programs just fill their buffers with random data or zeroes, after they're done with the confidential data. -- ... : [EMAIL PROTECTED] : And I see the elder races, : :.: putrid forms of man: : Jakob Østergaard : See him rise and claim the earth, : :OZ9ABN : his downfall is at hand. : :.:{Konkhra}...: - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/ -- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] Please read the FAQ at http://www.tux.org/lkml/