#14 systemd systems doesn't have /var/log/dmesg
Closed: Fixed None Opened 9 years ago by sakjur.

Steps to reproduce:

  1. Use Fedora 20
  2. Run fpaste --sysinfo --printonly
  3. Look at the DRM Information in the printout.

It says something like:

  • DRM Information (failed: "grep drm /var/log/dmesg"):
    N/A

If you replace the command with dmesg | grep drm, you'll get the expected result:

  • DRM Information (dmesg | grep drm):
    [ 214.046888] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [ 2096.590778] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [ 4169.250910] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [ 4677.899524] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [ 7049.880531] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [ 8822.565811] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [13149.005059] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [17028.757253] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [19414.354393] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [23094.290338] microcode joydev iwlwifi rtsx_pci serio_raw cfg80211 btusb bluetooth lpc_ich i2c_i801 mfd_core shpchp mei_me mei wmi rfkill binfmt_misc dm_crypt crct10dif_pclmul crc32_pclmul crc32c_intel i915 r8169 i2c_algo_bit ghash_clmulni_intel drm_kms_helper drm mii i2c_core video
    [23610.198967] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [25919.827457] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [27815.402933] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [29337.074406] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [37751.849726] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [40607.242272] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [40955.208557] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
    [48046.170046] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off

I have attached a patch for this.


This bug is related https://bugzilla.redhat.com/show_bug.cgi?id=1172857 (I didn't look here to see if it is reported already, just looked at bugzilla (will do next time (sorry))) Other issue which is not described here I can file another ticket if needed?

Just to comment on the patch (not an expert!), I suggested the same workaround in my bug report, but as I said there I think 'dmesg' is reading from limited size "buffer", so in case there is a lot of kernel messages, some important part can be lost in that method. (and mostly 'drm' stuff is initialized at "early boot", possibly some errors can appear later ...), and I used 'journalctl -k -b' to lists kernel log, but no luck to get the same output as dmesg (maybe I miss some extra 'journalctl' switch?)

Replying to [comment:2 bitlord]:

Just to comment on the patch (not an expert!), I suggested the same workaround in my bug report, but as I said there I think 'dmesg' is reading from limited size "buffer", so in case there is a lot of kernel messages, some important part can be lost in that method. (and mostly 'drm' stuff is initialized at "early boot", possibly some errors can appear later ...), and I used 'journalctl -k -b' to lists kernel log, but no luck to get the same output as dmesg (maybe I miss some extra 'journalctl' switch?)

About this 'journalctl -k' (it automatically adds '-b' for current session/boot (journalctl(1) man page))
And maybe that is alternative on f21+ (or even f20+?) ( journalctl -k | grep drm ).
It didin't work for me probably because of some issue with systemd(systemd-journald) (tracked here https://bugzilla.redhat.com/show_bug.cgi?id=1169269 )

Fixed in 0.3.7.4. Closing.

dmesg works, by the way, /var/log/dmesg doesn't exist.

Login to comment on this ticket.

Metadata