Bug 8034
Summary: | sata_nv broken by not using nolapic and acpi=off | ||
---|---|---|---|
Product: | ACPI | Reporter: | Florian Delizy (florian.delizy) |
Component: | Config-Interrupts | Assignee: | Zhang Rui (rui.zhang) |
Status: | REJECTED INSUFFICIENT_DATA | ||
Severity: | high | CC: | achrisjo, astarikovskiy, bill-osdl.org-bugzilla, bunk, fischer.timo, florian.delizy, htejun, protasnb, rui.zhang |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.20 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: |
lspci log on 2.6.20
dmesg on 2.6.20 acpidump on 2.6.20 panic boot log |
Description
Florian Delizy
2007-02-18 04:46:13 UTC
I forgot to mention (that was obvious on that mb), but I am running a x86_64 kernel (amd64). I got the inverse problem on MCP61. sata_nv induces kernel panic when acpi=off. (needs everything on) Could you please attach lspci, dmesg and acpidump outputs with the latest kernel... Snapshot of information on a screen then hanged might be useful too. Created attachment 11716 [details]
lspci log on 2.6.20
Created attachment 11717 [details]
dmesg on 2.6.20
Created attachment 11718 [details]
acpidump on 2.6.20
> Could you please attach lspci, dmesg and acpidump outputs with the latest
> kernel... Snapshot of information on a screen then hanged might be useful too.
Here you go, for the "Snapshot of information on a screen then hanged', I don't
get what you need, could you be a little more specific ?
Could you try to boot with acpi_use_timer_override option? With this option, the kernel panic not founding the rootfs ... It seems like nv_sata does not get called. Here is the end of the log : ACPI: (supports S0 S1 S2 S3 S4 S5) Time: tsc clocksource has been installed VFS: Cannot open root device "sda5" or unknown-block (0,0) Please append a correct "root=" boot option Kernel Panic - not syncing : VFS Unable to mount root fs on unknown-block(0,0) I don't see any message from the sata system in the log ... Any chance you could connect serial console to machine and send a bootlog? Sure, but still, the bootlog does not contain more information ... (I re-wrote more than 50% of the present lines). I guess I would have better luck in looking directly at the code (patching the kernel with kgdb, then breaking in the proper function ...). Do you have any clue of what could eventually cause that kind of thing, and/or a starting point in the code so I can start to look for it ? Anyway, i'll plug a serial cable to get the console log out of it (but not today). nolapic is about interrupt controller, and the only way how ACPI could affect detection of your hard drive is to set interrupts incorrectly. Thus you need to find, what interrupt setting makes your sata_nv disappear... It could be wrong interrupt line, it could be level vs. edge of interrupt issue... So full log will help here, mostly the place there interrupts are assigned. Also, it might be worth checking if your BIOS is new and interrupt settings in it are at default values. Florian, Were you able to use the serial and get the full boot log as requested in #10? You can also add "debug apic=debug" on the boot line so it hopefully will give more detail. Thanks. General: New kernels fails detecting hard-drive. Hardware: CPU:AMD Turion TL-58 (1.9 -GHz, 1-MB L2 cache) Chipset:Northbridge/Southbridge: nVidia MCP67M Disk drive: Hitachi HTS541616J9SA00 ATA ( reported by MS Vista device manager ) The machine specifications: http://h10025.www1.hp.com/ewfrf/wc/manualCategory?product=3466328&lc=en&cc=us&dlc=en&lang=en&cc=us ( i/o ports, mem-map, interrupts is listed ) ------------------------- Fedora core 6 x86_64 works with initial kernel, but receiving kernel upgrade package, it fails. Fedora7-x86_64, Fedora7-i386 - show the same as the following: ! Fedora7.9-x86_64. !- the following is what I have copied manually from what I may read from screen: ! . ata4: SATA link down (SStatus 0 SControl 300) pata_amd 0000:00:06.0: version 0.3.8 PCI: Setting latency timer of device 0000:00:06.0 to 64 scsi4: pata_amd scsi5: pata_amd ata5: PATA max UDMA/133 cmd 0x00000000000101f0 ctl 0x00000000000103f6 bmdma 0x00000000000130c0 irq 14 ata6: PATA max UDMA/133 cmd 0x0000000000010170 ctl 0x0000000000010376 bmdma 0x00000000000130c8 irq 15 ata5.00: ATAPI: MATSHITADVD-RAM UJ-851S, 1.50, max MWDMA2 ata5.00: ATAPI: configured for MWDMA2 ata6: port disabled. ignoring. scsi 4:0:0:0: CD-ROM MATSHITADVD-RAM UJ-851S, 1.50 PQ: 0 ANSI: 5 device-mapper: ioctl: 4.11.0-ioctl (2006-10-12) initialised: dm-devel@redhat.com Reading all physical volumes. This may take a while... No volume groups found Volume group "VolGroup00" not found Unable to access resume device (/dev/VolGroup00/LogVol01) mount: could not find filesystem '/dev/root' setuproot: moving /dev failed: No such file or directory Setuproot: error mounting /proc: No such file or directory Setuproot: error mounting /sys: No such file or directory switchroot: mount failed: No such file or directory Kernel panic - not syncing: Attempting to kill init! ----------------------- Identical configured kernels, here OpenSuse 10.2 with kernel 2.6.18.2-34 86x64 failed on a kernel 2.6.22.5-282 86x64 ! Following is manually retyped as what is shown on screen, OpenSuse 10.2, ! running 2.6.22.5-282-default ( cloned from working system 2.6.18.2-34-default ) ! It starts up identical after grub selection as far as able to see ( escape splash-screen) ! but the top lines scrolled out of screen..so I could not type it ... ! As the shell don't have lsmod, did cat/proc/modules instead. . . mice: PS/2 mouse device common for all mice input: AT Translated Set 2 keyboard as /class/input/input0 input: PC Speaker as /class/input/input1 Synaptics Touchpad, model: 1, fw: 6.3, id: 0x1a0b1, capa: 0xa04713/0x200000 input: SynPS/2 Synaptics TouchPat as /class/input/input2 NET: Registered protocol family 1 Freeing unused kernel memory: 324k freed Boot logging started on /dev/tty1(/dev/console) at Wed Sep 12 21:55:23 2007 Trying manual resume from /dev/sda5 Creating device nodes with udev Loading scsi_mod Loading processor ACPI: CPU0 (power states: C1[C1] C2[C2] C3[C3]) ACPI: Processor [CPU0] (support 8 throttling states ) ACPI Exception (processor_core-0787): AE_NOT_FOUND, Processor Device is not present [20070126] ACPI Exception (processor_core-0787): AE_NOT_FOUND, Processor Device is not present [20070126] Loading thermal ACPI Exception (thermal-0317): AE_BAD_DATA, No critical threshold [20070126] Loading libata Loading pata_amd scsi0 : pata_amd scsi1 : pata_amd ata1: PATA max UDMA/133 cmd 0x00000000000101f0 ctl 0x00000000000103f6 bmdma 0x00000000000130c0 irq 14 ata2: PATA max UDMA/133 cmd 0x0000000000010170 ctl 0x0000000000010376 bmdma 0x00000000000130c8 irq 15 ata1.00: ATAPI: MATSHITA DVD-RAM UJ-851S, 1.50, max MWDMA2 ata1.00: configured for MWDMA2 sr 0:0:0:0: CD-ROM MATSHITA DVD-RAM UJ-851S 1.50 PQ: 0 ANSI: 5 Loading jbd Loading mbcache Loading ext3 resume device /dev/sda not found (ignoring) Waiting for device /dev/sda7 to appear:..............................not found -- 2.6.18.2-34-default #1 SMP Mon Nov 27 11:46:27 UTC 2006 x86_64 x86_64 x86_64 GNU/Linux ---------------------- From OpenSuse 10.2 kernel 2.6.18.2-34 86x64 ( on working kernel ) lsmod Module Size Used by bridge 81328 0 iptable_filter 19968 0 ip_tables 39400 1 iptable_filter x_tables 37384 1 ip_tables joydev 28160 0 af_packet 57356 2 ipv6 357728 16 snd_pcm_oss 71680 0 snd_mixer_oss 35840 1 snd_pcm_oss snd_seq 82976 0 snd_seq_device 26516 1 snd_seq cpufreq_conservative 25608 0 cpufreq_ondemand 24592 1 cpufreq_userspace 24064 0 cpufreq_powersave 18688 0 powernow_k8 32416 1 freq_table 22912 1 powernow_k8 button 24736 0 battery 28168 0 ac 22792 0 apparmor 74264 0 aamatch_pcre 31232 1 apparmor nls_utf8 18944 2 ntfs 209032 2 loop 34064 0 dm_mod 81872 0 ohci1394 52040 0 ehci_hcd 51080 0 sdhci 35980 0 mmc_core 45056 1 sdhci ohci_hcd 38404 0 ieee1394 130552 1 ohci1394 usbcore 148064 2 ehci_hcd,ohci_hcd forcedeth 78864 0 snd_dummy 30848 1 snd_pcm 115464 2 snd_pcm_oss,snd_dummy snd_timer 44680 2 snd_seq,snd_pcm snd 89384 9 snd_pcm_oss,snd_mixer_oss,snd_seq,snd_seq_device,snd_dummy,snd_pcm,snd_timer soundcore 28192 1 snd snd_page_alloc 27792 1 snd_pcm ext3 167696 2 mbcache 27016 1 ext3 jbd 90872 1 ext3 edd 27912 0 fan 22408 0 sg 55080 0 sr_mod 34596 0 cdrom 54056 1 sr_mod sata_nv 30084 5 pata_amd 31520 0 libata 145056 2 sata_nv,pata_amd thermal 33552 0 processor 53992 2 powernow_k8,thermal sd_mod 39296 6 scsi_mod 173744 4 sg,sr_mod,libata,sd_mod --------------------------------------- Have sent a request for info/help to linux-nforce-bugs@nvidia.com linux-bugs@nvidia.com Got this from: "Allen Martin" <AMartin@nvidia.com> MCP67M device ids are in the 2.6.22.6 kernel. Make sure your SATA controller is in AHCI mode from the BIOS. Arne, Did it work for you with the above BIOS setting? Copying Tejun, this problem seems to be different from the one in original report. Might need a separate bugzilla. Fransois, can you provide more information? It would be better to have new bugzilla open for your case also. And I guess we are still awaiting a boot trace from Florian. I'll try the BIOS setting this evening (and will try to provide a boot trace as well) stay tuned hi, you wrote: > > Arne, > Did it work for you with the above BIOS setting? > NO. ---------------------- About my situation: I am only a "user" of Linux, and this was my first introduction to bugzilla. Had no idea how to report, where, but felt forced to do it - as I have been been 'ranting/screaming/etc' all over ;) I am "out of date", both with hardware and current development culture. In fact, I may need some tutoring. Now some words about the bug report: Description from Florian Delizy, did seem close to my results. It is assigned to "ACPI", which also seem correct. It relates to nVidia chips, serial ATA, and sofort. ( EU versus Microsoft, September 17th, 2007 - as that is somewhat linked to this situation. The ACPI draft included Hewlett-Packard, Intel, Microsoft, and Phoenix, Toshiba. But here we have "ACPI-2" (my invented term) So ? It is the next frontier. ) As only Microsoft has keys to the BIOS, I have no way to do any settings. Perhaps, I am missing something ? ( configure options in the kernel ? ) At the moment I am running 2.6.23-0.184.rc6.git4.fc8, writing these words on the machine. But what a way to get it started: - I toggle the power button, "interrupt" the system, - have to do it at the right place, and the right duration. May use several attempts. Quote Florian's initial description: "...the kernel panics not finding any root partition (it seems that sata_nv does not get initialized and does not probe anything in the log)." - this is what he 'sees' - the kernel hasn't started logging, and he only have the messages that is displayed. If you look at my entered comment #14 - section "Identical configured kernels,..", it gives some more info then Florian's example. ( it's difficult to remember, and I have lost the system...but..) a:) in /etc/grub.conf , I changed the kernel/vmlinux.... to verbose, instead of quiet. b:) ..hmm...could it be /etc/sysconfig ?? Anyway, there was a scsi file or header, in which I changed a similar (?? -q) entry, to be (?? -v -v ) , perhaps even more detailed verbose report. ( Note: my ?? marks above could have been -r, just don't recall ) Because Opensuse will give a 'shell', instead of simply exit with a 'kernel panic', I did manage to get quite a bit of info, which also included memory contents for each executed command. But I had no media to save it to. Not even a serial port. c:) But what did this tell me ? - the modules are loaded into memory, start executing - would first check the primary drive, which it finds ( it made a file system in ram, installed modules and system. Opensuse simply quit here, instead of closing with a "kernel panic". ) It did boot the disk, but it will first check the rest of the file system, CDROM, etc. The drive has already stopped in order to save power, and perhaps you wanted to run a live CD instead ? A think a command 'resume' is given, and THIS is when it writes: ata4: SATA link down (SStatus 0 SControl 300 ) // top of my first section comment#14 ( it is about here that I toggle the power button, and 'jump start it' - which on occation may be successful ) d:) Something killed the machine, and I lost those kernels. Had some results at a email-list, which I copied to that bug-report. Got the message: MCP67M would be in kernel2.6.22.6, and Fedora had released F8t1. But at the moment, I have no kernel source to work with, and I am uncertain if I will dig into it. My programming skills are in much doubt, the time it may take me to become familar with it, well..what good would that do ? So, I had to give bugzilla an attempt. Sooner or later, I will have learn some of it too. e:) This new "ACPI-2", well..it may not be anything new to you. But comparing specs, and other things, linux may have to decide upon either Intel or AMD, at a much earlier stage. I think that may be part of the plan. It's no problem for them - the OS is preinstalled. However, this isn't the real obstacle for linux. It is other aspects that will be a closed door. you wrote: > > Copying Tejun, this problem seems to be different from the one in original > report. Might need a separate bugzilla. > Well, I don't know what you refere to here. But I suspect that there are already several different 'bugs' that doesn't seem related. My reason for the last entry above, as my impression is that "the right hand may not know what the left is doing". This last part may be some speculation on my part, but I think it should be pretty obvious what is on the horizon. One last thing: perhaps if we compared actual models, and compared them - the picture would be easier to grasp ? ( it hasn't appeared all of a sudden, it is a ongoing development ) //ARNE bugme-daemon@bugzilla.kernel.org wrote: http://bugzilla.kernel.org/show_bug.cgi?id=8034 protasnb@gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |htejun@gmail.com ------- Comment #17 from protasnb@gmail.com 2007-09-17 22:49 ------- Arne, Did it work for you with the above BIOS setting? Copying Tejun, this problem seems to be different from the one in original report. Might need a separate bugzilla. Fransois, can you provide more information? It would be better to have new bugzilla open for your case also. And I guess we are still awaiting a boot trace from Florian. <br>hi,<br><br>you wrote:<br>><br>> Arne,<br>> Did it work for you with the above BIOS setting?<br>><br><br>NO.<br><br> ----------------------<br><br>About my situation:<br><br>I am only a "user" of Linux, and this was my first introduction to bugzilla.<br>Had no idea how to report, where, but felt forced to do it - as I have been<br>been 'ranting/screaming/etc' all over ;)<br>I am "out of date", both with hardware and current development culture. <br><br>In fact, I may need some tutoring. <br><br> <br>Now some words about the bug report:<br><br>Description from Florian Delizy, did seem close to my results.<br>It is assigned to "ACPI", which also seem correct.<br>It relates to nVidia chips, serial ATA, and sofort. <br><br><br>( EU versus Microsoft, September 17th, 2007 - as that is somewhat linked<br>to this situation. The ACPI draft included Hewlett-Packard, Intel, Microsoft,<br>and Phoenix, Toshiba. But here we have "ACPI-2" (my invented term) So ?<br>It is the next frontier. ) <br><br>As only Microsoft has keys to the BIOS, I have no way to do any settings. Perhaps,<br>I am missing something ? ( configure options in the kernel ? )<br><br>At the moment I am running 2.6.23-0.184.rc6.git4.fc8, writing these words on the<br>machine. But what a way to get it started:<br><br>- I toggle the power button, "interrupt" the system, - have to do it at the right place,<br>and the right duration. May use several attempts. <br><br>Quote Florian's initial description: "...the kernel panics not finding any root partition (it seems that sata_nv does not get initialized and does not probe anything in the log)."<br><br>- this is what he 'sees' - the kernel hasn't started logging, and he only have the messages that is displayed. <br><br>If you look at my entered comment #14 - section "Identical configured kernels,..", it<br>gives some more info then Florian's example. ( it's difficult to remember, and I have <br>lost the system...but..)<br><br>a:) <br>in /etc/grub.conf , I changed the kernel/vmlinux.... to verbose, instead of quiet.<br><br>b:)<br>..hmm...could it be /etc/sysconfig ?? Anyway, there was a scsi file<br>or header, in which I changed a similar (?? -q) entry, to be (?? -v -v ) , <br>perhaps even more detailed verbose report.<br>( Note: my ?? marks above could have been -r, just don't recall )<br><br>Because Opensuse will give a 'shell', instead of simply exit with a 'kernel panic', I did<br>manage to get quite a bit of info, which also included memory contents for each executed<br>command. But I had no media to save it to. Not even a serial port. <br><br>c:)<br>But what did this tell me ?<br><br>- the modules are loaded into memory, start executing - would first check the primary drive, which it finds ( it made a file system in ram, installed modules and system. <br>Opensuse simply quit here, instead of closing with a "kernel panic". )<br>It did boot the disk, but it will first check the rest of the file system, CDROM, etc.<br>The drive has already stopped in order to save power, and perhaps you wanted to run a live<br>CD instead ? A think a command 'resume' is given, and THIS is when it writes: <br><br>ata4: SATA link down (SStatus 0 SControl 300 ) // top of my first section comment#14 <br><br>( it is about here that I toggle the power button, and 'jump start it' - which on <br>occation may be successful )<br><br>d:)<br>Something killed the machine, and I lost those kernels. Had some results at a email-list, <br>which I copied to that bug-report. Got the message: MCP67M would be in kernel2.6.22.6,<br>and Fedora had released F8t1. But at the moment, I have no kernel source to work with,<br>and I am uncertain if I will dig into it. My programming skills are in much doubt,<br>the time it may take me to become familar with it, well..what good would that do ? <br>So, I had to give bugzilla an attempt. Sooner or later, I will have learn some of it too.<br><br>e:)<br>This new "ACPI-2", well..it may not be anything new to you. But comparing specs, and<br>other things, linux may have to decide upon either Intel or AMD, at a much earlier<br>stage. I think that may be part of the plan. It's no problem for them - the OS is preinstalled. However, this isn't the real obstacle for linux. It is other aspects that<br>will be a closed door. <br><br>you wrote:<br>><br>> Copying Tejun, this problem seems to be different from the one in original<br>> report. Might need a separate bugzilla.<br>><br><br>Well, I don't know what you refere to here. But I suspect that there are already<br>several different 'bugs' that doesn't seem related. My reason for the last entry <br>above, as my impression is that "the right hand may not know what the left is doing".<br>This last part may be some speculation on my part, but I think it should be pretty<br>obvious what is on the horizon.<br><br>One last thing: perhaps if we compared actual models, and compared them - the picture<br>would be easier to grasp ? ( it hasn't appeared all of a sudden, it is a ongoing development )<br><br>//ARNE<br><br><br><b><i>bugme-daemon@bugzilla.kernel.org</i></b> wrote:<blockquote class="replbq" style="border-left: 2px solid rgb(16, 16, 255); margin-left: 5px; padding-left: 5px;"> http://bugzilla.kernel.org/show_bug.cgi?id=8034<br><br><br>protasnb@gmail.com changed:<br><br> What |Removed |Added<br>----------------------------------------------------------------------------<br> CC| |htejun@gmail.com<br><br><br><br><br>------- Comment #17 from protasnb@gmail.com 2007-09-17 22:49 -------<br>Arne,<br>Did it work for you with the above BIOS setting?<br>Copying Tejun, this problem seems to be different from the one in original<br>report. Might need a separate bugzilla.<br>Fransois, can you provide more information? It would be better to have new<br>bugzilla open for your case also.<br>And I guess we are still awaiting a boot trace from Florian.<br><br><br>-- <br>Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email<br>------- You are receiving this mail because: -------<br>You are on the CC list for the bug, or are watching someone who is.<br></blockquote><br><p>  <hr size=1>Luggage? GPS? Comic books? <br> Check out fitting <a href="http://us.rd.yahoo.com/evt=48249/*http://search.yahoo.com/search?fr=oni_on_mail&p=graduation+gifts&cs=bz"> gifts for grads</a> at Yahoo! Search. Created attachment 12855 [details]
panic boot log
Here is the boot log you asked
I tried to find something related in the BIOS, but didn't find anything looking like "AHCI" ... all options related to Sata are (with their value) : 4th Sata Master: WDC WD2500KS-00M HDD Smart monitoring: Enabled Silicon Sata Disabled NVRAID Disabled Sata DMA Enabled I didn't find more options, I hope it will help Sorry about the mess. Don't try to edit on an unstable machine ;) - so, that's where the email disappeared ... //ARNE --------------------------------- Don't let your dream ride pass you by. Make it a reality with Yahoo! Autos. Sorry about the mess. Don't try to edit on an unstable machine ;)<br>- so, that's where the email disappeared ...<br>//ARNE<br> <p>  <hr size=1>Don't let your dream ride pass you by. <a href="http://us.rd.yahoo.com/evt=51200/*http://autos.yahoo.com/index.html;_ylc=X3oDMTFibjNlcHF0BF9TAzk3MTA3MDc2BHNlYwNtYWlsdGFncwRzbGsDYXV0b3MtZHJlYW1jYXI-"> Make it a reality</a> with Yahoo! Autos. Arne, so, the harddrive is detected on 2.6.18 (SL102) but it doesn't get detected if you upgrade the kernel to 2.6.22, right? This can be one of the following two things. 1. Distro's initrd generation is messed up and fails to load sata_nv. In SL102 emergency shell, run "cat /proc/modules" and see whether sata_nv is there. If not, run "insmod /lib/modules/{version}/kernel/drivers/ata/sata_nv.ko" and see whether devices get detected. 2. Something went wrong with sata_nv or other component and sata_nv can't detect the device anymore. In any case, full boot log will be very helpful in determining what's going on. If you have a second machine, setting up netconsole is usually the easiest way (Documentation/networking/netconsole.txt) to get boot log. I'm not sure whether this would work with root mount failure tho (module for network interface should either be in the kernel or initrd). Florian, you said that without "acpi=off" sata_nv detection doesn't happen at all, right? Can you roll you own kernel with sata_nv and SCSI modules built in (just take vendor kernel config file and change 'm' to 'y' to the options you need for the root device, network interface and netconsole) and post kernel boot log using netconsole? Your and Arne's cases have some similarity. Let's see whether it's initrd's fault or we have genuine driver regression. (In reply to comment #23) > Arne, so, the harddrive is detected on 2.6.18 (SL102) but it doesn't get > detected if you upgrade the kernel to 2.6.22, right? This can be one of the > following two things. > NOTE: I no longer have OpenSuse installed. Current install is Fedora8 test2 > 1. Distro's initrd generation is messed up and fails to load sata_nv. In > SL102 > emergency shell, run "cat /proc/modules" and see whether sata_nv is there. > If > not, run "insmod /lib/modules/{version}/kernel/drivers/ata/sata_nv.ko" and > see > whether devices get detected. > > 2. Something went wrong with sata_nv or other component and sata_nv can't > detect the device anymore. > > In any case, full boot log will be very helpful in determining what's going > on. > If you have a second machine, setting up netconsole is usually the easiest > way > (Documentation/networking/netconsole.txt) to get boot log. I'm not sure > whether this would work with root mount failure tho (module for network > interface should either be in the kernel or initrd). > BUT let me add some info regarding these previous kernels: a:) Release version Fedora core 6 - did get installed, but failed on receiving upgraded kernels. b:) Release Fedora7 + different test releases all failed. ( in the above, this include both i386 and x86_64 ) c:) However, I did download something from Cambridge?-and got some Fedora7 installed with ZEN. But as I don't have a clue what is involved, I gave it up. d:) A reason for this purchase, was to get wireless. While in contact with bcm43xx, I tried OpenSuse 10.2 which installed. Then downloaded a more recent kernel. Used clone-config, in order to get comparable versions. Did report this (comment #14) e:) The most troublesome issue has been the touchpad. Even after removing "tapping" ( gsynaptics or ksynaptics ) it is still troublesome. However it is a good indicator when things are very wrong, like it start executing applications on it's own. ( a hellish device that has cost me lot of grief ) ------------------ Comments to what you said: 1. I had the results as presented in (comment#19) 2. At bcm43xx - similar questions as yours came up. 3. Followed directions: 3.1 From YaST: libata and scsi_mod added to "INITRD MODULES" 3.2 executed : /sbin/mkinitrd 3.3 restarted: no apparent change 3.4 power off, restart hours later - didn't work. Machine dead, no lights, no display, no keys working. 3.5 Pulled out the RTC battery, and it did the trick. 3.6 Checked HP and found a report that did discribe this experience. Did install a BIOS upgrade ( done through Vista ), just in case. But also because perhaps it would give some new settings ? but nope. 3.7 Because of info (comment#16) I installed Fedora8 test2, as OpenSuse didn't have anything matching the kernel mentioned. In closing: - I think what I said in (comment#19) is correct. Have compared both situations, and the disk dies out - don't respond to "resume" as described. NOTE ! - you may have expected to see more things installed, more io to be reported. What I have reported - is kernels up until that latest OpenSuse. note (comment#16) - I have not supplied anything what the machine may give me now. If you compare with Florian, my chipset isn't in the kernel until recently. So, a lot of devices are not detected. The chipset enable that bridge, is what it seem like to me. This isn't a computer, it is a "movie/game/player". The chipset include encryptions/decoders on all "steams". It may be used for pay by download movies, and much more, mp3, you name it. Microsoft have the keys that may disconnect us. So ? - I am tempted to return it, see if something with a better keyboard and screen resolution can be found. But that isn't easy, as most are pretty much the same, geared towards the same marked. //ARNE (In reply to comment #23) > > If you have a second machine, setting up netconsole is usually the easiest > way > (Documentation/networking/netconsole.txt) to get boot log. I'm not sure > whether this would work with root mount failure tho (module for network > interface should either be in the kernel or initrd). > Have to look into it. I have no kernel-source, think it's something from rawhide, not sure where to get the source. Not sure where to get it. I simply wasn't prepared for all the activity and this sort of learning curve. //ARNE //ARNE Arne, sorry but can't really debug with the information you provided till now. I can't even determine whether it's a genuine driver/kernel bug or kernel upgrade went wrong. Please... 1. Roll your own vanilla kernel preferably 2.6.22.6 and build in all the drivers you need into it. 2. See whether you can reproduce the problem. If so, please report back here; otherwise, please report the problem to the respective distros. This bug report is getting very confusing. Currently there are two problems here. a. Routing goes wrong with ACPI. This can be worked around by "acpi=off nolapic". b. SATA controller isn't probed at all without "acpi=off" or under other circumstances. #a is verified the kernel log. I'm still not sure whether #b actually exists or not. If anyone can regenerate the problem, please post the failing boot log and the result of "ls /sys/bus/pci/drivers/" if possible. Arne, you can get the vanilla kernel source from http://kernel.org/ and ask google "kernel howto" to get information on how to build a kernel. Thanks. Hi, I did leave a message, but I don't see it here. Everything that can go wrong, did - so I do have some problems to solve. Lost X, etc. Didn't get to the patch stuff, as I need modules and other setting. More tomorrow. Had to do a new install with F8t2 from DVD. Things a much more difficult to "toggle start". But - having to boot up from CD, I have to do the same. That is a PATA drive. ( just thought I should mention it, as I had forgotten it ) Also, the place I toggle, is when the message it says "routing" on screen. Have to pass that. ( the specification from HP, says ioport, mem, and irq ) and it is different on AMD then on Intel. And I have compared irq's before and you seem to set them as on Intel. ( just wondered if this could be of help, or do I confuse it even more ?) hmm.. could the rescue shell on the DVD provide any info ? //ARNE hi, I have asked for assistance at the Fedora-test-site. There are problems with ACPI, that affects the touch-pad, screen, and more. Even in the most stable installs I have been able to do, - there has been problems with the ACPI. ( like touchpad, X driver, and more ) perhaps, lazy/non-lazy problems, etc. //ARNE (In reply to comment #24) > Florian, you said that without "acpi=off" sata_nv detection doesn't happen at > all, right? Yes > Can you roll you own kernel with sata_nv and SCSI modules built in Well, I *always* builld my own kernel, and, sata_nv and SCSI modules were builtin already (not module) > (just take vendor kernel config file and change 'm' to 'y' to the options you > need for the root device, network interface and netconsole) and post kernel > boot log using netconsole? The last bootlog is what you asked (precisely), (but by serial console) > Your and Arne's cases have some similarity. Let's > see whether it's initrd's fault or we have genuine driver regression. > I built 2.6.22 today, still have the problem (no change). If you need I can try to provide you with a kgdb patched version of the kernel and open a port from my firewall to connect to. (I have to see where I can find the patch). Nvidia based notebooks can have weird problems. Just Google for the fun people are having with HP Pavilion tx 1000 notebooks, for example. (The only thing that can make them work is the boot option "noapic" and then unexplained interrupts on IRQ 7 make the kernel shut down the USB driver's interrupt.) And there are no good open-source drivers for their video adapters, either, since they refuse to release any information about them. hi Chuck, Thank you ;) It didn't seem like I could get any attention to the problem, and yes, - I have checked around, there are a lot of people having problems. But basically, the install fails - due to some ACPI problem. Does this only affect Nvidia chips ? //ARNE hi, The strange thing is that earlier kernels does run, so what has changed with the distro ? What is it - distro or kernel that have changed ? //ARNE --------------------------------- Need a vacation? Get great deals to amazing places on Yahoo! Travel. hi,<br><br>The strange thing is that earlier kernels does run, so what has changed with<br>the distro ? What is it - distro or kernel that have changed ?<br><br>//ARNE<br><br><br><br><p>  <hr size=1>Need a vacation? <a href="http://us.rd.yahoo.com/evt=48256/*http://travel.yahoo.com/;_ylc=X3oDMTFhN2hucjlpBF9TAzk3NDA3NTg5BHBvcwM1BHNlYwNncm91cHMEc2xrA2VtYWlsLW5jbQ--">Get great deals to amazing places </a>on Yahoo! Travel. hi, I have just a day left of a 30 day trial of this product. Did resend an earlier mail Hate the problem, hate to give up - ...so ?? //ARNE --------------------------------- Got a little couch potato? Check out fun summer activities for kids.hi,<br>I have just a day left of a 30 day trial of this product. Did resend an earlier mail<br>Hate the problem, hate to give up - ...so ??<br>//ARNE<br><br><p>  <hr size=1>Got a little couch potato? <br> Check out fun <a href="http://us.rd.yahoo.com/evt=48248/*http://search.yahoo.com/search?fr=oni_on_mail&p=summer+activities+for+kids&cs=bz">summer activities for kids.</a> Florian, can you set up a minimal root fs on a usb disk or somewhere and report the result of 'dmesg' and 'lspci -nnvvv' when the nv controller is not detected? Setting up a minimal root fs isn't too difficult. Just put bash, lspci and dmesg binaries and all the libraries they need onto the filesystem and boot with "init=/bin/bash" kernel parameter. You can test minimal root fs works without rebooting by chrooting to it. If you can connect an IDE drive, just installing to an IDE drive should work too. Thanks. hi, Include a few files, dumps of the system at the moment. Just in hope it could be of some interest. - sorry that I haven't been able to patch a vanilla kernel yet, ( I would need time to figure out how to do it, it's years since I did anything like that ) -noapic in grub.conf helped a lot. //ARNE --------------------------------- Be a better Globetrotter. Get better travel answers from someone who knows. Yahoo! Answers - Check it out.hi,<br><br>Include a few files, dumps of the system at the moment.<br>Just in hope it could be of some interest. <br>- sorry that I haven't been able to patch a vanilla kernel<br>yet, ( I would need time to figure out how to do it, it's <br>years since I did anything like that ) <br>-noapic in grub.conf helped a lot. <br>//ARNE<br><p>  <hr size=1>Be a better Globetrotter. <a href="http://us.rd.yahoo.com/evt=48254/*http://answers.yahoo.com/dir/_ylc=X3oDMTI5MGx2aThyBF9TAzIxMTU1MDAzNTIEX3MDMzk2NTQ1MTAzBHNlYwNCQUJwaWxsYXJfTklfMzYwBHNsawNQcm9kdWN0X3F1ZXN0aW9uX3BhZ2U-?link=list&sid=396545469">Get better travel answers </a>from someone who knows.<br>Yahoo! Answers - Check it out. sorry, things look messy. Those attachments was .tgz , perhaps that is stripped off ? logs, ls(pci,mod,usb) dmesg, etc. for the running system with kernel 2.6.23-0.164.rc5.fc8, kernel 2.6.23-0.189.rc6.git8.fc8 //ARNE (In reply to comment #27) I have worked like crazy in order to supply you with something, while I still could. Done #1 - but did some mistake in configuration . I think. Lack some equipment in order to set up a console, not sure how I could have provided the results. With what they have done as Chuck explained, USB doesn't work, but else I don't recognize the machine. If you didn't get the .tgz and would like to see them, just ask. //ARNE ping response for comment# 36... pong for comment #36. Sorry I've been a little busy, I'll try to arrange that on Sunday Florian, any test result? Hello I encountered the same problem on my Notebook with the MCP67M chipset from nvidia that arne had (2.6.20 works, later not). I think it is a different bug, but I couldn't find it anywhere else... I used the ahci driver in the meantime but had several other caveats. This thread put me on the right track. I'm using the gentoo-sources by now. 2.6.20 worked flawless, later kernels didn't detect the nvidia sata chipset. what I did was the following: I compared the file drivers/ata/nvidia.c between the 2.6.20 and the 2.6.23 version (both with the gentoo chipset, but I didn't had any luck with the git or vanilla sources) I found following difference to mention: 2.6.20: static const struct pci_device_id nv_pci_tbl[] = { { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE2S_SATA), NFORCE2 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE3S_SATA), NFORCE3 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE3S_SATA2), NFORCE3 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_CK804_SATA), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_CK804_SATA2), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP04_SATA), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP04_SATA2), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP51_SATA), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP51_SATA2), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP55_SATA), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP55_SATA2), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SATA), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SATA2), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SATA3), GENERIC }, { PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID, PCI_CLASS_STORAGE_IDE<<8, 0xffff00, GENERIC }, { PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID, PCI_CLASS_STORAGE_RAID<<8, 0xffff00, GENERIC }, { } /* terminate list */ And 2.6.23: /* static const struct pci_device_id nv_pci_tbl[] = { { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE2S_SATA), NFORCE2 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE3S_SATA), NFORCE3 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE3S_SATA2), NFORCE3 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_CK804_SATA), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_CK804_SATA2), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP04_SATA), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP04_SATA2), CK804 }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP51_SATA), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP51_SATA2), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP55_SATA), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP55_SATA2), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SATA), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SATA2), GENERIC }, { PCI_VDEVICE(NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SATA3), GENERIC }, { } /*terminate list */ I simply copied the old version (from 2.6.20) to the newer kernel (2.6.23) compiled the kernel and TADA it worked. As I don't have any programming scills, please correct me. But as fas as I can see the MCP67M is simply not mentioned any longer (I think this PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID did the trick) but the source itself is working. I think the MCP67M just has to be added to the list of hardware (I don't know in the sata_nv.c or the in the pci_ids.h (is there the definition?) Please tell me if I'm on the right track, as this kind of fiddling around is completely new to me (even my first bugzilla entry). Greets Timo Will you please try to boot with the option of "pci=noacpi noapic" and attach the output of dmesg? It will be great if the debug function of pci and ACPI is enabled in kernel configuration. Thanks. close this bug as no response from the bug reporter. |