lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGXv+5EA9Ed2tyg2f3TSeXYdZOZ8a=VztaOc_UTFK2dEsENjJQ@mail.gmail.com>
Date:   Thu, 23 Feb 2023 18:15:28 +0800
From:   Chen-Yu Tsai <wenst@...omium.org>
To:     AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...labora.com>
Cc:     mturquette@...libre.com, sboyd@...nel.org, matthias.bgg@...il.com,
        johnson.wang@...iatek.com, miles.chen@...iatek.com,
        chun-jie.chen@...iatek.com, daniel@...rotopia.org,
        fparent@...libre.com, msp@...libre.com, nfraprado@...labora.com,
        rex-bc.chen@...iatek.com, zhaojh329@...il.com,
        sam.shih@...iatek.com, edward-jw.yang@...iatek.com,
        yangyingliang@...wei.com, granquet@...libre.com,
        pablo.sun@...iatek.com, sean.wang@...iatek.com,
        chen.zhong@...iatek.com, linux-kernel@...r.kernel.org,
        linux-clk@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org
Subject: Re: [PATCH v5 00/54] MediaTek clocks: full module build and cleanups

On Wed, Feb 22, 2023 at 5:25 PM AngeloGioacchino Del Regno
<angelogioacchino.delregno@...labora.com> wrote:
>
> Changes in v5:
>  - Fixed typo in MODULE_DEVICE_TABLE() for clk-mt8195-vpp0/vpp1
>
> Changes in v4:
>  - Rebased over clk-next branch as of 21/02/2023
>  - MT8135: Squashed removal of __initconst annotation with conversion
>    to platform drivers as suggested by Chen-Yu
>  - MT8135: Added mention of the introduction of the dummy clock in
>    the commit description of patch [54/54]
>  - Fixed Kconfig ordering issue for MT8195 IMGSYS dependant option
>
> Changes in v3:
>  - Added conversion to simple_probe for MT8135 clocks
>  - Reordered added Kconfig options for MT8195, MT8186
>  - Additional config options for MT8195, MT8186 are now default
>    enabled if parent COMMON_CLK_MT81{95,86} is enabled
>  - Added .remove() callback to MT2712 and MT7622 apmixedsys drivers
>    to avoid resource leakage on unbind/removal
>  - Dropped unwanted and useless change on mt7629-eth.c
>
> Changes in v2:
>  - Fixed issues on MT8183 (thanks Chen-Yu!)
>  - Changed builtin_platform_driver() -> module_platform_driver() for
>    MT8167 vdecsys clocks (as that was a mistake!)
>  - Some patches were split, some others were reordered
>  - Summarized: applied changes from Chen-Yu's review
>
> This is part 2 of the "MediaTek clocks cleanups and improvements" series,
> which was already picked.
>
> If reading this full cover letter is too boring for you, here's a short
> summary of the changes of this series:
>  - Added mtk_clk_pdev_probe() for mtk-mmsys probed clocks;
>  - Added divider clock support to common probe mechanism;
>  - Various cleanups here and there;
>  - Converted most clock drivers to platform_driver;
>  - MediaTek clocks can now be built as modules.
>
> NOTE: Applies on top of [1] and [2].
>
>
> Full blurb:
>
> This huge series adds more cleanups on top, reducing size and adding more
> commonization for clock drivers probe/remove, which also includes a new
> common probe mechanism for multimedia clock drivers that are usually
> probed by mtk-mmsys instead of a dt clock node: thanks to this, it was
> finally possible to convert almost all clock drivers to the common probe
> mechanism, which *finally again* makes us able to build all these drivers
> as modules!
>
> Since this looked like being *the* way forward, I went on converting some
> more drivers away from OF_CLK_DECLARE_DRIVER to full platform_driver(s),
> allowing for more (actually, almost all!) drivers to be built as modules.
>
> While at it, I also added some more consistency in macros usage by
> removing all of the duplicated full macro declaration for MediaTek gate
> clocks and replacing all of those with using the GATE_MTK macro instead,
> producing a nice reduction in amount of lines per file but, more
> importantly, improving readability and eventual future batch changes.
>
> This amount of commonization will also, in my opinion, greatly improve
> the review process for new clock drivers, as they will be mostly just a
> list of clocks and won't contain much new code, as it's all going to be
> handled in the common places, which also reduces chances to see new clock
> driver related bugs emerging on one SoC or the other.
>
> Since I don't own devices with all of the supported MediaTek SoCs, I
> could not test some of the conversions on real hardware... but I am
> confident that this will work as the drivers are *very* similar on a
> per-generation basis.
>
> This series was build-tested for all (both module and built-in build)
> and was manually tested on MT6795, MT8173, MT8192, MT8195.
>
> [1]: https://patchwork.kernel.org/project/linux-mediatek/list/?series=719067
> [2]: https://patchwork.kernel.org/project/linux-mediatek/patch/20230207014800.7619-2-moudy.ho@mediatek.com/
>
> AngeloGioacchino Del Regno (54):
>   clk: mediatek: clk-mtk: Switch to device_get_match_data()
>   clk: mediatek: clk-mtk: Introduce clk_mtk_pdev_{probe,remove}()
>   clk: mediatek: Migrate to mtk_clk_pdev_probe() for multimedia clocks
>   clk: mediatek: Add divider clocks to mtk_clk_simple_{probe,remove}()
>   clk: mediatek: mt2712: Migrate topckgen/mcucfg to
>     mtk_clk_simple_probe()
>   clk: mediatek: mt2712: Compress clock arrays entries to 90 columns
>   clk: mediatek: mt2712: Add error handling to
>     clk_mt2712_apmixed_probe()
>   clk: mediatek: mt2712: Move apmixedsys clock driver to its own file
>   clk: mediatek: mt2712-apmixedsys: Add .remove() callback for module
>     build
>   clk: mediatek: mt2712: Change to use module_platform_driver macro
>   clk: mediatek: mt8365: Move apmixedsys clock driver to its own file
>   clk: mediatek: mt8365: Convert simple_gate to mtk_gate clocks
>   clk: mediatek: mt8365: Join top_misc_mux_gates and top_misc_muxes
>     arrays
>   clk: mediatek: mt8365: Convert to mtk_clk_simple_{probe,remove}()
>   clk: mediatek: mt8167: Compress GATE_TOPx macros
>   clk: mediatek: mt8167: Move apmixedsys as platform_driver in new file
>   clk: mediatek: mt8167: Remove __initconst annotation from arrays
>   clk: mediatek: mt8167: Convert to mtk_clk_simple_{probe,remove}()
>   clk: mediatek: mt8183: Move apmixedsys clock driver to its own file
>   clk: mediatek: mt8183: Compress clocks arrays entries where possible
>   clk: mediatek: mt8183: Convert all remaining clocks to common probe
>   clk: mediatek: Consistently use GATE_MTK() macro
>   clk: mediatek: mt7622: Properly use CLK_IS_CRITICAL flag
>   clk: mediatek: mt7622: Move apmixedsys clock driver to its own file
>   clk: mediatek: mt7622-apmixedsys: Add .remove() callback for module
>     build
>   clk: mediatek: mt7622: Move infracfg to clk-mt7622-infracfg.c
>   clk: mediatek: mt7622: Convert to platform driver and simple probe
>   clk: mediatek: mt8516: Move apmixedsys clock driver to its own file
>   clk: mediatek: mt8516: Convert to platform driver and simple probe
>   clk: mediatek: mt8516: Allow building clock drivers as modules
>   clk: mediatek: Propagate struct device with
>     mtk_clk_register_dividers()
>   clk: mediatek: mt7986-apmixed: Use PLL_AO flag to set critical clock
>   clk: mediatek: mt7986-infracfg: Migrate to common probe mechanism
>   clk: mediatek: mt7986-eth: Migrate to common probe mechanism
>   clk: mediatek: mt8186-mcu: Migrate to common probe mechanism
>   clk: mediatek: Switch to module_platform_driver() where possible
>   clk: mediatek: Add MODULE_LICENSE() where missing
>   clk: mediatek: mt2712: Change Kconfig options to allow module build
>   clk: mediatek: Split MT8195 clock drivers and allow module build
>   clk: mediatek: Allow building MT8192 non-critical clocks as modules
>   clk: mediatek: Allow MT7622 clocks to be built as modules
>   clk: mediatek: Allow all MT8167 clocks to be built as modules
>   clk: mediatek: Allow all MT8183 clocks to be built as modules
>   clk: mediatek: Allow building most MT6765 clock drivers as modules
>   clk: mediatek: Allow building most MT6797 clock drivers as modules
>   clk: mediatek: Split configuration options for MT8186 clock drivers
>   clk: mediatek: mt8192: Move apmixedsys clock driver to its own file
>   clk: mediatek: Kconfig: Allow module build for core mt8192 clocks
>   clk: mediatek: Add MODULE_DEVICE_TABLE() where appropriate
>   clk: mediatek: mt8135: Move apmixedsys to its own file
>   clk: mediatek: mt8135: Properly use CLK_IS_CRITICAL flag
>   clk: mediatek: mt8135-apmixedsys: Convert to platform_driver and
>     module
>   clk: mediatek: mt8135: Join root_clk_alias and top_divs arrays
>   clk: mediatek: mt8135: Convert to simple probe and enable module build

Tested-by: Chen-Yu Tsai <wenst@...omium.org>

on MT8183, MT8192, MT8195. Unfortunately on MT8186 the patch series causes
the kernel to freeze during boot. I need to build a verbose kernel to figure
it out.

ChenYu

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ