Hi all, I''m at the end of my wits. I desperately need some ideas. Not long ago I posted here with problems booting xen on a P5Pdeluxe motherboard. When I boot this machine with vanilla debian etch, there are no problems. ata1: SATA link up 3,0 Gpb (SStatus 123 Scontrol 200) ata1: qc timeout (cmd 0xec) ata1: failed to identify (I/0 error, err_mask = 0x0104) ata1: port is slow to resond, please be patient ata1: failes to respond (30 secs) ata1: COMRESET failed, device not ready ata1: hardreset failded But this doesn''t always happen. 50% of the time, the process stalls for about 6 secs on. ACP: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (low level) -> IRQ 18 And then continues correctly. The machine also stalls during boot at usb hub detection, and NIC detection (or loading module). Well, I gave up, thought it was some weird hardware issue, and decided to use that machine as my own PC (because it boots etch just fine). So I started afresh with an asus P5LD2 and I''ve got exactly the same problem.!! The only thing that these two machines have in common is the powercord and I''ve changed that too! I''m not new to Xen, I''ve installed at least 15 since 3.0.1 These problems occur with my own compiled kernels and the debian packaged kernels. Any ideas please? Chris. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Chris Fanning wrote:> Hi all, > > I''m at the end of my wits. > I desperately need some ideas. > > Not long ago I posted here with problems booting xen on a P5Pdeluxe > motherboard. > When I boot this machine with vanilla debian etch, there are no > problems. > > ata1: SATA link up 3,0 Gpb (SStatus 123 Scontrol 200) > ata1: qc timeout (cmd 0xec) > ata1: failed to identify (I/0 error, err_mask = 0x0104) > ata1: port is slow to resond, please be patient > ata1: failes to respond (30 secs) > ata1: COMRESET failed, device not ready > ata1: hardreset failded > > But this doesn''t always happen. 50% of the time, the process stalls > for about 6 secs on. > ACP: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (low level) -> IRQ 18 > And then continues correctly. > > The machine also stalls during boot at usb hub detection, and NIC > detection (or loading module). > > Well, I gave up, thought it was some weird hardware issue, and > decided to use that machine as my own PC (because it boots etch just > fine). > > So I started afresh with an asus P5LD2 and I''ve got exactly the same > problem.!! > > The only thing that these two machines have in common is the > powercord and I''ve changed that too! > > I''m not new to Xen, I''ve installed at least 15 since 3.0.1 These > problems occur with my own compiled kernels and the debian packaged > kernels. > > Any ideas please? > > Chris.Sorry, no help from me. I just wanted to say I''m still in the same boat as you, although I use Fedora and a Dell Optiplex 745 machine. Mine boots about 5% of the time; I wish it would boot 50% of the time! My errors are similar to yours but not identical due to different distro. I can boot the vanilla Fedora 7 kernel, but not the Fedora 7 Xen kernel package or the manually compiled Xen kernel. What do we have in common? Error messages during or immediately before or after ata detection. Both the 965 and 945 chipsets lack native ata support, so it''s either not present (in my case) or tacked on with a 3rd party device by motherboard manufacturers (in your case). In both cases, however, it seems Xen can cause a lot of trouble with detecting ata, whether it''s physically present or not. Others have posted that Xen works fine with their 965 boards, so unfortunately this isn''t a consistent problem. I have determined that my problem seems to be occuring during the ramdisk''s (initrd) loading of modules, before the kernel even gets to start booting. There''s problems with the ata drivers. The thing is, no amount of remaking or fiddling with my initrd contents seems to fix this, but I''m no Linux guru so maybe there is a fix. Chris, does your system use an initrd or does it boot straight from the kernel? This might be useful information, although I don''t know what it would tell us at this point. I believe we''ve dismissed the previously suggested notion that this is a hardware problem, since you and I have had it on 3 separate models of boards. I''ve flashed my BIOS to the latest version, and done as much troubleshooting as I know how, but still no luck fixing it. All I can think to do is wait for the next release of a new Xen kernel. I''d love it if someone could give us a better option, though. Steve Brueckner, ATC-NY _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
I believe that certain parts of the chipset are unsupported in kernel 2.6.18. Ubuntu feisty has Xen ported to 2.6.19, and that will successfully boot on the Asus motherboard 100% of the time. I got it to boot and recognize all 7 internal sata drives I was using. There is a potential bug that may affect you if you use driver domains, however. I didn''t isolate the bug, but be aware that it potentially exists (softraid corruption when using a driver domain with 6 usb controllers and the only graphics card passed to it). Cheers, David On 7/3/07, Steve Brueckner <steve@atc-nycorp.com> wrote:> > Chris Fanning wrote: > > Hi all, > > > > I''m at the end of my wits. > > I desperately need some ideas. > > > > Not long ago I posted here with problems booting xen on a P5Pdeluxe > > motherboard. > > When I boot this machine with vanilla debian etch, there are no > > problems. > > > > ata1: SATA link up 3,0 Gpb (SStatus 123 Scontrol 200) > > ata1: qc timeout (cmd 0xec) > > ata1: failed to identify (I/0 error, err_mask = 0x0104) > > ata1: port is slow to resond, please be patient > > ata1: failes to respond (30 secs) > > ata1: COMRESET failed, device not ready > > ata1: hardreset failded > > > > But this doesn''t always happen. 50% of the time, the process stalls > > for about 6 secs on. > > ACP: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (low level) -> IRQ 18 > > And then continues correctly. > > > > The machine also stalls during boot at usb hub detection, and NIC > > detection (or loading module). > > > > Well, I gave up, thought it was some weird hardware issue, and > > decided to use that machine as my own PC (because it boots etch just > > fine). > > > > So I started afresh with an asus P5LD2 and I''ve got exactly the same > > problem.!! > > > > The only thing that these two machines have in common is the > > powercord and I''ve changed that too! > > > > I''m not new to Xen, I''ve installed at least 15 since 3.0.1 These > > problems occur with my own compiled kernels and the debian packaged > > kernels. > > > > Any ideas please? > > > > Chris. > > Sorry, no help from me. I just wanted to say I''m still in the same > boat as you, although I use Fedora and a Dell Optiplex 745 machine. > > Mine boots about 5% of the time; I wish it would boot 50% of the time! > My errors are similar to yours but not identical due to different > distro. I can boot the vanilla Fedora 7 kernel, but not the Fedora 7 > Xen kernel package or the manually compiled Xen kernel. > > What do we have in common? Error messages during or immediately > before or after ata detection. Both the 965 and 945 chipsets lack > native ata support, so it''s either not present (in my case) or tacked > on with a 3rd party device by motherboard manufacturers (in your > case). In both cases, however, it seems Xen can cause a lot of > trouble with detecting ata, whether it''s physically present or not. > Others have posted that Xen works fine with their 965 boards, so > unfortunately this isn''t a consistent problem. > > I have determined that my problem seems to be occuring during the > ramdisk''s (initrd) loading of modules, before the kernel even gets to > start booting. There''s problems with the ata drivers. The thing is, > no amount of remaking or fiddling with my initrd contents seems to fix > this, but I''m no Linux guru so maybe there is a fix. > > Chris, does your system use an initrd or does it boot straight from > the kernel? This might be useful information, although I don''t know > what it would tell us at this point. > > I believe we''ve dismissed the previously suggested notion that this is > a hardware problem, since you and I have had it on 3 separate models > of boards. I''ve flashed my BIOS to the latest version, and done as > much troubleshooting as I know how, but still no luck fixing it. > > All I can think to do is wait for the next release of a new Xen kernel. > I''d love it if someone could give us a better option, though. > > Steve Brueckner, ATC-NY > > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi,> > I have determined that my problem seems to be occuring during the > > ramdisk''s (initrd) loading of modules, before the kernel even gets to > > start booting. There''s problems with the ata drivers. The thing is, > > no amount of remaking or fiddling with my initrd contents seems to fix > > this, but I''m no Linux guru so maybe there is a fix. > >I''m no expert either. although we seem to have the same problem, I think mine is IRQ related. comparing dmesg and cat /proc/interrupts I could see (from memory) that the sata device conflicted with pci-x bus, so I changed the sata setting in the bios to compatible (S-ATA only). a USB hub conflicted with pci-x video card, so I disactivated usb funcionality in the bios. One NIC conflicted with an IRQ that I could see come up during boot using dmesg, but wasn''t present in /proc/interrupts, so I removed it anyway and activated the onboard NIC. (luckily that didn''t create any new conflicts). So, now I''ve got s server with no cdrom or usb, but Xen boots 100% of the time! (fingers crossed). After looking at sooo many threads I get the feeling that the hipervisor or dom0 (I don''t know which), gets a corrupt acpi table from the bios.> > Chris, does your system use an initrd or does it boot straight from > > the kernel? This might be useful information, although I don''t know > > what it would tell us at this point.I''m not too up on the boot process. kernel /xen.gz dom0_mem=512M module /vmlinuz-2.6.18-xen root=/dev/md2 ro console=tty0 module /initrd.img-xen-3.1.0 I think the hipervisor boots (don''t know how) and calls dom0 which then boots itself on the initramfs. or something like that.> > I have determined that my problem seems to be occuring during the > > ramdisk''s (initrd) loading of modules,Same here, it all happens before the switchroot. Hey Steve, if you''ve still got that machine around and can be bothered, who knows? I seemed to have stumbled out of a problem I still can believe I really had! Anyway, I''ve had two of these bad eggs in a row (I still haven''t tried this on p5pdelux) and it makes me wonder what I should be looking out for when I buy a motherboard. God''s blessing? Cheers. Chris. On 7/3/07, David Greenberg <dgrnbrg@mit.edu> wrote:> I believe that certain parts of the chipset are unsupported in kernel > 2.6.18. Ubuntu feisty has Xen ported to 2.6.19, and that will successfully > boot on the Asus motherboard 100% of the time. I got it to boot and > recognize all 7 internal sata drives I was using. There is a potential bug > that may affect you if you use driver domains, however. I didn''t isolate > the bug, but be aware that it potentially exists (softraid corruption when > using a driver domain with 6 usb controllers and the only graphics card > passed to it). > > Cheers, > David > > > On 7/3/07, Steve Brueckner <steve@atc-nycorp.com> wrote: > > Chris Fanning wrote: > > > Hi all, > > > > > > I''m at the end of my wits. > > > I desperately need some ideas. > > > > > > Not long ago I posted here with problems booting xen on a P5Pdeluxe > > > motherboard. > > > When I boot this machine with vanilla debian etch, there are no > > > problems. > > > > > > ata1: SATA link up 3,0 Gpb (SStatus 123 Scontrol 200) > > > ata1: qc timeout (cmd 0xec) > > > ata1: failed to identify (I/0 error, err_mask = 0x0104) > > > ata1: port is slow to resond, please be patient > > > ata1: failes to respond (30 secs) > > > ata1: COMRESET failed, device not ready > > > ata1: hardreset failded > > > > > > But this doesn''t always happen. 50% of the time, the process stalls > > > for about 6 secs on. > > > ACP: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (low level) -> IRQ 18 > > > And then continues correctly. > > > > > > The machine also stalls during boot at usb hub detection, and NIC > > > detection (or loading module). > > > > > > Well, I gave up, thought it was some weird hardware issue, and > > > decided to use that machine as my own PC (because it boots etch just > > > fine). > > > > > > So I started afresh with an asus P5LD2 and I''ve got exactly the same > > > problem.!! > > > > > > The only thing that these two machines have in common is the > > > powercord and I''ve changed that too! > > > > > > I''m not new to Xen, I''ve installed at least 15 since 3.0.1 These > > > problems occur with my own compiled kernels and the debian packaged > > > kernels. > > > > > > Any ideas please? > > > > > > Chris. > > > > Sorry, no help from me. I just wanted to say I''m still in the same > > boat as you, although I use Fedora and a Dell Optiplex 745 machine. > > > > Mine boots about 5% of the time; I wish it would boot 50% of the time! > > My errors are similar to yours but not identical due to different > > distro. I can boot the vanilla Fedora 7 kernel, but not the Fedora 7 > > Xen kernel package or the manually compiled Xen kernel. > > > > What do we have in common? Error messages during or immediately > > before or after ata detection. Both the 965 and 945 chipsets lack > > native ata support, so it''s either not present (in my case) or tacked > > on with a 3rd party device by motherboard manufacturers (in your > > case). In both cases, however, it seems Xen can cause a lot of > > trouble with detecting ata, whether it''s physically present or not. > > Others have posted that Xen works fine with their 965 boards, so > > unfortunately this isn''t a consistent problem. > > > > I have determined that my problem seems to be occuring during the > > ramdisk''s (initrd) loading of modules, before the kernel even gets to > > start booting. There''s problems with the ata drivers. The thing is, > > no amount of remaking or fiddling with my initrd contents seems to fix > > this, but I''m no Linux guru so maybe there is a fix. > > > > Chris, does your system use an initrd or does it boot straight from > > the kernel? This might be useful information, although I don''t know > > what it would tell us at this point. > > > > I believe we''ve dismissed the previously suggested notion that this is > > a hardware problem, since you and I have had it on 3 separate models > > of boards. I''ve flashed my BIOS to the latest version, and done as > > much troubleshooting as I know how, but still no luck fixing it. > > > > All I can think to do is wait for the next release of a new Xen kernel. > > I''d love it if someone could give us a better option, though. > > > > Steve Brueckner, ATC-NY > > > > > > _______________________________________________ > > Xen-users mailing list > > Xen-users@lists.xensource.com > > http://lists.xensource.com/xen-users > > > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On 7/3/07, David Greenberg <dgrnbrg@mit.edu> wrote:> I believe that certain parts of the chipset are unsupported in kernel > 2.6.18. Ubuntu feisty has Xen ported to 2.6.19, and that will successfully > boot on the Asus motherboard 100% of the time. I got it to boot and > recognize all 7 internal sata drives I was using. There is a potential bug > that may affect you if you use driver domains, however. I didn''t isolate > the bug, but be aware that it potentially exists (softraid corruption when > using a driver domain with 6 usb controllers and the only graphics card > passed to it).Hi David, Unfortunately, 2.6.19 didn''t help either. Here is the original P5PDeluxe thread http://lists.xensource.com/archives/html/xen-users/2007-06/msg00797.html Cheers. Chris.> Cheers, > David > > > On 7/3/07, Steve Brueckner <steve@atc-nycorp.com> wrote: > > Chris Fanning wrote: > > > Hi all, > > > > > > I''m at the end of my wits. > > > I desperately need some ideas. > > > > > > Not long ago I posted here with problems booting xen on a P5Pdeluxe > > > motherboard. > > > When I boot this machine with vanilla debian etch, there are no > > > problems. > > > > > > ata1: SATA link up 3,0 Gpb (SStatus 123 Scontrol 200) > > > ata1: qc timeout (cmd 0xec) > > > ata1: failed to identify (I/0 error, err_mask = 0x0104) > > > ata1: port is slow to resond, please be patient > > > ata1: failes to respond (30 secs) > > > ata1: COMRESET failed, device not ready > > > ata1: hardreset failded > > > > > > But this doesn''t always happen. 50% of the time, the process stalls > > > for about 6 secs on. > > > ACP: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (low level) -> IRQ 18 > > > And then continues correctly. > > > > > > The machine also stalls during boot at usb hub detection, and NIC > > > detection (or loading module). > > > > > > Well, I gave up, thought it was some weird hardware issue, and > > > decided to use that machine as my own PC (because it boots etch just > > > fine). > > > > > > So I started afresh with an asus P5LD2 and I''ve got exactly the same > > > problem.!! > > > > > > The only thing that these two machines have in common is the > > > powercord and I''ve changed that too! > > > > > > I''m not new to Xen, I''ve installed at least 15 since 3.0.1 These > > > problems occur with my own compiled kernels and the debian packaged > > > kernels. > > > > > > Any ideas please? > > > > > > Chris. > > > > Sorry, no help from me. I just wanted to say I''m still in the same > > boat as you, although I use Fedora and a Dell Optiplex 745 machine. > > > > Mine boots about 5% of the time; I wish it would boot 50% of the time! > > My errors are similar to yours but not identical due to different > > distro. I can boot the vanilla Fedora 7 kernel, but not the Fedora 7 > > Xen kernel package or the manually compiled Xen kernel. > > > > What do we have in common? Error messages during or immediately > > before or after ata detection. Both the 965 and 945 chipsets lack > > native ata support, so it''s either not present (in my case) or tacked > > on with a 3rd party device by motherboard manufacturers (in your > > case). In both cases, however, it seems Xen can cause a lot of > > trouble with detecting ata, whether it''s physically present or not. > > Others have posted that Xen works fine with their 965 boards, so > > unfortunately this isn''t a consistent problem. > > > > I have determined that my problem seems to be occuring during the > > ramdisk''s (initrd) loading of modules, before the kernel even gets to > > start booting. There''s problems with the ata drivers. The thing is, > > no amount of remaking or fiddling with my initrd contents seems to fix > > this, but I''m no Linux guru so maybe there is a fix. > > > > Chris, does your system use an initrd or does it boot straight from > > the kernel? This might be useful information, although I don''t know > > what it would tell us at this point. > > > > I believe we''ve dismissed the previously suggested notion that this is > > a hardware problem, since you and I have had it on 3 separate models > > of boards. I''ve flashed my BIOS to the latest version, and done as > > much troubleshooting as I know how, but still no luck fixing it. > > > > All I can think to do is wait for the next release of a new Xen kernel. > > I''d love it if someone could give us a better option, though. > > > > Steve Brueckner, ATC-NY > > > > > > _______________________________________________ > > Xen-users mailing list > > Xen-users@lists.xensource.com > > http://lists.xensource.com/xen-users > > > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users