Re: [PATCH v2 00/34] component: Make into an aggregate bus

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 07.10.2021 13:22, Andy Shevchenko wrote:


On Wednesday, October 6, 2021, Stephen Boyd <swboyd@xxxxxxxxxxxx <mailto:swboyd@xxxxxxxxxxxx>> wrote:

    This series is from discussion we had on reordering the device lists for
    drm shutdown paths[1]. I've introduced an 'aggregate' bus that we put
    the aggregate device onto and then we probe the aggregate device once
    all the components are probed and call component_add(). The probe/remove
    hooks are where the bind/unbind calls go, and then a shutdown hook is
    added that can be used to shutdown the drm display pipeline at the right
    time.

    This works for me on my sc7180 board. I no longer get a warning from i2c
    at shutdown that we're trying to make an i2c transaction after the i2c
    bus has been shutdown. There's more work to do on the msm drm driver to
    extract component device resources like clks, regulators, etc. out of
    the component bind function into the driver probe but I wanted to move
    everything over now in other component drivers before tackling that
    problem.

    I'll definitely be sending a v3 so this is partially a request for
    testing to shake out any more problems. Tested-by tags would be
    appreciated,
    and Acked-by/Reviewed-by tags too. I sent this to gregkh which may be
    incorrect but I don't know what better tree to send it all through.
    Maybe drm?

    I'll be faster at resending this next time, sorry for the long delay!


Yet another avoidance of mathematically proven device dependency graph...


Can we actually find and ask a mathematician to look into the problem and suggest real solution instead of all these ugly hacks: deferred probe (ugliest hack, how it even came into kernel?), component framework, custom approaches on how to see if devices are in the system (ASoC hack).

Master of Science in Math is OK? :)
The solution is simple - topological sort - this is what PM device list does, but it is possible only if there are no circular dependencies.

And in this case we have circular dependency.
In such case the solution in case of initialization is to split it to two phases - probe/bind (component framework), probe/attach (DSI bus) and it works. But PM still relies on topological sort. To make PM working with circular deps you can try to extend PM somehow (maybe replace list of devices with list of pairs (device, phase) ??? ), or to add notifiers to frameworks allowing providers to informs its consumers about state change. The latter approach I have advertised several times, even on OSS.

[1]: https://static.sched.com/hosted_files/osseu18/0f/deferred_problem.pdf
[2]: https://lwn.net/Articles/625454/


Regards
Andrzej



    Changes since v1
    (https://lore.kernel.org/r/20210520002519.3538432-1-swboyd@xxxxxxxxxxxx
    <https://lore.kernel.org/r/20210520002519.3538432-1-swboyd@xxxxxxxxxxxx>):
      - Use devlink to connect components to the aggregate device
      - Don't set the registering device as a parent of the aggregate device
      - New patch for bind_component/unbind_component ops that takes the
        aggregate device
      - Convert all drivers in the tree to use the aggregate driver approach
      - Allow one aggregate driver to be used for multiple aggregate devices

    [1]
    https://lore.kernel.org/r/20210508074118.1621729-1-swboyd@xxxxxxxxxxxx
    <https://lore.kernel.org/r/20210508074118.1621729-1-swboyd@xxxxxxxxxxxx>

    Stephen Boyd (34):
       component: Introduce struct aggregate_device
       component: Introduce the aggregate bus_type
       component: Move struct aggregate_device out to header file
       drm/msm: Migrate to aggregate driver
       component: Add {bind,unbind}_component() ops that take aggregate
         device
       drm/of: Add a drm_of_aggregate_probe() API
       drm/komeda: Migrate to aggregate driver
       drm/arm/hdlcd: Migrate to aggregate driver
       drm/malidp: Migrate to aggregate driver
       drm/armada: Migrate to aggregate driver
       drm/etnaviv: Migrate to aggregate driver
       drm/kirin: Migrate to aggregate driver
       drm/exynos: Migrate to aggregate driver
       drm/imx: Migrate to aggregate driver
       drm/ingenic: Migrate to aggregate driver
       drm/mcde: Migrate to aggregate driver
       drm/mediatek: Migrate to aggregate driver
       drm/meson: Migrate to aggregate driver
       drm/omap: Migrate to aggregate driver
       drm/rockchip: Migrate to aggregate driver
       drm/sti: Migrate to aggregate driver
       drm/sun4i: Migrate to aggregate driver
       drm/tilcdc: Migrate to aggregate driver
       drm/vc4: Migrate to aggregate driver
       drm/zte: Migrate to aggregate driver
       iommu/mtk: Migrate to aggregate driver
       mei: Migrate to aggregate driver
       power: supply: ab8500: Migrate to aggregate driver
       fbdev: omap2: Migrate to aggregate driver
       sound: hdac: Migrate to aggregate driver
       ASoC: codecs: wcd938x: Migrate to aggregate driver
       component: Get rid of drm_of_component_probe()
       component: Remove component_master_ops and friends
       component: Remove all references to 'master'

    Cc: Arnd Bergmann <arnd@xxxxxxxx <mailto:arnd@xxxxxxxx>>
    Cc: Benjamin Gaignard <benjamin.gaignard@xxxxxxxxxx
    <mailto:benjamin.gaignard@xxxxxxxxxx>>
    Cc: Chen Feng <puck.chen@xxxxxxxxxxxxx <mailto:puck.chen@xxxxxxxxxxxxx>>
    Cc: Chen-Yu Tsai <wens@xxxxxxxx <mailto:wens@xxxxxxxx>>
    Cc: Christian Gmeiner <christian.gmeiner@xxxxxxxxx
    <mailto:christian.gmeiner@xxxxxxxxx>>
    Cc: Chun-Kuang Hu <chunkuang.hu@xxxxxxxxxx
    <mailto:chunkuang.hu@xxxxxxxxxx>>
    Cc: Daniel Vetter <daniel.vetter@xxxxxxxx
    <mailto:daniel.vetter@xxxxxxxx>>
    Cc: Emma Anholt <emma@xxxxxxxxxx <mailto:emma@xxxxxxxxxx>>
    Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx
    <mailto:gregkh@xxxxxxxxxxxxxxxxxxx>>
    Cc: "Heiko Stübner" <heiko@xxxxxxxxx <mailto:heiko@xxxxxxxxx>>
    Cc: Inki Dae <inki.dae@xxxxxxxxxxx <mailto:inki.dae@xxxxxxxxxxx>>
    Cc: James Qian Wang (Arm Technology China) <james.qian.wang@xxxxxxx
    <mailto:james.qian.wang@xxxxxxx>>
    Cc: Jaroslav Kysela <perex@xxxxxxxx <mailto:perex@xxxxxxxx>>
    Cc: Joerg Roedel <joro@xxxxxxxxxx <mailto:joro@xxxxxxxxxx>>
    Cc: John Stultz <john.stultz@xxxxxxxxxx <mailto:john.stultz@xxxxxxxxxx>>
    Cc: Joonyoung Shim <jy0922.shim@xxxxxxxxxxx
    <mailto:jy0922.shim@xxxxxxxxxxx>>
    Cc: Jyri Sarha <jyri.sarha@xxxxxx <mailto:jyri.sarha@xxxxxx>>
    Cc: Kai Vehmanen <kai.vehmanen@xxxxxxxxxxxxxxx
    <mailto:kai.vehmanen@xxxxxxxxxxxxxxx>>
    Cc: Kyungmin Park <kyungmin.park@xxxxxxxxxxx
    <mailto:kyungmin.park@xxxxxxxxxxx>>
    Cc: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx
    <mailto:laurent.pinchart@xxxxxxxxxxxxxxxx>>
    Cc: <linux-fbdev@xxxxxxxxxxxxxxx <mailto:linux-fbdev@xxxxxxxxxxxxxxx>>
    Cc: <linux-omap@xxxxxxxxxxxxxxx <mailto:linux-omap@xxxxxxxxxxxxxxx>>
    Cc: <linux-pm@xxxxxxxxxxxxxxx <mailto:linux-pm@xxxxxxxxxxxxxxx>>
    Cc: Liviu Dudau <liviu.dudau@xxxxxxx <mailto:liviu.dudau@xxxxxxx>>
    Cc: Lucas Stach <l.stach@xxxxxxxxxxxxxx <mailto:l.stach@xxxxxxxxxxxxxx>>
    Cc: Mark Brown <broonie@xxxxxxxxxx <mailto:broonie@xxxxxxxxxx>>
    Cc: Maxime Ripard <mripard@xxxxxxxxxx <mailto:mripard@xxxxxxxxxx>>
    Cc: Neil Armstrong <narmstrong@xxxxxxxxxxxx
    <mailto:narmstrong@xxxxxxxxxxxx>>
    Cc: Paul Cercueil <paul@xxxxxxxxxxxxxxx <mailto:paul@xxxxxxxxxxxxxxx>>
    Cc: Philipp Zabel <p.zabel@xxxxxxxxxxxxxx
    <mailto:p.zabel@xxxxxxxxxxxxxx>>
    Cc: "Rafael J. Wysocki" <rafael@xxxxxxxxxx <mailto:rafael@xxxxxxxxxx>>
    Cc: Rob Clark <robdclark@xxxxxxxxx <mailto:robdclark@xxxxxxxxx>>
    Cc: Russell King <linux@xxxxxxxxxxxxxxx <mailto:linux@xxxxxxxxxxxxxxx>>
    Cc: Russell King <linux+etnaviv@xxxxxxxxxxxxxxx
    <mailto:linux+etnaviv@xxxxxxxxxxxxxxx>>
    Cc: Russell King <rmk+kernel@xxxxxxxxxxxxxxxx
    <mailto:rmk+kernel@xxxxxxxxxxxxxxxx>>
    Cc: Sandy Huang <hjc@xxxxxxxxxxxxxx <mailto:hjc@xxxxxxxxxxxxxx>>
    Cc: Saravana Kannan <saravanak@xxxxxxxxxx <mailto:saravanak@xxxxxxxxxx>>
    Cc: Sebastian Reichel <sre@xxxxxxxxxx <mailto:sre@xxxxxxxxxx>>
    Cc: Seung-Woo Kim <sw0312.kim@xxxxxxxxxxx
    <mailto:sw0312.kim@xxxxxxxxxxx>>
    Cc: Takashi Iwai <tiwai@xxxxxxxx <mailto:tiwai@xxxxxxxx>>
    Cc: Tian Tao <tiantao6@xxxxxxxxxxxxx <mailto:tiantao6@xxxxxxxxxxxxx>>
    Cc: Tomas Winkler <tomas.winkler@xxxxxxxxx
    <mailto:tomas.winkler@xxxxxxxxx>>
    Cc: Tomi Valkeinen <tomba@xxxxxxxxxx <mailto:tomba@xxxxxxxxxx>>
    Cc: Will Deacon <will@xxxxxxxxxx <mailto:will@xxxxxxxxxx>>
    Cc: Xinliang Liu <xinliang.liu@xxxxxxxxxx
    <mailto:xinliang.liu@xxxxxxxxxx>>
    Cc: Xinwei Kong <kong.kongxinwei@xxxxxxxxxxxxx
    <mailto:kong.kongxinwei@xxxxxxxxxxxxx>>
    Cc: Yong Wu <yong.wu@xxxxxxxxxxxx <mailto:yong.wu@xxxxxxxxxxxx>>

      drivers/base/component.c                      | 555 +++++++++++-------
      .../gpu/drm/arm/display/komeda/komeda_drv.c   |  20 +-
      drivers/gpu/drm/arm/hdlcd_drv.c               |  21 +-
      drivers/gpu/drm/arm/malidp_drv.c              |  21 +-
      drivers/gpu/drm/armada/armada_drv.c           |  23 +-
      drivers/gpu/drm/drm_drv.c                     |   2 +-
      drivers/gpu/drm/drm_of.c                      |  20 +-
      drivers/gpu/drm/etnaviv/etnaviv_drv.c         |  20 +-
      drivers/gpu/drm/exynos/exynos_drm_drv.c       |  21 +-
      .../gpu/drm/hisilicon/kirin/kirin_drm_drv.c   |  20 +-
      drivers/gpu/drm/imx/imx-drm-core.c            |  20 +-
      drivers/gpu/drm/ingenic/ingenic-drm-drv.c     |  24 +-
      drivers/gpu/drm/mcde/mcde_drv.c               |  23 +-
      drivers/gpu/drm/mediatek/mtk_drm_drv.c        |  20 +-
      drivers/gpu/drm/meson/meson_drv.c             |  21 +-
      drivers/gpu/drm/msm/msm_drv.c                 |  46 +-
      drivers/gpu/drm/omapdrm/dss/dss.c             |  17 +-
      drivers/gpu/drm/rockchip/rockchip_drm_drv.c   |  20 +-
      drivers/gpu/drm/sti/sti_drv.c                 |  20 +-
      drivers/gpu/drm/sun4i/sun4i_drv.c             |  26 +-
      drivers/gpu/drm/tilcdc/tilcdc_drv.c           |  28 +-
      drivers/gpu/drm/vc4/vc4_drv.c                 |  20 +-
      drivers/gpu/drm/zte/zx_drm_drv.c              |  20 +-
      drivers/iommu/mtk_iommu.c                     |  14 +-
      drivers/iommu/mtk_iommu.h                     |   6 +-
      drivers/iommu/mtk_iommu_v1.c                  |  14 +-
      drivers/misc/mei/hdcp/mei_hdcp.c              |  22 +-
      drivers/power/supply/ab8500_charger.c         |  22 +-
      drivers/video/fbdev/omap2/omapfb/dss/dss.c    |  20 +-
      include/drm/drm_of.h                          |   9 +-
      include/linux/component.h                     |  92 ++-
      sound/hda/hdac_component.c                    |  21 +-
      sound/soc/codecs/wcd938x.c                    |  20 +-
      33 files changed, 780 insertions(+), 488 deletions(-)


    base-commit: e4e737bb5c170df6135a127739a9e6148ee3da82
-- https://chromeos.dev <https://chromeos.dev>



--
With Best Regards,
Andy Shevchenko






[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux