Message ID | 20220415215640.2177403-1-lucas.demarchi@intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | [CI] HAX: drm/i915: Add softdep | expand |
On Fri, Apr 15, 2022 at 02:56:40PM -0700, Lucas De Marchi wrote: >CI is having several issues due to unbinding i915 while the sound driver >is still doing something. It's particularly harmful when it's not idle >and we attempt to remove i915. > >We still need to fix the snd driver so we can more reliably remove i915, >but this can't block other i915 tests. By adding a post softdep we make sure >that on module removal, kmod first removes snd_hda_intel, and only after >that removes i915. > >Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> ignore, it broke other platforms. Lucas De Marchi
diff --git a/drivers/gpu/drm/i915/i915_module.c b/drivers/gpu/drm/i915/i915_module.c index 65acd7bf75d0..4e3725a9c375 100644 --- a/drivers/gpu/drm/i915/i915_module.c +++ b/drivers/gpu/drm/i915/i915_module.c @@ -123,3 +123,5 @@ MODULE_AUTHOR("Intel Corporation"); MODULE_DESCRIPTION(DRIVER_DESC); MODULE_LICENSE("GPL and additional rights"); + +MODULE_SOFTDEP("post: snd_hda_intel");
CI is having several issues due to unbinding i915 while the sound driver is still doing something. It's particularly harmful when it's not idle and we attempt to remove i915. We still need to fix the snd driver so we can more reliably remove i915, but this can't block other i915 tests. By adding a post softdep we make sure that on module removal, kmod first removes snd_hda_intel, and only after that removes i915. Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> --- From some local tests I had success when making sure sound driver got unloaded before unloading i915. Submitting for CI and if nothing explodes we may consider it for topic/core-for-CI. drivers/gpu/drm/i915/i915_module.c | 2 ++ 1 file changed, 2 insertions(+)