From mboxrd@z Thu Jan 1 00:00:00 1970 From: CK Hu Subject: Re: [PATCH V6 0/8] make mt7623 clock of hdmi stable Date: Wed, 6 Mar 2019 17:48:24 +0800 Message-ID: <1551865704.1001.3.camel@mtksdaap41> References: <20190225020912.29120-1-wangyan.wang@mediatek.com> <1551837141.23995.13.camel@mtksdaap41> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1551837141.23995.13.camel@mtksdaap41> Sender: linux-kernel-owner@vger.kernel.org To: wangyan wang Cc: Michael Turquette , Stephen Boyd , Matthias Brugger , Philipp Zabel , David Airlie , Daniel Vetter , chunhui dai , Colin Ian King , Sean Wang , Ryder Lee , linux-clk@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-mediatek@lists.infradead.org, dri-devel@lists.freedesktop.org, srv_heupstream@mediatek.com List-Id: dri-devel@lists.freedesktop.org Hi, Wangyan: On Wed, 2019-03-06 at 09:52 +0800, CK Hu wrote: > On Mon, 2019-02-25 at 10:09 +0800, wangyan wang wrote: > > From: Wangyan Wang > > > > V6 adopt maintainer's suggestion. > > Here is the change list between V5 & V6 > > 1. change "unsigned char mux_flags;" to "u8 mux_flags;" to > > match with the struct in " clk: mediatek: add MUX_GATE_FLAGS_2". > > > > Hi, Wangyan: > > I'm not familiar with this clock system, so I still have some question > about it, if you could describe more clear, it would help us to speed up > this review process. > > In [1], I find the clock that dpi and hdmi_phy controls, > > dpi0: dpi@14014000 { > clocks = <&mmsys CLK_MM_DPI1_DIGL>, > <&mmsys CLK_MM_DPI1_ENGINE>, > <&topckgen CLK_TOP_TVDPLL>; > clock-names = "pixel", "engine", "pll"; > }; > > > hdmi_phy: phy@10209100 { > clocks = <&apmixedsys CLK_APMIXED_HDMI_REF>; > clock-names = "pll_ref"; > }; > > In [2], You say that to prevent changing tvdpll would let hdmi stable, > and this clock is controlled by dpi, why do you modify the control flow > in hdmi_phy? If these two have relationship, please describe more clear > because I'm not familiar with this clock system. > As discuss offline, in [3] we can find out that CLK_APMIXED_HDMI_REF is derived from tvdpll, so tvdpll is the parent clock of hdmi_phy_pll. Therefore, this series should modify hdmi_phy flow as well. [3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/clk/mediatek/clk-mt2701.c?h=v5.0#n973 Regards, CK > And I think that patch 'drm/mediatek: using new factor for tvdpll in > MT2701' is the major patch to prevent modifying tvdpll because it reduce > the factor case. Does MT8173 has the same problem? Just ask, I does not > require you to modify MT8173 part. > [1] > https://github.com/frank-w/BPI-R2-4.14/blob/663f7def421952eb49b2d698eadaff12d02622d2/arch/arm/boot/dts/mt7623.dtsi > [2] > http://lists.infradead.org/pipermail/linux-mediatek/2019-February/017693.html > > > Regards, > CK > > > > > > chunhui dai (8): > > drm/mediatek: recalculate hdmi phy clock of MT2701 by querying > > hardware > > drm/mediatek: move the setting of fixed divider > > drm/mediatek: using different flags of clk for HDMI phy > > drm/mediatek: fix the rate and divder of hdmi phy for MT2701 > > clk: mediatek: add MUX_GATE_FLAGS_2 > > clk: mediatek: using CLK_MUX_ROUND_CLOSEST for the clock of dpi1_sel > > drm/mediatek: using new factor for tvdpll in MT2701 > > drm/mediatek: fix the rate of parent for hdmi phy in MT2701 > > > > drivers/clk/mediatek/clk-mt2701.c | 4 +- > > drivers/clk/mediatek/clk-mtk.c | 2 +- > > drivers/clk/mediatek/clk-mtk.h | 20 ++++++--- > > drivers/gpu/drm/mediatek/mtk_dpi.c | 8 ++-- > > drivers/gpu/drm/mediatek/mtk_hdmi_phy.c | 34 ++++------------ > > drivers/gpu/drm/mediatek/mtk_hdmi_phy.h | 7 +--- > > drivers/gpu/drm/mediatek/mtk_mt2701_hdmi_phy.c | 56 +++++++++++++++++++++++--- > > drivers/gpu/drm/mediatek/mtk_mt8173_hdmi_phy.c | 23 +++++++++++ > > 8 files changed, 102 insertions(+), 52 deletions(-) > > >