[Intel-xe] [PATCH i-g-t] tests/core_hotunplug: Wait for device nodes to re-appear

[Kamil Konieczny]

Hi Janusz,

On 2023-08-22 at 09:58:58 +0200, Janusz Krzysztofik wrote:
> Sometimes our health check is not able to find a device node (cardX or
> renderY) after driver rebind or device rediscover.  Since device nodes are
> recreated in detmpvfs asynchronously, give it a bit of time to do its job
> before we try to reopen the device for health check.
> 
> (core_hotunplug:3612) DEBUG: Test requirement passed: priv->fd.drm == -1
> (core_hotunplug:3612) DEBUG: reopening DRM device for health check
> (core_hotunplug:3612) drmtest-DEBUG: Looking for devices to open using filter 0: sys:/sys/devices/pci0000:4a/0000:4a:02.0/0000:4b:00.0/0000:4c:01.0/0000:4d:00.0
> (core_hotunplug:3612) drmtest-DEBUG: Filter matched /dev/dri/card1 | /dev/dri/renderD128
> (core_hotunplug:3612) drmtest-DEBUG: Opened previously opened device: /dev/dri/card1
> (core_hotunplug:3612) DEBUG: running device sysfs healthcheck
> (core_hotunplug:3612) DEBUG: closing health checked device instance
> (core_hotunplug:3612) DEBUG: reopening render device for health check
> (core_hotunplug:3612) drmtest-DEBUG: Looking for devices to open using filter 0: sys:/sys/devices/pci0000:4a/0000:4a:02.0/0000:4b:00.0/0000:4c:01.0/0000:4d:00.0
> (core_hotunplug:3612) drmtest-DEBUG: Filter matched /dev/dri/card1 | /dev/dri/renderD128
> (core_hotunplug:3612) drmtest-INFO: Opened device: /dev/dri/renderD128
> (core_hotunplug:3612) DEBUG: running device sysfs healthcheck
> (core_hotunplug:3612) DEBUG: closing health checked device instance
> (core_hotunplug:3612) DEBUG: Test requirement passed: healthcheck(priv, false)
> (core_hotunplug:3612) DEBUG: Test requirement passed: priv->fd.drm_hc == -1
> (core_hotunplug:3612) igt_kmod-DEBUG: Module snd_hda_intel unloaded immediately
> (core_hotunplug:3612) INFO: Unloaded audio driver snd_hda_intel
> (core_hotunplug:3612) DEBUG: unbinding the driver from the device
> (core_hotunplug:3612) DEBUG: rebinding the driver to the device
> (core_hotunplug:3612) INFO: Realoading snd_hda_intel
> (core_hotunplug:3612) DEBUG: reopening DRM device for health check
> (core_hotunplug:3612) drmtest-DEBUG: Looking for devices to open using filter 0: sys:/sys/devices/pci0000:4a/0000:4a:02.0/0000:4b:00.0/0000:4c:01.0/0000:4d:00.0
> (core_hotunplug:3612) drmtest-DEBUG: Filter matched /dev/dri/card1 |
> (core_hotunplug:3612) drmtest-DEBUG: Opened previously opened device: /dev/dri/card1
> (core_hotunplug:3612) DEBUG: running device sysfs healthcheck
> (core_hotunplug:3612) DEBUG: closing health checked device instance
> (core_hotunplug:3612) DEBUG: reopening render device for health check
> (core_hotunplug:3612) drmtest-DEBUG: Looking for devices to open using filter 0: sys:/sys/devices/pci0000:4a/0000:4a:02.0/0000:4b:00.0/0000:4c:01.0/0000:4d:00.0
> (core_hotunplug:3612) drmtest-DEBUG: Filter matched /dev/dri/card1 |
> (core_hotunplug:3612) CRITICAL: Test assertion failure function local_drm_open_driver, file ../../../usr/src/igt-gpu-tools/tests/core_hotunplug.c:130:
> (core_hotunplug:3612) CRITICAL: Failed assertion: fd_drm >= 0
> (core_hotunplug:3612) CRITICAL: file descriptor fd_drm failed
> 
> Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/8830
> Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/500
> Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik at linux.intel.com>
> Cc: Petri Latvala <adrinael at adrinael.net>

Reviewed-by: Kamil Konieczny <kamil.konieczny at linux.intel.com>

> ---
>  tests/core_hotunplug.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/tests/core_hotunplug.c b/tests/core_hotunplug.c
> index 30d9a6a576..b254447c71 100644
> --- a/tests/core_hotunplug.c
> +++ b/tests/core_hotunplug.c
> @@ -501,6 +501,9 @@ static void node_healthcheck(struct hotunplug *priv, unsigned flags)
>  
>  static bool healthcheck(struct hotunplug *priv, bool recover)
>  {
> +	/* give udev some time to recreate device nodes before we continue */
> +	sleep(1);
> +
>  	/* device name may have changed, rebuild IGT device list */
>  	igt_devices_scan(true);
>  
> -- 
> 2.41.0
>