[OE-core] [PATCH 2/2] parselogs.py: whitelist more xserver related error

Richard Purdie richard.purdie at linuxfoundation.org
Wed Mar 11 08:26:28 UTC 2020


On Wed, 2020-03-11 at 10:13 +0800, changqing.li at windriver.com wrote:
> From: Changqing Li <changqing.li at windriver.com>
> 
> With default config, these errors always exist for
> core-image-sato, and xserver-nodm.server start successfully,
> these errors are not critially.
> 
> If set default syslog to rsyslog, all these errors
> will go into daemon.log and user.log, then test_parselog
> will fail, so whitelist these errors
> 
> Signed-off-by: Changqing Li <changqing.li at windriver.com>
> ---
>  meta/lib/oeqa/runtime/cases/parselogs.py | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/meta/lib/oeqa/runtime/cases/parselogs.py
> b/meta/lib/oeqa/runtime/cases/parselogs.py
> index 6e22520..2a5e5c6 100644
> --- a/meta/lib/oeqa/runtime/cases/parselogs.py
> +++ b/meta/lib/oeqa/runtime/cases/parselogs.py
> @@ -58,7 +58,15 @@ common_errors = [
>      "Failed to rename network interface",
>      "Failed to process device, ignoring: Device or resource busy",
>      "Cannot find a map file",
> -    "[rdrand]: Initialization Failed"
> +    "[rdrand]: Initialization Failed",
> +    "libGL error:",
> +    "[pulseaudio] authkey.c: Failed to open cookie file",
> +    "[pulseaudio] authkey.c: Failed to load authentication key",
> +    "FBIOPUT_VSCREENINFO failed, double buffering disabled",
> +    "xserver-nodm.service: Failed with result 'exit-code'",
> +    "xinit: server error",
> +    "matchbox-wm: X error warning",
> +    "X11 cannot support keycodes above 255"
>      ]
>  
>  video_related = [
> @@ -86,6 +94,8 @@ qemux86_common = [
>      'tsc: HPET/PMTIMER calibration failed',
>      "modeset(0): Failed to initialize the DRI2 extension",
>      "glamor initialization failed",
> +    "xf86OpenConsole: Switching VT failed",
> +    "Fatal server error:",

"Fatal server error" means the server didn't start correctly? I doubt
we want to whitelist that one.

With several of these you're right, they have been on the console for a
while and I'm worried that a) we don't see them in the logs and b) the
errors are occurring at all, we should really fix some of them, not
"hide" them.

Can we fix any of them rather than whitelisting?

If we really do want to "hide" them, I think at the very least we need
open bugs for some of them so we can track and fix the underlying
problems. We need the bugs open before any patch can merge.

Cheers,

Richard






More information about the Openembedded-core mailing list