http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c1 Takashi Iwai changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tiwai@suse.com, | |w.pelser@web.de Flags| |needinfo?(w.pelser@web.de) --- Comment #1 from Takashi Iwai --- Does the boot succeed without USB stick? And, is the kernel freeze, or does it still react to magic sysrq? Also, could you try with kernel-debug kernel (and without quiet bot option)? In anyway, it'd be helpful if you can get kernel logs at the hang. Thanks. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c3 --- Comment #3 from Takashi Iwai --- At least you can give the full kernel messages when you could boot normally. And please give "hwinfo --full" output, too. Last but not least, once after you succeed the boot without USB stick, what happens if you plug in the USB stick there? Does it freeze? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c5 --- Comment #5 from Walther Pelser --- Created attachment 639414 --> http://bugzilla.novell.com/attachment.cgi?id=639414&action=edit boot message -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c7 --- Comment #7 from Walther Pelser --- Created attachment 639416 --> http://bugzilla.novell.com/attachment.cgi?id=639416&action=edit hwinfo how? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c9 --- Comment #9 from Walther Pelser --- Sorry for the changes of the names of the attachments. "boot messages" is "boot messages.txt", "USB-Stick (p)lugged in later" is "boot messages plugged in later.txt", "hwinfo how?" is "hwinfo.txt", "YAST-hwinfo" is "YAST-hwinfo.txt". -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c11 --- Comment #11 from Walther Pelser --- Recovery-mode has the no-resume option. But it could not boot too. 2 screenshots are added from this. First from the boot screen, after waiting a lot; second from sysctl, which could be called from the switched on konsole. I guess, that the rpms are build in different ways, one from kernel.org, the other from SUSE. So there are different initrds.(There is no symvers build by kernel.org. kernel) Kernel 4.1 seemes to have a lot of timing problems f. e.: -DVB-T is stored in lib/firmware, but kernel could not find it, because he searches too early. Disconnecting the DVB-T stick, and then connecting it again helps. -Similar: See journalctl.jpeg, the raid-partitions could not be found, maybe the raid subsystem is not yet activated? -wicked is only randomly started completely. In most cases I had to start it after wards with YAST. My USB 3.0 controller is a PCIe card with a renesas chip on it. In kernel 4.1 there are a lot of changes for several renesas chips. ...... -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c13 --- Comment #13 from Walther Pelser --- Created attachment 639424 --> http://bugzilla.novell.com/attachment.cgi?id=639424&action=edit journalctl.jpeg -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c15 --- Comment #15 from Walther Pelser --- Here comes config-4.1.0-1-desktop.old from /boot and .config, which is used during kernel build. partitions.txt is from /proc/partitions -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c17 --- Comment #17 from Walther Pelser --- Created attachment 639426 --> http://bugzilla.novell.com/attachment.cgi?id=639426&action=edit config-4.1.0-1-desktop.old -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c19 --- Comment #19 from Neil Brown --- Could you please confirm precisely which versions of dracut and mdadm you have installed? rpm -q dracut mdadm -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c21 --- Comment #21 from Walther Pelser --- Created attachment 639553 --> http://bugzilla.novell.com/attachment.cgi?id=639553&action=edit dracut.log-20150625 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c23 --- Comment #23 from Walther Pelser --- No change with new dracut! In the meantime I compiled kernel 4.1.1. This kernel is running only with little problems. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c25 --- Comment #25 from Neil Brown --- I've had a look over all the details again. As you said there is a swap partition on the USB stick. When you try to boot without that plugged in, the initrd waits for it - because it really wants all swap partitions. To help with testing, it would be good to tell the system to ignore that swap partition. So boot with the USB plugged in (using any kernel that works), then turn of the swap to /dev/sdd2 (swapoff /dev/sdd2) and unplug the USB. Then make sure it isn't mentioned in /etc/fstab (maybe by uuid) and rerun mkinitrd. This should rebuild all initrds without any knowledge of the USB. Then you should be able to boot everything without the USB plugged in. So then boot the problematic kernel, plug in the USB, and try to see what happens. If the system locks up completely we will need to try to get information from sysrq. If it doesn't, then using 'ps' to find the process that has hung (probably in D state), and the "cat /proc/$PID/stack" to find out what it is doing. Hopefully some of those steps will get us nearer at least. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c27 --- Comment #27 from Walther Pelser --- Created attachment 640107 --> http://bugzilla.novell.com/attachment.cgi?id=640107&action=edit journalctl screenshot -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c29 --- Comment #29 from Walther Pelser --- Created attachment 640109 --> http://bugzilla.novell.com/attachment.cgi?id=640109&action=edit journalctl screenshot 3 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c31 --- Comment #31 from Walther Pelser --- Created attachment 640124 --> http://bugzilla.novell.com/attachment.cgi?id=640124&action=edit localmodconfig -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c33 --- Comment #33 from Walther Pelser --- Part of messages from a normal boot: 2015-07-05T12:07:05.493348+02:00 linux-z24t kernel: [ 2.644478] md: linear personality registered for level -1 2015-07-05T12:07:05.493350+02:00 linux-z24t kernel: [ 2.644484] md: raid0 personality registered for level 0 2015-07-05T12:07:05.493351+02:00 linux-z24t kernel: [ 2.644489] md: raid1 personality registered for level 1 Part of dracut emergency shell journalctl (taken from recovery mode.JPG): [ 7.559338] md: personality for level raid0 is not loaded! If this could not be loaded, partitions could not be read. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c35 --- Comment #35 from Walther Pelser --- Let me try to do some logical things. My kernel is now 4.1.1-2.g998afec-desktop from repositories/Kernel/Head/standard/ kernel-desktop-4.1.1-2.g998afec-x86_64.rpm . He is freezing. Running kernel is 4.1.1-1 desktop self comp. From kernel.org. 4.1.1 tar.xz If I compare /lib/modules/4.1.1-2.g998afec-desktop/kernel/drivers/md and /lib/modules/4.1.1-1-desktop/kernel/drivers/md , there are four drivers missing, because they are builtin (linear.ko and raid0.ko and raid1.ko and md-mod.ko), as I configured it and can find it in modules.builtin. (see comment #32) There are some added: dm-least-pending.ko and dm-log-writes.ko and dm-multipath.ko and dm-queue-length.ko and dm-service-time.ko. and md-cluster.ko . (due to config file) Takashi Iwai asked me, to change the config file, but I cannot see, how this should be helpful. The SUSE kernel kernel-deshtop-4.1.1-2.g998afec-x86_64.rpm is the generic one. And he should run in most cases also in mine too. What's about the line CONFIG_MD_AUTODETECT=y , which is missing in SUSE kernel , as you can see in comment #32 . Does this line have any function? This is the only difference, I could find. I would like to try a SUSE kernel with this line added. Possible? (I have to delete the SUSE kernel after such attempts, because grub2 tries always to resume with the highest kernel and this causes problems for me, as I often use the hibernate function.) -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c37 --- Comment #37 from Walther Pelser --- Last, but not .... I tried kernel 4.0.5-3.2, which comes up with tumbleweed 20150705-1.2. Same problems. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c39 --- Comment #39 from Walther Pelser --- Thank you for your answer. Now i got the information, you might have looked for!!!!!!!! First: it has nothing to do with the USB stick. I deleted the swap partition, from fstab too. I attached during boot nothing changed. But is has definitely something to do with the other swap partition! As it has something to with dracut and the SUSE version of it, I changed in /etc/dracut.conf.d/01-dist.conf the last line: # add_dracutmodules+=debug to add_dracutmodules+=debug. kernel 4.0.5-3.2 was installed with this new settings and booted and after some waiting, there came up the line, you can see on timeout.JPG( which is attached). After having downgraded YAST the partitioner is working again! So I can look into the partition table: The swap partition is part of an extended partition! SUSE kernel seems not to be able, to open this. Data of the extended partition shown by YAST(German version): Gerät: /dev/md126p2 Größe: 250.00 GiB Verschlüsselt: Nein Gerätepfad: Geräte-ID 1: md-uuid-b054006e:7e450d85:de257e42:4a16c54e-part2 FS-ID: 0x0F Extended Dateisystem: Dateisystem: Einhängepunkt: UUID: Kennung: Gerät: /dev/md126p5 Größe: 5.00 GiB Verschlüsselt: Nein Gerätepfad: Geräte-ID 1: md-uuid-b054006e:7e450d85:de257e42:4a16c54e-part5 FS-ID: 0x82 Linux swap Dateisystem: Dateisystem: Swap Einhängepunkt: swap UUID: df0343ee-74e3-4124-a8b2-7349d946e6a8 Kennung: swap Gerät: /dev/md126p6 Größe: 20.01 GiB Verschlüsselt: Nein Gerätepfad: Geräte-ID 1: md-uuid-b054006e:7e450d85:de257e42:4a16c54e-part6 FS-ID: 0x0C Win95 FAT32 LBA Dateisystem: Dateisystem: FAT Einhängepunkt: UUID: 2F8B-210D Kennung: Gerät: /dev/md126p7 Größe: 224.98 GiB Verschlüsselt: Nein Gerätepfad: Geräte-ID 1: md-uuid-b054006e:7e450d85:de257e42:4a16c54e-part7 FS-ID: 0x83 Linux native Dateisystem: Dateisystem: Ext4 Einhängepunkt: /home UUID: 70177251-9d78-4c95-ae66-7d829c8e6038 Kennung: home mp126p2 is mp126p5+mp126p6+mp126p7 So kernel has to add x2 in this case to the uuid, but there seems something to be wrong with it. The self compiled kernel has no problem with it, maybe I have builtin something? Or should be loaded to the SUSE kernel by dracut? I'm VERY interested in your reaction. Do you still need the initrd? In that case I have no idea, where to upload it, be maybe a compression should be possible? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c41 --- Comment #41 from Walther Pelser --- Attachment "Screenshot during "freeze"" is my "timeout.JPG". I do always expect, that the name of my file is copied too. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c43 --- Comment #43 from Walther Pelser --- Created attachment 640588 --> http://bugzilla.novell.com/attachment.cgi?id=640588&action=edit required data from comment 42 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c45 --- Comment #45 from Walther Pelser --- Hallo! There is another bug, I have filed. Probably there is a connection between them? Bug 938048. Here comment #6 from Stefan Hundhammer. In my eyes it could be a possible explanation of the reaction of the SUSE-kernel. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c46 --- Comment #46 from Neil Brown --- Sorry It's been so long ... I have had leave and a conference and ... I don't think the other bug is particularly related. parted should not be used during the boot process at all. My best guess is still that something is going wrong with the raid0.ko module - it isn't on the initrd. In my testing I cannot make this fail though. I assume you are booted with the RAID0 active when you install the new kernel??? The extra files you attached at my request all look perfectly normal and correct - I cannot see any problem there. So I'm at a loss .... does the problem still occur with the latest 4.3.0-rc4 kernel?? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c48 --- Comment #48 from Takashi Iwai --- (In reply to Walther Pelser from comment #47)

Thanks for your answer! 1. There are no "desktop"-kernels in repsitory/Kernel/Head or /Kernel/standard anymore. So I will wait or try a default-kernel later on.

Right, we deprecated kernel-desktop flavor. It's replaced with kernel-default now.

2. Now I am used to built my own kernel (act. version 4.1.10, localyesconfig, size 20,7 MB, built-time 27 min) and this kernel is always working fine (with some timing-problems caused by systemd). So it must be a SUSE-specific problem and not a raid0.ko one. But, as I seem to be the only one with it, it should be dropped or closed.

You should try localmodconfig instead of localyesconfig. We take only few SUSE-specific patches (except for ARM64 and Xen ones). The possible difference is rather whether it's a built-in or a module -- which influences on the timing. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c50 Neil Brown changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |trenn@suse.com --- Comment #50 from Neil Brown --- The contents of modules-load.d is not particularly relevant here. More relevant is the content of /lib/modules/4.3.rc4-2*-default/kernel/drivers/md/ This should have md-mod.ko and raid0.ko. If it doesn't (which I suspect to be the case), then that is the problem. You can easily test by running e.g. lsinitrd =boot/initrd-4.3.rc4-2*-default | grep raid0 .... Ah, I think I might know what is going wrong. You are booted with your all-yes-config kernel when you install the openSUSE kernel, aren't you? dracut (which builds the initrd) has some slightly complicated logic to decide whether to actually include a module in the initrd or not. The 'mdraid' dracut module asks the 'instmods' function (in /usr/lib/dracut somewhere) to install all modules in drivers/md/. instmods tries to be clever. For example if each module is currently loaded in the running kernel, it will be added to the initrd. If it could be loaded, but currently isn't, then it is not added to the initrd. Then there is some fairly complicated logic that only applies to modules with aliases (and raid0.ko does have aliases). I'm not 100% sure, but I think this logic is making unfounded assumptions (it is the checks in host_modalias[] in modules_is_host_only() in dracut_functions.sh) In any case, we can bypass all this logic by passing the "-A" flag to mkinitrd. This suppresses the 'hostonly' checks.... or at least that it what the documentation says. It seems that it lies. But we can call dracut directly So: as one final test, can you please run dracut -N --force /boot/initrd-$VERSION $VERSION where $VERSION is your kernel version, presumably 4.3.0-rc4-2.gefc53bd-default and then try to boot the openSUSE kernel. If you don't want to that is fine. I've managed to duplicate the bug now so I'm sure I can get it fixed. But having confirmation from you that this fixes your problem would be nice. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c52 --- Comment #52 from Walther Pelser --- So I did: (See latest attachment) But: It didn't boot! You are on the right way, I think. I noticed earlier, that dracut had changed all my older SUSE-kernels. In the beginning they were booting fine. Then I installed my own kernel and I could boot all my installed kernels. But one day there came a new dracut version or something else, I noticed that all initrds from all installed kernels were rewritten. From that time I could only boot with my own kernel. No older SUSE-kernel even after deleting and re-installation did work anymore. And no new SUSE-kernel did work too. And I also tried to rewrite dracut with force by choosing the SUSE-kernel. This did not change anything. So I could offer to delete all kernels from /boot and then install only the one mentioned above. It should work because dracut can no longer look into other kernels. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c54 --- Comment #54 from Neil Brown --- That is disappointing, and surprising. Deleting other kernels wouldn't help. dracut isn't looking into other kernels, it is looking into the currently running kernel. What does lsinitrd /boot/initrd-4.3.0-rc4-2.gefc53bd-default | grep drivers/md/raid show? I get raid0.ko raid1.ko raid10.ko raid456.ko listed. If you don't get that, then the "-N" flag to dracut isn't working. If you do get those, then there must be some other problem, other than not being able to load the raid0 module. We would need to get logs from the failed boot attempt like you did for comment #33. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c56 --- Comment #56 from Walther Pelser --- Now a normal comment My Os was broken. So I had to install a new tumbleweed-OS. As expected it is booting normally. Because of problems with my DVB-USB-Stick after hibernation I have to use again a self-compiled kernel with built-in module and built-in firmware. Which dracut-loglevel should I use, to get all changes in dracut. if there is there are changes during installation of a new kernel, while my own kernel is booted. It is obvious for me that dracut works fine, if a new own kernel is installed from a booted SUSE-Kernel. But the problems came up, when a SUSE-kernel was installed from a booted own kernel. So the SUSE-specifications in systemd and dracut may cause the problems. Dracut seems to have problems with them, when installing a new kernel, when the system was booted by my own kernel. So there should be some visible changes in the dracut-logs, if I have to to do so. The requested logs from comment 54 are impossible right now, but maybe seen sometime later on. I am open for another kind of approach to this problem, if You have a better one. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c59 --- Comment #59 from Walther Pelser --- Created attachment 653623 --> http://bugzilla.novell.com/attachment.cgi?id=653623&action=edit lsinitrd 1 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c61 --- Comment #61 from Walther Pelser --- Created attachment 653626 --> http://bugzilla.novell.com/attachment.cgi?id=653626&action=edit lsinitrd 3 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c63 --- Comment #63 from Walther Pelser --- Workaround is: dracut --kver 4.2.1-1-desktop --add-drivers raid0 --force This adds the dropped riad0.ko and I can get a bootable SUSE-kernel. I will add this to dracut.conf to avoid that this will be lost again, when initramrd is updated or another new kernel is installed. But a patch, which solves this bug, would be the better solution, I think. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c66 --- Comment #66 from Bernhard Wiedemann --- This is an autogenerated message for OBS integration: This bug (935993) was mentioned in https://build.opensuse.org/request/show/343477 13.2 / dracut -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 Swamp Workflow Management changed: What |Removed |Added ---------------------------------------------------------------------------- Whiteboard|ibs:running:1458:moderate |ibs:running:1458:moderate | |obs:running:4200:moderate -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c68 Fabian Vogt changed: What |Removed |Added ---------------------------------------------------------------------------- Status|IN_PROGRESS |RESOLVED Resolution|--- |FIXED --- Comment #68 from Fabian Vogt --- Ok, thanks. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c69 --- Comment #69 from Swamp Workflow Management --- openSUSE-SU-2015:2022-1: An update that solves one vulnerability and has two fixes is now available. Category: security (moderate) Bug References: 935338,935993,952491 CVE References: CVE-2015-0794 Sources used: openSUSE 13.2 (src): dracut-037-17.30.1 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=935993 http://bugzilla.novell.com/show_bug.cgi?id=935993#c70 --- Comment #70 from Swamp Workflow Management --- SUSE-SU-2015:2065-1: An update that solves one vulnerability and has three fixes is now available. Category: security (moderate) Bug References: 935338,935993,947518,952491 CVE References: CVE-2015-0794 Sources used: SUSE Linux Enterprise Server 12 (src): dracut-037-51.17.3 SUSE Linux Enterprise Desktop 12 (src): dracut-037-51.17.3 -- You are receiving this mail because: You are on the CC list for the bug.