| Message ID | 20190617113644.25621-1-amadeuszx.slawinski@linux.intel.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Fix driver reload issues | expand |
On Mon, Jun 17, 2019 at 01:36:33PM +0200, Amadeusz Sławiński wrote: > Hi, > > This series of patches introduces fixes to various issues found while > trying to unload all snd* modules and then loading them again. This > allows for modules to be really _modules_ and be unloaded and loaded on > demand, making it easier to develop and test them without constant > system reboots. Pierre? You did comment on the general concept in one of the patches but not on any of the patches directly.
On 6/25/19 7:04 AM, Mark Brown wrote: > On Mon, Jun 17, 2019 at 01:36:33PM +0200, Amadeusz Sławiński wrote: >> Hi, >> >> This series of patches introduces fixes to various issues found while >> trying to unload all snd* modules and then loading them again. This >> allows for modules to be really _modules_ and be unloaded and loaded on >> demand, making it easier to develop and test them without constant >> system reboots. > > Pierre? You did comment on the general concept in one of the patches > but not on any of the patches directly. I did review the patches internally and the v1. For the v2 I could only do an airport lounge review and didn't see any blatant issues, so feel free to take the following tag for the series. Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Hi, This series of patches introduces fixes to various issues found while trying to unload all snd* modules and then loading them again. This allows for modules to be really _modules_ and be unloaded and loaded on demand, making it easier to develop and test them without constant system reboots. There are some fixes in flow, either we don't initialize things before cleaning them up, clean up in wrong places or don't clean up at all. Other patches fix memory management problems, mostly things are not being freed. And finally there is few miscellaneous patches, please refer to specific patches to see what they do. This series was tested on SKL, BXT, GLK & KBL. Changes from previous patchset: * followed suggetion by Pierre in "ALSA: hdac: Fix codec name after machine driver is unloaded and reloaded" * dropped patches which were merged Amadeusz Sławiński (11): ASoC: Intel: Skylake: Initialize lists before access so they are safe to use ALSA: hdac: Fix codec name after machine driver is unloaded and reloaded ASoC: compress: Fix memory leak from snd_soc_new_compress ASoC: Intel: Skylake: Don't return failure on machine driver reload ASoC: Intel: Skylake: Remove static table index when parsing topology ASoC: Intel: Skylake: Add function to cleanup debugfs interface ASoC: Intel: Skylake: Properly cleanup on component removal ASoC: Intel: Skylake: Fix NULL ptr dereference when unloading clk dev ASoC: Intel: hdac_hdmi: Set ops to NULL on remove ASoC: topology: Consolidate how dtexts and dvalues are freed ASoC: topology: Consolidate and fix asoc_tplg_dapm_widget_*_create flow sound/hda/ext/hdac_ext_bus.c | 8 +- sound/soc/codecs/hdac_hdmi.c | 6 ++ sound/soc/intel/skylake/skl-debug.c | 9 ++ sound/soc/intel/skylake/skl-pcm.c | 16 ++-- sound/soc/intel/skylake/skl-ssp-clk.c | 16 ++-- sound/soc/intel/skylake/skl-topology.c | 50 ++++++----- sound/soc/intel/skylake/skl-topology.h | 2 + sound/soc/intel/skylake/skl.c | 7 +- sound/soc/intel/skylake/skl.h | 5 ++ sound/soc/soc-compress.c | 17 ++-- sound/soc/soc-topology.c | 114 ++++++++++++------------- 11 files changed, 136 insertions(+), 114 deletions(-)