[Bug 871817] New: kernel 3.14.0 very very sluggys (first time command run, has a huge delay)
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c0 Summary: kernel 3.14.0 very very sluggys (first time command run, has a huge delay) Classification: openSUSE Product: openSUSE 13.1 Version: Final Platform: x86-64 OS/Version: openSUSE 13.1 Status: ASSIGNED Severity: Major Priority: P5 - None Component: Kernel AssignedTo: jslaby@suse.com ReportedBy: rmilasan@suse.com QAContact: qa-bugs@suse.de Found By: L3 Blocker: --- I'm using kernel 3.14.0 from Kernel:stable/standard OBS repo and since yesterday, when I first install it, all commands have a 'strange' delay when running them: robert@viper:~> time lsusb <-- first run Bus 001 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub Bus 002 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 003: ID 046d:c05a Logitech, Inc. M90/M100 Optical Mouse Bus 002 Device 004: ID 413c:2107 Dell Computer Corp. real 0m3.406s user 0m0.090s sys 0m0.002s robert@viper:~> time lsusb <-- second run Bus 001 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub Bus 002 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 003: ID 046d:c05a Logitech, Inc. M90/M100 Optical Mouse Bus 002 Device 004: ID 413c:2107 Dell Computer Corp. real 0m0.090s user 0m0.086s sys 0m0.003s robert@viper:~> time ls <-- first run android bin Desktop Documents Downloads git ISO opensuse patches Pictures programming rpmbuild tests upstream VM vmware real 0m1.347s user 0m0.001s sys 0m0.000s robert@viper:~> time ls <-- second run android bin Desktop Documents Downloads git ISO opensuse patches Pictures programming rpmbuild tests upstream VM vmware real 0m0.002s user 0m0.002s sys 0m0.000s The delay happens when you run the command or when the commands ends it's process. Tried to strace this to see what happens, but seems to happen before even strace catches the first syscall or signals. The slow down is a a second a more, but like I said, don't see anything in the trace that would tell me from where is coming from. P.S. sorry for the poor explanation, but is really hard to say what exactly happens. What has been updated/upgraded: robert@viper:~> rpm -qa --last|head -n 20 kernel-desktop-devel-3.14.0-2.1.gfa168d7.x86_64 Thu 03 Apr 2014 09:38:19 AM CEST kernel-devel-3.14.0-2.1.gfa168d7.noarch Thu 03 Apr 2014 09:38:16 AM CEST kernel-desktop-3.14.0-2.1.gfa168d7.x86_64 Thu 03 Apr 2014 09:37:30 AM CEST libkpathsea6-6.2.0dev-11.6.3.x86_64 Wed 02 Apr 2014 02:26:37 PM CEST gnome-keyring-pam-32bit-3.10.1-6.1.x86_64 Mon 31 Mar 2014 01:11:48 PM CEST gnome-keyring-pam-3.10.1-6.1.x86_64 Mon 31 Mar 2014 01:11:48 PM CEST gnome-keyring-3.10.1-6.1.x86_64 Mon 31 Mar 2014 01:11:46 PM CEST libgck-modules-gnome-keyring-3.10.1-6.1.x86_64 Mon 31 Mar 2014 01:11:45 PM CEST icewm-1.3.7-26.4.1.x86_64 Mon 31 Mar 2014 01:11:44 PM CEST VirtualBox-4.3-4.3.10_93012_openSUSE123-1.x86_64 Thu 27 Mar 2014 03:39:58 PM CET kernel-firmware-20140319git-36.1.noarch Thu 27 Mar 2014 03:39:44 PM CET MozillaFirefox-28.0-17.1.x86_64 Thu 27 Mar 2014 03:35:49 PM CET mozilla-nss-certs-3.15.5-16.1.x86_64 Thu 27 Mar 2014 03:35:45 PM CET mozilla-nss-3.15.5-16.1.x86_64 Thu 27 Mar 2014 03:35:45 PM CET libsoftokn3-3.15.5-16.1.x86_64 Thu 27 Mar 2014 03:35:45 PM CET mozilla-nspr-4.10.4-8.1.x86_64 Thu 27 Mar 2014 03:35:44 PM CET libfreebl3-3.15.5-16.1.x86_64 Thu 27 Mar 2014 03:35:44 PM CET file-5.15-4.20.1.x86_64 Wed 26 Mar 2014 02:16:22 PM CET libmagic1-5.15-4.20.1.x86_64 Wed 26 Mar 2014 02:16:21 PM CET xinetd-2.3.15-2.4.1.x86_64 Wed 26 Mar 2014 02:16:20 PM CET NOTE: this happens both with 3.14.0-1.1 and 13.14.0-2.1 versions. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c1 --- Comment #1 from Jiri Slaby <jslaby@suse.com> 2014-04-03 11:34:34 UTC --- Cool, if you run htop (or top) in a second terminal, is there high I/O or CPU? Or something suspicious? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c2 --- Comment #2 from Robert Milasan <rmilasan@suse.com> 2014-04-03 11:46:25 UTC --- No, there seems to have no change, but on the other hand I have firefox, chrome and other apps running which use CPU and I/O, but still ls or lsusb or uname are slow in the terminal. I'll try in the console to see if there is any difference. BTW: if this helps, this is the output of cmdline: robert@viper:~> cat /proc/cmdline BOOT_IMAGE=/boot/vmlinuz-3.14.0-2.gfa168d7-desktop root=/dev/sda5 selinux=0 apparmor=0 resume=/dev/disk/by-uuid/987ef39a-a7eb-4c0c-a951-1a19f63e8293 splash=silent quiet rootfstype=ext4 libahci.ignore_sss=1 raid=noautodetect plymouth.enable=0 intel_idle.max_cstate=0 processor.max_cstate=0 idle=mwait nmi_watchdog=0 audit=0 net.ifnames=0 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c3 --- Comment #3 from Robert Milasan <rmilasan@suse.com> 2014-04-03 11:48:46 UTC --- Seems to happen in the console also, didn't happen with 'lsusb', 'free', but happened with 'uname -a' So here is how I've done it: login to the console as root # lsusb (runs fine, first run) # free (runs fine, first run) # uname -a (slow as hell, first run) Didn't run any other command or anything help. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c4 Robert Milasan <rmilasan@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P3 - Medium --- Comment #4 from Robert Milasan <rmilasan@suse.com> 2014-04-08 13:09:12 UTC --- BTW, something new also happens, seems that Firefox gets affected by this: robert@viper:~> strace -p 32415 Process 32415 attached futex(0x7fa31847214c, FUTEX_WAIT_PRIVATE, 322, NULL^CProcess 32415 detached Firefox gets stuck and needs to be killed. This happens already on 2 machines, initially was thinking is Firefox, but Firefox wasn't updated at all nor glibc or something related to Firefox. On the second machine (a Dell laptop), the commands don't seem to have the same issue as I initially reported, but I have the Firefox issue. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c5 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO InfoProvider| |rmilasan@suse.com --- Comment #5 from Jiri Slaby <jslaby@suse.com> 2014-04-14 20:11:58 UTC --- (In reply to comment #4)
futex(0x7fa31847214c, FUTEX_WAIT_PRIVATE, 322, NULL^CProcess 32415 detached
Hmm, 3.14.1 should fix that IMHO. Could you try after it gets build? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c6 Robert Milasan <rmilasan@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED InfoProvider|rmilasan@suse.com | --- Comment #6 from Robert Milasan <rmilasan@suse.com> 2014-04-16 06:40:50 UTC --- Jiri, the update didn't fix the issue, still behaves the same. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c7 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO InfoProvider| |rmilasan@suse.com --- Comment #7 from Jiri Slaby <jslaby@suse.com> 2014-04-16 08:37:38 UTC --- (In reply to comment #6)
Jiri, the update didn't fix the issue, still behaves the same.
Both firefox and the commandline? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c8 Robert Milasan <rmilasan@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED InfoProvider|rmilasan@suse.com | --- Comment #8 from Robert Milasan <rmilasan@suse.com> 2014-04-16 08:43:49 UTC --- Firefox didn't hang yet, but it doesn't hang all the time. The command line, yes, it's slow at least the first time. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c9 --- Comment #9 from Robert Milasan <rmilasan@suse.com> 2014-04-16 08:46:39 UTC --- Look at this: robert@viper:~> strace -T lspci execve("/sbin/lspci", ["lspci"], [/* 95 vars */]) = 0 <1.713959> brk(0) = 0xa6d000 <0.000019> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7fb1bcc1b000 <0.000014> access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) <0.000017> open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3 <0.000019> fstat(3, {st_mode=S_IFREG|0644, st_size=109626, ...}) = 0 <0.000013> mmap(NULL, 109626, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7fb1bcc00000 <0.000015> You see the time for execve? seem a bit too much, don't you think? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c10 --- Comment #10 from Robert Milasan <rmilasan@suse.com> 2014-04-16 08:49:11 UTC --- BTW, as a note here, I'm using 3.14 on a i386 machine, built by myself and don't see this issues at all. It's a small server, but haven't see any symptoms on that machine. On my other machine, the symptoms are not the same, the command line doesn't seem to be that affect, but still it's somewhat sluggish and Firefox hangs (haven't seen Firefox hang with kernel 3.14.1, but still there is time). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c11 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO InfoProvider| |rmilasan@suse.com --- Comment #11 from Jiri Slaby <jslaby@suse.com> 2014-04-16 09:38:04 UTC --- (In reply to comment #10)
BTW, as a note here, I'm using 3.14 on a i386 machine, built by myself and don't see this issues at all. It's a small server, but haven't see any symptoms on that machine.
But this is a different machine, right? I don't see the problem on my machine too. If it is the same machine and only the kernel is different, could you test with the -vanilla flavor of our kernels? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c12 Robert Milasan <rmilasan@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED InfoProvider|rmilasan@suse.com | --- Comment #12 from Robert Milasan <rmilasan@suse.com> 2014-04-16 10:02:01 UTC --- Just installed and tried kernel-vanilla and same thing. Is it possible some configuration issues? Something is enabled in the kernel which causes this? robert@viper:~> uname -a Linux viper.suse.cz 3.14.1-1.geafcebd-vanilla #1 SMP Mon Apr 14 20:10:59 UTC 2014 (eafcebd) x86_64 x86_64 x86_64 GNU/Linux One thing I've noticed is that now the slowness is at the end, meaning exiting the program or command. So when running 'time lsusb' for example, all is good, I'm getting real/user/sys info and looks good, but after 'sys' info is displayed by time command, it takes a while to actual finish (a couple of miliseconds to a second or two). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c13 --- Comment #13 from Robert Milasan <rmilasan@suse.com> 2014-04-16 11:44:04 UTC --- OK, got some good news at the moment. I've dropped some boot options: 'intel_idle.max_cstate=0 processor.max_cstate=0 idle=mwait' and now the system seems to work OK. No delays, no issues, works as expected. When idle=mwait option was set, I saw some error in dmesg: Malformed option 'idle' Checked the kernel code in arch/x86/kernel/process.c and seems mwait was dropped, only pool, nomwait and halt is possible. Got no idea why mwait was dropped, would have been nice to know :) Can idle option or the other cstate option create so many issues? I don't think the idle option has anything to do with the issue, as it was malformed, but the other 2 I'm sure was doing the nasty :) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c14 --- Comment #14 from Robert Milasan <rmilasan@suse.com> 2014-04-16 12:07:03 UTC --- I can confirm now that dropping 'intel_idle.max_cstate=0 processor.max_cstate=0' fixes the issue. I've tested kernel-vanilla and kernel-desktop, both work as expected (meaning no delays or sluggishness). At the moment I'm not sure which option 'intel_idle.max_cstate=0' or 'processor.max_cstate=0' is to blame. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c15 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO InfoProvider| |rmilasan@suse.com --- Comment #15 from Jiri Slaby <jslaby@suse.com> 2014-08-18 14:05:58 UTC --- Do you still see the issue with 3.16.1? Have you had a chance to investigate which option was to blame in the end? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c16 Robert Milasan <rmilasan@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED InfoProvider|rmilasan@suse.com | --- Comment #16 from Robert Milasan <rmilasan@suse.com> 2014-08-25 06:41:11 UTC --- At the moment I'm not seeing anything with 3.16.1, but to be totally honest my laptop is not exactly a reliable source of information, meaning it's a total s*it and need fixing or replacing. So lets close this as or INVALID or UPSTREAM, as u wish. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=871817 https://bugzilla.novell.com/show_bug.cgi?id=871817#c17 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |WORKSFORME --- Comment #17 from Jiri Slaby <jslaby@suse.com> 2014-08-25 06:55:54 UTC --- Rather works-for-me. If you happen to reproduce, feel free to reopen. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com