[Bug 1047726] New: i2c error messages fill kernel log like crazy (needs rate limit)
http://bugzilla.suse.com/show_bug.cgi?id=1047726 Bug ID: 1047726 Summary: i2c error messages fill kernel log like crazy (needs rate limit) Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: x86-64 OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: kernel-maintainers@forge.provo.novell.com Reporter: gp@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- On a notebook I ran into an issue with the touchpad "hanging" and eating loads of resources: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 909 root -51 0 0 0 0 D 14.89 0.000 1:15.58 irq/51-SYNA2B29 631 root 20 0 12056 5120 1572 S 2.128 0.031 0:07.23 haveged 841 root -51 0 0 0 0 D 0.709 0.000 3:43.81 irq/87-ITE8396: 1939 gp 20 0 709120 147156 120792 S 0.709 0.901 3:36.15 X 2642 gp 20 0 639044 46020 31036 S 0.709 0.282 0:45.56 gnome-terminal- This report is NOT about that per se, but rather the fact that i2c issues error messages into the log at a crazy rate: 08:29:32 ghazala kernel: i2c_hid i2c-ITE8396:00: failed to set a report to device. 08:29:32 ghazala kernel: i2c_designware i2c_designware.0: timeout waiting for bus ready 08:29:32 ghazala kernel: i2c_designware i2c_designware.0: timeout waiting for bus ready 08:29:32 ghazala kernel: i2c_hid i2c-ITE8396:00: failed to set a report to device. 08:29:32 ghazala kernel: i2c_designware i2c_designware.0: timeout waiting for bus ready 08:29:32 ghazala kernel: i2c_designware i2c_designware.0: timeout waiting for bus ready 08:29:32 ghazala kernel: i2c_hid i2c-ITE8396:00: failed to retrieve report from device. 08:29:32 ghazala kernel: i2c_designware i2c_designware.0: t^C # journalctl --since today | grep 'i2c.*timeout' | wc -l 333455 # journalctl --since today | grep 'i2c.*timeout' | uniq | wc -l 8232 That is, 40 identical errors are logged. Per second. "Forever" (until I shut down the system). At a minimum, can this be rate limited (to no more than one per second, or better every 5-10 seconds)? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1047726 http://bugzilla.suse.com/show_bug.cgi?id=1047726#c1 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tiwai@suse.com --- Comment #1 from Takashi Iwai <tiwai@suse.com> --- Adding the rate-limit should be trivial in both i2c-hid and i2c-designware, just replace dev_*() with dev_*_ratelimited(). It's not guaranteed to work effectively, though, as it depends on the rates to be printed. I can give it a shot if you want, though. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1047726 http://bugzilla.suse.com/show_bug.cgi?id=1047726#c2 --- Comment #2 from Takashi Iwai <tiwai@suse.com> --- BTW, we should track the root cause of these messages, individually from the rate-limit workaround. Please open another bug report, and put the hardware information details (at least hwinfo output). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1047726 http://bugzilla.suse.com/show_bug.cgi?id=1047726#c3 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |gp@suse.com Flags| |needinfo?(gp@suse.com) --- Comment #3 from Takashi Iwai <tiwai@suse.com> --- A kernel with the test patches is being built in OBS home:tiwai:bnc1047726. It should finish building in an hour or so. Please test it later. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1047726 http://bugzilla.suse.com/show_bug.cgi?id=1047726#c4 Gerald Pfeifer <gp@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(gp@suse.com) | --- Comment #4 from Gerald Pfeifer <gp@suse.com> --- (In reply to Takashi Iwai from comment #3)
A kernel with the test patches is being built in OBS home:tiwai:bnc1047726. It should finish building in an hour or so. Please test it later.
Thank you, Takashi! I installed this on the machine in question last Saturday, but we did not encounter this issue since then. (I.e., we did "lose" trackpad+touchscreen again, alas with a different failure mode and without the i2c error message in the log.) This notebook usually is a friend's, not mine, and her see her about once a week and asked her to note down time and date of any issues and will study the logs regularly. Would it make sense to still proceed upstream, or do you want to wait for a repeat occurrence? (It is not deterministic.) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1047726 http://bugzilla.suse.com/show_bug.cgi?id=1047726#c5 --- Comment #5 from Takashi Iwai <tiwai@suse.com> --- Well, I'd prefer seeing that it really improves before upstreaming. The ratelimit doesn't work always as expected (although it should be effective in this case). So yes, please keep eyes on it, you should have a kernel message "xxx XX output lines suppressed due to ratelimiting" after the relevant errors. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1047726 http://bugzilla.suse.com/show_bug.cgi?id=1047726#c6 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |jslaby@suse.com Resolution|--- |FIXED --- Comment #6 from Jiri Slaby <jslaby@suse.com> --- As there are no more responses, I assume this is fixed by now? -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com